Skip to main content
We thank Mr. Choudary, Mr. van Doorn, and Dr. Elshout for your interest in our article; active discussion of the medical literature is a cornerstone of medical education. Here are our responses to your remarks.
The decision to use a score ≥ 3 out of an ordinal scale 0-6 was not arbitrary; rather it was intended to include participants with moderate to severe knee pain with three specific considerations. 1) We aimed to detect a difference between groups based on our clinical and research experience while working within practical limitations of time and money. As for all such RCTs, this is an informed estimate that weighs potential effect at a given level of disease severity; we were pleased that our severity estimate was correct and we were able to detect differences between groups. 2) Patients with low or intermittent levels of pain may not be as adherent to an injection protocol, even one as minimally invasive as this one. Here again, our judgement proved correct; adherence was very high. 3) The 3/6 eligibility pain criterion allowed comparison to a prior benchmark study, facilitating comparison to the prolotherapy literature.1 In general, we suspect that prolotherapy with dextrose would also be effective for patients with lower levels of pain and would welcome efforts, necessarily larger to detect smaller differences, to test this hypothesis.
We selected eligibility criteria to best detect differences between groups in this efficacy trial, including the exclusion criterion of BMI ≥ 35 kg/m2. The eligibility criteria applied in an efficacy trial are intended to identify a population of interest in whom an intervention has the greatest likelihood to produce a clinically important and statistically significant effect.2 To improve the external validity in clinical settings, an effectiveness trial should be conducted as the next step, after knowing the internal validity of the efficacy trial.3 Therefore, while we acknowledged the exclusion of BMI ≥ 35 kg/m2 as a limitation to generalizability, we also conducted a moderation analysis to test on the effect of BMI. We found that in this study, baseline BMI was unlikely to affect treatment outcomes; the same is true in a prior RCT.1 However, we acknowledge that BMI may reasonably be expected to affect outcomes; a prior pilot study noted that BMI ≤ 25 kg/m2 (p=0.04) was associated with greater improvement in WOMAC scores.4 Additional work is needed here to fully assess co-morbidities and associations. Regarding “specific comorbidities and lifestyle factors of the excluded patients”, no one was excluded after randomization. Reasons for exclusion prior to randomization are given in Figure 1. Most “excluded” patients were simply those who declined to participate. While it is scientifically reasonable to be thorough, practical considerations sometimes prevent ultimate thoroughness. In this case they simply said “no thanks” and no further information was obtainable.
We don’t understand your concern about effect size. The effect size in this trial was measured by the difference-in-difference between DPT and NS. To clarify, the primary outcome of interest was the WOMAC pain subscale, not the WOMAC composite scale. If we apply the MCID of pain intensity as recommended by IMMPACT 20085, a 10% improvement is regarded as minimal important changes. In this trial, the difference-in-difference at 1 year was -10.7, equivalent to 11%, exceeding the MCID for pain. However, we purposely discussed the MCID of the WOMAC composite scale instead of the pain subscale, because we believe it’s important to give an overall benefits of DPT, not only on knee pain, but also knee function. While participants met MCID criteria, we would also caution against the imposition of an excessively granular approach to benchmarks of clinical improvement. Content experts warn that, given the complexity of knee OA pain, these benchmarks are best used as more general gauges of change.6
We trust our reply is useful and addresses concerns.
References:
1. Rabago D, Patterson JJ, Mundt M, et al. Dextrose prolotherapy for knee osteoarthritis: a randomized controlled trial. The Annals of Family Medicine 2013;11(3):229-37.
2. Hulley SB, Newman TB, Cummings SR. Choosing the study subjects: specification, sampling, and recruitment. Designing clinical research 2007;3:27-36.
3. Van Spall HGC, Toren A, Kiss A, et al. Eligibility Criteria of Randomized Controlled Trials Published in High-Impact General Medical JournalsA Systematic Sampling Review. JAMA 2007;297(11):1233-40. doi: 10.1001/jama.297.11.1233
4. Rabago D, Zgierska A, Fortney L, et al. Hypertonic dextrose injections (prolotherapy) for knee osteoarthritis: results of a single-arm uncontrolled study with 1-year follow-up. The Journal of Alternative and Complementary Medicine 2012;18(4):408-14.
5. Dworkin RH, Turk DC, Farrar JT, et al. Core outcome measures for chronic pain clinical trials: IMMPACT recommendations. Pain 2005;113(1):9-19.
6. Tubach F, Ravaud P, Baron G, et al. Evaluation of clinically relevant changes in patient reported outcomes in knee and hip osteoarthritis: the minimal clinically important improvement. Annals of the rheumatic diseases 2005;64(1):29-33.