Bloody diarrhoea of adults in Malawi: clinical features, infectious agents, and antimicrobial sensitivities.

In a prospective study, 132 hospital out-patients presenting with bloody diarrhoea ('cases') were evaluated in Malawi, Central Africa; 73 out-patient tuberculosis suspects acted as controls. Most (100/132, 76%) subjects reported an illness lasting < or = 5 d with > 5 bowel actions in the preceding 12 h; 39/132 (30%) reported use of systemic antimicrobial drugs in the preceding week; 57% (74/130) had a body mass index < 20; 4% (5/131) were febrile; and 18/130 (13%) had one or more sign(s) of dehydration. The 73 controls reported no diarrhoea and more systemic antimicrobial drug use (P = 0.0003), but were otherwise comparable to the subjects. All stool samples from controls and 38/124 (31%) from cases were macroscopically normal. Only 32% (40/124) of the cases had blood visible in the stool. Parasitic gut infections were found in 42/124 (34%) cases compared with 1/60 (2%) controls (P < 0.0001). The commonest parasite was Schistosoma mansoni. Bacterial cultures were positive in 32/124 (26%) of the subjects. Shigella dysenteriae (Sd) 1 accounted for 53% (17/32) of these. All bacterial isolates were sensitive in vitro to nalidixic acid and ciprofloxacin, while only 18% were sensitive to cotrimoxazole. Sd 1 with significant antimicrobial resistance continues to cause seasonal epidemics of dysentery in Malawi. During these, approximately two-thirds of patients presenting with bloody diarrhoea have no blood visible in the stool. Nalidixic acid remains the drug of choice but its use should be restricted to patients at greatest risk of complicated shigellosis.