This was a single-center double-blind randomized placebo-controlled trial on the use of mesalazine alongside standard medical and nutritional care in severely acutely malnourished children. Randomization was balanced 1:1.
Participants and setting
The study took place between June and November 2013 at the nutrition clinic of Baraka Health Centre, Mathare, Nairobi. The Baraka Health Centre (BHC) is run by ‘German Doctors’, a registered non-governmental organization, and provides free healthcare to children under five years old. Mathare is the second largest urban slum in Kenya, home to at least 200,000 people who mostly live in overcrowded iron-sheet housing with limited access to safe water and sanitation facilities .
Participants were recruited from among those self-presenting to BHC or via a program of active case-finding in the community conducted by local community health workers (CHWs). Eligible participants were children 12- to 60-months old with SAM, defined for the purposes of this study as mid-upper arm circumference (MUAC) <11.5 cm or bilateral pedal edema. They had uncomplicated SAM (that is, good appetite and no severe clinical illness) and were, therefore, eligible for outpatient management according to World Health Organization (WHO) guidelines . EED was inferred on the basis of stunting (height-for-age z-score < −2 using the WHO 2006 growth standards) and chronic inflammation (erythrocyte sedimentation rate (ESR) >20 mm/hour). Children were excluded if they had any of the following: HIV-infection, tuberculosis, bloody diarrhea, biochemical evidence of renal or hepatic impairment, thrombocytopenia or severe anemia. Children were also excluded if they were already receiving treatment for SAM from another center, if they had medical difficulties precluding normal feeding (for example, severe cerebral palsy), if they had known pre-existing renal disease, asthma, hypersensitivity to salicylates or if they were on medication known to interfere with the action of the study drug. The exclusion criteria were decided on the basis of contraindications listed in the Summary of Product Characteristics, pre-existing conditions that the investigator group felt increased the risk to participants (for example, HIV infection, bloody diarrhea, other overt infection requiring hospital admission), or likely futility in the presence of other major medical problems (for example, tuberculosis, cerebral palsy) . We did not consider concurrent or recent viral infection or administration of a live viral vaccine to be a contraindication to administration. Although Reye’s syndrome has historically been associated with salicylate (mainly aspirin) use in these circumstances, we were unable to find even a single report of Reye’s associated with mesalazine, and such cautions are not advised when it is used in the context of inflammatory bowel disease (IBD).
Screening, enrollment and randomization
Children, 12- to 60-months old, with SAM and stunting were regarded as potentially eligible and were referred to the study team for screening. If clinical eligibility was confirmed and informed consent for participation was provided by the child’s parent or guardian, venous blood was taken for HIV testing, full blood count, ESR, creatinine, liver function tests, and film for malaria parasites, if required. Final screening with blood results took place the following day, after which eligible children were enrolled by assigning the next consecutive study number.
A randomization schedule was developed in STATA (version 12.0) with variable block sizes (two, four and six) using the following code: ‘ralloc blknum blksiz Rx, nsubj(44) osize(3) ntreat(2) saving(mys) table’ . Allocations were assigned to study numbers by the trial statistician (GF). Sachets of mesalazine (Pentasa) granules and matched placebo were purchased from Ferring Pharmaceuticals (Saint-Prex, Switzerland) in 2 g foil sachets that were identical apart from labelling. Prior to initiation of the trial, sachets were disguised by application of opaque ‘black-out’ labels (Avery) and re-labelled (with the study number) by pharmacy staff independent of the trial team, according to the randomization schedule.
Dosing of drug and placebo was performed on the basis of weight, which required the contents of the 2 g sachets to be separated into smaller individual doses. Pharmacy technicians at BHC were trained to dispense the study drug using an electronic fine-scale balance (TX-323 L, Shimadzu). Doses were packed in foil pouches (purchased locally), heat-sealed to render them impermeable to light and air and labelled with the participant’s initials, study number and date. The study drug was dispensed weekly in order to minimize deterioration of the active product due to repackaging and account for changes in participants’ weight. Active and placebo granules were indistinguishable.
Participants were prescribed 30 mg/kg/day of mesalazine or placebo in three divided doses for the first seven days. Then, if tolerated (see below), the dosage was increased to 45 mg/kg/day for a further 21 days. They were followed up for a further 28 days after stopping the study drug (56 days total). Blood and stool were collected at days 7, 28 and 56. To account for excipients the dose of granules prescribed was 11 mg/kg three times daily for the first week, followed by 16.5 mg/kg three times daily. Because dispensing exactly to the milligram was not possible with a granular product, technicians dispensed in the range of the prescribed dose to prescribed dose + 5 mg.
If recognized side-effects of mesalazine occurred in the first week or the day 7 blood tests indicated deterioration in renal or hepatic function or blood dyscrasia (Grade 1 or 2 toxicity), the dosage was maintained at 30 mg/kg/day without unblinding. Such children were re-assessed after a week and their dose was escalated if or when it appeared to be safe to do so. The study drug was to be discontinued in the event of Grade 3 or 4 toxicity. Toxicity grades for biochemical indices were defined according to the US Division of Microbiology and Infectious Diseases’ Pediatric Toxicity Tables, 2007 . Carers were asked to withhold the study drug if the child developed diarrhea, blood in stools or unexplained bruising, and to bring the child for assessment as soon as possible. The study drug was suspended until diarrheal episodes had resolved.
All children received nutritional rehabilitation with ready-to-use therapeutic foods (RUTF) conforming to WHO/UNICEF standards until they were nutritionally cured of SAM (MUAC >11.5 cm and no edema at two consecutive weekly visits), alongside a seven-day course of amoxicillin and deworming with mebendazole or albendazole according to Kenyan national guidelines .
Primary outcomes were frequency of adverse events and compliance with the intervention, assessed via interview with the carer and weekly counting of full/empty sachets. Stool frequency and consistency was assessed at each study visit using a Kiswahili translation of the Bristol Stool Form Scale [see Additional file 1] . Secondary outcomes were time to recovery, growth and a panel of inflammatory markers (see below).
Blood tests designated for safety (full blood count, ESR, C-reactive protein, creatinine, alanine transaminase (ALT), aspartate aminotransferase (AST), gamma-glutamyl transpeptidase (GGT), bilirubin) and stool for microscopy were processed in a commercial good clinical laboratory practice (GCLP)-accredited laboratory in Nairobi (Pathologists Lancet Kenya) and results were provided the following day.
Plasma, serum and stool were kept on ice prior to freezing at −80°C pending batched enzyme-linked immunosorbent assay (ELISA) processing. The following were tested by ELISA following the manufacturers’ recommendations: fecal calprotectin (Bühlmann Laboratories AG (Schönenbuch, Switzerland) kits on stool following disruption by 30 seconds TissueLyser shaking (QIAGEN (Hilden, Germany), no beads)), plasma anti-endotoxin core immunoglobulin G (IgG EndoCAb, Hycult Biotech (Uden, The Netherlands)), plasma interferon-γ (Ebioscience (San Diego, California, USA)) and serum insulin-like growth factor-1 (R&D Systems (Minneapolis, Minnesota, USA)). Muliplex ELISA (Luminex (Austin, Texas, USA) MAGPIX system) was performed against the following targets in plasma: Eotaxin (chemokine (C-C motif) ligand (CCL)-11), GROα (growth regulated oncogene-α, chemokine (C-X-C motif) ligand (CXCL)-1), interferon-α, interleukin (IL)-1α, IL-1 receptor antagonist (IL-1RA), IL-7, IL-8, IL-10, IL-15, IL-17a, IL-22, IL-31, IP-10 (interferon-γ induced protein-10, CXCL10), MCP-1 (monocyte chemotactic protein-1, CCL2), MIP-1α (macrophage inflammatory protein-1α, CCL3), MIP-1β (CCL4), SDF-1 (stromal cell-derived factor-1, CXCL12) and tumor necrosis factor-β (TNFβ) (Ebioscience). Soluble CD14 (sCD14) was measured using an in-house ELISA (capture clone 55–3, detection clone 3-C39 both from BD (Franklin Lakes, New Jersey, USA), recombinant standard from Sigma-Aldrich Limited (Gillingham, United Kingdom)). Serum endotoxin was measured using HEK-Blue Endotoxin Detection Kit (Invivogen (San Diego, California, USA)): This test relies on a HEK293 cell line that has been stably transfected with Toll-like receptor 4 pathway genes and a secretory alkaline phosphatase that is transcriptionally regulated by NF-κB. Heat-inactivated sera (90°C for 30 minutes) were incubated in duplicate with cells for 24 hours in the presence of a detection reagent. Absorption at 620 nm was read against a standard curve.
Analysis was performed in STATA Version 12.0. We performed Mann–Whitney U tests or t-tests on log-transformed data between the arms at each time point. Growth in height and MUAC across the trial was calculated for individual participants in mm/day and compared using Mann–Whitney U tests. Fisher’s exact test was used to compare grouped variables. Comparison of the timing of adverse events was done using logrank. Raw ELISA data was analyzed in Graphpad Prism 6.0 prior to import into STATA and z-scores were calculated using WHO Anthro Version 3.2.2 STATA macros. Analyses were performed by intention to treat, except for blood and stool laboratory analyses, which were performed using data from all available specimens (that is, not including data missed due to withdrawal or failure to get specimens). For this pilot study, no cut-off was considered to indicate ‘statistical significance’, and P-values are provided throughout. Because the secondary analyses were intended to be exploratory and hypothesis-generating, post-hoc correction for multiple comparisons was not performed.
Sample size was set at 22 in each arm with reference to norms in Phase I and early Phase II research. No sample size calculation was performed and the study was not powered to formally address any outcomes at any given significance level.
All participants enrolled in the study had individual written informed consent provided by a parent or guardian. The study was approved by the Kenya Medical Research Institute (KEMRI) Ethical Review Committee, the Imperial College, London, Ethical Review Committee, and the Kenya Pharmacy & Poisons Board prior to initiation. Imperial College, London, was the sponsor. Clinical trials monitoring was performed by staff from the Clinical Trials Facility, KEMRI-Wellcome Trust Research Programme. An independent Data Safety and Monitoring Board (DSMB) was established and an independent consultant pediatrician acted as local safety monitor. Neither the sponsor nor any other party except the named investigators had any role in the design of the study, interpretation of the results, content of manuscripts or decision to publish. The trial was registered at http://clinicaltrials.gov/ct2/show/NCT01841099.