Host and microbial risk factors and pathophysiology of fatalRickettsia conorii infection in Portuguese patients

Rita de SOUSA1 Ana FRANCcedilA2 Soacutenia DOacuteRIA NOgraveBREGA3 Adelaide BELO4 MarioAMARO2 Tiago ABREU5 Joseacute POCcedilAS6 Paula PROENCcedilA7 Joseacute VAZ4 Jorge TORGAL8Faacutetima BACELLAR1 Nahed ISMAIL9 and David H WALKER9

1 Centro de Estudos de Vectores e Doenccedilas Infecciosas Instituto Nacional de Sauacutede Dr Ricardo Jorge LisboaPortugal Tel (351) 217508127Fax (351) 217508121 E-mail ritasousainsamin-saudept

2 Hospital Garcia de Orta EPE Almada

3 Hospital Fernando Fonseca Amadora

4 Centro Hospitalar do Baixo Alentejo Beja

5Hospital do Espiriacuteto SantoEPE Eacutevora

6Centro Hospitalar de SetuacutebalHospital de Setuacutebal EPE Setuacutebal

7Hospital Distrital de Faro Faro

8Departamento de Sauacutede Puacuteblica Faculdade de Ciecircncias Meacutedicas de Lisboa Lisboa

9Department of Pathology University of Texas Medical Branch Galveston TX USA

AbstractBackgroundmdashThe pathophysiologic mechanisms of severity of Mediterranean spotted fever(MSF) and the host and microbial risk factors for a fatal outcome are incompletely determined

MethodsmdashIn a prospective study of 140 patients with documented identification of the causativerickettsial strain admitted to 13 Portuguese hospitals during 1994minus2006 univariate and multivariateanalyses determined the risk factors for a fatal outcome

FindingsmdashSeventy one (51) patients were infected with Rickettsia conorii Malish strain and 69(49) with Israeli spotted fever (ISF) strain Patients were admitted to ICU (29) hospitalized asroutine inpatients (67) or managed as outpatients (4) Deaths occurred in 29 (21) adults Fataloutcome was significantly more likely for patients with ISF strain infection and alcoholism was arisk factor The pathophysiology of a fatal outcome involved significantly greater incidence ofpetechial rash gastrointestinal symptoms confusionobtundation dehydration tachypneahepatomegaly leukocytosis coagulopathy azotemia hyperbilirubinemia and elevated hepaticenzymes and creatine kinase Some but not all these were observed more often in ISF strain-infectedpatients

ConclusionsmdashAlthough fatalities and similar clinical manifestations occurred with both strainsISF strain was more virulent than Malish strain Multivariate analysis revealed that acute renal failureand hyperbilirubinemia were most strongly associated with a fatal outcome

para Correspondence should be addressed to David H Walker Department of Pathology Center for Biodefense and Emerging InfectiousDiseases University of Texas Medical Branch 301 University Blvd Galveston TX USA77555minus0609 Telephone 409772minus3989 Fax409772minus1850 Email dwalkerutmbeduWe declare that we have no conflict of interest

NIH Public AccessAuthor ManuscriptJ Infect Dis Author manuscript available in PMC 2009 August 15

Published in final edited form asJ Infect Dis 2008 August 15 198(4) 576ndash585 doi101086590211

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

KeywordsRickettsia conorii alcoholism Israeli spotted fever strain Malish strain eschar

IntroductionMediterranean spotted fever (MSF) Israeli spotted fever (ISF) Indian tick typhus andAstrakhan spotted fever are caused by distinct strains of Rickettsia conorii These strains exhibitminor antigenic and genotypic differences and have been hypothesized to cause distinctiveclinical signs and differences in severity of disease12

In Portugal MSF is caused by R conorii Malish and ISF strains3 The incidence of MSF was84105 inhabitants during 1989minus2005 a high rate compared with other endemic countries inthe Mediterranean basin In the last decade an increasing number of cases of the malignantform of MSF has been described in Portuguese patients45 Moreover based on confirmeddiagnoses in the Portuguese hospital database (ACSS) the number of admissions for MSFincreased from 176 patients in 1994 to 446 patients in 20046 The case fatality rate in the sameperiod was 3minus7 among hospitalized patients In 1997 the case fatality rate was very high intwo Portuguese hospitals and in the Beja district reached 324 A study conducted in BejaHospital by Sousa et al from 1994 to 1998 to identify the risk factors associated with a fataloutcome indicated that delay in antibiotic treatment of rickettsial infection was not theexplanation but the risk of death was related to co-morbidity such as diabetes mellitus It wassuspected that the fatal outcome might have been related to the ISF strain which was isolatedfor the first time in Portugal in 1997 from fatal cases7 The hypothesis was that the ISF strainmight cause different clinical manifestations than those caused by Malish strain which couldlead to a late clinical diagnosis and consequently a delay in treatment thus resulting in a highernumber of severe cases In 2005 Sousa et al evaluated the three main signs fever rash escharand severity in patients infected with ISF and Malish strains3 They observed no statisticallysignificant differences in the signs in particular the presence or absence of an eschar in 94Portuguese patients infected with Malish or ISF strains In Israel eschars caused by infectionwith ISF strain have been described in only 4 of cases8-10 As the latter studies were notexhaustive we conducted a larger study to determine the risk factors for death in Portuguesepatients with a diagnosis of MSF confirmed by identification of the infective strain

MethodsCase definition and data collection

Patients were admitted to Portuguese hospitals with a clinical diagnosis of MSF confirmed byisolation of Rickettsia from blood or detection of rickettsial DNA in skin biopsy by PCR at thePortuguese National Institute of Health during 1994minus2006 Additionally serum samples weretested for the presence of IgM and IgG antibodies against Rconorii Epidemiological clinicaland laboratory data were recorded on admission a) The onset of symptoms was defined as thefirst day of observing any of the following symptoms fever chills rash headache orgastrointestinal symptoms such as nausea vomiting and diarrhea b) alcoholism is defined asingestion of 120 gm ethanol per day and reports of family members andor family physicianabout the alcoholic habits of the patient c) respiratory insufficiency was defined as failure ofthe exchange of oxygen and carbon dioxide pO2lt60 mm Hg or pCO2 gt55 mmHg d)congestive heart failure was defined as classification as class III or above according to NYHAFunctional Classification for Congestive Heart Failure e) Appropriate antibiotic treatment wasdefined as therapy with at least one antibiotic to which R conorii is susceptible Severity ofdisease was scored based on patient admission to the intensive care unit (ICU) APACHE IIscore (data not shown) or death

de SOUSA et al Page 2

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

Laboratory confirmation of MSFRickettsial isolation from bloodmdashA heparinized blood sample was collected frompatients with presumptive clinical diagnosis of MSF at hospital admission Plasma was testedfor rickettsial antibodies by immunofluorescence assay (IFA) Blood culture to isolateRickettsia and strain identification were performed as previously described11

DNA detection in skin biopsiesmdashSkin (5 mm) punch biopsy was obtained upon admissionfrom the eschar or if not present from a rash lesion Informed consent was obtained from allsubjects and experiments were performed with approval of the ethical committees of thehospitals and the National Institute of Health DNA extraction PCR and strain characterizationwere performed as previously described3

SerologymdashThe diagnosis of R conorii infection was confirmed when acute serum containedIgM antibodies at a titer of ge32 andor IgG titer of ge128 or there was a four-fold increase intiter between acute and convalescent sera1213 The cut-off of a positive result for Rconorii infection was established by the National Institute of Health based on previous studiesin the Portuguese population and as an endemic country for MSF Convalescent serum wascollected at least 15 days after the acute serum

Statistical analysismdashExploratory data analysis was performed by using tables offrequencies and central tendency and dispersion measures for continuous variablesDifferences were evaluated by Fishers exact test or Pearson chi-square test for categoricalvariables and Mann-Whitney test when continuous variables did not show a Gaussiandistribution Differences were considered to be significant at p value le 005 The associationbetween death and each factor was measured by odds ratio (OR) obtained with univariatelogistic regression models The multiple logistic regression model was built by stepwiseselection using the criteria of plt005 for selection variables and pgt010 for backwardelimination The goodness-of-fit of the final model was evaluated by means of Pearson chi-square test and Hosmer-Lemeshow test Data analysis was carried out using Stata version 90software (Stata Corporation College Station Texas USA 2005)

ResultsBetween January 1994 and July 2006 the clinical diagnosis of MSF was confirmed in 140patients from 13 hospitals Most patients were from five hospitals from southern Portugal 23from Almada 39 from Beja 32 from Eacutevora 17 from Faro 17 from Setuacutebal The remaining 12patients were from eight other hospitals Most of the cases occurred in summer with 113 (81)cases occurring during July to September

MSF caused by infection with Malish and ISF strains of R conorii was diagnosed in 65 (46)males and 75 (54) females The median age was 605 years Six patients were lt18 years oldfive patients were lt 6 years old and one patient was 12 years old Forty (29) patients wereadmitted to an ICU 95 (67) to a ward and five (4) were treated as outpatients Deaths dueto infection with Malish and ISF strains of R conorii occurred in 29 (21) adults (15 malesand 14 females)

MSF was confirmed in 93 patients by rickettsial isolation from blood and in 47 patients bydetection of R conorii DNA by PCR Additionally serology was performed on 138 acute seraand 53 convalescent sera Of the 138 acute serum samples only four (3) patients had asignificant level of R conorii IgM of ge32 andor IgG of ge128 In these four patients thediagnoses were further confirmed by detection of rickettsial DNA in skin biopsy by PCR whilerickettsial culture was negative Forty-six patients were diagnosed by seroconversion between

de SOUSA et al Page 3

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

acute and convalescent sera which is considered as a strong evidence of recent infection Nosignificant level of R conorii IgM and IgG antibodies was detected in the fatal cases of MSF

The patient populations infected with R conorii ISF and Malish strains were similar in ageperiod of time between the onset of symptoms and first visit to the physician and length ofhospitalization of both fatal and non-fatal cases (Table 1) There was a delay of one day (amedian of five days vs four days) in treatment of patients infected with Malish strain comparedwith ISF strain No significant difference in the incidence of underlying diseases or history ofanimal contact was detected between patients infected with ISF or Malish strains (Table 2a)Although there was no difference in the seasonal distribution of patients infected with eitherISF or Malish strains history of a recognized tick bite was more common in patients infectedwith Malish strain than those infected with ISF strain (63 vs 32) (Table 2a)

Comparison of the clinical manifestations of MSF caused by Malish and ISF strains revealedtremendous overlap that would not permit clinical recognition of the strain involved Bothstrains were capable of causing disease that resulted in each of the signs symptoms andpathophysiologic conditions causing abnormal laboratory values of several markers of organdysfunction Moreover statistical analysis of the clinical signs symptoms and laboratory dataon admission revealed few differences between patients infected with Malish and ISF strains(Table 2a 2b) First an eschar was observed in a significantly higher percentage of patientswith Malish (60) than ISF strain (38) However it should be emphasized that many patientsinfected with Malish strain did not have an eschar (Table 2b) The anatomic locations of theeschars were similar in ISF and Malish infections An eschar was often found in the axillagroin (25) legs (25) and trunk (20) and only three patients infected with RconoriiMalish strain presented with more than one eschar Second ISF strain-infected patientssuffered nausea and vomiting more often than Malish strain-infected patients (Table 2a)suggesting significant gastrointestinal involvement in patients with ISF strain infection Thirda higher percentage of patients with ISF strain infection than Malish strain-infected patientshad significantly elevated serum levels of total bilirubin γ-glutamyl transferase and alkalinephosphatase among other altered parameters (Table 2b) suggesting greater hepaticinvolvement

Among children infected with R conorii four were infected with Malish strain and two withISF strain (Table 3) The most severely ill child was a 12 year old who was infected with Malishstrain and did not receive treatment until the tenth day of illness He developed encephalitisthrombocytopenia and had markedly elevated serum hepatic transaminases

The most important observation in this study was that there was statistically significantlygreater severity of disease in patients infected with R conorii ISF strain than Malish strain(Table 4a) The case fatality rate for ISF strain (29) was significantly greater (p = 002) thanfor Malish strain (13) and a greater portion of ISF patients (36) required admission to theICU than Malish patients (22 p=0061)

Analysis of the relationship of co-morbid host factors and a fatal outcome demonstrated thatalcoholism was a statistically significant host condition that was a risk factor for a fatal outcomeof MSF (Table 4b)

Analysis of the clinical and laboratory data of fatal and non-fatal cases provided a picture ofthe pathophysiology of severe MSF in a large series of patients documented specifically tohave been infected with R conorii by rickettsial isolation or molecular diagnosis and geneticidentification of the agent rather than by a diagnosis based on clinical manifestations orserology that does not distinguish among various SFG rickettsioses Fatal cases of etiologicallydocumented R conorii infection were more likely to have severe multi-organ systeminvolvement as indicated by significantly greater occurrence of a petechial rash nausea

de SOUSA et al Page 4

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

vomiting diarrhea prostration confusion andor obtundation dehydration tachypnea andhepatomegaly (Table 4a) Also fatal cases were more likely to have leukocytosis (gt 11300μl) prolonged partial thromboplastin time elevated serum concentrations of urea creatininebilirubin alanine transaminase γ-glutamyl transferase alkaline phosphatase and creatinekinase (Table 4c) at the time of admission Among the fatal cases five did not have rash oreschar at the time of admission In survivors group the presence of fever and rash weresignificantly more often present than in fatal cases Multivariate analysis identified threevariables as independent predictors associated with fatal outcome hyperbilirubinemia (OR2599 95 CI [490 13793] p lt 0001) acute renal failure (OR 1810 95 CI [189 minus17335] p= 0012) and absence of rash (OR 003 95 CI [000 minus 122] p = 0064) The resultsof the goodness-of-fit test (Hosmer- Lemeshow) showed that the model behaves with a goodfitness (p = 0 125)

Data on therapy were available in 129 adult patients (64 infected with Malish and 65 infectedwith ISF strain) Doxycycline was the drug used most often to treat MSF patients 57 (89)of patients infected with R conorii Malish strain and 46 (71 ) with ISF strain

Doxycycline was used in combination with a fluoroquinolone in two (3) patients infectedwith Malish strain and in seven (11) patients infected with ISF strain Therapy with afluoroquinolone alone was used in six (9) patients infected with ISF strain Among patientsinfected with Malish strain two (3) received no treatment two (3) were treated with a β-lactam drug and one (2) was treated with rifampin Among patients infected with ISF strainfive (8) patients received no treatment and one (2) was treated with a β-lactam drug Nostatistically significant differences (p=0472) were found in therapeutic regimens between thegroups of patients Among the seven patients who did not receive treatment two survived andfive died

Three children received treatment with a macrolide (azithromycin) one with chloramphenicoland one with doxycycline The six month old child recovered without specific treatment Nofatal cases occurred in children (Table 3)

DiscussionThis study represents the largest case series of patients with a confirmed diagnosis of MSFbased on documentation of R conorii infection and the largest series with identification of thestrain rather than only clinical or non-species specific serologic diagnosis Owing to thepresence of shared protein and lipopolysaccharide antigens among SFG rickettsiae it isextremely difficult to distinguish infections with closely related rickettsiae by serologic orimmunohistochemical methods1415 In the Mediterranean basin patients with similar clinicalmanifestations have been documented by rickettsial isolation to be infected with R sibiricamongolotimonae strain R akari and R massiliae all of whom would have producedantibodies crossreactive with R conorii11 1617 Isolation andor PCR detection followed bygenetic characterization can determine the genotype of the organism to the level of genusspecies and strain Previous studies by Sousa et al (Portugal) and Giammanco et al (Italy)compared some clinical manifestations of MSF caused by ISF and Malish strains318 This isthe first exhaustive study reporting the differences between patients infected with R conoriiMalish and ISF strains and also the microbial and host risk factors and clinical manifestationsassociated with a fatal outcome In general patients infected with ISF and Malish strains didnot show many differences in epidemiological clinical or laboratory data The most importantfindings documented are that ISF strain is more virulent than Malish strain in the populationinvestigated and alcoholism is a risk factor for fatal outcome in MSF The microbial pathogenicmechanism by which ISF strain causes more severe illness remains to be determined In regardto host-related risk factors previous studies have described alcoholic patients who developed

de SOUSA et al Page 5

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

severe MSF However the present study is the first statistical analysis with a representativesample that documents alcoholism as a risk factor for severity of the disease19

Our data suggest that gastrointestinal symptoms such as nausea vomiting and diarrhea (table4) are prominent manifestations in patients with fatal MSF However they might have beenrelated to central nervous system involvement if the intracranial pressure was increased (ieamong the fatal cases were many patients with confusionobtundation (67) and rickettsialinfection in severefatal cases may result in cerebral edema and increased intracranial pressurewhich are features of rickettsial encephalitis and nausea and vomiting could be gastrointestinalsymptoms as a reflection of CNS involvement) Thus the exact mechanism that accounts forthe presence of marked gastrointestinal symptoms in fatal cases is not yet clear Furthermorethere were significantly greater gastrointestinal manifestations such as nausea and vomiting inpatients with ISF strain infection It is possible that infection with virulent ISF strains causesinflammation of abdominal organs secondary to rickettsial growth in blood vessels in thesetissues or systemic or local production of high levels of pro-inflammatory cytokines andchemokines This conclusion is consistent with our previous study showing that patients withsevere MSF had very high dermal levels of pro-inflammatory TNF-α enzymes that mediateintracellular bacterial elimination as well as tissue injury (ie iNOS and IDO) and RANTESa chemokine that stimulates the migration and accumulation of T cells at the site of infection20

Multiple logistic regression analysis revealed an independent significant association betweenthe presence of hyperbilirubinemia and fatal outcome in a greater portion of ISF patients thanMalish patients Whether hyperbilirubinemia is related to cholestasis hemolysis or othermechanisms remains to be determined Nevertheless it is clearly evident that hepaticinvolvement marked by hepatomegaly and elevated serum ALT alkaline phosphatase and γ-glutamyl transferase levels is strongly associated with a fatal outcome Histological studieshave shown that the hepatic pathology in patients with MSF namely focal necrosis andmononuclear lobular and portal triad inflammation is not associated with fatal disease21 Thelack of correlation between pathology and fatal disease but the strong association between fataldisease and abnormal values of several laboratory markers of organ dysfunction confirm theclinical diagnosis of multi-system organ involvement as an important factor in fatal rickettsialdiseases Indeed there was significant evidence of greater injury in numerous organ systemsincluding injury to vascular endothelium in brain (confusionobtundation) lungs (tachypnea)the coagulation system (prolonged partial thromboplastin time) and skin (petechial rash) infatal cases than in non-fatal cases Systemic hypotension with decreased renal perfusion andglomerular filtration rate are also most likely the pathophysiologic mechanism of the greaterincidence of acute renal failure in fatal cases22

This study reveals a lower incidence of eschars in patients with MSF due to ISF compared toinfection with Malish strain The significant difference in eschar formation between MSF dueto ISF and Malish strain was not detected in our previous study which we attribute to a non-representative number of patients involved in our previous analysis Consistent with thisconclusion is the finding that many patients infected with Malish strain in this study also didnot have an eschar as described previously Furthermore the current study includes a largernumber of fatal cases infected with either Malish or ISF strain and some of them did not haveeschars Similarly the case fatality rate for ISF strain (29) in this study was significantlygreater (p = 002) than for Malish strain (13) and a greater portion of ISF patients (36)required admission to the ICU than Malish patients (22 p=0061) It is not yet clear whetherthe lower incidence of eschar in fatal cases due to infection with ISF strains is a prognosticfactor that indicates an ineffective or detrimental host immune response against Rickettsia ora confounding factor that causes delayed diagnosis of these patients and thus diseaseprogression We favor the first possibility because our data show that the ISF strain-infected

de SOUSA et al Page 6

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

patients received anti-rickettsial treatment earlier in the course of disease than Malish strain-infected patients

The substantial case fatality rate for MSF in Portugal differs remarkably from that reported insome other regions of the Mediterranean area mostly from countries that are near Portugalsuch as France and Spain The possibility that some Rickettsia strains are more virulent thanothers is a reasonable hypothesis to examine in such areas It is conceivable that another lessvirulent human pathogen such as R massiliae or R monacensis might be the etiologic agentof spotted fever rickettsiosis in some geographic locations such as in the northeast of Spain23 24

The reason for the greater occurrence of a history of tick bite in patients infected with Malishstrain is not evident Although different vector species or tick stages could be hypothesized tohave transmitted the infection without detection most of the cases occurred in similar monthsfor both strains and Rhipicephalus sanguineus is the only recognized tick host of R conoriiin Portugal25

The lower incidence of rash and eschar as well as Rickettsia-specific IgM or IgG antibodies infatal cases emphasizes the need to consider the diagnosis of MSF in the absence of these signsor serological markers in patients with unexplained febrile illness in endemic regions Thisstudy reveals that diagnosis by PCR amplification of rickettsial DNA from skin biopsy is auseful method to establish the etiologic agent PCR was the only method by which the diagnosiswas established in 47 patients Additionally eight patients were diagnosed by both PCR andblood culture Of a total of 109 skin biopsies examined by PCR in this study 55 (348)samples were positive and 54 (341) were negative Among the negative skin biopsiesevaluated by PCR 22 patients had rickettsial infection confirmed by serology

Further analysis of the genotype of the Rickettsia allows identification of the agent to the strainlevel with the potential for investigation of differences in specific hypothetical virulence factorsincluding adhesins (outer membrane proteins A and B) 26-28 actin mobility (RickA) 29 andmembranolytic enzymes (phospholipase D hemolysins A and C)30-32 Advances in diagnosisand knowledge of pathogenic mechanisms of rickettsioses require further attention

AcknowledgmentsWe would like to thank all physicians and laboratories from the Portuguese hospitals that have collaborated providingclinical data Teresa Luz Paulo Parreira and Liacutegia Chaiacutenho for providing excellent technical assistance and DorisBaker and Sherrill Hebert for secretarial assistance

This work was financially supported in part by a grant (AI021242) from the National Institute of Allergy and InfectiousDiseases

References1 Zhu Y Fournier PE Eremeeva M Raoult D Proposal to create subspecies of Rickettsia conorii based

on multi-locus sequence typing and an emended description of Rickettsia conorii BMC Microbiol2005511 [PubMed 15766388]

2 Walker DH Feng HM Saada JI et al Comparative antigenic analysis of spotted fever group rickettsiaefrom Israel and other closely related organisms Am J Trop Med Hyg 199552569ndash76 [PubMed7611567]

3 Sousa R Ismail N Doacuteria-Noacutebrega S et al The presence of eschars but not greater severity inPortuguese patients infected with Israeli spotted fever Ann N Y Acad Sci 20051063197ndash202[PubMed 16481514]

4 Sousa R Doacuteria S Bacellar F Torgal J Mediterranean spotted fever in Portugal Risk factors for fataloutcome in 105 hospitalized patients Ann N Y Acad Sci 2003990285ndash94 [PubMed 12860641]

de SOUSA et al Page 7

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

KeywordsRickettsia conorii alcoholism Israeli spotted fever strain Malish strain eschar

IntroductionMediterranean spotted fever (MSF) Israeli spotted fever (ISF) Indian tick typhus andAstrakhan spotted fever are caused by distinct strains of Rickettsia conorii These strains exhibitminor antigenic and genotypic differences and have been hypothesized to cause distinctiveclinical signs and differences in severity of disease12

In Portugal MSF is caused by R conorii Malish and ISF strains3 The incidence of MSF was84105 inhabitants during 1989minus2005 a high rate compared with other endemic countries inthe Mediterranean basin In the last decade an increasing number of cases of the malignantform of MSF has been described in Portuguese patients45 Moreover based on confirmeddiagnoses in the Portuguese hospital database (ACSS) the number of admissions for MSFincreased from 176 patients in 1994 to 446 patients in 20046 The case fatality rate in the sameperiod was 3minus7 among hospitalized patients In 1997 the case fatality rate was very high intwo Portuguese hospitals and in the Beja district reached 324 A study conducted in BejaHospital by Sousa et al from 1994 to 1998 to identify the risk factors associated with a fataloutcome indicated that delay in antibiotic treatment of rickettsial infection was not theexplanation but the risk of death was related to co-morbidity such as diabetes mellitus It wassuspected that the fatal outcome might have been related to the ISF strain which was isolatedfor the first time in Portugal in 1997 from fatal cases7 The hypothesis was that the ISF strainmight cause different clinical manifestations than those caused by Malish strain which couldlead to a late clinical diagnosis and consequently a delay in treatment thus resulting in a highernumber of severe cases In 2005 Sousa et al evaluated the three main signs fever rash escharand severity in patients infected with ISF and Malish strains3 They observed no statisticallysignificant differences in the signs in particular the presence or absence of an eschar in 94Portuguese patients infected with Malish or ISF strains In Israel eschars caused by infectionwith ISF strain have been described in only 4 of cases8-10 As the latter studies were notexhaustive we conducted a larger study to determine the risk factors for death in Portuguesepatients with a diagnosis of MSF confirmed by identification of the infective strain

MethodsCase definition and data collection

Patients were admitted to Portuguese hospitals with a clinical diagnosis of MSF confirmed byisolation of Rickettsia from blood or detection of rickettsial DNA in skin biopsy by PCR at thePortuguese National Institute of Health during 1994minus2006 Additionally serum samples weretested for the presence of IgM and IgG antibodies against Rconorii Epidemiological clinicaland laboratory data were recorded on admission a) The onset of symptoms was defined as thefirst day of observing any of the following symptoms fever chills rash headache orgastrointestinal symptoms such as nausea vomiting and diarrhea b) alcoholism is defined asingestion of 120 gm ethanol per day and reports of family members andor family physicianabout the alcoholic habits of the patient c) respiratory insufficiency was defined as failure ofthe exchange of oxygen and carbon dioxide pO2lt60 mm Hg or pCO2 gt55 mmHg d)congestive heart failure was defined as classification as class III or above according to NYHAFunctional Classification for Congestive Heart Failure e) Appropriate antibiotic treatment wasdefined as therapy with at least one antibiotic to which R conorii is susceptible Severity ofdisease was scored based on patient admission to the intensive care unit (ICU) APACHE IIscore (data not shown) or death

de SOUSA et al Page 2

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

Laboratory confirmation of MSFRickettsial isolation from bloodmdashA heparinized blood sample was collected frompatients with presumptive clinical diagnosis of MSF at hospital admission Plasma was testedfor rickettsial antibodies by immunofluorescence assay (IFA) Blood culture to isolateRickettsia and strain identification were performed as previously described11

DNA detection in skin biopsiesmdashSkin (5 mm) punch biopsy was obtained upon admissionfrom the eschar or if not present from a rash lesion Informed consent was obtained from allsubjects and experiments were performed with approval of the ethical committees of thehospitals and the National Institute of Health DNA extraction PCR and strain characterizationwere performed as previously described3

SerologymdashThe diagnosis of R conorii infection was confirmed when acute serum containedIgM antibodies at a titer of ge32 andor IgG titer of ge128 or there was a four-fold increase intiter between acute and convalescent sera1213 The cut-off of a positive result for Rconorii infection was established by the National Institute of Health based on previous studiesin the Portuguese population and as an endemic country for MSF Convalescent serum wascollected at least 15 days after the acute serum

Statistical analysismdashExploratory data analysis was performed by using tables offrequencies and central tendency and dispersion measures for continuous variablesDifferences were evaluated by Fishers exact test or Pearson chi-square test for categoricalvariables and Mann-Whitney test when continuous variables did not show a Gaussiandistribution Differences were considered to be significant at p value le 005 The associationbetween death and each factor was measured by odds ratio (OR) obtained with univariatelogistic regression models The multiple logistic regression model was built by stepwiseselection using the criteria of plt005 for selection variables and pgt010 for backwardelimination The goodness-of-fit of the final model was evaluated by means of Pearson chi-square test and Hosmer-Lemeshow test Data analysis was carried out using Stata version 90software (Stata Corporation College Station Texas USA 2005)

ResultsBetween January 1994 and July 2006 the clinical diagnosis of MSF was confirmed in 140patients from 13 hospitals Most patients were from five hospitals from southern Portugal 23from Almada 39 from Beja 32 from Eacutevora 17 from Faro 17 from Setuacutebal The remaining 12patients were from eight other hospitals Most of the cases occurred in summer with 113 (81)cases occurring during July to September

MSF caused by infection with Malish and ISF strains of R conorii was diagnosed in 65 (46)males and 75 (54) females The median age was 605 years Six patients were lt18 years oldfive patients were lt 6 years old and one patient was 12 years old Forty (29) patients wereadmitted to an ICU 95 (67) to a ward and five (4) were treated as outpatients Deaths dueto infection with Malish and ISF strains of R conorii occurred in 29 (21) adults (15 malesand 14 females)

MSF was confirmed in 93 patients by rickettsial isolation from blood and in 47 patients bydetection of R conorii DNA by PCR Additionally serology was performed on 138 acute seraand 53 convalescent sera Of the 138 acute serum samples only four (3) patients had asignificant level of R conorii IgM of ge32 andor IgG of ge128 In these four patients thediagnoses were further confirmed by detection of rickettsial DNA in skin biopsy by PCR whilerickettsial culture was negative Forty-six patients were diagnosed by seroconversion between

de SOUSA et al Page 3

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

acute and convalescent sera which is considered as a strong evidence of recent infection Nosignificant level of R conorii IgM and IgG antibodies was detected in the fatal cases of MSF

The patient populations infected with R conorii ISF and Malish strains were similar in ageperiod of time between the onset of symptoms and first visit to the physician and length ofhospitalization of both fatal and non-fatal cases (Table 1) There was a delay of one day (amedian of five days vs four days) in treatment of patients infected with Malish strain comparedwith ISF strain No significant difference in the incidence of underlying diseases or history ofanimal contact was detected between patients infected with ISF or Malish strains (Table 2a)Although there was no difference in the seasonal distribution of patients infected with eitherISF or Malish strains history of a recognized tick bite was more common in patients infectedwith Malish strain than those infected with ISF strain (63 vs 32) (Table 2a)

Comparison of the clinical manifestations of MSF caused by Malish and ISF strains revealedtremendous overlap that would not permit clinical recognition of the strain involved Bothstrains were capable of causing disease that resulted in each of the signs symptoms andpathophysiologic conditions causing abnormal laboratory values of several markers of organdysfunction Moreover statistical analysis of the clinical signs symptoms and laboratory dataon admission revealed few differences between patients infected with Malish and ISF strains(Table 2a 2b) First an eschar was observed in a significantly higher percentage of patientswith Malish (60) than ISF strain (38) However it should be emphasized that many patientsinfected with Malish strain did not have an eschar (Table 2b) The anatomic locations of theeschars were similar in ISF and Malish infections An eschar was often found in the axillagroin (25) legs (25) and trunk (20) and only three patients infected with RconoriiMalish strain presented with more than one eschar Second ISF strain-infected patientssuffered nausea and vomiting more often than Malish strain-infected patients (Table 2a)suggesting significant gastrointestinal involvement in patients with ISF strain infection Thirda higher percentage of patients with ISF strain infection than Malish strain-infected patientshad significantly elevated serum levels of total bilirubin γ-glutamyl transferase and alkalinephosphatase among other altered parameters (Table 2b) suggesting greater hepaticinvolvement

Among children infected with R conorii four were infected with Malish strain and two withISF strain (Table 3) The most severely ill child was a 12 year old who was infected with Malishstrain and did not receive treatment until the tenth day of illness He developed encephalitisthrombocytopenia and had markedly elevated serum hepatic transaminases

The most important observation in this study was that there was statistically significantlygreater severity of disease in patients infected with R conorii ISF strain than Malish strain(Table 4a) The case fatality rate for ISF strain (29) was significantly greater (p = 002) thanfor Malish strain (13) and a greater portion of ISF patients (36) required admission to theICU than Malish patients (22 p=0061)

Analysis of the relationship of co-morbid host factors and a fatal outcome demonstrated thatalcoholism was a statistically significant host condition that was a risk factor for a fatal outcomeof MSF (Table 4b)

Analysis of the clinical and laboratory data of fatal and non-fatal cases provided a picture ofthe pathophysiology of severe MSF in a large series of patients documented specifically tohave been infected with R conorii by rickettsial isolation or molecular diagnosis and geneticidentification of the agent rather than by a diagnosis based on clinical manifestations orserology that does not distinguish among various SFG rickettsioses Fatal cases of etiologicallydocumented R conorii infection were more likely to have severe multi-organ systeminvolvement as indicated by significantly greater occurrence of a petechial rash nausea

de SOUSA et al Page 4

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

vomiting diarrhea prostration confusion andor obtundation dehydration tachypnea andhepatomegaly (Table 4a) Also fatal cases were more likely to have leukocytosis (gt 11300μl) prolonged partial thromboplastin time elevated serum concentrations of urea creatininebilirubin alanine transaminase γ-glutamyl transferase alkaline phosphatase and creatinekinase (Table 4c) at the time of admission Among the fatal cases five did not have rash oreschar at the time of admission In survivors group the presence of fever and rash weresignificantly more often present than in fatal cases Multivariate analysis identified threevariables as independent predictors associated with fatal outcome hyperbilirubinemia (OR2599 95 CI [490 13793] p lt 0001) acute renal failure (OR 1810 95 CI [189 minus17335] p= 0012) and absence of rash (OR 003 95 CI [000 minus 122] p = 0064) The resultsof the goodness-of-fit test (Hosmer- Lemeshow) showed that the model behaves with a goodfitness (p = 0 125)

Data on therapy were available in 129 adult patients (64 infected with Malish and 65 infectedwith ISF strain) Doxycycline was the drug used most often to treat MSF patients 57 (89)of patients infected with R conorii Malish strain and 46 (71 ) with ISF strain

Doxycycline was used in combination with a fluoroquinolone in two (3) patients infectedwith Malish strain and in seven (11) patients infected with ISF strain Therapy with afluoroquinolone alone was used in six (9) patients infected with ISF strain Among patientsinfected with Malish strain two (3) received no treatment two (3) were treated with a β-lactam drug and one (2) was treated with rifampin Among patients infected with ISF strainfive (8) patients received no treatment and one (2) was treated with a β-lactam drug Nostatistically significant differences (p=0472) were found in therapeutic regimens between thegroups of patients Among the seven patients who did not receive treatment two survived andfive died

Three children received treatment with a macrolide (azithromycin) one with chloramphenicoland one with doxycycline The six month old child recovered without specific treatment Nofatal cases occurred in children (Table 3)

DiscussionThis study represents the largest case series of patients with a confirmed diagnosis of MSFbased on documentation of R conorii infection and the largest series with identification of thestrain rather than only clinical or non-species specific serologic diagnosis Owing to thepresence of shared protein and lipopolysaccharide antigens among SFG rickettsiae it isextremely difficult to distinguish infections with closely related rickettsiae by serologic orimmunohistochemical methods1415 In the Mediterranean basin patients with similar clinicalmanifestations have been documented by rickettsial isolation to be infected with R sibiricamongolotimonae strain R akari and R massiliae all of whom would have producedantibodies crossreactive with R conorii11 1617 Isolation andor PCR detection followed bygenetic characterization can determine the genotype of the organism to the level of genusspecies and strain Previous studies by Sousa et al (Portugal) and Giammanco et al (Italy)compared some clinical manifestations of MSF caused by ISF and Malish strains318 This isthe first exhaustive study reporting the differences between patients infected with R conoriiMalish and ISF strains and also the microbial and host risk factors and clinical manifestationsassociated with a fatal outcome In general patients infected with ISF and Malish strains didnot show many differences in epidemiological clinical or laboratory data The most importantfindings documented are that ISF strain is more virulent than Malish strain in the populationinvestigated and alcoholism is a risk factor for fatal outcome in MSF The microbial pathogenicmechanism by which ISF strain causes more severe illness remains to be determined In regardto host-related risk factors previous studies have described alcoholic patients who developed

de SOUSA et al Page 5

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

severe MSF However the present study is the first statistical analysis with a representativesample that documents alcoholism as a risk factor for severity of the disease19

Our data suggest that gastrointestinal symptoms such as nausea vomiting and diarrhea (table4) are prominent manifestations in patients with fatal MSF However they might have beenrelated to central nervous system involvement if the intracranial pressure was increased (ieamong the fatal cases were many patients with confusionobtundation (67) and rickettsialinfection in severefatal cases may result in cerebral edema and increased intracranial pressurewhich are features of rickettsial encephalitis and nausea and vomiting could be gastrointestinalsymptoms as a reflection of CNS involvement) Thus the exact mechanism that accounts forthe presence of marked gastrointestinal symptoms in fatal cases is not yet clear Furthermorethere were significantly greater gastrointestinal manifestations such as nausea and vomiting inpatients with ISF strain infection It is possible that infection with virulent ISF strains causesinflammation of abdominal organs secondary to rickettsial growth in blood vessels in thesetissues or systemic or local production of high levels of pro-inflammatory cytokines andchemokines This conclusion is consistent with our previous study showing that patients withsevere MSF had very high dermal levels of pro-inflammatory TNF-α enzymes that mediateintracellular bacterial elimination as well as tissue injury (ie iNOS and IDO) and RANTESa chemokine that stimulates the migration and accumulation of T cells at the site of infection20

Multiple logistic regression analysis revealed an independent significant association betweenthe presence of hyperbilirubinemia and fatal outcome in a greater portion of ISF patients thanMalish patients Whether hyperbilirubinemia is related to cholestasis hemolysis or othermechanisms remains to be determined Nevertheless it is clearly evident that hepaticinvolvement marked by hepatomegaly and elevated serum ALT alkaline phosphatase and γ-glutamyl transferase levels is strongly associated with a fatal outcome Histological studieshave shown that the hepatic pathology in patients with MSF namely focal necrosis andmononuclear lobular and portal triad inflammation is not associated with fatal disease21 Thelack of correlation between pathology and fatal disease but the strong association between fataldisease and abnormal values of several laboratory markers of organ dysfunction confirm theclinical diagnosis of multi-system organ involvement as an important factor in fatal rickettsialdiseases Indeed there was significant evidence of greater injury in numerous organ systemsincluding injury to vascular endothelium in brain (confusionobtundation) lungs (tachypnea)the coagulation system (prolonged partial thromboplastin time) and skin (petechial rash) infatal cases than in non-fatal cases Systemic hypotension with decreased renal perfusion andglomerular filtration rate are also most likely the pathophysiologic mechanism of the greaterincidence of acute renal failure in fatal cases22

This study reveals a lower incidence of eschars in patients with MSF due to ISF compared toinfection with Malish strain The significant difference in eschar formation between MSF dueto ISF and Malish strain was not detected in our previous study which we attribute to a non-representative number of patients involved in our previous analysis Consistent with thisconclusion is the finding that many patients infected with Malish strain in this study also didnot have an eschar as described previously Furthermore the current study includes a largernumber of fatal cases infected with either Malish or ISF strain and some of them did not haveeschars Similarly the case fatality rate for ISF strain (29) in this study was significantlygreater (p = 002) than for Malish strain (13) and a greater portion of ISF patients (36)required admission to the ICU than Malish patients (22 p=0061) It is not yet clear whetherthe lower incidence of eschar in fatal cases due to infection with ISF strains is a prognosticfactor that indicates an ineffective or detrimental host immune response against Rickettsia ora confounding factor that causes delayed diagnosis of these patients and thus diseaseprogression We favor the first possibility because our data show that the ISF strain-infected

de SOUSA et al Page 6

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

patients received anti-rickettsial treatment earlier in the course of disease than Malish strain-infected patients

The substantial case fatality rate for MSF in Portugal differs remarkably from that reported insome other regions of the Mediterranean area mostly from countries that are near Portugalsuch as France and Spain The possibility that some Rickettsia strains are more virulent thanothers is a reasonable hypothesis to examine in such areas It is conceivable that another lessvirulent human pathogen such as R massiliae or R monacensis might be the etiologic agentof spotted fever rickettsiosis in some geographic locations such as in the northeast of Spain23 24

The reason for the greater occurrence of a history of tick bite in patients infected with Malishstrain is not evident Although different vector species or tick stages could be hypothesized tohave transmitted the infection without detection most of the cases occurred in similar monthsfor both strains and Rhipicephalus sanguineus is the only recognized tick host of R conoriiin Portugal25

The lower incidence of rash and eschar as well as Rickettsia-specific IgM or IgG antibodies infatal cases emphasizes the need to consider the diagnosis of MSF in the absence of these signsor serological markers in patients with unexplained febrile illness in endemic regions Thisstudy reveals that diagnosis by PCR amplification of rickettsial DNA from skin biopsy is auseful method to establish the etiologic agent PCR was the only method by which the diagnosiswas established in 47 patients Additionally eight patients were diagnosed by both PCR andblood culture Of a total of 109 skin biopsies examined by PCR in this study 55 (348)samples were positive and 54 (341) were negative Among the negative skin biopsiesevaluated by PCR 22 patients had rickettsial infection confirmed by serology

Further analysis of the genotype of the Rickettsia allows identification of the agent to the strainlevel with the potential for investigation of differences in specific hypothetical virulence factorsincluding adhesins (outer membrane proteins A and B) 26-28 actin mobility (RickA) 29 andmembranolytic enzymes (phospholipase D hemolysins A and C)30-32 Advances in diagnosisand knowledge of pathogenic mechanisms of rickettsioses require further attention

AcknowledgmentsWe would like to thank all physicians and laboratories from the Portuguese hospitals that have collaborated providingclinical data Teresa Luz Paulo Parreira and Liacutegia Chaiacutenho for providing excellent technical assistance and DorisBaker and Sherrill Hebert for secretarial assistance

This work was financially supported in part by a grant (AI021242) from the National Institute of Allergy and InfectiousDiseases

References1 Zhu Y Fournier PE Eremeeva M Raoult D Proposal to create subspecies of Rickettsia conorii based

on multi-locus sequence typing and an emended description of Rickettsia conorii BMC Microbiol2005511 [PubMed 15766388]

2 Walker DH Feng HM Saada JI et al Comparative antigenic analysis of spotted fever group rickettsiaefrom Israel and other closely related organisms Am J Trop Med Hyg 199552569ndash76 [PubMed7611567]

3 Sousa R Ismail N Doacuteria-Noacutebrega S et al The presence of eschars but not greater severity inPortuguese patients infected with Israeli spotted fever Ann N Y Acad Sci 20051063197ndash202[PubMed 16481514]

4 Sousa R Doacuteria S Bacellar F Torgal J Mediterranean spotted fever in Portugal Risk factors for fataloutcome in 105 hospitalized patients Ann N Y Acad Sci 2003990285ndash94 [PubMed 12860641]

de SOUSA et al Page 7

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

Laboratory confirmation of MSFRickettsial isolation from bloodmdashA heparinized blood sample was collected frompatients with presumptive clinical diagnosis of MSF at hospital admission Plasma was testedfor rickettsial antibodies by immunofluorescence assay (IFA) Blood culture to isolateRickettsia and strain identification were performed as previously described11

DNA detection in skin biopsiesmdashSkin (5 mm) punch biopsy was obtained upon admissionfrom the eschar or if not present from a rash lesion Informed consent was obtained from allsubjects and experiments were performed with approval of the ethical committees of thehospitals and the National Institute of Health DNA extraction PCR and strain characterizationwere performed as previously described3

SerologymdashThe diagnosis of R conorii infection was confirmed when acute serum containedIgM antibodies at a titer of ge32 andor IgG titer of ge128 or there was a four-fold increase intiter between acute and convalescent sera1213 The cut-off of a positive result for Rconorii infection was established by the National Institute of Health based on previous studiesin the Portuguese population and as an endemic country for MSF Convalescent serum wascollected at least 15 days after the acute serum

Statistical analysismdashExploratory data analysis was performed by using tables offrequencies and central tendency and dispersion measures for continuous variablesDifferences were evaluated by Fishers exact test or Pearson chi-square test for categoricalvariables and Mann-Whitney test when continuous variables did not show a Gaussiandistribution Differences were considered to be significant at p value le 005 The associationbetween death and each factor was measured by odds ratio (OR) obtained with univariatelogistic regression models The multiple logistic regression model was built by stepwiseselection using the criteria of plt005 for selection variables and pgt010 for backwardelimination The goodness-of-fit of the final model was evaluated by means of Pearson chi-square test and Hosmer-Lemeshow test Data analysis was carried out using Stata version 90software (Stata Corporation College Station Texas USA 2005)

ResultsBetween January 1994 and July 2006 the clinical diagnosis of MSF was confirmed in 140patients from 13 hospitals Most patients were from five hospitals from southern Portugal 23from Almada 39 from Beja 32 from Eacutevora 17 from Faro 17 from Setuacutebal The remaining 12patients were from eight other hospitals Most of the cases occurred in summer with 113 (81)cases occurring during July to September

MSF caused by infection with Malish and ISF strains of R conorii was diagnosed in 65 (46)males and 75 (54) females The median age was 605 years Six patients were lt18 years oldfive patients were lt 6 years old and one patient was 12 years old Forty (29) patients wereadmitted to an ICU 95 (67) to a ward and five (4) were treated as outpatients Deaths dueto infection with Malish and ISF strains of R conorii occurred in 29 (21) adults (15 malesand 14 females)

MSF was confirmed in 93 patients by rickettsial isolation from blood and in 47 patients bydetection of R conorii DNA by PCR Additionally serology was performed on 138 acute seraand 53 convalescent sera Of the 138 acute serum samples only four (3) patients had asignificant level of R conorii IgM of ge32 andor IgG of ge128 In these four patients thediagnoses were further confirmed by detection of rickettsial DNA in skin biopsy by PCR whilerickettsial culture was negative Forty-six patients were diagnosed by seroconversion between

de SOUSA et al Page 3

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

acute and convalescent sera which is considered as a strong evidence of recent infection Nosignificant level of R conorii IgM and IgG antibodies was detected in the fatal cases of MSF

The patient populations infected with R conorii ISF and Malish strains were similar in ageperiod of time between the onset of symptoms and first visit to the physician and length ofhospitalization of both fatal and non-fatal cases (Table 1) There was a delay of one day (amedian of five days vs four days) in treatment of patients infected with Malish strain comparedwith ISF strain No significant difference in the incidence of underlying diseases or history ofanimal contact was detected between patients infected with ISF or Malish strains (Table 2a)Although there was no difference in the seasonal distribution of patients infected with eitherISF or Malish strains history of a recognized tick bite was more common in patients infectedwith Malish strain than those infected with ISF strain (63 vs 32) (Table 2a)

Comparison of the clinical manifestations of MSF caused by Malish and ISF strains revealedtremendous overlap that would not permit clinical recognition of the strain involved Bothstrains were capable of causing disease that resulted in each of the signs symptoms andpathophysiologic conditions causing abnormal laboratory values of several markers of organdysfunction Moreover statistical analysis of the clinical signs symptoms and laboratory dataon admission revealed few differences between patients infected with Malish and ISF strains(Table 2a 2b) First an eschar was observed in a significantly higher percentage of patientswith Malish (60) than ISF strain (38) However it should be emphasized that many patientsinfected with Malish strain did not have an eschar (Table 2b) The anatomic locations of theeschars were similar in ISF and Malish infections An eschar was often found in the axillagroin (25) legs (25) and trunk (20) and only three patients infected with RconoriiMalish strain presented with more than one eschar Second ISF strain-infected patientssuffered nausea and vomiting more often than Malish strain-infected patients (Table 2a)suggesting significant gastrointestinal involvement in patients with ISF strain infection Thirda higher percentage of patients with ISF strain infection than Malish strain-infected patientshad significantly elevated serum levels of total bilirubin γ-glutamyl transferase and alkalinephosphatase among other altered parameters (Table 2b) suggesting greater hepaticinvolvement

Among children infected with R conorii four were infected with Malish strain and two withISF strain (Table 3) The most severely ill child was a 12 year old who was infected with Malishstrain and did not receive treatment until the tenth day of illness He developed encephalitisthrombocytopenia and had markedly elevated serum hepatic transaminases

The most important observation in this study was that there was statistically significantlygreater severity of disease in patients infected with R conorii ISF strain than Malish strain(Table 4a) The case fatality rate for ISF strain (29) was significantly greater (p = 002) thanfor Malish strain (13) and a greater portion of ISF patients (36) required admission to theICU than Malish patients (22 p=0061)

Analysis of the relationship of co-morbid host factors and a fatal outcome demonstrated thatalcoholism was a statistically significant host condition that was a risk factor for a fatal outcomeof MSF (Table 4b)

Analysis of the clinical and laboratory data of fatal and non-fatal cases provided a picture ofthe pathophysiology of severe MSF in a large series of patients documented specifically tohave been infected with R conorii by rickettsial isolation or molecular diagnosis and geneticidentification of the agent rather than by a diagnosis based on clinical manifestations orserology that does not distinguish among various SFG rickettsioses Fatal cases of etiologicallydocumented R conorii infection were more likely to have severe multi-organ systeminvolvement as indicated by significantly greater occurrence of a petechial rash nausea

de SOUSA et al Page 4

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

vomiting diarrhea prostration confusion andor obtundation dehydration tachypnea andhepatomegaly (Table 4a) Also fatal cases were more likely to have leukocytosis (gt 11300μl) prolonged partial thromboplastin time elevated serum concentrations of urea creatininebilirubin alanine transaminase γ-glutamyl transferase alkaline phosphatase and creatinekinase (Table 4c) at the time of admission Among the fatal cases five did not have rash oreschar at the time of admission In survivors group the presence of fever and rash weresignificantly more often present than in fatal cases Multivariate analysis identified threevariables as independent predictors associated with fatal outcome hyperbilirubinemia (OR2599 95 CI [490 13793] p lt 0001) acute renal failure (OR 1810 95 CI [189 minus17335] p= 0012) and absence of rash (OR 003 95 CI [000 minus 122] p = 0064) The resultsof the goodness-of-fit test (Hosmer- Lemeshow) showed that the model behaves with a goodfitness (p = 0 125)

Data on therapy were available in 129 adult patients (64 infected with Malish and 65 infectedwith ISF strain) Doxycycline was the drug used most often to treat MSF patients 57 (89)of patients infected with R conorii Malish strain and 46 (71 ) with ISF strain

Doxycycline was used in combination with a fluoroquinolone in two (3) patients infectedwith Malish strain and in seven (11) patients infected with ISF strain Therapy with afluoroquinolone alone was used in six (9) patients infected with ISF strain Among patientsinfected with Malish strain two (3) received no treatment two (3) were treated with a β-lactam drug and one (2) was treated with rifampin Among patients infected with ISF strainfive (8) patients received no treatment and one (2) was treated with a β-lactam drug Nostatistically significant differences (p=0472) were found in therapeutic regimens between thegroups of patients Among the seven patients who did not receive treatment two survived andfive died

Three children received treatment with a macrolide (azithromycin) one with chloramphenicoland one with doxycycline The six month old child recovered without specific treatment Nofatal cases occurred in children (Table 3)

DiscussionThis study represents the largest case series of patients with a confirmed diagnosis of MSFbased on documentation of R conorii infection and the largest series with identification of thestrain rather than only clinical or non-species specific serologic diagnosis Owing to thepresence of shared protein and lipopolysaccharide antigens among SFG rickettsiae it isextremely difficult to distinguish infections with closely related rickettsiae by serologic orimmunohistochemical methods1415 In the Mediterranean basin patients with similar clinicalmanifestations have been documented by rickettsial isolation to be infected with R sibiricamongolotimonae strain R akari and R massiliae all of whom would have producedantibodies crossreactive with R conorii11 1617 Isolation andor PCR detection followed bygenetic characterization can determine the genotype of the organism to the level of genusspecies and strain Previous studies by Sousa et al (Portugal) and Giammanco et al (Italy)compared some clinical manifestations of MSF caused by ISF and Malish strains318 This isthe first exhaustive study reporting the differences between patients infected with R conoriiMalish and ISF strains and also the microbial and host risk factors and clinical manifestationsassociated with a fatal outcome In general patients infected with ISF and Malish strains didnot show many differences in epidemiological clinical or laboratory data The most importantfindings documented are that ISF strain is more virulent than Malish strain in the populationinvestigated and alcoholism is a risk factor for fatal outcome in MSF The microbial pathogenicmechanism by which ISF strain causes more severe illness remains to be determined In regardto host-related risk factors previous studies have described alcoholic patients who developed

de SOUSA et al Page 5

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

severe MSF However the present study is the first statistical analysis with a representativesample that documents alcoholism as a risk factor for severity of the disease19

Our data suggest that gastrointestinal symptoms such as nausea vomiting and diarrhea (table4) are prominent manifestations in patients with fatal MSF However they might have beenrelated to central nervous system involvement if the intracranial pressure was increased (ieamong the fatal cases were many patients with confusionobtundation (67) and rickettsialinfection in severefatal cases may result in cerebral edema and increased intracranial pressurewhich are features of rickettsial encephalitis and nausea and vomiting could be gastrointestinalsymptoms as a reflection of CNS involvement) Thus the exact mechanism that accounts forthe presence of marked gastrointestinal symptoms in fatal cases is not yet clear Furthermorethere were significantly greater gastrointestinal manifestations such as nausea and vomiting inpatients with ISF strain infection It is possible that infection with virulent ISF strains causesinflammation of abdominal organs secondary to rickettsial growth in blood vessels in thesetissues or systemic or local production of high levels of pro-inflammatory cytokines andchemokines This conclusion is consistent with our previous study showing that patients withsevere MSF had very high dermal levels of pro-inflammatory TNF-α enzymes that mediateintracellular bacterial elimination as well as tissue injury (ie iNOS and IDO) and RANTESa chemokine that stimulates the migration and accumulation of T cells at the site of infection20

Multiple logistic regression analysis revealed an independent significant association betweenthe presence of hyperbilirubinemia and fatal outcome in a greater portion of ISF patients thanMalish patients Whether hyperbilirubinemia is related to cholestasis hemolysis or othermechanisms remains to be determined Nevertheless it is clearly evident that hepaticinvolvement marked by hepatomegaly and elevated serum ALT alkaline phosphatase and γ-glutamyl transferase levels is strongly associated with a fatal outcome Histological studieshave shown that the hepatic pathology in patients with MSF namely focal necrosis andmononuclear lobular and portal triad inflammation is not associated with fatal disease21 Thelack of correlation between pathology and fatal disease but the strong association between fataldisease and abnormal values of several laboratory markers of organ dysfunction confirm theclinical diagnosis of multi-system organ involvement as an important factor in fatal rickettsialdiseases Indeed there was significant evidence of greater injury in numerous organ systemsincluding injury to vascular endothelium in brain (confusionobtundation) lungs (tachypnea)the coagulation system (prolonged partial thromboplastin time) and skin (petechial rash) infatal cases than in non-fatal cases Systemic hypotension with decreased renal perfusion andglomerular filtration rate are also most likely the pathophysiologic mechanism of the greaterincidence of acute renal failure in fatal cases22

This study reveals a lower incidence of eschars in patients with MSF due to ISF compared toinfection with Malish strain The significant difference in eschar formation between MSF dueto ISF and Malish strain was not detected in our previous study which we attribute to a non-representative number of patients involved in our previous analysis Consistent with thisconclusion is the finding that many patients infected with Malish strain in this study also didnot have an eschar as described previously Furthermore the current study includes a largernumber of fatal cases infected with either Malish or ISF strain and some of them did not haveeschars Similarly the case fatality rate for ISF strain (29) in this study was significantlygreater (p = 002) than for Malish strain (13) and a greater portion of ISF patients (36)required admission to the ICU than Malish patients (22 p=0061) It is not yet clear whetherthe lower incidence of eschar in fatal cases due to infection with ISF strains is a prognosticfactor that indicates an ineffective or detrimental host immune response against Rickettsia ora confounding factor that causes delayed diagnosis of these patients and thus diseaseprogression We favor the first possibility because our data show that the ISF strain-infected

de SOUSA et al Page 6

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

patients received anti-rickettsial treatment earlier in the course of disease than Malish strain-infected patients

The substantial case fatality rate for MSF in Portugal differs remarkably from that reported insome other regions of the Mediterranean area mostly from countries that are near Portugalsuch as France and Spain The possibility that some Rickettsia strains are more virulent thanothers is a reasonable hypothesis to examine in such areas It is conceivable that another lessvirulent human pathogen such as R massiliae or R monacensis might be the etiologic agentof spotted fever rickettsiosis in some geographic locations such as in the northeast of Spain23 24

The reason for the greater occurrence of a history of tick bite in patients infected with Malishstrain is not evident Although different vector species or tick stages could be hypothesized tohave transmitted the infection without detection most of the cases occurred in similar monthsfor both strains and Rhipicephalus sanguineus is the only recognized tick host of R conoriiin Portugal25

The lower incidence of rash and eschar as well as Rickettsia-specific IgM or IgG antibodies infatal cases emphasizes the need to consider the diagnosis of MSF in the absence of these signsor serological markers in patients with unexplained febrile illness in endemic regions Thisstudy reveals that diagnosis by PCR amplification of rickettsial DNA from skin biopsy is auseful method to establish the etiologic agent PCR was the only method by which the diagnosiswas established in 47 patients Additionally eight patients were diagnosed by both PCR andblood culture Of a total of 109 skin biopsies examined by PCR in this study 55 (348)samples were positive and 54 (341) were negative Among the negative skin biopsiesevaluated by PCR 22 patients had rickettsial infection confirmed by serology

Further analysis of the genotype of the Rickettsia allows identification of the agent to the strainlevel with the potential for investigation of differences in specific hypothetical virulence factorsincluding adhesins (outer membrane proteins A and B) 26-28 actin mobility (RickA) 29 andmembranolytic enzymes (phospholipase D hemolysins A and C)30-32 Advances in diagnosisand knowledge of pathogenic mechanisms of rickettsioses require further attention

AcknowledgmentsWe would like to thank all physicians and laboratories from the Portuguese hospitals that have collaborated providingclinical data Teresa Luz Paulo Parreira and Liacutegia Chaiacutenho for providing excellent technical assistance and DorisBaker and Sherrill Hebert for secretarial assistance

This work was financially supported in part by a grant (AI021242) from the National Institute of Allergy and InfectiousDiseases

References1 Zhu Y Fournier PE Eremeeva M Raoult D Proposal to create subspecies of Rickettsia conorii based

on multi-locus sequence typing and an emended description of Rickettsia conorii BMC Microbiol2005511 [PubMed 15766388]

2 Walker DH Feng HM Saada JI et al Comparative antigenic analysis of spotted fever group rickettsiaefrom Israel and other closely related organisms Am J Trop Med Hyg 199552569ndash76 [PubMed7611567]

3 Sousa R Ismail N Doacuteria-Noacutebrega S et al The presence of eschars but not greater severity inPortuguese patients infected with Israeli spotted fever Ann N Y Acad Sci 20051063197ndash202[PubMed 16481514]

4 Sousa R Doacuteria S Bacellar F Torgal J Mediterranean spotted fever in Portugal Risk factors for fataloutcome in 105 hospitalized patients Ann N Y Acad Sci 2003990285ndash94 [PubMed 12860641]

de SOUSA et al Page 7

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

acute and convalescent sera which is considered as a strong evidence of recent infection Nosignificant level of R conorii IgM and IgG antibodies was detected in the fatal cases of MSF

The patient populations infected with R conorii ISF and Malish strains were similar in ageperiod of time between the onset of symptoms and first visit to the physician and length ofhospitalization of both fatal and non-fatal cases (Table 1) There was a delay of one day (amedian of five days vs four days) in treatment of patients infected with Malish strain comparedwith ISF strain No significant difference in the incidence of underlying diseases or history ofanimal contact was detected between patients infected with ISF or Malish strains (Table 2a)Although there was no difference in the seasonal distribution of patients infected with eitherISF or Malish strains history of a recognized tick bite was more common in patients infectedwith Malish strain than those infected with ISF strain (63 vs 32) (Table 2a)

Comparison of the clinical manifestations of MSF caused by Malish and ISF strains revealedtremendous overlap that would not permit clinical recognition of the strain involved Bothstrains were capable of causing disease that resulted in each of the signs symptoms andpathophysiologic conditions causing abnormal laboratory values of several markers of organdysfunction Moreover statistical analysis of the clinical signs symptoms and laboratory dataon admission revealed few differences between patients infected with Malish and ISF strains(Table 2a 2b) First an eschar was observed in a significantly higher percentage of patientswith Malish (60) than ISF strain (38) However it should be emphasized that many patientsinfected with Malish strain did not have an eschar (Table 2b) The anatomic locations of theeschars were similar in ISF and Malish infections An eschar was often found in the axillagroin (25) legs (25) and trunk (20) and only three patients infected with RconoriiMalish strain presented with more than one eschar Second ISF strain-infected patientssuffered nausea and vomiting more often than Malish strain-infected patients (Table 2a)suggesting significant gastrointestinal involvement in patients with ISF strain infection Thirda higher percentage of patients with ISF strain infection than Malish strain-infected patientshad significantly elevated serum levels of total bilirubin γ-glutamyl transferase and alkalinephosphatase among other altered parameters (Table 2b) suggesting greater hepaticinvolvement

Among children infected with R conorii four were infected with Malish strain and two withISF strain (Table 3) The most severely ill child was a 12 year old who was infected with Malishstrain and did not receive treatment until the tenth day of illness He developed encephalitisthrombocytopenia and had markedly elevated serum hepatic transaminases

The most important observation in this study was that there was statistically significantlygreater severity of disease in patients infected with R conorii ISF strain than Malish strain(Table 4a) The case fatality rate for ISF strain (29) was significantly greater (p = 002) thanfor Malish strain (13) and a greater portion of ISF patients (36) required admission to theICU than Malish patients (22 p=0061)

Analysis of the relationship of co-morbid host factors and a fatal outcome demonstrated thatalcoholism was a statistically significant host condition that was a risk factor for a fatal outcomeof MSF (Table 4b)

Analysis of the clinical and laboratory data of fatal and non-fatal cases provided a picture ofthe pathophysiology of severe MSF in a large series of patients documented specifically tohave been infected with R conorii by rickettsial isolation or molecular diagnosis and geneticidentification of the agent rather than by a diagnosis based on clinical manifestations orserology that does not distinguish among various SFG rickettsioses Fatal cases of etiologicallydocumented R conorii infection were more likely to have severe multi-organ systeminvolvement as indicated by significantly greater occurrence of a petechial rash nausea

de SOUSA et al Page 4

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

vomiting diarrhea prostration confusion andor obtundation dehydration tachypnea andhepatomegaly (Table 4a) Also fatal cases were more likely to have leukocytosis (gt 11300μl) prolonged partial thromboplastin time elevated serum concentrations of urea creatininebilirubin alanine transaminase γ-glutamyl transferase alkaline phosphatase and creatinekinase (Table 4c) at the time of admission Among the fatal cases five did not have rash oreschar at the time of admission In survivors group the presence of fever and rash weresignificantly more often present than in fatal cases Multivariate analysis identified threevariables as independent predictors associated with fatal outcome hyperbilirubinemia (OR2599 95 CI [490 13793] p lt 0001) acute renal failure (OR 1810 95 CI [189 minus17335] p= 0012) and absence of rash (OR 003 95 CI [000 minus 122] p = 0064) The resultsof the goodness-of-fit test (Hosmer- Lemeshow) showed that the model behaves with a goodfitness (p = 0 125)

Data on therapy were available in 129 adult patients (64 infected with Malish and 65 infectedwith ISF strain) Doxycycline was the drug used most often to treat MSF patients 57 (89)of patients infected with R conorii Malish strain and 46 (71 ) with ISF strain

Doxycycline was used in combination with a fluoroquinolone in two (3) patients infectedwith Malish strain and in seven (11) patients infected with ISF strain Therapy with afluoroquinolone alone was used in six (9) patients infected with ISF strain Among patientsinfected with Malish strain two (3) received no treatment two (3) were treated with a β-lactam drug and one (2) was treated with rifampin Among patients infected with ISF strainfive (8) patients received no treatment and one (2) was treated with a β-lactam drug Nostatistically significant differences (p=0472) were found in therapeutic regimens between thegroups of patients Among the seven patients who did not receive treatment two survived andfive died

Three children received treatment with a macrolide (azithromycin) one with chloramphenicoland one with doxycycline The six month old child recovered without specific treatment Nofatal cases occurred in children (Table 3)

DiscussionThis study represents the largest case series of patients with a confirmed diagnosis of MSFbased on documentation of R conorii infection and the largest series with identification of thestrain rather than only clinical or non-species specific serologic diagnosis Owing to thepresence of shared protein and lipopolysaccharide antigens among SFG rickettsiae it isextremely difficult to distinguish infections with closely related rickettsiae by serologic orimmunohistochemical methods1415 In the Mediterranean basin patients with similar clinicalmanifestations have been documented by rickettsial isolation to be infected with R sibiricamongolotimonae strain R akari and R massiliae all of whom would have producedantibodies crossreactive with R conorii11 1617 Isolation andor PCR detection followed bygenetic characterization can determine the genotype of the organism to the level of genusspecies and strain Previous studies by Sousa et al (Portugal) and Giammanco et al (Italy)compared some clinical manifestations of MSF caused by ISF and Malish strains318 This isthe first exhaustive study reporting the differences between patients infected with R conoriiMalish and ISF strains and also the microbial and host risk factors and clinical manifestationsassociated with a fatal outcome In general patients infected with ISF and Malish strains didnot show many differences in epidemiological clinical or laboratory data The most importantfindings documented are that ISF strain is more virulent than Malish strain in the populationinvestigated and alcoholism is a risk factor for fatal outcome in MSF The microbial pathogenicmechanism by which ISF strain causes more severe illness remains to be determined In regardto host-related risk factors previous studies have described alcoholic patients who developed

de SOUSA et al Page 5

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

severe MSF However the present study is the first statistical analysis with a representativesample that documents alcoholism as a risk factor for severity of the disease19

Our data suggest that gastrointestinal symptoms such as nausea vomiting and diarrhea (table4) are prominent manifestations in patients with fatal MSF However they might have beenrelated to central nervous system involvement if the intracranial pressure was increased (ieamong the fatal cases were many patients with confusionobtundation (67) and rickettsialinfection in severefatal cases may result in cerebral edema and increased intracranial pressurewhich are features of rickettsial encephalitis and nausea and vomiting could be gastrointestinalsymptoms as a reflection of CNS involvement) Thus the exact mechanism that accounts forthe presence of marked gastrointestinal symptoms in fatal cases is not yet clear Furthermorethere were significantly greater gastrointestinal manifestations such as nausea and vomiting inpatients with ISF strain infection It is possible that infection with virulent ISF strains causesinflammation of abdominal organs secondary to rickettsial growth in blood vessels in thesetissues or systemic or local production of high levels of pro-inflammatory cytokines andchemokines This conclusion is consistent with our previous study showing that patients withsevere MSF had very high dermal levels of pro-inflammatory TNF-α enzymes that mediateintracellular bacterial elimination as well as tissue injury (ie iNOS and IDO) and RANTESa chemokine that stimulates the migration and accumulation of T cells at the site of infection20

Multiple logistic regression analysis revealed an independent significant association betweenthe presence of hyperbilirubinemia and fatal outcome in a greater portion of ISF patients thanMalish patients Whether hyperbilirubinemia is related to cholestasis hemolysis or othermechanisms remains to be determined Nevertheless it is clearly evident that hepaticinvolvement marked by hepatomegaly and elevated serum ALT alkaline phosphatase and γ-glutamyl transferase levels is strongly associated with a fatal outcome Histological studieshave shown that the hepatic pathology in patients with MSF namely focal necrosis andmononuclear lobular and portal triad inflammation is not associated with fatal disease21 Thelack of correlation between pathology and fatal disease but the strong association between fataldisease and abnormal values of several laboratory markers of organ dysfunction confirm theclinical diagnosis of multi-system organ involvement as an important factor in fatal rickettsialdiseases Indeed there was significant evidence of greater injury in numerous organ systemsincluding injury to vascular endothelium in brain (confusionobtundation) lungs (tachypnea)the coagulation system (prolonged partial thromboplastin time) and skin (petechial rash) infatal cases than in non-fatal cases Systemic hypotension with decreased renal perfusion andglomerular filtration rate are also most likely the pathophysiologic mechanism of the greaterincidence of acute renal failure in fatal cases22

This study reveals a lower incidence of eschars in patients with MSF due to ISF compared toinfection with Malish strain The significant difference in eschar formation between MSF dueto ISF and Malish strain was not detected in our previous study which we attribute to a non-representative number of patients involved in our previous analysis Consistent with thisconclusion is the finding that many patients infected with Malish strain in this study also didnot have an eschar as described previously Furthermore the current study includes a largernumber of fatal cases infected with either Malish or ISF strain and some of them did not haveeschars Similarly the case fatality rate for ISF strain (29) in this study was significantlygreater (p = 002) than for Malish strain (13) and a greater portion of ISF patients (36)required admission to the ICU than Malish patients (22 p=0061) It is not yet clear whetherthe lower incidence of eschar in fatal cases due to infection with ISF strains is a prognosticfactor that indicates an ineffective or detrimental host immune response against Rickettsia ora confounding factor that causes delayed diagnosis of these patients and thus diseaseprogression We favor the first possibility because our data show that the ISF strain-infected

de SOUSA et al Page 6

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

patients received anti-rickettsial treatment earlier in the course of disease than Malish strain-infected patients

The substantial case fatality rate for MSF in Portugal differs remarkably from that reported insome other regions of the Mediterranean area mostly from countries that are near Portugalsuch as France and Spain The possibility that some Rickettsia strains are more virulent thanothers is a reasonable hypothesis to examine in such areas It is conceivable that another lessvirulent human pathogen such as R massiliae or R monacensis might be the etiologic agentof spotted fever rickettsiosis in some geographic locations such as in the northeast of Spain23 24

The reason for the greater occurrence of a history of tick bite in patients infected with Malishstrain is not evident Although different vector species or tick stages could be hypothesized tohave transmitted the infection without detection most of the cases occurred in similar monthsfor both strains and Rhipicephalus sanguineus is the only recognized tick host of R conoriiin Portugal25

The lower incidence of rash and eschar as well as Rickettsia-specific IgM or IgG antibodies infatal cases emphasizes the need to consider the diagnosis of MSF in the absence of these signsor serological markers in patients with unexplained febrile illness in endemic regions Thisstudy reveals that diagnosis by PCR amplification of rickettsial DNA from skin biopsy is auseful method to establish the etiologic agent PCR was the only method by which the diagnosiswas established in 47 patients Additionally eight patients were diagnosed by both PCR andblood culture Of a total of 109 skin biopsies examined by PCR in this study 55 (348)samples were positive and 54 (341) were negative Among the negative skin biopsiesevaluated by PCR 22 patients had rickettsial infection confirmed by serology

Further analysis of the genotype of the Rickettsia allows identification of the agent to the strainlevel with the potential for investigation of differences in specific hypothetical virulence factorsincluding adhesins (outer membrane proteins A and B) 26-28 actin mobility (RickA) 29 andmembranolytic enzymes (phospholipase D hemolysins A and C)30-32 Advances in diagnosisand knowledge of pathogenic mechanisms of rickettsioses require further attention

AcknowledgmentsWe would like to thank all physicians and laboratories from the Portuguese hospitals that have collaborated providingclinical data Teresa Luz Paulo Parreira and Liacutegia Chaiacutenho for providing excellent technical assistance and DorisBaker and Sherrill Hebert for secretarial assistance

This work was financially supported in part by a grant (AI021242) from the National Institute of Allergy and InfectiousDiseases

References1 Zhu Y Fournier PE Eremeeva M Raoult D Proposal to create subspecies of Rickettsia conorii based

on multi-locus sequence typing and an emended description of Rickettsia conorii BMC Microbiol2005511 [PubMed 15766388]

2 Walker DH Feng HM Saada JI et al Comparative antigenic analysis of spotted fever group rickettsiaefrom Israel and other closely related organisms Am J Trop Med Hyg 199552569ndash76 [PubMed7611567]

3 Sousa R Ismail N Doacuteria-Noacutebrega S et al The presence of eschars but not greater severity inPortuguese patients infected with Israeli spotted fever Ann N Y Acad Sci 20051063197ndash202[PubMed 16481514]

4 Sousa R Doacuteria S Bacellar F Torgal J Mediterranean spotted fever in Portugal Risk factors for fataloutcome in 105 hospitalized patients Ann N Y Acad Sci 2003990285ndash94 [PubMed 12860641]

de SOUSA et al Page 7

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

vomiting diarrhea prostration confusion andor obtundation dehydration tachypnea andhepatomegaly (Table 4a) Also fatal cases were more likely to have leukocytosis (gt 11300μl) prolonged partial thromboplastin time elevated serum concentrations of urea creatininebilirubin alanine transaminase γ-glutamyl transferase alkaline phosphatase and creatinekinase (Table 4c) at the time of admission Among the fatal cases five did not have rash oreschar at the time of admission In survivors group the presence of fever and rash weresignificantly more often present than in fatal cases Multivariate analysis identified threevariables as independent predictors associated with fatal outcome hyperbilirubinemia (OR2599 95 CI [490 13793] p lt 0001) acute renal failure (OR 1810 95 CI [189 minus17335] p= 0012) and absence of rash (OR 003 95 CI [000 minus 122] p = 0064) The resultsof the goodness-of-fit test (Hosmer- Lemeshow) showed that the model behaves with a goodfitness (p = 0 125)

Data on therapy were available in 129 adult patients (64 infected with Malish and 65 infectedwith ISF strain) Doxycycline was the drug used most often to treat MSF patients 57 (89)of patients infected with R conorii Malish strain and 46 (71 ) with ISF strain

Doxycycline was used in combination with a fluoroquinolone in two (3) patients infectedwith Malish strain and in seven (11) patients infected with ISF strain Therapy with afluoroquinolone alone was used in six (9) patients infected with ISF strain Among patientsinfected with Malish strain two (3) received no treatment two (3) were treated with a β-lactam drug and one (2) was treated with rifampin Among patients infected with ISF strainfive (8) patients received no treatment and one (2) was treated with a β-lactam drug Nostatistically significant differences (p=0472) were found in therapeutic regimens between thegroups of patients Among the seven patients who did not receive treatment two survived andfive died

Three children received treatment with a macrolide (azithromycin) one with chloramphenicoland one with doxycycline The six month old child recovered without specific treatment Nofatal cases occurred in children (Table 3)

DiscussionThis study represents the largest case series of patients with a confirmed diagnosis of MSFbased on documentation of R conorii infection and the largest series with identification of thestrain rather than only clinical or non-species specific serologic diagnosis Owing to thepresence of shared protein and lipopolysaccharide antigens among SFG rickettsiae it isextremely difficult to distinguish infections with closely related rickettsiae by serologic orimmunohistochemical methods1415 In the Mediterranean basin patients with similar clinicalmanifestations have been documented by rickettsial isolation to be infected with R sibiricamongolotimonae strain R akari and R massiliae all of whom would have producedantibodies crossreactive with R conorii11 1617 Isolation andor PCR detection followed bygenetic characterization can determine the genotype of the organism to the level of genusspecies and strain Previous studies by Sousa et al (Portugal) and Giammanco et al (Italy)compared some clinical manifestations of MSF caused by ISF and Malish strains318 This isthe first exhaustive study reporting the differences between patients infected with R conoriiMalish and ISF strains and also the microbial and host risk factors and clinical manifestationsassociated with a fatal outcome In general patients infected with ISF and Malish strains didnot show many differences in epidemiological clinical or laboratory data The most importantfindings documented are that ISF strain is more virulent than Malish strain in the populationinvestigated and alcoholism is a risk factor for fatal outcome in MSF The microbial pathogenicmechanism by which ISF strain causes more severe illness remains to be determined In regardto host-related risk factors previous studies have described alcoholic patients who developed

de SOUSA et al Page 5

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

severe MSF However the present study is the first statistical analysis with a representativesample that documents alcoholism as a risk factor for severity of the disease19

Our data suggest that gastrointestinal symptoms such as nausea vomiting and diarrhea (table4) are prominent manifestations in patients with fatal MSF However they might have beenrelated to central nervous system involvement if the intracranial pressure was increased (ieamong the fatal cases were many patients with confusionobtundation (67) and rickettsialinfection in severefatal cases may result in cerebral edema and increased intracranial pressurewhich are features of rickettsial encephalitis and nausea and vomiting could be gastrointestinalsymptoms as a reflection of CNS involvement) Thus the exact mechanism that accounts forthe presence of marked gastrointestinal symptoms in fatal cases is not yet clear Furthermorethere were significantly greater gastrointestinal manifestations such as nausea and vomiting inpatients with ISF strain infection It is possible that infection with virulent ISF strains causesinflammation of abdominal organs secondary to rickettsial growth in blood vessels in thesetissues or systemic or local production of high levels of pro-inflammatory cytokines andchemokines This conclusion is consistent with our previous study showing that patients withsevere MSF had very high dermal levels of pro-inflammatory TNF-α enzymes that mediateintracellular bacterial elimination as well as tissue injury (ie iNOS and IDO) and RANTESa chemokine that stimulates the migration and accumulation of T cells at the site of infection20

Multiple logistic regression analysis revealed an independent significant association betweenthe presence of hyperbilirubinemia and fatal outcome in a greater portion of ISF patients thanMalish patients Whether hyperbilirubinemia is related to cholestasis hemolysis or othermechanisms remains to be determined Nevertheless it is clearly evident that hepaticinvolvement marked by hepatomegaly and elevated serum ALT alkaline phosphatase and γ-glutamyl transferase levels is strongly associated with a fatal outcome Histological studieshave shown that the hepatic pathology in patients with MSF namely focal necrosis andmononuclear lobular and portal triad inflammation is not associated with fatal disease21 Thelack of correlation between pathology and fatal disease but the strong association between fataldisease and abnormal values of several laboratory markers of organ dysfunction confirm theclinical diagnosis of multi-system organ involvement as an important factor in fatal rickettsialdiseases Indeed there was significant evidence of greater injury in numerous organ systemsincluding injury to vascular endothelium in brain (confusionobtundation) lungs (tachypnea)the coagulation system (prolonged partial thromboplastin time) and skin (petechial rash) infatal cases than in non-fatal cases Systemic hypotension with decreased renal perfusion andglomerular filtration rate are also most likely the pathophysiologic mechanism of the greaterincidence of acute renal failure in fatal cases22

This study reveals a lower incidence of eschars in patients with MSF due to ISF compared toinfection with Malish strain The significant difference in eschar formation between MSF dueto ISF and Malish strain was not detected in our previous study which we attribute to a non-representative number of patients involved in our previous analysis Consistent with thisconclusion is the finding that many patients infected with Malish strain in this study also didnot have an eschar as described previously Furthermore the current study includes a largernumber of fatal cases infected with either Malish or ISF strain and some of them did not haveeschars Similarly the case fatality rate for ISF strain (29) in this study was significantlygreater (p = 002) than for Malish strain (13) and a greater portion of ISF patients (36)required admission to the ICU than Malish patients (22 p=0061) It is not yet clear whetherthe lower incidence of eschar in fatal cases due to infection with ISF strains is a prognosticfactor that indicates an ineffective or detrimental host immune response against Rickettsia ora confounding factor that causes delayed diagnosis of these patients and thus diseaseprogression We favor the first possibility because our data show that the ISF strain-infected

de SOUSA et al Page 6

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

patients received anti-rickettsial treatment earlier in the course of disease than Malish strain-infected patients

The substantial case fatality rate for MSF in Portugal differs remarkably from that reported insome other regions of the Mediterranean area mostly from countries that are near Portugalsuch as France and Spain The possibility that some Rickettsia strains are more virulent thanothers is a reasonable hypothesis to examine in such areas It is conceivable that another lessvirulent human pathogen such as R massiliae or R monacensis might be the etiologic agentof spotted fever rickettsiosis in some geographic locations such as in the northeast of Spain23 24

The reason for the greater occurrence of a history of tick bite in patients infected with Malishstrain is not evident Although different vector species or tick stages could be hypothesized tohave transmitted the infection without detection most of the cases occurred in similar monthsfor both strains and Rhipicephalus sanguineus is the only recognized tick host of R conoriiin Portugal25

The lower incidence of rash and eschar as well as Rickettsia-specific IgM or IgG antibodies infatal cases emphasizes the need to consider the diagnosis of MSF in the absence of these signsor serological markers in patients with unexplained febrile illness in endemic regions Thisstudy reveals that diagnosis by PCR amplification of rickettsial DNA from skin biopsy is auseful method to establish the etiologic agent PCR was the only method by which the diagnosiswas established in 47 patients Additionally eight patients were diagnosed by both PCR andblood culture Of a total of 109 skin biopsies examined by PCR in this study 55 (348)samples were positive and 54 (341) were negative Among the negative skin biopsiesevaluated by PCR 22 patients had rickettsial infection confirmed by serology

Further analysis of the genotype of the Rickettsia allows identification of the agent to the strainlevel with the potential for investigation of differences in specific hypothetical virulence factorsincluding adhesins (outer membrane proteins A and B) 26-28 actin mobility (RickA) 29 andmembranolytic enzymes (phospholipase D hemolysins A and C)30-32 Advances in diagnosisand knowledge of pathogenic mechanisms of rickettsioses require further attention

AcknowledgmentsWe would like to thank all physicians and laboratories from the Portuguese hospitals that have collaborated providingclinical data Teresa Luz Paulo Parreira and Liacutegia Chaiacutenho for providing excellent technical assistance and DorisBaker and Sherrill Hebert for secretarial assistance

This work was financially supported in part by a grant (AI021242) from the National Institute of Allergy and InfectiousDiseases

References1 Zhu Y Fournier PE Eremeeva M Raoult D Proposal to create subspecies of Rickettsia conorii based

on multi-locus sequence typing and an emended description of Rickettsia conorii BMC Microbiol2005511 [PubMed 15766388]

2 Walker DH Feng HM Saada JI et al Comparative antigenic analysis of spotted fever group rickettsiaefrom Israel and other closely related organisms Am J Trop Med Hyg 199552569ndash76 [PubMed7611567]

3 Sousa R Ismail N Doacuteria-Noacutebrega S et al The presence of eschars but not greater severity inPortuguese patients infected with Israeli spotted fever Ann N Y Acad Sci 20051063197ndash202[PubMed 16481514]

4 Sousa R Doacuteria S Bacellar F Torgal J Mediterranean spotted fever in Portugal Risk factors for fataloutcome in 105 hospitalized patients Ann N Y Acad Sci 2003990285ndash94 [PubMed 12860641]

de SOUSA et al Page 7

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

severe MSF However the present study is the first statistical analysis with a representativesample that documents alcoholism as a risk factor for severity of the disease19

Our data suggest that gastrointestinal symptoms such as nausea vomiting and diarrhea (table4) are prominent manifestations in patients with fatal MSF However they might have beenrelated to central nervous system involvement if the intracranial pressure was increased (ieamong the fatal cases were many patients with confusionobtundation (67) and rickettsialinfection in severefatal cases may result in cerebral edema and increased intracranial pressurewhich are features of rickettsial encephalitis and nausea and vomiting could be gastrointestinalsymptoms as a reflection of CNS involvement) Thus the exact mechanism that accounts forthe presence of marked gastrointestinal symptoms in fatal cases is not yet clear Furthermorethere were significantly greater gastrointestinal manifestations such as nausea and vomiting inpatients with ISF strain infection It is possible that infection with virulent ISF strains causesinflammation of abdominal organs secondary to rickettsial growth in blood vessels in thesetissues or systemic or local production of high levels of pro-inflammatory cytokines andchemokines This conclusion is consistent with our previous study showing that patients withsevere MSF had very high dermal levels of pro-inflammatory TNF-α enzymes that mediateintracellular bacterial elimination as well as tissue injury (ie iNOS and IDO) and RANTESa chemokine that stimulates the migration and accumulation of T cells at the site of infection20

Multiple logistic regression analysis revealed an independent significant association betweenthe presence of hyperbilirubinemia and fatal outcome in a greater portion of ISF patients thanMalish patients Whether hyperbilirubinemia is related to cholestasis hemolysis or othermechanisms remains to be determined Nevertheless it is clearly evident that hepaticinvolvement marked by hepatomegaly and elevated serum ALT alkaline phosphatase and γ-glutamyl transferase levels is strongly associated with a fatal outcome Histological studieshave shown that the hepatic pathology in patients with MSF namely focal necrosis andmononuclear lobular and portal triad inflammation is not associated with fatal disease21 Thelack of correlation between pathology and fatal disease but the strong association between fataldisease and abnormal values of several laboratory markers of organ dysfunction confirm theclinical diagnosis of multi-system organ involvement as an important factor in fatal rickettsialdiseases Indeed there was significant evidence of greater injury in numerous organ systemsincluding injury to vascular endothelium in brain (confusionobtundation) lungs (tachypnea)the coagulation system (prolonged partial thromboplastin time) and skin (petechial rash) infatal cases than in non-fatal cases Systemic hypotension with decreased renal perfusion andglomerular filtration rate are also most likely the pathophysiologic mechanism of the greaterincidence of acute renal failure in fatal cases22

This study reveals a lower incidence of eschars in patients with MSF due to ISF compared toinfection with Malish strain The significant difference in eschar formation between MSF dueto ISF and Malish strain was not detected in our previous study which we attribute to a non-representative number of patients involved in our previous analysis Consistent with thisconclusion is the finding that many patients infected with Malish strain in this study also didnot have an eschar as described previously Furthermore the current study includes a largernumber of fatal cases infected with either Malish or ISF strain and some of them did not haveeschars Similarly the case fatality rate for ISF strain (29) in this study was significantlygreater (p = 002) than for Malish strain (13) and a greater portion of ISF patients (36)required admission to the ICU than Malish patients (22 p=0061) It is not yet clear whetherthe lower incidence of eschar in fatal cases due to infection with ISF strains is a prognosticfactor that indicates an ineffective or detrimental host immune response against Rickettsia ora confounding factor that causes delayed diagnosis of these patients and thus diseaseprogression We favor the first possibility because our data show that the ISF strain-infected

de SOUSA et al Page 6

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

patients received anti-rickettsial treatment earlier in the course of disease than Malish strain-infected patients

The substantial case fatality rate for MSF in Portugal differs remarkably from that reported insome other regions of the Mediterranean area mostly from countries that are near Portugalsuch as France and Spain The possibility that some Rickettsia strains are more virulent thanothers is a reasonable hypothesis to examine in such areas It is conceivable that another lessvirulent human pathogen such as R massiliae or R monacensis might be the etiologic agentof spotted fever rickettsiosis in some geographic locations such as in the northeast of Spain23 24

The reason for the greater occurrence of a history of tick bite in patients infected with Malishstrain is not evident Although different vector species or tick stages could be hypothesized tohave transmitted the infection without detection most of the cases occurred in similar monthsfor both strains and Rhipicephalus sanguineus is the only recognized tick host of R conoriiin Portugal25

The lower incidence of rash and eschar as well as Rickettsia-specific IgM or IgG antibodies infatal cases emphasizes the need to consider the diagnosis of MSF in the absence of these signsor serological markers in patients with unexplained febrile illness in endemic regions Thisstudy reveals that diagnosis by PCR amplification of rickettsial DNA from skin biopsy is auseful method to establish the etiologic agent PCR was the only method by which the diagnosiswas established in 47 patients Additionally eight patients were diagnosed by both PCR andblood culture Of a total of 109 skin biopsies examined by PCR in this study 55 (348)samples were positive and 54 (341) were negative Among the negative skin biopsiesevaluated by PCR 22 patients had rickettsial infection confirmed by serology

Further analysis of the genotype of the Rickettsia allows identification of the agent to the strainlevel with the potential for investigation of differences in specific hypothetical virulence factorsincluding adhesins (outer membrane proteins A and B) 26-28 actin mobility (RickA) 29 andmembranolytic enzymes (phospholipase D hemolysins A and C)30-32 Advances in diagnosisand knowledge of pathogenic mechanisms of rickettsioses require further attention

AcknowledgmentsWe would like to thank all physicians and laboratories from the Portuguese hospitals that have collaborated providingclinical data Teresa Luz Paulo Parreira and Liacutegia Chaiacutenho for providing excellent technical assistance and DorisBaker and Sherrill Hebert for secretarial assistance

This work was financially supported in part by a grant (AI021242) from the National Institute of Allergy and InfectiousDiseases

References1 Zhu Y Fournier PE Eremeeva M Raoult D Proposal to create subspecies of Rickettsia conorii based

on multi-locus sequence typing and an emended description of Rickettsia conorii BMC Microbiol2005511 [PubMed 15766388]

2 Walker DH Feng HM Saada JI et al Comparative antigenic analysis of spotted fever group rickettsiaefrom Israel and other closely related organisms Am J Trop Med Hyg 199552569ndash76 [PubMed7611567]

3 Sousa R Ismail N Doacuteria-Noacutebrega S et al The presence of eschars but not greater severity inPortuguese patients infected with Israeli spotted fever Ann N Y Acad Sci 20051063197ndash202[PubMed 16481514]

4 Sousa R Doacuteria S Bacellar F Torgal J Mediterranean spotted fever in Portugal Risk factors for fataloutcome in 105 hospitalized patients Ann N Y Acad Sci 2003990285ndash94 [PubMed 12860641]

de SOUSA et al Page 7

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

patients received anti-rickettsial treatment earlier in the course of disease than Malish strain-infected patients

The substantial case fatality rate for MSF in Portugal differs remarkably from that reported insome other regions of the Mediterranean area mostly from countries that are near Portugalsuch as France and Spain The possibility that some Rickettsia strains are more virulent thanothers is a reasonable hypothesis to examine in such areas It is conceivable that another lessvirulent human pathogen such as R massiliae or R monacensis might be the etiologic agentof spotted fever rickettsiosis in some geographic locations such as in the northeast of Spain23 24

The reason for the greater occurrence of a history of tick bite in patients infected with Malishstrain is not evident Although different vector species or tick stages could be hypothesized tohave transmitted the infection without detection most of the cases occurred in similar monthsfor both strains and Rhipicephalus sanguineus is the only recognized tick host of R conoriiin Portugal25

The lower incidence of rash and eschar as well as Rickettsia-specific IgM or IgG antibodies infatal cases emphasizes the need to consider the diagnosis of MSF in the absence of these signsor serological markers in patients with unexplained febrile illness in endemic regions Thisstudy reveals that diagnosis by PCR amplification of rickettsial DNA from skin biopsy is auseful method to establish the etiologic agent PCR was the only method by which the diagnosiswas established in 47 patients Additionally eight patients were diagnosed by both PCR andblood culture Of a total of 109 skin biopsies examined by PCR in this study 55 (348)samples were positive and 54 (341) were negative Among the negative skin biopsiesevaluated by PCR 22 patients had rickettsial infection confirmed by serology

Further analysis of the genotype of the Rickettsia allows identification of the agent to the strainlevel with the potential for investigation of differences in specific hypothetical virulence factorsincluding adhesins (outer membrane proteins A and B) 26-28 actin mobility (RickA) 29 andmembranolytic enzymes (phospholipase D hemolysins A and C)30-32 Advances in diagnosisand knowledge of pathogenic mechanisms of rickettsioses require further attention

AcknowledgmentsWe would like to thank all physicians and laboratories from the Portuguese hospitals that have collaborated providingclinical data Teresa Luz Paulo Parreira and Liacutegia Chaiacutenho for providing excellent technical assistance and DorisBaker and Sherrill Hebert for secretarial assistance

This work was financially supported in part by a grant (AI021242) from the National Institute of Allergy and InfectiousDiseases

References1 Zhu Y Fournier PE Eremeeva M Raoult D Proposal to create subspecies of Rickettsia conorii based

on multi-locus sequence typing and an emended description of Rickettsia conorii BMC Microbiol2005511 [PubMed 15766388]

2 Walker DH Feng HM Saada JI et al Comparative antigenic analysis of spotted fever group rickettsiaefrom Israel and other closely related organisms Am J Trop Med Hyg 199552569ndash76 [PubMed7611567]

3 Sousa R Ismail N Doacuteria-Noacutebrega S et al The presence of eschars but not greater severity inPortuguese patients infected with Israeli spotted fever Ann N Y Acad Sci 20051063197ndash202[PubMed 16481514]

4 Sousa R Doacuteria S Bacellar F Torgal J Mediterranean spotted fever in Portugal Risk factors for fataloutcome in 105 hospitalized patients Ann N Y Acad Sci 2003990285ndash94 [PubMed 12860641]

de SOUSA et al Page 7

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

5 Amaro M Bacellar F Franca A Report of eight cases of fatal and severe Mediterranean spotted feverin Portugal Ann N Y Acad Sci 2003990331ndash43 [PubMed 12860647]

6 Administraccedilatildeo Central do Sistema de Sauacutede Portuguese Ministry of Health wwwacssmin-saudeptwwwacssmin-saudept

7 Bacellar F Beati L Franca A Pocas J Regnery R Filipe A Israeli spotted fever Rickettsia (Rickettsiaconorii complex) associated with human diseases in Portugal Emerg Infect Dis 19995835ndash6[PubMed 10603225]

8 Gross EM Yagupsky P Torok V Goldwasser RA Resurgence of Mediterranean spotted fever Lancet1982131107

9 Regev-Yochay G Segal E Rubinstein E Glucose-6-phosphate dehydrogenase deficiency Possibledeterminant for a fulminant course of Israeli spotted fever Isr Med Assoc J 20002781ndash2 [PubMed11344735]

10 Gross EM Yagupsky P Israeli rickettsial spotted fever in children A review of 54 cases Acta Trop19874491ndash6 [PubMed 2884843]

11 Sousa R Barata C Vitorino L et al Rickettsia sibirica isolation from a patient and detection in ticksPortugal Emerg Infect Dis 2006121103ndash8 [PubMed 16836827]

12 La Scola B Raoult D Laboratory diagnosis of rickettsioses Current approaches to diagnosis of oldand new rickettsial diseases J Clin Microbiol 1997352715ndash27 [PubMed 9350721]

13 Bacellar F Sousa R Santos A Santos-Silva M Parola P Boutonneuse fever in Portugal 1995minus2000Data of a state laboratory Eur J Epidemiol 200318275ndash7 [PubMed 12800955]

14 Cwikel BJ Ighbarieh J Sarov I Antigenic polypeptides of Israeli spotted fever isolates comparedwith other spotted fever group rickettsiae Ann N Y Acad Sci 1990590381ndash8 [PubMed 2378464]

15 Goldwasser RA Klingberg W Steiman Y Swartz TA The isolation of strains of rickettsiae of thespotted fever group in Israel and their differentiation from other members of the group byimmunofluorescence methods Scand J Infect Dis 1974653ndash6 [PubMed 4208205]

16 Radulovic S Feng H-M Morovic M et al Isolation of Rickettsia akari from a patient in a regionwhere Mediterranean spotted fever is endemic Clin Infect Dis 199622216ndash20 [PubMed 8838175]

17 Vitale G Mansueto S Rolain JM Raoult D Rickettsia massiliae human isolation Emerg Infect Dis200612174ndash175 [PubMed 16634183]

18 Giammanco GM Vitale G Mansueto S Capra G Caleca MP Ammatuna P Presence of Rickettsiaconorii subsp israelensis the causative agent of Israeli spotted fever in Sicily Italy ascertained ina retrospective study J Clin Microbiol 2005436027ndash31 [PubMed 16333093]

19 Raoult D Zuchelli P et al Incidence clinical observations and risk factors in the severe form ofMediterranean spotted fever among patients admitted to hospital in Marseilles 1983minus1984 J Infect198612111ndash16 [PubMed 3701097]

20 De Sousa R Ismail N Nobrega SD et al Intralesional expression of mRNA of interferon-γ tumornecrosis factor-α interleukin-10 nitric oxide synthase indoleamine-2 3-dioxygenase and RANTESis a major immune effector in Mediterranean spotted fever rickettsiosis J Infect Dis 2007196770ndash81 [PubMed 17674321]

21 Walker DH Staiti A Mansueto S Triangali G Frequent occurrence of hepatic lesions in boutonneusefever Acta Trop 198643175ndash81 [PubMed 2874714]

22 Walker DH Mattern WD Acute renal failure in Rocky Mountain spotted fever Arch Intern Med1979139443ndash48 [PubMed 434998]

23 Font-Creus B Bella-Cueto F Espejo-Arenas E et al Mediterranean spotted fever a cooperativestudy of 227 cases Rev Infect Dis 19857635ndash42 [PubMed 3903943]

24 Jado I Oteo J Aldaacutemiz M et al Rickettsia monacensis and human disease Spain Emerg Infect Dis2007131405ndash7 [PubMed 18252123]

25 De Sousa R Santos- Silva M Santos AS et al Rickettsia conorii Israeli tick typhus strain isolatedfrom Rhipicephalus sanguineus ticks in Portugal Vector Borne Zoonotic Dis 20077444ndash47[PubMed 17767403]

26 Li H Walker DH rOmpA is a critical protein for the adhesion of Rickettsia rickettsii to host cellsMicrobial Pathogen 199824289ndash298

de SOUSA et al Page 8

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

27 Martinez JJ Seveau S Veiga E Matsuyama S Cossart P Ku70 a component of DNA-dependentprotein kinase is a mammalian receptor for Rickettsia conorii Cell 20051231013ndash23 [PubMed16360032]

28 Uchiyama T Kawano H Kusuhara Y The major outer membrane protein rOmpB of spotted fevergroup rickettsiae functions in the rickettsial adherence to and invasion of Vero cells Microbes andInfect 20068801ndash809 [PubMed 16500128]

29 Gouin E Egile C Dehoux P et al The RickA protein of Rickettsia conorii activates the Arp23complex Nature 2004427457ndash61 [PubMed 14749835]

30 Renesto P Dehoux P Gouin E Touqui L Cossart P Raoult D Identification and characterization ofa phospholipase D-superfamily gene in rickettsiae J Infect Dis 20031881276ndash83 [PubMed14593584]

31 Whitworth T Popov VL Yu X-J Walker DH Bouyer DH Expression of the Rickettsia prowazekiipld or tlyC gene in Salmonella enterica serovar typhimurium mediates phagosomal escape InfectImmun 2005736668ndash73 [PubMed 16177343]

32 Radulovic S Troyer JM Beier MS Lau AOT Azad AF Identification and molecular analysis of thegene encoding Rickettsia typhi hemolysin Infect Immun 1999676104ndash8 [PubMed 10531273]

de SOUSA et al Page 9

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 10

Table 1Age and intervals (days) of events for patients infected by Malish or Israel spotted fever rickettsial strain

Status Variable R conorii Malish strain R conorii ISF strain p-value

Survived Age (median) 63 years 57 years 03261st symptom to 1st visit 4 days 3 days 01891st visit to specific therapy 0 days 0 days 02411st symptom to specific therapy 5 days 4 days 0028

emsp Length of stay 7 days 6 days 0081

Fatal Age (median) 66 years 575 years 01141st symptom to 1st visit 4 days 3 days 06181st visit to specific therapy 0 days 05 days 03711st symptom to specific therapy 45 days 4 days 0899Length of stay 2 days 1 days 0203

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 11

Table 2aComparative analysis of recognized potential epidemiologic exposure clinical signs and symptoms and underlyingdiseases of patients infected by R conorii Malish strain versus Israeli spotted fever strain

Epidemiologic exposure clinicalcharacteristics and co-morbidities

Total Patients (n)a R conorii Malishstrain n ()b

R conoriiISFstrain n ()b

p-value

Epidemiologic exposureAnimal contact 88 44 (94) 36 (88) 0465Tick bite history 66 20 (63) 11 (32) 0014Clinical signs and symptomsFever 137 64 (94) 65 (94) 1Eschar 123 38 (60) 23 (38) 0015Rash 137 65 (96) 63 (91) 0463Type of rash Maculopapular 120 61(94) 59 (94) 1 Petechial 8 4 (6) 4 (6) 1Asthenia 80 39 (95) 35 (90) 0426Headache 77 32 (78) 28 (78) 0977Myalgia 85 38 (84) 34 (85) 0943Arthralgia 69 22 (63) 16 (47) 0187Abdominal Pain 70 7 (19) 12 (36) 0101Anorexia 65 27 (73) 19 (68) 0653Nausea 75 14 (35) 22 (63) 0016Vomiting 84 13 (30) 23 (56) 0017Diarrhea 79 11 (28) 17 (44) 0135Prostration 81 27 (64) 26 (67) 0822Confusion obtundation 100 19 (38) 15 (30) 0398Dehydration 72 14 (39) 17 (47) 0475Tachypnea 63 10 (38) 18 (49) 0452Chest pain 53 3 (14) 6 (19) 1Hepatomegaly 60 9 (29) 9 (31) 0866Splenomegaly 57 2 (7) 3 (11) 0660Co-morbiditiesAlcoholism 103 9 (18) 9 (17) 0892Tobacco use 87 6 (15) 7 (15) 0939Diabetes 107 5 (10) 4 (7) 0737Cardiac failure 105 8 (16) 6 (11) 0443Respiratory failure 85 2 (5) 4 (9) 0677Renal failure 85 3 (7) 0 0116Hypertension 89 10 (22) 7 (16) 0449

aTotal patients with available data

bthe percentage is based on total patients infected by each strain

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 12

Table 2bComparative analysis of admission laboratory data of patients infected by R conorii Malish strain versus Israeli spottedfever strain

Laboratory Observation Total Patientsa(n) R conorii Malishstrain n ()b

R conorii ISFstrain n ()b

p- value

Anemia lt11gdl 118 13 (22) 8 (13) 0197Leukopenia lt 4400μl 121 6 (10) 10 (16) 0571Leukocytosis gt 11300μl 121 12 (20) 10 (16) 0571Thrombocytopenia (Platelets PLT) 122 45 (76) 55 (87) 0113Prothrombin time (PT) gt 13 seg 84 19 (48) 24 (55) 0544Partial thromboplastin time (PTT)PTT gt35 seg 95 17 (35) 22 (47) 0598Glucose gt 110 mgdl 101 33 (66) 29 (57) 0346Urea 112 29 (56) 32 (53) 0796Creatinine gt 12 mgdl 120 36 (60) 32 (53) 0461Sodium lt 145 mmoll 120 30 (52) 26 (42) 0283Potassium lt 35 mmoll 116 10 (18) 7 (11) 0308Total bilirubin gt12 mgdl 109 14 (26) 25 (45) 0047Alanine transaminase (ALT) 120 37 (64) 45 (73) 0301Aspartate transaminase (AST) 112 42 (79) 53 (90) 0119γ- Glutamyl transferase (GGT) 77 17 (50) 32 (74) 0027Alkaline phosphatase (ALP) 88 10 (26) 24 (49) 0026Creatine kinase (CK) 75 19 (51) 13 (34) 0133C-reactive protein (CRP)CRP gt 05 mgdl 66 34 (97) 31 (100) 0343

aChildren were excluded from analyses of laboratory data

bThe percentage is based on total patients infected by each strain

The abnormal values were considered out of normal range (lower or elevated values) limits established by each different hospital laboratory

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 13Ta

ble

3C

linic

al a

nd la

bora

tory

obs

erva

tions

and

trea

tmen

t in

6 ch

ildre

n w

ith M

edite

rran

ean

spot

ted

feve

r

Patie

nt 1

Patie

nt 2

Patie

nt 3

Patie

nt 4

Patie

nt 5

Patie

nt 6

Infe

ctin

g St

rain

ISF

ISF

Mal

ish

Mal

ish

Mal

ish

Mal

ish

Age

(yea

r or m

onth

)4y

6y1y

5 m

onth

6 m

onth

12y

1st sy

mpt

oms t

o sp

ecifi

c th

erap

y(d

ays)

24

4

10

Adm

issi

on to

dis

char

ge (d

ays)

09

99

011

Feve

rye

sye

sye

sye

sye

sye

sEs

char

nono

nono

noye

sR

ash

mac

ulop

apul

arpe

tech

ial

mac

ulop

apul

arm

acul

opap

ular

mac

ulop

opul

arm

acul

opap

ular

Hea

dach

eye

sM

yalg

iaye

sye

sG

astro

inte

stin

al si

gns

vom

iting

vom

iting

Hep

atom

egal

ySp

leno

meg

aly

yes

Com

plic

atio

nsen

ceph

aliti

sTh

erap

eutic

saz

ithro

myc

indo

xycy

clin

eaz

ithro

myc

inch

lora

mph

enic

olno

t tre

ated

azith

rom

ycin

Lab

orat

ory

data

Leuc

ocyt

es (c

ells

μl)

8920

2260

9930

178

0013

870

Plat

elet

s (ce

lls μ

l)17

700

010

600

013

700

017

700

088

000

Ure

a (m

gdl

)21

1527

--

Cre

atin

ine

(mg

dl)

04

05

03

06

-A

LT (U

L)

5624

634

7-

826

AST

(UL

)58

285

365

-11

17

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 14Ta

ble

4aU

niva

riate

ana

lysi

s of e

pide

mio

logi

c an

d cl

inic

al fi

ndin

gs in

fata

l and

non

-fat

al M

SF c

ases

Epi

dem

iolo

gic

and

clin

ical

char

acte

rist

ics

Tot

al p

atie

nts (

n)Su

rviv

ors n

()

Fata

l cas

es n

()

Odd

s Rat

io (O

R)

adju

sted

by

stra

in95

C

Ip-

val

ue

Ric

ketts

ial s

train

Mal

ish

7162

(87

)9

(13

)1

mdashmdash

IS

F69

49 (7

1)

20 (2

9)

281

118

minus 6

72

002

0A

nim

al c

onta

ct88

69 (9

0)

11 (1

00

)a

aa

Tick

bite

his

tory

6627

(50

)4

(33

)0

680

17 minus

27

20

587

Feve

r13

710

6 (9

7)

23 (8

2)

011

023

minus 0

54

000

6Es

char

123

53 (5

2)

8 (3

6)

059

022

minus 1

56

058

9R

ash

137

105

(97

)23

(79

)0

120

03 minus

05

20

005

Type

of r

ash

Mac

ulop

apul

ar10

2 (9

7)

18 (7

8)

1-

-

P

etec

hial

3 (3

)

5 (2

2)

109

32

29 minus

53

830

003

Ast

heni

a80

59 (9

4)

15 (8

8)

063

009

7 minus

412

10

633

Hea

dach

e77

47 (7

6)

13 (8

7)

234

043

minus 1

273

032

6M

yalg

ia85

58 (8

3)

14 (9

3)

308

035

minus 2

719

031

1A

rthra

lgia

6931

(55

)7

(54

)1

190

33 minus

42

30

792

Abd

omin

al P

ain

7011

(29

)8

(57

)4

561

23 minus

16

850

023

Ano

rexi

a65

37 (6

6)

9 (8

2)

258

046

minus 1

450

028

2N

ause

a75

25 (4

0)

11 (9

2)

125

148

minus 1

069

40

021

Vom

iting

8422

(33

)14

(82

)7

51

88 minus

30

120

004

Dia

rrhe

a79

16 (2

6)

12 (7

1)

638

181

minus 2

242

000

4Pr

ostra

tion

8136

(58

)17

(89

)6

731

37 minus

33

030

019

Con

fusi

on o

btun

datio

n10

022

(27

)12

(67

)9

262

61 minus

32

910

001

Deh

ydra

tion

7218

(33

)13

(76

)7

551

95 minus

29

290

003

Tach

ypne

a63

11 (2

5)

17 (8

9)

291

75

35 minus

158

97

lt00

01C

hest

pai

n53

6 (1

5)

3 (2

5)

179

035

minus 9

18

048

6H

epat

omeg

aly

6012

(24

)6

(67

)7

371

46 minus

37

040

015

Sple

nom

egal

y57

4 (8

)

1 (1

7)

200

017

minus 2

377

058

4

a Odd

s rat

io c

ould

not

be

calc

ulat

ed si

nce

ther

e w

ere

no fa

tal c

ases

with

nor

mal

val

ues

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 15Ta

ble

4bC

ompa

rison

of e

xist

ence

of p

re-e

xist

ing

co-m

orbi

ditie

s in

fata

l and

non

-fat

al c

ases

of M

edite

rran

ean

spot

ted

feve

r un

ivar

iate

ana

lysi

s

Co-

mor

bidi

ties

Tot

al P

atie

ntsa (n

)Su

rviv

ors n

()

Fata

l cas

es n

()

Odd

s Rat

io (O

R)

adju

sted

by

stra

in95

C

Ip-

val

ue

Alc

ohol

ism

103

11 (1

3)

7 (4

7)

699

41

99 minus

24

640

002

Toba

cco

use

8710

(14

)3

(18

)1

286

030

minus 5

54

073

5D

iabe

tes

107

8 (9

)

1 (6

)

074

50

08 minus

66

50

793

Car

diac

failu

re10

511

(13

)3

(18

)1

773

041

minus 7

60

044

0R

espi

rato

ry fa

ilure

855

(7

)1

(6

)0

705

007

minus 6

79

076

3C

hron

ic re

nal f

ailu

re85

3 (4

)

0-

--

Hyp

erte

nsio

n89

13 (1

8)

4 (2

5)

183

90

48 minus

70

70

375

a Tota

l pat

ient

s with

ava

ilabl

e da

ta

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 16Ta

ble

4cU

niva

riate

ana

lysi

s of l

abor

ator

y fin

ding

s in

fata

l and

non

-fat

al M

SF c

ases

Lab

orat

ory

Cha

ract

eris

tics

Tot

al P

atie

ntsa (n

)Su

rviv

ors n

()

Fata

l cas

es n

()

Odd

s Rat

io(O

R) a

djus

ted

by st

rain

95

CI

p- v

alue

Ane

mia

lt11

gd

l11

818

(20

)3

(11

)0

590

15 minus

22

00

424

Leuk

open

ia lt

440

0μl

121

13 (1

4)

3 (1

2)

130

031

minus 5

59

071

6Le

ukoc

ytos

is gt

113

00μ

l12

110

(11

)12

(46

)11

13

337

minus 3

680

lt00

01Th

rom

bocy

tope

nia

(Pla

tele

ts P

LT)

122

73 (7

8)

27 (9

6)

68

086

minus 5

378

006

9Pr

othr

ombi

n tim

e (P

T) gt

13

seg

8434

(50

)9

(60

)2

140

59 minus

77

40

246

Parti

al th

rom

bopl

astin

tim

e (P

TT)

PTT

gt35

seg

9522

(29

)17

(85

)36

78

453

minus 2

984

40

001

Glu

cose

gt 1

10 m

gdl

101

52 (6

8)

10(4

2)

036

014

minus 0

95

003

8U

rea

11

236

(42

)25

(93

)21

02

450

minus 9

842

lt00

01C

reat

inin

e gt

12

mg

dl12

042

(46

)26

(93

)20

81

444

minus 9

746

lt00

01So

dium

lt 1

45 m

mol

l12

048

(52

)8

(29

)0

390

15 minus

09

90

047

Pota

ssiu

m lt

35

mm

oll

116

15 (1

6)

2 (8

)

049

010

minus 2

36

037

5To

tal b

iliru

bin

gt12

mg

dl10

916

(19

)23

(92

)45

53

962

minus 2

150

lt00

01A

lani

ne tr

ansa

min

ase

(ALT

)12

058

(62

)24

(89

)4

601

26 minus

16

710

021

Asp

arta

te tr

ansa

min

ase

(AST

)11

270

(81

)25

(96

)4

730

59 minus

38

670

147

Glu

tam

yl tr

ansf

eras

e (G

GT)

77

34 (5

6)

15 (9

4)

958

116

minus 7

892

003

6A

lkal

ine

phos

phat

ase

(ALP

)88

19 (2

7)

15 (8

8)

175

23

59 minus

85

44lt0

001

Cre

atin

e ki

nase

(CK

)75

20 (3

5)

12 (6

7)

506

150

minus 1

709

000

9C

-rea

ctiv

e pr

otei

n (C

RP)

CR

P gt

05

mg

dl66

52 (9

8)

13 (1

00

)b

bb

a Chi

ldre

n w

ere

excl

uded

from

ana

lyse

s of l

abor

ator

y da

ta

b Odd

s rat

io c

ould

not

be

calc

ulat

ed si

nce

ther

e w

ere

no fa

tal c

ases

with

nor

mal

val

ues

The

abno

rmal

val

ues w

ere

cons

ider

ed o

ut o

f nor

mal

rang

e (lo

wer

or e

leva

ted

valu

es) l

imits

est

ablis

h by

eac

h di

ffer

ent h

ospi

tal l

abor

ator

y

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 10

Table 1Age and intervals (days) of events for patients infected by Malish or Israel spotted fever rickettsial strain

Status Variable R conorii Malish strain R conorii ISF strain p-value

Survived Age (median) 63 years 57 years 03261st symptom to 1st visit 4 days 3 days 01891st visit to specific therapy 0 days 0 days 02411st symptom to specific therapy 5 days 4 days 0028

emsp Length of stay 7 days 6 days 0081

Fatal Age (median) 66 years 575 years 01141st symptom to 1st visit 4 days 3 days 06181st visit to specific therapy 0 days 05 days 03711st symptom to specific therapy 45 days 4 days 0899Length of stay 2 days 1 days 0203

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 11

Table 2aComparative analysis of recognized potential epidemiologic exposure clinical signs and symptoms and underlyingdiseases of patients infected by R conorii Malish strain versus Israeli spotted fever strain

Epidemiologic exposure clinicalcharacteristics and co-morbidities

Total Patients (n)a R conorii Malishstrain n ()b

R conoriiISFstrain n ()b

p-value

Epidemiologic exposureAnimal contact 88 44 (94) 36 (88) 0465Tick bite history 66 20 (63) 11 (32) 0014Clinical signs and symptomsFever 137 64 (94) 65 (94) 1Eschar 123 38 (60) 23 (38) 0015Rash 137 65 (96) 63 (91) 0463Type of rash Maculopapular 120 61(94) 59 (94) 1 Petechial 8 4 (6) 4 (6) 1Asthenia 80 39 (95) 35 (90) 0426Headache 77 32 (78) 28 (78) 0977Myalgia 85 38 (84) 34 (85) 0943Arthralgia 69 22 (63) 16 (47) 0187Abdominal Pain 70 7 (19) 12 (36) 0101Anorexia 65 27 (73) 19 (68) 0653Nausea 75 14 (35) 22 (63) 0016Vomiting 84 13 (30) 23 (56) 0017Diarrhea 79 11 (28) 17 (44) 0135Prostration 81 27 (64) 26 (67) 0822Confusion obtundation 100 19 (38) 15 (30) 0398Dehydration 72 14 (39) 17 (47) 0475Tachypnea 63 10 (38) 18 (49) 0452Chest pain 53 3 (14) 6 (19) 1Hepatomegaly 60 9 (29) 9 (31) 0866Splenomegaly 57 2 (7) 3 (11) 0660Co-morbiditiesAlcoholism 103 9 (18) 9 (17) 0892Tobacco use 87 6 (15) 7 (15) 0939Diabetes 107 5 (10) 4 (7) 0737Cardiac failure 105 8 (16) 6 (11) 0443Respiratory failure 85 2 (5) 4 (9) 0677Renal failure 85 3 (7) 0 0116Hypertension 89 10 (22) 7 (16) 0449

aTotal patients with available data

bthe percentage is based on total patients infected by each strain

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 12

Table 2bComparative analysis of admission laboratory data of patients infected by R conorii Malish strain versus Israeli spottedfever strain

Laboratory Observation Total Patientsa(n) R conorii Malishstrain n ()b

R conorii ISFstrain n ()b

p- value

Anemia lt11gdl 118 13 (22) 8 (13) 0197Leukopenia lt 4400μl 121 6 (10) 10 (16) 0571Leukocytosis gt 11300μl 121 12 (20) 10 (16) 0571Thrombocytopenia (Platelets PLT) 122 45 (76) 55 (87) 0113Prothrombin time (PT) gt 13 seg 84 19 (48) 24 (55) 0544Partial thromboplastin time (PTT)PTT gt35 seg 95 17 (35) 22 (47) 0598Glucose gt 110 mgdl 101 33 (66) 29 (57) 0346Urea 112 29 (56) 32 (53) 0796Creatinine gt 12 mgdl 120 36 (60) 32 (53) 0461Sodium lt 145 mmoll 120 30 (52) 26 (42) 0283Potassium lt 35 mmoll 116 10 (18) 7 (11) 0308Total bilirubin gt12 mgdl 109 14 (26) 25 (45) 0047Alanine transaminase (ALT) 120 37 (64) 45 (73) 0301Aspartate transaminase (AST) 112 42 (79) 53 (90) 0119γ- Glutamyl transferase (GGT) 77 17 (50) 32 (74) 0027Alkaline phosphatase (ALP) 88 10 (26) 24 (49) 0026Creatine kinase (CK) 75 19 (51) 13 (34) 0133C-reactive protein (CRP)CRP gt 05 mgdl 66 34 (97) 31 (100) 0343

aChildren were excluded from analyses of laboratory data

bThe percentage is based on total patients infected by each strain

The abnormal values were considered out of normal range (lower or elevated values) limits established by each different hospital laboratory

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 13Ta

ble

3C

linic

al a

nd la

bora

tory

obs

erva

tions

and

trea

tmen

t in

6 ch

ildre

n w

ith M

edite

rran

ean

spot

ted

feve

r

Patie

nt 1

Patie

nt 2

Patie

nt 3

Patie

nt 4

Patie

nt 5

Patie

nt 6

Infe

ctin

g St

rain

ISF

ISF

Mal

ish

Mal

ish

Mal

ish

Mal

ish

Age

(yea

r or m

onth

)4y

6y1y

5 m

onth

6 m

onth

12y

1st sy

mpt

oms t

o sp

ecifi

c th

erap

y(d

ays)

24

4

10

Adm

issi

on to

dis

char

ge (d

ays)

09

99

011

Feve

rye

sye

sye

sye

sye

sye

sEs

char

nono

nono

noye

sR

ash

mac

ulop

apul

arpe

tech

ial

mac

ulop

apul

arm

acul

opap

ular

mac

ulop

opul

arm

acul

opap

ular

Hea

dach

eye

sM

yalg

iaye

sye

sG

astro

inte

stin

al si

gns

vom

iting

vom

iting

Hep

atom

egal

ySp

leno

meg

aly

yes

Com

plic

atio

nsen

ceph

aliti

sTh

erap

eutic

saz

ithro

myc

indo

xycy

clin

eaz

ithro

myc

inch

lora

mph

enic

olno

t tre

ated

azith

rom

ycin

Lab

orat

ory

data

Leuc

ocyt

es (c

ells

μl)

8920

2260

9930

178

0013

870

Plat

elet

s (ce

lls μ

l)17

700

010

600

013

700

017

700

088

000

Ure

a (m

gdl

)21

1527

--

Cre

atin

ine

(mg

dl)

04

05

03

06

-A

LT (U

L)

5624

634

7-

826

AST

(UL

)58

285

365

-11

17

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 14Ta

ble

4aU

niva

riate

ana

lysi

s of e

pide

mio

logi

c an

d cl

inic

al fi

ndin

gs in

fata

l and

non

-fat

al M

SF c

ases

Epi

dem

iolo

gic

and

clin

ical

char

acte

rist

ics

Tot

al p

atie

nts (

n)Su

rviv

ors n

()

Fata

l cas

es n

()

Odd

s Rat

io (O

R)

adju

sted

by

stra

in95

C

Ip-

val

ue

Ric

ketts

ial s

train

Mal

ish

7162

(87

)9

(13

)1

mdashmdash

IS

F69

49 (7

1)

20 (2

9)

281

118

minus 6

72

002

0A

nim

al c

onta

ct88

69 (9

0)

11 (1

00

)a

aa

Tick

bite

his

tory

6627

(50

)4

(33

)0

680

17 minus

27

20

587

Feve

r13

710

6 (9

7)

23 (8

2)

011

023

minus 0

54

000

6Es

char

123

53 (5

2)

8 (3

6)

059

022

minus 1

56

058

9R

ash

137

105

(97

)23

(79

)0

120

03 minus

05

20

005

Type

of r

ash

Mac

ulop

apul

ar10

2 (9

7)

18 (7

8)

1-

-

P

etec

hial

3 (3

)

5 (2

2)

109

32

29 minus

53

830

003

Ast

heni

a80

59 (9

4)

15 (8

8)

063

009

7 minus

412

10

633

Hea

dach

e77

47 (7

6)

13 (8

7)

234

043

minus 1

273

032

6M

yalg

ia85

58 (8

3)

14 (9

3)

308

035

minus 2

719

031

1A

rthra

lgia

6931

(55

)7

(54

)1

190

33 minus

42

30

792

Abd

omin

al P

ain

7011

(29

)8

(57

)4

561

23 minus

16

850

023

Ano

rexi

a65

37 (6

6)

9 (8

2)

258

046

minus 1

450

028

2N

ause

a75

25 (4

0)

11 (9

2)

125

148

minus 1

069

40

021

Vom

iting

8422

(33

)14

(82

)7

51

88 minus

30

120

004

Dia

rrhe

a79

16 (2

6)

12 (7

1)

638

181

minus 2

242

000

4Pr

ostra

tion

8136

(58

)17

(89

)6

731

37 minus

33

030

019

Con

fusi

on o

btun

datio

n10

022

(27

)12

(67

)9

262

61 minus

32

910

001

Deh

ydra

tion

7218

(33

)13

(76

)7

551

95 minus

29

290

003

Tach

ypne

a63

11 (2

5)

17 (8

9)

291

75

35 minus

158

97

lt00

01C

hest

pai

n53

6 (1

5)

3 (2

5)

179

035

minus 9

18

048

6H

epat

omeg

aly

6012

(24

)6

(67

)7

371

46 minus

37

040

015

Sple

nom

egal

y57

4 (8

)

1 (1

7)

200

017

minus 2

377

058

4

a Odd

s rat

io c

ould

not

be

calc

ulat

ed si

nce

ther

e w

ere

no fa

tal c

ases

with

nor

mal

val

ues

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 15Ta

ble

4bC

ompa

rison

of e

xist

ence

of p

re-e

xist

ing

co-m

orbi

ditie

s in

fata

l and

non

-fat

al c

ases

of M

edite

rran

ean

spot

ted

feve

r un

ivar

iate

ana

lysi

s

Co-

mor

bidi

ties

Tot

al P

atie

ntsa (n

)Su

rviv

ors n

()

Fata

l cas

es n

()

Odd

s Rat

io (O

R)

adju

sted

by

stra

in95

C

Ip-

val

ue

Alc

ohol

ism

103

11 (1

3)

7 (4

7)

699

41

99 minus

24

640

002

Toba

cco

use

8710

(14

)3

(18

)1

286

030

minus 5

54

073

5D

iabe

tes

107

8 (9

)

1 (6

)

074

50

08 minus

66

50

793

Car

diac

failu

re10

511

(13

)3

(18

)1

773

041

minus 7

60

044

0R

espi

rato

ry fa

ilure

855

(7

)1

(6

)0

705

007

minus 6

79

076

3C

hron

ic re

nal f

ailu

re85

3 (4

)

0-

--

Hyp

erte

nsio

n89

13 (1

8)

4 (2

5)

183

90

48 minus

70

70

375

a Tota

l pat

ient

s with

ava

ilabl

e da

ta

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 16Ta

ble

4cU

niva

riate

ana

lysi

s of l

abor

ator

y fin

ding

s in

fata

l and

non

-fat

al M

SF c

ases

Lab

orat

ory

Cha

ract

eris

tics

Tot

al P

atie

ntsa (n

)Su

rviv

ors n

()

Fata

l cas

es n

()

Odd

s Rat

io(O

R) a

djus

ted

by st

rain

95

CI

p- v

alue

Ane

mia

lt11

gd

l11

818

(20

)3

(11

)0

590

15 minus

22

00

424

Leuk

open

ia lt

440

0μl

121

13 (1

4)

3 (1

2)

130

031

minus 5

59

071

6Le

ukoc

ytos

is gt

113

00μ

l12

110

(11

)12

(46

)11

13

337

minus 3

680

lt00

01Th

rom

bocy

tope

nia

(Pla

tele

ts P

LT)

122

73 (7

8)

27 (9

6)

68

086

minus 5

378

006

9Pr

othr

ombi

n tim

e (P

T) gt

13

seg

8434

(50

)9

(60

)2

140

59 minus

77

40

246

Parti

al th

rom

bopl

astin

tim

e (P

TT)

PTT

gt35

seg

9522

(29

)17

(85

)36

78

453

minus 2

984

40

001

Glu

cose

gt 1

10 m

gdl

101

52 (6

8)

10(4

2)

036

014

minus 0

95

003

8U

rea

11

236

(42

)25

(93

)21

02

450

minus 9

842

lt00

01C

reat

inin

e gt

12

mg

dl12

042

(46

)26

(93

)20

81

444

minus 9

746

lt00

01So

dium

lt 1

45 m

mol

l12

048

(52

)8

(29

)0

390

15 minus

09

90

047

Pota

ssiu

m lt

35

mm

oll

116

15 (1

6)

2 (8

)

049

010

minus 2

36

037

5To

tal b

iliru

bin

gt12

mg

dl10

916

(19

)23

(92

)45

53

962

minus 2

150

lt00

01A

lani

ne tr

ansa

min

ase

(ALT

)12

058

(62

)24

(89

)4

601

26 minus

16

710

021

Asp

arta

te tr

ansa

min

ase

(AST

)11

270

(81

)25

(96

)4

730

59 minus

38

670

147

Glu

tam

yl tr

ansf

eras

e (G

GT)

77

34 (5

6)

15 (9

4)

958

116

minus 7

892

003

6A

lkal

ine

phos

phat

ase

(ALP

)88

19 (2

7)

15 (8

8)

175

23

59 minus

85

44lt0

001

Cre

atin

e ki

nase

(CK

)75

20 (3

5)

12 (6

7)

506

150

minus 1

709

000

9C

-rea

ctiv

e pr

otei

n (C

RP)

CR

P gt

05

mg

dl66

52 (9

8)

13 (1

00

)b

bb

a Chi

ldre

n w

ere

excl

uded

from

ana

lyse

s of l

abor

ator

y da

ta

b Odd

s rat

io c

ould

not

be

calc

ulat

ed si

nce

ther

e w

ere

no fa

tal c

ases

with

nor

mal

val

ues

The

abno

rmal

val

ues w

ere

cons

ider

ed o

ut o

f nor

mal

rang

e (lo

wer

or e

leva

ted

valu

es) l

imits

est

ablis

h by

eac

h di

ffer

ent h

ospi

tal l

abor

ator

y

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 11

Table 2aComparative analysis of recognized potential epidemiologic exposure clinical signs and symptoms and underlyingdiseases of patients infected by R conorii Malish strain versus Israeli spotted fever strain

Epidemiologic exposure clinicalcharacteristics and co-morbidities

Total Patients (n)a R conorii Malishstrain n ()b

R conoriiISFstrain n ()b

p-value

Epidemiologic exposureAnimal contact 88 44 (94) 36 (88) 0465Tick bite history 66 20 (63) 11 (32) 0014Clinical signs and symptomsFever 137 64 (94) 65 (94) 1Eschar 123 38 (60) 23 (38) 0015Rash 137 65 (96) 63 (91) 0463Type of rash Maculopapular 120 61(94) 59 (94) 1 Petechial 8 4 (6) 4 (6) 1Asthenia 80 39 (95) 35 (90) 0426Headache 77 32 (78) 28 (78) 0977Myalgia 85 38 (84) 34 (85) 0943Arthralgia 69 22 (63) 16 (47) 0187Abdominal Pain 70 7 (19) 12 (36) 0101Anorexia 65 27 (73) 19 (68) 0653Nausea 75 14 (35) 22 (63) 0016Vomiting 84 13 (30) 23 (56) 0017Diarrhea 79 11 (28) 17 (44) 0135Prostration 81 27 (64) 26 (67) 0822Confusion obtundation 100 19 (38) 15 (30) 0398Dehydration 72 14 (39) 17 (47) 0475Tachypnea 63 10 (38) 18 (49) 0452Chest pain 53 3 (14) 6 (19) 1Hepatomegaly 60 9 (29) 9 (31) 0866Splenomegaly 57 2 (7) 3 (11) 0660Co-morbiditiesAlcoholism 103 9 (18) 9 (17) 0892Tobacco use 87 6 (15) 7 (15) 0939Diabetes 107 5 (10) 4 (7) 0737Cardiac failure 105 8 (16) 6 (11) 0443Respiratory failure 85 2 (5) 4 (9) 0677Renal failure 85 3 (7) 0 0116Hypertension 89 10 (22) 7 (16) 0449

aTotal patients with available data

bthe percentage is based on total patients infected by each strain

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 12

Table 2bComparative analysis of admission laboratory data of patients infected by R conorii Malish strain versus Israeli spottedfever strain

Laboratory Observation Total Patientsa(n) R conorii Malishstrain n ()b

R conorii ISFstrain n ()b

p- value

Anemia lt11gdl 118 13 (22) 8 (13) 0197Leukopenia lt 4400μl 121 6 (10) 10 (16) 0571Leukocytosis gt 11300μl 121 12 (20) 10 (16) 0571Thrombocytopenia (Platelets PLT) 122 45 (76) 55 (87) 0113Prothrombin time (PT) gt 13 seg 84 19 (48) 24 (55) 0544Partial thromboplastin time (PTT)PTT gt35 seg 95 17 (35) 22 (47) 0598Glucose gt 110 mgdl 101 33 (66) 29 (57) 0346Urea 112 29 (56) 32 (53) 0796Creatinine gt 12 mgdl 120 36 (60) 32 (53) 0461Sodium lt 145 mmoll 120 30 (52) 26 (42) 0283Potassium lt 35 mmoll 116 10 (18) 7 (11) 0308Total bilirubin gt12 mgdl 109 14 (26) 25 (45) 0047Alanine transaminase (ALT) 120 37 (64) 45 (73) 0301Aspartate transaminase (AST) 112 42 (79) 53 (90) 0119γ- Glutamyl transferase (GGT) 77 17 (50) 32 (74) 0027Alkaline phosphatase (ALP) 88 10 (26) 24 (49) 0026Creatine kinase (CK) 75 19 (51) 13 (34) 0133C-reactive protein (CRP)CRP gt 05 mgdl 66 34 (97) 31 (100) 0343

aChildren were excluded from analyses of laboratory data

bThe percentage is based on total patients infected by each strain

The abnormal values were considered out of normal range (lower or elevated values) limits established by each different hospital laboratory

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 13Ta

ble

3C

linic

al a

nd la

bora

tory

obs

erva

tions

and

trea

tmen

t in

6 ch

ildre

n w

ith M

edite

rran

ean

spot

ted

feve

r

Patie

nt 1

Patie

nt 2

Patie

nt 3

Patie

nt 4

Patie

nt 5

Patie

nt 6

Infe

ctin

g St

rain

ISF

ISF

Mal

ish

Mal

ish

Mal

ish

Mal

ish

Age

(yea

r or m

onth

)4y

6y1y

5 m

onth

6 m

onth

12y

1st sy

mpt

oms t

o sp

ecifi

c th

erap

y(d

ays)

24

4

10

Adm

issi

on to

dis

char

ge (d

ays)

09

99

011

Feve

rye

sye

sye

sye

sye

sye

sEs

char

nono

nono

noye

sR

ash

mac

ulop

apul

arpe

tech

ial

mac

ulop

apul

arm

acul

opap

ular

mac

ulop

opul

arm

acul

opap

ular

Hea

dach

eye

sM

yalg

iaye

sye

sG

astro

inte

stin

al si

gns

vom

iting

vom

iting

Hep

atom

egal

ySp

leno

meg

aly

yes

Com

plic

atio

nsen

ceph

aliti

sTh

erap

eutic

saz

ithro

myc

indo

xycy

clin

eaz

ithro

myc

inch

lora

mph

enic

olno

t tre

ated

azith

rom

ycin

Lab

orat

ory

data

Leuc

ocyt

es (c

ells

μl)

8920

2260

9930

178

0013

870

Plat

elet

s (ce

lls μ

l)17

700

010

600

013

700

017

700

088

000

Ure

a (m

gdl

)21

1527

--

Cre

atin

ine

(mg

dl)

04

05

03

06

-A

LT (U

L)

5624

634

7-

826

AST

(UL

)58

285

365

-11

17

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 14Ta

ble

4aU

niva

riate

ana

lysi

s of e

pide

mio

logi

c an

d cl

inic

al fi

ndin

gs in

fata

l and

non

-fat

al M

SF c

ases

Epi

dem

iolo

gic

and

clin

ical

char

acte

rist

ics

Tot

al p

atie

nts (

n)Su

rviv

ors n

()

Fata

l cas

es n

()

Odd

s Rat

io (O

R)

adju

sted

by

stra

in95

C

Ip-

val

ue

Ric

ketts

ial s

train

Mal

ish

7162

(87

)9

(13

)1

mdashmdash

IS

F69

49 (7

1)

20 (2

9)

281

118

minus 6

72

002

0A

nim

al c

onta

ct88

69 (9

0)

11 (1

00

)a

aa

Tick

bite

his

tory

6627

(50

)4

(33

)0

680

17 minus

27

20

587

Feve

r13

710

6 (9

7)

23 (8

2)

011

023

minus 0

54

000

6Es

char

123

53 (5

2)

8 (3

6)

059

022

minus 1

56

058

9R

ash

137

105

(97

)23

(79

)0

120

03 minus

05

20

005

Type

of r

ash

Mac

ulop

apul

ar10

2 (9

7)

18 (7

8)

1-

-

P

etec

hial

3 (3

)

5 (2

2)

109

32

29 minus

53

830

003

Ast

heni

a80

59 (9

4)

15 (8

8)

063

009

7 minus

412

10

633

Hea

dach

e77

47 (7

6)

13 (8

7)

234

043

minus 1

273

032

6M

yalg

ia85

58 (8

3)

14 (9

3)

308

035

minus 2

719

031

1A

rthra

lgia

6931

(55

)7

(54

)1

190

33 minus

42

30

792

Abd

omin

al P

ain

7011

(29

)8

(57

)4

561

23 minus

16

850

023

Ano

rexi

a65

37 (6

6)

9 (8

2)

258

046

minus 1

450

028

2N

ause

a75

25 (4

0)

11 (9

2)

125

148

minus 1

069

40

021

Vom

iting

8422

(33

)14

(82

)7

51

88 minus

30

120

004

Dia

rrhe

a79

16 (2

6)

12 (7

1)

638

181

minus 2

242

000

4Pr

ostra

tion

8136

(58

)17

(89

)6

731

37 minus

33

030

019

Con

fusi

on o

btun

datio

n10

022

(27

)12

(67

)9

262

61 minus

32

910

001

Deh

ydra

tion

7218

(33

)13

(76

)7

551

95 minus

29

290

003

Tach

ypne

a63

11 (2

5)

17 (8

9)

291

75

35 minus

158

97

lt00

01C

hest

pai

n53

6 (1

5)

3 (2

5)

179

035

minus 9

18

048

6H

epat

omeg

aly

6012

(24

)6

(67

)7

371

46 minus

37

040

015

Sple

nom

egal

y57

4 (8

)

1 (1

7)

200

017

minus 2

377

058

4

a Odd

s rat

io c

ould

not

be

calc

ulat

ed si

nce

ther

e w

ere

no fa

tal c

ases

with

nor

mal

val

ues

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 15Ta

ble

4bC

ompa

rison

of e

xist

ence

of p

re-e

xist

ing

co-m

orbi

ditie

s in

fata

l and

non

-fat

al c

ases

of M

edite

rran

ean

spot

ted

feve

r un

ivar

iate

ana

lysi

s

Co-

mor

bidi

ties

Tot

al P

atie

ntsa (n

)Su

rviv

ors n

()

Fata

l cas

es n

()

Odd

s Rat

io (O

R)

adju

sted

by

stra

in95

C

Ip-

val

ue

Alc

ohol

ism

103

11 (1

3)

7 (4

7)

699

41

99 minus

24

640

002

Toba

cco

use

8710

(14

)3

(18

)1

286

030

minus 5

54

073

5D

iabe

tes

107

8 (9

)

1 (6

)

074

50

08 minus

66

50

793

Car

diac

failu

re10

511

(13

)3

(18

)1

773

041

minus 7

60

044

0R

espi

rato

ry fa

ilure

855

(7

)1

(6

)0

705

007

minus 6

79

076

3C

hron

ic re

nal f

ailu

re85

3 (4

)

0-

--

Hyp

erte

nsio

n89

13 (1

8)

4 (2

5)

183

90

48 minus

70

70

375

a Tota

l pat

ient

s with

ava

ilabl

e da

ta

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 16Ta

ble

4cU

niva

riate

ana

lysi

s of l

abor

ator

y fin

ding

s in

fata

l and

non

-fat

al M

SF c

ases

Lab

orat

ory

Cha

ract

eris

tics

Tot

al P

atie

ntsa (n

)Su

rviv

ors n

()

Fata

l cas

es n

()

Odd

s Rat

io(O

R) a

djus

ted

by st

rain

95

CI

p- v

alue

Ane

mia

lt11

gd

l11

818

(20

)3

(11

)0

590

15 minus

22

00

424

Leuk

open

ia lt

440

0μl

121

13 (1

4)

3 (1

2)

130

031

minus 5

59

071

6Le

ukoc

ytos

is gt

113

00μ

l12

110

(11

)12

(46

)11

13

337

minus 3

680

lt00

01Th

rom

bocy

tope

nia

(Pla

tele

ts P

LT)

122

73 (7

8)

27 (9

6)

68

086

minus 5

378

006

9Pr

othr

ombi

n tim

e (P

T) gt

13

seg

8434

(50

)9

(60

)2

140

59 minus

77

40

246

Parti

al th

rom

bopl

astin

tim

e (P

TT)

PTT

gt35

seg

9522

(29

)17

(85

)36

78

453

minus 2

984

40

001

Glu

cose

gt 1

10 m

gdl

101

52 (6

8)

10(4

2)

036

014

minus 0

95

003

8U

rea

11

236

(42

)25

(93

)21

02

450

minus 9

842

lt00

01C

reat

inin

e gt

12

mg

dl12

042

(46

)26

(93

)20

81

444

minus 9

746

lt00

01So

dium

lt 1

45 m

mol

l12

048

(52

)8

(29

)0

390

15 minus

09

90

047

Pota

ssiu

m lt

35

mm

oll

116

15 (1

6)

2 (8

)

049

010

minus 2

36

037

5To

tal b

iliru

bin

gt12

mg

dl10

916

(19

)23

(92

)45

53

962

minus 2

150

lt00

01A

lani

ne tr

ansa

min

ase

(ALT

)12

058

(62

)24

(89

)4

601

26 minus

16

710

021

Asp

arta

te tr

ansa

min

ase

(AST

)11

270

(81

)25

(96

)4

730

59 minus

38

670

147

Glu

tam

yl tr

ansf

eras

e (G

GT)

77

34 (5

6)

15 (9

4)

958

116

minus 7

892

003

6A

lkal

ine

phos

phat

ase

(ALP

)88

19 (2

7)

15 (8

8)

175

23

59 minus

85

44lt0

001

Cre

atin

e ki

nase

(CK

)75

20 (3

5)

12 (6

7)

506

150

minus 1

709

000

9C

-rea

ctiv

e pr

otei

n (C

RP)

CR

P gt

05

mg

dl66

52 (9

8)

13 (1

00

)b

bb

a Chi

ldre

n w

ere

excl

uded

from

ana

lyse

s of l

abor

ator

y da

ta

b Odd

s rat

io c

ould

not

be

calc

ulat

ed si

nce

ther

e w

ere

no fa

tal c

ases

with

nor

mal

val

ues

The

abno

rmal

val

ues w

ere

cons

ider

ed o

ut o

f nor

mal

rang

e (lo

wer

or e

leva

ted

valu

es) l

imits

est

ablis

h by

eac

h di

ffer

ent h

ospi

tal l

abor

ator

y

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 12

Table 2bComparative analysis of admission laboratory data of patients infected by R conorii Malish strain versus Israeli spottedfever strain

Laboratory Observation Total Patientsa(n) R conorii Malishstrain n ()b

R conorii ISFstrain n ()b

p- value

Anemia lt11gdl 118 13 (22) 8 (13) 0197Leukopenia lt 4400μl 121 6 (10) 10 (16) 0571Leukocytosis gt 11300μl 121 12 (20) 10 (16) 0571Thrombocytopenia (Platelets PLT) 122 45 (76) 55 (87) 0113Prothrombin time (PT) gt 13 seg 84 19 (48) 24 (55) 0544Partial thromboplastin time (PTT)PTT gt35 seg 95 17 (35) 22 (47) 0598Glucose gt 110 mgdl 101 33 (66) 29 (57) 0346Urea 112 29 (56) 32 (53) 0796Creatinine gt 12 mgdl 120 36 (60) 32 (53) 0461Sodium lt 145 mmoll 120 30 (52) 26 (42) 0283Potassium lt 35 mmoll 116 10 (18) 7 (11) 0308Total bilirubin gt12 mgdl 109 14 (26) 25 (45) 0047Alanine transaminase (ALT) 120 37 (64) 45 (73) 0301Aspartate transaminase (AST) 112 42 (79) 53 (90) 0119γ- Glutamyl transferase (GGT) 77 17 (50) 32 (74) 0027Alkaline phosphatase (ALP) 88 10 (26) 24 (49) 0026Creatine kinase (CK) 75 19 (51) 13 (34) 0133C-reactive protein (CRP)CRP gt 05 mgdl 66 34 (97) 31 (100) 0343

aChildren were excluded from analyses of laboratory data

bThe percentage is based on total patients infected by each strain

The abnormal values were considered out of normal range (lower or elevated values) limits established by each different hospital laboratory

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 13Ta

ble

3C

linic

al a

nd la

bora

tory

obs

erva

tions

and

trea

tmen

t in

6 ch

ildre

n w

ith M

edite

rran

ean

spot

ted

feve

r

Patie

nt 1

Patie

nt 2

Patie

nt 3

Patie

nt 4

Patie

nt 5

Patie

nt 6

Infe

ctin

g St

rain

ISF

ISF

Mal

ish

Mal

ish

Mal

ish

Mal

ish

Age

(yea

r or m

onth

)4y

6y1y

5 m

onth

6 m

onth

12y

1st sy

mpt

oms t

o sp

ecifi

c th

erap

y(d

ays)

24

4

10

Adm

issi

on to

dis

char

ge (d

ays)

09

99

011

Feve

rye

sye

sye

sye

sye

sye

sEs

char

nono

nono

noye

sR

ash

mac

ulop

apul

arpe

tech

ial

mac

ulop

apul

arm

acul

opap

ular

mac

ulop

opul

arm

acul

opap

ular

Hea

dach

eye

sM

yalg

iaye

sye

sG

astro

inte

stin

al si

gns

vom

iting

vom

iting

Hep

atom

egal

ySp

leno

meg

aly

yes

Com

plic

atio

nsen

ceph

aliti

sTh

erap

eutic

saz

ithro

myc

indo

xycy

clin

eaz

ithro

myc

inch

lora

mph

enic

olno

t tre

ated

azith

rom

ycin

Lab

orat

ory

data

Leuc

ocyt

es (c

ells

μl)

8920

2260

9930

178

0013

870

Plat

elet

s (ce

lls μ

l)17

700

010

600

013

700

017

700

088

000

Ure

a (m

gdl

)21

1527

--

Cre

atin

ine

(mg

dl)

04

05

03

06

-A

LT (U

L)

5624

634

7-

826

AST

(UL

)58

285

365

-11

17

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 14Ta

ble

4aU

niva

riate

ana

lysi

s of e

pide

mio

logi

c an

d cl

inic

al fi

ndin

gs in

fata

l and

non

-fat

al M

SF c

ases

Epi

dem

iolo

gic

and

clin

ical

char

acte

rist

ics

Tot

al p

atie

nts (

n)Su

rviv

ors n

()

Fata

l cas

es n

()

Odd

s Rat

io (O

R)

adju

sted

by

stra

in95

C

Ip-

val

ue

Ric

ketts

ial s

train

Mal

ish

7162

(87

)9

(13

)1

mdashmdash

IS

F69

49 (7

1)

20 (2

9)

281

118

minus 6

72

002

0A

nim

al c

onta

ct88

69 (9

0)

11 (1

00

)a

aa

Tick

bite

his

tory

6627

(50

)4

(33

)0

680

17 minus

27

20

587

Feve

r13

710

6 (9

7)

23 (8

2)

011

023

minus 0

54

000

6Es

char

123

53 (5

2)

8 (3

6)

059

022

minus 1

56

058

9R

ash

137

105

(97

)23

(79

)0

120

03 minus

05

20

005

Type

of r

ash

Mac

ulop

apul

ar10

2 (9

7)

18 (7

8)

1-

-

P

etec

hial

3 (3

)

5 (2

2)

109

32

29 minus

53

830

003

Ast

heni

a80

59 (9

4)

15 (8

8)

063

009

7 minus

412

10

633

Hea

dach

e77

47 (7

6)

13 (8

7)

234

043

minus 1

273

032

6M

yalg

ia85

58 (8

3)

14 (9

3)

308

035

minus 2

719

031

1A

rthra

lgia

6931

(55

)7

(54

)1

190

33 minus

42

30

792

Abd

omin

al P

ain

7011

(29

)8

(57

)4

561

23 minus

16

850

023

Ano

rexi

a65

37 (6

6)

9 (8

2)

258

046

minus 1

450

028

2N

ause

a75

25 (4

0)

11 (9

2)

125

148

minus 1

069

40

021

Vom

iting

8422

(33

)14

(82

)7

51

88 minus

30

120

004

Dia

rrhe

a79

16 (2

6)

12 (7

1)

638

181

minus 2

242

000

4Pr

ostra

tion

8136

(58

)17

(89

)6

731

37 minus

33

030

019

Con

fusi

on o

btun

datio

n10

022

(27

)12

(67

)9

262

61 minus

32

910

001

Deh

ydra

tion

7218

(33

)13

(76

)7

551

95 minus

29

290

003

Tach

ypne

a63

11 (2

5)

17 (8

9)

291

75

35 minus

158

97

lt00

01C

hest

pai

n53

6 (1

5)

3 (2

5)

179

035

minus 9

18

048

6H

epat

omeg

aly

6012

(24

)6

(67

)7

371

46 minus

37

040

015

Sple

nom

egal

y57

4 (8

)

1 (1

7)

200

017

minus 2

377

058

4

a Odd

s rat

io c

ould

not

be

calc

ulat

ed si

nce

ther

e w

ere

no fa

tal c

ases

with

nor

mal

val

ues

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 15Ta

ble

4bC

ompa

rison

of e

xist

ence

of p

re-e

xist

ing

co-m

orbi

ditie

s in

fata

l and

non

-fat

al c

ases

of M

edite

rran

ean

spot

ted

feve

r un

ivar

iate

ana

lysi

s

Co-

mor

bidi

ties

Tot

al P

atie

ntsa (n

)Su

rviv

ors n

()

Fata

l cas

es n

()

Odd

s Rat

io (O

R)

adju

sted

by

stra

in95

C

Ip-

val

ue

Alc

ohol

ism

103

11 (1

3)

7 (4

7)

699

41

99 minus

24

640

002

Toba

cco

use

8710

(14

)3

(18

)1

286

030

minus 5

54

073

5D

iabe

tes

107

8 (9

)

1 (6

)

074

50

08 minus

66

50

793

Car

diac

failu

re10

511

(13

)3

(18

)1

773

041

minus 7

60

044

0R

espi

rato

ry fa

ilure

855

(7

)1

(6

)0

705

007

minus 6

79

076

3C

hron

ic re

nal f

ailu

re85

3 (4

)

0-

--

Hyp

erte

nsio

n89

13 (1

8)

4 (2

5)

183

90

48 minus

70

70

375

a Tota

l pat

ient

s with

ava

ilabl

e da

ta

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 16Ta

ble

4cU

niva

riate

ana

lysi

s of l

abor

ator

y fin

ding

s in

fata

l and

non

-fat

al M

SF c

ases

Lab

orat

ory

Cha

ract

eris

tics

Tot

al P

atie

ntsa (n

)Su

rviv

ors n

()

Fata

l cas

es n

()

Odd

s Rat

io(O

R) a

djus

ted

by st

rain

95

CI

p- v

alue

Ane

mia

lt11

gd

l11

818

(20

)3

(11

)0

590

15 minus

22

00

424

Leuk

open

ia lt

440

0μl

121

13 (1

4)

3 (1

2)

130

031

minus 5

59

071

6Le

ukoc

ytos

is gt

113

00μ

l12

110

(11

)12

(46

)11

13

337

minus 3

680

lt00

01Th

rom

bocy

tope

nia

(Pla

tele

ts P

LT)

122

73 (7

8)

27 (9

6)

68

086

minus 5

378

006

9Pr

othr

ombi

n tim

e (P

T) gt

13

seg

8434

(50

)9

(60

)2

140

59 minus

77

40

246

Parti

al th

rom

bopl

astin

tim

e (P

TT)

PTT

gt35

seg

9522

(29

)17

(85

)36

78

453

minus 2

984

40

001

Glu

cose

gt 1

10 m

gdl

101

52 (6

8)

10(4

2)

036

014

minus 0

95

003

8U

rea

11

236

(42

)25

(93

)21

02

450

minus 9

842

lt00

01C

reat

inin

e gt

12

mg

dl12

042

(46

)26

(93

)20

81

444

minus 9

746

lt00

01So

dium

lt 1

45 m

mol

l12

048

(52

)8

(29

)0

390

15 minus

09

90

047

Pota

ssiu

m lt

35

mm

oll

116

15 (1

6)

2 (8

)

049

010

minus 2

36

037

5To

tal b

iliru

bin

gt12

mg

dl10

916

(19

)23

(92

)45

53

962

minus 2

150

lt00

01A

lani

ne tr

ansa

min

ase

(ALT

)12

058

(62

)24

(89

)4

601

26 minus

16

710

021

Asp

arta

te tr

ansa

min

ase

(AST

)11

270

(81

)25

(96

)4

730

59 minus

38

670

147

Glu

tam

yl tr

ansf

eras

e (G

GT)

77

34 (5

6)

15 (9

4)

958

116

minus 7

892

003

6A

lkal

ine

phos

phat

ase

(ALP

)88

19 (2

7)

15 (8

8)

175

23

59 minus

85

44lt0

001

Cre

atin

e ki

nase

(CK

)75

20 (3

5)

12 (6

7)

506

150

minus 1

709

000

9C

-rea

ctiv

e pr

otei

n (C

RP)

CR

P gt

05

mg

dl66

52 (9

8)

13 (1

00

)b

bb

a Chi

ldre

n w

ere

excl

uded

from

ana

lyse

s of l

abor

ator

y da

ta

b Odd

s rat

io c

ould

not

be

calc

ulat

ed si

nce

ther

e w

ere

no fa

tal c

ases

with

nor

mal

val

ues

The

abno

rmal

val

ues w

ere

cons

ider

ed o

ut o

f nor

mal

rang

e (lo

wer

or e

leva

ted

valu

es) l

imits

est

ablis

h by

eac

h di

ffer

ent h

ospi

tal l

abor

ator

y

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 13Ta

ble

3C

linic

al a

nd la

bora

tory

obs

erva

tions

and

trea

tmen

t in

6 ch

ildre

n w

ith M

edite

rran

ean

spot

ted

feve

r

Patie

nt 1

Patie

nt 2

Patie

nt 3

Patie

nt 4

Patie

nt 5

Patie

nt 6

Infe

ctin

g St

rain

ISF

ISF

Mal

ish

Mal

ish

Mal

ish

Mal

ish

Age

(yea

r or m

onth

)4y

6y1y

5 m

onth

6 m

onth

12y

1st sy

mpt

oms t

o sp

ecifi

c th

erap

y(d

ays)

24

4

10

Adm

issi

on to

dis

char

ge (d

ays)

09

99

011

Feve

rye

sye

sye

sye

sye

sye

sEs

char

nono

nono

noye

sR

ash

mac

ulop

apul

arpe

tech

ial

mac

ulop

apul

arm

acul

opap

ular

mac

ulop

opul

arm

acul

opap

ular

Hea

dach

eye

sM

yalg

iaye

sye

sG

astro

inte

stin

al si

gns

vom

iting

vom

iting

Hep

atom

egal

ySp

leno

meg

aly

yes

Com

plic

atio

nsen

ceph

aliti

sTh

erap

eutic

saz

ithro

myc

indo

xycy

clin

eaz

ithro

myc

inch

lora

mph

enic

olno

t tre

ated

azith

rom

ycin

Lab

orat

ory

data

Leuc

ocyt

es (c

ells

μl)

8920

2260

9930

178

0013

870

Plat

elet

s (ce

lls μ

l)17

700

010

600

013

700

017

700

088

000

Ure

a (m

gdl

)21

1527

--

Cre

atin

ine

(mg

dl)

04

05

03

06

-A

LT (U

L)

5624

634

7-

826

AST

(UL

)58

285

365

-11

17

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 14Ta

ble

4aU

niva

riate

ana

lysi

s of e

pide

mio

logi

c an

d cl

inic

al fi

ndin

gs in

fata

l and

non

-fat

al M

SF c

ases

Epi

dem

iolo

gic

and

clin

ical

char

acte

rist

ics

Tot

al p

atie

nts (

n)Su

rviv

ors n

()

Fata

l cas

es n

()

Odd

s Rat

io (O

R)

adju

sted

by

stra

in95

C

Ip-

val

ue

Ric

ketts

ial s

train

Mal

ish

7162

(87

)9

(13

)1

mdashmdash

IS

F69

49 (7

1)

20 (2

9)

281

118

minus 6

72

002

0A

nim

al c

onta

ct88

69 (9

0)

11 (1

00

)a

aa

Tick

bite

his

tory

6627

(50

)4

(33

)0

680

17 minus

27

20

587

Feve

r13

710

6 (9

7)

23 (8

2)

011

023

minus 0

54

000

6Es

char

123

53 (5

2)

8 (3

6)

059

022

minus 1

56

058

9R

ash

137

105

(97

)23

(79

)0

120

03 minus

05

20

005

Type

of r

ash

Mac

ulop

apul

ar10

2 (9

7)

18 (7

8)

1-

-

P

etec

hial

3 (3

)

5 (2

2)

109

32

29 minus

53

830

003

Ast

heni

a80

59 (9

4)

15 (8

8)

063

009

7 minus

412

10

633

Hea

dach

e77

47 (7

6)

13 (8

7)

234

043

minus 1

273

032

6M

yalg

ia85

58 (8

3)

14 (9

3)

308

035

minus 2

719

031

1A

rthra

lgia

6931

(55

)7

(54

)1

190

33 minus

42

30

792

Abd

omin

al P

ain

7011

(29

)8

(57

)4

561

23 minus

16

850

023

Ano

rexi

a65

37 (6

6)

9 (8

2)

258

046

minus 1

450

028

2N

ause

a75

25 (4

0)

11 (9

2)

125

148

minus 1

069

40

021

Vom

iting

8422

(33

)14

(82

)7

51

88 minus

30

120

004

Dia

rrhe

a79

16 (2

6)

12 (7

1)

638

181

minus 2

242

000

4Pr

ostra

tion

8136

(58

)17

(89

)6

731

37 minus

33

030

019

Con

fusi

on o

btun

datio

n10

022

(27

)12

(67

)9

262

61 minus

32

910

001

Deh

ydra

tion

7218

(33

)13

(76

)7

551

95 minus

29

290

003

Tach

ypne

a63

11 (2

5)

17 (8

9)

291

75

35 minus

158

97

lt00

01C

hest

pai

n53

6 (1

5)

3 (2

5)

179

035

minus 9

18

048

6H

epat

omeg

aly

6012

(24

)6

(67

)7

371

46 minus

37

040

015

Sple

nom

egal

y57

4 (8

)

1 (1

7)

200

017

minus 2

377

058

4

a Odd

s rat

io c

ould

not

be

calc

ulat

ed si

nce

ther

e w

ere

no fa

tal c

ases

with

nor

mal

val

ues

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 15Ta

ble

4bC

ompa

rison

of e

xist

ence

of p

re-e

xist

ing

co-m

orbi

ditie

s in

fata

l and

non

-fat

al c

ases

of M

edite

rran

ean

spot

ted

feve

r un

ivar

iate

ana

lysi

s

Co-

mor

bidi

ties

Tot

al P

atie

ntsa (n

)Su

rviv

ors n

()

Fata

l cas

es n

()

Odd

s Rat

io (O

R)

adju

sted

by

stra

in95

C

Ip-

val

ue

Alc

ohol

ism

103

11 (1

3)

7 (4

7)

699

41

99 minus

24

640

002

Toba

cco

use

8710

(14

)3

(18

)1

286

030

minus 5

54

073

5D

iabe

tes

107

8 (9

)

1 (6

)

074

50

08 minus

66

50

793

Car

diac

failu

re10

511

(13

)3

(18

)1

773

041

minus 7

60

044

0R

espi

rato

ry fa

ilure

855

(7

)1

(6

)0

705

007

minus 6

79

076

3C

hron

ic re

nal f

ailu

re85

3 (4

)

0-

--

Hyp

erte

nsio

n89

13 (1

8)

4 (2

5)

183

90

48 minus

70

70

375

a Tota

l pat

ient

s with

ava

ilabl

e da

ta

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 16Ta

ble

4cU

niva

riate

ana

lysi

s of l

abor

ator

y fin

ding

s in

fata

l and

non

-fat

al M

SF c

ases

Lab

orat

ory

Cha

ract

eris

tics

Tot

al P

atie

ntsa (n

)Su

rviv

ors n

()

Fata

l cas

es n

()

Odd

s Rat

io(O

R) a

djus

ted

by st

rain

95

CI

p- v

alue

Ane

mia

lt11

gd

l11

818

(20

)3

(11

)0

590

15 minus

22

00

424

Leuk

open

ia lt

440

0μl

121

13 (1

4)

3 (1

2)

130

031

minus 5

59

071

6Le

ukoc

ytos

is gt

113

00μ

l12

110

(11

)12

(46

)11

13

337

minus 3

680

lt00

01Th

rom

bocy

tope

nia

(Pla

tele

ts P

LT)

122

73 (7

8)

27 (9

6)

68

086

minus 5

378

006

9Pr

othr

ombi

n tim

e (P

T) gt

13

seg

8434

(50

)9

(60

)2

140

59 minus

77

40

246

Parti

al th

rom

bopl

astin

tim

e (P

TT)

PTT

gt35

seg

9522

(29

)17

(85

)36

78

453

minus 2

984

40

001

Glu

cose

gt 1

10 m

gdl

101

52 (6

8)

10(4

2)

036

014

minus 0

95

003

8U

rea

11

236

(42

)25

(93

)21

02

450

minus 9

842

lt00

01C

reat

inin

e gt

12

mg

dl12

042

(46

)26

(93

)20

81

444

minus 9

746

lt00

01So

dium

lt 1

45 m

mol

l12

048

(52

)8

(29

)0

390

15 minus

09

90

047

Pota

ssiu

m lt

35

mm

oll

116

15 (1

6)

2 (8

)

049

010

minus 2

36

037

5To

tal b

iliru

bin

gt12

mg

dl10

916

(19

)23

(92

)45

53

962

minus 2

150

lt00

01A

lani

ne tr

ansa

min

ase

(ALT

)12

058

(62

)24

(89

)4

601

26 minus

16

710

021

Asp

arta

te tr

ansa

min

ase

(AST

)11

270

(81

)25

(96

)4

730

59 minus

38

670

147

Glu

tam

yl tr

ansf

eras

e (G

GT)

77

34 (5

6)

15 (9

4)

958

116

minus 7

892

003

6A

lkal

ine

phos

phat

ase

(ALP

)88

19 (2

7)

15 (8

8)

175

23

59 minus

85

44lt0

001

Cre

atin

e ki

nase

(CK

)75

20 (3

5)

12 (6

7)

506

150

minus 1

709

000

9C

-rea

ctiv

e pr

otei

n (C

RP)

CR

P gt

05

mg

dl66

52 (9

8)

13 (1

00

)b

bb

a Chi

ldre

n w

ere

excl

uded

from

ana

lyse

s of l

abor

ator

y da

ta

b Odd

s rat

io c

ould

not

be

calc

ulat

ed si

nce

ther

e w

ere

no fa

tal c

ases

with

nor

mal

val

ues

The

abno

rmal

val

ues w

ere

cons

ider

ed o

ut o

f nor

mal

rang

e (lo

wer

or e

leva

ted

valu

es) l

imits

est

ablis

h by

eac

h di

ffer

ent h

ospi

tal l

abor

ator

y

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 14Ta

ble

4aU

niva

riate

ana

lysi

s of e

pide

mio

logi

c an

d cl

inic

al fi

ndin

gs in

fata

l and

non

-fat

al M

SF c

ases

Epi

dem

iolo

gic

and

clin

ical

char

acte

rist

ics

Tot

al p

atie

nts (

n)Su

rviv

ors n

()

Fata

l cas

es n

()

Odd

s Rat

io (O

R)

adju

sted

by

stra

in95

C

Ip-

val

ue

Ric

ketts

ial s

train

Mal

ish

7162

(87

)9

(13

)1

mdashmdash

IS

F69

49 (7

1)

20 (2

9)

281

118

minus 6

72

002

0A

nim

al c

onta

ct88

69 (9

0)

11 (1

00

)a

aa

Tick

bite

his

tory

6627

(50

)4

(33

)0

680

17 minus

27

20

587

Feve

r13

710

6 (9

7)

23 (8

2)

011

023

minus 0

54

000

6Es

char

123

53 (5

2)

8 (3

6)

059

022

minus 1

56

058

9R

ash

137

105

(97

)23

(79

)0

120

03 minus

05

20

005

Type

of r

ash

Mac

ulop

apul

ar10

2 (9

7)

18 (7

8)

1-

-

P

etec

hial

3 (3

)

5 (2

2)

109

32

29 minus

53

830

003

Ast

heni

a80

59 (9

4)

15 (8

8)

063

009

7 minus

412

10

633

Hea

dach

e77

47 (7

6)

13 (8

7)

234

043

minus 1

273

032

6M

yalg

ia85

58 (8

3)

14 (9

3)

308

035

minus 2

719

031

1A

rthra

lgia

6931

(55

)7

(54

)1

190

33 minus

42

30

792

Abd

omin

al P

ain

7011

(29

)8

(57

)4

561

23 minus

16

850

023

Ano

rexi

a65

37 (6

6)

9 (8

2)

258

046

minus 1

450

028

2N

ause

a75

25 (4

0)

11 (9

2)

125

148

minus 1

069

40

021

Vom

iting

8422

(33

)14

(82

)7

51

88 minus

30

120

004

Dia

rrhe

a79

16 (2

6)

12 (7

1)

638

181

minus 2

242

000

4Pr

ostra

tion

8136

(58

)17

(89

)6

731

37 minus

33

030

019

Con

fusi

on o

btun

datio

n10

022

(27

)12

(67

)9

262

61 minus

32

910

001

Deh

ydra

tion

7218

(33

)13

(76

)7

551

95 minus

29

290

003

Tach

ypne

a63

11 (2

5)

17 (8

9)

291

75

35 minus

158

97

lt00

01C

hest

pai

n53

6 (1

5)

3 (2

5)

179

035

minus 9

18

048

6H

epat

omeg

aly

6012

(24

)6

(67

)7

371

46 minus

37

040

015

Sple

nom

egal

y57

4 (8

)

1 (1

7)

200

017

minus 2

377

058

4

a Odd

s rat

io c

ould

not

be

calc

ulat

ed si

nce

ther

e w

ere

no fa

tal c

ases

with

nor

mal

val

ues

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 15Ta

ble

4bC

ompa

rison

of e

xist

ence

of p

re-e

xist

ing

co-m

orbi

ditie

s in

fata

l and

non

-fat

al c

ases

of M

edite

rran

ean

spot

ted

feve

r un

ivar

iate

ana

lysi

s

Co-

mor

bidi

ties

Tot

al P

atie

ntsa (n

)Su

rviv

ors n

()

Fata

l cas

es n

()

Odd

s Rat

io (O

R)

adju

sted

by

stra

in95

C

Ip-

val

ue

Alc

ohol

ism

103

11 (1

3)

7 (4

7)

699

41

99 minus

24

640

002

Toba

cco

use

8710

(14

)3

(18

)1

286

030

minus 5

54

073

5D

iabe

tes

107

8 (9

)

1 (6

)

074

50

08 minus

66

50

793

Car

diac

failu

re10

511

(13

)3

(18

)1

773

041

minus 7

60

044

0R

espi

rato

ry fa

ilure

855

(7

)1

(6

)0

705

007

minus 6

79

076

3C

hron

ic re

nal f

ailu

re85

3 (4

)

0-

--

Hyp

erte

nsio

n89

13 (1

8)

4 (2

5)

183

90

48 minus

70

70

375

a Tota

l pat

ient

s with

ava

ilabl

e da

ta

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 16Ta

ble

4cU

niva

riate

ana

lysi

s of l

abor

ator

y fin

ding

s in

fata

l and

non

-fat

al M

SF c

ases

Lab

orat

ory

Cha

ract

eris

tics

Tot

al P

atie

ntsa (n

)Su

rviv

ors n

()

Fata

l cas

es n

()

Odd

s Rat

io(O

R) a

djus

ted

by st

rain

95

CI

p- v

alue

Ane

mia

lt11

gd

l11

818

(20

)3

(11

)0

590

15 minus

22

00

424

Leuk

open

ia lt

440

0μl

121

13 (1

4)

3 (1

2)

130

031

minus 5

59

071

6Le

ukoc

ytos

is gt

113

00μ

l12

110

(11

)12

(46

)11

13

337

minus 3

680

lt00

01Th

rom

bocy

tope

nia

(Pla

tele

ts P

LT)

122

73 (7

8)

27 (9

6)

68

086

minus 5

378

006

9Pr

othr

ombi

n tim

e (P

T) gt

13

seg

8434

(50

)9

(60

)2

140

59 minus

77

40

246

Parti

al th

rom

bopl

astin

tim

e (P

TT)

PTT

gt35

seg

9522

(29

)17

(85

)36

78

453

minus 2

984

40

001

Glu

cose

gt 1

10 m

gdl

101

52 (6

8)

10(4

2)

036

014

minus 0

95

003

8U

rea

11

236

(42

)25

(93

)21

02

450

minus 9

842

lt00

01C

reat

inin

e gt

12

mg

dl12

042

(46

)26

(93

)20

81

444

minus 9

746

lt00

01So

dium

lt 1

45 m

mol

l12

048

(52

)8

(29

)0

390

15 minus

09

90

047

Pota

ssiu

m lt

35

mm

oll

116

15 (1

6)

2 (8

)

049

010

minus 2

36

037

5To

tal b

iliru

bin

gt12

mg

dl10

916

(19

)23

(92

)45

53

962

minus 2

150

lt00

01A

lani

ne tr

ansa

min

ase

(ALT

)12

058

(62

)24

(89

)4

601

26 minus

16

710

021

Asp

arta

te tr

ansa

min

ase

(AST

)11

270

(81

)25

(96

)4

730

59 minus

38

670

147

Glu

tam

yl tr

ansf

eras

e (G

GT)

77

34 (5

6)

15 (9

4)

958

116

minus 7

892

003

6A

lkal

ine

phos

phat

ase

(ALP

)88

19 (2

7)

15 (8

8)

175

23

59 minus

85

44lt0

001

Cre

atin

e ki

nase

(CK

)75

20 (3

5)

12 (6

7)

506

150

minus 1

709

000

9C

-rea

ctiv

e pr

otei

n (C

RP)

CR

P gt

05

mg

dl66

52 (9

8)

13 (1

00

)b

bb

a Chi

ldre

n w

ere

excl

uded

from

ana

lyse

s of l

abor

ator

y da

ta

b Odd

s rat

io c

ould

not

be

calc

ulat

ed si

nce

ther

e w

ere

no fa

tal c

ases

with

nor

mal

val

ues

The

abno

rmal

val

ues w

ere

cons

ider

ed o

ut o

f nor

mal

rang

e (lo

wer

or e

leva

ted

valu

es) l

imits

est

ablis

h by

eac

h di

ffer

ent h

ospi

tal l

abor

ator

y

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 15Ta

ble

4bC

ompa

rison

of e

xist

ence

of p

re-e

xist

ing

co-m

orbi

ditie

s in

fata

l and

non

-fat

al c

ases

of M

edite

rran

ean

spot

ted

feve

r un

ivar

iate

ana

lysi

s

Co-

mor

bidi

ties

Tot

al P

atie

ntsa (n

)Su

rviv

ors n

()

Fata

l cas

es n

()

Odd

s Rat

io (O

R)

adju

sted

by

stra

in95

C

Ip-

val

ue

Alc

ohol

ism

103

11 (1

3)

7 (4

7)

699

41

99 minus

24

640

002

Toba

cco

use

8710

(14

)3

(18

)1

286

030

minus 5

54

073

5D

iabe

tes

107

8 (9

)

1 (6

)

074

50

08 minus

66

50

793

Car

diac

failu

re10

511

(13

)3

(18

)1

773

041

minus 7

60

044

0R

espi

rato

ry fa

ilure

855

(7

)1

(6

)0

705

007

minus 6

79

076

3C

hron

ic re

nal f

ailu

re85

3 (4

)

0-

--

Hyp

erte

nsio

n89

13 (1

8)

4 (2

5)

183

90

48 minus

70

70

375

a Tota

l pat

ient

s with

ava

ilabl

e da

ta

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 16Ta

ble

4cU

niva

riate

ana

lysi

s of l

abor

ator

y fin

ding

s in

fata

l and

non

-fat

al M

SF c

ases

Lab

orat

ory

Cha

ract

eris

tics

Tot

al P

atie

ntsa (n

)Su

rviv

ors n

()

Fata

l cas

es n

()

Odd

s Rat

io(O

R) a

djus

ted

by st

rain

95

CI

p- v

alue

Ane

mia

lt11

gd

l11

818

(20

)3

(11

)0

590

15 minus

22

00

424

Leuk

open

ia lt

440

0μl

121

13 (1

4)

3 (1

2)

130

031

minus 5

59

071

6Le

ukoc

ytos

is gt

113

00μ

l12

110

(11

)12

(46

)11

13

337

minus 3

680

lt00

01Th

rom

bocy

tope

nia

(Pla

tele

ts P

LT)

122

73 (7

8)

27 (9

6)

68

086

minus 5

378

006

9Pr

othr

ombi

n tim

e (P

T) gt

13

seg

8434

(50

)9

(60

)2

140

59 minus

77

40

246

Parti

al th

rom

bopl

astin

tim

e (P

TT)

PTT

gt35

seg

9522

(29

)17

(85

)36

78

453

minus 2

984

40

001

Glu

cose

gt 1

10 m

gdl

101

52 (6

8)

10(4

2)

036

014

minus 0

95

003

8U

rea

11

236

(42

)25

(93

)21

02

450

minus 9

842

lt00

01C

reat

inin

e gt

12

mg

dl12

042

(46

)26

(93

)20

81

444

minus 9

746

lt00

01So

dium

lt 1

45 m

mol

l12

048

(52

)8

(29

)0

390

15 minus

09

90

047

Pota

ssiu

m lt

35

mm

oll

116

15 (1

6)

2 (8

)

049

010

minus 2

36

037

5To

tal b

iliru

bin

gt12

mg

dl10

916

(19

)23

(92

)45

53

962

minus 2

150

lt00

01A

lani

ne tr

ansa

min

ase

(ALT

)12

058

(62

)24

(89

)4

601

26 minus

16

710

021

Asp

arta

te tr

ansa

min

ase

(AST

)11

270

(81

)25

(96

)4

730

59 minus

38

670

147

Glu

tam

yl tr

ansf

eras

e (G

GT)

77

34 (5

6)

15 (9

4)

958

116

minus 7

892

003

6A

lkal

ine

phos

phat

ase

(ALP

)88

19 (2

7)

15 (8

8)

175

23

59 minus

85

44lt0

001

Cre

atin

e ki

nase

(CK

)75

20 (3

5)

12 (6

7)

506

150

minus 1

709

000

9C

-rea

ctiv

e pr

otei

n (C

RP)

CR

P gt

05

mg

dl66

52 (9

8)

13 (1

00

)b

bb

a Chi

ldre

n w

ere

excl

uded

from

ana

lyse

s of l

abor

ator

y da

ta

b Odd

s rat

io c

ould

not

be

calc

ulat

ed si

nce

ther

e w

ere

no fa

tal c

ases

with

nor

mal

val

ues

The

abno

rmal

val

ues w

ere

cons

ider

ed o

ut o

f nor

mal

rang

e (lo

wer

or e

leva

ted

valu

es) l

imits

est

ablis

h by

eac

h di

ffer

ent h

ospi

tal l

abor

ator

y

J Infect Dis Author manuscript available in PMC 2009 August 15

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

NIH

-PA Author Manuscript

de SOUSA et al Page 16Ta

ble

4cU

niva

riate

ana

lysi

s of l

abor

ator

y fin

ding

s in

fata

l and

non

-fat

al M

SF c

ases

Lab

orat

ory

Cha

ract

eris

tics

Tot

al P

atie

ntsa (n

)Su

rviv

ors n

()

Fata

l cas

es n

()

Odd

s Rat

io(O

R) a

djus

ted

by st

rain

95

CI

p- v

alue

Ane

mia

lt11

gd

l11

818

(20

)3

(11

)0

590

15 minus

22

00

424

Leuk

open

ia lt

440

0μl

121

13 (1

4)

3 (1

2)

130

031

minus 5

59

071

6Le

ukoc

ytos

is gt

113

00μ

l12

110

(11

)12

(46

)11

13

337

minus 3

680

lt00

01Th

rom

bocy

tope

nia

(Pla

tele

ts P

LT)

122

73 (7

8)

27 (9

6)

68

086

minus 5

378

006

9Pr

othr

ombi

n tim

e (P

T) gt

13

seg

8434

(50

)9

(60

)2

140

59 minus

77

40

246

Parti

al th

rom

bopl

astin

tim

e (P

TT)

PTT

gt35

seg

9522

(29

)17

(85

)36

78

453

minus 2

984

40

001

Glu

cose

gt 1

10 m

gdl

101

52 (6

8)

10(4

2)

036

014

minus 0

95

003

8U

rea

11

236

(42

)25

(93

)21

02

450

minus 9

842

lt00

01C

reat

inin

e gt

12

mg

dl12

042

(46

)26

(93

)20

81

444

minus 9

746

lt00

01So

dium

lt 1

45 m

mol

l12

048

(52

)8

(29

)0

390

15 minus

09

90

047

Pota

ssiu

m lt

35

mm

oll

116

15 (1

6)

2 (8

)

049

010

minus 2

36

037

5To

tal b

iliru

bin

gt12

mg

dl10

916

(19

)23

(92

)45

53

962

minus 2

150

lt00

01A

lani

ne tr

ansa

min

ase

(ALT

)12

058

(62

)24

(89

)4

601

26 minus

16

710

021

Asp

arta

te tr

ansa

min

ase

(AST

)11

270

(81

)25

(96

)4

730

59 minus

38

670

147

Glu

tam

yl tr

ansf

eras

e (G

GT)

77

34 (5

6)

15 (9

4)

958

116

minus 7

892

003

6A

lkal

ine

phos

phat

ase

(ALP

)88

19 (2

7)

15 (8

8)

175

23

59 minus

85

44lt0

001

Cre

atin

e ki

nase

(CK

)75

20 (3

5)

12 (6

7)

506

150

minus 1

709

000

9C

-rea

ctiv

e pr

otei

n (C

RP)

CR

P gt

05

mg

dl66

52 (9

8)

13 (1

00

)b

bb

a Chi

ldre

n w

ere

excl

uded

from

ana

lyse

s of l

abor

ator

y da

ta

b Odd

s rat

io c

ould

not

be

calc

ulat

ed si

nce

ther

e w

ere

no fa

tal c

ases

with

nor

mal

val

ues

The

abno

rmal

val

ues w

ere

cons

ider

ed o

ut o

f nor

mal

rang

e (lo

wer

or e

leva

ted

valu

es) l

imits

est

ablis

h by

eac

h di

ffer

ent h

ospi

tal l

abor

ator

y

J Infect Dis Author manuscript available in PMC 2009 August 15