18 The heterogeneity across studies was tested by using I-square and Cochran’s Q tests. A P value <.10 for chi-square testing of the Q statistic or an I-square >50% was regarded as the existence of significant heterogeneity. 19 We performed a subgroup analysis
according to the different dosages, regimens, and preparations of PRP, as well as the severity of knee degenerative lesions. A sensitivity analysis was conducted by removing some studies with extreme effect size values to observe whether the action caused serious changes in the overall buy CB-839 result. We used a funnel plot and the Begg’s test to examine the publication bias, which was defined as the tendency for positive trials to be published and the tendency for negative and null trials not to be published. 20 All analyses were performed by using Stata 10.0. a Of the 73 nonduplicate citations identified from the literature, 18 clinical trials were screened for eligibility (fig 1). One study21 was excluded because PRP was introduced by performing a miniarthrotomy (not by an injection technique), and the other study22 was removed because of an inability to NVP-BEZ235 cell line extract data from box plots. An assessment of the remaining 16 articles revealed that 8 used a single-arm, open-label, and prospective follow-up design.23, 24, 25, 26, 27, 28, 29 and 30 Two
quasi-experimental studies31 and 32 and 4 randomized controlled trials33, 33, 34, 35 and 36 compared PRP with HA injections, 1 randomized controlled trial compared different doses of PRP with normal saline,37 and 1 quasi-experimental trial compared a single-spinning approach of PRP with Gemcitabine molecular weight a double-spinning approach.38 The 16 included trials comprised 26 treatment arms, of which
18 used PRP treatments, 7 administered HA, and 1 used saline for placebo controls. Regarding knee-specific outcome measures, we extracted data from IKDC in 8, KOOS in 1, and WOMAC in 7 of the 16 studies. The 16 included studies had a total enrollment of 1543 patients, 840 of whom (54.4%) were men (tables 1 and 2). The duration from the onset of knee pain to registration in each trial was listed from 3 months to more than 1 year. The follow-up period ranged from 6 to 24 months, and the latest point of assessment for most trials was at 12 months after PRP injections. Most studies recruited patients with knee OA with a severity less than grade III on the Kellgren-Lawrence (KL) scale, and some of them also enrolled participants affected by cartilage degenerative lesions with a grade of 0 on the KL scale. Compared with the preinjection condition, we found a pooled effect size of 2.31 (95% CI, 1.53–3.09) at 2 months, 2.52 (95% CI, 1.94–3.09) at 6 months, and 2.88 (95% CI, .97–4.79) at 12 months, which all favored the status after PRP treatment (fig 2). If we deleted an outlier with an extremely high effect size,24 the beneficial effects from PRP injections remained, with an effect size of 1.84 (95% CI, 1.53–3.09) at 2 months, 2.19 (95% CI, 1.73–2.