1. Characteristics and quality of the included studies
As listed in Table 1, a total of 8 prospective studies published from 2001 to 2019 were finally enrolled into the present analysis [29, 32-38], including 7 RCTs and 1 prospective case control study. 3 studies were multi-center studies (NCT02645708, NCT02643134, one not registered). 1 study was after RIRS and the rest 7 studies were after ESWL. A total of 1065 subjects were included in the present study, 528 in the physical therapy group, and 537 in control group.
The baseline information of gender (OR：0.96, 95% CI: 0.71-1.28, p=0.76), age (MD: 0.17, 95% CI: -0.11-0.46, p=0.24) and BMI (MD: 0.17, 95% CI: 0.26-0.60, p=0.45) were comparable in over population, details were showed in Table 1. The stone sizes were all less than 2cm prior to lithotripsy, except not available in one study. The details of stone locations and stone fragments numbers after lithotripsy were not available in all the studies. And LE of all RCTs was 1b and the score of methodological quality ranged from 3 to 5. Risk of bias assessment was described in Figure 2.
Details of different physical techniques, including the techniques name, machine used, time to first session physical therapy, time for each session of percussion, inversion angle, water drinking, whether medicine applied and the follow-up were presented in Table 2. To be specific, 4 studies with use of EPVL (external physical vibration lithecbole), 2 with PDI (percussion, diuresis and inversion), 1 with MP (mechanical percussion), and 1 with HDI (hydration diuresis and inversion). Discrepancies were noted in different PDI techniques, while EPVL had a similar protocol. Drinking 500-3000ml water was advised in 7 studies prior to the physical therapy to promote stone expulsion by diuresis. As for the time to perform the first session physical therapy, 4 studies performed physical therapy immediately after ESWL, 3 within postoperative 1 week, and 1 in 3 months after ESWL. 1 to 4 sessions of physical therapy in total were performed. Combined with percussion, an inversion angle of 12-60° tilt was selected. Percussion parameter was definite in EPVL, with a power of 40W, vibration frequency of 2800-3500 blows per minute, amplitude of 5mm, while these parameters were not available or consistent in other physical therapy procedures. The percussion lasted 10-20 minutes for each session. The definition of stone free status was no non-intervention of stone fragments under radiography (KUB, CT or ultrasonography).
2. Meta-analysis results
2.1 stone free rate
As shown in Figure. 3.A, a higher SFR was noted in physical therapy group (OR：3.50, 95% CI:2.55-4.81, p<0.0001) than non-intervention group. There was no significant difference in heterogeneity (I2 = 0%, p = 0.54). Sensitivity analysis in Figure 3.C showed that omitting any study would not change the final results or cause large elevation of quantitative difference. But minimal publication bias was detected in contour-enhanced funnel plots (t=1.85, p=0.113, Figure 3.B), since the study of Pace et.al showed a marked deviation when compared with other studies. Given that the study of Pace et.al was performed in 2000, which was much earlier than others, and the results was far abnormal when compared to others, we decided to remove it from the final pool analysis. After excluding this study from final analysis, SFR in physical therapy group was still higher when compared to non-intervention group (OR: 3.38, 95%CI: 2.45-4.66, p<0.0001), and there was no significant difference in heterogeneity (I2= 0%, p= 0.76, Figure 4). Additionally, higher rate of first two-day stone expulsion was observed in physical therapy group than non-intervention group (OR： 2.07, 95% CI: 1.36-3.16, p = 0.0007), as showed in Figure 5.A.
Overall complication rate was comparable as showed in Figure 6 (OR: 0.84; 95% CI: 0.62-1.13; p = 0.237), there were no significant difference in terms of hematuria (OR: 0.84; 95% CI: 0.54-1.29; p = 0.423), dizziness (OR: 2.88; 95% CI: 0.89-9.39; P = 0.078), lumbago (OR: 0.61; 95% CI: 0.31-1.19; P=0.146) and urinary tract infection (OR: 0.73; 95% CI: 0.39-1.36; P= 0.328), respectively.
3. Subgroup meta-analysis
3.1 SFR in different physical therapy
As presented in Figure 4, when different techniques of physical therapy were classified into EPVL and PDI, the SFR no matter in EPVL (OR: 3.47; 95% CI: 2.24-5.37; p < 0.0001) or in PDI (OR: 3.24; 95% CI: 2.01-5.21; p < 0.0001) was higher than SFR in non-intervention group.
3.2 SFR in different time point
As depicted in Figure 5.B, an overall higher SFR was observed in physical therapy group (OR: 2.48; 95% CI: 1.99-3.10; p < 0.0001) when compared to non-intervention group. Specifically, in the first week (OR: 1.93; 95% CI: 1.35-2.76; p =0.0003), in second week (OR: 2.59; 95% CI: 1.83-3.66; p < 0.0001), and in first month (OR: 3.51; 95% CI: 2.17-3.10; p < 0.0001).
3.3 Influence of stone location to SFR
A higher SFR in physical therapy group was noted in the overall analysis (OR: 3.19; 95% CI: 2.27-4.50; p < 0.0001) in terms of different stone locations. In the subgroup meta-analysis of the stone fragments location (Figure 7), physical intervention group owned a higher SFR in LCS (OR: 3.51; 95% CI: 2.21-5.55; p < 0.0001) and UPS (OR: 2.79; 95% CI: 1.62-4.81; p= 0.0002). UCS and MCS were mixed into pooled analysis as others due to few information, no significant difference was noted (OR: 3.39; 95% CI: 0.77-15.03; p=0.108).