Recognition of AIDS by health personnel in rural south-Rwanda.

The ability of rural health personnel to recognise AIDS related symptoms and signs according to the WHO clinical case definition (CCD) and its modified Rwandan version was tested in 4141 clinically suspected cases in South-Rwanda. The sensitivities of these CCDs for AIDS in adults were 33% (36%), the specificities 78% (76%), and the positive predictive values (ppv) 46%. For AIDS in children the sensitivities of the CCDs were 13% (16%), the specificities 94% (90%), and the ppv 44% (38%). While the specificities did not differ from those found in studies conducted by trained physicians, the low sensitivities and predictive values demand improvement of the training of the health personnel to diagnose AIDS related symptoms and signs, especially where laboratory tests are not available.

PIP: In order to test the ability of health care workers in rural Rwanda to recognize the signs and symptoms of AIDS according to the World Health Organization clinical case definition (CCD) and to the slightly modified Rwandan version, serum samples and completed questionnaires were collected from all 4141 clinically-suspected cases of AIDS from 62 rural health centers and hospitals in Butare province during 1991. Of the 3669 patients older than 12 years, 964 met the CCD for adults, and 444 were seropositive. 1040 met the modified Rwandan version of the CCD, and 482 were seropositive. Of the 472 children, 35 met the CCD for children, with 16 seropositive. 53 met the Rwandan CCD, and 20 of these were seropositive. Sensitivities of the CCD and the Rwandan CCD, respectively, were 33 and 36% in adults and 13 and 16% in children. The respective specificities were 78 and 76% in adults and 94 and 90% in children. The positive predictive values were, therefore, 46% in adults and 44 and 38%, respectively, in children. These low sensitivity and positive predictive values will result in many cases of AIDS going undiagnosed. Whereas no single symptom had a sensitivity and specificity high enough to be used for diagnoses, an episode of herpes zoster is highly indicative of AIDS. Since this diagnosis is unequivocal, this strong association may prove useful in the diagnosis of AIDS. Because laboratory tests will remain unavailable in certain settings, health personnel must receive proper training to improve their ability to diagnose AIDS following the clinical case definition.

HIV-1 infection among malnourished children in Butare, Rwanda.

In order to investigate the relationship between human immunodeficiency virus (HIV-1) infection and protein-energy malnutrition (PEM), all 101 malnourished children who were admitted to the Department of Pediatrics of the National University Hospital between February and July of 1989 (median age = 2.5 years), and who were accompanied by their mother were screened for HIV-1 antibody. Mothers were also screened and interviewed. Mother-child pairs were followed-up 2 years later to determine mortality and clinical status. Fourteen per cent of malnourished children were HIV-1 seropositive. Only one seropositive child had a seronegative mother. This child had a history of multiple blood transfusions and injections. Among children above 15 months of age, HIV-1 seropositivity was more common among marasmic children than among malnourished children presenting with oedema at admission to the hospital. Also, HIV-1 infection was found more frequently among chronically malnourished children (low height for age and weight for age) than among acutely malnourished children (low weight for height). Mortality during the 2-year follow-up was 75 per cent among HIV-1 seropositive children and 23 per cent among HIV-1 seronegatives (mortality density ratio = 6.2; 95 per cent confidence interval = 2.2-17.4). Severe, chronic PEM should always alert health workers to the possible diagnosis of pediatric AIDS, and its implications for treatment and prognosis.