This study was a two-parallel arm, blinded, randomized controlled trial. The trial was conducted in accordance with the Declaration of Helsinki and the Guidelines for Good Clinical Practice at the Evidence-Based Medicine Center of the Fourth Affiliated Hospital of Harbin Medical University, a clinical research center in China. Patients were recruited between August 2012 and November 2013. The study was approved by the Medical Ethics Committee of the Fourth Affiliated Hospital of Harbin Medical University. Eligible participants were randomly allocated to the baclofen group or the placebo group at a 1:1 allocation ratio and received treatment for 5 days, with 15 days of follow-up. This study complied with CONSORT guidelines.
Inclusion and exclusion criteria
We included participants aged 18 to 65 years with persistent hiccups (more than 48 hours but less than 1 month) after stroke. In addition, participants were required not to have taken baclofen within 15 days before study entry and were required to sign an informed-consent form.
The exclusion criteria included persistent hiccups mainly associated with cancer , multiple sclerosis , meningitis , brain abscess , traumatic brain injury , encephalitis , spinal cord lesions , kidney failure , pneumonia , laryngitis , and cardiorespiratory arrest ; rejection of baclofen therapy; and failure to complete clinical treatment.
Randomization was performed by using a computer, and group-allocation instructions were concealed in opaque sealed envelopes. For assignment to the baclofen or placebo group, participants were required to select one of the sequentially numbered envelopes.
Participants and recruitment
We planned to conduct our research in the Fourth Affiliated Hospital of Harbin Medical University. In preparing for this research, we found that this center had offered baclofen or placebo treatment to 30 people between August 2012 and November 2013. This enabled a fair test of the feasibility criteria, and if recruitment was good, potentially permitted a reliable calculation of the effect size of the intervention for computing the subsequent sample size. Those individuals who elected to undergo baclofen treatment were informed about the research and given an information sheet. From those agreed to participate, consent was given at the next appointment. After the clinical assessment, participants were randomized to receive baclofen or a placebo, both of which were given by therapists, who were trained in their administration.
Participants recruited to the baclofen group received 10 mg 3 times daily for 5 days. Participants assigned to the placebo group received 10 mg of a placebo 3 times daily for 5 days. The participants were then asked to identify any obstacles to the goal attainment and to consider the possibility that the therapy could help them overcome these obstacles. This was intended to motivate the participants to engage in the therapy.
The placebo tablets used in this study were composed of hydroxypropyl cellulose, pregelatinized starch, dextrin, starch, silicon dioxide, magnesium stearate, sodium carboxymethyl starch, sucrose, talcum powder, gelatin, silicone oil, and Chinese wax. The tablets were produced by Changzhou Siyao Pharmaceuticals Co. Ltd. and had the same dose, taste, and color as baclofen.
Patients outcomes were evaluated by using the Chinese Medicine Medical Association standard criteria  and classified as follows: cure, cessation of persistent hiccups within the intervention period, with no relapse in one week; improvement, reduction in the frequency and severity of hiccups and other symptoms (such as abdominal discomfort); and no effect, no amelioration of hiccups.
Data collection and analysis
In addition to the clinical outcomes, the efficacy and side effects in the two groups were recorded and analyzed. We collected data on the number of eligible participants with persistent hiccups, willingness of the participants to be randomized, and compliance with the intervention, and aimed to estimate the effect size for a fully powered trial.
Sample size and analysis
This feasibility study estimated that a sample of 30 participants would be sufficient to provide data to answer our study questions . Clinical outcome data were analyzed by using an “intention-to- treat” approach, and the initial analysis examined the demographic and baseline characteristics of patients randomized to the trial. Between-group differences in categoric data were assessed by using the Fisher Exact test or Mann–Whitney U test; the t test was used for continuous data. Relative risks and 95% confidence intervals were also reported. Levels of significance were reported at P < 0.05. Descriptive statistics were used to assess the feasibility questions. Analysis was performed by a statistician who was blinded to the study group.