Risk factors for HIV-1 infection among women in the Arusha region of Tanzania.

Risk factors for HIV-1 infection among women were assessed through a population-based cross-sectional study in the Arusha region of northern Tanzania. The study participants were obtained by randomly selecting 10-household clusters from Unga limited, the town of Babati, and the roadside village of Matufa, which are urban, semi-urban, and rural communities, respectively. Informed verbal consent for participation in an interview and in HIV-1 testing was sought from each respondent. Blood samples were collected from each consenting individual for HIV-1 antibody testing using enzyme-linked immunosorbent assay (ELISA), and all positive sera were confirmed using repeated ELISA tests. Information of risk factors was obtained through the interview process using a structured questionnaire. Of the 567 women who gave blood samples, 48 (8.5%) were HIV-1 positive. The HIV-1 seroprevalence rates among women in the urban area, the semi-urban area, and the rural village were 14.4%, 6.9% and 2.3%, respectively. Factors associated with significantly higher HIV-1 seroprevalence were urban residence; history of having traveled out of the Arusha region within Tanzania, as well as having traveled abroad; having multiple sex partners; and having sexual intercourse under the influence of alcohol. Women who reported ever having used condoms had significantly higher probability of being infected with HIV-1 than those who had never used condoms, suggesting that condom use may be a marker of high-risk sexual behavior and that condom use is probably not adhered to in a way that consistently protects against HIV-1 infection. These results suggest the need for health education interventions aimed at increasing appropriate and consistent condom use and reduction of the number of sexual partners.

PIP: A population-based cross-sectional study conducted in northern Tanzania's Arusha region found women to be at even greater risk of human immunodeficiency virus (HIV) infection than men. Three diverse sites--Unga limited (urban), the town of Babati (semi-urban), and the roadside village of Matufa (rural)--were selected. 862 (567 women and 295 men) out of 2216 eligible respondents both completed a questionnaire and provided a blood sample for HIV testing. Overall HIV-1 seroprevalence was significantly higher for women (8.5%) than men (2.0%), with a range from 2.3% among women in the rural village of Matufa to 14.4% among urban women in Unga limited. Marital status, occupation, age at first intercourse, educational status, religion, and travel within the Arusha region were not associated with a significant increase in HIV infection; however, travel out of the region or out of Tanzania was a significant risk factor. Also significant was having two or more sexual partners in the last six months or past five years. Finally, HIV rates were significantly higher among women who used condoms and those who had sexual intercourse under the influence of alcohol. The high HIV seroprevalence rate identified in this study among women may reflect selection bias and other potential risk factors (e.g., previous history of sexually transmitted diseases) that were not assessed. Of particular concern is the likelihood, given condom use as an HIV risk factor, that condoms are being used incorrectly or inconsistently in this population. Since Tanzanian women lack power in sexual relations, HIV/AIDS preventive interventions must target men as well as women.

Hemoglobin concentration in pregnant women. Experience from Moshi, Tanzania.

BACKGROUND: Anemia in pregnancy is common in Tanzania. Previous information on distribution of hemoglobin concentration in total populations is scarce. METHODS: Standard measurement of hemoglobin concentration in consecutive first visit attenders at the antenatal care clinic in Moshi, 1991-1994, for a total of 1800, divided into three series of 600 each. RESULTS: Hemoglobin concentrations ranged from 41 to 146 g/l, mean 96.9 and s.d. 16.3. For international comparison, 74.5% were below 110 g/l and 7.0% below 70 g/l. The median was 97.0; other percentiles were 10th: 75.0, 90th: 116.0. The distribution did not change with mother's age, but in the years 1991 and 1994 there were more low hemoglobin values than in the intervening years. CONCLUSIONS: Low hemoglobin concentrations are prevalent among pregnant women in Moshi, Tanzania. Seven percent were found to have severe anemia, defined as a hemoglobin concentration below 70 g/l.

Prevalence of HIV-1 infection in urban, semi-urban and rural areas in Arusha region, Tanzania.

OBJECTIVE: To estimate HIV-1 prevalence among the adult population in the Arusha region, northern Tanzania. DESIGN: Cross-sectional study. METHODS: Clusters of 10 households were randomly selected from urban, semi-urban and rural areas in the Arusha region. Informed verbal consent for participation in the interview and HIV-1 test were obtained from the respondents. HIV-1 antibodies were tested using enzyme-linked immunosorbent assay (ELISA) and all ELISA-positive sera were confirmed using Western blot. Approximately 66.0% of the eligible registered adults (aged 15-54 years) participated in the interview, of whom 56.6% provided blood samples for HIV-1 testing. RESULTS: HIV-1 prevalence rates among the study population from the low and high socioeconomic status urban areas, semi-urban area and rural village were 10.7, 5.2, 2.2 and 1.6%, respectively. HIV-1 prevalence was significantly higher among women (6.5%) than men (1.7%), and infection rates were consistently higher among women than men in all areas studied except in the rural village. Divorced and separated individuals had significantly higher HIV-1 prevalence (14.8%) compared with married and cohabiting individuals (4.0%). Significantly more individuals with multiple sexual partners reported regular condom use (19.3%) than those with one sexual partner (6.4%). CONCLUSION: HIV-1 prevalence in the Arusha region was higher among the urban population than among the rural population. Women living in urban areas with low socioeconomic status appeared to be associated with increased risk of HIV-1 infection.

Demographic and AIDS-related characteristics of consenters to a population-based HIV-survey: results from a pilot study in Arusha, Tanzania.

The aim was to compare demographic and AIDS-related characteristics of people who consented to HIV-testing as part of a population survey with those who did not consent to HIV-testing. Subjects, aged 15-54 years, living in ten randomly selected clusters of households in one ward of Arusha town were asked to participate in a structured interview and to provide a blood sample for HIV-testing. Measurements included demographic variables and AIDS-related factors, such as knowledge of AIDS and sexual behaviour, and HIV-testing with Western Blot confirmation. Sixty-two percent of the eligible population (N = 600) participated in the interview, while 38% consented to HIV-testing. Odds ratio analysis techniques were used to compare consenters and non-consenters. More women than men participated in the interview (OR = 4.23), and participating men were younger than non-participating men (29.1 vs 31.9 years). Subjects who had attended secondary school were underrepresented among the HIV- test consenters (OR = 0.40). No other demographic or AIDS-related differences were observed between HIV-test consenters and non-consenters. At least in this pilot, non- consenters did not appear to be at any higher risk for HIV-infection than the consenters. Future population-based HIV-surveys might have to compromise on maximizing participation rate in order to secure informed, non-coerced consent from participants.

PIP: Tanzania is one of the countries in East Africa most severely affected by the HIV/AIDS epidemic. Population-based HIV surveys have been given high priority by the research community in Tanzania because of their ability to provide information on the progression of the epidemic. One major methodological problem with such surveys, however, is the potential for low participation rates which may, in turn, challenge the validity of the data. The authors compared the demographic and AIDS-related characteristics of people who consented to HIV testing as part of a population survey with those who did not consent to such testing. One ward of Arusha was approached in 1991 through the local government and party leaders in which 360 men and 378 women aged 15-54 years were initially registered as eligible to participate in the study. 138 of the individuals had moved, were traveling outside of town for the duration of the study, or had died, and were therefore ineligible to participate in the study. 372 of the remaining 600 individuals consented to participate in a structured interview; 77.8% of the women and 43.9% of men. Only eight people actually refused to be interviewed. 148 others who were not interviewed had work obligations which had them away from home, while information is lacking on reasons for the nonparticipation of 72 subjects. Men who consented to be interviewed were of mean age 29.1 years compared to non-consenters who were of mean age 31.9 years. 227 of the subjects interviewed gave their consent to be tested for HIV-1. Odds ratio analysis techniques were used to compare consenters and non-consenters. In addition to the sex and age differences between the two groups, subjects with secondary or higher education were more likely to refuse HIV testing than subjects with less formal education. No other demographic or AIDS-related differences were observed between HIV test consenters and non-consenters. Non-consenters did not appear to be at any higher risk for HIV infection than consenters. These study results indicate the existence of a potential conflict between achieving a high participation rate and obtaining informed noncoerced consent for participation in population-based HIV surveys. This problem may be increasing as populations become more aware of the HIV/AIDS epidemic and its consequences.