Epidemiology and control of gonococcal ophthalmia neonatorum.

UpdatelLe point Epidemiology and control of gonococcal ophthalmia neonatorum* M. Laga,' A. Meheus,2 & P. Piot' From a public health point of view gonococcal ophthalmia neonatorum (GCON) is important as it can rapidly lead to blindness. The frequency ofGCON is determined by the prevalence ofmaternal gonococcal infection. In most industrialized countries the prevalence of gonorrhoea in pregnant women is less than 1 %; in developing countries the rates are between 3% and 15%, more than 50% being due to penicillinase- producing Neisseria gonorrhoeae strains (PPNG). The rate of transmission from mother to newborn is between 30% and 50%. Strategies for the control of GCON include: (1) prevention of gonococcal infection in women of childbearing age, (2) detection and treatment ofgonococcal infection in pregnant women, (3) eye prophylaxis in the newborn at birth, and (4) diagnosis and treatment of GCON. Eye prophylaxis by the instillation immediately after birth of either 1% silver nitrate eye drops or 1% tetracycline eye ointment is very effective. This reduces the GCON incidence by 80% to 95% and is highly cost-effective, particularly in high-risk settings. Epidemiology Disease manifestation Ophthalmia neonatorum (ON) is defined as a purulent conjunctivitis in infants less than 30 days old, where the Gram stain of an eye smear shows at least one polymorphonuclear leukocyte per high-power field. Of the many causes, the two main ones are Neisseria gonorrhoeae and Chlamydia trachomatis. Although there are different patterns of disease, the manifesta- tions produced by any one agent are not sufficiently distinctive to lead to an etiologic diagnosis from the clinical signs alone. Gonococcal ophthalmia neonatorum (GCON) tends to appear earlier and to be more severe than the chlamydial infection. Differences in the incidence of ON, by etiology, obtained in Nairobi from a cross- sectional clinic-based study and a population-based cohort study are shown in Table 1; C. trachomatis was the most frequently identified cause when 1019 mother- infant pairs were followed up for 1 month, while N. r A resume of this article in French appears on pages 476-477. Department of Microbiology, Institute of Tropical Medicine, Antwerp, Belgium. ' Programme of Sexually Transmitted Diseases, World Health Organization, 1211 Geneva 27, Switzerland. Requests for reprints should be sent to this author. Reprint No. 5001 gonorrhoeae accounted for the majority of cases seen at the sexually transmitted diseases (STD) clinic. GCON begins between 1 and 13 days from birth and is mostly bilateral and purulent, the conjunctivae and eyelids being oedemic and hyperaemic. If untreated, diffuse epithelial oedema of the cornea gives it a hazy greyish appearance. Coarse white opacities (infiltra- tions) appear near the border of the cornea and the sclera, and can enlarge and become ulcerated by the end of the second or third week. These ulcerations can lead to perforation of the eyeball with loss of vision. When new blood vessels invade the cornea, corneal scarring may occur. The initiation of effective treat- ment dramatically changes the course and outcome of the disease, usually with recognizable improvement within 24 hours. It is difficult to estimate the risk of blindness associated with gonococcal ophthalmia. In 1880, over a hundred years ago, in Stuttgart the incidence of GCON varied from 1% to 14% and 20-79% of children in institutions for the blind had a history of GCON. After introduction of ocular prophylaxis at birth, GCON dropped dramatically and the per- centage of institutionalized children whose blindness was caused by GCON steadily declined. At present, in many developing countries the incidence ofGCON is still high although blindness in children is not reported to be highly prevalent; however, since blind children Bulletin of the World Health Organization, 67 (5): 471-478 (1989) World Health Organization 1989 471 M. Laga et al. Table 1: Comparison of the Incidence and etiology of ophthalmia neonatorum (ON) In Nairobi from a cohort study and clinic-based survey' Cohort studyb Incidence per 100 live births Percentage of all ON (n= 181) Cross-sectional study:c percentage of all ON (n= 149) Gonococcal ophthalmia 2.8 12% 43% Chlamydial ophthalmia 7.3 32% 13% Gonococcal and chlamydial ophthalmia 0.8 3% 4% Non-gonococcal, non-chlamydial ophthalmia 12.3 53% 40% Total 23.2 ' Data from references/10 and 17. q : £ a bCohort of 1019 mothers and infants followed up for the first month of life. c Survey of Infants with ON presenting at the sexually transmitted diseases clinic in Nairobi. have a higher mortality rate, the true extent of the give an estimate of the potential complications in problem would not be apparent in prevalence surveys. puerperal women and neonates. Table 2 summarizes In a clinic-based study on 64 neonates with GCON, the prevalence of gonorrhoea in pregnant women from 16% proved to have corneal involvement (10). It is not surveys in different countries. The reported prevalence yet clear which factors are responsible for a more from the USA shows wide geographical variations fulminant course of disease; virulence of some strains ranging from 0.6% to 7.6% in the different popula- or unusual susceptibility of the host, e.g., cases among tions studied. Generally, much higher prevalence rates premature babies, have been suggested. were observed in populations in the USA than in other industrialized countries where they are usually below Gonococcal Infections 1%. In most African countries the prevalence rates Prevalence In pregnant women. Data on the prevalence among antenatal clinic patients range from 3% to as of gonorrhoea among women receiving antenatal care high as 22%. The few published results from other Table 2: Prevalence of gonococcal Infections among pregnant women worldwide Gonococcal Place No. tested cervical infection (%) Year/reference Africa: Cameroon: Yaounde 296 14 1984/11 Rural area 22 Gabon: Masuku' 530 5.5 1984/29 Gambia: Bakau 6.7 1984/19 Ghana: Accra 148 3.4 1985/3 Kenya: Nairobi 3751 6.5 1986/17 South Africa: Bloemfontein 1200 11.7 1986/28 Zambia: Lusaka - 11.2 1986/14 Asia: Malaysia: Kuala Lumpur 744 0.54 1981/12 Europe: Norway: Trondheim 686 0.3 1986/24 United Kingdom: Cardiff 625 0.16 1975/25 Glasgow 1000 0.2 Newcastle 311 0 USA: Minnesota 6464 2.7 1978/8 San Francisco 6854 0.6 1987/26 Seattle 543 2 1986/13 New York 1082 2.6 1985/20 Formerly called Franceville. 472 Epidemiology and control of gonococcal ophthalmia neonatorum continents suggest that some but not all countries have a prevalence rate similar to that in Africa. Since the mid-1970s, worldwide dissemination of penicillinase-producing N. gonorrhoeae (PPNG) strains and of strains with chromosomally-mediated penicillin and tetracycline resistance increased progres- sively. The proportion of gonococcal infection due to PPNG strains is summarized in Table 3, which shows the need for new approaches to gonorrhoea therapy. Transmlialon rate from mother to baby. The neonate acquires GCON during delivery through the infected birth canal. Occasionally the disease has been trans- mitted to infants delivered by Caesarian section after prolonged rupture of the membranes (27) The transmission rate from mother to child in the absence of ocular prophylaxis has been estimated in two prospective studies in Africa (Table 4). In Kenya (Nairobi) GCON developed in 28 out of 67 babies whose mothers had N. gonorrhoeae infections, a trans- mission rate of 42% (seven babies were lost for follow-up), while in Cameroon, 30% ofexposed babies (12/40) developed GCON; the incidence rates are therefore 3.6 and 4 per 100 live births, respectively. These incidence rates are very high, compared with those from Western countries (average, 0.06%). Asymp- tomatic carriage of N. gonorrhoeae in the eyes with minimum inflammation has been described (23). In the above-mentioned studies from Africa no children were Table 3: Percentage of gonococcal infetion due to penicil- ilnase-producing N. gonorrhoeae (PPNG) In various coun- tries Country PPNG strains (%) Year/Reference Belgium 6 1986/15 Netherlands 11 1986/7 United Kingdom 2 1985/1 USA 0.5 1982/6 Kenya 55 1986/17 Zambia 41 1986/14 Rwanda (Kigali) 51.9 1984/4 Tanzania (Dar es Salaam) 19.2 1984/21 identified with an asymptomatic gonococcal infection in the eyes. The study from Kenya demonstrated that the transmission rates of PPNG and non-PPNG strains were similar. When the mother is infected with both N. gonorrhoeae and C. trachomatis at the same time, the gonococcal transmission rate to the newborn is significantly higher (68% instead of 31%, P<0.01) (17). This study also showed that postpartum endo- metritis in the mother was a significant risk factor for GCON which may be explained by differences in pathogenicity of the strains, by both conditions arising from maternal chorioamnionitis, or by factors protect- ing the mothers against postpartum upper genital tract infection which might also protect the newborns against ocular infection. Conjunctival infection is the commonest clinical form of N. gonorrhoeae infection in the newborn, but the mucous membranes ofthe vagina, the pharynx, the rectum and the ear canal may also be colonized. Isolation rates of N. gonorrhoeae from the pharynx in neonates with gonococcal ON range from 7% to 15% (10, 17). The transmission rate of N. gonorrhoeae from the maternal cervix to extraocular sites has never been determined in the absence ofGCON and the natural history ofextra-ocular gonococcal infection is unclear. The possibility of extra-ocular colonization of N. gonorrhoeae in infants with gonococcal ON has thera- peutic implications and GCON should therefore always be treated systemically. Strategies for control Four different strategies are available for controlling GCON in neonates. (1) Primary prevention Pregnant and non-pregnant women (or their partners) who are at risk can protect themselves from acquiring gonococcal infections through behavioural modifica- tion and/or the use of barrier contraceptives. Primary prevention has not generally been successful in the Table 4: Prevalence of maternal gonococcal Infections and transmission rate from mother to child In the absence of ocular prophylaxis No. of neonates Transmission Incidence per Place and reference Prevalence (%) presenting with GCON rate (%) 100 live births Kenya (Nairobi) (17) 7 (67/1019)' 28 42b 3.6 Cameroon (Yaound6) (11) 14 (40/296) 12 30 4 ' Figures in parentheses indicate the No. of women infected/No. of women screened. b This figure is a minimal estimate since 7 out of 67 newborns were lost for follow-up; the real transmission rate ranges between 42% and 52%. 473 M. Laga et al. control of STD in the past, but the increase in incurable viral infections that are transmitted sexually, such as AIDS, may renew interest in primary preven- tion of STD. Data demonstrating a reduction in the prevalence of gonococcal infections in women as a result of behavioural change and condom use are not yet available. (2) Screening/case-fInding and treatment of gono- coccal Infections during pregnancy There are two basic approaches for early detection and/or treatment depending on the availability of laboratory services and resources: (a) identification of the infection through laboratory confirmation (mass screening or selective screening), and (b) mass treat- ment (indiscriminate or selective) without laboratory confirmation. * MN" screning. Unfortunately countries with the highest prevalence of infection, and thus the greatest need for detection programmes are in general those that are the least able to afford and perform screening. Routine screening for gonococcal infections in preg- nancy has been abandoned in many European coun- tries (because of very low prevalences), but is still practised in the USA. It has not yet been introduced in most developing countries because of lack of diagnos- tic facilities. The minimum prevalence necessary for a cost-effective screening programme has not been defined, but considering the preventable maternal and neonatal complications, it would probably be low (about 1%). Another complication is that women at high risk for gonococcal infections often do not attend antenatal care. The best way for laboratory diagnosis of gonor- rhoea is by culture, but a cheaper alternative is direct microscopic examination of a Gram stain of the cervical discharge. The latter is not a very valid test, the positive predictive value in cases where the prevalence of gonococcal infections in women is 5%, 15%, and 25% being 29%, 47% and 63%, respectively. Thus the use of Gram-stain screening is only advocated in areas with a high prevalence of gonococcal infections in women (> 15%). * Selectve screening. The cost-effectiveness of screen- ing programmes can be increased by concentrating on high-risk groups. These groups can be defined by epidemiologic risk profiling, presence of symptoms, or presence of clinical signs of cervicitis. In a series of 1000 pregnant women in Nairobi, being single and residing in certain areas were risk factors for gonococcal infection (17). * Nma treatment. Mass treatment would be indicated only if it was shown to be more cost-effective than early detection programmes, or if detection is impossible. Since in many areas the prevalences of gonococcal infections in pregnant women exceed 5% (see Table 2), indiscriminate mass treatment could be the most effective strategy if all maternal and perinatal com- plications resulting from maternal gonococcal infec- tions are taken into account. Although the two above-mentioned strategies are complex, expensive and operationally very difficult, they offer benefits besides lowering the incidence of GCON, such as reducing other complications (post- partum endometritis) related to N. gonorrhoeae. Rates of reinfection should be examined in order to assess the need for repeated mass treatment or inclusion of sex partners. Selected treatment regimens should be evaluated for efficacy, development of resistance (in areas with high proportions of PPNG), side-effects, and cost. (3) Ocular prophylaxis at birth The transmission of N. gonorrhoeae from the maternal cervix to the newborn's eyes can be interrupted by the use of eye drops or ointment immediately after birth. Three different regimens have been recommen- ded: silver nitrate, 1% eye drops; tetracycline, 1% eye ointment; and erythromycin, 0.5% eye ointment. Only the first two have been evaluated prospectively in areas with a high proportion of multiresistant gonococcal strains, including penicillinase-producing N. gonor- rhoeae. The results from Kenya demonstrated equal efficacy of silver nitrate drops and tetracycline oint- ment in the prevention of GCON, the attack rate among exposed newborns given silver nitrate and tetracycline being 7% and 3%, respectively (Table 5). These findings are consistent with risk estimates of transmission from mother to child when prophylaxis was given (2, 27). In the other studies (all retrospective) on the efficacy ofthe three regimens in preventing GCON, the prevalence of gonococcal infections in the mother was often unknown and the expected rate of GCON was too low to discern any protective effect. Among the many explanations for the failure of ocular prophylaxis to prevent neonatal ophthalmia are acquisition of infection in utero following pro- longed rupture ofmembranes, failure to instil the agent directly into the conjunctival sac, flushing of the eye after administration of silver nitrate (to prevent chemical conjunctivitis), postpartum acquisition of GCON either by autoinoculation or from other infec- ted persons, and the failure to differentiate chlamydial conjunctivitis from GCON. Silver nitrate is cheap, but is toxic if overconcent- 474 Epidemiology and control of gonococcal ophthalmia neonatorum Table 5: Attack rates of GCON among exposed newborns reelving silver nitrate, tetracycline, and no prophylaxl' Silver nitrate Tetracycline No prophylaxisb Attack rates of GCON (%)C 7.0 (5/71)d 3.0 (2/66) 46.6 (28/60) Efficacy compared to no prophylaxis 83% 93% a Includes only exposed infants seen at follow-up visits; data from reference 18. b Infants from a historical cohort. c The difference between the silver nitrate and tetracycline groups was 4.0% (95% confidence interval: -3.4 to 11.4). d Figures in parentheses are the No. of newborns with gonococcal ophthalmia/No. of newborns exposed to N. gonorrhoeae. rated (through bad preservation). Single-dose ampoules are much more expensive and less easily available. Tetracycline is nontoxic and may remain longer in the eye because it is an ointment; multidose preparations (which did not cause complications in the Nairobi trial) are cheap and widely available in developing countries. Erythromycin ointment is expensive and not available in many poor countries. It has been shown that a delay in prophylaxis of more than 4 hours after birth is associated with a 4-5-fold increase in risk ofGCON (22). Prophylaxis should therefore be given as soon after birth as practical, preferably within one hour, for both hospital and home births. Traditional birth attendants' kits should include a single-dose dispensing system for eye prophylaxis. There is no doubt that such prophylaxis is operationally the most feasible strategy and the most cost-effective approach ofthe four. In areas with a high prevalence ofgonococcal infections in pregnant women prophylaxis at birth should be reinforced or (re)intro- duced immediately as an initial step in the reduction of neonatal morbidity related to STD in pregnancy. (4) Diagnosis and treatment of GCON Early diagnosis and appropriate treatment of gono- coccal ophthalmia is important since the infection can rapidly lead to blindness. The presence of a high proportion of penicillin-resistant strains requires a more expensive and less available treatment regimen in many areas of the world. Some countries with a low prevalence of gono- coccal infections (e.g., Netherlands) have adopted the strategy of diagnosis and treatment of GCON in the newborns. The presence of neonatal infection is the indicator for an infection in the parents, and screening during pregnancy is not performed. In very low prevalence areas, with good coverage of health services, this strategy may be the most cost-effective, but has never been evaluated systematically. Unfortunately, in most countries where gonococcal infections are highly prevalent, the diagnostic facilities and appropriate treatment regimens are not available, and control programmes should therefore focus on ocular pro- phylaxis. Proposed interventions Cultural acceptability, political feasibility and psychosocial effects A major objection to the use of silver nitrate therapy has been the high incidence ofchemical conjunctivitis. Some developing countries even abandoned ocular prophyl- axis because of this. The problem can be overcome by using tetracycline ointment instead of silver nitrate drops, which is at least as effective and has no side- effects. Instilling drops or ointment in the eyes of a new- born may decrease visual alertness of the infant during the first hours of life. These problems are of concern to those who believe this may impair maternal-infant bonding by reducing eye contact (S). However, any potential impairment ofmaternal-infant bonding does not outweigh the increased risk of GCON from delaying the instillation of ocular prophylaxis. Cost-effectiveness and feasibIlfty The strategy of detection and treatment of gonococcal infections in pregnancy has the advantage of not only reducing GCON but also other maternal complications. It is, however, a very expensive (if the goal is to reduce GCON) and operationally complicated strategy, and it is currently not feasible in most areas of the developing countries. The cost of the use of ocular prophylaxis has been compared with the cost of early diagnosis and treat- ment of GCON in a group of 1000 women, with a prevalence of gonococcal infection of 10% (Table 6). The cost for 1000 prophylactic regimens is US$100 for silver nitrate 1% drops (single-dose wax ampoules) and $50 for tetracycline 1% ointment (multidose tubes for 10 babies, $0.50 per tube). It is estimated that despite ocular prophylaxis, 7% and 3% of the babies in the silver nitrate and tetracycline group, respec- 475 M. Laga et al. Table 6: Estimate of costs for the control ofGCON among 1000 pregnantwomen with prevalences of gonococcal Infection of 10% and 1% Silver nitrate Tetracycline No prophylaxis Cost of 1 dose US$ 0.10 $0.05 (single-dose (multidose ampoule) tubes) Cost for 1000 neonates $100 $50 Attack rates of GCON among newborns' 7% 3% 47% 10% prevalence: Cost of treatment of GCON (single dose, $5)b $35 $15 $235 Total cost $135 $65 $235 Cost per adverse outcome avertedc $2.90 $1.40 $5 1% prevalence: Cost of treatment of GCON (single dose, $5)b $3.50 $1.50 $23.50 Total cost $103.50 $51.50 $23.50 Cost per adverse outcome avertedc $22 $11 $5 ' Rates based on Nairobi clinical trial (17, 18); see Table 5. bOptimal treatment for GCON is ceftriaxone 125 mg, IM single dose (16). cAdverse outcome includes only GCON here. tively, will develop GCON, compared to 47% of those who receive no prophylaxis. The price of one treatment regimen for GCON varies from $5 (ceftriaxone 125 mg, intramuscular single dose) to $2 (kanamycin 75 mg+ topical tetra- cycline) (16). The assumption is that each case of GCON would be brought to the health care facilities and receive appropriate treatment. The total cost of the three programmes (silver nitrate, tetracycline, and no prophylaxis) for a group of 1000 women (with a 10% prevalence of GCON) is $135, $65 and $235, respectively, resulting in a cost per adverse outcome averted of $2.90, $1.40 and $5 (see Table 6). Indirect costs ofvisual impairment due to late or inadequate treatment are not taken into account. It is clear that the strategy of ocular prophylaxis is more cost-effective than early diagnosis and treatment. It is furthermore more convincing on humanitarian grounds, especially in areas where availability of efficacious drugs are scarce. In low-prevalence areas of maternal gonococcal infection (<1%), the price of ocular prophylaxis is higher than the price of treatment of actual cases with GCON (see Table 6). However, the potential risk of blindness (if not adequately treated) warrants further use of ocular prophylaxis in all countries where the coverage of health care is not optimal. Resume Epidemlologle et prevention de l'ophtalmle du nouveau-ne Les etiologies les plus importantes de l'ophtalmie du nouveau-ne sont N. gonorrhoeae et C. trachomatis. 476 Du point de vue de la sante publique, l'ophtalmie gonococcique (OGNN) est prioritaire puisque cette affection peut mener rapidement a la cecite chez le nouveau-ne. La frequence de l'OGNN depend de la prevalence de l'infection gonococcique chez les femmes enceintes. Dans la plupart des pays industrialises, cette prevalence est inferieure a 1%. Dans les pays en developpement, elle est souvent comprise entre 3% et 15%, et plus de la moitie des infections sont dues a des souches de N. gonorrhoeae productrices des penicill- inase (NGPP). Cela signifie que la penicilline n'est plus efficace pour traiter l'OGNN et que si des antibioti- ques efficaces, qui sont plus coiuteux, ne sont pas disponibles, le risque de cecite est considerable pour le nouveau-ne. Celui-ci est infecte a l'accouchement pendant le passage dans la filiere genitale. Le taux de transmission du gonocoque du col uterin aux yeux du nouveau-ne est de 30 a 50%. Les strategies de lutte contre l'OGNN sont: 1) prevention de la gonococcie chez les femmes en age de procreer, 2) depistage et traitement de la gonococcie chez la femme enceinte, 3) prophylaxie oculaire chez le nouveau-ne immediatement apr's la naissance, et 4) diagnostic et traitement de l'OGNN. La femme peut se proteger de la gonococcie et des autres maladies sexuellement transmissibles (MST) par un comportement sexuel adapte et l'utilisation du preservatif par le partenaire. Le depistage de la gonococcie chez la femme enceinte suivi d'un traitement adequat previent aussi bien les complications chez la femme que chez le nouveau-ne (OGNN). Mais la prevalence de la gono- coccie etant basse dans la plupart des pays industria- lises, cette strategie a un rapport coiut/efficacite trop faible. Un depistage selectif, c'est-a-dire dans des sous-groupes de femmes enceintes a risque eleve de Epidemiology and control of gonococcal ophthalmia neonatorum gonococcie, peut etre envisage. Dans les pays en developpement, le facteur limitant pour le depistage est le manque d'infrastructures de laboratoire. Le test recommande est la culture du gonocoque, peu reali- sable en dehors de quelques centres de reference dans ces pays. Dans des conditions de haute prevalence de la gonococcie chez la femme enceinte, un traitement systematique pourrait eventuellement etre envisage. L'instauration d'une prophylaxie oculaire imme- diatement apres la naissance est tres efficace. Par l'utilisation soit de collyre de nitrate d'argent a 1%, soit de pommade ophtalmique de tetracycline a 1%, l'incidence de l'OGNN diminue de 80 a 95% et cette strategie a un rapport coutt/efficacite eleve, particuliere- ment lorsque la prevalence de la gonococcie chez la femme enceinte est elevee. Le diagnostic et le traitement adequat comme strategie principale de lutte contre l'OGNN ne sont acceptables que dans les pays a faible prevalence de gonococcie chez la femme enceinte et oiu la couverture et la qualite des services de sante sont optimales. II est donc clair que la prophylaxie oculaire reste la strategie de choix dans tous les pays et regions otu la couverture sanitaire laisse encore a desirer. References 1. Adhr, M.W. Epidemiology and treatment of penicillinase- producing Neisseria gonorrhoeae. In: Oriel, D.J. & Harris, J.R.W., ed. Recent advances in sexually transmitted diseases. Edinburgh, Churchill Livingstone, 1985, pp. 23-38. 2. Armstrong, J.H. et al. Ophthalmia neonatorum: a chart review. Pediatrics, 57: 884-892 (1976). 3. Bental, C. et al. Genital infections with Chiamydia trachomatis and Neisseria gonorrhoeae in Ghanaian women. Genitourinary medicine, 61: 48-50 (1985). 4. Bogaerts, J. et al. In vitroantimicrobial sensitivity of N. gonorrhoeae from Rwanda. Genitourinary medicine, 62: 217-220 (1986). 5. Butterfield, P.M. et al. Effects of silver nitrate on initial visual behavior. American journal ofdiseases ofchildren, 132: 426 (1978). 6. Centers for Disease Control. Global distribution of penicillinase-producing Neisseria gonorrhoeae (PPNG). Morbidity & mortality weekly report, 32: 1-3 (1982). 7. CommIttee on Health Care. [Guidelines on sexually transmitted diseases.] The Hague, 1986 (in Dutch). 8. Edwards, LE. at al. Gonorrhea in pregnancy. American joumal ofobstetrics and gynecology, 132: 637-641 (1978). 9. Fransn, L et al. Parents of infants with ophthalmia neonatorum: a high-risk group for sexually transmitted diseases. Sexually transmitted diseases, 12: 150-154 (1985). 10. Fransen, L d al. Ophthalmia neonatorum in Nairobi, Kenya: the roles of Neisseria gonorrhoeae and Chlamydia trachomatis. Joumal of infctious diseases 153: 862-8 (1986). 11. Gabega, F.P. et al. Gonococcal ophthalmia neonatorum: the case for prophylaxis in tropical Africa. Bulletin ofthe World Health Organization, 62- 95-6 (1984). 12. Goh, T.H. et al. Screening for gonorrhoea in a prenatal clinic in southeast Asia. Sexually transmitted diseases, 8: 67-69 (1981). 13. Gravet, M.G. t al. Independent association of bacterial vaginosis and Chlamydia btmdomats infection with adverse pregnancy outcome. Journal of the American Medical Association, 256: 1899-1903 (1986). 14. HIra, S.K. Sexually transmitted diseases-a menace to mothers and children. World health forum, 7: 243-247 (1986). 15. Institute for Hygiene and Epidemiology. [Surveillance of infectious diseases through a network of microbio- logy laboratories.] Report No. 54, 1986 (in Dutch). 16. Laga, M. et al. Single-dose therapy of gonococcal ophthalmia neonatorum with ceftriaxone. New England journal of medicine, 315: 1382-1385 (1986). 17. Laga, M. et al. Epidemiology of ophthalmia neonatorum in Kenya. Lancet, 2: 1145-1148 (1986). 18. Laga, M. et al. Prophylaxis of gonococcal and chlamy- dial ophthalmia neonatorum. A comparison of silver nitrate and tetracycline. New England journal of medi- cine, 318: 653-657 (1988). 19. Mabey, D.C.W. t al. Sexually-transmitted diseases among randomly selected attenders at an antenatal clinic in the Gambia. British journal of venereal diseases, 60: 331-336 (1984). 20. McMlllan, J.A. et al. Efficacy of maternal screening and therapy in the prevention of chlamydia infection of the newborn. Infection, 13: 263-266 (1985). 21. Maselle, W.B. et al. Penicillinase-producing N. gonor- rhoeae in Dar es Salaam, Tanzania. African journal of sexually transmitted diseases, 1: 71-75 (1984). 22. Muhe, L. & Tafarl, N. Is there a critical time for prophylaxis against neonatal gonococcal ophthalmia? Genitourinary medicine, 62: 356-357 (1986). 23. Podgore, J.K. & Holmes, K.K. Ocular gonococcal infection with minimal or no inflammatory response. Journal of the American Medical Association, 246: 242-243 (1981). 24. Skeldetd, F.E. & Deen, A. The prevalence of Chlamydia trachomatis in the cervix of puerperal women, and its consequences for the outcome of pregnancy. Scandi- navian journal of primary health care, 4: 209-212 (1986). 25. Sparks, R.A. et al. Antenatal screening for candidiasis, trichomoniasis and gonorrhoea. British journal of venereal diseases, 51: 110-115 (1975). 26. Sweet, R.L. et al. Chlamydia trachomatis infection and pregnancy outcome. American journal of obstetrics and gynecology, 156: 824-833 (1987). 27. Thompson, T.R. et al. Gonococcal ophthalmia neona- torum: relationship of time of infection to relevant control measures. Journal of the American Medical Association, 228: 186-188 (1974). 28. Welgemoed, N.C. et al. Prevalence of Neisseria gonor- rhoeae infection in patients attending an antenatal clinic. South African mnedical joumal, 09: 32_34 (1986). 29. Yvert, F. et al. Les infections gonococciques au Gabon, Haut-Ogooud. Pathologie biologie, 32: 80-1 (1984). 477