Study Design
We performed a prospective, randomized, open-label, single-center, parallel group controlled, non-inferiority study in the First Affiliated Hospital of Wannan Medical College (Yijishan Hospital) from Dec 1, 2020 to Dec 30, 2021. All patients provided written informed consent and confirmed their willingness to participate.
Participants
Participants diagnosed with SAT were evaluated in terms of eligibility for inclusion in the study. The diagnosis of SAT was based on clinical features, physical examination and laboratory test results, which includes elevated ESR, serum free-thyroxine (fT4), and free triiodothyronine (fT3) levels with a suppressed thyroid-stimulating hormone (TSH) level. The diagnosis and resolution of SAT were confirmed by thyroid ultrasonography in all patients.
In this study, patients with the side effects of steroids and pregnancy were excluded.
Randomization and Blinding
We randomly assigned patients to the 1 of the 2 groups in a 1:1 ratio using random block with stratification (http://www.randomization.com). A study coordinator was only involved in randomization and allocation. Randomized numbers were sealed in opaque envelopes in order and kept concealed until the study day. The investigator made a phone call to the study coordinator the day before patient enrollment to confirm the patient's assignment. Assessment of study outcomes was analyzed by the investigator. The purpose of this study was to explore the efficacy and safety of different doses of PDN for the treatment of SAT, and complete blinding would not be meaningful for either the investigators or the patients. Therefore, this study was conducted as an open-label study in which only laboratory personnel were blinded and no randomization groups were allowed to be removed.
Procedures
The dose tapering protocol is 15-10-5-2.5 and 30-20-15-10-5 mg per week for the LD and RD group, respectively. Treatment duration has been extended by 1 or 2 weeks with the continuation of the same dose in the previous week in patients who complained having an increased pain during the dose tapering. The dose was reduced from the week of normalized ESR (about 7-14 days), and the first reduction was 1/3 of the dose. Based on clinical experience and examination results, most patients experienced complete disappearance of pain within 5-10 days after PDN treatment, either at the small initial dose or at the regular initial dose. Therefore, if patients experienced recurrence of symptoms during the dose reduction, we adjusted them to the regular initial dose.
After the patients were enrolled, we conducted weekly face-to-face or telephone follow-up visits. The results of the follow-up visits were recorded on the patients' informed consent forms. Each participant received a final visit before the completion of the trial.
Outcomes
The primary outcome of the study was assessed the clinical efficacy of different doses of PDN in the treatment of SAT, which was the time (days) required for PDN treatment. The secondary outcomes were: (I)percentages of participants experiencing relapse during and after treatment; (II)the mean score for the Morisky Medication Adherence Scale-8©(MMAS-8); (III)time(days) required for symptoms to resolve;(IV) cumulative PDN dose (mg);(V) the mean erythrocyte sedimentation rate (ESR) at 2 weeks and baseline.
Safety outcomes included the occurrence of all adverse drug events, which was elevated blood pressure, elevated blood glucose, mood swings, facial acne, insomnia, weight gain, hypothyroidism, etc.)
Statistical analysis
Primary and secondary outcomes will be analyzed based on the full-analysis-set, with missing values carried forward using the last observation and a supportive sensitivity analysis using the last follow-up. Safety outcomes will be analyzed based on the safety data set, including the set of all subjects who received at least one treatment after randomization. All outcomes will also be analyzed according to each protocol set. The primary outcome was the time (days) required for PDN treatment. The purpose of this study was to show that low-dose PDN therapy may be sufficient to achieve complete recovery and better outcomes in SAT. Therefore, we hypothesized that the patients randomized to LD group is non-inferior to that of patients randomized to RD group. To assess for non-inferiority, non-inferiority delta of 7(days)was chosen for the following reasons. In previous study(6,7), the mean time was around 36 days with a standard deviation of 12 days. Given the Δ of 7, we calculated the sample size to test the null hypothesis H0: μ1 − μ0 > 7 versus the alternative hypothesis H1: μ1 − μ0 ≤ 7, where, μ1 − μ0 is the mean difference of the time period between the LD group and the RD group (non-inferiority test). Based on 80% power and α = .05 significance level, a total of 40 participants were included in each group, for a total of 80 participants, considering a 5% drop-out rate. However, because of the study was conducted in the midst of a Covid-19 virus pandemic, which prevented some patients from coming to the hospital and the approximate annual number of hospitalized patients with SAT in our department was 40. Finally, only 45 participants were recruited in this study.
All statistical analyses were performed with Statistical Package for the Social Sciences (SPSS) v 26.0 software (SPSS, Inc, Chicago, IL). All tests were performed with the a-level set to 0.05 (2-tailed). Data are presented as mean±standard deviation, median (inter-quartile range), or numbers (%) as appropriate. The difference between the randomized groups on baseline variables was assessed using standardized mean difference (SMD). We assessed the effect of LD group vs. RD group on the primary outcome, using a 2-sample t test. Non-inferiority was claimed when the upper limit of the 2-sided 95% confidence interval was below the non-inferiority margin. Time period, the mean score for the MMAS-8, neck pain, tenderness at the region of thyroid gland and cumulative PDN dose(mg) were presented as point estimate and corresponding 95% confidence interval (4). We assessed the treatment effect on normally distributed continuous variables using the 2-sample t test.
Randomized groups were compared on the percentage of relapse during treatment and after treatment using a χ² test or Fisher’s exact test where appropriate, which were also used to assess differences in the incidence of adverse events between treatments. We also used Kaplan-Meier curves and log-rank test to analyze the incidence of recurrent symptoms during treatment and the relapse rate after discontinuation of the drug; P<0.05 was considered a statistically significant difference.