This collaborative multicentre academic clinical trial with funding from the Research Council of Norway under a research grant on Global Health and Vaccination Research (GLOBVAC) and CISMAC (Centre for Intervention Sciences in Maternal and Child Health), Norway, will be executed under a Department of Biotechnology (DBT), Govt of India Program of Cooperation between DBT, Indian Institutions and the Norwegian Institutions under the agreement in Science and Technology between Government of Republic of India and Government of Kingdom of Norway. The aim of this individually randomized double-blind placebo controlled parallel group superiority trial is to measure the efficacy of zinc administered orally as an adjunct to standard therapy to infants aged 3 days to 2 months hospitalized with clinical severe infection in reducing case fatality during hospitalization. We also propose to estimate the efficacy of zinc on reducing the extended case fatality risk, that is the risk of death until 12 weeks from the day of enrolment. The secondary objectives are to evaluate the efficacy of 10 mg of elemental zinc administered orally as an adjunct to standard antibiotic therapy to infants aged 3 days up to 2 month (59 days) hospitalized with clinical severe infection against the following i) failure of primary treatment defined as a need to change antibiotics or requirement for life support or death, ii) time to cessation of signs of clinical severe infection, iii) time to failure of primary treatment, iv) time to discharge, and v) death or severe illness at any time after discharge from hospital until 12 weeks from day of enrolment. We also propose to measure the immunobiological effects of zinc, to interrogate the stool for enteropathogens and characterization of intestinal microbiome/ metagenome and to evaluate the health and economic consequences of providing zinc as adjunct to standard treatment to young infants hospitalized with “clinical severe infection. The trial will be conducted over a period of 5 years. The recruitment for this multicentre study will take place in 7 hospitals, 5 in Delhi, India and 2 in Nepal. The participants will be randomized to receive zinc or placebo in a 1:1 allocation ratio. The intervention will be administered daily at 12 hourly intervals from the time of enrolment for 14 days and will be followed up till discharge and until 12 weeks from the day of enrolment. If the results of this study are consistent with our earlier trial, then this study would contribute evidence towards revising treatment recommendations for low resource settings in South Asia and elsewhere.

 Brief Statistical Analysis Plan for Zinc as an adjunct for the treatment of clinical severe infection in infants younger than 2 months (doi: 10.1186/s40360-017-0162-5)

Main research question: Does oral zinc used as an adjunct to standard treatment reduce risk of death in infants with clinical severe infection?

Hypothesis: Daily oral administration of 10 mg elemental zinc as an adjunct to standard therapy to infants aged 3 to 59 days hospitalized for clinical severe infection (CSI) will result in a relative mortality risk reduction (i.e. efficacy) of at least 30% both during hospitalization (case fatality) as well as for the period up to 12 weeks from the day of enrolment (extended case fatality).

Primary objectives: Estimate the efficacy of the adjunct zinc treatment for CSI against

1.              case fatality

2.              extended case fatality

Secondary objectives: Estimate the efficacy of the adjunct zinc treatment for CSI:

i.            against treatment failure during initial hospitalization, i.e. required life support (ventilation or vasoactive drugs) or need to change antibiotics because of persistence of CSI signs after 48h of enrolment, or worsening of existing or appearance of new CSI signs or death

ii.           against death at any time after discharge from hospital until 12 completed weeks from enrolment

iii.          against severe illness requiring hospitalisation at any time after discharge from hospital until 12 completed weeks from enrolment.

iv.          with respect to time to cessation of CSI signs (beginning of a 48-hour period with no CSI signs)

v.           with respect to time to discharge from hospital

We will compare the risk of case fatality [primary objective 1 (1^o#1)], extended case fatality (1^o#2) and treatment failure (2^o#i), death from discharge to 12 completed follow-up weeks (2^o#ii), and of rehospitalization for severe illness (2^o#iii) between the two trial arms to arrive at estimates of relative risk (RR); efficacy (1-RR), risk difference (RD) and number needed-to-treat (NNT=1/RD). As a supplementary analysis, we will also estimate the time to death from enrollment while in hospital (1^o#1) and to the end of 12 weeks after enrollment (1^o#2). Such time-to-event analysis will also be used to estimate the efficacy of adjunct zinc therapy with respect to time to cessation of CSI signs (2^o#iv) as well as to hospital discharge (2^o#v).

Graphic presentations and analytic procedures

We will in the trial profile present all screened patients down to those randomized and whose data were analyzed. The baseline table will present the characteristics of the enrolled participants by trial arm. The information will be used to evaluate whether any baseline differences of prognostic factors should be adjusted for (confounding) and depict the type of CSI patients the study findings pertain to (external validity). We will in the supplement also present a baseline table for all randomized patients separately for the 240 patients enrolled in the two hospitals in Nepal and the 2913 patients enrolled in the 5 Indian hospitals. The baseline tables will include but will not necessarily be limited to the following characteristics:

Mother and patient’s family:

Mother’s education (years)

[Any other maternal/family characteristics]

Patient:

Age

Mean (SD) age

Categorized in 3-6 days, 7-28 days and 29-59 days)

Sex

Weight in kg (Mean [SD])

Weight for age Z scores (< -2 Z-score) (n[%])

History

Antimicrobial therapy for current illness received before admission

Clinical signs

Axillary temperature > 38^oC

Stopped feeding well

Severe chest in-drawing

Grunting

Fast breathing (i.e. respiratory rate >60/min)

Convulsions

Movement only when stimulated

Refusal to feed

Diarrhea

Laboratory parameters

Hemoglobin (g/dL)

Total Leukocyte Count (cells/mL)

Absolute Neutrophil Count (cells/mL)

Band cell: Neutrophil ratio

Micro-ESR(mm)

CRP (mg/L)

Plasma zinc concentration (µg/dL) median (IQR)

Statistical Analysis

Using generalized linear models of the binomial family with log and identity links we will estimate the RRs (and corresponding efficacies) and RDs (and corresponding NNT), respectively, for the two primary objectives (death during hospitalization and death until 12 completed weeks after enrolment) as well as for the secondary objectives i (treatment failure), ii (death from discharge until 12 completed weeks after enrolment) and iii (post-discharge rehospitalization for severe illness). Treatment failure will be reported both by (i) first event and by (ii) worst event.

In the supplementary analysis, we will estimate the corresponding hazard ratios for time to death from enrolment while in hospital (1^o#1) and to the end of 12 weeks after enrolment (1^o#2), right-censoring children who die on the day of their demise. Patients will also be censored when the caretaker withdraws consent for study continuation or the infant for other reasons could not be followed up for as long as we intended.

Such supplementary time-to-event analysis will also be used to estimate the efficacy of adjunct zinc therapy with respect to time to death at any time after discharge from hospital until 12 completed weeks from enrolment (2^o#ii), time to cessation of CSI signs (2^o#iv), and time to hospital discharge (2^o#v).

We will also use Poisson regression or negative binomial regression, to address secondary objectives.

Zinc may exert an effect only after some time. In an exploratory analysis, the above analyses will therefore be repeated where the outcomes are redefined to occur only when they take place after 24 h of administering the first dose of zinc.

Subgroup analyses

We will estimate the efficacy of zinc in various subgroups based on:

Parameters on which the randomization was stratified:

A.     presence or absence of diarrhea on admission,

B.     study hospital and country

Other parameters:

C.     Age at enrolment (3-6 days vs. ≥ 7 days)

D.     Positive vs. negative septic screen[1]

We will display the RDs for each of the above-mentioned strata and depict them in a forest plot. We will also estimate the heterogeneity of RDs using interaction terms in our regression models to determine the absolute excess risk due to interaction.

Our primary analysis will use an intention-to-treat approach where all infants assigned into the zinc or placebo arm of the trial and whose outcome is known will be included. We will also consider Instrumental Variable Analyses (IVA) attempting to shed light on what may be a better estimate of the intrinsic efficacy of adjunct zinc therapy for CSI, i.e. the efficacy of zinc had it been given to all children in the scheduled doses and intervals. In our IVA, the random allocation will be the instrument and actual amount of zinc administered to each baby over the first 5 days of treatment will be the exposure variable. We will also perform a simple per protocol analysis in which patients who received less than 50% of the projected doses in the first 5 days after enrolment will be excluded from the analyses.

[1]Positive septic screen: Presence of any two of the following laboratory parameters: total leucocyte count <5000mL^-1; absolute neutrophil count <1500mL^-1; band cell: neutrophil ratio > 0.2; micro erythrocyte sedimentation rate >15 mm at 1 hour; C-reactive protein levels >1 mg/dL.

Negative septic screen: Presence of a maximum of one of the above laboratory parameters.