- Email: [email protected]
IMPORTANCE OF HLA ANTIGEN SPLITS FOR KIDNEY TRANSPLANT MATCHING
Saturday 9 July
IMPORTANCE OF HLA ANTIGEN SPLITS FOR KIDNEY TRANSPLANT MATCHING GERHARD OPELZ FOR THE COLLABORATIVE TRANSPLANT STUDY
Department of Transplantation Immunology, Institute of Immunology, University of Heidelberg, Heidelberg, Federal Republic of Germany Whether
matching for HLA antigen "splits" results in better transplant outcome than matching for "broad" HLA antigens was investigated in 30 000 first cadaver kidney transplants. At Summary
three years, there was an 18% difference between the survival rates of grafts with 0 or 4 mismatches among transplants typed for HLA-A and B antigen splits whereas the difference in transplants typed for broad antigens was only 2%. Analysis of HLA-A, B antigens together with HLA-DR antigens showed an even greater advantage of matching for antigen splits: the difference in survival at three years between grafts with 0 or 6 mismatches for HLA-A, B, DR was 31% when antigen splits were analysed, in contrast to a 6% difference with broad antigens. These results indicate that typing for HLA antigen splits is important in renal transplantation, that the potential benefit of HLA matching in renal transplantation is greater than currently accepted, and that HLA typing and kidney allocation routines must be refined in order to exploit this potential. Introduction TYPING for HLA antigens can be done for so-called broad antigen specificities or for their subunits which are commonly called antigen splits.! For example, individuals possessing the HLA antigen HLA-A9 can be characterised further as possessing either HLA-A23 or HLA-A24, two subunits or "splits" of HLA-A9. Because the serological reagents for determination of antigen splits are scarce, typing for splits is more difficult and more expensive than typing for broad antigens. The general view is that, for the
1988
purpose of cadaver kidney matching, consideration of antigen splits is not necessary. Organ exchange organisations base their allocation of kidneys on broad HLA antigen assignments, and many transplant centres are not typing for antigen splits. This strategy found support in a
report by Sanfilippo et al2who found that, in a series of 3800 patients, the correlation of HLA-A, B matching with graft survival was similar whether broad antigens or antigen splits were analysed; they advocated matching for a small number of serologically "cross-reacting" HLA antigens, arguing that, apart from reducing the cost of HLA typing, this would carry the advantage that recipients and donors could be "matched" more easily because fewer antigen specificities would have to be considered. In this report we examine a much larger series of transplants. The results are relevant to recent controversial reports on the efficacy of HLA matching in cyclosporintreated kidney transplant recipients3-7 and to the continuing discussion about equitable distribution of cadaver
organs.8-l0 Patients and Methods The data on which this analysis was based were contributed by 240 centres that are participating in the Collaborative Transplant Study.ll HLA typings were performed in the individual centres’ laboratories. Typing and baseline medical data were reported to the study centre at the University of Heidelberg. Clinical follow-up information was requested at 3, 6, 9, 12, 24, and 36 months after
transplantation. The transplants were separated into two categories according to whether typing for HLA-A, B antigen splits had or had not been done. One category consisted of transplants in which any of the following antigen splits were reported in the recipient or donor and in which broad antigens were not reported: HLA-A 23, 24, 25, 26, 29,30,31,32,33,34,66,68,69, and HLA-B 38, 39, 44, 45, 49, 50, 51,52,54,55,56,57,58,60,61,62,63,64,65,71,72. Antigens for which no splits are known1 were counted as equivalent to splits. The second category consisted of transplants where any of the following broad antigens were reported: HLA-A 9,10,19,28, and HLA-B 5, 12,14,15,16,17,21,22,40,70. When a recipient was reported with broad antigens and a donor with antigen splits, or vice versa, the 8602
©
62
antigen splits were converted by the computer to the appropriate broad antigens and the transplant was counted in the broad antigen category. For transplants with HLA-DR typing information, the antigens HLA-DR 1-10 were analysed. Splits of HLA-DR antigens were not considered because typing for splits was not done in
most
laboratories that submitted data.
The donors
analysis was restricted to first transplants from cadaver performed from 1982 to 1987. Graft survival rates were computed by actuarial methods and statistical significance was estimated by log-rank or weighted regression analysis12 as indicated. In the latter, the dependent variable was the percentage survival at three years, the independent variable was the number of mismatched HLA antigens, and the weighting factor was the number of patients studied. We estimated the standard error of the regression coefficient from the standard error of the percentage survival, and statistical significance was judged on the basis of the ratio of the regression coefficient to its standard error. No exclusions were
made.
Fig 1-Influence
ofHLA-A and HLA-B mismatches in
patients studied are indicated for each curve. The horizontal line at 70% was
Fig 2-Analysis of HLA-A, HLA-B, and HLA-DR mismatches to
Matching for HLA-A and HLA-B antigens correlated with graft survival in transplants that were typed for antigen splits but not in transplants typed for broad antigens (fig 1). When antigen splits were analysed, grafts with 0 mismatches had an 18% higher success rate at three years than grafts The analysis of with 4 mismatches (p log rank <0’0001). broad antigens showed a difference of only 2% (NS). The three-year success rates of transplants with 0, 1, or 2 HLA-DR mismatches were: 65% (SE 0-7),62% (SE 0-6), and 58% (SE 0-9), respectively. When HLA-DR mismatches were added to the number of mismatches for HLA-A and HLA-B, a striking potentiating effect on the correlation with graft survival was observed among transplants typed for HLA-A, B antigen splits. The three-year survival rate of transplants with 0 mismatches for
transplants typed for antigen splits or broad antigens.
The number of mismatched HLA antigens and the number of drawn to allow for better comparison.
Antigen splits and broad antigens refer
Results
in first cadaver
HLA-A and HLA-B locus
antigens.
kidney transplants.
63
Fig 3-Effect of mismatching for HLA antigen splits and broad antigens in cyclosporin-treated patients. HLA A, B, DR was 80% (SE 2-6) (SE 2-7) rate with 6 mismatches (p
compared with a 49% log rank <0’0001). In contrast, among transplants typed for broad HLA-A, B antigens, the effect of matching on graft outcome was much less pronounced (fig 2). Grafts with 0 mismatches had only a 6% higher survival rate at three years than grafts with 6 mismatches (NS). Fig 3 shows that the results described above are applicable to patients who were immunosuppressed with cyclosporin. Among transplants typed for HLA-A, B antigen splits, the correlation of matching with outcome was excellent (p regression < 00001) whereas in transplants typed for broad antigens the results were unimpressive. Two factors are likely to have contributed to the observed difference in results. It can be assumed that laboratories that typed for antigen splits possessed greater expertise in HLA serology than those typing for broad antigens.
Consequently, erroneous HLA typing results may have been contained in the data reported by the latter. Of particular interest from a scientific viewpoint is whether a mismatch between recipient and donor that is evident when splits but not when broad antigens are considered leads to an increased rejection rate. To investigate this, we programmed the computer to convert the antigen splits (reported by "expert" laboratories) into broad antigens, and we reanalysed those transplants in which the conversion resulted in a reduced number of mismatches. Had matching for broad antigens been sufficient, the conversion procedure should not have affected the analysis. However, as shown in fig 4, the correlation of matching with graft outcome deteriorated. Transplants that were converted to well matched grafts (0 or 1 broad antigen mismatch) fared less well than 0 or 1 mismatch grafts as determined with antigen splits-ie, mismatches for antigen splits are immunogenic.
Fig 4-Comparison of HLA matching correlation obtained with antigen splits with that obtained when splits were converted to broad antigens.
64
Discussion Our data provide evidence that typing for HLA antigen splits is important in renal transplantation. In the light of this finding, previous reports on a lack of correlation between HLA matching and graft survival must be re-examined.3-7 Since existing kidney sharing policies are based on matching for broad HLA antigens, it can be assumed that the role of matching for broad antigens was analysed. Because the patient numbers were small, it is not surprising that the weak influence of broad antigen matching went undetected. In our own previous studies, HLA antigens were analysed as reported by the transplant centres, either as splits or as broad antigens. It is evident from the current results that the correlation of matching with graft outcome that we foundYnt3 was due primarily to the contribution of transplants that were matched for HLA antigen splits. This observtion is important. Some transplant programmes no longer pay attention to HLA matching3,4 and matching is rapidly losing influence at others.6,7,14 As shown herein, even immunosuppression with cyclosporin does not overrride the strong effect of HLA matching if studied correctly. A departure from HLA matching, therefore, seems entirely inappropriate. A practical consequence of our results is that the standard of HLA typing must be improved at centres where typing is at present done for broad antigens only. Waiting lists of future transplant recipients need to be updated with respect to antigen splits and computer matching programs need to be adjusted accordingly. Certainly it is easier to find a "good match" for broad antigens than for antigen splits because fewer antigen specificities must be considered. However, as shown herein, typing for antigen splits is required if the full potential of HLA matching is to be realised. To provide well matched kidneys for as many patients as possible, the need for large pools of waiting recipients will be even greater in the future than today. National and international cooperation for kidney allocation will become essential for obtaining transplants that are matched for HLA antigen splits. In the current analysis, a comparison of matching for broad antigens and antigen splits was conducted for antigens of the HLA-A and HLA-B loci. The effect of antigen splitting was potentiated by the inclusion of HLA-DR antigens in the mismatch calculation. Difficulties in defining HLA-DR antigen splits by current serological means prevented us from extending the splitting analysis to HLA-DR. Recent DNA biochemical data show that several new antigen splits can be recognised on the HLA-DR locus. IS From the results shown herein, it seems likely that typing for HLA-DR splits will further improve the correlation of HLA matching with graft survival. To clarify this complex issue, an international cooperative project to study the relevance of HLA-DR splits as determined by DNA restriction-fragment-length-polymorphism typing has been initiated within the Collaborative Transplant
Study. The contributions of doctors, nurses, coordinators, technicians, and secretaries at the 240 transplant centres participating in the Collaborative Transplant Study are gratefully acknowledged. This study was funded in part by grants from Bundesministerium fur Forschung und Technologie, Deutsche Stiftung Organtransplantation, and Deutsche Krebshilfe, Federal Republic of Germany. Computer hardware and software support was provided by IBM Scientific Center Heidelberg.
SINGLE DOSE PHENOBARBITONE PREVENTS CONVULSIONS IN CEREBRAL MALARIA S. LOOAREESUWAN N. J. WHITE R. E. PHILLIPS P. CHANTHAVANICH D.A. WARRELL
Bangkok Hospital for Tropical Diseases, Faculty of Tropical Medicine, Mahidol University, Bangkok 10400, Thailand; and Tropical Medicine Unit, Nuffield Department of Clinical Medicine, University of Oxford 48 patients over 6 years of age with strictly defined cerebral malaria were randomised to receive either a single intramuscular injection of phenobarbitone (3·5 mg/kg) or placebo in a double-blind, placebo-controlled study. Phenobarbitone significantly reduced the incidence of subsequent convulsions from 54% to 12·5%, without adverse effects. A single intramuscular injection of phenobarbitone is a simple, cheap, and effective method for prevention of convulsions in cerebral malaria.
Summary
Introduction
CONVULSIONS associated with malaria are an important of morbidity and mortality in the tropics. About half of all adult patients admitted to hospital with cerebral malaria will have a convulsion after admission.l,2 The incidence in children, in whom the seizure threshold is lower, was 78% in recent prospective studies from Malawi
cause
(Taylor T, Molyneaux ME, personal communication), and convulsions on or after admission were associated with a bad prognosis. Epileptic seizures are almost invariably associated with a deterioration in the level of consciousness and commonly lead to aspiration pneumonia, an important cause of death in cerebral malaria.3 The long-term sequelae of convulsions associated with malaria in adults are not known, but in children may include a predisposition to
C. OPELZ: REFERENCES 1. Albert ED, Baur MP, Mayr WR, eds. Histocompatibility testing 1984. Berlin Springer-Verlag, 1984. 2. Sanfilippo F, Vaughn WK, Light JA, Le For WM. Lack of influence of donor-recipient differences in subtypic HLA-A,B antigens (splits) on the outcome of cadaver renal transplantation. Transplantation 1985, 39: 151-56. 3. Harris KR, Digard N, Gosling DC, et al. Azathioprine and cyclosporin: Different tissue matching cntena needed* Lancet 1985; u: 802-05. 4. Lundgren G, Groth CG, Albrechtsen D, et al. HLA matching and pretransplant blood transfusions in cadaveric renal transplantation-a changing picture with cyclosporin. Lancet 1986; ii: 66-69. 5. Lundgren G, Albrechtsen D, Brynger H, et al. Role of HLA matching and pretransplant blood transfusions in cyclosporine-treated recipients of cadaveric renal allografts: 2- to 3-year results Transplant Proc 1987; 19: 3614-18. 6. Slapak M. Triple and quadruple unmunosuppressive therapy in organ transplantation. Lancet 1987; ii: 958-60. 7. Kerman RH, Van Buren CT, Lewis RM, Kahan BD. Successful transplantation of 100 nontransfused cyclosponne-treated primary recipients of cadaveric renal allografts Transplantation 1988; 45: 37--40. 8. Hardy MA. Prospective HLA typing is helpful in cadaveric renal transplantationMaybe ? Transplant Proc 1987; 19: 144--48. 9. Mendez R. A national allocation system. Transplant Proc 1988,20 (suppl): 1014-16. 10. Morris PJ. Are there ethical problems posed by multiple small transplant centres? Transplant Int 1988; 1 : 47-48. 11. Opelz G for the Collaborative Transplant Study. Effect of HLA matching in 10.000 cyclosponne-treated cadaver kidney transplants. Transplant Proc 1987; 19: 641--46 12. Dunn OJ, Clark VA, eds. Applied statistics Analysis of variance and regression New York. Wiley, 1974: 236. 13. Opelz G for the Collaborative Transplant Study. The benefit of exchanging donor kidneys among transplant centers. N Engly Med 1988; 318: 1289-92. 14. Salaman JR, Ross WB. Exchanging kidney transplants-is it worth it? Lancet 1987; I 1480-81. J. 15. Bidwell J. DNA-RFLP analysis and genotyping of HLA-DR and DQ antigens. 7mMtUtm/ Todav 1988: 9: 1-6.
IMPORTANCE OF HLA ANTIGEN SPLITS FOR KIDNEY TRANSPLANT MATCHING GERHARD OPELZ FOR THE COLLABORATIVE TRANSPLANT STUDY
Department of Transplantation Immunology, Institute of Immunology, University of Heidelberg, Heidelberg, Federal Republic of Germany Whether
matching for HLA antigen "splits" results in better transplant outcome than matching for "broad" HLA antigens was investigated in 30 000 first cadaver kidney transplants. At Summary
three years, there was an 18% difference between the survival rates of grafts with 0 or 4 mismatches among transplants typed for HLA-A and B antigen splits whereas the difference in transplants typed for broad antigens was only 2%. Analysis of HLA-A, B antigens together with HLA-DR antigens showed an even greater advantage of matching for antigen splits: the difference in survival at three years between grafts with 0 or 6 mismatches for HLA-A, B, DR was 31% when antigen splits were analysed, in contrast to a 6% difference with broad antigens. These results indicate that typing for HLA antigen splits is important in renal transplantation, that the potential benefit of HLA matching in renal transplantation is greater than currently accepted, and that HLA typing and kidney allocation routines must be refined in order to exploit this potential. Introduction TYPING for HLA antigens can be done for so-called broad antigen specificities or for their subunits which are commonly called antigen splits.! For example, individuals possessing the HLA antigen HLA-A9 can be characterised further as possessing either HLA-A23 or HLA-A24, two subunits or "splits" of HLA-A9. Because the serological reagents for determination of antigen splits are scarce, typing for splits is more difficult and more expensive than typing for broad antigens. The general view is that, for the
1988
purpose of cadaver kidney matching, consideration of antigen splits is not necessary. Organ exchange organisations base their allocation of kidneys on broad HLA antigen assignments, and many transplant centres are not typing for antigen splits. This strategy found support in a
report by Sanfilippo et al2who found that, in a series of 3800 patients, the correlation of HLA-A, B matching with graft survival was similar whether broad antigens or antigen splits were analysed; they advocated matching for a small number of serologically "cross-reacting" HLA antigens, arguing that, apart from reducing the cost of HLA typing, this would carry the advantage that recipients and donors could be "matched" more easily because fewer antigen specificities would have to be considered. In this report we examine a much larger series of transplants. The results are relevant to recent controversial reports on the efficacy of HLA matching in cyclosporintreated kidney transplant recipients3-7 and to the continuing discussion about equitable distribution of cadaver
organs.8-l0 Patients and Methods The data on which this analysis was based were contributed by 240 centres that are participating in the Collaborative Transplant Study.ll HLA typings were performed in the individual centres’ laboratories. Typing and baseline medical data were reported to the study centre at the University of Heidelberg. Clinical follow-up information was requested at 3, 6, 9, 12, 24, and 36 months after
transplantation. The transplants were separated into two categories according to whether typing for HLA-A, B antigen splits had or had not been done. One category consisted of transplants in which any of the following antigen splits were reported in the recipient or donor and in which broad antigens were not reported: HLA-A 23, 24, 25, 26, 29,30,31,32,33,34,66,68,69, and HLA-B 38, 39, 44, 45, 49, 50, 51,52,54,55,56,57,58,60,61,62,63,64,65,71,72. Antigens for which no splits are known1 were counted as equivalent to splits. The second category consisted of transplants where any of the following broad antigens were reported: HLA-A 9,10,19,28, and HLA-B 5, 12,14,15,16,17,21,22,40,70. When a recipient was reported with broad antigens and a donor with antigen splits, or vice versa, the 8602
©
62
antigen splits were converted by the computer to the appropriate broad antigens and the transplant was counted in the broad antigen category. For transplants with HLA-DR typing information, the antigens HLA-DR 1-10 were analysed. Splits of HLA-DR antigens were not considered because typing for splits was not done in
most
laboratories that submitted data.
The donors
analysis was restricted to first transplants from cadaver performed from 1982 to 1987. Graft survival rates were computed by actuarial methods and statistical significance was estimated by log-rank or weighted regression analysis12 as indicated. In the latter, the dependent variable was the percentage survival at three years, the independent variable was the number of mismatched HLA antigens, and the weighting factor was the number of patients studied. We estimated the standard error of the regression coefficient from the standard error of the percentage survival, and statistical significance was judged on the basis of the ratio of the regression coefficient to its standard error. No exclusions were
made.
Fig 1-Influence
ofHLA-A and HLA-B mismatches in
patients studied are indicated for each curve. The horizontal line at 70% was
Fig 2-Analysis of HLA-A, HLA-B, and HLA-DR mismatches to
Matching for HLA-A and HLA-B antigens correlated with graft survival in transplants that were typed for antigen splits but not in transplants typed for broad antigens (fig 1). When antigen splits were analysed, grafts with 0 mismatches had an 18% higher success rate at three years than grafts The analysis of with 4 mismatches (p log rank <0’0001). broad antigens showed a difference of only 2% (NS). The three-year success rates of transplants with 0, 1, or 2 HLA-DR mismatches were: 65% (SE 0-7),62% (SE 0-6), and 58% (SE 0-9), respectively. When HLA-DR mismatches were added to the number of mismatches for HLA-A and HLA-B, a striking potentiating effect on the correlation with graft survival was observed among transplants typed for HLA-A, B antigen splits. The three-year survival rate of transplants with 0 mismatches for
transplants typed for antigen splits or broad antigens.
The number of mismatched HLA antigens and the number of drawn to allow for better comparison.
Antigen splits and broad antigens refer
Results
in first cadaver
HLA-A and HLA-B locus
antigens.
kidney transplants.
63
Fig 3-Effect of mismatching for HLA antigen splits and broad antigens in cyclosporin-treated patients. HLA A, B, DR was 80% (SE 2-6) (SE 2-7) rate with 6 mismatches (p
compared with a 49% log rank <0’0001). In contrast, among transplants typed for broad HLA-A, B antigens, the effect of matching on graft outcome was much less pronounced (fig 2). Grafts with 0 mismatches had only a 6% higher survival rate at three years than grafts with 6 mismatches (NS). Fig 3 shows that the results described above are applicable to patients who were immunosuppressed with cyclosporin. Among transplants typed for HLA-A, B antigen splits, the correlation of matching with outcome was excellent (p regression < 00001) whereas in transplants typed for broad antigens the results were unimpressive. Two factors are likely to have contributed to the observed difference in results. It can be assumed that laboratories that typed for antigen splits possessed greater expertise in HLA serology than those typing for broad antigens.
Consequently, erroneous HLA typing results may have been contained in the data reported by the latter. Of particular interest from a scientific viewpoint is whether a mismatch between recipient and donor that is evident when splits but not when broad antigens are considered leads to an increased rejection rate. To investigate this, we programmed the computer to convert the antigen splits (reported by "expert" laboratories) into broad antigens, and we reanalysed those transplants in which the conversion resulted in a reduced number of mismatches. Had matching for broad antigens been sufficient, the conversion procedure should not have affected the analysis. However, as shown in fig 4, the correlation of matching with graft outcome deteriorated. Transplants that were converted to well matched grafts (0 or 1 broad antigen mismatch) fared less well than 0 or 1 mismatch grafts as determined with antigen splits-ie, mismatches for antigen splits are immunogenic.
Fig 4-Comparison of HLA matching correlation obtained with antigen splits with that obtained when splits were converted to broad antigens.
64
Discussion Our data provide evidence that typing for HLA antigen splits is important in renal transplantation. In the light of this finding, previous reports on a lack of correlation between HLA matching and graft survival must be re-examined.3-7 Since existing kidney sharing policies are based on matching for broad HLA antigens, it can be assumed that the role of matching for broad antigens was analysed. Because the patient numbers were small, it is not surprising that the weak influence of broad antigen matching went undetected. In our own previous studies, HLA antigens were analysed as reported by the transplant centres, either as splits or as broad antigens. It is evident from the current results that the correlation of matching with graft outcome that we foundYnt3 was due primarily to the contribution of transplants that were matched for HLA antigen splits. This observtion is important. Some transplant programmes no longer pay attention to HLA matching3,4 and matching is rapidly losing influence at others.6,7,14 As shown herein, even immunosuppression with cyclosporin does not overrride the strong effect of HLA matching if studied correctly. A departure from HLA matching, therefore, seems entirely inappropriate. A practical consequence of our results is that the standard of HLA typing must be improved at centres where typing is at present done for broad antigens only. Waiting lists of future transplant recipients need to be updated with respect to antigen splits and computer matching programs need to be adjusted accordingly. Certainly it is easier to find a "good match" for broad antigens than for antigen splits because fewer antigen specificities must be considered. However, as shown herein, typing for antigen splits is required if the full potential of HLA matching is to be realised. To provide well matched kidneys for as many patients as possible, the need for large pools of waiting recipients will be even greater in the future than today. National and international cooperation for kidney allocation will become essential for obtaining transplants that are matched for HLA antigen splits. In the current analysis, a comparison of matching for broad antigens and antigen splits was conducted for antigens of the HLA-A and HLA-B loci. The effect of antigen splitting was potentiated by the inclusion of HLA-DR antigens in the mismatch calculation. Difficulties in defining HLA-DR antigen splits by current serological means prevented us from extending the splitting analysis to HLA-DR. Recent DNA biochemical data show that several new antigen splits can be recognised on the HLA-DR locus. IS From the results shown herein, it seems likely that typing for HLA-DR splits will further improve the correlation of HLA matching with graft survival. To clarify this complex issue, an international cooperative project to study the relevance of HLA-DR splits as determined by DNA restriction-fragment-length-polymorphism typing has been initiated within the Collaborative Transplant
Study. The contributions of doctors, nurses, coordinators, technicians, and secretaries at the 240 transplant centres participating in the Collaborative Transplant Study are gratefully acknowledged. This study was funded in part by grants from Bundesministerium fur Forschung und Technologie, Deutsche Stiftung Organtransplantation, and Deutsche Krebshilfe, Federal Republic of Germany. Computer hardware and software support was provided by IBM Scientific Center Heidelberg.
SINGLE DOSE PHENOBARBITONE PREVENTS CONVULSIONS IN CEREBRAL MALARIA S. LOOAREESUWAN N. J. WHITE R. E. PHILLIPS P. CHANTHAVANICH D.A. WARRELL
Bangkok Hospital for Tropical Diseases, Faculty of Tropical Medicine, Mahidol University, Bangkok 10400, Thailand; and Tropical Medicine Unit, Nuffield Department of Clinical Medicine, University of Oxford 48 patients over 6 years of age with strictly defined cerebral malaria were randomised to receive either a single intramuscular injection of phenobarbitone (3·5 mg/kg) or placebo in a double-blind, placebo-controlled study. Phenobarbitone significantly reduced the incidence of subsequent convulsions from 54% to 12·5%, without adverse effects. A single intramuscular injection of phenobarbitone is a simple, cheap, and effective method for prevention of convulsions in cerebral malaria.
Summary
Introduction
CONVULSIONS associated with malaria are an important of morbidity and mortality in the tropics. About half of all adult patients admitted to hospital with cerebral malaria will have a convulsion after admission.l,2 The incidence in children, in whom the seizure threshold is lower, was 78% in recent prospective studies from Malawi
cause
(Taylor T, Molyneaux ME, personal communication), and convulsions on or after admission were associated with a bad prognosis. Epileptic seizures are almost invariably associated with a deterioration in the level of consciousness and commonly lead to aspiration pneumonia, an important cause of death in cerebral malaria.3 The long-term sequelae of convulsions associated with malaria in adults are not known, but in children may include a predisposition to
C. OPELZ: REFERENCES 1. Albert ED, Baur MP, Mayr WR, eds. Histocompatibility testing 1984. Berlin Springer-Verlag, 1984. 2. Sanfilippo F, Vaughn WK, Light JA, Le For WM. Lack of influence of donor-recipient differences in subtypic HLA-A,B antigens (splits) on the outcome of cadaver renal transplantation. Transplantation 1985, 39: 151-56. 3. Harris KR, Digard N, Gosling DC, et al. Azathioprine and cyclosporin: Different tissue matching cntena needed* Lancet 1985; u: 802-05. 4. Lundgren G, Groth CG, Albrechtsen D, et al. HLA matching and pretransplant blood transfusions in cadaveric renal transplantation-a changing picture with cyclosporin. Lancet 1986; ii: 66-69. 5. Lundgren G, Albrechtsen D, Brynger H, et al. Role of HLA matching and pretransplant blood transfusions in cyclosporine-treated recipients of cadaveric renal allografts: 2- to 3-year results Transplant Proc 1987; 19: 3614-18. 6. Slapak M. Triple and quadruple unmunosuppressive therapy in organ transplantation. Lancet 1987; ii: 958-60. 7. Kerman RH, Van Buren CT, Lewis RM, Kahan BD. Successful transplantation of 100 nontransfused cyclosponne-treated primary recipients of cadaveric renal allografts Transplantation 1988; 45: 37--40. 8. Hardy MA. Prospective HLA typing is helpful in cadaveric renal transplantationMaybe ? Transplant Proc 1987; 19: 144--48. 9. Mendez R. A national allocation system. Transplant Proc 1988,20 (suppl): 1014-16. 10. Morris PJ. Are there ethical problems posed by multiple small transplant centres? Transplant Int 1988; 1 : 47-48. 11. Opelz G for the Collaborative Transplant Study. Effect of HLA matching in 10.000 cyclosponne-treated cadaver kidney transplants. Transplant Proc 1987; 19: 641--46 12. Dunn OJ, Clark VA, eds. Applied statistics Analysis of variance and regression New York. Wiley, 1974: 236. 13. Opelz G for the Collaborative Transplant Study. The benefit of exchanging donor kidneys among transplant centers. N Engly Med 1988; 318: 1289-92. 14. Salaman JR, Ross WB. Exchanging kidney transplants-is it worth it? Lancet 1987; I 1480-81. J. 15. Bidwell J. DNA-RFLP analysis and genotyping of HLA-DR and DQ antigens. 7mMtUtm/ Todav 1988: 9: 1-6.