The utility of verbal autopsies for identifying HIV-1-related deaths in Haitian children.

OBJECTIVE: To determine whether deaths among Haitian infants born to HIV-1-seronegative women could be distinguished from deaths among children born to HIV-1-seropositive women using the verbal autopsy technique. METHODS: Mothers of 315 Haitian children who died were interviewed about events leading to the child's death. Three physicians independently reviewed interview data and determined the probable cause of death without knowledge of maternal HIV-1 status or hospital records. The underlying causes of death assigned to the infants were analyzed to determine whether maternal HIV status could be predicted. RESULTS: There was good agreement among the physicians (kappa = 0.62) and 90% agreement between hospital records and the verbal autopsy diagnosis. Compared with children born to HIV-1-seronegative women, deaths in children born to HIV-1-seropositive mothers were more likely to be ascribed to a presumptive diagnosis of AIDS (37 versus 21%; P = 0.01). The sensitivity and specificity of verbal autopsies for identifying deaths associated with maternal HIV-1 infection ranged from 37 to 59% and from 69 to 79%, respectively, depending on the classification system used. The predictive positive value of a death believed to be consistent with pediatric HIV-1 infection was 26-30% and the predictive negative value was 85-90%. CONCLUSION: Verbal autopsies may be useful for distinguishing certain causes of death, but have limited utility for distinguishing deaths associated with maternal HIV-1 infection from deaths among children born to HIV-1-seronegative women.

Lack of association between maternal antibodies to V3 loop peptides and maternal-infant HIV-1 transmission.

We investigated the association between maternal antibodies to HIV-1 peptides in pregnant women and the acquisition of HIV-1 infection by their offspring. Pregnant HIV-1-infected Haitian women were tested for the presence of antibodies against peptides of 14-17 amino acid length from the V3 loop region of strains IIIb and MN. Antibody testing was performed in two separate laboratories by enzyme-linked immunosorbent assay (ELISA). Peptides from four regions of the V3 loop were synthesized in several different laboratories and the purity confirmed by high performance liquid chromatography (HPLC). The mothers of infants who acquired HIV-1 infections did not differ significantly from the mothers of uninfected infants in the prevalence or concentration of antibodies to any of the 15 peptides evaluated. Additional studies are indicated to determine if neutralizing antibodies or other immunologic parameters are associated with maternal-infant HIV-1 transmission.

Diagnosis and differentiation of HTLV-I and HTLV-II infection by enzyme immunoassays using synthetic peptides.

Synthetic peptides from the major envelope protein of HTLV-I (ENV-I, amino acid 177-213) and HTLV-II (ENV-II, amino acid 173-209) and a conserved region of the transmembrane protein (TM, amino acid 378-402) were used as antigens in microtiter plate enzyme immunoassays (EIA) to detect and discriminate antibodies to HTLV-I and II. The ENV-I and ENV-II peptide EIAs were able to correctly discriminate HTLV-I and II infections in 17 of 18 subjects whose infections were determined by a gene amplification method. Sera from 100 of 107 subjects with serologically confirmed infection with HTLV-I/II and 0 of 218 seronegative controls reacted with one or more of the peptides (sensitivity, 93.5%; specificity, 100%). Ninety-six of the 100 peptide positive sera reacted exclusively with either the ENV-I or the ENV-II peptide, thereby differentiating the two viral infections. The pattern of reactivity to the ENV peptides was distinct in different populations. Patients attending an Emergency Department, who had a history of drug abuse, and male inmate entering a correctional facility only had antibody reactivity to the ENV-II peptide. Subjects from Haiti and patients with HTLV-associated neurological disease only had antibody reactivity to the ENV-I peptide. Peptide-based enzyme immunoassays that distinguish antibodies to HTLV-I and HTLV-II will facilitate studies of the epidemiology of HTLV.

Prevalence of HIV-1 DNA and p24 antigen in breast milk and correlation with maternal factors.

Breast milk specimens from human immunodeficiency virus type 1 (HIV-1)-seropositive and HIV-1-seronegative women were examined for the presence of HIV-1 p24 antigen by the antigen capture method and for viral DNA using the polymerase chain reaction. HIV-1 DNA was present in 70% of milk specimens collected from 47 HIV-seropositive women 0-4 days after delivery and in approximately 50% of specimens collected 6 and 12 months postpartum. p24 antigen, present in 24% of milk specimens collected from 37 seropositive women within the first 4 days postpartum, was not detected in any of the subsequent specimens. The presence of HIV-1 DNA or p24 antigen in milk was not significantly associated with maternal CD4 lymphocyte count, beta 2-microglobulin level, or fulfillment of the AIDS clinical case definition. Although the correlation of either HIV-1 proviral DNA or p24 antigen with the presence of infectious virus is not known, these data indicate the need for additional studies examining the role of breastfeeding in maternal-infant transmission of HIV-1.

Human immunodeficiency virus neutralizing antibodies in sera from North Americans and Africans.

Neutralizing antibodies (NAs) against four isolates of human immunodeficiency virus type 1 (HIV-1) were assayed in HIV-1 antibody positive sera from the United States, Haiti, Zimbabwe, and Zaire. Overall, there were NAs detected in 95, 81, 60, and 73% of sera with reciprocal geometric mean titers (GMTs) of 626, 23, 10, and 20, respectively, against HIV-1MN, HIV-1IIIB, HIV-1RF, and HIV-1Z3. Sera from North America had significantly higher NA titers against HIV-1MN. In each country, the highest antibody titers observed were against the MN strain. Otherwise, sera from the U.S. neutralized most strongly HIV-1IIIB, sera from Zaire neutralized most strongly HIV-1Z3, and sera from Zimbabwe had equal titers against all three viruses. The differences between countries were reflected in analyses of NA titers of subgroups classified on the basis of clinical status, indicating that the differences were not likely to be related to differences in clinical status of the patients being tested. Some of this antigenic variation is reflective of known genetic diversity, while some is not. The results suggest that undefined preserved and variable regions containing neutralization epitope(s) exist. These data do not indicate a need to define antigenic subtypes of HIV-1 at present. The existence of conserved neutralization epitope(s) may indicate the potential for broad immunogenicity of appropriately selected vaccine antigens.

HIV-1 in Haitian women 1982-1988. The Cite Soleil/JHU AIDS Project Team.

Pregnant Haitian women (n = 4,474) residing in a periurban slum were interviewed to identify risk factors for sexually transmitted diseases and sera were tested to identify antibodies to HIV-1 and syphilis. The seroprevalence rates for antibodies to HIV-1 increased from 8.9% in 1986 to 9.9% in 1987 and 10.3% in 1988. Sera obtained in 1982 from 533 mothers of young infants in the same community revealed that 7.8% were HIV-1 seropositive. Of women pregnant for the first time in 1986-1988, 6.6% were HIV-1 seropositive and 6.0% had a positive VDRL. The highest seropositivity rates (greater than 15%) were noted in women 20 to 29 years of age with a history of two or more sexual partners in the year prior to pregnancy. Factors independently associated with HIV-1 seropositivity in pregnant women by logistic regression analysis included being unmarried, age 20-29 years, having had more than one sex partner in the year prior to pregnancy, a positive serologic test for syphilis, and smoking. A dose-response effect was noted in the association between HIV-1 seropositivity and smoking. The association between smoking and HIV-1 infections could be confounded by unrecognized behavioral factors or due to a biologic effect of smoking. The continuing high HIV-1 seropositivity rates in pregnant women indicate that current preventive measures are insufficient and increased control efforts are urgently needed.

Childhood survival in Haiti: protective effect of measles vaccination.

To evaluate the impact of measles vaccination on survival of children residing in a periurban slum in Haiti, a total-population survey was conducted 2.5 years after completion of a one-time study of the serologic response to measles vaccine administered in the same population. Pregnancy histories from the 16,400 women in the population revealed that 1499 children had been born during a 7-month interval that would have made them eligible for participation in the measles vaccine program. Of these children, 1381 (92.1%) survived to 9 months of age, the median age that measles vaccine had been administered. Seventy-three infants had died between 9 and 39 months of age. Mortality of infants who were seronegative before receiving measles vaccine was significantly lower (P = .0013) than that of unvaccinated infants (3/235 vs 70/1056, respectively). Other factors positively associated with survival between 9 and 39 months of age included socioeconomic status (P = .0002), maternal literacy (P = .0020), maternal knowledge and use of oral rehydration solution (P = .0002), and an interval of greater than 24 months to the birth of the next younger sibling (P = .0012). Multivariate stepwise logistic regression analysis was used to evaluate the independent association of measles vaccination by adjusting for other factors that also correlated with survival and that might have been associated with maternal seeking of vaccinations.(ABSTRACT TRUNCATED AT 250 WORDS)

Epidemiologic studies of measles, measles vaccine, and subacute sclerosing panencephalitis.

Histories obtained in 350 of 375 clinically cofirmed cases of subacute sclerosing panencephalitis (SSPE) reported to a national registry showed that 292 patients had measles and 58 had no history of measles. Forty of the latter patients received live, attenuated measles virus vaccine. In patients with a history of measles, measles illness occurred before age 2 years in 46%, and a mean of 7.0 years before onset of SSPE. In contrast, there was no relationship of SSPE with age at vaccination in 35 of the 40 patients historically associated with measles vaccine, and SSPE occurred a mean of 3.3 years after vaccination. Based on estimated national measles morbidity data and national measles vaccine distribution data, the risk of SSPE following measles vaccination (0.5 to 1.1 cases/106) appears to be less than the risk following measles (5.2 to 9.7 cases/106). Because live measles vaccine is highly effective in preventing measles illness and a high proportion of children in the United States have received measles vaccine, these data are consistent with the observed downward trend in SSPE incidence since 1969.

Differences in antibody response to whole-cell pertussis vaccines.

It has been assumed that whole-cell pertussis vaccines (WCVs) commercially distributed in the United States are of comparable immunogenicity, as all must comply with established standards for licensure. However, we have recently noted significant differences in antibody responses between groups of infants receiving the two WCVs commercially available in the United States. In separate studies performed concurrently under similar protocols at Vanderbilt and Johns Hopkins universities, infants were randomized to receive either an acellular pertussis vaccine or WCV. The acellular pertussis vaccine used at the two sites was identical, but the WCVs were from different manufacturers. Antibody responses to acellular pertussis vaccine did not differ between the two studies; responses to WCV differed dramatically, with infants receiving the Lederle WCV producing a 46-fold increase in antibody to pertussis toxin, compared with a 2.4-fold increase for infants receiving the Connaught WCV (P = .00003). Evaluation of other comparative data sets that were available provided further support for the conclusion that the two commercially available WCVs consistently differed in their ability to induce antibody to pertussis toxin. These findings have important implications for the design and interpretation of clinical trials comparing acellular and WCV products.

Serum antibody response to pneumococcal vaccine in children with nephrotic syndrome.

Children with nephrotic syndrome are susceptible to serious pneumococcal disease and may be immunodeficient on the basis of abnormal humoral immune responses to natural antigens or immunoglobulin loss during relapse. As part of an ongoing study to evaluate pneumococcal anticapsular antibody concentration and immunologic competence, 27 steroid-responsive and six steroid-resistant patients with nephrotic syndrome, and 12 age-matched control subjects, were vaccinated with polyvalent pneumococcal vaccine. Antibody responders were defined as patients with at least a twofold increase in antibody after vaccination as well as an antibody concentration greater than 200 ng of anticapsular antibody nitrogen per milliliter (ngN/ml) after vaccination. Pneumococcal antibody concentrations before and after vaccination were significantly depressed in steroid-resistant patients when compared with control subjects (P less than .002) and with the steroid-responsive nephrotic syndrome group (P less then .001). Steroid-responsive nephrotic children who were not receiving corticosteroid therapy at the time of vaccination had significantly higher antibody concentrations to five pneumococcal types before vaccination and to seven types after vaccination compared with control subjects (P less than .05). Fewer steroid-responsive patients receiving corticosteroids achieved antibody concentrations greater than or equal to 200 ngN/ml against type 19F compared with patients not receiving steroids or with control subjects (P less than .05). These results suggest that pneumococcal vaccine is immunogenic in children with steroid-responsive nephrotic syndrome and may protect these patients from disease due to pneumococcal types contained in the vaccine.

Bacillus Calmette-Guérin complications in children born to HIV-1-infected women with a review of the literature.

OBJECTIVE: To compare the risk of complications following Bacillus Calmette-Guérin (BCG) vaccination among children by maternal and infant HIV-1 infection status as part of an investigation of an outbreak of BCG complications. METHODS: A nonconcurrent cohort study of BCG complications among 125 infants born to HIV-1 seropositive and 166 infants born to HIV-1 seronegative mothers was conducted in Cité Soleil, Haiti. Infants were examined at regular intervals until 15 months of age, and complications from BCG were documented. An investigation of BCG vaccination practices was conducted. RESULTS: Mild or moderate complications occurred among 16 of 166 (9.6%) infants born to HIV-1 seronegative mothers compared with 4 of 13 HIV-1-infected infants (30.8%, P = .04) and 10 of 75 (13.3%, P = .39) uninfected infants born to HIV-1-infected mothers. No serious complications were noted. The outbreak of complications was associated with administration of 2.0 to 2.5 times the recommended dose of BCG vaccine. CONCLUSIONS: This and five other cohort studies indicate that there may be a small increased risk of complications following BCG vaccination among HIV-1-infected children, but the reactions are usually mild and the risk does not outweigh the benefits of BCG vaccination in populations at high risk of tuberculosis during infancy and childhood.

Hepatitis B vaccine administered to children and adolescents at yearly intervals.

OBJECTIVE: Hepatitis B vaccines are usually administered on a schedule of 0, 1 to 2, and 6 months. Longer intervals between the second and third doses have been studied, but the effectiveness of hepatitis B vaccine administered at intervals of >2 months between the first and second doses have not been studied. Our objective was to compare the antibody response in recipients of Engerix-B hepatitis B vaccine administered at 12-month intervals to the response to vaccine administered at 0-, 1-, and 6-month intervals. METHODS: A total of 389 children, 5 through 16 years of age, were randomized to receive Engerix-B (10 mg) at a schedule of either 0-, 1-, and 6-month intervals or 0-, 12-, and 24-month intervals. Blood was drawn before and 1 month after the third dose. RESULTS: Immediately before the third dose of vaccine, 92.3% of children who received vaccine on the 0-, 1-, and 6-month schedule and 88.8% of children who received the 0-, 12-, and 24-month schedule had antibody to hepatitis B surface (anti-HBs) antigen concentrations >/=10 mIU/mL. Of the children in the 0-, 1-, and 6-month schedule, 95% received the third dose according to protocol versus 90% of those in the 0-, 12-, 24-month schedule. The geometric mean anti-HBs concentration just before the third dose for recipients of the 0-, 1-, and 6-month schedule (117.9 mIU/mL) was somewhat lower than that for the children who had received vaccine on the 0-, 12-, and 24-month schedule (162.1 mIU/mL). One month after the third dose, >98% of all children had anti-HBs concentrations >/=10 mIU/mL and high geometric mean antibody concentrations were observed in both groups: 5687 mIU/mL for children on the 0-, 1-, and 6-month schedule and 3159 mIU/mL for children on the 0-, 12-, and 24-month schedule. Body mass index was correlated inversely with final antibody concentration, but age was not a factor after adjustment for body mass index. DISCUSSION: Engerix-B administered on a 0-, 12-, and 24-month schedule is highly immunogenic. Providers should consider this alternate immunization schedule for children who are at low risk of immediate exposure to hepatitis B infections.

Safety of combination vaccines: perception versus reality.

BACKGROUND: Combination vaccines contain multiple antigens to protect against several diseases simultaneously and have simplified the delivery of childhood immunizations. Children are healthier today because of the use of combination vaccines, and the United States is benefiting from record low numbers of vaccine-preventable diseases. Despite obvious benefits, concerns and misconceptions exist regarding the safety and efficacy of combination vaccines. METHODS: A review of the pediatric literature to dispel the common misperceptions and potential barriers to combining vaccines. RESULTS: Assurance that combination vaccines approved by the United States Food and Drug Administration undergo extensive testing will help to alleviate concerns regarding safety and efficacy of combination vaccines. Food and Drug Administration standards are rigorous and require that combination vaccines be as safe and effective as each component of the vaccine administered separately. Combination vaccines have been available for >50 years, and lessons learned during this time are continuously applied to the development and use of new products. CONCLUSIONS: Children will benefit from new combination vaccines because fewer injections will be required to protect against vaccine-preventable diseases, allowing for the introduction of new vaccines into the immunization schedule and prevention of additional diseases.

Discussion of Immunization Practices Advisory Committee/American Academy of Pediatrics recommendations for universal infant hepatitis B vaccination.

The Committee on Infectious Diseases of the American Academy of Pediatrics and the Advisory Committee on Immunization Practices of the United States Public Health Service have revised the recommendations for hepatitis B control to include universal immunization of infants. Previous selective strategies targeting high risk groups for immunization have failed to prevent the significant morbidity and mortality associated with hepatitis B virus infection. The committees considered several alternative policy options before selecting universal immunization of infants as the most effective strategy for preventing the disease and its sequelae in the United States. More than 30 other countries have also adopted universal infant immunization programs. All children should receive three doses of vaccine by the time they are 18 months of age. Scheduling of the three doses is flexible. Two hepatitis B vaccines are currently available in the United States, and both are highly effective and safe. Dosage depends on several factors, including manufacturer, age at vaccination, maternal serologic status for hepatitis B surface antigen and presence of underlying disease. The committees continue to recommend immunization of high risk older children and adults, and immunization of adolescents is encouraged where resources are available.

Pediatric acquired immunodeficiency syndrome: special considerations for developing nations.

PIP: WHO estimates the existence of 1 million AIDS cases and over 500,000 AIDS deaths over the past 10 years. An estimated 9-11 million people are infected, including 3 million women and 500,000 infants and children. 80% of infected women and children live in sub-Saharan Africa. Their numbers are expected to continue to rise. HIV-1 will probably lead to the death of 3 million or more women and 2.7 million children over the next decade, and will become the leading cause of death for men and women aged 15-49 in major North, Central, and South American, West European, and sub-Saharan African cities. Accordingly, pediatric AIDS threatens much of what has been achieved in improving developing country child survival over the past 20 years. Concerted efforts must be made on an international scale to prevent and control the growing pandemic. This paper discusses the magnitude of HIV-1 infection and AIDS, maternal-infant transmission, breast-feeding, blood transfusion, diagnosis, clinical features and medical management, impact on child survival, and prevention and control.^ieng

Successful immunization of infants at 6 months of age with high dose Edmonston-Zagreb measles vaccine. Cite Soleil/JHU Project Team.

A group of 2097 Haitian infants 6 to 11 months of age were randomized to receive Schwarz or Edmonston-Zagreb strain measles vaccines containing 10- to 500-fold more vaccine viral particles than standard potency vaccines. No unusual adverse reactions were noted. Edmonston-Zagreb vaccines were more effective than equivalent doses of Schwarz vaccines as measured by the proportion of vaccinated children with measles antibody concentrations greater than or equal to 200 mIU/ml 2 months after vaccination and the persistence of antibody at 18 to 24 months of age. High titer Edmonston-Zagreb vaccine administered at 6 months of age induced antibody concentrations greater than or equal to 200 mIU/ml in 83% of infants by plaque reduction neutralization and 93% of infants by enzyme-linked immunosorbent assay with high rates of antibody persistence at 12 to 24 months of age. The World Health Organization recommends high titer Edmonston-Zagreb measles vaccines for routine use at 6 months of age in areas where measles is an important cause of mortality in young infants.

Comparison of high titer Edmonston-Zagreb, Biken-CAM and Schwarz measles vaccines in Peruvian infants.

In an effort to identify the optimal dose and strain of measles vaccination for early immunization, Peruvian infants were randomly assigned to receive one of three measles vaccines in varying doses at 5 to 6 or 8 to 9 months of age. Edmonston-Zagreb vaccines were significantly (P < 0.001) more immunogenic than equivalent or higher titers of Schwarz or Biken-CAM vaccines as determined by neutralization antibody response 3 months after vaccination. Eighty-two percent of infants who received high titer Edmonston-Zagreb vaccine at 5 to 6 months of age developed protective concentrations of measles antibody, a response rate similar to that observed after standard titer Schwarz (81%) or high titer Biken-CAM vaccine (81%) at 8 to 9 months of age. No significant differences in the rates of fever, rash or other adverse events were noted by vaccine group 10 to 14 days after vaccination. Although the high titer vaccines are more immunogenic in young infants than standard vaccines, long term safety must be assured before these vaccines can be put into widespread use.

Inactivated poliovirus vaccine alone or sequential inactivated and oral poliovirus vaccine in two-, four- and six-month-old infants with combination Haemophilus influenzae type b/hepatitis B vaccine.

BACKGROUND: Advisory committees have recommended the increased use of inactivated poliovirus vaccine (IPV) for children. OBJECTIVES: The purpose of this study was to assess the safety and immunogenicity of three schedules using IPV administered with diphtheria and tetanus toxoids and whole cell pertussis vaccines in a dual-chambered syringe. Children also received a combination of Haemophilus influenzae type b (Hib) and hepatitis B vaccines (COMVAX). METHODS: All infants (n = 295) received IPV and COMVAX at 2 and 4 months of age and IPV, oral poliovirus vaccine (OPV) or both vaccines at 6 months and OPV at 15 months of age. RESULTS: After two doses of IPV 96 to 100% of infants had antibodies to poliomyelitis viruses types 1, 2 and 3, and after a third dose of vaccine (IPV or OPV) all but one child had antibodies to all three poliovirus types. After two doses of COMVAX 89 and 96% of children had protective levels of antibody to Hib and hepatitis B, respectively. CONCLUSIONS: IPV is highly immunogenic in a two-dose schedule. Administration of a third dose of IPV or a dose of OPV at 6 months of age is highly effective. Simultaneous administration of a combination H. influenzae type b/hepatitis B vaccine did not interfere with the response to IPV.

Use of ethnographic research for instrument development in a case-control study of immunization use in Haiti.

A rapid ethnographic assessment of barriers to health service utilization was conducted to identify maternal factors predicting use of child immunizations in Haiti. Methods included four focus group interviews, four natural group interviews, individual interviews with 14 health care providers and participant observation at vaccination posts. Analysis of qualitative data identified five categories of maternal factors associated with immunization use: competing priorities, low motivation, socioeconomic constraints, perceived accessibility of services, fears about health or social consequences and knowledge and folk beliefs related to vaccines. Selected variables among these factors were incorporated into a survey instrument designed to compare mothers of completely vaccinated children (cases) with mothers of incompletely vaccinated children (controls). The questionnaire was administered to 299 randomly selected mothers (217 cases, 82 controls). Bivariate and multivariate analyses found that of the factors identified through ethnographic research, only vaccine-related knowledge was significantly associated with immunization status. The utility and constraints of using ethnographic research for instrument development in epidemiological studies are discussed.

PIP: A rapid ethnographic assessment consisting of 4 focus groups, 4 natural group interviews, 14 provider interviews and participant observation rally posts was combined into a questionnaire which was tested by chi square for use in predicting which Haitian mothers utilize immunization services. The study population were mothers of 1800 children 12-23 months old in the Mirebalais Area Community Health program in the central highlands of Haiti. The program uses the rally post system, with teams visiting villages at 6-week intervals. Audio tapes and written notes were reduced to lists of all possible barriers to maternal utilization, and then regrouped into 16 categories, under 5 topics: competing priorities, lack of motivation, socioeconomic constraints, perceived accessibility of posts, fears of health consequences and knowledge of folk beliefs about vaccines. 22 questionnaire items were then developed, which were edited by the overall project director without input from the ethnographer. The questions were administered to 299 mothers, 82 with incompletely immunized, and 217 with completely immunized children. The questions which significantly predicted complete vaccination by bivariate analysis were knowledge of the name of 1 or more vaccines or illnesses, the recommended number of doses, and the correct age to begin vaccinations. It was likely that time demands from subsistence farming and income generating activities also affected service utilization, but the women probably interpreted the question on employment incorrectly.

Risk factors for trichiasis in women in Kongwa, Tanzania: a case-control study.

Women are at a greater risk compared to men for blinding complications from trachoma. In order to evaluate risk factors in women, 205 cases of trichiasis (TT) were selected from 11 villages in rural Tanzania. Each case of trichiasis was matched to two women of the same age and from the same village, who had no clinical signs of trachoma. Factors associated with trichiasis in a conditional logistic regression included history of trichiasis in the women's mother (odds ratio [OR] = 3.6; 95% confidence interval [CI]: 2.0-6.5); sleeping in a room with a cooking fire during childbearing years (OR = 1.8; 95% CI: 1.2-2.8); a home of wood and earth during childbearing years (OR = 2.1; 95% CI: 1.3-3.3); no adult education classes (OR = 2.2; 95% CI: 1.4-3.4); and five or more deaths among her children (OR = 2.6; 95% CI: 1.3-5.1). Detailed measures of prolonged exposure to child care as a young girl and as a mother showed no significant difference between cases and controls. Results from this study characterize women at high risk for severe disease and eventual blindness in a trachoma endemic area.

PIP: Ophthalmologists analyzed data on 205 women with ingrowing eyelashes caused by trachomatous scarring (trichiasis) with data on 410 age- and village-matched controls to determine risk factors for trichiasis in trachoma hyperendemic areas where women face a greater risk of trichiasis than do men. All women lived in the Kongwa subdistrict of Dodoma Region, Tanzania. Cases were 3.6 times more likely to have a mother who also had trichiasis. Family history of fathers or other relatives was not associated with trichiasis, however. Cases tended to have never married (odds ratio [OR] = 3.7), suggesting that in a society that expects marriage women with trichiasis were socially and economically marginalized. Mothers with trichiasis were more apt to experience at least 5 deaths of children than controls (OR = 2.6). The researchers assumed that a factor common to both trichiasis and increased child deaths (e.g., socioeconomic status) accounted for this association. Controls were more likely to have attended adult education classes than cases (OR = 2.2), yet attendance at primary school was not a predictor for trichiasis. Women who had slept in a room with a cooking fire during their reproductive years were more prone to trichiasis than women who did during childhood or as current practice (OR -1.8), indicating a more profound pathogenic response (scarring) to infection with Chlamydia trachomatis during pregnancy. Living in a wood/earth (i.e., traditional) home during childbearing years was also a significant predictor of trichiasis (OR = 2.1). These findings suggest that improving living conditions, educational opportunities, and socioeconomic status during the childbearing years would reduce the risk for trichiasis and blindness.