Co-trimoxazole or multivitamin multimineral supplement for post-discharge outcomes after severe anaemia in African children: a randomised controlled trial

Abstract

Background: Severe anaemia is a leading cause of paediatric admission to hospital in Africa; post-discharge outcomes remain poor, with high 6-month mortality (8%) and re-admission (17%). We aimed to investigate post-discharge interventions that might improve outcomes.

Methods: Within the two-stratum, open-label, multicentre, factorial randomised TRACT trial, children aged 2 months to 12 years with severe anaemia, defined as haemoglobin of less than 6 g/dL, at admission to hospital (three in Uganda, one in Malawi) were randomly assigned, using sequentially numbered envelopes linked to a second non-sequentially numbered set of allocations stratified by centre and severity, to enhanced nutritional supplementation with iron and folate-containing multivitamin multimineral supplements versus iron and folate alone at treatment doses (usual care), and to co-trimoxazole versus no co-trimoxazole. All interventions were administered orally and were given for 3 months after discharge from hospital. Separately reported randomisations investigated transfusion management. The primary outcome was 180-day mortality. All analyses were done in the intention-to-treat population; follow-up was 180 days. This trial is registered with the International Standard Randomised Controlled Trial registry, ISRCTN84086586, and follow-up is complete.

Findings: From Sept 17, 2014, to May 15, 2017, 3983 eligible children were randomly assigned to treatment, and followed up for 180 days. 164 (4%) were lost to follow-up. 1901 (95%) of 1997 assigned multivitamin multimineral supplement, 1911 (96%) of 1986 assigned iron and folate, and 1922 (96%) of 1994 assigned co-trimoxazole started treatment. By day 180, 166 (8%) children in the multivitamin multimineral supplement group versus 169 (9%) children in the iron and folate group had died (hazard ratio [HR] 0·97, 95% CI 0·79-1·21; p=0·81) and 172 (9%) who received co-trimoxazole versus 163 (8%) who did not receive co-trimoxazole had died (HR 1·07, 95% CI 0·86-1·32; p=0·56). We found no evidence of interactions between these randomisations or with transfusion randomisations (p>0·2). By day 180, 489 (24%) children in the multivitamin multimineral supplement group versus 509 (26%) children in the iron and folate group (HR 0·95, 95% CI 0·84-1·07; p=0·40), and 500 (25%) children in the co-trimoxazole group versus 498 (25%) children in the no co-trimoxazole group (1·01, 0·89-1·15; p=0·85) had had one or more serious adverse events. Most serious adverse events were re-admissions, occurring in 692 (17%) children (175 [4%] with at least two re-admissions).

Interpretation: Neither enhanced supplementation with multivitamin multimineral supplement versus iron and folate treatment or co-trimoxazole prophylaxis improved 6-month survival. High rates of hospital re-admission suggest that novel interventions are urgently required for severe anaemia, given the burden it places on overstretched health services in Africa.

Funding: Medical Research Council and Department for International Development.

Figure 1

Trial profile All children for whom co-trimoxazole prophylaxis should have been prescribed according to WHO or national guidelines (eg, HIV-infected children) received it regardless of randomisation.

Figure 2

Mortality and re-admissions through 180 days Mortality (A) and re-admission (C) in the multivitamin multimineral supplement and iron and folate groups, and mortality (B) and re-admission (D) in the co-trimoxazole and no co-trimazole groups. *These deaths occurred between timepoints on the x-axis. † Considered to have attended the 180-day visit as seen within the 120–240-day visit window: 3459 (99%) of 3484 seen at day 170 or later.

Figure 3

Proportions of children with haemoglobin concentrations of less than 6 g/dL, 6–9 g/dL, and more than 9 g/dL over 180 days in multivitamin multimineral supplement and iron and folate groups (A) and co-trimoxazole and no co-trimazole groups (B) Numbers can be found in the appendix (p 38).