BackgroundVital registration and cause of death reporting is incomplete in the countries in which the HIV epidemic is most severe. A reliable tool that is independent of HIV status is needed for measuring the frequency of AIDS deaths and ultimately the impact of antiretroviral therapy on mortality.Methods and FindingsA verbal autopsy questionnaire was administered to caregivers of 381 adults of known HIV status who died between 1998 and 2003 in Manicaland, eastern Zimbabwe. Individuals who were HIV positive and did not die in an accident or during childbirth (74%; n = 282) were considered to have died of AIDS in the gold standard. Verbal autopsies were randomly allocated to a training dataset (n = 279) to generate classification criteria or a test dataset (n = 102) to verify criteria. A rule-based algorithm created to minimise false positives had a specificity of 66% and a sensitivity of 76%. Eight predictors (weight loss, wasting, jaundice, herpes zoster, presence of abscesses or sores, oral candidiasis, acute respiratory tract infections, and vaginal tumours) were included in the algorithm. In the test dataset of verbal autopsies, 69% of deaths were correctly classified as AIDS/non-AIDS, and it was not necessary to invoke a differential diagnosis of tuberculosis. Presence of any one of these criteria gave a post-test probability of AIDS death of 0.84.ConclusionsAnalysis of verbal autopsy data in this rural Zimbabwean population revealed a distinct pattern of signs and symptoms associated with AIDS mortality. Using these signs and symptoms, demographic surveillance data on AIDS deaths may allow for the estimation of AIDS mortality and even HIV prevalence.

Histopathological and molecular studies suggest that different histological subtypes (histotypes) of ovarian cancer have different aetiologies. Few studies have been large enough to explore reliably the effect of tubal ligation (sterilization), which has been associated with a reduced overall risk of ovarian cancer, on different tumour histotypes. In a prospective study of 1.1 million UK women without prior cancer or bilateral oophorectomy, 8,035 ovarian cancers occurred during mean follow‐up of 13.8 years. Using a Cox proportional hazards model, we estimated adjusted relative risks of ovarian cancer associated with tubal ligation. Overall, there was substantial heterogeneity in tumour risk associated with tubal ligation for the four main histotypes, serous, endometrioid, mucinous and clear cell (heterogeneity: p < 0.0001). For serous tumours, the most common histotype (n = 3,515), risks differed significantly between high‐grade (RR: 0.77, 95% CI: 0.67–0.89) and low‐grade tumours (RR: 1.13, 95% CI: 0.89–1.42); heterogeneity: p = 0.007. Relative risks were almost halved for endometrioid (n = 690, RR: 0.54, 95% CI: 0.43–0.69) and clear cell tumours (n = 401, RR: 0.55, 95% CI: 0.39–0.77), but there was no association between tubal ligation and mucinous tumours (n = 836, RR: 0.99, 95% CI: 0.84–1.18). For the main tumour histotypes we found little variation of risk by timing of tubal ligation. The significant differences by tumour histotype are unlikely to be due to confounding and are consistent with hypotheses that high‐grade and low‐grade serous tumours have different origins, and that some endometrioid and clear cell tumours might arise from cells and/or carcinogens travelling through the fallopian tubes.

Ovarian cancer risk is known to be reduced amongst women who have had children, but reported associations with breastfeeding are varied. Few studies have had sufficient power to explore reliably these associations by tumour histotype. In a prospective study of 1.1 million UK women, 8719 developed ovarian cancer during follow‐up. Cox regression yielded adjusted relative risks (RRs) overall and by tumour histotype amongst women with different childbearing patterns. Nulliparous women had a 24% greater ovarian cancer risk than women with one child, with significant heterogeneity by histotype (p = 0.01). There was no significant increase in serous tumours, a modest increase in mucinous tumours, but a substantial increase in endometrioid (RR = 1.49, 95% CI: 1.18‐1.89) and clear‐cell tumours (RR = 1.68, 1.29‐2.20). Among parous women, each additional birth was associated with an overall 6% reduction in ovarian cancer risk; this association also varied by histotype (p = 0.0006), with the largest reduction in risk for clear‐cell tumours (RR per birth = 0.75, 0.65‐0.85, p < 0.001) and weak, if any, effect for endometrioid, high‐grade serous, or mucinous tumours. We found little association with age at first or last birth. There was about a 10% risk reduction per 12‐months breastfeeding (RR = 0.89, 0.84‐0.94, p < 0.001), with no significant heterogeneity by histotype, but statistical power was limited. In this large prospective study, ovarian cancer risk associated with parity varied substantially by tumour histotype. Nulliparity was associated with a substantially greater overall risk than expected from the effect of a single birth, especially for clear cell and endometrioid tumours, perhaps suggesting that infertility is associated with these histotypes.