ABSTRACT.
Azithromycin is a promising alternative to amoxicillin in the management of uncomplicated severe acute malnutrition (SAM) as it can be administered as a single dose and has efficacy against several pathogens causing infectious disease and mortality in children under 5. In this pilot trial, we aimed to establish the feasibility of a larger randomized controlled trial and provide preliminary evidence comparing the effect of azithromycin to amoxicillin on weight gain in children with uncomplicated SAM. We enrolled children 6–59 months old with uncomplicated SAM at six healthcare centers in Burkina Faso. Participants were randomized to a single dose of azithromycin or a 7-day course of amoxicillin and followed weekly until nutritional recovery and again at 8 weeks. Apart from antibiotics, participants received standard of care, which includes ready-to-use therapeutic food. Primary feasibility outcomes included enrollment potential, refusals, and loss to follow-up. The primary clinical outcome was weight gain (g/kg/day) over 8 weeks. Outcome assessors were masked. Between June and October 2020, 312 children were screened, 301 were enrolled with zero refusals, and 282 (93.6%) completed the 8-week visit. Average weight gain was 2.5 g/kg/day (standard deviation [SD] 2.0) in the azithromycin group and 2.6 (SD 1.7) in the amoxicillin group (mean difference −0.1, 95% CI −0.5 to 0.3, P = 0.63). Fewer adverse events were reported in the azithromycin group (risk ratio 0.50, 95% CI 0.31–0.82, P = 0.006). With strong enrollment and follow-up, a fully powered trial in this setting is feasible.
INTRODUCTION
Severe acute malnutrition (SAM) is estimated to impact 17 million children under 5 and is responsible for 7% of mortality in this age group.1,2 Infectious disease is common in children with SAM, who face an increased risk of mortality from diarrhea, measles, and pneumonia.3,4 However, identifying infection in this population can be challenging, since SAM suppresses the immune response, resulting in poor correlation between clinical signs and the presence of infection.4–6 Given this potential for asymptomatic infection, the WHO recommends that children with uncomplicated SAM be treated with a broad-spectrum oral antibiotic in addition to ready-to-use therapeutic food (RUTF).7
Many national nutritional guidelines recommend oral amoxicillin in the management of uncomplicated SAM, but evidence on the role of amoxicillin in this population is mixed. A trial in Malawi randomized children to amoxicillin, cefdinir, or placebo for 7 days and found that both antibiotics increased nutritional recovery and decreased mortality relative to placebo.8 Another trial in Niger compared a 7-day course of amoxicillin to placebo and did not find a difference in nutritional recovery or mortality between treatment groups.9 Children receiving amoxicillin in the Niger trial were less likely to be transferred to inpatient care and experienced a shorter time to recovery compared with placebo.9 Both trials found that children who received antibiotics experienced increased weight gain compared with children receiving placebo. Meta-analyses of the amoxicillin groups from these trials have produced confidence intervals consistent with no effect of amoxicillin on recovery (lower bound of 1.00) as well as up to a 6% increase in recovery with amoxicillin compared with placebo (upper bounds of 1.05 and 1.06).10,11 Another meta-analysis pooled both antibiotics examined in the two studies and estimated a 6% increase in recovery in the groups receiving any antibiotic compared with placebo.12 The difference in results in these two trials may indicate that antibiotics may not be required for uncomplicated SAM in some settings or that antibiotics other than amoxicillin should be considered.
Azithromycin may provide several advantages over amoxicillin in the management of uncomplicated SAM. A cluster-randomized trial in Niger demonstrated a substantial reduction in mortality with a single dose of azithromycin at 20 mg/kg in children without established infection.13 A subgroup analysis of this trial suggested that the administration of a single dose of azithromycin to severely malnourished children may reduce the difference in mortality between malnourished children and their well-nourished peers.14 In addition, azithromycin has a long half-life, which may decrease the burden of infectious organisms for several weeks.15 Indeed, the reported duration of population-level reductions in infectious disease burden ranges from 2 weeks to 6 months after a single azithromycin distribution.16–18 With the potential for longer term protection, azithromycin could be administered during routine SAM outpatient visits as a single dose, removing the need to rely on caregivers to administer multiple doses.19 Finally, amoxicillin is more commonly used than macrolide antibiotics for routine treatment in many settings experiencing high malnutrition.20,21 As such, the potential for the selection for antimicrobial resistance and thus reduced efficacy may be greater with amoxicillin compared with azithromycin.22
In this pilot randomized trial, we aimed to establish the feasibility of conducting a full-scale randomized controlled trial to evaluate different antibiotics for uncomplicated SAM in this setting. In addition, we aimed to gather preliminary evidence comparing the effect of azithromycin to amoxicillin on weight gain and nutritional recovery in the management of uncomplicated SAM in Burkina Faso. We hypothesized that children randomized to receive azithromycin will have both increased weight gain and nutritional recovery compared with amoxicillin 8 weeks after admission to the nutritional program.
MATERIALS AND METHODS
Design overview.
This individual-randomized trial was designed to establish the feasibility of a larger trial and to provide preliminary evidence on the effect of azithromycin compared with amoxicillin on weight gain and nutritional recovery in the management of uncomplicated SAM in children 6–59 months old in Burkina Faso. Eligible children presenting to nutritional programs in the study area were randomized to a single dose of azithromycin or a 7-day course of amoxicillin upon admission. All children received standard outpatient treatment of uncomplicated SAM according to Burkina Faso guidelines.23 Participants were followed weekly at routine follow-up visits until nutritional recovery and all participants were asked to return for a final study visit at 8 weeks. The study protocol has been previously published.24
Patient involvement.
Patients were not involved in the design or conduct of this research.
Study setting and participants.
Participants were enrolled at six Center de Santé et de Promotion Sociale (CSPS) with nutritional programs for SAM from a catchment area of 54 communities in Boromo District, Burkina Faso. Center de Santé et de Promotion Sociales are government-run primary health facilities that provide basic preventive care and treatment. All children under 5 years of age receive care free of charge subsidized by the government. The majority of communities in the catchment area are rural, agricultural, and impacted by seasonal food insecurity which occurs from roughly August to October each year.
Each enrollment site screened children presenting to the CSPS for eligibility. Community health workers also screened children in the study catchment areas for SAM using mid-upper arm circumference (MUAC), though this screening was not tracked and these children were also officially screened for study eligibility after presenting to the CSPS. Children 6–59 months old living in the study area were eligible for enrollment if they presented with: a weight-for-height Z-score (WHZ) < −3 or MUAC < 11.5 cm; no antibiotic use or SAM treatment in the prior 7 days; no clinical complications requiring antibiotic or inpatient treatment, including bipedal edema; no congenital abnormality or chronic illness; sufficient appetite; and written informed consent.
Randomization and masking.
The randomization sequence was generated by the University of California, San Francisco (UCSF) data team in R (R Foundation for Statistical Programming, Vienna, Austria) and implemented through an electronic data collection system. After a baseline assessment, enrolled children were randomized 1:1 to azithromycin or amoxicillin without stratification or blocking. Participants and study personnel administering treatment were not masked to treatment assignment given logistical constraints and the lack of placebo. Outcome data collectors were masked to treatment assignment. Allocation concealment from outcome assessors was achieved by restricting access to the randomization assignment and ensuring that a separate masked team member collected outcome data.
Interventions.
Azithromycin was administered as a single directly observed weight-based dose (20 mg/kg) at the time of enrollment (Azithrin oral suspension 200 mg/5 mL, Strides Shasun Ltd, Bangalore, India). Amoxicillin was administered as a 7-day course at a dosage of 80 mg/kg of body weight per day divided into two daily doses (Amoxicillin syrup 250 mg/5 mL, Reyoung Pharmaceutical, Shandong, China) according to national guidelines.23 Center de Santé et de Promotion Sociale study personnel administered and directly observed the first dose of amoxicillin at enrollment and caregivers administered remaining doses at home. Apart from antibiotics, all enrolled children received standard outpatient treatment of uncomplicated SAM according to the guidelines of the government of Burkina Faso,23 which include provision of RUTF (Plumpy’Nut, Nutriset, Malaunay, France), antimalarials if positive for malaria by rapid diagnostic test, antiparasitics, missing vaccinations, and vitamin A supplementation. Each RUTF sachet contains 500 kcal, 12.8 g of protein, 30.3 g of lipids, 45 g of carbohydrates, and micronutrients.
During the course of the trial, Burkina Faso faced a national shortage of RUTF, which impacted some participants in this trial. The stockouts were reported to the DSMC and institutional review boards in both countries. As country-wide stockouts of RUTF are a regular occurrence in this setting, the standard of care stipulates that cooking demonstrations should be made to caregivers when RUTF is unavailable. Caregivers of children enrolled during a period of stockout received cooking demonstrations and nutritional advice per national guidelines.
Caregivers were instructed to report adverse events experienced within 7 days of enrollment. Follow-up visits included a survey of adverse events, including fever, diarrhea, vomiting, abdominal pain, skin rash, and constipation. Serious adverse events were defined as death, hospitalization, or any other life-threatening situation and were reported to the Medical Monitor within 24 hours to determine whether the event was possibly related to the study drug.
Participant timeline and data collection.
At enrollment, the baseline assessment involved a questionnaire on socioeconomic status and feeding practices in addition to anthropometric assessments, including height (ShorrBoard Infant/Child Measuring Board, Weigh and Measure, LLC, Olney, MD), weight (Seca 874dr scale, Seca, Germany), and MUAC (Shorr Child MUAC tape, Weigh and Measure, LLC). Height was measured to the nearest 0.1 cm, weight was measured to the nearest 0.1 kg, and MUAC was measured to the nearest 1 mm. All anthropometric measurements were taken in triplicate and the median was used for analyses. The scale was calibrated daily before study activities using a 2 kg test weight.
Children were followed weekly (±2 days) at routine follow-up visits until nutritional recovery and again at 8 weeks (−5 days, +21 days) after enrollment. Anthropometry, vital status, adverse events, and clinical examination outcomes were recorded at each visit. Data on malaria were collected and reported separately.25 All data were collected electronically using a mobile application (CommCare by Dimagi, Cambridge, MA) and uploaded to a secure, cloud-based server.
Outcomes.
The primary feasibility outcomes included 1) enrollment potential defined as average number of participants enrolled per site per day, 2) refusals calculated as the percentage of eligible participants refusing to participate, and 3) loss to follow-up defined as the percentage of enrolled participants with incomplete follow-up.
The primary clinical outcome was weight gain defined as grams per kilogram per day (g/kg/day) over the 8-week study period. Secondary outcomes included nutritional recovery by 8 weeks (WHZ ≥ −2 and/or MUAC ≥ 125 mm on two consecutive visits, based on the criterion used for enrollment), nonresponse at 8 weeks, transfer to inpatient care, mortality, clinical signs of infection, height-for-age Z-score (HAZ), MUAC, weight-for-age Z-score (WAZ), and WHZ. Anthropometric outcomes were defined according to the 2006 WHO Child Growth Standards.26 Sensitivity analyses included the use of alternative definitions for nutritional recovery. One sensitivity analysis defined nutritional recovery as WHZ ≥ −2 or MUAC ≥ 125 mm on two consecutive visits or WHZ ≥ −2 or MUAC ≥ 125 mm at the final study visit. Another sensitivity analysis compared time to recovery by group.
Sample size and statistical considerations.
As a pilot trial, a sample size of 300 (150 per group) was chosen pragmatically to balance resource constraints against the objectives of the trial. Given this sample size, we estimated having 80% power to detect a 27% increase in weight gain (g/kg/day) in children receiving azithromycin compared with amoxicillin. Assumptions for this sample size calculation were based on estimates from a trial comparing amoxicillin to placebo for children with uncomplicated SAM in Niger.9 These assumptions include an average weight gain of 4.9 g/kg/day over 8 weeks in the amoxicillin group, a standard deviation of 3.9 g/kg/day, loss to follow-up of 10%, and an alpha of 0.05.
Feasibility outcomes were summarized descriptively in aggregate and by group. The primary clinical outcome analysis compared weight gain by group from baseline to 8 weeks using a linear regression model with treatment group as the sole covariate. Similar sensitivity analyses compared period-specific weight gain by group from week 0 to 1, 1 to 2, 2 to 4, and 4 to8. Secondary binary outcomes like nutritional recovery and clinical outcomes were analyzed with log-binomial regression. Modified Poisson models with robust standard errors were used if the log-binomial models failed to converge.27 Other secondary anthropometric outcomes (weight, height, MUAC, WAZ, WHZ, HAZ) were analyzed as continuous variables at all time points and overall using linear regression models adjusted for baseline measurements. Time to event outcomes were compared by group using Cox proportional hazards regression. No adjustments were made for multiple comparisons as all outcomes were prespecified. All analyses were intention-to-treat and conducted in Stata 15.1 (StataCorp, College Station, TX) or R (R Foundation for Statistical Programming, Vienna, Austria).
Ethical approval and trial oversight.
Ethical approval was obtained from the Comité Institutionnel d’Ethique du Center de Recherche en Santé de Nouna and the UCSF, before study activities began. Written informed consent was obtained from a caregiver for the participation of each child. An infectious disease specialist at UCSF served as the Medical Monitor to provide clinical oversight on the study design and to monitor serious adverse events. A DSMC approved the protocol before the study began and monitored study progress, data collection, and adverse events through quarterly progress reports and annual meetings with the study team. The trial was registered at clinicaltrials.gov (NCT03568643). Participants were not provided compensation for participation in the trial.
RESULTS
Between June 3, 2020 and October 9, 2020, 312 children presenting to enrollment sites were screened for eligibility and 301 were enrolled (Figure 1), resulting in an average of more than two participants enrolled overall per day or almost one participant enrolled per site per day. No caregivers refused to have their child participate. Of those enrolled, 161 were randomized to receive a single dose of oral azithromycin and 140 to a 7-day course of oral amoxicillin. Two hundred eighty-two children (93.7%) completed the 8-week follow-up visit and were included in the primary analysis, with 150 (93.2%) in the azithromycin group and 132 (94.2%) in the amoxicillin group. Overall, baseline characteristics were balanced by treatment group (Table 1). In both groups at baseline, 164 (54.5%) of children presented with a WHZ < −3 and 243 (80.7%) presented with MUAC < 11.5 cm.
Figure 1.
Flow of study participants. CONSORT diagram of flow of study participants from enrollment through analysis by treatment group.
Table 1.
Baseline characteristics by treatment group*
| Characteristic | Total (N = 301) | Azithromycin (N = 161) | Amoxicillin (N = 140) |
|---|---|---|---|
| Enrollment site (n, %) | |||
| Boromo | 48 (16.0%) | 27 (16.8%) | 21 (15.0%) |
| Farah | 74 (24.6%) | 42 (26.1%) | 32 (22.9%) |
| Ouahabou | 41 (13.6%) | 19 (11.8%) | 22 (15.7%) |
| Pa | 54 (17.9%) | 26 (16.2%) | 28 (20.0%) |
| Poura | 34 (11.3%) | 18 (11.2%) | 16 (11.4%) |
| Siby | 50 (16.6%) | 29 (18.0%) | 21 (15.0%) |
| Enrollment criterion† | |||
| MUAC < 11.5 cm only | 101 (33.6%) | 62 (38.5%) | 39 (27.9%) |
| WHZ < −3 SD only | 40 (13.3%) | 17 (10.6%) | 23 (16.4%) |
| Both MUAC < 11.5 cm and WHZ < 3 SD | 160 (53.2%) | 82 (50.9%) | 78 (55.7%) |
| Age, months (mean, SD) | 16.7 (8.4) | 16.9 (8.4) | 16.5 (8.2) |
| Female sex (N, %) | 169 (56.2%) | 92 (57.1%) | 77 (55.0%) |
| Currently breastfeeding (N, %) | 212 (70.4%) | 108 (67.1%) | 104 (74.3%) |
| Exclusively breastfeeding (N, %) | 21 (7.0%) | 10 (6.2%) | 11 (7.9%) |
| No. of children under five in the household (mean, SD) | 1.9 (1.6) | 2 (1.7) | 1.9 (1.5) |
| Household water source (N, %) | |||
| Running water | 97 (32.2%) | 50 (31.1%) | 47 (33.6%) |
| Protected well | 121 (40.2%) | 68 (42.2%) | 53 (37.9%) |
| Unprotected well | 29 (9.6%) | 18 (11.2%) | 11 (7.9%) |
| Well (not specified) | 54 (17.9%) | 26 (16.2%) | 28 (20.0%) |
| Household sanitation (N, %) | |||
| Flush toilets | 26 (8.6%) | 14 (8.7%) | 12 (8.6%) |
| Improved pit latrine | 10 (3.3%) | 4 (2.5%) | 6 (4.3%) |
| Slab latrine | 138 (45.9%) | 67 (41.6%) | 71 (50.7%) |
| Unimproved latrine | 81 (26.9%) | 52 (32.3%) | 29 (20.7%) |
| Open defecation | 46 (15.3%) | 24 (14.9%) | 22 (15.7%) |
| Mother’s age, years (mean, SD) | 26.1 (6.1) | 26.0 (6.2) | 26.2 (6.0) |
| Literate mother (N, %) | 54 (17.9%) | 30 (18.6%) | 24 (17.1%) |
| No. times mother pregnant (mean, SD) | 2.3 (2.0) | 2.2 (1.9) | 2.5 (2.1) |
| No. of mother’s children still alive (mean, SD) | 1.9 (1.6) | 1.8 (1.5) | 2.0 (1.8) |
| Weight, kg (mean, SD) | 6.7 (1.3) | 6.8 (1.3) | 6.6 (1.3) |
| Height, cm (mean, SD) | 71.9 (6.6) | 72.2 (6.6) | 71.6 (6.6) |
| HAZ | |||
| Mean score (mean, SD) | −2.4 (1.6) | −2.4 (1.5) | −2.5 (1.6) |
| HAZ < −3 (N, %) | 113 (37.5%) | 55 (34.2%) | 58 (41.4%) |
| MUAC | |||
| Mean circumference, cm (mean, SD) | 11.2 (0.5) | 11.2 (0.4) | 11.2 (0.5) |
| < 11.5 cm (N, %) | 243 (80.7%) | 136 (84.5%) | 107 (76.4%) |
| WAZ | |||
| Mean score (mean, SD) | −3.5 (1.1) | −3.4 (1.0) | −3.5 (1.1) |
| Score below −3 (N, %) | 211 (70.1%) | 115 (71.4%) | 96 (68.6%) |
| WHZ | |||
| Mean score (mean, SD) | −3.1 (1.0) | −3.0 (1.1) | −3.1 (1.0) |
| Score below −3 (N, %) | 164 (54.5%) | 84 (52.2%) | 80 (57.1%) |
HAZ = height-for-age Z-score; MUAC = mid-upper arm circumference; SD = standard deviation; WAZ = weight-for-age Z-score; WHZ = weight-for-height Z-score.
Percentages may not total to 100% due to rounding.
MUAC and WHZ assessments to determine eligibility occurred before the baseline anthropometric assessments. Numbers of children meeting MUAC and/or WHZ thresholds may differ between these two timepoints (eligibility screening and baseline anthropometric assessment).
Weight gain velocity (g/kg/day) is summarized by group overall (primary outcome) and by time point in Table 2 and Figure 2. On average, children in both treatment groups gained weight at all time points, with the greatest weight gain in the first week (Table 2). Average weight gain over the 8-week period was 2.5 g/kg/day (SD 2.0) in the azithromycin group and 2.6 g/kg/day (SD 1.7) in the amoxicillin group. At the final study visit, there was no difference in weight gain when comparing groups (mean difference −0.1, 95% CI −0.5–0.3, P = 0.63). No differences were demonstrated in weight gain at individual time points or in other anthropometric outcomes when compared by treatment group (Table 2).
Table 2.
Longitudinal anthropometric outcomes by treatment group
| Outcome | Azithromycin mean (SD) | Amoxicillin mean (SD) | Unadjusted mean difference, Azithromycin vs. Amoxicillin (95% CI) | Unadjusted P value* | Adjusted mean difference, Azithromycin vs. Amoxicillin (95% CI)† | Adjusted P value*† |
|---|---|---|---|---|---|---|
| Weight gain (g/kg/day)‡ | ||||||
| Week 1 | 5.9 (7.9) | 5.0 (9.2) | 0.8 (−1.2–2.8) | 0.41 | NA | NA |
| Week 2 | 3.3 (6.2) | 3.6 (7.4) | −0.3 (−1.9–1.3) | 0.71 | NA | NA |
| Week 4 | 2.0 (4.0) | 2.5 (4.0) | −0.5 (−1.5–0.5) | 0.29 | NA | NA |
| Week 8 | 1.8 (2.1) | 2.0 (2.2) | −0.3 (−0.8–0.3) | 0.33 | NA | NA |
| Overall | 2.5 (2.0) | 2.6 (1.7) | −0.1 (−0.5–0.3) | 0.63 | NA | NA |
| Weight (kg) | ||||||
| Week 1 | 7.1 (1.3) | 6.9 (1.3) | 0.15 (−0.16–0.47) | 0.33 | 0.05 (−0.04–1.04) | 0.30 |
| Week 2 | 7.2 (1.3) | 7.0 (1.3) | 0.22 (−0.09–0.53) | 0.17 | 0.03 (−0.07–0.14) | 0.51 |
| Week 4 | 7.4 (1.4) | 7.3 (1.4) | 0.10 (−0.23–0.43) | 0.54 | 0.004 (−0.13–0.13) | 0.95 |
| Week 8 | 7.7 (1.4) | 7.7 (1.4) | 0.05 (−0.28–0.38) | 0.75 | −0.03 (−0.19–0.13) | 0.72 |
| Height (cm) | ||||||
| Week 1 | 72.4 (6.7) | 71.8 (6.6) | 0.6 (−1.0–2.1) | 0.46 | 0.13 (−0.11–0.37) | 0.30 |
| Week 2 | 72.6 (6.8) | 71.9 (6.5) | 0.7 (−0.8–2.2) | 0.37 | 0.03 (−0.29–0.34) | 0.86 |
| Week 4 | 72.7 (6.6) | 72.6 (6.5) | 0.06 (−1.5–1.6) | 0.94 | −0.07 (−0.40–0.26) | 0.69 |
| Week 8 | 73.7 (6.2) | 73.5 (6.5) | 0.2 (−1.3–1.6) | 0.84 | −0.05 (−0.44–0.34) | 0.80 |
| MUAC (cm) | ||||||
| Week 1 | 11.5 (0.5) | 11.5 (0.5) | 0.02 (−0.1–0.1) | 0.76 | 0.02 (−0.06–0.10) | 0.58 |
| Week 2 | 11.7 (0.6) | 11.6 (0.6) | 0.09 (−0.05–0.2) | 0.20 | 0.08 (−0.03–0.18) | 0.17 |
| Week 4 | 11.9 (0.7) | 11.8 (0.6) | 0.1 (−0.04–0.3) | 0.14 | 0.12 (−0.02–0.25) | 0.10 |
| Week 8 | 12.4 (0.8) | 12.3 (0.8) | 0.1 (−0.07–0.3) | 0.23 | 0.10 (−0.07–0.28) | 0.24 |
| WAZ | ||||||
| Week 1 | −3.16 (1.1) | −3.30 (1.1) | 0.15 (−0.10–0.39) | 0.25 | 0.05 (−0.07–0.17) | 0.40 |
| Week 2 | −3.00 (1.1) | −3.16 (1.1) | 0.17 (−0.08–0.42) | 0.19 | 0.02 (−0.12–0.16) | 0.74 |
| Week 4 | −2.83 (1.1) | −2.91 (1.1) | 0.07 (−0.19–0.33) | 0.60 | −0.05 (−0.21–0.10) | 0.50 |
| Week 8 | −2.58 (1.1) | −2.65 (1.1) | 0.07 (−0.19–0.33) | 0.59 | −0.03 (−0.21–0.15) | 0.75 |
| WHZ | ||||||
| Week 1 | −2.63 (1.1) | −2.76 (0.9) | 0.13 (−0.10 to −0.37) | 0.25 | 0.07 (−0.09–0.23) | 0.38 |
| Week 2 | −2.39 (1.1) | −2.58 (1.0) | 0.19 (−0.05–0.42) | 0.12 | 0.09 (−0.09–0.27) | 0.33 |
| Week 4 | −2.18 (1.1) | −2.35 (1.0) | 0.17 (−0.08–0.43) | 0.18 | 0.07 (−0.13–0.28) | 0.47 |
| Week 8 | −1.89 (1.2) | −1.95 (1.0) | 0.06 (−0.21–0.32) | 0.68 | −0.005 (−0.24–0.23) | 0.97 |
| HAZ | ||||||
| Week 1 | −2.39 (1.5) | −2.53 (1.6) | 0.14 (−0.21–0.49) | 0.44 | 0.02 (−0.14–0.18) | 0.81 |
| Week 2 | −2.41 (1.5) | −2.50 (1.6) | 0.09 (−0.27–0.45) | 0.62 | −0.06 (−0.24–0.13) | 0.54 |
| Week 4 | −2.37 (1.5) | −2.32 (1.5) | −0.06 (−0.41–0.30) | 0.76 | −0.17 (−0.33 to −0.01) | 0.04 |
| Week 8 | −2.30 (1.5) | −2.37 (1.5) | 0.07 (−0.27–0.41) | 0.69 | −0.05 (−0.24–0.14) | 0.59 |
CI = confidence interval; HAZ = height-for-age Z-score; MUAC = mid-upper arm circumference; SD = standard deviation; WAZ = weight-for-age Z-score; WHZ = weight-for-height Z-score.
*P values calculated using t test (g/kg/day outcomes) or linear regression models.
†Adjusted models include baseline measurements as a covariate; weight gain velocity outcome calculation includes adjustment for baseline and so a separate adjusted model is not presented for this weight gain.
‡Weekly weight gain is interval-specific weight gain velocity (e.g., Baseline to Week 1; Week 1 to Week 2; Week 2 to Week 4; Week 4 to Week 8); overall weight gain is weight gain velocity from Baseline to Week 8.
Figure 2.
Weight gain velocity in g/kg/day by treatment interval and treatment group. Week 1 indicates weight gain (g/kg/day) from Baseline to Week 1, Week 2 is Week 1 to Week 2, and so on. All children had measurements collected at Baseline, Weeks 1–4, and Week 8. Measurements were only collected on Weeks 5–7 for children who had not recovered and thus not presented.
By 8 weeks, 56 children (37.3%) in the azithromycin group and 48 children (36.4%) in the amoxicillin group met the definition for nutritional recovery (Table 3; risk ratio 1.03, 95% CI 0.76–1.40, P = 0.87). In a sensitivity analysis that also included children who met the definition for recovery at the final visit, 98 children (65.3%) in the azithromycin group and 87 children (65.9%) in the amoxicillin group had recovered. The sensitivity analysis also found no difference in nutritional recovery by treatment group (Table 3; risk ratio 0.99, 95% CI 0.84–1.18, P = 0.92). Similarly, additional sensitivity analyses found no difference in time to nutritional recovery by group with either definition of nutritional recovery (original definition: hazard ratio 1.05, 95% CI 0.71–1.53, P = 0.82; sensitivity analysis definition: hazard ratio 1.02, 95% CI 0.77–1.35, P = 0.90).
Table 3.
Clinical outcomes and adverse events by treatment group
| Outcome | Azithromycin | Amoxicillin | Risk ratio* (95% CI) | P value |
|---|---|---|---|---|
| Nutritional recovery, original definition† | 56 (37.3%) | 48 (36.4%) | 1.03 (0.76–1.40) | 0.87 |
| Nutritional recovery, sensitivity definition‡ | 98 (65.3%) | 87 (65.9%) | 0.99 (0.84–1.18) | 0.92 |
| Nonresponse at 8 weeks§ | 89 (58.9%) | 80 (58.8%) | 1.00 (0.74–1.36) | 0.99 |
| Death | 2 (1.3%) | 1 (0.8%) | 1.75 (0.16–19.2) | 0.65 |
| Sought medical care by 8 weeks | 27 (17.3%) | 29 (21.0%) | 0.82 (0.51–1.32) | 0.42 |
| Hospitalized by 8 weeks | 1 (0.6%) | 1 (0.7%) | 0.88 (0.06–14.1) | 0.93 |
| Recent symptoms of infection, 2 weeks after enrollment | ||||
| 3 + watery stools in the last 24 hours | 2 (1.3%) | 2 (1.5%) | 0.89 (0.13–6.25) | 0.91 |
| 14 days of diarrhea | 0 | 0 | N/A | N/A |
| Blood/mucus in stool | 0 | 0 | N/A | N/A |
| Difficulty breathing | 0 | 3 (2.2%) | N/A | N/A |
| Tachypnea‖ | 6 (3.9%) | 7 (5.2%) | 0.76 (0.26–2.22) | 0.62 |
| Adverse events, 2 weeks after enrollment | ||||
| Any | 21 (13.7%) | 37 (27.2%) | 0.50 (0.31–0.82) | 0.006 |
| Fever | 16 (10.5%) | 23 (16.9%) | 0.62 (0.34–1.12) | 0.11 |
| Diarrhea | 5 (3.3%) | 12 (8.8%) | 0.37 (0.13–1.03) | 0.06 |
| Vomiting | 5 (3.3%) | 4 (2.9%) | 1.11 (0.30–4.06) | 0.87 |
| Abdominal pain | 0 | 3 (2.2%) | N/A | N/A |
| Skin rash | 1 (0.7%) | 1 (0.7%) | 0.89 (0.06–14.1) | 0.93 |
| Constipation | 0 | 1 (0.7%) | N/A | N/A |
*Estimated using modified Poisson model with robust standard errors.27
†Defined as MUAC > 12.5 cm and/or WHZ > −2 on two consecutive visits, corresponding to the admission criterion (MUAC and/or WHZ).
‡Defined as MUAC > 12.5 cm and/or WHZ > −2 on two consecutive visits and/or at their final measurement.
§Defined as MUAC < 12.5 cm and/or WHZ < −2 at the 8-week visit, corresponding to the admission criterion (MUAC and/or WHZ).
‖Defined as a respiratory rate > 50 breaths/minute for children aged 6–11 months and > 40 breaths/minute for children aged 12–59 months.
Of the 301 enrolled children, caregivers of 58 (19.3%) reported adverse events within the first 2 weeks of enrollment. Adverse events were significantly different by treatment group, with 21 children (13.7%) in the azithromycin group and 37 children (27.2%) in the amoxicillin group experiencing any adverse event over 2 weeks (Table 3; P = 0.006). No differences were identified in comparisons of individual adverse events by group (Table 3). Overall, five serious adverse events were reported during the study period, all of which were deaths or hospitalizations determined to be unrelated to the study drug. No differences were found in other clinical outcomes by treatment group, including nonresponse, death, and hospitalization (Table 3).
DISCUSSION
The results of this pilot trial establish the feasibility of a fully powered efficacy trial to evaluate antibiotics for uncomplicated SAM in this setting. Given the burden of malnutrition and the community-based presence of the nutritional programs in this district in Burkina Faso, this pilot enrolled 301 children in six sites over 4 months, confirming the potential ease and speed of enrollment in a larger trial. Nearly 94% of children completed the final 8-week follow-up visit and > 90% of children completed each interim follow-up visit. The excellent follow-up in this pilot reduces concerns about underpowered or biased outcome comparisons associated with loss to follow-up in a larger trial. In addition, the strong follow-up sets the stage for longer term follow-up time points in a future trial. Evidence indicates that postdischarge relapse is common, but research and treatment programs have typically focused on short-term outcomes during active management.28
We were unable to detect differences in weight gain or nutritional recovery between the azithromycin and amoxicillin groups. On average, children in both groups experienced 2.6 g/kg/day of weight gain and 37% of children reached nutritional recovery by 8 weeks, both of which are lower than expected by programs or reported in other studies. Nutritional programs target 8 g/kg/day of weight gain and expect > 75% of children with SAM to recover during management.29 Although other studies have also found lower weight gain and nutritional recovery in children with uncomplicated SAM in sub-Saharan Africa compared with program targets, these other study populations reported greater weight gain than this pilot trial and greater recovery proportions overall.8,9,30 In a sensitivity analysis, the proportion of children achieving nutritional recovery in this trial was more similar to those seen in other West African settings.9,30 The sensitivity analysis included children who met the recovery definition at the final visit rather than requiring two consecutive visits, suggesting that programmatic decisions about discharge and recovery may differ from the official guidelines. At the same time, Burkina Faso bears a particularly large burden of severe malnutrition which could underpin slower or lower recovery in this population.2,31,32 In this particular trial, the shortage of RUTF, a common occurrence in nutritional programs in this setting, could also have contributed to low weight gain and recovery. A future trial might consider ensuring a constant supply of RUTF to avoid this limitation, although this approach would widen the gap between the trial and real-world setting, affecting the generalizability and interpretation of results. Alternatively, though the 8-week primary outcome time point was chosen to align with previous trials,8 it is possible that a longer time frame is needed to capture recovery in this setting.
Fewer adverse events were reported in children receiving azithromycin compared with amoxicillin. The difference appears to be driven by fewer events of diarrhea and fever in children receiving azithromycin, although these individual comparisons by group were not statistically significant at an alpha of 0.05. Azithromycin is known for its safety profile, has been well tolerated by children in community-based programs for trachoma and studies of child survival, and is known for its efficacy against common childhood infections.13,15,33 This finding is also consistent with results from a cluster-randomized trial that compared community distributions of azithromycin to placebo in children in Niger.13 In this setting, a survey of adverse events in infants after community drug distributions found fewer cases of diarrhea in children in communities receiving azithromycin compared with placebo (RR 0.68, 95% CI 0.49–0.96, P = 0.03).34 This trial also found evidence of a reduction in the relative abundance of a range of gut bacteria in children in communities receiving azithromycin compared with placebo,35 including pathogenic Campylobacter which is a common cause diarrheal disease in children.36–38 Amoxicillin is associated with a similar adverse events profile to azithromycin, and the previous trials of amoxicillin for uncomplicated SAM found more adverse events in the amoxicillin arm that our trial.8,10 This may in part be due to the different definitions of adverse events and different monitoring time periods used across studies.
The strengths of the study include the randomized design and the low loss to follow-up. Although a placebo was not used to maintain masking of participants or personnel administering treatment, outcome assessors remained masked, limiting the threat of bias in data collection. Limitations include the shortage of RUTF during enrollment, which could have influenced nutritional outcomes and contributed to the lower overall weight gain and recovery seen in this trial. While this impacts the broader generalizability of results, we do not believe the RUTF shortage affected treatment groups differentially. As this pilot aimed primarily to determine the feasibility of conducting a larger trial, the trial was underpowered to detect a modest difference in nutritional outcomes in two groups receiving antibiotics. Previous trials have used nutritional recovery as a primary outcome to align with program targets, yet this dichotomized outcome may not fully capture the continuous recovery process experienced by malnourished children. As a continuous outcome, weight gain velocity would provide greater power to detect a difference in groups and may better characterize the dynamic nature of the treatment process. Another limitation is the overlap in reported symptoms of infection and adverse events. We attempted to separate reporting of common adverse events associated with these antibiotics and more serious infectious symptoms, but there remains quite a bit of similarity. A future study may consider reporting all clinical complications and adverse events together to avoid this. This pilot was also limited by the lack of resources to sufficiently track adherence to amoxicillin, adherence to RUTF, or the use of other antibiotics for other conditions, each of which will be important to consider in future work. In addition, as azithromycin is not currently readily available at CSPS-level health centers in Burkina Faso, procurement will be another consideration in the feasibility of operationalizing this intervention beyond a trial setting, if azithromycin is found to be efficacious in a larger trial. Broadly, the results of the present study may be generalizable to children with uncomplicated SAM in government-run nutritional programs in settings with a similar profile and burden of malnutrition.
Overall, the results of this pilot study establish the feasibility and rationale for conducting a randomized controlled trial to evaluate antibiotics for uncomplicated SAM in this setting. Although we were unable to demonstrate a difference in nutritional outcomes by group, both groups experienced weight gain and nutritional recovery and the azithromycin group experienced fewer adverse events. Given its ease of dosing and safety profile, azithromycin may be a viable alternative antibiotic to consider in the adjunctive treatment of uncomplicated SAM.
ACKNOWLEDGMENTS
We would like to thank the study teams at the CSPS enrollment sites for their contributions to this trial. The authors would also like to thank the Data and Safety Monitoring Committee: Nisha Acharya (chair; MD, University of California, San Francisco), Emily Smith (ScD, MPH, George Washington University), and Christine Stewart (PhD, University of California, Davis).
REFERENCES
- 1. UNICEF , 2019. The State of the World’s Children 2019. Children, Food, and Nutrition: Growing Well in a Changing World. New York, NY: UNICEF. [Google Scholar]
- 2.Black RE et al. 2013. Maternal and child undernutrition and overweight in low-income and middle-income countries. Lancet 382: 427–451. [DOI] [PubMed] [Google Scholar]
- 3.Alcoba G Kerac M Breysse S Salpeteur C Galetto-Lacour A Briend A Gervaix A , 2013. Do children with uncomplicated severe acute malnutrition need antibiotics? A systematic review and meta-analysis. PLOS ONE 8: e53184. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Bourke CD Berkley JA Prendergast AJ , 2016. Immune dysfunction as a cause and consequence of malnutrition. Trends Immunol 37: 386–398. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Page AL et al. 2013. Infections in children admitted with complicated severe acute malnutrition in Niger. PLOS ONE 8: e68699. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Jones KD Berkley JA , 2014. Severe acute malnutrition and infection. Paediatr Int Child Health 34 (Suppl 1): S1–S29. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. World Health Organization , 2013. Guideline: Updates on the Management of Severe Acute Malnutrition in Infants and Children. Geneva, Switzerland: WHO. [PubMed] [Google Scholar]
- 8.Trehan I Goldbach HS LaGrone LN Meuli GJ Wang RJ Maleta KM Manary MJ , 2013. Antibiotics as part of the management of severe acute malnutrition. N Engl J Med 368: 425–435. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Isanaka S et al. 2016. Routine amoxicillin for uncomplicated severe acute malnutrition in children. N Engl J Med 374: 444–453. [DOI] [PubMed] [Google Scholar]
- 10.Trehan I Schechtman KB Manary MJ , 2016. Amoxicillin for severe acute malnutrition in children. N Engl J Med 375: 191. [DOI] [PubMed] [Google Scholar]
- 11.Million M Lagier JC Raoult D , 2017. Meta-analysis on efficacy of amoxicillin in uncomplicated severe acute malnutrition. Microb Pathog 106: 76–77. [DOI] [PubMed] [Google Scholar]
- 12.Das JK Salam RA Saeed M Kazmi FA Bhutta ZA , 2020. Effectiveness of interventions for managing acute malnutrition in children under five years of age in low-income and middle-income countries: a systematic review and meta-analysis. Nutrients 12: 116. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Keenan JD et al. 2018. Azithromycin to reduce childhood mortality in sub-Saharan Africa. N Engl J Med 378: 1583–1592. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.O’Brien KS et al. 2020. Biannual azithromycin distribution and child mortality among malnourished children: a subgroup analysis of the MORDOR cluster-randomized trial in Niger. PLoS Med 17: e1003285. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Langtry HD Balfour JA , 1998. Azithromycin—a review of its use in paediatric. Drugs 56: 273–297. [DOI] [PubMed] [Google Scholar]
- 16.Coles CL Levens J Seidman JC Mkocha H Munoz B West S , 2012. Mass distribution of azithromycin for trachoma control is associated with short-term reduction in risk of acute lower respiratory infection in young children. Pediatr Infect Dis J 31: 341–346. [DOI] [PubMed] [Google Scholar]
- 17.Coles CL Seidman JC Levens J Mkocha H Munoz B West S , 2011. Association of mass treatment with azithromycin in trachoma-endemic communities with short-term reduced risk of diarrhea in young children. Am J Trop Med Hyg 85: 691–696. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Phiri MD. et al. , 2021. The duration of protection from azithromycin against malaria, acute respiratory, gastro-intestinal and skin infections when given alongside seasonal malaria chemoprevention: secondary analyses of data from a clinical trial in Houndé, Burkina Faso and Bougouni, Mali. Clin Infect Dis 73: e2379–e2386. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Blumer JL , 2005. Evolution of a new drug formulation: the rationale for high-dose, short-course therapy with azithromycin. Int J Antimicrob Agents 26 (Suppl 3): S143–S147. [DOI] [PubMed] [Google Scholar]
- 20.Rogawski ET et al. 2017. Use of antibiotics in children younger than two years in eight countries: a prospective cohort study. Bull World Health Organ 95: 49–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Sie A et al. 2018. Effect of antibiotics on short-term growth among children in Burkina Faso: a randomized trial. Am J Trop Med Hyg 99: 789–796. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Bell BG Schellevis F Stobberingh E Goossens H Pringle M , 2014. A systematic review and meta-analysis of the effects of antibiotic consumption on antibiotic resistance. BMC Infect Dis 14: 13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. 2014. Protocole Nationale: Prise en charge integrée de la malnutrition aiguë (PCIMA). Ouagadougou, Burkina Faso: Ministere de la Santé Burkina Faso. [Google Scholar]
- 24.O’Brien KS Sié A Dah C Ourohire M Arzika AM Boudo V Lebas E Godwin WW Arnold BF Oldenburg CE , 2021. Azithromycin for uncomplicated severe acute malnutrition: study protocol for a pilot randomized controlled trial. Pilot Feasibility Stud 7: 97. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Sie A et al. 2022. Azithromycin versus amoxicillin and malarial parasitemia among children with uncomplicated severe acute malnutrition: a randomized controlled trial. Am J Trop Med Hyg 106: 351–355. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. WHO Multicentre Growth Reference Study Group , 2006. WHO Child Growth Standards: Length/Height-for-Age, Weight-for-Age, Weight-for-Length, Weight-for-Height and Body Mass Index for-Age: Methods and Development. Geneva, Switzerland: World Health Organization, 312. [Google Scholar]
- 27.Zou G , 2004. A modified poisson regression approach to prospective studies with binary data. Am J Epidemiol 159: 702–706. [DOI] [PubMed] [Google Scholar]
- 28.Stobaugh HC Mayberry A McGrath M Bahwere P Zagre NM Manary MJ Black R Lelijveld N , 2019. Relapse after severe acute malnutrition: a systematic literature review and secondary data analysis. Matern Child Nutr 15: e12702. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. 2011. The Sphere Handbook: Humanitarian Charter and Minimum Standards in Humanitarian Response. Geneva, Switzerland: The Sphere Project. [Google Scholar]
- 30.Desyibelew HD Bayih MT Baraki AG Dadi AF , 2020. The recovery rate from severe acute malnutrition among under-five years of children remains low in sub-Saharan Africa. A systematic review and meta-analysis of observational studies. PLOS ONE 15: e0229698. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Belesova K Gasparrini A Sié A Sauerborn R Wilkinson P , 2017. Household cereal crop harvest and children’s nutritional status in rural Burkina Faso. Environ Health 16: 65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Belesova K Gasparrini A Sié A Sauerborn R Wilkinson P , 2017. Annual crop-yield variation, child survival, and nutrition among subsistence farmers in Burkina Faso. Am J Epidemiol 187: 242–250. [DOI] [PubMed] [Google Scholar]
- 33.Taylor HR Burton MJ Haddad D West S Wright H , 2014. Trachoma. Lancet 384: 2142–2152. [DOI] [PubMed] [Google Scholar]
- 34.Oldenburg CE et al. 2018. Safety of azithromycin in infants under six months of age in Niger: a community randomized trial. PLoS Negl Trop Dis 12: e0006950. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Doan T et al. 2019. Gut microbiome alteration in MORDOR I: a community-randomized trial of mass azithromycin distribution. Nat Med 25: 1370–1376. [DOI] [PubMed] [Google Scholar]
- 36.Jimenez SG Heine RG Ward PB Robins-Browne RM , 1999. Campylobacter upsaliensis gastroenteritis in childhood. Pediatr Infect Dis J 18: 988–992. [DOI] [PubMed] [Google Scholar]
- 37.Bourke B Chan VL Sherman P , 1998. Campylobacter upsaliensis: waiting in the wings. Clin Microbiol Rev 11: 440–449. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Sire JM Garin B Chartier L Fall NK Tall A Seck A Weill FX Breurec S Vray M , 2013. Community-acquired infectious diarrhoea in children under 5 years of age in Dakar, Senegal. Paediatr Int Child Health 33: 139–144. [DOI] [PubMed] [Google Scholar]