Minimally Invasive Percutaneous Nephrolithotomy versus Retrograde Intrarenal Surgery for Upper Urinary Stones: A Systematic Review and Meta-Analysis

Minimally invasive percutaneous nephrolithotomy (mini-PCNL) and retrograde intrarenal surgery (RIRS) are both alternatives for PCNL to treat renal calculi. This study is aimed at comparing the stone-free rate (SFR) and other surgery parameters of two approaches for treating upper urinary calculi. We performed this meta-analysis in September 2016 by searching studies about mini-PCNL and RIRS for treating upper urinary calculi in various databases, and RevMan v.5.3 was applied. Three randomized controlled trials and ten nonrandomized trials were included, involving a total of 1317 patients. Meta-analysis showed that mini-PCNL group led to a higher SFR [odds ratio: 1.96; 95% confidence interval: 1.46–2.64; P < 0.00001] but brought a larger postoperative decrease in hemoglobin levels compared with RIRS. RIRS provided a shorter hospital time. There was no significant difference in operation time. Higher postoperative complications were detected in the mini-PCNL, but the difference was not significant. Grade I and III complications did not vary between two procedures, but grade II complications were of lower incidence in RIRS group. In the light of these results, compared with RIRS, mini-PCNL provided significantly higher SFR and efficiency quotient for managing calculi; however, it resulted in higher incidence of postoperative complications, larger hemoglobin drops, and longer hospital stay.


Introduction
Kidney calculi is a common urological disorder which is characterized by high recurrence rate [1]. The stone movement leading to renal colic and the obstruction by calculi could result in kidney function loss. Recently, the incidence of kidney calculi has been on the rise in China, probably caused by the changed climate and environment. For releasing the obstruction, urologists choose different treatments for different size calculi diameter from less than 0.6 cm to more than 3.0 cm. As the guidelines recommend, percutaneous nephrolithotomy (PCNL), of which standard access tracts are [24][25][26][27][28][29][30], is a recommended management of patients with renal or ureteral stones more than 20 mm or and for smaller stones (10-20 mm) of the lower pole stones when anatomic factors make extracorporeal shockwave lithotripsy (ESWL) unfavorable. Although PCNL is suggested as a standard method for its excellent stone-free rate, there is still few surgical drawbacks that may compromise its efficacy [1]. Table 1, we rated the level of evidence (LE) of every included trail according to the Oxford Centre for Evidence-Based Medicine Criteria [24]. The qualities of nonrandomized controlled trials (non-RCTs) were assessed according to Newcastle-Ottawa Scale (NOS), and RCTs qualities were assessed by the Jadad scale [25,26]. Two reviewers performed the procedure independently and all disagreements were resolved by consensus.

Data Extraction and Statistical
Analysis. All metaanalyses were performed to assess the overall outcomes using Review Manager Software (RevMan v.5.3, Cochrane Collaboration, Oxford, UK). Extracted data for the analysis included stone-free rate, operative time, hemoglobin drop, length of hospital stay, and postoperative complication rate. If standard deviations were not reported we estimated data according to Hozo SP [27], and if standard deviation (SD) could not be calculated from available data, we asked authors to supply the data. For evaluating dichotomous variables, we chose odds ratio (OR), and, for continuous parameters, weighted mean difference (WMD) or standardized mean difference was used. The Cochrane chi-square test and inconsistency ( 2 ) were used to evaluate the heterogeneity among studies. Pooled estimates were calculated with the fixed-effect model for stone-free rate, and the other outcomes were calculated with the random-effect model due to the heterogeneity. Funnel plots were generated using RevMan v.5.3 to evaluate published bias of literatures. In addition, we included highquality studies (RCTs and high score non-RCTs (NOS: 7 of 9 points)) into meta-analysis for sensitive analysis.

Study Characteristics and Quality.
Thirteen trails were selected for analysis including a total of 701 mini-PCNL cases (53.2%) and 616 RIRS cases (46.8%) in our meta-analysis. The literature screening process is shown in Figure 1.
The baseline characteristics of the included studies are shown in Tables 2 and 3. Surgical technique for mini-PCNL differed in aspect of tract size and lithotripsy. As shown in Table 1, the studies consisted of three RCTs (LE: 2b) [11,16,22], two matched-pair analysis trails (LE: 3b) [13,20], and eight case control studies (LE: 3b) including 4 prospective [14,17,21,23] and 4 retrospective studies [12,15,18,19]. In terms of the methodological quality, eight and two of the nonrandomized studies were relatively high (NOS: 7 of 9 points and 6 of 9 points) and medium (NOS: 5 of 9 points and 4 of 9 points), respectively. The three RCTs were of medium quality (Jadad scale: 3 of 5 points). In 3 studies, mini-PCNL and RIRS were compared between patients with a single stone. There were also two studies comparing mini-PCNL with RIRS for proximal ureteral stones, whereas the rest were compared for renal calculi. In general, the preoperative demographic characteristics such as mean age (51.80 ± 14.80 versus 49.9 ± 14.29 year) were comparable between mini-PCNL and RIRS. (SFR). SFR was evaluated with fixedeffect model, and the comparison of stone-free rate between the mini-PCNL and RIRS group is shown in Figure 2. All involved studies reported postoperative stone-free rate, and the result suggested that mini-PCNL group provided a significantly higher SFR than RIRS group (OR: 1.96; 95% CI, 1.46-2.64; < 0.00001) with no significant homogeneity ( 2 = 42%). As the stone location subgroup meta-analysis shows (Figure 3), mini-PCNL has significantly higher SFR than RIRS in any location (OR: 2.13; 95% CI, 1.53-2.96; < 0.00001). Figure 4 shows that mini-PCNL can lead to higher stone clearance in both 1-2 cm (OR: 2.01; 95% CI, 1.27-3.19;

Operative Time.
Twelve studies combined had reported operative time, which was evaluated with random effect model. As the meta-analysis result shows in Figure 5(a), there was no remarkable difference between mini-PCNL and RIRS (WMD, −2.21; 95% CI, −12.62-8.20; = 0.68).

Hospital Stay.
In 10 studies there were available data to extract in terms of hospital stay, which was analyzed by random effect model. As shown in Figure 5(b), hospital stay for RIRS is shorter than mini-PCNL (WMD: 1.63 d; 95% CI, 0.98-2.28; < 0.00001).

Hemoglobin (Hb) Drop.
Hb drop was analyzed by random effect model, and the result was shown in Figure 5(c). Six studies provided accessible data about Hb drop and the metaanalysis shows that RIRS led to less Hb drop than mini-PCNL (WMD, 0.60; 95% CI, 0.32-0.88; < 0.0001).

Complication
Rate. All studies reported available data for the assessment of the complications between mini-PCNL and RIRS group. Random effect model and OR were used for statistical analysis and the result is shown in Figure 6. RIRS has a lower complication rate than mini-PCNL (OR: 1.62; 95% CI, 0.92-2.88; = 0.10). Furthermore, we analyzed the complications according to Clavien-Dindo Classification (Table 4) to evaluate minor indisposition and major complication [28,29]. As the results show, there is no significant difference between grade I and grade III complications between the two groups (OR: 1.24, 95% CI, 0.66-2.32, = 0.51; OR: 1.41, 95% CI, 0.97-2.04, = 0.77); however, we observed a significantly lower incidence of grade II complications in RIRS group (OR: 1.63; 95% CI, 1.01-2.63; = 0.04) (Figure 7).

Sensitivity Analysis.
The sensitivity analysis suggested that the results of this meta-analysis were relatively stable (Table 5). When only RCTs and high score non-RCTs (NOS: 7 of 9 points) were included, most of the outcomes including stone-free rate, operative time, total postoperative complications, and grade I and III surgery complications were not greatly changed. Meanwhile, significant differences of grade II complications, hemoglobin drop, and hospital stay between two groups were not found because of the reduced sample capacity. It is notable that even if the significant differences  were no longer detectable in the sensitivity analysis, the tendency of meta-analysis stayed in the same direction.

4.7.
Publish Bias Analysis. The funnel plot (Figure 8) showed an apparent asymmetry, which suggested the existence of a potential publication bias.

Discussion
With high technological advancement, urologists who take charge of urolithiasis are in possession of high technique instruments, which leads to safer and more effective lithotripsy. So far PCNL is considered to be the recommended therapy for large stones > 2.0 cm by both AUA and EAU        guidelines. Furthermore, with the development of the "mini-PCNL" procedure, smaller access sheaths (≤20 F) are becoming increasingly popular for its relative safety. Besides, recent reports suggested that RIRS is a safer approach that could lead to less complications and Hb drop than normal tract PCNL. We conducted this meta-analysis to systematically analyze the outcomes of two miniature procedures, mini-PCNL and RIRS, which cause considerably lesser trauma than standard PCNL, to find which one could lead to better efficacy and safety. And, to the best of our knowledge, this meta-analysis is an update analysis comparing these two modern minimally invasive approaches applying for upper urinary stone.
SFR is the most important parameter for estimating the efficacy of two approaches. According to the synthesis analysis of data, mini-PCNL has a higher SFR than RIRS group though there were various imaging modalities to identify. Stone-free rates are correlated with the lithotripsy and the location or size of stones. Seven inclusive studies used only laser to dispose stones and others made use of pneumatic or ultrasound waves to fragment calculi.  To evaluate the different locations that may impact SFR of two procedures, a subgroup analysis was performed. As Figure 3 shows, the proximal and low pole subgroups did not show any remarkable advantage of two approaches, while all locations showed that mini-PCNL has a significant advantage in SFR. Additionally, one stone size subgroup analysis was performed to estimate the impact on meta-analysis; results showed that mini-PCNL has more efficiency stone clearance in both 1-2 cm and >2 cm groups ( Figure 4). Besides, mini-PCNL carries high efficiency quotient (EQ) ( Table 6), which was related to SFR, percentage retreatment, and percentage requiring an auxiliary procedure (as the following formula) [30,31], reported by three included studies (Table 5) [14,16,22]. However, De et al. had performed a meta-analysis that compared PCNL and RIRS for managing kidney stone and the results showed that RIRS can provide higher stone-free rates compared with mPCNL, which was opposite to our results [10]. It should be noted that only 5 literatures were included in the previous study, and the "mPCNL" in this study referred to micronephroscope which is 4.85 Fr and mini-PCNL from 11 to 19 Fr. This diversity of definition and sample size may result in the outcomes' difference between our meta-analysis and the previous study. Including more relative studies, the outcome would become more reliable.
(1)  Operative times were reported in 12 involved studies, and six studies indicated that mini-PCNL spent shorter operating time, while four studies favored RIRS. The overall metaanalysis showed that two procedures brought no significantly varied operation time; meanwhile, we noticed the heterogeneity in this section was as high as 97%, mainly led by Kumar [12,13,16,22]. If the four studies were excluded, heterogeneities would be declined to 70%, and the preference would favor mini-PCNL procedure. Operative time is closely related to nuance in surgical techniques and doctors' experience, different surgeons in different centers provided a large variation in operative time, and a significant heterogeneity was proved from twelve inclusive studies.
The overall analysis found that RIRS resulted in shorter hospital stay than mini-PCNL group. The reason for this   [14] 0.904 0.523 -Kumar et al. [16] 0.915 0.842 0.01 Gu et al. [22] 0.830 0.500 -EQ = efficiency quotient. difference might be less invasive caused by RIRS. Moreover, it carries lower complication rate and hemoglobin drop. The size of the tract is one key factor for blood loss during endourology surgery, so mini-PCNL with miniature tract can reduce bleeding and the risk of blood transfusion compared to normal tract PCNL [32]. Besides, the overall analysis of the literature suggested that RIRS resulted in less hemoglobin drop than mini-PCNL. Accordingly, RIRS has a high efficiency for the management of intrarenal stones with a slight complication to patients [33,34].
All trails have made the comparison of postoperative morbidity between mini-PCNL and RIRS. The results showed that RIRS provided a lower complication rate than mini-PCNL; however, the difference had no significance. The complications of mini-PCNL are similar to those of PCNL; bleeding, pain, and fever are very common [35][36][37][38].
Furthermore, we performed a subgroup meta-analysis of postoperative complications, classifying them into grades I, II, and III based on Clavien-Dindo Classification, between the two groups [28]. As Table 4 shows, grade I represents the morbidities that needed no pharmacological or surgical treatment, which could easily occur after operation, and grade III means complications requiring surgical, endoscopic, or radiological intervention, which rarely occur after lithotripsy operation. Thus, we did not observe a remarkable difference in comparison of grade I. Grade III complications were only observed in 5 studies, and the result showed that RIRS has a potential safety on severe complications. As for grade II complications, mini-PCNL has a significantly higher rate than RIRS according to our meta-analysis, which means RIRS was probably safer with respect to middle or severe morbidities after operation, and, in term of light deviations, the incidences of mini-PCNL and RIRS were similar.
There are several limitations in the present meta-analysis. In our systematic review and meta-analysis, we included the currently available comparative studies. Although we have done the sensitivity analysis to show that the results were relatively stable, there is still some bias of our conclusion caused by non-RCTs. Besides, heterogeneities among involved literatures, which may relate to diverse calculi size and location, different tract size, and lithotripsies, could lead to some limitations in our meta-analysis. In addition, most of the included trials failed to describe complications with the same criteria and blinding procedures in detail, and this might lead to conclusion bias, as the more details the literatures describe, the more credible the conclusion will be concluded. However, to the best of our knowledge, this study is one update review and meta-analysis to compare mini-PCNL and RIRS for treating renal calculi. We believe the results of the present meta-analysis could help urologists make better clinical decisions to manage stone disease patients.

Conclusions
From this meta-analysis, we found that both mini-PCNL and RIRS can provide safe and effective treatment for renal calculi patients. In the light of these results, compared with RIRS, mini-PCNL provided significantly higher stonefree rate and efficiency quotient for management of upper urinary calculi, however, could increase the incidence of postoperative complications and the average hospital stay.

Disclosure
This article does not contain any studies with humans or animals performed by any of the authors. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.