Induction immunotherapy in pediatric heart transplant recipients: a multicenter study

10
PEDIATRICS Induction Immunotherapy in Pediatric Heart Transplant Recipients: A Multicenter Study Robert J. Boucek, Jr., MD, a David Naftel, PhD, b Mark M. Boucek, MD, c Richard Chinnock, MD, d Robert W. Morrow, MD, e Elfriede Pahl, MD, f Sharon DiSano, RN, g and the Pediatric Heart Transplant Study Group Background: The efficacy and safety of induction immunotherapy with antithymocyte antibody preparations (IND) in pediatric heart transplantation is controversial. Experimentally, recipient age is an important determinant of immune responses. The effects on induction immunotherapy were determined by an analysis of outcomes of 465 pediatric (age , 18 years) heart recipients that either did or did not receive IND in the first week post-transplant. Methods: The outcomes of 2 groups who received either OKT3 (n 5 101) or rabbit polyclonal antithymocyte serum (N/R-ATS, n 5 105) were compared with 255 recipients who did not receive antithymocyte antibodies. The study population were all heart recipients enrolled in the Pediatric Heart Transplant Study Group (PHTS) between January 1993 and December 1995 and followed up to 36 months. Results: Overall mortality and death due to rejection were lowest with N/R-ATS IND (8/105 and 1/105, respectively) compared with the no-induction group (58/255 and 8/ 255, respectively) or the OKT3 group (22/101 and 7/101, respectively) with significance of p 5 0.001 and 0.06 respectively. Late mortality beyond 30 days after transplant was lowest with N/R-ATS IND compared with the OKT3 and no IND (p 5 0.01). Induction did not affect cumulative infections, deaths due to infection, or the frequency of malignancies. Patients excluded from N/R-ATS induction had the highest mortality (18/ 43), suggesting that the protocol’s exclusion criteria identified a high-risk group. To minimize potential effect(s) of exclusion bias, patients transplanted at centers participating in the N/R-ATS induction trial were reanalyzed with a post hoc intent-to- treat analysis assigning patients by center (IND or no IND) irrespective of actual treatment. With this analysis overall mortality was 18% for N/R-ATS centers, 21% for OKT3 centers, and 26% for centers not using IND (p 5 0.3). The mortalities of recipients , 6 months old at transplant were lowest at centers using N/R-ATS and OKT3 IND compared to centers not using IND (p 5 0.04). Cumulative rejection (0.8 vs 1.2 rejection/pt/year, p 5 0.01) and freedom from rejection death (99% vs 93% at From the University of South Florida/All Children’s Hospital, St. Petersburg, Florida; a University of Alabama, Birmingham, Alabama; b University of Colorado/The Children’s Hospital, Denver, Colorado; c Loma Linda University Medical Center, Loma Linda, California; d University of Arkansas/Arkansas Children’s Hospital, Little Rock, Arkansas; e Children’s Memo- rial Hospital, Chicago, Illinois; f and All Children’s Hospital, St. Petersburg, Florida a , g Corresponding author: Robert J. Boucek, Jr., MD, Professor of Pediatrics, Cardiology, Director, Pediatric Heart Transplant Program, University of South Florida/All Children’s Hospital, 880 Sixth Street South, Suite 260, St. Petersburg, Florida 33701-4827. Telephone: 813-892-8984. Fax: 813-897-4803. E- mail: [email protected] Submitted June 30, 1998; accepted Dec. 7, 1998. Copyright © 1999 by the International Society for Heart and Lung Transplantation. 1053-2498/99/$–see front matter PII S1053-2498(98)00076-X 460

Transcript of Induction immunotherapy in pediatric heart transplant recipients: a multicenter study

PEDIATRICS

Induction Immunotherapy inPediatric Heart TransplantRecipients: A Multicenter StudyRobert J. Boucek, Jr., MD,a David Naftel, PhD,b Mark M. Boucek, MD,c

Richard Chinnock, MD,d Robert W. Morrow, MD,e Elfriede Pahl, MD,f

Sharon DiSano, RN,g and the Pediatric Heart Transplant Study Group

Background: The efficacy and safety of induction immunotherapy with antithymocyteantibody preparations (IND) in pediatric heart transplantation is controversial.Experimentally, recipient age is an important determinant of immune responses. Theeffects on induction immunotherapy were determined by an analysis of outcomes of 465pediatric (age , 18 years) heart recipients that either did or did not receive IND in thefirst week post-transplant.

Methods: The outcomes of 2 groups who received either OKT3 (n 5 101) or rabbitpolyclonal antithymocyte serum (N/R-ATS, n 5 105) were compared with 255recipients who did not receive antithymocyte antibodies. The study population were allheart recipients enrolled in the Pediatric Heart Transplant Study Group (PHTS)between January 1993 and December 1995 and followed up to 36 months.

Results: Overall mortality and death due to rejection were lowest with N/R-ATS IND(8/105 and 1/105, respectively) compared with the no-induction group (58/255 and 8/255, respectively) or the OKT3 group (22/101 and 7/101, respectively) with significanceof p 5 0.001 and 0.06 respectively. Late mortality beyond 30 days after transplant waslowest with N/R-ATS IND compared with the OKT3 and no IND (p 5 0.01). Inductiondid not affect cumulative infections, deaths due to infection, or the frequency ofmalignancies. Patients excluded from N/R-ATS induction had the highest mortality (18/43), suggesting that the protocol’s exclusion criteria identified a high-risk group. Tominimize potential effect(s) of exclusion bias, patients transplanted at centersparticipating in the N/R-ATS induction trial were reanalyzed with a post hoc intent-to-treat analysis assigning patients by center (IND or no IND) irrespective of actualtreatment. With this analysis overall mortality was 18% for N/R-ATS centers, 21% forOKT3 centers, and 26% for centers not using IND (p 5 0.3). The mortalities ofrecipients , 6 months old at transplant were lowest at centers using N/R-ATS andOKT3 IND compared to centers not using IND (p 5 0.04). Cumulative rejection (0.8vs 1.2 rejection/pt/year, p 5 0.01) and freedom from rejection death (99% vs 93% at

From the University of South Florida/All Children’s Hospital, St.Petersburg, Florida;a University of Alabama, Birmingham,Alabama;b University of Colorado/The Children’s Hospital,Denver, Colorado;c Loma Linda University Medical Center,Loma Linda, California;d University of Arkansas/ArkansasChildren’s Hospital, Little Rock, Arkansas;e Children’s Memo-rial Hospital, Chicago, Illinois;f and All Children’s Hospital, St.Petersburg, Floridaa,g

Corresponding author: Robert J. Boucek, Jr., MD, Professor ofPediatrics, Cardiology, Director, Pediatric Heart Transplant

Program, University of South Florida/All Children’s Hospital,880 Sixth Street South, Suite 260, St. Petersburg, Florida33701-4827. Telephone: 813-892-8984. Fax: 813-897-4803. E-mail: [email protected]

Submitted June 30, 1998; accepted Dec. 7, 1998.Copyright © 1999 by the International Society for Heart and

Lung Transplantation.1053-2498/99/$–see front matter PII S1053-2498(98)00076-X

460

year 1, p 5 0.02) of the N/R-ATS centers were lower compared to OKT3 centers butwere not different from centers not using IND.

Conclusion: Following orthotopic transplantation, induction immunotherapy can exertthe enduring benefit of reducing late deaths, a possible surrogate for rejection severity,in recipients less than 6 months of age, while not being associated with higher rates ofinfectious or malignant complications. Since polyclonal anti-T cell antibodypreparations appears superior to OKT3 induction in pediatric recipients, the efficacy ofATS induction should be further evaluated in a randomized prospective study inpediatric heart recipients. J Heart Lung Transplant 1999;18:460–469.

Pediatric heart transplantation is an effectivetherapeutic modality for children with end-stage orinoperable cardiac disease.1,2 Acute rejection canlead to graft dysfunction, graft failure, and death.3,4

As many as one-quarter of the pediatric transplantrecipients will have acute rejection that is eitherrecurrent or does not resolve despite standard im-munosuppressive agents. Thus, reducing the overallincidence and/or severity of rejection remains animportant therapeutic goal in pediatric cardiactransplant recipients.

One strategy to reduce the incidence and/or se-verity of rejection is induction therapy. Prophylacticuse of cytolytic antibodies in the immediate post-transplant period has been effective in animal mod-els.5 Clinical trials in adult cardiac transplant recip-ients have been reported with prophylacticadministration of polyclonal antithymocyte prepara-tions such as N/R-ATS, ATGAM, antilymphocyteglobulin (ALG), and monoclonal antithymocytepreparations such as OKT3. The effects observedhave been mainly to delay the onset of rejection6,7

without demonstrating any enduring benefit.The efficacy of induction immunotherapy in pe-

diatric heart transplantation has not been resolved.Preliminary reports have suggested that induction

immunotherapy with N/R-ATS reduced the rate ofrejection8-10 and overall mortality in recipients, themajority of which were less than 6 years old.11

Induction immunotherapy with OKT3, however, didnot appear to affect the rate of rejection in olderpediatric heart recipients.12 These observations sug-gest that recipient age can affect the efficacy ofinduction immunotherapy. In the initial experimen-tal animal models in which induction immunother-apy had achieved tolerance, the recipients weredevelopmentally immature.13 Immaturity of the im-mune system is not an intrinsic property of thenewborn immune system, but the immature natureof the antigen presenting cells can determinewhether the outcome of antigen exposure is toler-ance or immune response.14 Recent analyses ofrejection from PHTS preliminarily indicated thatage at transplant less than 6 months reduced post-transplant mortality from rejection.15 To date, theinteraction of recipient age on induction immuno-therapy in man has not been reported.

The primary aim of this study was to determine ifanti-T cell induction immunotherapy in infants andchildren had an effect on patient and graft survival,and the frequency and severity of rejection. Corol-lary aims were to determine the effect of recipientage on the efficacy of anti-T cell induction therapy,

TABLE I Patient profiles grouped by inductiontreatment status

Induction therapy Patients CentersAge (yrs) at Tx

mean 6 SD

ATS 105 4 3.3 6 4.9OKT3 101 9 7.6 6 6.7ATGam 4 3 2.5 6 4.7None (ATS center)* 43 4 1.7 6 3.8None (non-ATS center) 212 14 6.3 6 6.3

Total 465 21 5.5 6 6.3

p , .0001*Excluded per protocol.

TABLE II Mortality of recipients by inductiontreatment status

Induction therapy n

Deaths

No. %

ATS 105 8 8%OKT3 101 22 22%None (ATS center) 43 18 42%None (non-ATS center) 212 54 25%

Total 461 102 22%

p , .0001

The Journal of Heart and Lung Transplantation Boucek et al. 461Volume 18, Number 5

and to compare the safety and efficacy of OKT3 andN/R-ATS. The results were preliminarily reported.11

METHODS

A total of 465 recipients of orthotopic cardiactransplants under the age of 18 years of age wereenrolled and followed in the PHTS between January1, 1993 and December 31, 1995. Of these, 210pediatric patients received induction immunother-

apy. Induction immunotherapy was defined as cyto-lytic therapy given in the immediate post-transplantperiod in the absence of rejection. Three cytolytictherapies were used: N/R-ATS (Applied MedicalResearch, Nashville, TN), OKT3 (Ortho Biotech,Raritan, NJ), and ATGAM (Upjohn, Kalamazoo,MI). The outcomes of 2 of the induction therapiesused, N/R-ATS (n 5 105) and OKT3 (n 5 101) wereretrospectively reviewed and compared with 255

TABLE III Cause of death by induction treatment status

Cause of death N

Induction therapy None

ATS OKT3 ATS center non-ATS center

n % of 8 n % of 22 n % of 18 n % of 54

Early graft failure 27 2 25% 2 9% 8 44% 15 28%Rejection 19 1 12% 6 27% 3 17% 8 15%Infection 20 3 38% 6 27% 3 17% 8 15%Sudden 12 1 12% 3 14% 1 6% 7 13%Non-specific graft failure 6 0 — 1 5% 1 6% 4 7%CAD 3 0 — 0 — 0 — 3 6%Other 15 1 12% 4 18% 2 11% 9 17%

Total 102 8 100% 22 100% 18 100% 54 100%

FIGURE 1 The effects of induction immunotherapy on overall survival (%) are presentedas a Kaplan-Meier actuarial with log-rank testing. The 2 groups were identified as receivingeither N/R-ATS (ATS) or OKT3 induction immunotherapy. Two groups were identified asnot receiving induction immunotherapy based on whether the patients were or were not ata center where N/R-ATS induction was also used. The number of patients in each of the 4groups is shown in parentheses.

462 Boucek et al. The Journal of Heart and Lung TransplantationMay 1999

pediatric recipients who did not receive inductiontherapy (Table I). Only 4 recipients were givenATGAM IND, and this group was not considered inthe analysis because of the small number of patients.All patients received immunosuppression regimenswith anti-inflammatory corticosteroids, calcinurininhibitor (cyclosporine A, Sandimmune), and anti-proliferative therapy (azathioprine) at the discretionof each recipient’s physician.

N/R-ATS was administered at a dose of 0.5 mL/kgintravenously daily for a total of 5 consecutive days.Forty-three of the 252 pediatric recipients did notreceive induction at institutions participating in aphase 1 trial of N/R-ATS. Ten recipients wereexcluded based on study exclusion criteria: positiveskin test (n 5 1), active infection at the time forinduction (n 5 4), or open sternotomy (n 5 5).Eight patients died prior to receipt of the first dose.The remainder of the 43 were not given N/R-ATS,either because the drug was not available in thepharmacy at the time of induction (n 5 6) orbecause informed consent was not obtained or it wasrefused (n 5 20). Because of the potential forselection bias, the results of all 43 patients wereconsidered separately. In addition, selected analyseswere performed as intent to treat, combining pa-tients receiving, and excluded from, induction ther-apies.

Statistical AnalysisThe data was initially examined with simple contin-gency tables, chi-square tests for comparing propor-tions, t tests, and cumulative frequency distributions.The Nelson method for estimating cumulativeevents in the presence of censoring was used forrejection episodes. Kaplan-Meier actuarial tech-niques were combined with the log-rank test foranalysis of death.

RESULTSThe mean follow-up time of all survivors was 17months (range, 1 to 36 months). The age at trans-plant was different between the 4 groups (p ,0.0001; Table I). There was a small difference in theduration of follow-up among survivors. The dura-tions of follow-up were: 17.7 6 1 month in the groupthat received N/R-ATS induction, 14.2 6 2.3months for the group that did not receive inductionat N/R-ATS centers, 19.4 6 1.18 months for thegroup that received OKT3 induction, and 15.6 60.85 months for the group that did not receiveinduction (p 5 0.03).

Survival

Overall survival of the 4 groups was significantlydifferent (p , 0.0001, see Table II, chi-square test)with the highest survival in the group receiving

FIGURE 2 The effects of induction immunotherapy on overall survival (%) of patients inthe 4 groups who survived at least 30 days are presented as a Kaplan-Meier actuarial withlog-rank testing.

The Journal of Heart and Lung Transplantation Boucek et al. 463Volume 18, Number 5

N/R-ATS induction therapy. Of the 102 (102/465)pediatric patients who died, only 8 of 105 (8%)patients receiving N/R-ATS induction died com-pared with 54 of 212 (25%) patients who did notreceive induction (Table II). Induction with OKT3did not improve overall survival compared withpatients at institutions who did not routinely useinduction (Table II). At those institutions whereinduction therapy was administered, the group that

was excluded and did not receive induction had thehighest mortality. Eighteen of the 43 patients ex-cluded from receipt of N/R-ATS per protocol died,a 42 % mortality (Table II), and 1 of 4 patientsexcluded from OKT3 died, a 25% mortality. Whensurvival rates were estimated over the follow-upperiod by actuarial survival analysis (Figure 1),N/R-ATS induction had the highest survival com-pared to the other groups. Multivariate analysisconfirmed that no N/R-ATS induction both at N/R-ATS centers (p 5 0.004) and at all other centersTABLE IV Number of recipients rejection free by

induction treatment status

Induction therapy n

Patients with atleast onerejectionepisode

No. %

ATS 105 66 63%OKT3 101 64 63%None (ATS center) 43 26 60%None (non-ATS center) 212 118 56%

Total 461 274 59%

p 5 .5

TABLE V Number of malignancies by inductiontreatment status

Induction therapy n

Malignancies

No. %

ATS 105 2 2%OKT3 101 3 3%None (ATS center) 43 0 0%None (non-ATS center) 212 5 2%

Total 461 10 2%

p 5 .7

FIGURE 3 The effects of induction immunotherapy on rejection mortality (% free fromrejection death) are presented as a Kaplan-Meier actuarial. The number of patients whosedeath was directly attributed to rejection and the patient number for each of the 4 groupsare shown in parentheses. Log-rank testing was performed for all 4 groups, designated asoverall, and for ATS vs other groups.

464 Boucek et al. The Journal of Heart and Lung TransplantationMay 1999

(p 5 0.003) are early risk factors of death aftertransplant. Previous surgical intervention was a con-stant risk factor for death after transplant.

The timing of the deaths was analyzed for the 4groups. The largest number of deaths were attrib-uted to early graft failure (27/102) not related torejection or infection (Table III). In the groupexcluded from N/R-ATS, mortality was largely early,with 12 of the 18 deaths in the first 30 days post-transplant and 10 in the first 24 hours post-trans-plant, as might be anticipated by the exclusioncriteria. The early mortalities of the other groupswere not different. Induction immunotherapy didaffect the outcome of recipients who survivedgreater than 30 days. Mortality rates for recipientswho survived greater than 30 days were as follows: 4of 96 recipients (; 4%) who received N/R-ATSinduction died; 13 of 89 recipients (; 15%) whoreceived OKT3 died; and 49 of 175 recipients (;28%) who did not receive induction at non-ATScenters died; 6 of 31 recipients who did not receiveinduction at the ATS centers died (p 5 0.01;Figure 2).

Causes of death other than early graft failure thatwere identified in this study were rejection (n 5 19),infection (n 5 20), non-specific graft failure (n 5 6),sudden (n 5 12), and coronary artery disease (n 5

3). Only 16 deaths could not be classified into thesegroups.

Rejection

A total of 483 rejection episodes—defined as inten-sified immunosuppression associated with diagnos-tic biopsy (334, ; 70%) and/or echocardiogram(124, ; 26%)—were diagnosed in 277 pediatricpatients. Death due to rejection was low but was lessfor the group receiving N/R-ATS (1/105) comparedto the other groups (p 5 0.05) when analyzed byfreedom from death due to rejection (Figure 3). Themortality of rejection (rejection deaths/number ofrejection episodes) was the highest in the groupreceiving OKT3 (6/123; Table V) and lowest in thegroup receiving N/R-ATS (1/114). The group ex-cluded from N/R-ATS had 3 deaths due to rejection(17% of deaths in this group) and the highestmortality from rejection (6 rejection deaths perrejection episode, Table III). The mortality of rejec-tion in the other groups were not different (TableIII). Freedom from rejection death with intent-to-treat analysis was significantly less for N/R-ATSthan OKT3 induction (99% vs 93% at year 1; p 50.02).

Induction therapy did not appear to affect theincidence and frequency of rejection when com-

FIGURE 4 The effect of induction immunotherapy on the number of rejection episodesper patient are presented for the 4 groups. Log-rank testing was performed comparing ATSvs other groups.

The Journal of Heart and Lung Transplantation Boucek et al. 465Volume 18, Number 5

pared with no induction. The number of recipientsfree from rejection was not different between thegroups (Table IV). In an analysis that censors fordeath, the cumulative number of rejection episodesfor the 4 patient groups was not different (Figure 4).However, when compared with OKT3, N/R-ATSdid reduce rejection frequency (0.8 vs 1.2 rejection/pt/year, p 5 0.01, intent-to-treat analysis).

Infection

A total of 439 episodes of infection—defined as lifethreatening and/or requiring IV therapy—were di-agnosed and reported. Induction therapy did notaffect the incidence or frequency of infections.Cumulative infection rates at 24 months were notdifferent, 1.2 and 1.5 for N/R-ATS and OKT3induction groups, respectively, and 1.4 and 1.1 forgroups not receiving induction at N/R-ATS andother institutions, respectively. Overall mortalitydue to infection and cumulative infection frequencywas not significantly increased by induction therapy(Figure 5). CMV infection were not different (p 50.23) between induction and no induction groupswith CMV infections reported in 21 of 206 recipi-ents receiving induction and 24 of 255 recipients notreceiving induction.

Malignancy

A total of 10 malignancies were reported over thefollow-up period with no significant differences be-tween the 4 groups (Table V). In specific, inductionimmunotherapy did not increase the frequency ofmalignancies over the interval of follow-up. How-ever, both the short duration of follow-up and thesmall number of malignancies limit the power of thecomparison.

DISCUSSION

This study has shown that pediatric cardiac recipi-ents who received induction immunotherapy withN/R-ATS had significantly lower mortality com-pared to recipients who did not receive any induc-tion protocol (Figure 1). The effect of N/R-ATSinduction appeared to be enduring, as N/R-ATSinduction reduced “late” deaths (Figure 2, sur-vival . 30 days). This study suggests that N/R-ATSand OKT3 induction were not comparable. Induc-tion protocols using OKT3 did not improve survivalcompared to the patient group who receive N/R-ATS induction (Figure 1). With an intent-to-treatanalysis, freedom from rejection and death fromrejection were significantly less for N/R-ATS thanOKT3 induction.

FIGURE 5 The effects of induction immunotherapy on infection mortality (% free frominfection death) are presented as a Kaplan-Meier actuarial. Log-rank testing wasperformed comparing all 4 groups, designated as overall, and ATS vs other groups.

466 Boucek et al. The Journal of Heart and Lung TransplantationMay 1999

The mechanism for the apparent reduction inmortality with induction therapy (Figure 1) is notknown, but one possible mechanism is immuno-modulation, as there is a reduction in mortality dueto rejection (p 5 0.06; Figure 3). Both monoclonaland polyclonal anti-T cell antibodies result in imme-diate lymphocytopenia and can have long-term im-munomodulatory effects.5 Recently, CD4 T cellnumbers have been reported to be reduced for up to60 months following polyclonal anti-T cell antibodyadministration to renal recipients (adult).16 Theselong-lasting (3-year) immunomodulatory effectscould be related to the development of “tolerance.”That cumulative rejection (Figure 4) and the num-ber of recipients who were rejection free (Table IV)were not different raises questions about an endur-ing immunologic benefit of induction immunother-apy in this study.

The most immature recipients—those less than 6month of age—had the largest mortality differencebetween induction (24 deaths in 121 infants, 19%)and no induction (19 deaths in 52 infants, 37%).Induction immunotherapy reduced overall mortality

for recipients under 6 months old (p 5 0.04, Figure7). In preliminary phase 1 and 2 trials in pediatriccardiac recipients, N/R-ATS induction immunother-apy appears to be efficacious. In a randomized,placebo-controlled trial with pediatric heart recipi-ents (n 5 59), N/R-ATS, when used in an inductionprotocol, decreased overall rejection frequency by ;50% (1.72 vs 0.85 rejection episodes/pt in year 1).8,12

N/R-ATS significantly decreased the number ofpatients with at least 1 episode of rejection from68% (17/25) in the placebo group to only 38%(13/34).8,12

The monoclonal anti-T cell antibody OKT3 ap-pears to have little enduring benefit and possiblyunacceptable risks when used for induction in pedi-atric12 and adult cardiac transplant recipients.17

Polyclonal anti-T cell antibodies such as N/R-ATShave been reported to be either no different than,6

or superior to,18 OKT3 for induction-prophylacticimmunotherapy in adult cardiac transplant. Differ-ences in the immunomodulatory effects of monoclo-nal OKT3, which is a potent T-cell activator, andpolyclonal anti-T cell antibodies, which also contain

FIGURE 6 The effects of induction immunotherapy on overall mortality (% survival) arepresented as a Kaplan-Meier actuarial. Patients transplanted at centers participating in theN/R-ATS induction trial were analyzed with a post hoc intent-to-treat analysis assigningpatients by center irrespective of actual treatment. Patients (n 5 148) from centers usingN/R-ATS were assigned to the ATS group. Patients (n 5 107) from centers using OKT3were assigned to the OKT3 group. Patients from institutions not using either N/R-ATS orOKT3 induction were assigned to the group designated in this figure as none. Log-ranktesting was performed for all 4 groups.

The Journal of Heart and Lung Transplantation Boucek et al. 467Volume 18, Number 5

“blocking” antibodies such as anti-CD4, may ac-count for these differences.5 This study did notidentify a significant increase in either infectious(Figure 5) or malignant (Table IV) complications ofinduction immunotherapy in pediatric heart recipi-ents. This study does forward new evidence thatpolyclonal anti-T cell antibody preparations may besuperior to OKT3 induction in pediatric recipientsin reducing rejection frequency.

Limitations of the Study

The power of the multicenter design used in thisanalysis must be balanced with its limitations. Inspecific, the maintenance immunosuppression andthe treatment of rejection episodes were not stan-dardized between participating institutions. Sinceinduction tended to be an institutional decision,some of the observed differences between inductionand no induction may be attributed to institutionaldifferences (Figure 6).

Initial analysis of the data indicated that highestmortality (; 42%) following cardiac transplantationwas in recipients who were excluded from partici-pating in N/R-ATS or OKT3 trials (Figure 1).

Because this group’s mortality was so different, apossible interpretation of results shown in Figure 1is that the N/R-ATS protocol’s exclusion criteriaintroduced an inadvertent exclusion bias. This inter-pretation is supported by the post hoc center-basedintent-to-treat analysis of mortality which was notsignificantly affected by induction (Figure 6). How-ever, late mortality also was reduced by induction(Figure 2) and it would be more difficult to “select”for late mortality. Furthermore, approximately halfof the group of 43 infants who did not receiveinduction would not be considered at high risk andwould not a priori contribute to selection biasfavoring N/R-ATS induction. Thus, while this posthoc center-based intent-to-treat analysis reduces thelikelihood of a type 1 error, it can increase thelikelihood of a type 2 error and magnify centerdifferences.

Despite these limitations, we believe that fromthese analyses a prospective randomized trial ofinduction immunotherapy with polyclonal anti-Tcell antibodies vs placebo is justified in children. Inaddition, a direction identified for future study is thehypothesis that immunologic immaturity is a clini-

FIGURE 7 The effects of induction immunotherapy on overall mortality (% survival) forrecipients that are less than 6 months at the time of transplant are presented in a Kaplan-Meier actuarial format. The number of patients whose death was directly attributed torejection and the patient number for each of the 4 groups are shown in parentheses.Patients transplanted at centers in the N/R-ATS induction trial (n 5 89) and in centersusing OKT3 induction (n 5 32) were compared with centers not using induction (n 5 52).

468 Boucek et al. The Journal of Heart and Lung TransplantationMay 1999

cally significant determinant of efficacy of polyclonalanti-T cell induction protocol.

The authors express their sincere appreciation for theefforts of all of the transplant coordinators at the partic-ipating centers. All of them should be cited for theircontributions to the collection of the factual base for thisstudy. The editorial assistance of Terra Sroka is gratefullyacknowledged.

REFERENCES1. Bailey LL, Concepcion W, Shattuck H, et al. Method of heart

transplantation for treatment of hypoplastic left heart syn-drome. J Thorac Cardiovasc Surg 1986;92(1):1-5.

2. Boucek MM, Hodgkins DD, Mathis CM, Kanakriyeh MS,Boucek RJ Jr, Gundry SR, Bailey LL. Accuracy of echocar-diographic rejection surveillance in infant cardiac transplan-tation. J Heart Transplant 1990;9:63.

3. Myles JL, Ratliff NB, McMahon JT, et al. Reversibility ofmyocyte injury in moderate and severe acute rejection incyclosporine-treated cardiac transplant patients. Arch PatholLab Med 1987;111:947-52.

4. Fricker FJ, Griffith II, Hardesty RL, et al. Experience withheart transplantation in children. Pediatrics 1987;79:138-46.

5. Bonnefoy-Bernard N, Revillard JP. Mechanisms of immuno-suppression induced by antithymocyte globulins and OKT3.J Heart Lung Transplant 1996;15:435-442.

6. Frist WH, Merrill WH, Eastburn TE, Atkinson JB, StewartJR, Hammon JW, Bender HW. Unique antithymocyte serumversus induction immunotherapy after heart transplantation.J Heart Transplant 1990;9:489-494.

7. Carey J, Frist WH. Use of polyclonal antilymphocytic prep-arations for prophylaxis in heart transplantation. J HeartTransplant 1990;9:297-300.

8. Boucek MM, Mathis CM, Chinnock RE, Kanakriyeh MS,Nehlsen-Cannarella S, Gundry SR, Bailey LL, Boucek RJJr. Polyclonal versus monoclonal anti-T cell antibody

therapy in infant heart transplantation. J Heart Transplant1991;10:161.

9. Boucek MM, Mathis CM, Labeck LK, Nehlsen-Cannarella S,Gundry S, Bailey L. Prophylactic anti-thymocyte serum re-duces rejection frequency after pediatric heart transplanta-tion. J Heart Lung Transplant 1992;11:205.

10. Boucek MM, Mathis CM, Gundry SR, Bailey L. Antithymo-cyte serum prophylaxis: enduring reduction of rejectionepisodes in pediatric heart transplantation. J Heart LungTransplant 1994;13:S35.

11. Boucek RJ Jr, Naftel D MD, Boucek MM, Chinnock R,Morrow RW, Pahl E, DiSano S. Induction immunotherapy inpediatric heart transplant recipients: a multi-institutionalevaluation. J Heart Lung Trans 1998;16:24.

12. Shaddy RA, Bullock EA, Morwessel NJ, Hannon DW, et al.Murine monoclonal CD3 antibody (OKT3)-based early re-jection prophylaxis in pediatric heart transplantation. J HeartLung Transplant 1993;12:434-9.

13. Billingham RE, Brent L, Medawar PB. Actively acquiredtolerance of foreign cells. Nature 1953;172:603-606.

14. Ridge JP, Fuchs EJ, Matzinger P. Neonatal tolerance re-visted: turning on newborn T cells with dendritic cells.Science 1996:217:1723-1729.

15. Rotando K, Naftel D, Boucek RJ Jr, et al. Allograft rejectionfollowing cardiac transplantation in infants and children: amulti-institutional study. J Heart Lung Transplant 1996;15:S80.

16. Muller TF, Grebe SO, Neuman C, Heymanns J, Radsak L,Sprebger H, Lange H. Persistent long-term changes in lym-phocyte subsets induced by polyclonal antibodies. Transplan-tation 1997;64(10):1432-1437.

17. Frank RA, Sanchez JA, Addonizio LJ, Hsu DT, Rose EA,Micher RE. OKT3 induction therapy increases early infec-tions and does not reduce early rejections in pediatric hearttransplantation. J Heart Lung Transplant 1991;10:161.

18. Carrier M, Jenicek M, Pelletier. Value of monoclonal anti-body OKT3 in solid organ transplantation: a meta-analysis.Transplantation Proc 1992;24:2586-2591.

The Journal of Heart and Lung Transplantation Boucek et al. 469Volume 18, Number 5