Trial design
Single centre, balanced randomized [1:1], single blind, two arms, parallel-group clinical trial was conducted to compare the antipyretic effectiveness and side effects of single dose of oral and rectal paracetamol.
Participants
Children aged between 2-6 years admitted to the Paediatric Unit at the District General Hospital Gampaha were enrolled between June 2017 to December 2017. Eligibility criteria for study inclusion were: Children aged between 2-6 years; Duration of fever for a maximum of 3 days irrespective of the cause of fever and irrespective of prior administration of any anti-pyretic or any other medicines; Documented axillary temperature > 100° F at the time of recruitment. Exclusion criteria were: Children with known allergy to paracetamol; Conditions that preclude oral (example: vomiting) and rectal (example: diarrhoea) administration of medicines; Children with reduced level of consciousness; Children who had possible overdose of paracetamol; Children who had any anti pyretic within 8 hours of admission.
Interventions
Children meeting inclusion and exclusion criteria and whose parents provided the consent, were allocated into two study arms using simple randomization. Arm 1: Single dose of paracetamol syrup 15mg/kg was given orally. The same preparation (syrup containing 120 mg/5ml of paracetamol) available in the Hospital Pharmacy was used throughout the study. The dose was calculated in mg and then converted to ml. The required dose was given by a calibrated sterile syringe. Arm 2: Single dose of paracetamol suppository 30 mg/kg was given rectally. The same preparation was used throughout the study. Rectal paracetamol is available as 125 mg, 250 mg, 500 mg suppositories, it was used in suppository form. As partition of the suppository to pieces was difficult, round up method was used. After allocation of children into each group, the recommended dose of Paracetamol was given. Dose was calculated to the weight. Therefore, before administration of drug weight of all the participants were measured by accurate digital weighing scale throughout the study. The intervention was administered by trained nurses and the outcomes was assessed by the principal investigator who was made unaware of the interventional status.
Temperature measurements
Principal investigator who was blind to identity of the intervention, measured the temperature once before the administration of paracetamol and also in 15, 30, 60, 90, 120,150,180 minutes ’time after the administration of paracetamol. Same thermometer was used for the entire study. Principal investigator followed the standard protocol in measuring the temperature and recorded the readings in the questionnaire. Tepid sponging was not given by care givers during trial duration three hours as it may contribute to temperature reduction and can unduly influence the research outcome.
Outcomes
Primary outcome: Time taken to fever reduction by at least 1°F (rate of temperature reduction)
Secondary outcome: Maximum antipyresis following administration of single dose of Paracetamol within 3 hours’ time.
Sample size
135 participants were recruited into each group.
Formula of calculating sample size for detecting clinical superiority:
n = [(Z α/2 + Z β)2× {2(ó)2}]/ (μ1 - μ2)2 [17,18]
n = sample size required in each group,
μ1 = mean change in intervention group = 2
μ2 = mean change in the control group = 1.7
μ1-μ2 = clinically significant difference = 0.3
ó = standard deviation = 0.84
Z α/2: The SND value corresponding to 5% significance level, taken as 1.96
Z β: The SND value of the Type II error when power is 80%, taken as 0.84
= 2 (1.96 + 0.84) 2 x 0.842
0.32
= 2 x 7.84 x 0.7
0.09
= 121
Assuming a response rate of 90%
N= 121/ 0.9 = 134 per each arm
Randomisation and blinding
Block randomization was done using a computer-generated random letter sequence. A Co - investigator and the parents of the children were aware about the route of the drug given. But the principal investigator who is recording the temperature was blinded for the intervention.
Statistical methods
Data analysis was done using Statistical Package for Social Sciences (SPSS) software version 21. Univariate analysis was performed using Chi Square test to compare categorical variables and independent samples t- test and one-way ANOVA were used to compare means between categories. Pearson correlation coefficient was used to determine correlations. Multivariate analysis was done using multiple logistic regression models. p value of < 0.05 was considered as statistically significant.