Testing of interventions aimed

at early empirical management of sepsis is limited by a lack of validated criteria for making syndromic diagnosis of sepsis in these settings. Here we show that using a modification of international criteria,6, 8, 23 and 24 amongst adults with a clinical suspicion of severe infection, patients with a diagnosis of sepsis, and those at highest risk of death, can be identified. Reduced systolic blood pressure, reduced percentage oxygen saturation and a low haemoglobin were independent risk factors for death amongst the whole cohort, with male sex, decreased temperature, reduced GCS, reduced haemoglobin and increased respiratory rate being predictive of severe sepsis; prospective validation of these factors, which are line with observations made in a similar adult population in Uganda,4 may enable development of a locally applicable risk stratification tool. We found that the mortality MEK inhibitor from severe sepsis was 50% compared with 17% for patients with sepsis. This mortality is higher than that

reported from industrialised countries where critically ill patients are predominantly treated in intensive care units12 and 25 and is likely to have been even higher had we had access to outcome data at 30 days.4 Almost two thirds of the cohort presented with sepsis as their index presentation for HIV and notably, most of the patients with known positive HIV status were Selleck Cyclopamine already on ART. Having been on ART for more than 90 days reduced mortality by almost two thirds compared with those who had commenced treatment more recently; of these, two thirds of those with sepsis, and three quarters of those with severe sepsis died, a mortality rate similar to that of their ART naive counterparts. This is likely to have a considerable impact

in Malawi and in similar countries. CHIR-99021 concentration Malawi has an estimated national HIV seroprevalence of 12%26 and since 2004, has benefitted from large scale antiretroviral therapy (ART) rollout, together with widespread implementation of cotrimoxazole prophylaxis.27 By the end of 2009, 271,105 HIV-infected individuals had been registered on the national ART programme of whom 73% remained alive and on treatment.28 Nonetheless our data suggest that in addition to the time needed for immune restitution, sepsis contribute to this adverse outcome, in addition to intercurrent malnutrition and anaemia,29 and unmasking of underlying TB.30 Previous studies in Malawi have demonstrated superadded bacterial infections to be a major cause of death during the initial two months of TB therapy and that septic patients with Mycobacterium tuberculosis bacteraemia have a high in-hospital mortality. 31, 32, 33 and 34Thus mycobacterial infection may have contributed to a poor outcome in our cohort but were unable to determine this with any certainty.