|
nif:isString
|
-
This was a multi-centric, double blind placebo-controlled randomised trial conducted in outpatient departments of 8 referral hospitals in India. Individual institutional ethics committees' as well as the IndiaClen Institutional Review Board approved the study. It was designed for equivalence of two modes of therapy with a predefined range of equivalence as an interval from −5% to +5%. The trial was registered with the Clinical Trials registry (NCT00407394). The protocol for this trial and supporing CONSORT checklist are available as supporting information; see Checklist S1 and Protocol S1. Two-stage screening was done (verbal followed by standardised screening) for inclusion and exclusion criteria. Children aged 2–59 months were verbally screened for complaints of cough, rapid respiration, or difficulties in breathing. Excluded were those with signs of WHO defined severe pneumonia or very severe disease[7], other conditions requiring antibiotics therapy, clinically recognized congenital heart disease, chronic systemic disorders, hospitalisation in past 2 weeks, use of antibiotics in previous 2 days, measles within the last month, history of penicillin allergy and prior enrolment in the study. Those who were verbal screening positive and had no exclusion criteria were examined for study entry criteria, which were audible or auscultatory wheeze and WHO defined fast breathing, that is respiratory rate of ≥50 per minute (for age 2–11 months) or ≥40 per minute (for age 12–59 months) [10]. Those who fulfilled the study entry criteria were given a maximum of three doses of salbutamol (400 microgm/dose) by either a nebuliser or metered dose inhaler and their respiratory rate was reassessed. If fast breathing persisted above WHO defined age specific cut-offs, a chest radiograph was taken for diagnosing radiological pneumonia, defined as presence of fluids, parenchymal infiltrates or consolidation [10]. Those with radiological pneumonia were treated according to the standard hospital treatment guidelines. The rest were invited to participate in the study. Those who accepted the invitation were requested to provide informed written parental consent in local language. Local languages used were site specific and were Hindi for Lucknow, Delhi and Chandigarh sites, Marathi for Nagpur and Mumbai sites, Tamil for Chennai and Vellore site and Malayali for Trivandrum site. Children of parents consenting for participation were randomised to receive amoxycillin or placebo and continued with oral salbutamol. Coordinating centre ensured standardised training and quality assurance. All the participants were followed up on day- 4 and between days 11–14 days. Home visits were done within 24 hours for those who failed to report on the appointed days. Children who failed on therapy, developed adverse reactions or withdrew consent were treated according to standard hospital guidelines.
Participants received either amoxycillin, 31–54 mg/Kg/day, in three divided doses, or placebo, which looked and tasted similar to amoxycillin, in three divided doses orally. In addition all participants also received oral salbutamol, 2.5 ml or 5 ml thrice a day for ages 2–11 months and 12–59 months, respectively (each 5 ml = 2 mg salbutamol). Patients with fever also received paracetemol (10–15 mg/kg/dose) in addition to the study intervention.
The primary objective was to compare the proportions of eligible cases that achieve clinical cure on day- 4 on three-days of treatment with either oral amoxycillin or placebo. Our secondary objective was to compare the proportions that clinically relapsed within the next 11–14 days, as well as identify the determinants, if any, of clinical failure. Secondary laboratory outcome was the proportions positive for respiratory syncycial virus (RSV) in nasopharyngeal aspirate at enrolment.
Main outcome measure was clinical cure on day- 4, defined as those who were not assessed to have clinical failure. Clinical failure was defined as either (i) development of WHO-defined severe pneumonia or very severe disease (with or without wheeze) before or on day- 4 assessment, or (ii) oxygen saturation on pulse oximetry <90% before or on day- 4 assessment or (iii) axillary temperature >101 degrees F on day- 4 assessment, or (iv) persistence of WHO-defined non-severe pneumonia on day- 4 assessment, or (v) presence of wheeze on day- 4 assessment. Loss to follow up on day- 4 or withdrawal from the study at any time after recruitment was also considered as therapy failure in the intention-to-treatment (ITT) analysis. Clinical relapse was defined as cases, which were clinically cured on day-4 assessment, but at days-11-14 follow-up showed signs of WHO-defined pneumonia.
This was calculated for therapeutic equivalence of two treatments (placebo and amoxycillin). It was assumed that the overall failure rate would be 17% if the treatments were equivalent [5]. We used δ = 0.05 to be the range of equivalence for the difference in failure rates. To detect this difference using Type I error probability (alpha) of 0.05 and power of 0.9, the required size of each treatment group was 970.
A professional not associated with clinical care (GGA) generated the randomization scheme at Lucknow. Randomised block sizes of two to four were used to avoid any uneven distribution of patients between two interventions. Unlabelled medicines were placed in serially numbered opaque envelopes by the pharmacy. Doses were provided for a maximum weight of 20 Kg. Extra doses were kept for administration to a child who vomited within 30 minutes of dosing. Health care workers, patients as well as research staff and investigators were blinded.
Adherence was assessed by pill count on day- 4 follow-up. Non-adherence was defined as intake of less than 7 doses.
A chest radiograph was taken prior to randomisation. Nasopharyngeal aspirate was tested at enrolment for RSV using DirectogenR kit (manufactured by Becton Dickinson and Company, Sparks MD 21152, USA).
Primary analyses were done on ITT basis. Per protocol analysis was also done on subjects with complete follow-up and adherence to treatment. We have used the two one-sided tests (TOST) procedure as well as the confidence interval approach for assessing therapeutic equivalence of two treatments [11]. Baseline and other characteristics were compared between the two regimens. Association of clinical failure with different characteristics of patients was assessed. Chi-squared test was used for testing two-sided hypotheses, whereas Z- test was used to test one-sided hypotheses. A p value of <0.05 was considered to be statistically significant. Crude odds ratios and 95% confidence intervals were computed. The covariates, which were found to be significant or nearly significant (p≤0.10) in univariate analysis, were included to construct a multivariate model for assessing determinants of treatment failure by forward stepwise logistic regression. In the later model, we also investigated the possibility of variables acting as effect modifier for the risk factors. Final model included covariates that were statistically significant. The primary outcome measure was also assessed for each site and the overall odds ratio after stratifying for site was obtained. Mantel-Haenszel statistic for testing independence of clinical failure with interventions, adjusting for site was used. Breslow-Day statistic for testing three-factor interaction of clinical failure, interventions and site was used. Data were analyzed using statistical software “Statistical Package for Social Sciences” (SPSS) 11.0 version.
|
![[This page as RDF]](http://data.gesis.org/somesci/static/rdf-icon.gif){kind=icon}