A case-control study of stillbirths at a teaching hospital in Zambia, 1979-80: serological investigations for selected infectious agents

Bulletin ofthe World Health Organization, 62 (5): 803 - 808 (1984) i World Health Organization 1984 A case-control study of stillbirths at a teaching hospital in Zambia, 1979-80: serological investigations for selected infectious agents THERESA E. WATTS,' SANDRA A. LARSEN,2 & STUART T. BROWN3 Sera were obtained from 266 mothers of singleton stillborn babies (cases) and 266 mothers of live-born babies (controls), matchedfor parity, who delivered at the University Teaching Hospital, Lusaka, Zambia, between October 1979 andApril 1980. Tests wereper- formed on 262 samples from cases and 261 from controls. The microhaemagglutination assay for Treponema pallidum (MHA-TP) was reactive in 54%o of cases and 29% of con- trols; the rapid plasma reagin (RPR) 18-mm circle card test was reactive at a dilution of 1:16 or greater in 29% of cases and in 3.5% of controls. Both these differences are highly significant. Serafrom cases and controls were further examinedfor evidence of cytomegalovirus, human (alpha) herpesvirus, hepatitis B virus, toxoplasma, andplasmodium infections. The only difference between sera from cases and controls was that cytomegalovirus antibody titres > 1:1024 occurred more often among cases. There was no relationship between antibody titre and birth weight. The results of this study emphasize the importance of screening pregnant women for syphilis. Treatment of those found to be infected should help prevent stillbirths due to syphilis. The University Teaching Hospital is the only hospi- tal delivery facility for the population of over 500 000 people in Lusaka, Zambia. Three health centres in Lusaka also perform a few uncomplicated deliveries each year. More than 20 000 women deliver at the teaching hospital each year, accounting for more than 80Wo of all deliveries in Lusaka. Staff in the Depart- ment of Obstetrics regularly review all stillbirths occurring in the hospital, but the cause of death can- not be identified for nearly 50Oo of them. This study was undertaken to try to determine the factors associ- ated with stillbirths in an effort to identify targets for prevention programmes. The study methods and the findings relating to antenatal care have been reported previously (1). The present paper describes the results of serological tests for infectious agents. PATIENTS AND METHODS All mothers of singleton stillbirths at the University Teaching Hospital between October 1979 and May Senior Lecturer and Head, Department of Community Health, University Teaching Hospital, P.O. Box 50110, Lusaka, Zambia. 2 Chief, Treponema Research Branch, STD Laboratory Program Office, Center for Infectious Diseases, Centers for Disease Control, Atlanta, GA, USA. 3 Coordinator, International VD Control Activities, Venereal Disease Control Division, Center for Prevention Services, Centers for Disease Control, Atlanta, GA, USA. 1980 were identified. Each mother was paired with the next mother of the same parity to deliver a live-born child. A specimen of blood was withdrawn from each of these mothers into a Vacutainer. After clotting, the blood samples were centrifuged and the sera separ- ated and stored at - 20 °C, on the same day. These sera were transported frozen to the Centers for Disease Control (CDC), Atlanta, GA, USA, for test- ing. Serological tests for syphilis Qualitative and quantitative rapid plasma reagin (RPR) card tests" were performed with 18-mm circle cards using standard procedures (2). The micro- haemagglutination assay for antibody to Treponema palliduitn (MHA-TP)" was performed according to the manufacturer's directions (3). Initially, 100 sera were tested both by the RPR and MHA-TP tests. Since all sera that were reactive in the RPR card test were also MHA-TP reactive, the remaining sera were screened with the MHA-TP, and only those giving a positive reaction were tested with the RPR card test. Only sera that were reactive in the qualitative RPR card tests were tested quantitatively. " Hynson, Wescott and Dunning, Ba timore, MD, USA. Trade names and commercial sources are used for product identification only, and do not represent an endorsemer t by the US Department of Health and Human Services or the Public Health Service. " Sera Tek, Ames Co., Elkhart, IN, USA. 4460 -803- T. E. WATTS ET AL. In addition, routine serological tests for syphilis were requested by the hospital staff for some of these women. The hospital laboratory staff tested these routine specimens using the Venereal Disease Research Laboratory (VDRL) slide test antigen. In this procedure, the test results were read without the aid of a microscope. Other serological tests A sample of 50 paired sera was examined for anti- bodies to cytomegalovirus and human (alpha) herpes- virus (herpes simplex virus) using indirect haemag- glutination assay (IHA) procedures, as described pre- viously (4, 5). Hepatitis B surface antigen (HBsAg) and antibodies to hepatitis B core antigen (anti-HBc) were sought in another sample of 50 paired sera, using previously described radioimmunoassay procedures with commercially available reagents (6). A quanti- tative indirect haemagglutination assay employing Plasmodiumfalciparuin antigen was used to identify antibodies against malaria in a sample of 100 sera selected from cases and controls (7). IHA titres of less than 1: 16 were considered non-reactive. Antibody to Toxoplasma gondii was sought using a quantitative immunofluorescent assay (IFA)" in 426 sera (8). Since no significant differences in antibody dis- tribution between cases and controls were observed for any of these tests, no additional sera were tested. RESU LTS Sera were collected from 266 cases (mothers of still- born babies) and 266 controls. Unfortunately some serum containers broke, and in some cases the quantity of serum was not sufficient to perform all the tests; thus sera were available for testing for a total of 523 cases and controls. Syphilis Of the 262 sera from mothers of stillborn babies, 102 (39)o) were reactive in the RPR card test, com- ' International Diagnostic Technology, Santa Clara, CA, USA. pared with 39 (1507o) of 261 control sera (Table 1). This difference is highly significant (X2 = 38.3, P < 0.01). RPR card test titres were also significantly higher among the cases than among the controls (X2 = 23.6; P < 0.002) (Fig. 1). Of the 79 women with an RPR card test titre > 1: 32, only 6 had live-born babies. Of these, 3 had received treatment for syphilis during pregnancy and the others had low-birth-weight babies. Overall, RPR card titres > 1: 16 were related to the parity of the women: of 120 primiparae, 18%7o had RPR titres > 1: 16 compared with 1 507o of mothers of parity 3 or 4, and 10.60%o of mothers of parity 5 or greater (dif- ference between proportions = 2.78; P < 0.05). A total of 141 sera from 262 cases (54%7o) were reactive in the MHA-TP test, compared with 76 (2901o) of the 261 controls (X2 = 35; P < 0.01); 36 sera from the cases (13.7%) and 34 sera from the controls (13%o) had a reactive MHA-TP and non-reactive RPR. Results of the routine VDRL slide test carried out in the hospital laboratory were available for comparison with the MHA-TP and RPR card test results for 369 sera (Table 2). Only 73 (630%o) of the 115 sera reactive to both MHA-TP and RPR card tests were reactive in the modified VDRL slide test performed in the University Teaching Hospital laboratory. However, 67 (88%o) of the 76 sera with RPR titres of 1: 16 or greater had reactive VDRL 30 . M 25 . 0 20 - O 15 . 10 5. ----- --.- -_-_ CONROLS 1:16 1:16 1:32 1:64 1:128 1:256 TITRE TO llPR Fig. 1. Titres of reactive controls. 102 RPR-reactive cases and 39 RPR- Table 1. Results of quantitative RPR card test for mothers of stillborn babies and controls, Lusaka, 1979-80 RPR card test reciprocal titre No. No. tested non-reactive 1 2 4 8 16 32 > 32 Cases 262 160 9 5 6 7 2 9 64 Controls 261 222 8 10 5 7 3 3 3 Total 523 382 17 15 11 14 5 12 67 804 INFECTIOUS AGENTS ASSOCIATED WITH STILLBIRTFHS IN ZAMBIA Table 2. Comparison of results of the modified VDRL slide test and the MHA-TP and RPR card tests for 369 sera from cases and controls, Lusaka, 1979-80 MHA-TP reactive Total Modified VDRL RPR card RPR card MHA-TP slide test reactive non-reactive non-reactive No. % Reactive 73 (67)" 1 10 84 22.8 Non-reactive 42 (9)" 45 198 285 77.2 Total 115 46 208 369 " Figure in parentheses gives number with RPR titre > 1: 16. Table 3. Indirect haemagglutination titres of antibody against cytomegalovirus in 50 women with stillbirths and 50 women with live births, Lusaka, 1979-80 Cytomegalovirus IHA reciprocal titre < 32 32 64 128 256 512 1024 > 2048 Cases 8 4 1 7 9 6 6 9 Controls 6 5 7 8 8 11 2 3 Total 14 9 8 15 17 17 8 12 Table 4. Indirect haemagglutination titres of antibody against human (alpha) herpesvirus among 50 women with stillbirths and 50 with live births, Lusaka, 1979-80 Human (alpha) herpesvirus IHA reciprocal titre <32 32 64 128 256 512 1024 2048 >4096 Cases 6 0 5 1 5 6 10 12 5 Controls 3 0 5 4 7 8 14 6 3 Total 9 0 10 5 12 14 24 18 8 tests. Of 208 sera that were non-reactive in the MHA- TP test, 198 (95/o) were also non-reactive in the VDRL test carried out at the hospital. We did not identify any relationship between the results of the serological tests for syphilis and the results of serological tests for other infectious diseases. Cytomegalovirus Sera from 50 cases and 50 controls were examined for evidence of cytomegalovirus infection (Table 3). Of these, 86% had cytomegalovirus titres > 1: 32. There was no significant difference in the distribution of titres between cases and controls. However, titres > 1:1024 were significantly more common in the cases than in the controls (X2 = 5.01; P < 0.05). The birth weight of the babies was not related to the serum titres. Human (alpha) herspesvirus No difference was found in IHA titres of antibodies to human (alpha) herpesvirus between the cases and controls (Table 4); 91 O of the population had titres > 1:32. Hepatitis B Serological tests for anti-HBc and HBsAg were done on sera from 50 cases and 50 controls (Table 5). No difference was found between these groups for evidence of hepatitis B infection; anti-HBc was detected in 48% of sera from cases and controls. 805 8r. E. WATTS ET AL. Table 5. Results of serological tests for anti-HBc and HBsAg among 50 women with stillbirths and 50 with live births, Lusaka, 1979-80 Anti-HBc-positive Anti-HBc- negative HBsAg-positive HBsAg-negative Cases 24" 3 23 Controls 28" 5 17 Total 52 8 40 ' A low titre of HBsAg was detected in one case and one con- trol, but could not be verified by repeat testing. Table 7. Immunofluorescent antibody titres against Toxoplasma gondii among 213 women with a stillbirth and 213 with a live birth, Lusaka, 1979-80 IFA toxoplasma reciprocal titre Non-reactive 16 32 64 228 > 256 Cases 181 8 16 4 2 2 Control 179 1 1 10 9 3 1 Total 360 19 26 13 5 3 DISCUSSION Table 6. Indirect haemagglutination titres of antibody against malaria among 51 women with a stillbirth and 49 women with a live birth, Lusaka, 1979-80 IHA reciprocal titre Non-reactive 16-256 > 512 Cases 24 15 12 Controls 24 18 7 Total 48 33 19 Malaria Sera from 51 cases and 49 controls were examined for antibodies against P.falciparum by the IHA test (Table 6); 52070 of these mothers had sera that were reactive in this test. No differences in test reactivity were found between cases and controls. We found no difference in birth weights between mothers with non-reactive tests and mothers with IHA-malaria titres > 1:512. In each group the birth weights were lower in the stillborn babies compared with the live-born. In these sera, the titres of the 77 samples taken in November and December were higher than those of the 23 samples taken in April, but the numbers tested were too few to identify small dif- ferences with any confidence. Toxoplasma Sera from 213 cases and 213 controls were examined for antibodies to toxoplasma. Only 15.3%o of mothers had reactive IFA tests; no differences were identified between cases and controls (Table 7). On the assumption that titres > 1:256 indicate recent infection, then only 3 women (0.7%o) had become infected during pregnancy. The high prevalence of sera reactive in the MHA- TP test in this series is alarming. This test is highly specific (98%o or higher) for syphilis in the USA (3) but there has been no study to confirm that it is equally specific in the Zambian population. Many women in these study groups had a reactive MHA-TP and a non-reactive RPR card test, with no difference between cases and controls. These serological find- ings may reflect old treponemal infection or an un- identified non-treponemal factor that causes re- activity; however, Zambia has not been recognized as a focus of yaws or endemic syphilis. This study failed to find a relationship between the MHA-TP reaction and serological evidence of other infections. Un- explained reactive MHA-TP and non-reactive RPR card tests were also found in a large proportion (24%) of a "normal" population in Swaziland (9); similar results were found in a large proportion (13%) of the controls in the present study. The relationship observed between stillbirths and high-titre RPR card test seroreactivity was the strongest association observed (a titre > 1: 32 carried a relative risk of 21). There was no significant relationship between lower titre RPR card tests or MHA-TP reactivity and stillbirths. A high titre in a non-treponemal serological test, particularly one of 1: 32 or more for the RPR card test, usually indicates early active disease (10). Thus, the study findings are consistent with the hypothesis that untreated early syphilis in pregnancy is an important cause of still- births in Zambia, as was the case in the era preceding penicillin therapy (11). Although no woman in this series had clinical symptoms of syphilis and only 28 had a positive VDRL test recorded before delivery (1), there is considerable evidence of syphilis in infants examined in the University Teaching Hospital; a prevalence of 6.4% was found in unselected consecutive deliveries, and a rate of 8.6% in infants under 3 months old admitted to the hospital (12-14). 806 INFECTIOUS AGENTS ASSOCIATFED WI [H S[ILLBIRTFHS IN ZAMBIA The diagnosis of syphilis presents a problem in many developing countries. The test usually employed is the VDRL slide test, modified to suit the local conditions. Although the VDRL slide test per- formed in the hospital laboratory identified nearly 9007o of sera with an RPR card test titre of more than 1:16, many patients with lower titres were missed. Low sensitivity may be expected when the agglutination reaction on the slide is detected without the use of a microscope. Reactivity in the VDRL slide test, as performed in the hospital laboratory, occurred in nearly 507o of sera found to be non- reactive in both the MHA-TP and RPR card tests. Other workers have reported higher test specificity (98-99%) when careful quality control measures are followed (3, 12). In addition to the technical problems of serological testing, pregnant women in Lusaka may not be screened for syphilis because laboratory facilities are not available at all antenatal clinics and containers for blood or laboratory requests may be unavailable; also, many mothers attend the clinic infrequently or late. In this series of 532 women, only 38% of cases and 56% of controls had had a VDRL test before delivery. Of the 18 cases found positive, 6 received treatment, compared with 8 of 10 controls found posi- tive (1). With the present uncertainty regarding the significance of low-titre reactions, it would be advisable to retest women with low reactions and give treatment if the titre in the second test is the same or higher. After delivery, follow-up of the women becomes increasingly difficult. Multiple partners are common in the urban areas of Zambia, and the higher proportion of strongly reactive women among primiparae suggests that syphilis is contracted early during the child-bearing period. Age and parity were closely related in this sample of 532 women. The prevalence of toxoplasma infection during pregnancy was lower in this series than the 6.4 per 1000 found in the USA and the 3.0 per 1000 in France (16). Abortions and perinatal deaths resulting from congenital toxoplasma infection have been reported, but there was no difference between cases and con- trols in this study. It has been suggested that coincident infection with toxoplasma and cyto- megalovirus may increase the risk of stillbirth (17), but this could not be demonstrated in the 50 paired sera examined in this study. In one study in the United Republic of Tanzania (18), cytomegalovirus antibodies were found in all pregnant mothers; the proportion was nearly as high in this study. Only for the high titres ( 1: 1024) was there a difference between cases and controls, which may indicate that cytomegalovirus infection during pregnancy could cause death of the fetus. Cyto- megalovirus infection was not related to birth weight in this study, although others have made this associ- ation (19, 20). The present study did not seek to evaluate mental retardation or other more common consequences of perinatal cytomegalovirus infection. Significant levels of malaria antibodies were found in 520%o of mothers, despite the malaria control pro- gramme in the Lusaka urban area; this may be because many mothers were from the new urban townships around the periphery of Lusaka where con- trol is less effective. No relationship between malarial antibody titres and birth weight was found in this study. Workers in Gambia found an inverse relation- ship in late pregnancy between antibody titre and parasite density (21). Although more high titres were observed in November and December than in April, this may only reflect the fact that, with the unstable malaria that occurs in Zambia, the mother has had a shorter period of exposure to malaria transmission, which increases around October. Thus it does not appear that an indirect fluorescent antibody titre of I :256 or greater is a good indicator of malaria infection in the last 6 months of pregnancy, as was found in a general population (22). Nearly half the women in this study had evidence of hepatitis B infection, but only 807o were carriers. There was no relationship between stillbirths and infection with hepatitis B. Syphilis was the most important infection associ- ated with stillbirths in this study. It is an eminently preventable condition. A large proportion of these mothers could be identified if a simple screening test was available that could be performed by staff in the antenatal clinic. The importance of obtaining a sero- logical test for syphilis in pregnancy and providing immediate treatment is emphasized. The RPR card test may be suitable for introduction into antenatal clinics as it can be read during the clinic, so that there would be no delay in instituting treatment. ACKNOWLEDGEMENTS This study was made possible by a grant from the World Health Organization. We thank the mothers and staff in the Department of Obstetrics who participated in this study, Dr Kagan who kindly transported the sera to Atlanta, the staff of several laboratories of the Centers for Disease Control, Atlanta, who performed the serological tests, and Ms Kabulansando for typing the manuscript. 807 808 T. E. WATTS ET AL. RtSUMt ETUDE CAS-TEMOIN RELATIVE A DES MORTINAISSANCES OBSERVEES DANS UN H6PITAL UNIVERSITAIRE DE ZAMBIE EN 1979-80: EXAMENS SEROLOGIQUES A LA RECHERCHE D'UN CERTAIN NOMBRE D'AGENTS INFECTIEUX. Des 6chantillons de serum ont ete pr6lev6s sur 266 meres ayant mis au monde un unique enfant mort-ne (qui cons- tituent les cas) et sur 266 autres ayant accouche d'enfants vivants (qui constituent les t6moins), appari6es quant a la parite et qui ont accouch6 a l'h6pital universitaire de Lusaka, Zambie, entre octobre 1979 et avril 1980. Les examens s6rologiques ont port6 sur 262 echantillons provenant des cas 6tudi6s et sur 261 provenant des t6moins. L'6preuve de microh6magglutination a la recherche de Treponema pallidum (MHA-TP) s'est revel&e positive chez 54% des cas et chez 29% des temoins; 1'6preuve rapide des r6agines plasmatiques (RPR) pratiqu6e sur carte (cercle de 18 mm) a donne lieu a une r6action a la dilution de 1/16 et au-dela chez 29% des cas et 3,5% des t6moins. Dans les deux cas ces differences sont tout a fait significatives. Les s6rums pr6leves sur les cas et les temoins ont 6ga- lement fait l'objet d'examens & la recherche du cytomegalo- virus (CMV), de l'herpesvirus humain alpha, du virus de l'hepatite B et des toxoplasmes et plasmodies. La seule diff6rence relev6e entre les cas et les temoins tenait au fait que chez les cas, les titres d'anticorps anti-CMV etaient plus fr6quemment superieurs ou 6gaux a 1/24. Ces anticorps 6taient pr6sents dans 86% des prelvements alors que les anticorps dirig6s contre l'herpesvirus humain alpha l'etaient dans 91% des serums. On a decel6 la pr6sence d'anticorps dirig6s contre l'antigene central du virus HB dans 48% des s6rums; l'antigene de surface du virus HB etait identifie dans 8% des prelevements. Dans 15%W des s6rums on a cons- tat6 la pr6sence d'anticorps anti-toxoplasmes. Les titres d'anticorps antiplasmodies mesur6s en immunofluorescence etaient superieurs ou egaux a 1/16 dans 50% des s6rums. Aucune corr6lation entre les titres d'anticorps et le poids de naissance n'a ete constatee. Les r6sultats de cette etude soulignent l'importance d'un depistage de la syphilis chez les femmes enceintes. En fait, en traitant les patientes syphilitiques, on pourrait r6duire la mortinatalite due a cette maladie. REFERENCES I. WATTS, T. E. & HARRIS, R. R. A case-control study of stillbirths at a teaching hospital in Zambia, 1979-80: antenatal factors. Bulletin of the World Health Organ- ization, 60: 971-979 (1982). 2. US PUBLIC HEALTH SERVICE. Manual of tests for syphilis, 1969. Washington, DC, Department of Health, Education, and Welfare, 1969 (Public Health Service Publication No. 411). 3. JAFFE, H. W. ET AL. Haemagglutination tests for syphi- litic antibodies. American journal ofclinicalpathology, 70: 230-233 (1978). 4. WARNER, J. L. ET AL. Cytomegalovirus. In: Rose, N. R. & Friedman, H., ed., Manual of clinical immu- nology. Washington, DC, American Society of Micro- biology, 1980, pp. 622-627. 5. STEWART, J. A. & HERMANN, K. L. Herpes simplex virus. In: Rose, N. R. & Friedman, H., ed., Manual of clinical immunology, Washington, DC, American Society of Microbiology, 1980, pp. 614-619. 6. FRANCIS, D. P. ET AL. The prevention of hepatitis B with vaccine. Report of the Centers for Disease Control multi-center efficacy trial among homosexual men. Annals of internal medicine, 97: 362-366 (1982). 7. MATTHEWS, H. H. ET Al.. The indirect hemagglutina- tion test for malaria. Evaluation of antigens prepared from Plasmodium falciparum and Plasmodium vivax. A merican journal of tropical medicine and hygiene, 24: 417-422 (1975). 8. WAI.s, K. W. & BARNHARDT, E. R. Titration of human serum antibodies to Toxoplasma gondii with a simple fluorometric assay. Journal of clinical microbiology, 7: 234-235 (1978). 9. URSI, J. P. iT Al.. Syphilis in Swaziland. A serological survey. British journal of venereal diseases, 57: 95-99 (1981). 10. LUGER, A. Diagnosis of syphilis. Bulletin of the World Health Organization, 59: 647-654 (1982). 11. INGRAHAM, N. R. The value of penicillin alone in the prevention and treatment of congenital syphilis. Acta dermato-venereologica, 31 (Suppl. 24): 60-88 (1951). 12. HIRA, S. K. ET AL. Congenital syphilis in Lusaka- 1. Incidence in a general nursery ward. East African medical journal, 59: 241-246 (1982). 13. BHAT, G. J. ET AL. Congenital syphilis in Lusaka-2. Incidence in the neonatal intensive-care unit. East African medical journal, 59: 306-310 (1982). 14. RATNAM, A. V. ET AL. Syphilis in pregnant women in Zambia. British journal of venereal diseases, 58: 355-358 (1982). 15. OMER EL FADIL, E. ET AL. Evaluation of VDRL test in Sudanese blood donors. Tropical doctor, 12: 61-62 (1982). 16. CHARLES, D. & FINLAND, M. Obstetric and perinatal infections. Philadelphia, Lea and Febiger, 1973. 17. DEMIAN, S. D. E. ET AL. Coexistent congenital cyto- megalovirus and toxoplasmosis in a stillborn. American journal of diseases of children, 125: 420-421 (1973). 18. Intra-uterine infections. Amsterdam, Associated Scientific Publishers, 1973 (CIBA Foundation Sympo- sium, No. 10). 19. WATERSON, A. P. Virus infections (other than rubella) during pregnancy. British medical journal, 2: 564-566 (1979). 20. DUDGEON, J. A. Infective causes of human malforma- tions. British medical bulletin, 32: 77-83 (1976). 21. BRAY, R. S. & ANDERSON, M. J. Falciparum malaria and pregnancy. Transactions of the Royal Society of Tropical Medicine and Hygiene, 73: 427-431 (1979). 22. DRAPER, C. C. & SIRR, S. S. Serological investigations in retrospective diagnosis of malaria. British medical journal, 280: 1575-1576 (1980).