Committee 12

Urinary Incontinence in the Developing World:The Obstetric Fistula

Chairman

L. L. WALL (USA),

Members

S.D. ARROWSMITH (USA),

N. D. BRIGGS (NIGERIA),

A. LASSEY (GHANA)

893

1. UROLOGIC INJURY

2. GYNECOLOGIC INJURY

3. THE RECTO-VAGINAL FISTULA

4. ORTHOPEDIC TRAUMA

5. NEUROLOGIC INJURY

6. DERMATOLOGIC INJURY

7. SOCIAL CONSEQUENCES OF PROLONGED

OBSTRUCTED LABOR

1. THE FISTULA COMPLICATED BY URETHRAL

DAMAGE

2. URINARY DIVERSION FOR THE IRREPARABLE

FISTULA

REFERENCE

XI. CONCLUSIONS AND RECOMMENDATIONS

X. DEALING WITH THE BACKLOGOF SURGICAL CASES

IX. PREVENTION OF OBSTETRICFISTULAS

VIII. COMPLICATED CASES ANDTECHNICAL SURGICAL

QUESTIONS

VII. SURGICAL TECHNIQUE FORFISTULA CLOSURE

VI. EARLY CARE OF THE FISTULAPATIENT

V. THE CLASSIFICATION OFOBSTETRIC FISTULAS

IV. THE ÇOBSTRUCTED LABORINJURY COMPLEXÈ

III. EPIDEMIOLOGY OF THE OBSTETRIC FISTULA

II. THE RELATIONSHIP OFOBSTETRIC FISTULAS TO MATERNAL MORTALITY

I. LEVELS OF EVIDENCE CONCER-NING OBSTETRIC FISTULAS

894

CONTENTS Chapter 12

People in the developing and the industrialized worldsshare a common humanity and with this comes a com-mon susceptibility to the pathophysiological processesthat can lead to urinary incontinence; however, there isone continence issue that is both unique to and particu-larly prevalent in developing countries: the obstetricfistula. Although once common in Western Europe andthe United States, the obstetric fistula is virtually unk-nown in these regions today. The prevalence of thiscondition has also fallen precipitously in the moreindustrialized nations of Asia and Latin America; butfistulas nonetheless remain prevalent and problematicin Africa and in the less developed regions of Asia andOceania. This fact alone requires special attention, allthe moreso because the obstetric fistula is uniqueamong the causes of urinary incontinence in that it canbe completely and reliably prevented by the provisionof appropriate health care. It is this lack of appropriatehealth careÑ-specifically the lack of appropriate heal-th care for pregnant women in developing countriesÑ-that is responsible for the widespread prevalence of thisdevastating condition in certain areas of the worldtoday. Quite clearly, the obstetric fistula has vanishedfrom industrialized nations due to the creation in thosecountries of efficient and effective systems of materni-ty care which provide prompt emergency obstetricalservices to women when complications arise in labor.The problem of obstetric fistulas in the developingworld will not be solved until similarly effective sys-tems of maternal health care are created there. Unfortu-nately, ÒSafe MotherhoodÓ has largely become anÒorphanÓ initiative (Rosenfield and Maine 1985; Gra-ham 1998, Weil and Fernandez 1999). In virtually noother area in which health statistics are commonly col-lected is the disparity between the industrialized and thedeveloping worlds so great as in the area of maternalhealth (AbouZahr and Royston 1991). This situationremains one of the most glaring, and one of the mostneglected, issues of international social injustice in theworld today. For this reason, the committee has chosen

ÇIn vast areas of the world, in South East Asia, inBurma, in India, in parts of Central America, SouthAmerica and Africa 50 million women will bringforth their children this year in sorrow, as inancient Biblical times, and exposed to grave dan-gers. In consequence, today as ever in the past,uncounted hundreds of thousands of youngmothers annually suffer childbirth injuries; inju-ries which reduce them to the ultimate state ofhuman wretchedness.

Consider these young women. Belonging generallyto the age group 15-23 years, and thus at the verybeginning of their reproductive lives, they are moreto be pitied even than the blind, for the blind cansometimes work and marry. Their desolation des-cends below that of the lepers, who though scarred,crippled and shunned, may still marry and find use-ful work to do. The blind, the crippled and thelepers, with lesions obvious to the eye and thereforeappealing to the heart, are all remembered andcared for by great charitable bodies, national andinternational.

Constantly in pain, incontinent of urine or faeces,bearing a heavy burden of sadness in discoveringtheir child stillborn, ashamed of a rank personaloffensiveness, abandoned therefore by their hus-bands, outcasts of society, unemployable except inthe fields, they live, they exist, without friends andwithout hope.

Because their injuries are pudendal, affecting thoseparts of the body which must be hidden from viewand which a woman may not in modesty easilyspeak, they endure their injuries in silent shame.No charitable organization becomes aware of them.Their misery is utter, lonely, and complete.È

RHJ Hamlin and E. Catherine Nicholson, 1966

895

Urinary Incontinence in the Developing World:The Obstetric Fistula

L. L. WALL

S.D. ARROWSMITH, N. D. BRIGGS, A. LASSEY1

1. The committee would like to thank Dr. Jonathan Karshima andDr. Gunter Rienhart, who served as consultants to the committee

to focus its discussion of incontinence in the developingworld on the obstetric fistula as the first and most pres-sing continence priority, rather than attempt to recapi-tulate our understanding of all forms of incontinencewithin the particular context of non-industrializedcountries.

This chapter provides an overview of the pathophysio-logy of obstetric fistula formation and discusses therelationship of this condition to the broader issue ofmaternal mortality with which it is intimately linked.This chapter also summarizes the most important issuesin the treatment of obstetric fistulas, and suggests direc-tions for much-needed future research in this area.

Scientific data on the problem of obstetric fistulas, theirprevalence and forms of treatment, are limited by unu-sual historical circumstances. Prior to the middle of the19th Century, an obstetric vesicovaginal fistula wasgenerally regarded as an incurable and hopeless condi-tion. It was only after the work of the American surgeonJ. Marion Sims and his colleague and successor Tho-mas Addis Emmett (Sims 1852, Emmett 1868, Harris1950, Zacharin 1988) that surgical cure of this condi-tion could be undertaken with confidence. As obstetricsdeveloped as a scientific specialty in the first half of the20th Century, maternal mortality underwent a precipi-tous decline throughout Europe, the United States, andother developed nations (Loudon 1992a, 1992b). Thismeant that the obstetric fistula vanished from the clini-cal and social experience of the West just as it was star-ting to become rigorously scientific. As a result of thesehistorical circumstances, the Western medical literatureon obstetric fistulas is old and generally uncritical bycurrent scientific criteria. This literature consists almostentirely of anecdotes, case series (some quite large),and personal experiences reported by dedicated sur-geons who have labored in remote corners of the worldwhile facing enormous clinical challenges with scantyor absent resources at their disposal (Evidence Levels 4and 5). The committee charged with producing thisreport was able to locate only a handful of articles, allquite recent, that rise to a higher level of evidence. Forexample, there appears to be only one prospective, ran-domized clinical trial in the literature on vesico-vaginalfistulas in the developing world (Tomlinson and Thorn-ton 1998), and only one comparative study of surgicaltechnique (Rangnekar et. al. 2000). This fact underlineshow the entire problem of obstetric fistulas in develo-ping countries has been neglected by the bioscientificmedical community of the industrialized world. Thepaucity of well-designed studies makes it impossible to

produce a sophisticated meta-analysis of hard scientificdata on the fistula problem at this time. While acknow-ledging this difficulty, the members of the committeefeel that it is still possible to arrive at a consensus viewof the broad outlines of the fistula problem. This reportis therefore a general summary of the current state ofour knowledge regarding obstetric fistulas in the deve-loping world and the challenges thus presented. Itshould be regarded as a point of departure for furtherworkÑ-and further actionÑ- while recognizing thatmany important issues remain unclear and urgentlyrequire more intensive scientific study.

The commonly accepted definition of a maternal deathis Çthe death of a woman while pregnant or within 42days of termination of pregnancy, irrespective of theduration and the site of the pregnancy, from any causerelated to or aggravated by the pregnancy or its mana-gement but not from accidental or incidental causesÈ(Abour Zahr and Royston 1991). The most commonmeasure of maternal mortality used for internationalcomparative purposes is the maternal mortality ratio:the number of maternal deaths per 100,000 live births.The available statistics show huge discrepancies bet-ween the developed and the developing worlds (WHO1996). The overall world maternal mortality ratio isestimated at 430 maternal deaths per 100,000 livebirths. In more developed regions of the world the ratiois 27 deaths per 100,000 live births, contrasted with 480deaths per 100,000 live births in less developed regions.The numbers are substantially worse for Africa: 870 forthe continent as a whole, 950 in middle Africa, 1,020 inWest Africa, and 1,060 in East Africa. In northernEurope and North America there are 11 maternal deathsper 100,000 live births. There are many problems asso-ciated with the collection of maternal mortality statis-tics, especially in developing countries, and all suchstatistics are acknowledged to be underestimates tosome (usually to a substantial) degree. For an indivi-dual woman, a more important statistic than the mater-nal mortality ratio is her lifetime risk of pregnancy-rela-ted death. This statistic is a function of the risk of dyingin any particular pregnancy multiplied by the number oftimes she is likely to become pregnant. The risks aretherefore highest in areas of high fertility where accessto emergency obstetric care is poor. Overall, the globallifetime risk of maternal death is 1 in 60. In more deve-loped regions, the risk is only 1 in 1800; in less develo-ped regions the risk is 1 in 48. In North America orNorthern Europe, a woman has a lifetime risk of pre-gnancy-related death of approximately 1 in 4,000; in

II. THE RELATIONSHIP OFOBSTETRIC FISTULAS TO MATERNAL MORTALITY

I. LEVELS OF EVIDENCE CONCER-NING OBSTETRIC FISTULAS

896

Africa, the risk is 1 in 16, and in the poorest parts ofAfrica a womanÕs lifetime risk of dying as the result ofpregnancy or childbirth is as high as 1 in 7. The cur-rently accepted world maternal mortality statistics aresummarized in Table 1 (WHO 1996).

The majority of maternal deaths are due to five princi-pal causes: hemorrhage, sepsis, hypertensive disordersof pregnancy, unsafe abortion, and obstructed labor(AbouZahr and Royston 1991). The vast majority offistulas are due to obstructed labor. Not surprisingly,obstetric fistulas are most prevalent in areas wherematernal mortality is high and where obstructed labor isa major contributor to maternal deaths. These are areaswhere access to emergency obstetric care is poor; cor-respondingly, accurate epidemiological information isalso poor in these regionsÑ-a continuing point of diffi-culty in the evaluation of maternal mortality in generaland in the evaluation of obstetric fistulas in particular.

The problem of obstetric fistula formation is linkeddirectly to that of maternal mortality. Maternal mortali-

ty is embedded in a complex network of social issuesthat have to do with the social status of women, the dis-tribution and availability of healthcare resources, per-ceptions about the nature and importance of maternalhealth problems, and the social, economic and politicalinfrastructures of developing countries. Indeed, it iscommonly said that obstetric fistulas result from thecombination of Çobstructed labor and obstructed trans-portation.È Thaddeus and Maine (1994) have articula-ted the concept of three Çstages of delayÈ that result inmaternal mortality: delay in deciding to seek care,delay in arriving at a health care facility, and delay inreceiving adequate care once a woman arrives at such afacility. All of these factors are present in the formationof an obstetric fistula. Women in labor are often neglec-ted in the hopes that Çeverything will come out allrightÈ on its own. Other women refuse to seek care forfear they will be perceived as ÒweakÓ or Òcowardly.ÓFrequently the seriousness of the situation is not appre-ciated or help is not sought for fear of incurring highfinancial costs. Even if help is sought, poor roads and

897

Table 1 : Revised estimates of maternal mortality by United Nations regions (WHO 1996)

Maternal Mortality Number of Maternal Lifetime Risk ofRatio (Maternal deaths Deaths Maternal Death,per 100,000 live births) 1 in:

World total 430 585,000 60More developed regions* 27 4,000 1,800Less developed regions 480 582,000 48

Africa 870 235,000 16Eastern Africa 1,060 97,000 12Middle Africa 950 31,000 14Northern Africa 340 16,000 55Southern Africa 260 3,600 75Western Africa 1,020 87,000 12

Asia* 390 323,000 65Eastern Asia 95 24,000 410South-central Asia 560 227,000 35South-eastern Asia 440 56,000 55Western Asia 320 16,000 55

Europe 36 3,200 1,400Eastern Europe 62 2,500 730Northern Europe 11 140 4,000Southern Europe 14 220 4,000Western Europe 17 350 3200

Latin America & the Caribbean 190 23,000 130Caribbean 400 3,200 75Central America 140 4,700 170South America 200 15,000 140

Northern America 11 500 3,700

Oceania* 680 1,400 26Australia-New Zealand 10 40 3,600Melanesia 810 1,400 21

*Australia, New Zealand and Japan have been excluded from the regional totals but are included in the total for developedcountries. Figures may not add to total due to rounding.

inadequate public transportation often result in longdelays in reaching a health care facility, and even if thelaboring woman arrives at a hospital, inadequate facili-ties or untrained personnel may preclude the perfor-mance of an emergency cesarean section or meeting herother medical needs.

The three Òstages of delayÓ keep women on what issometimes termed the Òroad to maternal death.Ó Insimilar fashion it seems that there is also a Çroad to obs-tetric fistulaÈ that begins when young girls grow up innutritionally marginal circumstances, are marriedaround the age of menarche, become pregnant whilestill adolescents, and labor at home either alone orunder the care of untrained birth attendants for prolon-ged periods of time and with inadequate access to emer-gency obstetrical care. In addition, many become vic-tims of harmful traditional medical practices that fur-ther complicate matters. The obstetric fistula pathwayis summarized in Figure 1.

In industrialized countries most vesico-vaginal fistulas(VVF) are the result of radiation therapy or surgery(Latzko 1942; Everett and Mattingly 1956; Counsellorand Haigler 1956; Radman 1961; Moir 1966; Masseeet. al. 1964; Weed 1967; Taylor and Droegemueller1967; Goodwin and Scardino 1980; Enzelsberger andGitsch 1991; Langkilde et. al. 1999). In the developingworld, on the other hand, most fistulas occur from theneglect of obstetric complications. As such, they occurunder very different circumstances. As Gharoro andAbedi remarked in their short case series of fistulasfrom Benin City, Nigeria (1999), ÇAny pregnant patientcould develop VVF from substandard care.È

To date there has never been a comprehensive world-wide survey designed to determine precisely whereobstetric fistulas occur. Questions regarding the inci-dence and prevalence of obstetric fistulas have neverbeen included on the standardized demographic andhealth surveys (DHS) that are carried out to evaluatepopulation characteristics and overall health status indeveloping countries. Virtually no population-basedsurveys have been carried out in countries where thereappears to be a high incidence and high prevalence ofobstetric fistulas. Furthermore, the urinary and/or fecalincontinence that accompanies fistula formation makesits sufferers social outcasts, pushing them to the mar-gins of society where they are ignored, further obscu-ring the true extent of the problem. There is a very realneed for ongoing scientific research in this area. Untilsuch work is forthcoming, we are left with only indirect

means of describing the epidemic. In a short survey ofavailable information on obstetric fistulas published bythe World Health Organization in 1991 that encompas-sed a literature review and correspondence with over250 individuals, institutions and organizations in deve-loping countries, a map was created showing the distri-bution of countries where obstetric fistulas had beenreported (Figure 2). The committee members, from per-sonal experience and contact with other workers in thefield, know that this map should include virtually all ofAfrica and south Asia, the less developed parts of Ocea-nia, Latin America, and the Middle East; and, we sus-pect (though we cannot prove) the more remote regionsof Central Asia and selected isolated areas of the formerSoviet Union and Soviet-dominated eastern Europe.

The true magnitude of the fistula problem worldwide isunknown, but it is clearly enormous. The situation inNigeria may be cited as an example. Arrowsmith(1994), writing from the plateau region of central Nige-ria, noted that Çthe local popular press estimates thatthe region may harbor up to 150,000 victims of vesico-vaginal fistula.È Harrison, also writing from northernNigeria, reported a vesico-vaginal fistula rate of 350cases per 100,000 deliveries at a university teachinghospital (1985). Karshima, who has carried out village-based survey work on obstetric fistulas in the middlebelt of Nigeria, suspects that there may be as many as400,000 unrepaired fistulas in Nigeria (J. Karshima,personal communication, 2001), and the NigerianFederal Minister for Women Affairs and Youth Deve-lopment, Hajiya Aisha M.S. Ismail, has estimated thenumber of unrepaired vesico-vaginal fistulas in Nigeriaat between 800,000 and 1,000,000 (personal communi-cation, 2001).

The data on maternal morbidity (non-fatal obstetriccomplications) in developing countries are poor, but itis obvious that the number of serious morbid episodesor Çnear missesÈ greatly exceeds the number of mater-nal deaths in the developing nations (Prual et. al.,1998). In parts of the world where a womanÕs lifetimerisk of maternal death is high, a womanÕs lifetime riskof suffering serious maternal morbidity (including obs-tetric fistula) may be extraordinarily high. In one of thefew studies that has looked at the issue of maternalmorbidity, Fortney and Smith calculated the ratios ofserious morbidities to maternal mortalities in Indonesia,Bangladesh, India and Egypt. For each maternal death,they calculated that there were 149, 259, 300 and 591serious morbidities in these respective countries, and112, 114, 24, and 67 life-threatening morbidities res-pectively (Fortney and Smith, 1996). In the face of suchsobering statistics, Danso and colleagues suggested thatin regions of the world where obstructed labor is amajor contributor to maternal mortality, the obstetricfistula rate may approach the maternal mortality rate

III. EPIDEMIOLOGY OF THE OBSTETRIC FISTULA

898

899

Figure 1 : The obstetric fistula pathway: Origins and consequences (©Worldwide Fund for Mothers Injured in Childbirth,used by permission).

Figure 2 : Countries from which obstetric vesico-vaginal fistulas have been reported (WHO 1991). The prevalence isactually greater than this map indicates.

Low socio-economic status of women

Childbearing before pelvic growth is complete

Limited social roles

Early marriage

Obstructed labor

Stigmatization Isolation and loss of social support

Divorce or separation Worsening poverty

Worsening malnutrition Suffering, illness, and premature death

" Obstructed Labor Injury Complex "Fetal Death

Fistula formation Complex urologic injury

Vaginal scarring and stenosis Secondary infertility

Musculoskeletal injury Footdrup

Chronic skin irritationOffensive odor

Harmful traditional practices

Cephalopelvic disproportion

Lack of emergency obstetric services

Relatively large fetus ormalpresentation

Malnutrition

Fecal incontinence Urinary incontinence

Illiteracy and lack of formal education

(Danso et. al., 1996). Other authors have reached simi-lar conclusions (Hilton and Ward, 1998).

In the absence of good population-based epidemiologi-cal studies, most information on fistulas has come fromlarge series of patients seen at teaching hospitals or atdedicated fistula centers. Just as there are reasons to bewary of hospital-based statistics in surveys of maternalmortality (AbouZahr and Royston 1991), there are rea-sons to be cautious about considering hospital-basedsurveys of fistula patients to be representative of theover-all population of women with fistulas from obs-tructed labor. Answers to the following questionsappear to be crucial in delineating the true extent of thefistula problem worldwide:

¥ Where do fistula patients come from?

¥ What is their social background?

¥ What is their educational level?

¥ What is the age at marriage of women who develop fistulas?

¥ What is the distribution of fistulas by age and parityat the time of their occurrence?

¥ In which pregnancy do obstetric fistulas occur?

¥ What proportion of women who develop fistulas hadaccess to obstetric care, and if such care was avai-lable, why was it not used in a timely fashion?

¥ Where do women who develop fistulas deliver andwho attends those deliveries?

¥ How long are these women in labor before they reco-gnize that labor is obstructed and seek help, and howlong is the delay before effective help is obtained?

¥ What is their marital/social status at the time oflabor?

¥ How long must women who develop fistulas waitbefore they undergo an attempt at surgical repair,and at what age do they present for repair?

¥ What is the social status of fistula victims after theydevelop this problem?

¥ Does this status change after fistula closure?

¥ What role do harmful traditional beliefs and prac-tices play in the genesis of prolonged obstructedlabor and fistula formation?

Although detailed answers to these questions awaitdetailed population-based research, preliminaryimpressions can be obtained from surveys taken of fis-tula patients in hospital settings in several countries.The overall impression is that fistula patients comefrom poor rural areas where infrastructure developmentis rudimentary and access to health careÑparticularly

access to basic midwifery and emergency obstetric ser-vicesÑis lacking. Fistula patients tend to be youngwomen, many of whom married very early, of short sta-ture, poorly educated, married to farmers or petty tra-ders who themselves have little or no formal education.They typically have had little or no access to prenatalcare, and even if they have had access to antenatalscreening, they have often nonetheless delivered athome attended by family members or traditional mid-wives. If they have sought help from trained midwivesor medical doctors, this often occurs late in labor afterserious complications have already set in.

Several papers from Africa support this picture of thefistula patient. Kelly and Kwast (1993b) reviewed a10% sample of fistula patients seen at the Addis AbabaFistula hospital between 1983 and 1988, by reviewingthe records of every tenth patient. The mean age ofthese women was 22.4 years (range 9 Ð 45 years); 42%were less than 20 years of age, and 65% were less than25 years of age. Of the women surveyed, 52.3% hadbeen deserted by their husbands and 21.5% had to liveby begging for food. Nearly 30% had delivered alone,and the mean duration of labor was 3.9 days (range 1-6days).

In one review of 150 fistulas from Ghana, 91.5% werethe result of obstructed labor and 8.5% were the resultof complications of difficult gynecological surgery(Danso et. al. 1996). Nearly 53% of the obstetric fistu-la patients were under the age of 25 and nearly 43% ofpatients were primigravid. Interestingly, nearly 25% ofthe patients had a parity of 5 or more, indicating thatlabor can become obstructed even in women who havepreviously delivered vaginally. This probably repre-sents the tendency for birth weights to increase withsuccessive pregnancies, as well as the effects of agingon changes in pelvic anatomy.

Information from several studies in northern Nigeria issimilar. Tahzib reviewed records of 1,443 patients withvesico-vaginal fistulas seen in the clinics at AhmaduBello University in Zaria between 1969 and 1980. Deli-very occurred at home in 64.4% of these women.Among women who developed fistulas, 54.8% wereunder the age of 20 and only 22.7% were older than 25years. The majority of women developed a fistula intheir first pregnancy (52%), and 21.5 % of fistulapatients were parity 4 or greater. Two case control stu-dies from northern Nigeria have found fistula patientsto be shorter, to be of lower socioeconomic status, andto have less education than control patients without fis-tulas (Ampofo, Otu, and Uchebo 1990; Onolemhemhenand Ekwempu 1999).

Virtually every paper on vesico-vaginal fistulas fromdeveloping countries demonstrates that the most com-mon cause of fistula formation by far is prolonged obs-

900

tructed labor (Amr 1998; Arrowsmith 1994; Azia 1965;Bird 1967; Coetzee and Lithgow 1966; Gharoro andAbedi 1999; Goh 1998; Hilton and Ward 1998; Iloaba-chie 1992; Kelly 1992; Lavery 1955; Krishnan 1949;Lawson 1972; Mahfouz 1957; Mustafa and Rushwan1971; Naidu 1962; Tahzib 1983; Tahzib 1985; Naidu1962; Waaldijk 1989; Yenen and Babuna 1965) accoun-ting for between 76% and 97% of fistulas in most largeseries. Additional important causes include injuries sus-tained during complicated gynecologic surgery perfor-med under very difficult circumstances, accidents suchas penetrating injuries to the vagina and bladder invol-ving cattle horns or impalement by falling on a stick,infection (particularly lymphogranuloma venereum, butalso diphtheria, measles, schistosomiasis, infected cen-tipede bites, etc.), and cervical cancer. In their series of309 urinary fistulas treated in Pietermaritzburg, Natal,South Africa between 1954 and 1963, Coetzee andLithgow (1966) reported 248 as resulting from obstruc-ted labor (80%), 9 (3%) from complications of pelvicsurgery, and 52 (17%) from advanced carcinoma of thecervix. Particularly troubling are fistulas caused bysexual abuse, rape, or from attempts forcibly to enlargethe vaginal introitus of child brides so that sexual rela-tions can commence (Tahzib 1985; Muleta andWilliams 1999).

In some parts of the world, harmful traditional practicesare also responsible for fistula formation. For example,in northern Nigeria, a harmful traditional practice cal-led gishiri-cutting is responsible for fistula formation in6 Ð 13% of cases (Ampofo, Otu and Uchebo 1990; Tah-zib 1983, 1985). Gishiri is the Hausa word for Çsalt.È Itis often used to refer to the encrustations of salt thatoccur on the outsides of porous water-jars as theircontents evaporate. Gishiri is an important ethnomedi-cal condition in Hausa traditional medicine. The beliefis that an imbalance of salty or sweet foodstuffs cancause a ÇfilmÈ to grow over the womanÕs vagina, cau-sing a variety of gynecological complaints, the mostimportant of these being difficult labor (Wall 1988).When this diagnosis is made, surgical treatment is oftenundertaken. A midwife or barber is summoned. A sharpobject such as a knife, razor blade, or piece of brokenglass is inserted into the vagina, and a series of randomcuts is made to alleviate the postulated obstruction andÇopen the wayÈ for the baby to come out. Serious infec-tion, life-threatening hemorrhage, and fistulas frequent-ly result from this practice. The gishiri fistula typicallypresents as a direct longitudinal slit in the bladder neckand urethra, occasionally presenting as a similar poste-rior injury affecting the rectum.

Another traditional practice that have been reported toproduce fistulas is the insertion of caustic substancesinto the vagina, either as part of a traditional herbalremedy for a gynecologic condition (Lawson 1968) or

as part of traditional puerperal practices to ÇhelpÈ thevagina return to its nulliparous state. The latter practiceis a part of the traditional folk medicine of several Arabcountries (Kingston 1957, El Guindi 1962, Frith 1960).The practice was summarized by Betty Underhill in her1964 report of 65 such cases seen at the BahrainGovernment Hospital between 1957 and 1963 thus:

ÇFor hundreds of years it has been the custom of Arabwomen to pack the vagina with salt for the first weekafter delivery. This is popularly supposed to restore thevagina to its nulliparous state and to add to the hus-bandÕs sexual pleasure. Since the practice has survivedfor so long it must be presumed that this desirable effectis sometimes achieved. It is certainly true, however, thatin a great many cases the end results are devastatingand it is these unhappy women who are driven to hos-pital for relief. É The substance used is crude rock saltsold in the bazaar. A piece of salt roughly the size andshape of an egg is pushed into the vagina daily for 2 to15 days after the delivery. An additional incentive to theuse of salt is its supposed antiseptic property, and somewomen are said to employ it between pregnancies as acontraceptive.

The immediate effect of the pessary on the hyperaemicvagina is the production of severe inflammation whichgoes on to ulceration. Healing of the ulcers leads to avery severe fibrosis and the vagina becomes partly orcompletely occluded by a substance with an almost car-tilaginous texture. The walls of the upper third of thevagina are held apart by the cervix and the lower thirdis constantly pulled upon by the levator ani muscle. Thelax middle third is commonly the area affected by fibro-sis and the patient is left with a tiny lower vagina 2 to5 cm long, then a block or cord of fibrous tissue with anupper chamber indented by the cervix.È

The fibrosis that results is extensive enough to produceobstructed labor in many cases (Fahmy 1962). Fistulasmay result from either obstructed labor (Fahmy 1962)or by direct chemical action. As Naim has written, ÇThemethod of salt packing determines the site and types ofthe fistulae that may develop. Thus in Kuwait, wherethe whole vagina is packed with rock salt, combinedrectovaginal and vesicovaginal fistulae develop due torock salt being in contact with both the anterior andposterior vaginal walls. In Saudi Arabia, where piecesof rock salt are inserted high in the vagina, rectovaginalfistulae only were met with.È In some cases the vaginais completely occluded, leading to the formation of ahematocolpos, dysmenorrhea, infertility and relatedgynecological conditions (Underhill 1964).

Much popular concern has focused on the problem ofÇfemale circumcisionÈ (Çfemale genital mutilationÈ) inrecent years (for example, Abdalla 1982; Aziz 1980;Boddy 1982; El-Dareer 1982; Gruenbaum 2001; Tou-

901

bia 1994; WHO 1998). While these practices shouldrightly be condemned by the medical community asbeing markedly injurious to the health of women andfemale children, there is absolutely no evidence in theworld literature to suggest that practices of this type arethe major cause of fistula formation in developingcountries. Although fistulas are common in communi-ties where various forms of female genital mutilationare practiced, the former are not usually caused direct-ly by the latter. Rather, these two phenomena bothreflect the low social status of women in these coun-tries, which in turn is reflected by a striking under-investment in womenÕs reproductive health care inthose countries, with a consequent lack of access toemergency obstetric care and poor gynecologic health.

Where female genital mutilation is practiced, it contri-butes to the genesis of fistulas in two ways. First, as inthe case of gishiri-cutting among the Hausa, surgicalattempts to modify the female genitalia in any fashioncan cause direct injury to the bladder or urethra. Whensuch procedures are carried out without anestheticunder unsterile conditions by unskilled practitionerswho do not understand female pelvic anatomy, the risksof catastrophic complications are great. Fistula forma-tion may occur in these cases through direct surgicaltrauma to the urinary tract. Secondly, in cases wherelarge amounts of vulvar or vaginal tissue are removed,particularly in the case of infibulation (ÇPharaonic cir-cumcisionÈ) where the vaginal introitus is sewn almostcompletely shut or where excessive vulvar scarring hasoccurred as a result of these practices, the vaginal out-let may become extremely narrow and constricted bydense scar tissue. This may prevent expulsion of thefetus at the end of the second stage of labor; indeed, ananterior episiotomy is usually required in order to effectvaginal delivery in such women. If the outlet cannot beopened sufficiently to allow passage of the fetus, obs-tructed labor may result and may end in fistula forma-tion. However, obstructed labor from this cause wouldonly occur at the end of labor.

The available evidence suggests that in the majority ofcases in which fistulas occur, labor becomes obstructedin the mid-pelvis or at the pelvic brim, rather than at thepelvic outlet. Further evidence for the primary impor-tance of obstetric factors in the development of vesico-vaginal fistulas can be seen in the fact that the Hausa,who have one of the highest prevalences of obstetricfistula formation in the world, do not practice Çfemalecircumcision.È Comparative radiographic studies sug-gest that the Nigerian pelvis is considerably smallerthan that of European (Welsh) women, particularly atthe pelvic inlet, least significantly at the pelvic outlet(Kolawole, Adamu and Evans 1978). This suggests thatthere are anatomic differences that predispose Africanwomen to obstructed labor (Briggs 1983). Because

sexual maturity and adult height are reached beforegrowth of the pelvic dimensions has been completed,pregnancies that occur in early adolescence are morelikely to be complicated by obstructed labor than thosethat occur in older women (Moerman 1982). Early mar-riage and accompanying early pregnancy are also lin-ked to a relatively low social status of women, tyingthese factors together in one more way.

Obstructed labor occurs when the presenting fetal partcannot pass through the maternal bony pelvis. The pre-senting part then becomes wedged against the maternalpelvic bones, compressing the soft tissues in between(Figure 3). The uterine contractions force the presen-ting part deeper into the pelvis, compressing the mater-nal soft tissues more forcibly. If this process is not relie-ved by surgical intervention, the blood supply to theentrapped soft tissues becomes compromised, ultimate-ly resulting in tissue death and fistula formation.

The pathophysiology of obstructed labor was describedclearly and succinctly by Mahfouz in 1930:

IV. THE ÇOBSTRUCTED LABORINJURY COMPLEXÈ

902

Figure 3 : Obstructed labor from absolute cephalo-pelvicdisproportion, from William Smellie's Sett of AnatomicalTables with Explanations, and an Abridgment of the Practi-ce of Midwifery. London, 1754

ÇThe process by which a fistula develops after labour isthe following. When labour becomes difficult, onaccount of disproportion between the pelvis and pre-senting part, or when the presentation is abnormal, theuterine contractions increase in strength and endea-vour to force the presenting part through the brim. Themembranes protrude unduly in the vagina, and prema-ture rupture occurs. In consequence of early ruptureand disproportion the full force of the uterine contrac-tions is directly exerted upon the foetus and the presen-ting part is forced against the brim of the pelvis or getstightly impacted therein. The vesico-vaginal septum,and the cervix if the latter is not dilated, will be tightlycompressed against the back of the symphysis pubis.The uterus in such cases usually passes into a state oftonic contractions which prevents any remission in thepressure exerted on the soft parts. As a result of thecontinued pressure the tissues undergo necrosis andslough away. The duration of compression in suchcases is usually very long, but I have seen cases inwhich a fistula developed after 3 hours of compressiononly. At about the fifth day of the puerperium the sloughbegins to separate and urine dribbles involuntarily intothe vagina. É

The slough that develops from this pressure necrosismost commonly results in a vesico-vaginal fistula(Figure 4). Once this has occurred, the unfortunatewoman suffers from constant, unremitting loss of urine.She can attempt to stay this flow with the use of rags orcloth, but she can never get rid of it. J. Marion Sims,who first developed a consistently successful method offistula closure, expressed well the situation of thepatient with a fistula (1852):

ÇIts diagnosis is sufficiently easy. Incontinence of urine,following a tedious labour after a lapse of from one tofifteen days, will always prove its existence. But to

determine the exact size, shape, and relative position ofthe artificial opening require some nicety of examina-tion. The consequences of the involuntary discharge ofurine are indeed painful. The vagina may become infla-med, ulcerated, encrusted with urinary calculi, andeven contracted; while the vulva, nates, and thighs aremore or less excoriated, being often covered with pus-tules having a great resemblance to those produced bytartar emetic. These pustules sometimes degenerateinto sloughs, causing loss of substance, and requiring along time to heal. The clothes and bedding of the unfor-tunate patient are constantly saturated with thedischarge, thus exhaling a disagreeable effluvium, alikedisgusting to herself and repulsive to others.

The accident, per se, is never fatal; but it may well beimagined that a lady of keen sensibilities so afflicted,and excluded from all social enjoyment, would preferdeath. A case of this kind came under my observation afew years since, where the lady absolutely pined awayand died, in consequence of her extreme mortificationon ascertaining that she was hopelessly incurable.È

The location and nature of the maternal injury thatresults from prolonged obstructed labor is a function ofthe force and duration of the compression that occurs,as well as the level at which labor becomes obstructed(Figure 5). As Mahfouz remarked (1930): ÒThe situa-tion of the fistula depends to a great extent on the stateof the cervix, when impaction and compression occur,and also on the plane of impaction. If pressure and com-pression occur before the cervix is pulled up over thehead, the vault of the vagina and the cervical tissuesmay be involved in the slough. The resulting fistula willbe vesico-cervico-vaginal, or uretero-vaginal, as thecase may be.Ó The location of the fistula is often deter-mined by the particular configuration of the womanÕspelvis. Thus, John St. George noted (1969):

903

Figure 4 : Typical mid-vaginal vesico-vaginal fistula fromobstructed labor. A metal sound passed through the urethracan clearly be seen inside the bladder. (©Worldwide Fundfor Mothers Injured in Childbirth, used by permission).

Figure 5 : Types of genito-urinary fistulas. The locationdepends on the point at which labor becomes obstructed. A.Utero-vesical fistula. B. Cervico-vesical fistula. C. Mid-vaginal vesico-vaginal fistula. D. Vesico-vaginal fistulainvolving the bladder neck. E. Urethro-vaginal fistula.From Elkins (1994).

ÇVesico-vaginal fistulae occurred more often in primi-gravidae (often very young) than in multiparae. Deeptransverse arrest of the fetal head was commoner inprimiparae with an android type of pelvis, and therefo-re the site of the fistula was more often at the bladderneck or was juxta-urethral. Primiparae were also per-mitted longer labour by their folk because of theiryouth, and hence their fistulae were extensive. On theother hand, labour in multigravidae was often obstruc-ted at the inlet because of a secondary flat pelvis; inthese cases mid-vaginal or juxta-cervical fistulae,which were more amenable to surgery, were more com-mon.È

Many women who develop an obstetric vesico-vaginalfistula have been delivered vaginally with the use ofinstruments or have had an abdominal delivery byCesarean section. As early as the 19th Century, obste-tricians were often blamed for creating fistulas throughthe injudicious use of operative techniques. Althoughincompetent surgical delivery is undoubtedly respon-sible for some fistulas, the vast majority of patientsundergoing operative delivery have been in labor for aprolonged period of time and have already suffered theischemic insult that ultimately leads to fistula forma-tion. The mode of delivery is therefore often unrelatedto the development of the fistula that subsequentlyoccurs (Emmett 1879). As Das and Sengupta noted intheir series from India (1969):

ÇPressure necrosis is responsible to a great extent forthe fistula. Most of the instrumental labour group werealso prolonged labour cases and would probably haveresulted in vesico-vaginal fistula even without instru-mentation. Therefore, all cases were not caused by ins-trumental delivery as such but were due to pressurenecrosis that had already existed.È

The prevalence of women who have had prolongedlabors and then undergone operative delivery is in parta function of increasing access to obstetric care, even ifthat access is not timely enough to prevent the develop-ment of a fistula. In his review of vesico-vaginal fistu-las in Jordan, M.F. Amr (1998) summarized the rela-tionship between womenÕs health care and fistula for-mation with these words: ÇThe incidence of urinary fis-tula reflects the standard of obstetric and gynaecologi-cal care and its availability to the population. Most fis-tulae are due to obstetric causes such as prolonged andobstructed labour, difficult instrumental delivery andother obstetric manipulations, and ruptured uterus.È

The vast majority of women in non-industrialized coun-tries who develop a vesico-vaginal fistula do so as theresult of prolonged obstructed labor from cephalo-pel-vic disproportion. The presenting fetal part is wedgedinto the pelvis, trapping the womanÕs soft tissues bet-ween two bony plates which effectively shut off the

blood supply to the affected tissues. This results inextensive tissue necrosis which frequently destroys thevesico-vaginal septum and results in fistula formation;however, it is important to realize that obstructed laborproduces a broad-spectrum Òfield injuryÓ that mayaffect many parts of the pelvis. The damage that occursis not limited to vesico-vaginal or recto-vaginal fistulaformation alone. This fact is not generally appreciatedby most Western doctors. Although the fistula thatresults is usually the dominant clinical injury, it iswrong to focus solely on the Çhole in the bladderÈ tothe exclusion of the other consequences of obstructedlabor. Women who have sustained an obstetric fistulaare usually injured in multiple ways, all of whichimpact their lives and well-being, and all of which mustbe considered in their care. Arrowsmith, Hamlin andWall (1996) have called this spectrum of injury theÇobstructed labor injury complexÈ (Table 2). Theunderstanding that one must treat the Òwhole personÓwith a fistulaÑ-and not just her injured bladder or rec-tumÑ-is the single most important concept in fistulacare. Doing this effectively requires some understan-ding of the multi-system consequences of prolongedobstructed labor.

1. UROLOGIC INJURY

Obstetric vesicovaginal fistulas are not caused by tea-ring or laceration of the bladder; rather they result fromischemic injury. Normal tissue perfusion is disrupted bycompression of the soft tissues by the fetal head. Thisleads to ischemia, tissue death, subsequent necrosis,and fistula formation, along with varying degrees ofinjury to surrounding tissues which, although they havenot died, nonetheless are often not completely healthyeither, having suffering a non-fatal but debilitatingischemic vascular injury as well. This process has animpact at various levels throughout the urinary tract.

a) Bladder. The most familiar injury from obstructedlabor is the vesicovaginal fistula. The loss of bladdertissue from pelvic ischemia during obstructed laboraffects both the technique needed for, as well as thefunctional outcome of, fistula repair. Loss of bladdertissue is one of the main reasons why obstetric fistularepair is technically difficult. The surgeon must try toclose large defects in the bladder often with only smallremnants of residual bladder tissue. Although there areas yet no basic histologic studies of the tissue surroun-ding obstetric fistulas, it seems clear that these tissueshave themselves sustained significant damage duringobstructed labor. The fistula itself develops in an areawhich has sustained enough damage to become necro-tic; the process by which this occurs also appears toaffect another, variable zone surrounding the fistula inwhich the tissues have been damaged but not killed. In

904

repairing obstetric fistulas, it is this unhealthy surroun-ding tissue that must be used to close the defectÑ-aproblem that has led many fistula surgeons over thegenerations to the brink of despair (Figure 6).

In some cases pressure necrosis may destroy virtuallythe entire bladder, so that when the defect is finally clo-sed the afflicted woman is left with a remarkably small(30 - 50 ml) bladder that remains virtually functionless.Because most of the innervation of the bladder runsthrough the base and trigone, ischemic injury to theseareas probably also produces an element of neuropathicbladder dysfunction. Basic scientific studies confirmingthis hypothesis have yet to be undertaken. Clinicalexperience with fistula patients also suggests that blad-der compliance may be altered by the extensive fibroticchanges that often take place. To date there has beenonly one urodynamic study of post-repair fistulapatients (Schleicher et. al 1993), but compliance wasnot measured in this study and the number of patients

evaluated (18) is too small to allow meaningful genera-lizations to be made. There is a great need for furtherinvestigation of these issues; unfortunately, those hos-pitals most likely to see large numbers of patients withobstetric fistulas lack the resources for such urologicinvestigation. The potential use of bladder augmenta-tion operations to restore the functional bladder capaci-ty of such patients remains largely uninvestigated.Although there are well-established surgical techniquesfor bladder augmentation that can be used in the healthcare systems of industrialized countries, for the mostpart those developing countries in which fistulas aremost likely to be major problems are precisely thosecountries in which the capacity is lacking to performsuch operations safely, competently, and successfully.At least for the present, complex bladder augmentationoperations (cecocystoplasty, ileocystoplasty, gastrocys-toplasty, etc.) do not appear to be feasible therapies formost women with contracted bladders after obstetricfistula repair. One possible solution to this problemmay lie in the use of Òauto-augmentationÓ operations inwhich a portion of the detrusor muscle is removed, per-mitting the underlying bladder mucosa to expand, inessence creating a large bladder diverticulum (Kennel-ly et. al. 1994; Snow and Cartwright 1996; Leng et. al.1999). At present the potential of these techniques inthe treatment of fistula patients with small fibrotic blad-ders remains completely unexplored within the contextof the developing world.

b) Urethra. The ischemic changes produced by obs-tructed labor often have a devastating impact on ure-thral function. The great Egyptian fistula surgeonNaguib Mahfouz was well aware of this problem. As hewrote in 1930:

I have carefully examined 100 patients suffering from

905

Table 2 : Spectrum of Injuries Seen in the "Obstructed Labor Inju-ry Complex"

Urologic InjuryVesicovaginal fistulaUrethrovaginal fistulaUreterovaginal fistulaUterovesical fistulaComplex combinations of fistulasUrethral damage, including complete urethral lossStress urinary incontinenceSecondary hydroureteronephrosisChronic pyelonephritisRenal failure

Gynecologic InjuryAmenorrheaVaginal stenosisCervical damage, including complete cervical destructionSecondary pelvic inflammatory diseaseSecondary infertility

Gastrointestinal InjuryRectovaginal fistulaAcquired rectal atresiaAnal sphincter incompetence

Musculoskeletal InjuryOsteitis pubis

Neurological InjuryFoot-dropComplex neuropathic bladder dysfunction

Dermatological InjuryChronic excoriation of skin from maceration in urine/feces

Fetal InjuryApproximately 95% fetal case mortality rate

Social InjurySocial isolationDivorceWorsening poveryMalnutritionDepression, sometimes suicide

Figure 6 : Large mid-vaginal vesico-vaginal fistula withscarring. The ureters, which have been catheterized for bet-ter visualization, lie directly on the margins of the fistula.(©Worldwide Fund for Mothers Injured in Childbirth, usedby permission).

what was termed vesico-vaginal fistula during the last10 years. I found by careful examination and measure-ment of the urethra that the sloughing which ultimatelyled to the formation of the fistulae had in more than halfthe cases involved from one-third to half of the urethra.This is not to be wondered at, since in most cases ofobstructed delivery in which the bladder is pulled upabove the brim of the pelvis the urethra is pulled upwith it. If the seat of obstruction happens to be at thebrim of the pelvis, the neck of the bladder and a smallportion of the upper third of the urethra seldom escapescompression.. In cases in which the presenting part isimpacted in the cavity of the pelvis, or detained at theoutlet, the entire urethral canal will be lying in the planeof compression. In some of these cases the urethrasloughs away completely.È

The finest centers in the world report fistula closurerates in excess of 90%, yet many patients who have hada successful fistula repair continue to have severe uri-nary incontinence. Although the bladder defect hasbeen closed successfully, many patients have defective,injured urethras which are often foreshortened, fibrotic,functionless ÇdrainpipesÈ densely bound in scar tissue.Patients with these findings may remain almost totallywet in spite of fistula closure, and perhaps as many as30% of fistula patients have some element of persistentstress incontinence after repair (Hudson and Henrickse1975; Hassim and Lucas 1974; Schleicher et. al. 1993).The development of successful techniques for dealingwith this problem remains an unmet challenge in fistu-la surgery (Hilton et. al., 1998).

Loss of the urethra traditionally has been the most fea-red form of obstetric fistula. Complete urethral lossoccurs in about 5% of fistula patients, with about 30%of fistula patients sustaining partial urethral injury.Mahfouz stated (1930) that fistulas Çin which the wholeurethra has sloughedÈ are Çthe most troublesome ofall.È The experience of subsequent surgeons seems tobear this out. In a 1980 series based on 1,789 fistulapatients, Sister Ann Ward reported that only 26 caseswere inoperable; but in all 26 urethral loss was present.In urethral fistula repair, the surviving tissues must bereassembled not just as a tube, but as a supple, functio-nal organ that serves both as a conduit for urine as wellas a ÇgatekeeperÈ ensuring that the passage of urineoccurs only at socially appropriate times and places.There are no comparative surgical studies that evaluatediffering techniques of urethral reconstruction inpatients with obstetric fistulas. Work of this kind isbadly needed.

c) Ureters. Ureterovaginal fistulas from direct injury tothe distal ureter during obstructed labor are uncommon,comprising only about 1% of fistula cases. Dependingon the amount of tissue that is lost at the bladder base,

the ureteral orifices can be found in bizarre locations,ranging from the lateral vaginal walls all the way up tothe level of the vesicourethral junction and the pubicarch (Figure 6). Aberrant ureteral locations of this kindcan easily be missed on clinical examination and areone cause of persistent incontinence after otherwiseÇsuccessfulÈ fistula closure. Standard urological toolssuch as ureteral stents are usually not available in hos-pitals in the developing world, and most of the surgeonswho work in such hospitals are not trained in Çurolo-gicÈ techniques such as ureteral reimplantation (Waal-dijk 1995).

d) Kidneys. The incidence of secondary injury to theupper urinary tract in fistula patients has received littlestudy, but this phenomenon appears to be clinicallyimportant. Clinical experience suggests that renal failu-re is a common cause of death in women with obstetricfistulas. Upper tract damage could result from chronicascending infection, obstruction from distal ureteralscarring, or even from reflux in very young patients.Lagundoye et al (1976) found that 49% of fistulapatients had some abnormality of the kidneys whenintravenous urograms were performed. Most of thepathology that was detected consisted of minor calycealblunting, but 34% of patients had hydroureter, 9.7%had ureteral deviation, four patients had bladder stones,and 10 patients had a non-functioning kidney. In thedeveloping world, hospitals typically have neither thelaboratory capability to detect azotemia, nor the radio-graphic facilities to diagnose hydroureteronephrosis. Itis clear, however, that injury to the kidneys is a com-mon complication in patients with obstetric fistulas.

2. GYNECOLOGIC INJURY

a) Vagina. An impaction of the fetal head seriousenough to cause ischemic injury to the bladder will alsocause ischemic injury to the vagina, which is likewisetrapped between the two bony surfaces. The necroticareas that develop subsequently heal with varyingdegrees of scarring. A small sonographic study by Ade-tiloye and Dare (2000) detected fibrotic changes in 32%of fistula patients and minor vaginal wall fibrosis inanother 36%. Vaginal injuries in fistula patients arearrayed along a spectrum the includes only small focalbands of scar tissue on one end all the way to virtualobliteration of the vaginal cavity on the other. Roughly30% of fistula patients require some form of vagino-plasty at the time of fistula repair.

The degree of vaginal injury has several importantimplications. In the first instance, severe vaginal injuryresults in loss of substantial portions of the vagina. Inmany instances the scarring is such that vaginal inter-course is simply not possible. There is virtually noinformation available on the sexual functioning of fis-

906

tula patients, yet this is obviously an important concernin healthy marital relationships and undoubtedly contri-butes to the high rates of separation and divorce thatappear common among these women. Surgical repair offistulas in women with extensive vaginal scarring oftenrequires the use of flaps and tissue grafts in order toclose the fistula. Little work has been done to assesswhether or not sexual function normalizes in womenwho have had such operations. The presence of scarringthat requires the use of plastic surgical techniques ofthis kind markedly reduces the effectiveness of surgicalrepair when fistula closure is attempted by surgeonswho lack experience in reconstructive gynecologic sur-gery. Although several papers have described varioustechniques for vaginoplasty that may be required in fis-tula patients (Dick and Strover 1971; Hoskins et. al.1984; Margolis et. al 1994), there is a pressing need toinvestigate the role of vaginal plastic surgery at the timeof fistula repair and to evaluate subsequent sexual func-tioning in patients who require surgery of this kind.

Vaginal scarring impacts more than just sexual functio-ning. The presence of vaginal scarring appears to be animportant prognostic factor in determining the likeli-hood both of successful fistula closure, and also for thedevelopment of debilitating urinary stress incontinenceafter otherwise successful fistula repair. In one unpubli-shed series of 26 fistula patients with severe vaginalscarring, 57.7 percent suffered from stress incontinenceafter fistula repair and 23.5 percent had a persistent orrecurrent fistula This would compare to an expectedstress incontinence rate of around 26% and a failed fis-tula closure rate of about 7% in the overall populationof fistula patients (Arrowsmith, personal communica-tion). As Lawson Tait remarked in his book, Diseases ofWomen (1879): ÇOne case, which I utterly failed toimprove in any way, had the whole vagina destroyed bysloughing, so that the rectum, ureters, and uterus ope-ned into a common cloaca about two inches deep, withwalls of cartilaginous hardness. In such cases the dama-ge is nearly always very extensive and very difficult toremedy.È Kelly and Kwast have also reported worse-ning surgical outcome in fistula patients who have vagi-nal scarring than in those without such findings(1993a).

Cervix, Uterus, and Future Reproductive Performan-ce. Many patients sustain severe cervical damage aswell as vaginal injury in the course of obstructed labor.When fistula patients are examined, a completely nor-mal cervix is rarely seen. The presence of cervical inju-ry would also help explain the apparently high preva-lence of pelvic inflammatory disease encounteredamong these patients. In the worst cases, prolongedobstructed labor may result in complete cervical des-truction, leaving the patient with no identifiable cervi-cal tissue at all. Unfortunately, detailed descriptions of

the condition of the cervix have not been included inthe series of fistulas published to date. Since cervicalcompetence is such an important factor in future repro-ductive performance, this is yet another clinical areathat demands further study.

A review of the menstrual histories of 998 patients withobstetric fistulas in Ethiopia (Arrowsmith et. al. 1996)showed 63.1 percent were amenorrheic. Other studieshave shown amenorrhea rates from 25% to 44% (Aima-khu 1974; Bieler and Schnabel 1976; Evoh and Akinla1978). Many of these patients undoubtedly have hypo-thalamic or pituitary dysfunction (Bieler and Schnabel1973). While the high incidence of amenorrhea in VVFpatients is widely recognized, only one unpublishedstudy has been done to look specifically at uterinepathology in the VVF population. Dosu Ojengbede ofthe University of Ibadan (personal communication)performed hysteroscopy on fistula patients in Nigeriaand found that intrauterine scarring and AshermanÕssyndrome are common. The combination of widespreadamenorrhea, vaginal scarring, and cervical destructionleads to a tremendous problem of secondary infertilityamong these patients. To date, there have been noserious scientific efforts to explore treatment of cervicaland uterine damage in VVF patients.

Subsequent reproductive performance of women whohave had an obstetric vesicovaginal fistula has beenanalyzed in a few articles (Naidu and Krishna 1963;Aimakhu 1974; Evoh and Akinla 1978; Emembolu1992). Emembolu analyzed the subsequent reproducti-ve performance of 155 fistula patients delivered atAhmadu Bello University Teaching Hospital in Zaria,Nigeria, between January 1986 and December, 1990.This series included pregnancies in 75 women whobecame pregnant after successful fistula closure and 80women who became pregnant while still afflicted withan unrepaired fistula that had occurred in a previouspregnancy. The data presented do not allow one todetermine the subsequent fertility rates of women whodevelop a fistula, but clearly indicate that women can,and do, become pregnant after sustaining an obstetricfistula. The proportion of booked pregnancies receivingantenatal care was higher in the repaired group (73%)than in the unrepaired group (51%), and reproductiveperformance was better (but still dismal) in thosepatients who had had a fistula repair. Of the 69 patientswhose fistulas had been repaired, there was a recurren-ce in 8 (11.6%), and among those undergoing a trial ofvaginal delivery, the fistula recurrence rate was nearly27%. In women with unrepaired fistulas who did notregister for prenatal care, maternal mortality and mor-bidity in subsequent pregnancies was high and severe,and reflects the conditions that led to fistula formationin the first instance:

907

ÇMrs. A.A., a 25-year-old unbooked para 1+0 (neona-tal death) whose VVF was as yet unrepaired was admit-ted at 40 weeks gestation in the second stage of labor.She had been in labor at home for over 20 h beforeadmission. Vaginal examination revealed that the vagi-na was markedly stenosed and the patient in advancedobstructed labor with an intrauterine fetal death. Atcesarean section, the uterus was found to have ruptu-red. The macerated female fetus weighting 2800 g wasdelivered and the patient had a repair of the uterus withbilateral tubal ligation. She however died on the sixthpost-operative day of persistent anemia and overwhel-ming infection due to Klebsiella species.

Mrs. M.A., a 28-year old unbooked para 1+0 (stillbir-th) who had a residual vesico-vaginal fistula wasadmitted in labor at 34 weeks gestation. She had beenin labor at home for about 19 h prior to admission inthe second stage of labor. Vaginal examination revealedmarked vaginal stenosis admitting only one finger. Adiagnosis of advanced obstructed labor with fetal dis-tress was made. At cesarean section, a severelyasphyxiated female baby weighing 2200 g with Apgarscore of 1 and 2 at 1 and 5 min was delivered. The babydied within 30 min of delivery. The mother subsequent-ly developed post-partum eclampsia with septicemiafrom Staphylococcus aureus and died on the fourthpost-operative day.

The commonest maternal morbidity, excluding recur-rence of VVF, was hemorrhage requiring blood transfu-sion in 35 patients (27.3%). Others included ruptureduterus in 3 unbooked patients whose fistulae had notbeen repaired, bladder injury at cesarean section in1.6% and acute renal failure in 0.8%. The maternalcomplication occurred more frequently in the patientswhose fistulae had not been repaired and who were alsounbooked.È

The largest series is that of Aimakhu, who analyzedsubsequent reproductive performance in 246 womenwho underwent successful fistula closure at UniversityCollege Hospital in Ibadan, Nigeria, between 1957 and1966. Only 48 patients became pregnant following fis-tula repair with a total of 65 pregnancies. All but 6 ofthese were managed at University College Hospital.

Five patients had aborted prior to the 16th week of ges-tation, leaving only 60 viable pregnancies. The planwas to perform elective Cesarean section on all patientswho became pregnant after fistula surgery, but only 49Cesarean operations were carried out. The results of thevaginal deliveries were not encouraging. Aimakhusummarized the results in this fashion (1974):

ÇOf the eight vaginal deliveries, six were our bookedpatients, and they were all admitted in the second stageof labor; one patient had in fact delivered the first of

her twins before arrival. There were no maternal deathsamong these six patients and there was no reopening ofthe fistulae after delivery. Five of the seven babies sur-vived. The seventh patient with vaginal delivery had aninstrumental vaginal delivery elsewhere. The baby wasstillborn and the fistula recurred. At the beginning ofthe repair of the new fistula in our unit, this patient diedfrom cardiac arrest. The eighth patient with vaginaldelivery sustained her first fistula in her first delivery in1952. This was successfully repaired elsewhere 4 yearslater. In 1962 she was allowed a vaginal delivery inanother hospital. The baby was stillborn and the fistu-la recurred.. Following another successful repair of thefistula she became pregnant again. She defaulted froma cesarean section offered her in our unit. She deliveredat home another stillborn baby and the fistula wasagain reopened.È

Patients who underwent cesarean delivery fared better.There were 49 fetuses delivered and 47 survived. Therewas no recurrent fistula among women previouslyrepaired who became pregnant and had a subsequentcesarean section. There was one maternal death frompulmonary embolism in a woman who underwent aemergency delivery at 32 weeks gestation due to a pro-lapsed fetal umbilical cord.

3. THE RECTO-VAGINAL FISTULA

Rectovaginal fistulas appear to be significantly lesscommon than vesico-vaginal fistulas. In case series ofpatients presenting with vesico-vaginal fistulas, bet-ween 6% and 24%% have a combined recto-vaginaland vesico-vaginal fistula (Figure 7) (Arrowsmith1994; Ashworth 1973; Aziz 1965; Bird 1967; Coetzeeand Lithgow 1966; Emmett 1878; Hilton and Ward1998; Mustafa and Rushwan 1971; Naidu 1962; Tahzib1983, 1985; Waaldijk 1989). In most series, isolatedrecto-vaginal fistulas are less common than combinedfistulas. Indeed, most series do not even mention isola-ted recto-vaginal fistulas as a clinical phenomenon. In aseries of patients from Turkey reported by Yenen andBabuna (1965), 7.1% had recto-vaginal fistulas and6.5% had ÇcombinedÈ fistulas. From the report it is notclear if this latter figure was comprised solely of recto-vaginal and vesico-vaginal fistulas, or if it includedother combinations of urinary tract fistulas as well(vesico-cervico-vaginal, urethro-vaginal, etc.). Kellyand Kwast (1993), reporting data from the Addis AbabaFistula Hospital, noted a 15.2% prevalence of combi-ned fistulas, and a 6.8% prevalence of isolated rectova-ginal fistulas in that population. Ethiopia appears tohave one of the highest rates of recto-vaginal fistulasreported in the literature. Whether this relates to speci-fic obstetric characteristics of the Ethiopian populationor whether this relates to other social factorsÑ-such as

908

the cases of rape and sexual abuse of young Ethiopiangirls reported by Muleta and Williams (1999)Ñisunclear.

Since the pubic symphysis poses an obstruction to deli-very through the anterior pelvis, in normal birth mecha-nics the fetal head is normally forced posteriorlytowards the rectum, anus, and perineum towards theend of the second stage of labor. In non-obstructedlabor, direct laceration of the perineum is not uncom-mon, occasionally resulting in a complete perineal tearwith complete disruption of the anal sphincter. If this isnot repaired, a complete perineal tear with sphincterdisruption can create a recto-vaginal fistula at the analoutlet. This mechanism of fistula formation seems morelikely to account for low recto-vaginal fistulas, whereasrecto-vaginal fistulas higher in the pelvis would seemmore likely to be caused by direct tissue compressionfrom obstructed labor. As Mahfouz noted in 1938:

ÇThe process by which faecal fistula forms after labourdiffers greatly from that which leads to the formation ofa urinary fistula. Sloughing, due to pressure-necrosisproduced by impaction of the presenting part, which

accounts for the overwhelming majority of urinary fis-tulae, is seldom the cause of faecal fistula. It accountedfor 2 case only in my series of 75. The majority of theremaining cases were the result of a complete tear ofperineum which extended into the recto-vaginal sep-tum. The lacerated edges of the perineum united spon-taneously in the lower part where the tissues were fle-shy, but remained ununited at the upper end where tis-sues were thin. This results in a permanent communica-tion between the vagina and rectum at the upper end ofthe healed tear.È

In Das and SenguptaÕs series of 135 obstetric fistulasfrom India, there were 12 patients with recto-vaginalfistulas, 6 patients with complete perineal tears, 1patient with both a recto-vaginal fistula and a completeperineal laceration, and 1 patient with a recto-vaginalfistula as well as a fistula-in-ano, giving a posterior fis-tula rate of 20/135 or 14.8%.

Because recto-vaginal fistulas appear to be less com-mon than genito-urinary fistulas in the developingworld, less attention has been paid to describing tech-niques for their repair in this setting. Mahfouz recom-mended different approaches to recto-vaginal fistulas,depending on their location (1938).

ÇThe methods of treatment of faecal fistula differ accor-ding to their site. If the fistula is situated at the vaginaloutlet, incorporated in or lying immediately above anincompletely healed perineal tear, the perineum shouldbe cut through. In other words, the recto-vaginal fistu-la is converted into a complete tear of the perineum andis dealt with as such. The vaginal and rectal walls arenext separated from one another by a transverse inci-sion. This separation should be carried well above theupper edge of the fistula. The rent in the rectum is nowcarefully sutured with catgut. The sutures should notpierce the mucous membrane of the gut. The next stepis to unite the levator ani muscles in the middle line sothat a thick mass of tissue is interposed between thelines of sutures in the vagina and rectum respectively.The cut ends of the sphincter should now be very care-fully brought together and the perineum reconstructedin the usual manner.

In dealing with rectal fistulae situated at a distancefrom the perineum the latter should not be cut through.These fistulae should be dealt with by a flap-splittingoperation performed on the same principles employedin operating on urinary fistulae. The separation of therectal from the vaginal wall should be carried until apoint well beyond the upper and lower limits of the fis-tulae. In rectal fistulae this separation can be effectedmore easily, and much more widely, than separation ofthe bladder in urinary fistulae.È

These general principles still prevail. For example,

909

Figure 7 : Large combined vesicovaginal and rectovaginalfistula. (©Worldwide Fund for Mothers Injured in Child-birth, used by permission).

Arrowsmith (1994) recommended oral cathartics andcleansing enemas (where the fistulas were smallenough to allow the patient to retain some enema fluid),followed by wide mobilization of the fistula, a two-layered rectal closure, with perineal reconstruction andthe use of Martius bulbocavernosus fat flaps for largedefects. These general principles have also been endor-sed by Hudson (1970).

More problematic is the very high recto-vaginal fistula,where the fistula occurs in the upper vagina (Bentley1973; Eden 1914; Lawson 1972; Mahfouz 1934). Suchcases are particularly troublesome if the vagina is verystenotic or the fistula is fixed and tethered in such amanner as to prevent its being dissected free andbrought down low enough so that a vaginal repair canbe attempted. In these cases, the route of repair must betransabdominal. Before an abdominal attempt at recto-vaginal fistula repair, the patient should always have atemporary transverse colostomy performed prior to fis-tula repair and undergo as thorough a bowel preparationregimen as is possible. In HudsonÕs series of 88 highrectovaginal fistulas, successful closure was signifi-cantly more likely if a colostomy had been performedprior to attempted fistula closure (1970). Each abdomi-nal repair operation is likely to be different, due to thecomparative rarity of the high rectovaginal obstetric fis-tula and the variable severity and extent of the concur-rent pathology that is likely to accompany it, but ingeneral the surgeon should attempt to separate the vagi-na from the rectum, close the rectum in multiple layers,interpose a pedicled graft of omentum between the rec-tum and the vagina and, where possible, close the vagi-na separately (Bentley 1973).

Clinical experience suggests that Çdouble fistulasÈ(combined vesico-vaginal and recto-vaginal fistulas)are more difficult to repair than solitary rectovaginalfistulas (Hudson 1970). In patients with double fistulas,the amount of normal tissue available for use as vaginalflaps is often very limited, and the reduction in vaginalspace that occurs after successful closure of a doublefistula can be formidable. The presence of a rectovagi-nal fistula also decreases the success rate for repairingthe vesicovaginal fistula that is present.

Rectovaginal fistula repair presents special problemsfor the fistula surgeon, particularly in cases where thefistula is high or complex. Many gynecologic surgeonsmay not be comfortable performing a colostomy, letalong more complex colon and rectal operations such asa colo-anal pull-through procedure. Colostomy surgeryrequires access to more sophisticated anesthesia capa-bility than vaginal fistula repair, and in many cultures acolostomy is no more acceptable than a rectovaginalfistula, particularly as useable colostomy appliances areoften difficult or impossible for most fistula patients to

obtain. From a research perspective, the issue of analsphincter function in obstetric fistula patients remainsvirtually unaddressed, both with regard to the presenceof injuries in women who only have a vesico-vaginalfistula, as well as in patients who have had successfulrecto-vaginal fistula repair.

4. ORTHOPEDIC TRAUMA

Ischemic injury from obstructed labor not only affectspelvic organs, but also the pelvis itself. These changesare most pronounced in the pubic symphysis. The nor-mal radiography of the symphysis pubis has been des-cribed in detail by Vix and Ryu (1971). In obstructedlabor, the pubic bones are often directly involved as oneside of the bony vise in which the vulnerable soft tis-sues are trapped. In fistulas where large amounts ofbladder tissue are lost, the periosteum of the pubic archcan often be palpated directly through the fistula defect.It is these cases in which ischemic damage to the pubicbones is most likely to be demonstrable. In a study of312 Nigerian women with obstetric vesicovaginal fistu-las Cockshott (1973) noted bony abnormalities in 32percent on plain pelvic radiographs. The findings inclu-ded bone resorption, marginal fractures and bone spurs,bony obliteration of the symphysis, and wide (>1 cm)symphyseal separation. Most of these changes appearto be the result of avascular necrosis of the pubic sym-physis. Their long-term significance remains uncertainand further study is required.

5. NEUROLOGIC INJURY

Another tragic injury associated with obstetric fistulaformation is foot-drop (Waaldjik and Elkins 1994). Therelationship between difficult labor and neurologicalinjury has been known for centuries, and the conditionwas traditionally called Çobstetric palsyÈ (Sinclair1952). Women with this condition have an inability todorsiflex the foot and therefore walk with a seriouslimp, dragging their injured foot, and using a stick forsupport [Figure 8]. In SinclairÕs paper on maternal obs-tetric palsy in South Africa (1952), he made the com-ment that ÇThere are no records of this lesion associa-ted with vaginal fistulae, where there has been prolon-ged pressure by the foetal skull in the lower part of thepelvis.È This statement is clearly wrong. In Waaldijkand ElkinsÕ review of 947 fistula patients, nearly 65%of those studied prospectively had evidence of peronealinjury either by history or physical examination (1994).The prevalence of clinical footdrop among patientsseen at the Addis Ababa Fistula Hospital is about 20%(Arrowsmith, Hamlin and Wall, 1996). Various theorieshave been proposed for the etiology of this condition.In general clinical series of peroneal nerve palsy, themost common etiologies appear to be direct trauma

910

from ankle inversion, fractures of the hip, femur, fibulaor tibia; knee injuries, alcoholic neuropathies, and avariety of miscellaneous or idiopathic causes. In obste-tric patients the lesion most likely develops from one ofthree causes: prolapse of an intravertebral disk, pressu-re from the fetal head on the lumbo-sacral nerve trunkin the pelvis leading to direct compression of the per-oneal nerve, or direct trauma to the peroneal nerve fromprolonged squatting and pushing in the second stage oflabor (Sinclair 1952; Colachis et. al. 1994, Reif 1988).Of the three possible etiologies, the last seems the mostlikely; however, to date no one has performed electro-myographic studies of the lower extremities in fistulapatients with foot-drop to evaluate the neurophysiologi-cal abnormalities that are present. Foot-drop has alsobeen associated with trauma sustained in difficult for-ceps deliveries, particularly mid-pelvic rotations, butagain, as others have emphasized ÇThe peripheral nervelesion following instrumental delivery may have deve-loped in any case and forceps were but incidental or atthe most a precipitating factor in border-line casesÈ(Sinclair 1952). The prognosis for recovery from thisinjury is unclear, as there are no proper prospective stu-dies of women who have developed this condition.Waaldijk and Elkins suggest that most patients recoversome or all of their nerve function spontaneously withintwo years of the injury (1994); however 13% showedpersistent signs of nerve trauma. In some cases theaffected women are almost completely crippled withbilateral lesions and suffer tremendously from the addi-tional burden imposed by immobility on someone whoalready suffers from intractable urinary incontinence.Physiotherapy and the use of posterior splints improve

the condition of some patients. Others may require sur-gical intervention: the use of posterior tibialis tendontransfer is a well-established procedure for patientswith foot-drop from other causes (such as leprosy), andit may be that this method will be useful in treatingwomen with unresponsive obstetric palsy as well (Hall1977; Richard 1989).

6. DERMATOLOGIC INJURY

The condition of the skin, which is in constant, unre-mitting contact with a stream of urine or feces, is one ofthe most bothersome problems for the fistula patient[Figure 9]. The problems thus encountered have beeneloquently described since the early days of fistula sur-gery. The great American gynecologist Thomas AddisEmmett summarized the clinical situation in 1868:

ÇUnless the greatest care has been given to cleanliness,the sufferer, in a few weeks after receiving the injury,becomes a most loathsome object. From the irritationof the urine, the external organs of generation becomeexcoriated and oedematous, with the same conditionextending over the buttocks and down the thighs. Thelabia are frequently the seat of deep ulcerations andoccasionally of abscesses. The mucous membrane ofthe vagina is in part lost, and the abraded surfacerapidly becomes covered at every point with a sabulousor offensive phosphate deposit from the urine. If theloss of tissue has been extensive, the inverted posteriorwall of the bladder protrudes in a semi-strangulatedcondition, more or less incrusted with the same deposit,and bleeding readily. This deposit will frequently accu-mulate to such an extent in the vagina that the suffererbecomes unable to walk or even to stand upright,without the greatest agony.

The deposit must be carefully removed as far as pos-sible by means of a soft sponge, and the raw surfacebrushed over with a weak solution of nitrate of silver. If,at any point, it cannot be at first removed without cau-sing too much bleeding, the deposit itself must be trea-ted in the same manner, or coated with the solid stick.Warm sitz-baths add greatly to the comfort of the suffe-rer. The vagina must be washed out several times a daywith a large quantity of tepid water. After bathing, it isbest for the patient to protect herself by freely anointingthe outlet of the vagina and the neighboring parts withany simple ointment. She must be instructed to washher napkins thoroughly when saturated with urine, andnot simply to dry them for after-use. Time, and increa-sed comfort of the patient, are gained by judiciousattention to such details.

About every fifth day, the excoriated surfaces yetunhealed should be protected with the solution of nitra-te of silver; and it is frequently necessary to pursue thesame general course for many weeks, before the parts

911

Figure 8 : Fistula patient with foot-drop, manifested by aninability to dorsiflex the foot. In particular note the traumato the toes, as well as the constant puddle of urine in whichshe walks. (©Worldwide Fund for Mothers Injured inChildbirth, used by permission).

can be brought into a perfectly healthy condition. Thispoint is not reached until not only the vaginal wall, butalso the hypertrophied and indurated edges of the fistu-la, have attained a natural color and density. This is thesecret of success; but the necessity is rarely apprecia-ted; without it, the most skillfully performed operationis almost certain to fail.

When the proper condition has been brought about, thesurgeon may then be able to decide upon some definiteplan of procedure for the closing of the fistulaÈ.

Further work remains to be done on determining theoptimal regimen of skin care for patients with fistulaswithin the context of developing countries. It seemsclear that efforts of this kind will produce markedimprovements in the reduction of patient suffering,even before fistula closure is undertaken.

7. SOCIAL CONSEQUENCES OF PROLONGED

OBSTRUCTED LABOR

Although physicians tend to think in terms of clinicallydefinable injuries, much of the suffering that fistulapatients endure is a result of the social consequences oftheir condition. It is vitally important to understand thesocial context in which these injuries occur (Wall1998). Although there is no such thing as a monolithicÇAfrican cultureÈ, in many areas where fistulas areendemic there appear to be recurring patterns that allowfor some general observations to be made regarding theposition of women. In countries where there are highfistula rates, women generally have a lower social sta-tus than men. Often a womanÕs role in family life cen-ters around a strong obligation to satisfy the sexual

needs of her husband and to provide him with offspring(preferably male), and both men and women are depen-dent on their children for care when they become old.In many African societies, women are often called uponto perform heavy manual labor, tending the fields, car-rying water and firewood. ReligionÑbe it Islam, Chris-tianity, or a traditional African religionÑ-also plays amore central role in day-to-day life than is common inWestern countries. Each of these areas of human life isaffected in a profoundly negative way by the injuriessustained in obstructed labor.

a) Marriage and family life

In societies where obstetric fistulas are still prevalent, awomanÕs role in life is defined almost exclusively interms of marriage, childbearing, and the family life thatresults (Wall 1998). Because most women in thesesocieties appear to accept this role at present, the inabi-lity to have children or to satisfy her husbandÕs sexualneeds may diminish her own sense of self worth. Vagi-nal injuries often make intercourse impossible, and theconstant stream of urine makes it otherwise unpleasant.As members of agrarian societies, women are oftenexpected to contribute long hours of hard labor workingon the family farm. Foot drop and associated pelvicinjuries may make the satisfactory performance of thesetasks impossible. The woman, formerly a productivelaborer, then becomes an economic burden as an inva-lid. The combination of all of these factors often leadsto a gradual disintegration of the marriage over time,which then ends with complete rupture of the relation-ship. The existing data suggest that large numbers offistula patients become divorced or separated from theirhusbands, particularly when it becomes evident thattheir condition is chronic, rather than transient. Unpu-blished data from Ethiopia suggest that almost 50 per-cent of VVF victims are divorced or separated (S.D.Arrowsmith, personal communication). MurphyÕsresearch documented similar findings in northern Nige-ria (1981).

Infertility is a devastating problem to couples in anyculture, but it is difficult to underestimate the importan-ce of fertility in African societies. Large families are asource of pride and a symbol of affluence. Since thesocial, economic, and political lives of these societiesare still dominated in many respects by ties of kinship,not having offspring is a disaster on many fronts. Fur-thermore, large families may be the only source ofreliable (or affordable) farm labor, a critical economicfactor in societies based on peasant agriculture. Becau-se governmental social welfare programs are unreliableor non-existent in most African countries, children arethe only hope one has for security in old age. The fetalmortality in obstructed labor is staggering, generallyabove 90%. Since obstructed labor is most commonly a

912

Figure 9 : Encrustation of the vulva with uric acid salts, asthe result of the constant trickle of urine from the fistula.(©Worldwide Fund for Mothers Injured in Childbirth, usedby permission).

complication of a womanÕs first labor, and since she isusually infertile thereafter, the majority of fistula vic-tims (about 70%) have no living children. Not only isthis a disaster for the affected woman, but from her hus-bandÕs perspective, it puts his whole future in jeopardyas well. Faced with this prospect, many men find iteasier to rid themselves of their damaged wives andseek other, fertile, spouses.

Although obstructed labor is most common as a compli-cation of a first pregnancy, it can occur in any pregnancyif the baby is too large, presents wrong, or other compli-cations arise. In a review of 121 obstetric fistulas fromKumasi, Ghana, Danso and colleagues found that 43% ofpatients had had three or more pregnancies, and that 25%had had five or more (Danso et. al. 1996). Similarly, atthe Addis Ababa Fistula Hospital, nearly 25% of fistulapatients have had three or more pregnancies, and 11 per-cent have had six or more children (Kelly and Kwast,1993). What happens to the living children when theirmotherÕs life is ruined in this fashion? Important indirecthealth consequences of fistula formation such as thishave been largely unexplored.

b) Religious and social implications

In some parts of the world, such as northern Nigeria,married women live under a system of Çwife seclusionÈin which their social contacts are severely restricted totheir immediate family and female neighbors (Wall1998). The offensive odor that accompanies total urina-ry and/or fecal incontinence usually curtails even thislimited opportunity for social interaction. In order todeal with the never-ending problem of foul smells andomnipresent urine and fecal loss, the families of thesepatients may remove them from the main family dwel-ling into a peripheral hut, sometimes even forcing themto live outdoors. Not uncommonly, they are often for-ced out of the family compound altogether over time.

Because of the nature of her injury, a fistula patient sim-ply cannot maintain normal hygiene, no matter howhard she tries. This fact has an enormous impact on allaspects of her life, including her participation in reli-gious or spiritual life. Fistula women are generallyregarded as both physically and ritually unclean. ManyAfrican religious groups, especially Muslims, requirepersonal cleanliness as a prerequisite for worship. Thisoften excludes them from participation in religious acti-vities (often a central concern of African social life);this further diminishes their sense of self-worth andsocial connectedness.

The tragedy faced by these women was described veryeloquently by Kelsey Harrison, who had extensiveexperience with their plight as a result of his long tenu-re as Professor of Obstetrics and Gynaecology atAhmadu Bello University in Zaria, Nigeria (1983):

ÇWhether in hospital or outside, their own society goesto great lengths to ostracize these girls, an action dis-paraged by outsiders but considered reasonable bymost local people including certain (male) health wor-kers. Their point of view is straightforward enough. Thesafe delivery of a healthy baby is always an occasionfor great rejoicing and at the naming ceremony held onthe eighth day after birth, the whole family and localcommunity celebrate. In the case of the girl with anobstetric fistula, the baby is usually stillborn and thistogether with the fact that her odour is offensive meansthat such celebrations cannot take place. Soon, herincontinence becomes confused with venereal disease,and the affected family feels a deep sense of shame. Theconsequences are devastating: the girl is initially kepthidden; subsequently, she finds it difficult to maintaindecent standards of hygiene because water for washingis generally scarce; divorce becomes inevitable anddestitution follows, the girl being forced to beg for herlivelihood. So traumatic is this experience that evenwhen cured, some girls never regain their self esteem.Ó

In considering the overall impact of the fistula problem,one should consider the life-time burden of sufferingthat this condition presents. The vast majority of thesewomen are young and develop their fistula in their firstpregnancy. In a series of over 9,000 fistula victims inEthiopia the average patient age was 19 yrs (Arrows-mith, Hamlin and Wall, 1996). Most of the injuriesassociated with prolonged obstructed labor cause per-manent disability without causing early death. Themembers of the committee have seen patients presen-ting for surgery over 40 years after the initial injury: thewasted years of human life represented by such cases ismind-numbing.

Professor Abbo Hassan Abbo, Professor of Obstetricsand Gynaecology at the University of Khartoum in theSudan, himself an international authority on fistulas,tells a powerfully poignant story of a group of Somaliwomen with fistulas who, in despair, chained them-selves together and jumped off the dock in Mogadishuin a mass suicide because their suffering had becomeunendurable. (A.H. Abbo, personal communication).

The social ramifications of having a fistula is a majorco-morbidity that is easy to forget unless one has seenthe patient in her own social setting. This is why theproper care of fistula victims must be based on a holis-tic approach that pays as much attention to healing thepsycho-social wounds inflicted on these women as itdoes to curing their physical injuries. Programs forwomen with obstetric fistulas must encompass educa-tion, literacy training, the development of social net-works, and the provision of skills with which to earn anadequate livelihood, if the social problems that these

913

women face are to be overcome (Bangser 1999). Somewomen are so injured that they can never return to theirhome villages, and many have such serious medicalproblems that they linger around fistula centers andbecome permanent residents. At the Addis Ababa Fistu-la Hospital, many of these women become nursingaides, taking care of other fistula patients. In a uniquesocial experiment, that institution is developing a per-manent residential farm for the most severely injured ofthese women, in essence forming a new womenÕs com-munity based on shared experiences and the Çsiste-rhood of sufferingÈ that has resulted (E.C. Hamlin, per-sonal communication).

At the present time, there is no generally accepted, stan-dardized system for describing, classifying, or stagingobstetric fistulas. The situation remains much the sameas it was in 1951, when Bayard Carter and colleagueslamented, ÇIt is difficult to interpret and to comparemany of the reported series of fistulas. There is a realneed for standard methods to describe the fistulas, theactual operations, and results.È McConnachie (1958)eloquently summed up the problems surrounding dis-cussions of fistula repair in the following words:

ÇIt is common to find that each author has either usedhis own form of classification based solely on the ana-tomical structures involved, or the size of the fistula, oreven one of convenience. But no two writers seem tohave reached any common grounds for agreement otherthan that a urinary fistula is present. This lack of agree-ment has prevented comparison between individualworkersÕ material or results, based on a common scien-tific classification. Such classification needs to be fullyet simple, workable in its details so that it will not becumbersome, to be based on the three principal anato-mical structures involved in urinary leakage [bladder,urethra, ureters], and to take into account the clinicalaspects of site, size, accessibility and condition of thetissues involved in attempted repair.È

In describing their experience with fistula repair, manysurgeons have put forth various systems or rationalesfor grouping cases together. The earliest such systemwas that of Sims (1852) who classified fistulas accor-ding to their location in the vagina: 1) urethro-vaginalfistulas, where the defect was confined to the urethra; 2)fistulas situated Çat the bladder neck or root of the ure-thra, destroying the trigone;È 3) fistulas involving thebody and floor of the bladder; and 4) utero-vesical fis-tulas where the opening of the fistula communicatedwith the uterine cavity or cervical canal. Succeeding

authors have added to, modified, and re-grouped thecategories of fistulas in a variety of ways (Emmett1868, Mahfouz 1930; Thomas 1945; Krishnan 1949;Lawson 1968; McConnachie 1958; Moir 1961; Hamlinand Nicholson 1969; Hilton and Ward 1998; Waaldijk1994, 1995). Mahfouz (1930) listed the main points tobe considered in the evaluation of a fistula as follows:

¥ The situation, size, form and variety of the fistula

¥ The scarring of the vagina and its effect on the mobi-lity of the fistula

¥ The attachment of the fistula to the pelvic walls

¥ The condition of the urethral sphincter and permea-bility of the internal orifice of the urethra (occasio-nally the urethra is completely occluded by fibrosis)

¥ The location of the ureteral orifices, and their rela-tion to the edges of the fistula

¥ The presence of complications such as recto-vaginalfistula, inflammatory lesions of the pelvis, vagina,vulva or peritoneum

¥ The presence of more than one fistula

There is an urgent need to develop a standardized,generally-accepted system for describing and classi-fying fistulas, to aid communication between fistulacenters and to facilitate research in this field. The com-mittee has not taken upon itself the task of developingsuch as system, but recommends that the Standardiza-tion Committee of the International Continence Societymake this a priority in the future. The committee feelsthat the following factors should be taken into accountin the development of a standardized ICS ClassificationSystem:

¥ The system should provide a simple yet precisemethod of describing fistula location and size.Although the size of the fistula may be quite impres-sive in cases of prolonged obstructed labor, theactual size of the opening is probably less importantin the overall prognosis than other factors. As JohnSt. George noted in 1969, ÇThis [size] is not veryimportant as long as other factors are favourable. Ithas been found that an opening of 1 cm or less withunhealthy tissues, fixed to bone within a stenosedvagina, is worse than an opening of 5 cm or morewith lax vaginal walls, no fixity and a wide vaginalintroitus. Also, a small bladder neck fistula is foundto be more difficult and less successful at repair thana large mid-vaginal fistula.È

¥ The system should make some attempt to assess theimpact of the fistula on function; i.e. how does thefistula impact the bladder neck, urethra, ureters, or(in the case of rectovaginal fistulas) the anus and itssphincter mechanism?

V. THE CLASSIFICATION OFOBSTETRIC FISTULAS

914

¥ The presence, location, and degree of vaginal scar-ring should be quantified, as this appears to bedirectly related to the probability of successful fistu-la closure. Many authors have commented on thegrim prognosis for fistulas complicated by vaginalscarring. Das and Sengupta noted (1969) ÇMassivescarring is frequently present and is definitely a greathandicap for a successful vaginal operation. Annularconstriction ring around the vagina, distal to the fis-tula, with fixity to the pubic bones with cartilage-nous margins are very unfavourable signs. But, for-tunately it is noted that once the neighbouring bandsof adhesions are released, the bladder can be broughtdown easily and anatomic apposition made.È

Any classification system for fistulas should be basedon criteria that correlate with the prognosis for success-ful surgical repair. This will require looking at thewhole patient, rather than simply looking at just the fis-tula. In a comparative study of failed fistula repairs car-ried out in Addis Ababa, Kelly and Kwast (1993a)found a statistically significant association (p < 0.001)with failed fistula repair and a ruptured uterus, with aprevious history of failed surgical repair (especiallyrepairs carried out at other, non-specialist, institutions),with the presence of limb contractures, with the needfor special preoperative feeding in order to become fitenough for surgery, and with the presence of a Çcom-plicatedÈ fistula requiring more extensive and morecomplicated surgical intervention. These authors belie-ved that these factors reflected a more extensive injuryfrom prolonged obstructed labor. McConnachie (1958)suggested that failures in fistula surgery were due toprevious failed repairs with increased scar tissue for-mation, differences in the grade and type of fistulasreported in various series, to inexperienced operatorstaking on cases beyond their surgical capabilities, andÇresidual urinary sepsis and alkalinity.È Further compa-rative studies are needed to ascertain if these factors arereliably associated with surgical failure, and to determi-ne if other, as yet unascertained, factors are also relatedto a poorer surgical prognosis.

¥ There should be a standardized definition of Çsuc-cessÈ in fistula surgery. In his series of 303 fistulas,McConnachie (1958) refused to regard the patient asÇcuredÈ even though the fistula was ÇclosedÈ unlessÇshe also has complete urinary continence andcontrol.È This principle reduced his overall success(Çclosed and dryÈ) rate from 79.5% to 65.1%. Per-haps as many as one third of women who undergosuccessful fistula closure still have significant incon-tinence due to the presence of sphincteric damageand/or other persistent alterations in bladder or ano-rectal function. To classify as a ÇsuccessÈ a womanwho has a closed fistula but continuous, debilitatingstress incontinence from a non-functioning urethra isboth dishonest and clinically unhelpful.

¥ In describing surgical success rates, the time atwhich this assessment is made must be clearly spe-cified. Long-term outcome data on women who haveundergone fistula repair surgery are almost entirelylacking in the literature. Given the difficult circum-stances from which most fistula patients come, andthe difficulties of travel and communication in ruralAfrica, this is not surprising. As a result of these pro-blems, most ÇsuccessesÈ are evaluated as such onlyat the time of discharge from hospital, with almostno further follow-up. This may well be misleading asto actual outcome. Coetzee and Lithgow (1966) defi-ned ÇcureÈ of a patient with a vesicovaginal fistulaas follows:

ÇFor a 100% cure the following conditions must befully satisfied: 1) The patient should have completecontinence by day and by night, and to achieve this ithas been found that the bladder should hold a minimumof 170 ml. 2) No stress incontinence should be present.3) The vagina should allow normal coitus without dys-pareunia. 4) Traumatic amenorrhea should not result.5) The patient should be able to bear children.È

¥ Achieving standards this high in patients who haveundergone extensive pelvic trauma from obstructedlabor will not be easy, but honesty is the first prere-quisite if progress is to be made with true scientificintegrity.

¥ General agreement should be reached concerningwhat constitutes an Çirreparable fistulaÈÑthat is,which women should be offered urinary and/or colo-nic diversion as their first surgical procedure, ratherthan undergo an attempt at fistula closure? The com-mittee recommends that the term Çirreparable fistu-laÈ be used in preference to the phrase Çhopeless fis-tula,È which is sometimes encountered in these dis-cussions.

¥ Any system for the classification or staging of obs-tetric fistulas must emphasize low-technology clini-cal assessment, since such a system will have to beused in areas of the world with extremely limitedresources. Fistulas afflict the worldÕs poorestwomen, not the affluent elites of industrialized eco-nomies.

¥ In view of the multi-system pathology produced byobstructed labor, any classification system shouldnote the presence of associated conditions related tothe primary pathophysiological process in obstructedlabor, such as the presence of neurologic injury,damage to the pubic symphysis, the concurrent pre-sence of a recto-vaginal fistula, the presence of chro-nic skin ulcers, nutritional state, etc.

¥ Because most of the suffering endured by fistula vic-tims is the result of social isolation and abandonment

915

and the subsequent loss of self-esteem and economicdeprivation that results from this social isolation, theevaluation of women with obstetric fistulas shouldinclude information on these social Çco-morbidi-ties.È The fistula problem in the developing world isinextricably intertwined with its social milieu, andthis must be openly acknowledged.

Although many patients with vesico-vaginal fistulasappear for care many years after they have sustainedtheir injury, others arrive at a medical facility in obs-tructed labor, or shortly after the fistula appears. Thetraditional teaching has been that an attempt at repairshould be deferred for three months until the extent ofthe injury has fully manifested itself and any infec-tion/inflammation in the injured tissues has resolved.However, it also appears to be true that some fistulasmight be prevented by prompt treatment of women whoarrive after obstructed labor, that some fistulas mightclose spontaneously if the bladder is drained for a pro-longed period of time, and that a subset of fistulasmight even be amenable to early closure. John St. Geor-ge (1969) believed that prompt medical treatment ofpatients after obstructed labor could prevent post-par-tum sepsis and promote the healing of injured tissues,thus preventing some fistulas altogether, or at leastminimize the extent of injury that developed. He advo-cated vigorous local care of the injured tissues andprompt antibiotic treatment as soon as such patientswere seen. He described one particularly noteworthycase:

ÇThe patient, who was aged 18, was brought to the hos-pital three days after delivery, following labour whichlasted seven days. She was incontinent of urine andfaeces. She was very pale (haemoglobin concentration3.7 g per 100 ml), dehydrated, toxic and unable to walk.Vaginal examination revealed a gaping opening withoffensive purulent lochia and discharge. The lateralwalls of the vagina had sloughed, revealing both pubicrami and the right ischial tuberosity. The perineum wastorn, leaving an indefinite anal sphincter. A large vesi-co-vaginal fistula and a small recto-vaginal fistula,both with infected edges, were seen. Apart from bloodtransfusions, parenteral antibiotics and high proteindiet were given: the vagina was douched twice dailywith warm Dettol solution; this was followed each timeby insertion of layers of ofra tulle gauze. After threeweeks, the local treatment was given only once daily foranother three weeks. When the patient was examinedunder anesthesia ten weeks after admission, and beforedischarge from hospital, the vaginal walls had healed

over and returned to normal; the perineum had healed,leaving only a wide vaginal introitus; the recto-vaginalfistula had healed and the vesico-vaginal fistula wasreduced to 1 cm diameter with mobile surroundingedges. She returned for repair three months later, andthis was successfully carried out. Prevention of sepsiswas the chief factor in such a transformation as occur-red in this case, as in many others, and repair wouldotherwise have failed.È

Further research on the effectiveness of immediatelocal care after fistula formation should be encouraged.Waaldijk (1994) found that 50 - 60% of smaller fistulas(< 2 cm) would heal spontaneously, if prompt prolon-ged bladder drainage was started within three months ofthe initial injury. If the fistula did not close with simplecatheter management, he performed an early closure.Using this regimen, he reported a successful closurerate in 92% of cases, with continence in 94% of thosein whom the fistula was closed successfully. Theseresults are encouraging, but await replication at othercenters.

In general, specific issues of surgical technique areamong the least important issues in obstetric fistularepair today. In practiced hands, skilled fistula surgeonsroutinely achieve fistula closure rates of 80% or better.Multiple papers reporting large case series support thiscontention (Abbott 1950; Amr 1998; Arrowsmith 1994;Ashworth 1973; Aziz 1965; Bird 1967; Coetzee andLithgow 1966; Gharoror and Abedi 1999; Goh 1998;Hamlin and Nicholson 1966, 1969; Hilton and Ward1998; Iloabachie 1992; Kelly 1992; Krishnan 1949;Lavery 1955; Lawson 1972; Mahfouz 1929, 1930,1938, 1957; Moir 1966; Mustafa and Rushwan 1971;Naidu 1962; Waaldijk 1989; Yenen and Babuna 1965).There is a fairly general consensus concerning the basicprinciples of fistula repair, which can be summarized asfollows:

¥ The best chance for successful fistula closure is atthe first operation. In their large series of 2,484 fis-tula patients, Hilton and Ward (1998) reported suc-cessful fistula closure in 82.8% of patients at the firstattempt. Successful closure was achieved in only65% of those patients who required two or more ope-rations.

¥ The fistula should be widely mobilized from the sur-rounding tissues at the time of repair, so that fistulaclosure can be achieved without tension on the siteof repair.

VII. SURGICAL TECHNIQUE FORFISTULA CLOSURE

VI. EARLY CARE OF THE FISTULAPATIENT

916

¥ The repair must be Çwater-tightÈ at the time of clo-sure. If it is not, failure is virtually certain. The sim-plest way to test this is to instill a solution of coloredwater into the bladder at the time of fistula closureand make certain that no leakage can be demonstra-ted. If leakage occurs, the repair should be takendown and repeated or reinforced until no leakage canbe demonstrated. As J. Marion Sims wrote in his ori-ginal paper on fistula repair (1854), Çif a single dropof urine finds its way through the fistulous orifice, itis sure to be followed by more, and thus a failure tosome extent is almost inevitable.È

¥ The fistula should be closed in multiple layers, avoi-ding over-lapping lines of suture, whenever this canbe achieved.

¥ After fistula repair, the bladder should be emptied byprolonged continuous catheter drainage in order toprevent distention of the bladder and increased ten-sion on the suture lines. The traditional duration ofbladder drainage is 14 days, but no comparativetrials have been carried out to see if shorter durationof bladder drainage (for example, 7 or 10 days) isassociated with increased risk of failed repair.Research on the optimal duration of bladder draina-ge is important, and has obvious consequences forfistula centers with large clinical volumes. If, forexample, the duration of post-operative catheteriza-tion could be decreased from 14 days to 10 dayswithout a significant increase in failure rates, thecenter could increase the number of fistula patientsundergoing surgical repair by almost 30%.

¥ Especially in the case of fistulas due to prolongedobstructed labor where the injury is the result of pro-longed ischemia and tissue necrosis, successful clo-sure is enhanced by the use of tissue grafts (bulbo-cavernosus graft, gracilis muscle graft, etc.) whichbring a new blood supply to the site of repair. Thereis one retrospective paper in the literature that eva-luated the repair of comparable fistulas with andwithout the use of Martius bulbocavernosus flapsand demonstrated substantially higher rates of suc-cessful closure when such a graft was employed(Rangnekar et. al. 2000).

Each fistula is unique, and an ability to improvise in theface of unexpected findings or complications is a virtuethat every fistula surgeon must strive to develop. It isclearly not possible to illustrate here every differenttype of fistula and all of the various techniques that maybe employed to close them. However, a generallyaccepted techniqueÑ-based largely on the work of Drs.Reginald and Catherine Hamlin at the Addis Ababa Fis-tula Hospital in Ethiopia where nearly 20,000 obstetricfistulas have now been repairedÑ can be described andillustrated as follows [Figures 10 - 29]. The first prere-

quisite for successful fistula repair is meticulous atten-tion to detail. As Abbot aptly (if somewhat quaintly)noted (1950), ÇThere must be no attempt to operate onthese cases with one eye on the clock and the other onthe tea wagon. In fact, the operator upon vesico-vaginalfistulae should combine the traits of daintiness, gentle-ness, neatness and dexterity of the pekinese, with thetenacity and perseverance of the English bulldog.È

The position for fistula surgery depends upon the natureand location of the fistula to be repaired. For the vastmajority of straightforward fistulas (especially for mid-vaginal fistulas) a high lithotomy position with the but-tocks pulled well over the edge of the operating table,provides excellent exposure [Figure 10]. Surgery in thisposition is easy to perform under spinal anesthesia,which is the cheapest and easiest form of anesthesia forÇlow technologyÈ settings in developing countries. Tooperate in the knee-chest position is relatively uncom-fortable for patients and can compromise pulmonaryfunction. Performing operations in the knee-chest posi-tion generally requires intubation of the patient, the useof general anesthesia, and continuous ventilation. Trans-abdominal surgery with the patient in the supine posi-tion is rarely needed, except for certain complex fistulas.An abdominal approach increases both the cost of sur-gery and the likelihood of complications, such as woundinfections. When doing fistula surgery in the developingworld, failure to use a trans-vaginal approach requiresspecial justification, such as cases in which additionalintra-abdominal pathology must be addressed.

The first requirement for successful fistula repair isadequate exposure of the operative field. Figure 11

917

Figure 10 : Exaggerated lithotomy position for fistula repair.The patient is positioned at 35 - 45 degrees, head down posi-tion, with the use of shoulder supports and the buttocks pul-led over the edge of the table. (©Worldwide Fund for MothersInjured in Childbirth, used by permission).

depicts a typically narrow, scarred vagina of the typethat often develops after obstructed labor. The vesico-vaginal fistula is not clearly visible as it is obscured byscar tissue and a constriction ring. When this scar tissueis released by performing vaginal relaxing incisions,the fistula can be seen and repaired. The fistula depic-ted here is a straightforward mid-vaginal vesico-vagi-nal fistula.

It is generally preferable to identify and catheterize theureters in most operations so that they are readily iden-tifiable throughout the course of the operation. This canusually be done by passing catheters through the fistu-la into the ureters under direct vision [Figure 12]. The

purpose of ureteral catheterization is ensure that theureters are not inadvertently ligated during the fistularepair, with subsequent renal damage or death. (Thiswas the cause of death in SimsÕ only operative fatality;unfortunately, it occurred in a very public demonstra-tion of his technique for fistula closure in London at theSamaritan Hospital; see McKay 1922:518). Althoughsome fistula surgeons prefer to leave ureteral cathetersin place for up to 14 days after surgery, current Westernurological practice would suggest that such catheterscan be removed immediately at the end of the case, orwithin a day or two after surgery at most.

Once the fistula is exposed and the ureters are identi-fied, it is important to mobilize the fistula fully so thatit may be closed without tension. Figure 13 demons-trates the first move in mobilizing a mid-vaginal fistu-la. The posterior border of the fistula is incised and theincision is carried out laterally onto the vaginal side-walls. The incision extends only through the vaginalepithelium, not into the bladder itself. Wide vaginal dis-section will allow complete mobilization of the fistula.

Following the initial vaginal incision, the posteriorvaginal flap is developed, always keeping the course ofthe ureters in mind [Figure 14].

Continued mobilization of the fistula is achieved byextending the incision circumferentially around the fis-tula, then anteriorly towards the urethra [Figure 15]. Asbefore, the incision extends only through the vaginalepithelium, not into the bladder. When this portion ofthe operation has been completed, the bladder shouldbe freed completely from the vagina. When this hasbeen accomplished, the ureteral catheters can be passedthrough back into the bladder and brought out throughthe urethra to keep them out of the operative field.

The anterior vaginal flaps then are developed widely tomobilize the fistula more completely [Figure 16]. Dueto the presence of scarring and tethering of the fistula,it is generally useful to carry the anterior dissectionupwards behind the pubic symphysis, opening theretropubic Space of Retzius and detaching the bladderfrom its supports in this area. This allows full mobiliza-tion of the fistula [Figure 17]. The anterior vaginal flapscan be sutured out of the operative field using Çstaysutures.È

Additional tension may be taken off the fistula at thispoint by performing a partial bladder suspension bypassing sutures from the bladder up behind the pubicsymphysis and anchoring the sutures in the symphysealperiosteum [Figure 18]. Persistent stress incontinence isrelatively common in patients after otherwise success-ful fistula closure. It has been argued that suture place-ment of this type reduces the prevalence of this problempost-operatively (Hudson, Hendrickse, and Ward 1975;

918

Figure 11 : Placement of labial stay sutures helps increaselateral exposure. Incisions through bands of scar tissue areoften necessary to allow insertion of retractors. The upperillustration shows the vagina prior to relaxing incisions.The lower illustration shows enhanced exposure afterrelaxing incisions. (©Worldwide Fund for Mothers Injuredin Childbirth, used by permission).

919

Figure 12 : Ureters are identified and cannulated, if pos-sible. The use of intravenous diuretics and/or vital dyes canbe used to improve ureteric localization, if needed.(©Worldwide Fund for Mothers Injured in Childbirth, usedby permission).

Figure 13 : Dissection of the fistula begins at its posteriormargin by making an incision that runs from one vaginalsidewall to the other. Accentuation of tissue planes in somecases may be aided by local infiltration of saline or localanesthetic. (©Worldwide Fund for Mothers Injured inChildbirth, used by permission).

Figure 14 : Complete mobilization of the fistula is conti-nued by raising posterior vaginal flaps. The mobilization iscarried out to each vaginal sidewall as well as to the cervix(if one is still present). Special care must be taken at the 4and 7 oÕclock positions to avoid possible injury to the ure-ters. (©Worldwide Fund for Mothers Injured in Childbirth,used by permission).

Figure 15 : The mobilizing incision is extended circumfe-rentially around the fistula. When this is completed, ananterior midline incision is made in the vagina and exten-ded up towards the external urethral meatus. (©WorldwideFund for Mothers Injured in Childbirth, used by permis-sion).

920

Figure 16 : The anterior vaginal flaps should be mobilizedall the way from the edge of the fistula to the pubic arch.(©Worldwide Fund for Mothers Injured in Childbirth, usedby permission).

Figure 17 : Once the anterior flaps have been fully develo-ped, the endopelvic fascia is perforated and the space ofRetzius is entered, to complete the full mobilization of thefistula. This portion of the operation is often technicallychallenging. The apices of the anterior vaginal flaps can beretracted out of the operative field with stay sutures toimprove visualization of the fistula. At this point, the blad-der base and vagina should be completely separated fromone another. It is generally not necessary (or advisable) totrim or Òpare backÓ the edges of the fistula. This onlyincreases the size of the defect and decreases the amount ofremaining bladder tissue. (©Worldwide Fund for MothersInjured in Childbirth, used by permission).

Figure 18 : Closure of the fistula begins by placing ancho-ring sutures lateral to the bladder defect and superiorlythrough the periosteum of the pubic arch. The sutures aredesigned to decreases tension across the line of bladder clo-sure. (©Worldwide Fund for Mothers Injured in Childbirth,used by permission).

Hassim and Lucas 19740); however, no controlled stu-dies have yet been carried out that document this asser-tion.

Once the fistula has been mobilized as fully as possible,fistula closure can begin. There are many different tech-niques by which this task may be accomplished. Oneproven technique is to close the first layer of the blad-der with a continuous, running, interlocking stitch of anabsorbable suture [Figure 19]. The initial suture biteshould be placed beyond the lateral margin of the fistu-la and the last suture should be placed in a similar posi-tion on the opposite side of the fistula. Although Simsinitially Çpared backÈ the edges of the fistula in his clo-sure technique, there seems to be no need to do this inthe vast majority of obstetric fistulas: a better philoso-phy is to preserve as much bladder tissue as possible,particularly when attempting to close extensive fistulas.

After initial fistula closure has been completed, the pri-mary suture line should be reinforced by a second layerof closure if at all possible. Ideally this should be donein such a fashion that the second row of sutures imbri-cates the initial closure, rolling more bladder tissueover the first line of closure in order to protect it [Figu-re 20]. When the second layer has been closed, the inte-grity of the repair should be checked by instilling 150 -250 ml of water colored with indigo carmine, methyle-ne blue, or another suitable dye. If there is no leakageof colored water, the fistula repair can be assumed to beÇwater tight,È and the operation can proceed. If leaka-

ge is observed, the repair should be taken down andrepeated until no more leakage is observed.

Extensive experience with fistula repairs has led manysurgeons to believe that successful fistula closure ismarkedly enhanced by the use of a bulbocavernosus fatpad (Martius) graft (Hamlin and Nicholson 1969;Elkins et. al 1990; Baines et. al. 1976; Punekar et. al1999; Rangnekar et. al. 2000; Given and Acosta 1989;Fitzpatrick and Elkins 1993; Zacharin 1980; Shaw1949; Martius 1956). There is one small comparativesurgical study that documents this finding (Rangnekaret. al. 2000).

Development of the fat pad graft should begin with avertical midline skin incision on the left or right labiummajus, extending from the base of the mons pubis toabout the level of the middle of the vaginal introitus[Figure 21]. Sharp dissection with a surgical scissors isused to expose a central ÇcordÈ of labial fat, and thisdissection is carried down to the deep fascial layer[Figure 22]. Even in thin, malnourished women, dis-section of this fat is always possible. Once this ÇcordÈof fat has been identified and dissected, it is cross-clam-ped superiorly with a single clamp and transected[Figure 23]. The superior stump is suture-ligated toachieve hemostasis.

Next, the superior aspect of the graft is grasped gentlywith an Allis forcep and the cord of fat is further mobi-lized down to its base. It is important to protect the baseof the pedicle to insure that its blood supply remainsintact [Figure 24]. At this point, surgical scissors areused to dissect a tunnel which extends from the base ofthe fat padÕs pedicle into the vagina, traveling betweenthe vaginal epithelium and the pubic arch [Figure 25].The tunnel should be large enough to allow passage ofthe surgeonÕs finger from the labial defect down to thebladder base [Figure 26].

Prior to passing the graft through the tunnel, fouranchoring stitches of absorbable suture should be pla-ced into the muscularis of the bladder at the 2,4, 8, and10 oÕclock positions. This allows the Martius graft to beanchored into place securely against the repaired fistu-la, protecting it and bringing in a new blood supply tonourish the surgical site [Figure 27]. The graft is thenpassed through the tunnel to the bladder base and isanchored into position using the previously placedanchoring stitches. The fistula site should be complete-ly covered at this point [Figure 28].

Finally, the stay sutures which had been holding theanterior vaginal flaps out of the surgical field are relea-sed and the vaginal defect is closed as an Òinverted TÈ[Figure 29]. The ureteric catheters can now be removedand a vaginal pack placed, if desired.

921

Figure 19 : Fistula closure. The bladder defect is closed intwo layers using absorbable suture. If any tension isencountered during bladder closure, the initial dissectionhas been inadequate and should be revised. The first closu-re can be made with either interrupted sutures or a conti-nuous running, interlocking, suture. (©Worldwide Fundfor Mothers Injured in Childbirth, used by permission).

922

Figure 20 : Fistula closure. A second set of sutures is pla-ced, imbricating the second suture line over the first. Whenthis has been completed, Çwater tightÈ closure of the fistulashould be confirmed by gentling filling the bladder with asolution of colored water or sterile infant feeding formula tocheck of leakage. If any leakage is noted, the repair shouldbe taken down and reclosed until no leakage can bedemonstrated. (©Worldwide Fund for Mothers Injured inChildbirth, used by permission).

Figure 22 : Martius flap. Lateral and medial dissection ofthe bulbocavernosus fat is made down to the deep fasciallayer, exposing a ÒcordÓ of labial fat. (©Worldwide Fundfor Mothers Injured in Childbirth, used by permission).

Figure 23 : Martius flap. Once the ÒcordÓ of fat has beenidentified and dissected, it is cross-clamped superiorly anddivided. The superior stump is suture-ligated to insurehemostasis. (©Worldwide Fund for Mothers Injured inChildbirth, used by permission).

Figure 21 : Martius bulbocavernosus fat flap. A verticalskin incision is made in the labium majus from the base ofthe mons to the level of the middle of the vaginal introitus.(©Worldwide Fund for Mothers Injured in Childbirth, usedby permission).

923

Figure 24 : Martius flap. The superior aspect of the graft isgently grasped with an Allis forcep and the labial fat is fur-ther mobilized down to its base, which is left intact to main-tain its blood supply. (©Worldwide Fund for Mothers Inju-red in Childbirth, used by permission).

Figure 25 : Martius flap. A tunnel is dissected from the baseof the labial fat graft into the vagina, running between thevaginal epithelium and the pubic arch.. (©Worldwide Fundfor Mothers Injured in Childbirth, used by permission)..

Figure 27 : Martius flap. Prior to passing the graft throughthe tunnel, four anchoring stitches of absorbable materialare placed into the muscularis of the bladder at the 2, 4, 8,and 10 oÕclock positions. (©Worldwide Fund for MothersInjured in Childbirth, used by permission).

Figure 26 : Martius flap. The final caliber of the tunnelshould be large enough to allow passage of the surgeonÕsfinger from the labial defect down to the bladder base.(©Worldwide Fund for Mothers Injured in Childbirth, usedby permission).

1. THE FISTULA COMPLICATED BY URETHRAL

DAMAGE

Virtually all authors with extensive experience in themanagement of obstetric fistulas comment on the greatdifficulty in achieving post-operative continence inpatients who have had extensive damage to the urethra,even if the fistula defect itself can be closed successful-ly. J. Chassar Moir, the great British gynecologist,referred (1965) to the worst of these cases as Çcircum-ferentialÈ fistulas, which Çinvolve a destruction of thebladder neck not only on the vaginal side but, in manyinstances, on the pubic side as well. The result is a cir-cumferential sloughing with subsequent discontinuityof the urethra and bladder; the intervening tissue ismerely the epithelium that has grown over, and becomeadherent to, the periosteum of the back of the pubicbone.È Fistulas involving this level of destruction aredaunting, and are rarely seen in developed countries.According to Moir, the three great problems involved indealing with this type of fistula are:

1. extremely difficult exposure;

2. technical difficulty in dissecting the tissue remnantsfrom the pubic bone; and

3. difficulty in joining the bladder neck to the urethralremnant or stump, if, indeed, any portion of the urethrais still intact.

The basic principles of technique needed to deal withthis type of injury are complete mobilization of thebladder so that it can be drawn down low enough tocreate a tension-free anastomosis with the urethral rem-nant. Freeing the urethral remnants from their adheren-ce to the pubic bone may require a suprapubic incisionwith dissection from above in order to accomplish this.In such cases Moir took care to reinforce the bladderneck with buttressing sutures, and generally brought ina MartiusÕ graft for better support and a renewed bloodsupply.

If only the posterior portion of the urethra had beensloughed and the anterior portion of the urethra wasintact, Moir (1964) advocated a different technique forurethral reconstruction. In this technique, a thin cathe-ter was stitched into position to serve as a splint for thenew urethra. The margins of the urethral bed were freedfrom the vagina and were mobilized to allow them to bepulled together over the underlying catheter withouttension. The vaginal incision was extended above the

VIII. COMPLICATED CASES ANDTECHNICAL SURGICAL

QUESTIONS

924

Figure 28 : Martius flap. The graft is then brought throughthe tunnel and anchored securely into place against therepaired fistula. The site of fistula repair should be comple-tely covered at this point. (©Worldwide Fund for MothersInjured in Childbirth, used by permission).

Figure 29 : Vaginal closure. The stay sutures holding thevaginal flaps are released and the vaginal defect is closed asan Òinverted TÓ using absorbable sutures. (©WorldwideFund for Mothers Injured in Childbirth, used by permis-sion).

bladder neck and then reinforced with Çinrollingstitches.È The repair was usually buttressed with a Mar-tius bulbocavernosus fat graft, after which the vaginawas closed over the repair with vertical mattress suturesto achieve a Çbroad appositionÈ of the vaginal wall.The bladder was then drained for 10 days and the vagi-nal sutures were removed after 21 days. Although hereported good success with this technique, with 23 of34 women (67%) having Çperfect or near perfectcontrolÈ six months after surgery, 8 of 34 (24%) hadpersistent stress incontinence, and 9% had no improve-ment, reconfirming the view that persistent stressincontinence remains a significant problem for manywomen after successful fistula closure. Similar tech-niques with similar results have been reported by otherauthors (Noble 1901; Symmonds 1969; Symmonds andHills 1978)

Various authors have described neourethral reconstruc-tion using bladder flaps (Barnes and Wilson 1949;Flocks and Culp 1953; Su 1969; Quartey 1972; Tana-gho and Smith 1972). All of these operations are basedupon transabdominal techniques; however, a transvagi-nal approach to neourethral reconstruction using ananterior bladder flap technique was described byElkins, Ghosh and co-workers (1992). In this tech-nique, a neo-urethra is created by mobilizing a flapfrom the anterior bladder, which is then rolled into atube. In this technique, the anterior and lateral edges ofthe fistula are freed up and the space of Retzius is ente-red transvaginally beneath the pubic bone. The anteriorbladder is then pulled down into the vagina and mobili-zed. A 3 cm incision is made into the bladder and theanterior bladder wall is then rolled around a 16 Fr.Foley catheter to create a tube. After this is tackeddown, a similar incision is made on the other side tocomplete mobilization of the tube. The anterior surfaceof the neourethra is then sutured in two layers and theposterior edge of the fistula is closed transversely, alsoin two layers. The neourethra is reattached to the poste-rior edge of the pubic symphysis, and a Martius graft isplaced, before reapproximating the vaginal epithelium.This technique resulted in successful closure of the fis-tula in 18 of 20 cases, 4 of whom had severe stressincontinence post-operatively.

Based on their extensive experience with fistulas inAddis Ababa, Ethiopia, in 1969 Hamlin and Nicholsonintroduced the concept of the Çdifficult urinary fistulaÈto describe the complicated aspects of the problem tou-ched on by Moir. According to them, the Çdifficult fis-tulaÈ

ÇÉis a complex of several grave injuries occurringtogetherÑ-namely, a) total destruction of the urethra(all walls), the remaining tissue being merely fibrousconnective tissue and squamous epithelium which hasgrown over and become adherent to the periosteum on

the back of the pubic bones; b) an extensive sloughingof the bladder neck and trigone sometimes so large asto cause one or both ureteric orifices to open directlyinto the vagina; and c) fibrosis to an incredible degreewhich 1) narrows the vagina to the diameter of one fin-gerbreadth, and 2) binds the remains of the bladderhigh up to the descending pubic rami and to the pubicsymphysis. In a word, no part of the patientÕs lower uri-nary tract has escaped some degree of damage. This isthe fistula which daunts the hearts of most observerswho see it for the first time. É The gynaecologist boldenough to attempt the classical flap-splitting operationfor a case like this soon discovers that he is operatingin an area as confined and almost as inaccessible as theinside of the toe of a leather shoe. He will find himselffreeing the bladder of scar and the lateral fixation of itstorn edges by touch only. É Within the vagina nothingexists É except, almost quite literally, skin and bone.È

In such cases, Hamlin and Nicholson recommendedconstructing a new urethra by creating a new ÇinnerÈurethra using the skin and fibrous connective tissuecovering the pubic bones and the inferior border of thepubic symphysis. In this technique, two lateral verticalincisions about 2 cm part are made in the skin, and leftand right skin flaps were then created and reflectedmedially until their edges could be joined togetherwithout tension in the midline underneath the urinarycatheter that had been placed in the bladder. When joi-ned, these fragile skin flaps are rolled into a tube. Atthis point the neourethra is a fragile, untenable creation.As the authors noted, ÇSuch a fragile neourethra, stan-ding unsupported, would almost certainly necrose, andeven if it survived would not restore any worthwhiledegree of function.È The neourethra was then reinfor-ced using a gracilis muscle flap taken from the thigh,preserving its neurovascular pedicle. The gracilis ten-don is pulled through a tunnel in the thigh that crossesthe ischiopubic ramus at the level of the urethra and isguided into the vagina, under the pubic symphysis, andis sutured to the anterior lip of the cervix, the lateralvaginal fascia, and the fibrous connective tissue cove-ring the periosteum of the ischiopubic rami and thepubic symphysis. Once this has been accomplished,additional grafting is necessary using a Martius flapwhich is then covered with skin flaps.

Using this technique, the authors reported no deathsand only one Çcomplete failureÈ in 50 operations, thiscase being due to failure of the blood supply to the gra-cilis muscle flap. In some cases small urethro-vaginalfistulas remained, which were repaired at a subsequentoperation. Surprisingly, only 8 women (16%) develo-ped ÇsevereÈ stress incontinence after this reconstruc-tion, four of whom regained ÇsatisfactoryÈ continenceover time, and four of whom required an operation forstress incontinence. In the latter four patients, only two

925

of these operations were completely successful. Sixpatients (12%) developed a urethral stricture, three ofwhich were successfully treated by passage of a soundand three of which required surgical correction. Theremaining 35 patients (70%) were discharged homewithin six weeks of surgery clinically cured or withmild residual stress incontinence which did not appearto be clinically bothersome for them.

Complete urethral loss from obstructed labor remains adaunting surgical challenge, to which an ideal solutionhas yet to be found.

2. URINARY DIVERSION FOR THE IRREPARABLE

FISTULA

Although there is a general consensus in the literaturethat some fistulas simply cannot be repaired with resto-ration of full continence, there is no general agreementas to which fistulas should be treated initially by pri-mary urinary diversion rather than an attempt at fistulaclosure. Similarly, there are no accepted criteria in fai-led cases to dictate when further attempts at closureshould be abandoned and the patient should be offeredsome form of urinary diversion as a treatment. Thedilemma has been summarized by Hodges (1999):

Many centres which perform VVF repairs have a smallgroup of ÇproblemÈ patients who have failed to gaincontinence despite often repeated attempts by differentsurgeons. Urinary diversion, with all its disadvantages,is seen as an admission of defeat by the VVF surgeonand so this decision is reached reluctantly. This reluc-tance is compounded by the concern of performingsuch a major procedure in often basic conditions.However, in the best interests of the patient, eventuallythe experienced VVF surgeon must admit when allattempts to gain continence have failed and considerurinary diversion.È

Because urinary diversion tends to be a Çhigh technolo-gyÈ approach to fistula management, its use in coun-tries that do not have a well-developed nursing infra-structure to support the ongoing care of such patientssuggests that this technique should be used with extre-me caution. For example, transplantation of the uretersinto an ileal conduit requires the use of an external col-lecting device. The use of such appliances may well beunacceptable in the local culture and patients are likelyto experience significant difficulty in obtaining suitableexternal appliances and may have trouble performinggood stoma care. The result of such a policy could wellbe simply to transpose the fistula from the vagina to theabdomen! Likewise, if continent urinary diversions areperformed with the creation of a catheterizable stoma,the problem of clean intermittent self-catheterizationremains. This can be compounded by loss of the cathe-

ter or the development of stomal stenosis, with urinaryretention, reservoir breakdown, sepsis, and death.Hodges (1999) has reported a series of seven patientswith intractable fistulas who were treated in Uganda bycontinent urinary diversion using a Mitrofanoff proce-dure in which the appendix is mobilized as the cathete-rizable stoma. There was one death 6 days after surge-ry, apparently from coincidental complications ratherthan as a direct result of the operative technique. Theother six patients were reported as doing well up to 14months after surgery. Because of the possibility of thecomplications already alluded to, Çpatients are encou-raged to remain near a hospital which can deal with anylikely complications; five of the six patients have takenthis advice. All patients carry spare catheters and a let-ter clearly explaining the nature of the procedure andthe emergency treatment if there is a complication.ÈWhether these arrangements would be suitable foremployment in different environments in other coun-tries remains unknown.

The ÇtraditionalÈ operation for dealing with the irrepa-rable fistula has usually been uretero-sigmoidostomy inwhich the ureters are transplanted into the sigmoidcolon, which then functions as a reservoir for both urineand feces. The long-term consequences of this opera-tion, such as anastomotic leakage with peritonitis, ure-teral stenosis and hydronephrosis, acute and chronicpylonephritis, electrolyte imbalances, diarrhea, long-term renal failure, and the development of adenocarci-noma of the colon at the site of ureteral implantation,are well-known. These complications should give com-passionate surgeons pause before recommending opera-tions of this kind.

Several small series of fistula patients who have under-gone ureterosigmoidostomy have been reported in theliterature (Attah and Ozumba 1993, Humphries et. al.1961, Thompson 1945, Foda 1959). These series aresomewhat difficult to evaluate, as the numbers ofpatients reported are small and follow-up is often unsa-tisfactory. For example, Das and Sengupta (1969)resorted to urinary diversion and ureteral transplanta-tion in 20 of 135 fistula patients that they reported inone series from India. Of these patients, two died ofuremia after the surgeryÑone of the 9th and one on the21st post-operative day after a stormy post-operativecourse. They did not know the follow-up of the otherpatients. Case fatality rates as high as 38.5% have beenreported following these procedures in developingcountries (Thompson 1945); although such grim statis-tics tend to come from the older surgical literature, theconditions of surgical practice and the facilities avai-lable for patient care may not have changed appreciablyin those countries in which fistulas are still prevalent.

In recent years it has been discovered that many of the

926

disadvantages of ÇclassicalÈ ureterosigmoidostomycould be overcome by modifications of technique, theso-called ÇMainz IIÈ pouch (Fisch et. al 1993; Gerharzet. al. 1998). In this modification the anterior colon isopened 12 cm distal and proximal to the rectosigmoidjunction, and a side-to-side anastomosis is made. Thisdetubularization of the bowel reduces the force of colo-nic contractions and creates a Çlow pressureÈ systemthat does not appear to predispose such patients to thedevelopment of hydronephrosis. The bowel is attachedto the sacral promontory for stabilization, and the ure-ters are mobilized bilaterally and transplanted into thecolon through a 4-5 cm submucosal tunnel on eachside. Aside from the continuing need to monitor suchpatients for the development of hyperchloremic meta-bolic acidosis and the possibility of the growth of mali-gnant polyps at the uretro-colonic implantation site,patients undergoing the Mainz II procedure have donequite well, and almost all have had socially acceptableurinary continence and improved body image compa-red to patients undergoing urinary diversion using anileal-conduit.

Good long-term follow-up studies of patients who haveundergone urinary diversion of this type for irreparablefistulas in developing countries are virtually non-exis-tent. If ureterosigmoidostomy is contemplated for a fis-tula patient, it is essential that she have normal analsphincter control pre-operatively. This can be assessedby giving a large volume enema as a continence test tosee how long she can retain it. It is also essential thatpatients be given enough time and counseling tounderstand the possible consequences of this type ofsurgery for their lives and future health. Treatment ofintestinal parasites and a thorough bowel-prep prior tosurgery are obvious essential components of this typeof therapy. The optimal use of such techniques in fistu-la patients remains to be determined; however, it isclear that Çsalvage therapyÈ of this type will continue tohave a place in the treatment of some patients with obs-tetric fistulas for many decades to come.

The ultimate strategy for dealing with obstetric fistulasshould be to prevent them entirely. Indeed, this is pre-cisely how the Western world solved the problemwithin its borders. Because obstetric fistulas are tiedclosely to overall maternal mortality, the best way toreduce fistula formation is to provide essential obstetricservices at the community level with prompt access toemergency obstetric services at the first referral level.The success of this strategy has been amply demonstra-ted by Loudon (1992), by Maine (1991) and the essen-

tial elements of such obstetric care have been elabora-ted in some detail by the World Health Organization(1986). Figure 30 shows the historical trends in mater-nal mortality in the United States, England and Wales,and the Netherlands since 1920 (Loudon 1992a). At thebeginning of the 20th Century, maternal mortality inWestern Europe and North America was similar to thatwhich currently exists in the developing world. Theintroduction of antibiotics, blood transfusion, safeCesarean section, better transporation and improvedaccess to care, along with the professionalization ofobstetric care and midwifery services, led to a dramaticand continued decrease in maternal mortality in allthese countries (Loudon 1992a, 1992b, 2000). Duringthe same period the obstetric fistula virtually vanishedfrom the experience of the industrialized world. Elimi-nation of obstetric fistulas from the developing worldwill require that their health care systems undergo asimilar transformation.

The committee believes that the fundamental compo-nents of such a program include the following:

¥ Promotion of breastfeeding and the elimination ofchildhood infections which hamper growth, such asgastroenteritis, respiratory infections, along withimmunization against the six Çkiller diseasesÈ ofchildhood: measles, diphtheria, tetanus, polio, per-tussis, tuberculosis.

¥ Adequate childhood nutrition to allow young womento achieve full pelvic growth before childbearingbegins

IX. PREVENTION OF OBSTETRICFISTULAS

927

Figure 30 : Historical trends in maternal mortality in theUnited States, England and Wales, and the Netherlands,1920 - 1960. (Loudon 1992a).

¥ Delay in childbearing until full pelvic growth iscompleted. There is substantial evidence to suggestthat education is the best way to shelter girls frompremature childbearing.

¥ Provision of family life education, education aboutwomenÕs health, sex education, and contraception toadolescent girls

¥ Elimination of traditional customs that promoteearly marriage. While respect for the principle ofindividual autonomy would argue for the eliminationof early betrothal and arranged marriages, even incases where such practices continue there should bea mutually-understood and agreed-upon delay in theconsummation of such marriages until the youngwoman has reached full pelvic maturity. There is noadvantage to any community in having early adoles-cent pregnancies.

¥ Supervision of the labor of every pregnant womanby a trained birth attendant. For developing countriesthis requires a commitment to developing culturally-acceptable community midwifery programs and anexpansion of midwifery training services everywhe-re.

¥ Monitoring of every labor with the use of parto-grams to detect cephalo-pelvic disproportion earlyand to prevent the development of obstructed labor.There is overwhelming evidence that simple techno-logy can accomplish this goal (Kwast 1994).

¥ Prompt, universal access to emergency obstetric careat the first referral level. This should be a fundamen-tal goal of every health care system in the world.This will require removal of cultural and institutio-nal, as well as physical, barriers (Prevention ofMaternal Mortality Network 1995; Maine 1991;Thaddeus and Maine 1994)

¥ Universal basic education for women. There is sub-stantial evidence that the education of women playsa major role in promoting maternal health, reducingmaternal mortality, and eliminating obstetric fistulas(Harrison 1997). These goals appear to be achievedlargely through better access to and utilization oflife-saving health care services; however, it must beemphasized that for education to be effective inachieving these goals, effective health care servicesmust first exist. Maternal mortality, and with it obs-tetric fistula formation, is largely a problem of theworldÕs poor: the affluent countries of the world bearsubstantial responsibility for allowing this situationto continue when relatively low cost, low technolo-gy interventions exist that could prevent it (Rosen-field and Maine 1985; Weil and Fernandez 1999)

¥ Finally, education of men concerning the importance

of womenÕs reproductive health for their own fami-lies in particular and for the community at large.Because men control a disproportionate share ofsocial resources in almost every developing country,they must understand that ÇwomenÕs healthÈ is notÇjust a womenÕs issue,È but is one in which they toomust become intimately involved.

The WHO Technical Working Group on the preventionand treatment of obstetric fistulas which met in Genevafrom April 17-21, 1989 naively suggested that a plancould be put forth by which the backlog of existing fis-tula cases could be cleared up within five years (WHO1989). Greater experience now suggests that solvingthe problem of cases awaiting surgery will require ageneration or more. Not only is the backlog of unrepai-red cases huge, but the absence of adequate maternalhealth care and emergency obstetrical services meansthat large numbers of new cases are continually occur-ring in those parts of the world where the fistula pro-blem is greatest, thus adding continually to the burdenof injury, which is not diminishing.

The special nature of the injuries produced by obstruc-ted labor, the stigmatizing and socially isolating natureof the injury, and the long periods of rehabilitation nee-ded before operation and the length of nursing carerequired after surgery, strongly argue in favor of thecreation of specialized fistula centers in areas of theworld where fistulas are highly prevalent. Ideally, eachcountry or region in which this problem exists shouldhave its own dedicated center, and centers in neighbo-ring countries should be encouraged to collaborate insharing information and establishing common proto-cols for research and training. There is an urgent needto create an international network of fistula centers inthe developing world that can collaborate with one ano-ther in advancing the care of patients who have sustai-ned an obstetric fistula.

Why are specialized fistula units needed? The firstargument is the efficiencies obtained from Çeconomiesof scale.È A Çfocused factoryÈ that does nothing exceptrepair fistulas and take care of the related problemsstemming from obstructed labor will develop specialexpertise in managing those problems. Such a centercan deal with more cases more efficiently than a smal-ler service attached to a general hospital. Such a centeralso allows for the concentration of a sufficient volumeof cases for meaningful training programs to be establi-shed through which additional fistula surgeons and fis-tula nurses can be developed for other centers in deve-loping countries.

X. DEALING WITH THE BACKLOGOF SURGICAL CASES

928

Second, the large volume of existing cases in develo-ping countries justifies this approach in terms of thesheer numbers of patients involved. The Addis AbabaFistula Hospital in Ethiopia, for example, has repairedapproximately 20,000 cases to date, and still there is noshortage of women needing services.

Third, fistula units in general hospitals must competefor scarce resources with the needs of the general medi-cal and surgical population. Since fistula repairs arerarely emergencies, it is difficult to book and keep sche-duled operating theater time in the face of ongoingemergencies such a road traffic accidents, incarceratedhernia, intestinal perforations from typhoid or parasites,women in obstructed labor who need a surgical delive-ry, and so on. In settings where the health care systemis already overwhelmed and where resources are scar-ce, scheduled fistula repairs are continually gettingÇbumpedÈ from the operating schedule by more acuteemergencies. While this is perfectly understandable, itaugurs poorly for the development of an efficient fistu-la service in such a setting. Furthermore, because fistu-la patients traditionally have required two weeks ofpost-operative care for catheter drainage, they are moreÇbed intensiveÈ than other surgical cases. In LaveryÕs(1955) series of 160 obstetric fistula cases from SouthAfrica, the average duration of stay in hospital was 57.3days, with a range of 11 to 264 days. In an environmentwith constant pressure to turn over hospital beds morequickly, fistula patients are likely to receive short shrift.

Fourth, fistula patients tend to integrate poorly intogeneral hospital wards. By the time most such patientsarrive for surgical treatment, they have been cast outand stigmatized by the society in which they live. Theyare psychologically and spiritually vulnerable, and theyare socially offensive due to the odor that surroundsthem. Many of these women are further stigmatized byunwarranted beliefs about why they developed a fistulain the first place: it is assumed by many to be a punish-ment for some offense against God (such as a pregnan-cy originating from an adulterous relationship) or as theresult of a hideous venereal disease, which some consi-der contagious. Fear of the unknown breeds hostilitytowards these patients in the community at large.

Fifth, for reasons cited above, these patients do better ina communal environment. It is a great psychologicalrelief to fistula patients to realize that they are notalone, to meet fellow sufferers with whom they canshare experiences. A Çsisterhood of sufferingÈ developsin a dedicated fistula center that is immensely benefi-cial in restoring psychological health and hope to thesewomen. At the Addis Ababa Fistula Hospital, most ofthe nursing care is provided by former fistula patients,who provide a level of empathetic nursing unequaledanywhere in the world.

It seems unlikely that large, dedicated fistula centerssimilar to that which exists in Addis Ababa will bedeveloped in every country where there is great needwithin the foreseeable future. What then can be done?For the reasons enumerated above, it is important thatfacilities be created that are dedicated exclusively to thecare of women with obstetric fistulas. Such facilitiesneed not be identical everywhere; their capabilitiescould be stratified into those capable of dealing withÇsimpleÈ cases, and those dealing with more complica-ted Çhigh riskÈ fistulas (Elkins et. al., 1994).

Hamlin and NicholsonÕs concept of the ÇdifficultÈ fis-tula has already been alluded to (1969). It is unrealisticto expect (and undesirable to encourage) a neophyte fis-tula surgeon to tackle cases of this complexity in asmall, low volume fistula unit. However, the majorityof obstetric fistulas seen in developing countries are notthis complicated, and the size of the fistula has littlebearing on the degree of urinary incontinence and resul-ting disability that the affected woman experiences:The same amount of urine runs out of an unscarred,freely mobile, 1 cm mid-vaginal fistula as runs out of a5 cm fistula in which the bladder neck and proximalurethra have been destroyed, the vagina stenosed, andthe fistula scarred against the pubic symphysis in densebands of fibrosis. A relatively unskilled fistula surgeoncould easily fix the former, thereby restoring continen-ce and hope to the afflicted patient, whereas he or shewould likely get into serious difficulties in an attempt tooperate on the latter.

Elkins and Wall (1996) have demonstrated that trainedobstetrician-gynecologists and general surgeons canquickly become proficient with the basic principles offistula surgery if they are given appropriate training andsupervision through an intensive Çshort courseÈ in fis-tula repair: it does not take years of additional trainingto create capable fistula surgeons. Although such trai-nees may not be fully able to tackle difficult or ÇhighriskÈ cases, they can certainly close the less complica-ted fistulas (Hamlin and Nicholson 1969; Elkins et. al.1994). What is required in order to achieve this level ofskill is an adequate volume of surgical cases that can bedone under supervision while the basic techniques aremastered. Thus, it does seem feasible to create a Çtie-redÈ system of fistula centers within countries wherethat need exists. The key to making such programswork is trained, committed personnel who have ade-quate local resources. Small programs can be startedwith a few dedicated beds in a general hospital. If addi-tional resources are provided, such programs canexpand to occupy a dedicated ward at the same facility.Ultimately, the program can become transformed intoan entirely separate facility dedicated exclusively to thecare and rehabilitation of fistula patients.

929

The following recommendations can be made for thecreation of specialized fistula centers in developingcountries:

¥ For the reasons outlined above, the creation of dedi-cated fistula centers should be encouraged.

¥ Fistula centers should be located close to an all-wea-ther road and good public transportation to facilitateaccess by patients. Wherever possible, such centersshould be located in the geographic region in whichfistulas are most prevalent, rather than in capitalcities which have few fistulas and which thereforerequire patients to travel long distances to obtaincare.

¥ Fistula centers should be located reasonably close toa general hospital so that emergency cases can bereferred elsewhere, thus protecting the overall mis-sion of the fistula program.

¥ Each fistula center should have a dedicated opera-ting theater, ideally one in which two or more opera-tions can be carried out simultaneously to maximizethe turnover of cases.

¥ An in-patient ward of adequate size should be provi-ded to deal with the expected number of cases. Award of 40 or 50 beds should allow 500 to 1,000operative cases to be treated each year.

¥ A Çstep-down unitÈ or hostel should be located at thefistula center to allow for pre-operative nutritionaland educational support for fistula victims, and toallow better post-operative monitoring of non-acutepatients.

¥ Care for fistula patients should be provided free ofcharge as a matter of principle. Fistula victimsoverwhelmingly tend to be young, poor, uneducatedwomen, often deserted by husband and family, andfrom rural areas. There are many tragic stories ofwomen who have suffered with fistulas for decadesbecause they were unable to scrape together bus fareto a hospital or the cost of surgical supplies. Manystudies of health care in developing countries havedemonstrated that the institution of Çuser feesÈ dra-matically reduces the utilization of services, oftenwith worsening health outcomes (Ekwempu 1990;Prevention of Maternal Mortality Network 1995).The total cost of supplies and services for fistula

repair is generally quite modest ($150 - $200).Although this still represents a large expenditure indeveloping countries, set against the value obtainedfrom giving a woman back her life, it representsexcellent cost-effective treatment. In order to bearthese expenses, most fistula centers will probablyrequire some degree of external funding throughinternational medical aid; efforts in this directionshould be encouraged. As the famous British sur-geon Lawson Tait wrote in 1889, ÇI have alreadysaid that operations for vaginal fistulae are rarelypaid for, except in gratitude, because the patients arenearly always poor. I must have operated on two orthree hundred cases, and I have not yet been remu-nerated to an extent which would pay for the instru-ments I have bought for the purpose.È

¥ Fistula centers will require an on-site kitchen andlaundry facilities. In African general hospitals, manyof these services are typically provided by the fami-ly and relatives of the patient. The nature of the inju-ry sustained by fistula victims means that many havebeen abandoned and lack this fundamental resourcefor coping with their condition. Fistula centersshould therefore be prepared to provide the essen-tials of nutrition and laundry. Due to the vagaries ofwater and electric power in developing countries,fistula centers should have their own diesel genera-tors and self-contained water supply, wherever fea-sible.

¥ Facilities and surgical care should be providedwithin a framework of Çlow technologyÈ medicine.Almost all surgical procedures can be providedunder simple spinal or ether anesthesia using stan-dard surgical instruments, simple suture, and trans-urethral catheter drainage. Clinical laboratory requi-rements should be kept to a minimum. Cinder blockconstruction, open wards, sheet metal roofing andsimple architecture can be used to keep constructionand operating costs to a minimum.

¥ Adequate housing for doctors and nursing staffshould be provided, within the appropriate expecta-tions of the local culture. If current and formerpatients are trained to work as nursing assistants,they can be housed acceptably in dormitory-stylehousing.

930

¥ The precise extent of the fistula problem in develo-ping countries is unknown. The available evidencesuggests that at a minimum, hundreds of thousands(if not several millions) of women are afflicted withthis condition worldwide, most especially in sub-Saharan Africa. The enormous burden of sufferingcaused by fistulas is borne principally by youngwomen living under conditions where their opportu-nities for education, economic prosperity, self-deter-mination, and Çthe pursuit of happiness,È are limitedby poverty, illiteracy, restricted social roles, theabsence of adequate reproductive health services,and political disenfranchisement. Fistulas are prima-rily a condition that afflicts societyÕs Çhave nots,Èand the prevalence of obstetric fistulas closely tracksworld maternal mortality statistics, especially inthose areas where obstructed labor is a principalcontributing cause of maternal death. The continuedprevalence of obstetric fistulas represents a tragicwaste of some of the most precious of the worldÕshuman resources: its young women.

¥ In theory, obstetric fistulas are completely preven-table. When they do occur, they should be curable inover 90% of cases by the provision of appropriate,low-technology medical and surgical care.

¥ The fistula problem has been shamefully neglectedby the governments of those countries in which theyoccur, by local health services, by non-governmentalorganizations, by international health organizationssuch as WHO, and by the foreign aid and internatio-nal development agencies of the worldÕs wealthycountries. The fistula problem has been almost uni-formly neglected by the worldÕs ÇSafe MotherhoodÈinitiative, which itself has been largely ineffectual inreducing maternal death in developing countries. Thefistula problem has been, and still is, an Çorphan.ÈThis situation is intolerable, and must be changed.

¥ There is a great need for village-based communitystudies of the incidence and prevalence of obstructedlabor and fistula formation. It is clear that most fis-tulas arise from the combination of Çobstructed laborand obstructed transportation,È but much more workis needed to understand the social context in whichobstetric emergencies arise and how they are dealtwith in developing countries. Nonetheless, theurgent needs of pregnant women should not be sacri-ficed on the altar of epidemiological research; rather,more attention should be paid to improving emer-gency treatment for obstetric complications at exis-

ting referral facilities, to upgrading peripheral facili-ties to provide essential life-saving obstetric care, toeducating the community about the danger signs ofobstetric complications, and to working with com-munity leaders to improve access to emergency obs-tetric care in areas where maternal mortality andobstetric fistula rates are high.

¥ There is no standard classification system for obste-tric fistulas. The committee recommends that theICS take this task upon itself and that a sub-commit-tee of the committee on the standardization of termi-nology be appointed to begin dealing with this task.

¥ Although the solution to the fistula problem will ulti-mately come from the provision of essential obstetricservices for all the worldÕs women, the current needsof those women who have already developed an obs-tetric fistula cannot be ignored. The committeerecommends that specialized fistula centers shouldbe created in all countries where obstetric fistulas areprevalent. Women with fistulas should have access toprompt, high quality surgical reconstruction whichshould be provided free of charge. Such facilitiesshould serve as centers of compassionate excellenceand should provide high quality patient care, medicaland nursing education, and clinical research as part oftheir mandate for existence. It is unlikely that essen-tial obstetric services will be available to all of theworldÕs women within the foreseeable future; there-fore, development of a sustainable clinical infrastruc-ture to deal with the effects of maternal morbidity islikely to be an essential need in reproductive medici-ne for the rest of the 21st Century.

¥ Although current surgical techniques consistentlyresult in fistula closure rates of 80% - 95% of cases,there has been almost no scientifically rigorousresearch carried out on most of the persistent techni-cal surgical questions raised by fistulas. Urgent workis needed in this area. Among the issues that shouldbe addressed are:

1. The problem of persistent stress incontinence aftersuccessful fistula closure;

2. The management of dramatically reduced bladdercapacity in patients with large fistulas who haveundergone successful closure;

3. The best technique for rectovaginal fistula repair;

4. The role of urinary diversion for the ÇincurableÈ fis-tula, usually involving some form of ureterosimoi-dostomy;

5. The role of vaginoplasty in patients with vaginal atre-sia from obstructed labor;

XI. CONCLUSIONS AND RECOMMENDATIONS

931

Abbott DH. 1950 The repair of vesico-vaginal fistulae. East AfricanMedical Journal 27109-118.

Abdalla RHD. 1982 Sisters in Affliction: Circumcision and Infibula-tion of Women in Africa. London: Zed Press.

AbouZahr C, Royston E. 1991 Maternal Mortality: A Global Fact-book. Geneva: World Health Organization.

AbouZahr C. 1998 Maternal mortality overview. Pp. 111-164 in:Murray CJL and Lopez AD (Eds), Health Dimension of Sex andReproduction: The global burden of sexually transmitted diseases,HIV, maternal conditions, perinatal disorders, and congenital anoma-lies. Boston: Harvard University Press. Published by the HarvardSchool of Public Health on behalf of the World Health Organizationand the World Bank

Aimakhu VE. 1974 Reproductive functions after the repair of obste-tric vesicovaginal fistulae. Fertility and Sterility 25:586-591.

Ampofo K, Otu T, Uchebo G. 1990 Epidemiology of vesico-vaginalfistulae in northern Nigeria. West African Journal of Medicine 9:98-102.

Amr MF. 1998 Vesico-vaginal fistula in Jordan. European Journal ofObstetrics and Gynecology and Reproductive Biology 80:201-203.

Arrowsmith SD. 1994 Genitourinary reconstruction in obstetric fistu-las. Journal of Urology 152:403-6.

Arrowsmith S, Hamlin EC, Wall LL. 1996 ÇObstructed labor injurycomplex:È Obstetric fistula formation and the multifaceted morbidityof maternal birth trauma in the developing world. Obstetrical andGynecological Survey 51:568-574.

Ashworth FL. 1973 Urinary vaginal fistulae: A series of 152 patientstreated in a small hospital in Ghana. West African Medical JournalApril:39-43.

Attah CA, Ozumba BC. 1993 Management of unrepairable urinaryvaginal fistulae in a developing country. Australia New Zealand Jour-nal of Surgery 63:217-220.

Aziz FA. 1965 Urinary fistulae from obstetrical trauma. Journal of Obs-tetrics and Gynaecology of the British Commonwealth 72:765-768.1980 Gynecologic and obstetric complications of female circumci-sion. International Journal of Gynaecology and Obstetrics 17:560-563.

Baines REM, Orford HJL, Theron JLL. 1976 The repair of vesicova-ginal fistulae by means of omental slings and grafts. South AfricanMedical Journal 50:959-961.

Bangser M, Gumodoka B, Berege Z. 1999 A comprehensive approa-ch to vesico-vaginal fistula: A project in Mwanza, Tanzania, pp.157-165 in M Berer and TKS Ravindram (Eds); Safe Motherhood Initia-tives: Critical Issues.. Oxford: Blackwell Science, for ReproductiveHealth Matters.

Barnes RW, Wilson WM. 1949 Reconstruction of the urethra with atube from bladder flap. Urologic and Cutaneous Review 53:604-606.

Bentley RJ. 1973 Abdominal repair of high rectovaginal fistula. Jour-nal of Obstetrics and Gynaecology of the British Commonwealth80:364-367.

Bieler EU, Schnabel T. 1976 Pituitary and ovarian function in womenwith vesicovaginal fistulae after obstructed and prolonged labour.South African Medical Journal 50:257-266.

Bird GC. 1967 Obstetric vesico-vaginal and allied fistulae. Journal ofObstetrics and Gynaecology of the British Commonwealth 74:749-752.

Bland KG, Gelfand M. 1970 The influence of urinary bilharziasis onvesico-vaginal fistula in relation to causation and healing. Transactionsof the Royal Society of Tropical Medicine and Hygiene 64:588-592.

Boddy J. 1982 Womb as oasis: The symbolic context of Pharaoniccircumcision in rural northern Sudan. American Ethnologist 9:682-698.

Briggs ND. 1983 Occipito-Posterior Positions: Predictive Indices onthe Outcome of Labor. M.D. Thesis, University of Lagos..

Carter B, Palumbo L, Creadick RN, Ross RA. 1952 Vesicovaginalfistulas. American Journal of Obstetrics and Gynecology 63:479-496.

Cockshott WP. 1973 Pubic changes associated with obstetric vesico-vaginal fistulae. Clinical Radiology 24:241-247.

Coetzee T, Lithgow DM. 1966 Obstetric fistulae of the urinary tract.Journal of Obstetrics and Gynaecology of the British Commonweal-th 73:837-844.

Colachis SC III, Pease WS, Johnson WE. 1994 A preventable causeof foot drop during childbirth. American Journal of Obstetrics andGynecology 171:270-272.

Cottingham J, Royston E. 1991 Obstetric Fistulae: A Review of Avai-lable Information. Geneva: World Health Organization (WHO/ MCH/MSM91.5)

Counsellor VS, Haigler FH. 1956 Management of urinary-vaginal fis-tula in 253 cases. American Journal of Obstetrics and Gynecology72:367-376.

Danso KA, Martey JO, Wall LL, Elkins TE. 1996 The epidemiologyof genitourinary fistulae in Kumasi, Ghana, 1977-1992. InternationalUrogynecology Journal 7:117-120.

Das RK, Sengupta SK. 1969 Vesico-vaginal fistula of obstetric ori-gin. Journal of Obstetrics and Gynaecology of India 19:383-389.

Dick JS, Strover RM. 1971 Vaginoplasty following vesicovaginal fis-tula repair: a preliminary report. South African Medical Journal45:617-620.

Eden TW. 1914 A case of superior recto-vaginal fistula dealt with bythe abdominal route after preliminary colostomy. Journal of Obste-trics and Gynaecology of the British Empire 26:175-185.

Ekwempu CC, Maine D, Olorukoba MB, Essien ES, Kisseka MN.1990 Structural adjustment and health in Africa. Lancet 336:56-7..

El-Dareer A. 1982 Woman, Why Do You Weep? Circumcision and ItsConsequences. London: Zed Press.

El-Guindi I. 1962 Vaginal atresia in Saudi Arabia. Journal of Obste-trics and Gynaecology of the British Commonwealth 69:996-998.

Elkins TE. 1994 Surgery for the obstetric vesicovaginal fistula: Areview of 100 operations in 82 patients.. American Journal of Obste-trics and Gynecology 170:1108-1120.

Elkins TE, DeLancy JOL, McGuire EJ. 1990 The use of modifiedMartius graft as an adjunctive technique in vesicovaginal and recto-vaginal fistula repair. Obstetrics and Gynecology 90:727-733.

REFERENCES

6. The best method of caring for patients who presentwith a ÇfreshÈ fistula a few days or weeks afterdelivery, including the role of early repair in selec-ted obstetric fistulas;

7. Future reproductive function in patients who havesustained an obstetric fistula;

8. The proper role and choice of flaps and grafts infistula surgery;

9. The optimum duration of post-operative bladderdrainage after fistula repair;

10. The role of nutritional support in improving theoutcomes of patients with obstetric fistulas; and

11. The role of physical therapy and reconstructiveorthopedic surgery in the management of womenwith obstetric foot-drop from prolonged obstructedlabor.

932

Elkins TE, Ghosh TS, Tagoe GA, Stocker R. 1992 Transvaginalmobilization and utilization of the anterior bladder wall to repair vesi-covaginal fistulas involving the urethra. Obstetrics and Gynecology79:455-60.

Elkins TE, Mahama E, OÕDonnell P, Fort D, Park RC. 1994 Reco-gnition and management of patients with high-risk vesicovaginal fis-tulas: Implications for teaching and research. Internatiional Urogyne-cology Journal 5:183-187.

Elkins TE, Wall LL. 1996 Report of a pilot project on the rapid trai-ning of pelvic surgeons in techniques of obstetric vesico-vaginal fis-tula repair in Ghana and Nigeria. Journal of Pelvic Surgery 2:182-186.

Emembolu J.. 1992 The obstetric fistula: Factors associated withimproved pregnancy outcome after a successful repair. InternationalJournal of Gynecology and Obstetrics 39:205-212.

Emmet TA. 1868 Vesico-Vaginal Fistula from Parturition and OtherCauses: With Cases of Recto-Vaginal Fistula. New York: WilliamWood.

1879 The necessity for early delivery, as demonstrated by the analy-sis of one hundred and sixty-one cases of vesico-vaginal fistula. Tran-sactions of the American Gynecological Society 3:114-34.

Enzelsberger H, Gitsch E. 1991 Surgical management of vesicovagi-nal fistulas according to Chassar MoirÕs method. Surgery, Gynecolo-gy and Obstetrics 173;183-186.

Everett HS, Mattingly RF. 1956 Vesicovaginal fistula. AmericanJournal of Obstetrics and Gynecology 72:712-724.

Evoh NJ, Akinla O. 1978 Reproductive performance after the repairof obstetric vesico-vaginal fistulae. Annals of Clinical Research10:303-306.

Fahmy K. 1962 Cervical and vaginal atresia due to packing the vagi-na with salt after labor. American Journal of Obstetrics and Gyneco-logy 84:1466-1469

Fisch M, Wammack R, Muller SC, Hohenfellner R. 1993 The MainzPouch II (sigma rectum pouch). Journal of Urology 149: 258- 263.

Fitzpatrick C, Elkins TE. 1993 Plastic surgical techniques in therepair of vesicovaginal fistulas: A review. International Urogynecolo-gy Journal 4:287-295.

Flocks RH, Culp DA. 1953 A modification of technique for anasto-mosing membranous urethra and bladder neck following total prosta-tectomy. Journal of Urology 69:411-415

Foda MS. 1959 Evaluation of methods of treatment of urinary fistu-lae in women: A report of 220 cases. Journal of Obstetrics and Gynae-cology of the British Empire 66:372-381.

Fortney JA, Smith JB (Eds). 1996 The Base of the Iceberg: Prevalen-ce and Perceptions of Maternal Morbidity in Four Developing Coun-tries (The Maternal Morbidity Network). Research Triangle Park,NC: Family Health International.

Frith K. 1960 Vaginal atresia of Arabia. Journal of Obstetrics andGynaecology of the British Empire 67:82-85

Gerharz EW, Kohl UN, Weingartner K, Kleinhans BJ, Melekos MD,Riedmiller H. 1998 Experience with the Mainz modification of ure-terosigmoidostomy. British Journal of Surgery 85:1512-1516.

Gharoro EP, Abedi HO. 1999 Vesico-vaginal fistula in Benin City,Nigeria. International Journal of Gynecology and Obstetrics 64:313-314.

Given FT Jr, Acosta AA. 1989 The Martius procedureÑbulbocaver-nosus fat flap: A review. Obstetrical and Gynecological Survey45:34-40.

Goh JTW. 1998 Genital tract fistula repair on 116 women. AustraliaNew Zealand Journal of Obstetrics and Gynaecology 38:158-161.

Goodwin WE, Scardino PT. 1980 Vesicovaginal and ureterovaginalfistulas: A summary of 25 years of experience. Journal of Urology123:370-374.

Graham W. 1998 The scandal of the century. British Journal of Obs-tetrics and Gynaecology 105:375-6.

Gruenbaum E. 2001 The Female Circumcision Controversy: AnAnthropological Perspective. Philadelphia: University of Pennsylva-nia Press.

Hall G. 1977 A review of drop-foot corrective surgery. LeprosyReview 48:184-192.

Hamlin RHJ, Nicholson EC. 1966 Experiences in the treatment of600 vaginal fistulas and in the management of 80 labours which havefollowed the repair of these injuries. Ethiopian Medical Journal4(5):189-192. 1969 Reconstruction of urethra totally destroyed inlabour. British Medical Journal 1:147-1150

Harris, S. 1950 WomanÕs Surgeon: The Life Story of J. Marion Sims.New York: Macmillan.

Harrison KA. 1983 Obstetric fistula: one social calamity too many.British Journal of Obstetrics and Gynaecology 90:385-386.

1985 Child-bearing, health, and social priorities: A survey of 22,774consecutive deliveries in Zaria, northern Nigeria. British Journal ofObstetrics and Gynaecology 92(Suppl 5):1-119.

Hassim AM, Lucas C. 1974 Reduction in the incidence of stressincontinence complicating fistula repair. British Journal of Surgery61:461-465.

Hilton P, Ward A. 1998 Epidemiological and surgical aspects of uro-genital fistulae: A review of 25 yearsÕ experience in southeast Nige-ria. International Urogynecology Journal 9:189-194.

Hilton P, Ward A, Molloy M, Umana O. 1998 Periurethral injectionof autologous fat for the treatment of post-fistula repair stress incon-tinence: A preliminary report. International Urogynecology Journal9:118-121.

Hodges AM. 1999 The Mitrofanoff urinary diversion for complexvesicovaginal fistulae: experience from Uganda. BJU International84:436-439.

Hoskins WJ, Park RC, Long R, Artman LE, McMahon EB. 1984Repair of urinary tract fistulas with bulbocavernosus myocutaneousflaps. Obstetrics and Gynecology 63:588-593.

Hudson CN. 1970 Acquired fistulae between the intestine and the vagi-na. Annals of the Royal College of Surgeons of England 46:20-40.

Hudson CN, Henrickse JP, Ward A. 1975 An operation for restorationof urinary incontinence following total loss of the urethra. BritishJournal of Obstetrics and Gynaecology 82:501- 504.

Humphries SV. 1961 Transplantation of ureters for vesico-vaginal fis-tula. South African Medical Journal 35:643-646.

Iloabachie GC. 1992 260 cases of juxta cervical fistula. East AfricanMedical Journal 69:188-190.

Kelly J. 1992 Vesico-vaginal and recto-vaginal fistulae. Journal of theRoyal Society of Medicine 85:257-258.

Kelly J, Kwast BE 1993a Obstetric vesicovaginal fistulas: Evaluationof failed repairs. International Urogynecology Journal 4:271-3.1993b Epidemiological study of vesicovaginal fistulas in Ethiopia.International Urogynecology Journal 4:278-281.

Kennelly MJ, Gormley EA, McGuire EJ. 1994 Early clinical expe-rience with adult bladder auto-augmentation. Journal of Urology152:303-306.

Kingston AE. 1957 The vaginal atresia of Arabia. Journal of Obste-trics and Gynaecology of the British Empire 64:836-839.

Kolawole TM, Adam SP, Evans KT. 1978 Comparative pelvimetricmeasurements in Nigerian and Welsh women. Clinical Radiology29:85-90.

Krishnan RG. 1949 A review of a series of 100 cases of vesicovagi-nal fistulae. Journal of Obstetrics and Gynaecology of the BritishEmpire 56:22-27.

Kwast BE. 1994 World Health Organization partograph in manage-ment of labour. Lancet 343:1399-1404

933

Lagundoye SB, Bell D, Gill G, Ogunbode O. 1976 Urinary tractchanges in obstetric vesico-vaginal fistulae: A report of 216 cases stu-died by intravenous urography. Clinical Radiology 27:531-539.

Langkilde NC, Pless TK, Lundbeck F, Nerstrom B. 1999 Surgicalrepair of vesicovaginal fistulae: A ten-year retrospective study. Scan-dinavian Journal of Urology and Nephrology 33:100-103.

Latzko W. 1942 Postoperative vesicovaginal fistulas: Genesis andtherapy. American Journal of Surgery 58:211-228.

Lavery DWP. 1955 Vesico-vaginal fistulae: A report on the vaginalrepair of 160 cases. Journal of Obstetrics and Gynaecology of the Bri-tish Empire 62:530-539.

Lawson JB. 1968 Birth-canal injuries. Proceedings of the RoyalSociety of Medicine 61:22-24. 1972a Rectovaginal fistulae followingdifficult labour. Proceedings of the Royal Society of Medicine65:283-286. 1972b Vesical fistulae into the vaginal vault. BritishJournal of Urology 44:623-631.

Leng WW, Blalock J, Fredriksson WH, English SF, McGuire EJ.1999 Enterocystoplasty or detrusor myectomy? Comparison of indi-cations and outcomes for bladder augmentation. Journal of Urology161:758-763.

Loudon I 1992a The transformation of maternal mortality. BritishMedical Journal 305:1557-1560. 1992b Death in Childbirth: AnInternational Study of Maternal Care and Maternal Mortality, 1800-1950. Oxford: Clarendon Press. 2000 Maternal mortality in the pastand its relevance to developing countries today. American Journal ofClinical Nutrition 2000;72(suppl)241S-246S.

McConnachie ELF. 1958 Fistulae of the urinary tract in the female: Aproposed classification. South African Medical Journal 32:524-527.

McKay WJS. 1922 Lawson Tait: His Life and Work. New York:William Wood and Co.

Mahfouz N. 1929 Urinary and rectovaginal fistulae in women. Jour-nal of Obstetrics and Gynaecology of the British Empire 36:581-589.1930 Urinary fistulae in women. Journal of Obstetrics and Gynaeco-logy of the British Empire 37:566-576. 1934 A new technique in dea-ling with superior recto-vaginal fistulae. Journal of Obstetrics andGynaecology of the British Empire 41:579-587. 1938 Urinary andfaecal fistulae. Journal of Obstetrics and Gynaecology of the BritishEmpire 45:405-424. 1957 Urinary Fistulae in women. Journal of Obs-tetrics and Gynaecology of the British Empire 64:23-34.Maine DM. 1991 Safe Motherhood Programs: Options and Issues.New York: Center for Population and Family Health, Columbia Uni-versity School of Public Health.Margolis T, Elkins TE, Seffah J, Oparo-Addo HS, and Fort D. 1994Full-thickness Martius grafts to preserve vaginal depth as an adjunctin the repair of large obstetric fistulas. Obstetrics and Gynecology84:148-152.Martius HR. 1956 Martius Gynecological Operations, With Empha-sis on Topographic Anatomy. Boston: Little, Brown, and Company..Translated and edited by Milton L. McCall and Karl A. Bolten (fromDie Gynakologischen Operationen und ihre Topographisch-Anatomi-schen Grundlagen, 7th ed). Massee JS, Welch JS, Pratt JH, Symmonds RE. 1964 Management ofurinary-vaginal fistula: Ten year survey. Journal of the AmericanMedical Association 190:124-128.

Moerman ML. 1982 Growth of the birth canal in adolescent girls.American Journal of Obstetrics and Gynecology. 143:528-32

Moir JC. 1964 Reconstruction of the urethra. Journal of Obstetricsand Gynaecology of the British Commonwealth 71:349-359. 1965The Sir Kendarnath Das Oration: The ÇcircumferentialÈ vesico-vagi-nal fistula. Journal of Obstetrics and Gynaecology of India 15:441-448. 1966 Vesicovaginal fistula: Thoughts on the treatment of 350cases. Proceedings of the Royal Society of Medicine 59:1019-1022.

Muleta M, Williams G. 1999 Postcoital injuries treated at the AddisAbaba Fistula Hospital, 1991-1997. Lancet 354:2051-2052.

Murphy M. 1981 Social consequences of vesico-vaginal fistula innorthern Nigeria. Journal of Biosocial Science 13:139-150.

Murphy M, Baba TM. 1981 Rural dwellers and health care in nor-thern Nigeria. Social Science and Medicine 15A:265-271.

Mustafa AZ, Rushwan HME. 1971 Acquired genito-urinary fistulaein the Sudan.. Journal of Obstetrics and Gynaecology of the BritishCommonwealth 78:1039-1043.

Naidu PM. 1962 Vesico-vaginal fistulae: An experience with 208cases. Journal of Obstetrics and Gynaecology of the British Com-monwealth 69:311-316.

Naidu PM, Krishna S. 1963 Vesico-vaginal fistulae and certain obs-tetric problems arising subsequent to repair. Journal of Obstetrics andGynaecology of the British Commonwealth 70:473-475.

Naim A. 1965 Vaginal fistulae complicating salt atresia in Arabia.Alexandria Medical Journal 11:218-226.

Noble CP. 1901 The new formation of the female urethra, with reportof a case. American Journal of Obstetrics 43:170-178.

Onolemhemhen DO, Ekwempu CC. 1999 An investigation of socio-medical risk factors associated with vaginal fistula in northern Nige-ria. Women and Health 28;103-116. Prevention of Maternal Mortali-ty Network. 1995 Situation analyses of emergency obstetric care:Examples from eleven operations research projects in West Africa.Social Science and Medicine 40:657-667.

Prual A, Huguet D, Garbin O, Rabe G. 1998 Severe obstetric morbi-dity of the third trimester, delivery, and early puerperium in Niamey(Niger); African Journal of Reproductive Health 2:10-19.

Punekar SV, Buch DN, Soni AB, Swami G, Rao SR, Kinne JS,Karhadkar SS. 1999 MartiusÕ labial fat pad interposition and its modi-fication in complex lower urinary fistulae. Journal of PostgraduateMedicine 45:69-73.

Quartey JKM. 1972 Bladder rotation flap for repair of difficult vesi-covaginal fistulas. Journal of Urology 107:60-62.

Radman HM. 1961 Vesicovaginal fistula. American Journal of Obs-tetrics and Gynecology 82:1238-1242..

Rangnekar NP. Imdad Ali N. Kaul SA. Pathak HR. 2000 Role of theMartius procedure in the management of urinary-vaginal fistulas.Journal of the American College of Surgeons 191:259-63.

Reif ME. 1988 Bilateral common peroneal nerve palsy secondary toprolonged squatting in natural childbirth. Birth 15:100-102.

Richard BM. 1989 Interosseous transfer of tibialis posterior for com-mon peroneal nerve palsy. Journal of Bone and Joint Surgery Ð Briti-sh Volume 71:834-7.

Rosenfeld A, Maine D. 1985 Maternal mortalityÑa neglected trage-dy. Where is the ÔMÕ in ÇMCH?Õ Lancet 2:83-85. St. George, J.1969Factors in the prediction of successful vaginal repair of vesico-vaginal fistulae. Journal of Obstetrics and Gynaecology of the Briti-sh Commonwealth 76:741-745.

Schleicher DJ, Ojengbede OHA, Elkins TE. 1993 Urologic evalua-tion after closure of vesicovaginal fistulas. International Urogyneco-logy Journal 4:262-265.

Shaw W. 1949 The Martius bulbo-cavernosus interposition operation.British Medical Journal 2:1261-1264.

Sims JM 1852 On the treatment of vesico-vaginal fistula. AmericanJournal of the Medical Sciences (new series) 23:59-82.

Sinclair RSC. 1952 Maternal obstetric palsy. South African MedicalJournal 26:708-714.

Snow BW, Cartwright PC. 1996 Bladder autoaugmentation. UrologicClinics of North America 23:323- 331.

Su CT. 1969 A flap technique for repair of vesicovaginal fistula. Jour-nal of Urology 102:56-59.

Symmonds RE. 1969 Loss of the urethral floor with total urinaryincontinence: A technique for urethral reconstruction. American Jour-nal of Obstetrics and Gynecology 103:665-678.

Symmonds RE, Hill LM. 1978 Loss of the urethra: A report on 50patients. American Journal of Obstetrics and Gynecology 130:130-138.

Tahzib F 1983 Epidemiological determinants of vesico-vaginal fistu-

934

las. British Journal of Obstetrics and Gynaecology 90:387-391 1985Vesicovaginal fistula in Nigerian children. Lancet 2:1291-3.

Tait L. 1879 Diseases of Women, second edition. New York: WilliamWood and Co. 1889 Diseases of Women and Abdominal Surgery.London.

Tanagho EA, Smith DR. 1972 Clinical evaluation of a surgical tech-nique for the correction of complete urinary incontinence. Journal ofUrology 107:402-411.

Taylor ES, Droegemueller W. 1967 Repair of urinary vaginal fistulas.Obstetrics and Gynecology 30:674-678.

Thaddeus S, Maine D. 1994 ÇToo far to walk:È maternal mortality incontext. Social Science and Medicine 38:1091-1110.

Thompson DJ. 1945 Vesicovaginal fistula: A review of 42 cases trea-ted by plastic repair or ureterosigmoidostomy. Journal of Obstetricsand Gynaecology of the British Empire 52:271-277.

Tomlinson AJ. Thornton JG. 1998 A randomised controlled trial ofantibiotic prophylaxis for vesico-vaginal fistula repair. British Journalof Obstetrics and Gynaecology. 105:397-9.

Toubia N. 1994 Female circumcision as a public health issue. NewEngland Journal of Medicine 331:712-716.

Underhill BML. 1964 Salt-induced vaginal stenosis of Arabia. Jour-nal of Obstetrics and Gynaecology of the British Commonwealth71:293-298.

Vix VA, Ryu CY. 1971 The adult symphysis pubic: normal and abnor-mal. American Journal of Roentgenology, Radium Therapy, andNuclear Medicine 112:517-525.

Waaldijk K. 1989 The (surgical) management of bladder fistula in775 women in Northern Nigeria. Thesis: University of Utrecht,Netherlands. 1994a Step-By-Step Surgery of Vesicovaginal Fistulas.Edinburgh: Campion Press. 1994b The immediate surgical manage-ment of fresh obstetric fistulas with catheter and/or early closure.International Journal of Gynecology and Obstetrics 45:11-16. 1995Surgical classification of obstetric fistulas. International Journal ofGynecology and Obstetrics 49:161-163.

Waaldijk K, Elkins TE. 1994 The obstetric fistula and peroneal nerveinjury: an analysis of 947 consecutive patients. International Urogy-necology Journal 5:12-14

Walker AHC. 1954 The management of urinary fistulas in a primiti-ve population. Surgery Gynecology and Obstetrics 99:301-309

Wall LL 1988 Hausa Medicine: Illness and Well-being in a West Afri-can Culture. Durham, NC: Duke University Press. 1998 Deadmothers and injured wives: the social context of maternal morbidityand mortality among the Hausa of northern Nigeria. Studies in Fami-ly Planning 29:341-359.

Ward A. 1980 Genito-urinary fistulae: a report on 1,789 cases. Pro-ceedings of the 2nd International Congress on Obstetrics and Gynae-cology, Lagos, Nigeria.

Weed JC. 1967 Management of vesicovaginal fistula: Experiencewith 75 cases. American Journal of Obstetrics and Gynecology97:1071-1075.

Weil O, Fernandez H. 1999 Is safe motherhood an orphan initiative?Lancet 354:940-943.

World Health Organization. 1986 Essential Obstetric Functions atFirst Referral level to Reduce Maternal Mortality: Report of a Tech-nical Working Group, June 23-27, 1986. Geneva: WHO. (FHE/86..4)1989 The Prevention and Treatment of Obstetric Fistulae: Report of aTechnical Working Group, Geneva, 17-21 April, 1989. Geneva:Division of Family Health, World Health OrganizationWHO/FHE/89.5 1998 Female Genital Mutilation: An Overview.Geneva: WHO, 1998.

World Health Organization and UNICEF. 1996 Revised 1990 Esti-mates of Maternal Mortality: A New Approach. Geneva: WHO.

Yenen E, Babuna C. 1965 Genital fistula: A study based on 197consecutive cases. Obstetrics and Gynecology 26:219-224.

Zacharin RF. 1980 Grafting as a principle in the surgical managementof vesicovaginal and rectovaginal fistulae. Australia New ZealandJournal of Obstetrics and Gynaecology 20:10-17. 1988 Obstetric Fis-tula. New York: Springer-Verlag.

935