Comparison of the stone-free rates of minipercutaneous nephrolithotomy, standard percutaneous nephrolithotomy, and retrograde intrarenal surgery for renal stones: A systematic review and network meta-analysis

Background: Retrograde intrarenal surgery (RIRS) and percutaneous nephrolithotomy (PCNL) are performed to treat renal stones. PCNL is effective for kidney stones >2 cm but is invasive compared with RIRS. Miniature PCNL (mPCNL) has been used as an alternative treatment for conventional PCNL, and employs a miniature endoscope of 11–18 Fr. We conducted a systematic review of published studies regarding the RIRS, PCNL, and mPCNL treatment modalities, and performed a network meta-analysis of the success or stone-free rates. Methods: The data collected up to January 2016 were searched using PubMed and EMBASE, and references were searched electronically. Two researchers used data extraction formats to extract data on the stone-free or success rates, study design, number of subjects and characteristics, and treatments for renal stones (i.e., RIRS, PCNL, and mPCNL). To evaluate the quality of the studies, the Downs and Black checklist, which is an observational research quality evaluation tool, was used and analyzed by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement. Results: Twenty-five studies were used to compare the stone-free or success rates of RIRS, PCNL, and mPCNL for renal stones. Six comparisons of PCNL and mPCNL, seven of mPCNL and RIRS, and 12 of RIRS and PCNL were analyzed. No difference in the stone-free rate was found between PCNL and mPCNL (OR: 0.95; 95% CI: 0.51–1.9) in network mata-analysis. Between RIRS and mPCNL, the stone-free rate of RIRS was lower than mPCNL (OR: 0.41; 95% CI: 0.021–0.82). Comparison of RIRS and PCNL revealed that RIRS was also lower than PCNL in terms of stone-free rate (OR: 0.43; 95% CI: 0.22–0.82). In the ranking analysis, rankogram showed that mPCNL was ranked as No. 1 and PCNL was ranked as No. 2. The P-score was 0.820 for mPCNL, 0.680 for PCNL and 0 for

treatment of renal stones. However, PCNL and mPCNL showed no difference in the treatment outcome.

Background
Mini-percutaneous nephrolithotomy (mPCNL) was developed for the treatment of pediatric patients with renal stones [1]. In 1997, Helal et al. performed mPCNL for the first time in a 2-year-old pediatric patient using a 15-Fr Hickman peel-away sheath [2]. The current definition of mPCNL implies PCNL performed using access sheaths with a diameter of 14-20 Fr [3]. mPCNL has advantages over standard PCNL in terms of blood loss, postoperative pain, and renal parenchymal damage, which are complications that occur due to the use of larger instruments in standard PCNL [4]. Despite these advantages, however, mPCNL is not considered a preferred technique [5,6]. Additionally, the European Association of Urology (EAU) Urolithiasis Guidelines reports that mPCNL requires a longer operating time and states that additional studies should be conducted on the treatment outcomes [7].
The EAU guidelines recommend extracorporeal shock wave lithotripsy (ESWL) and retrograde intrarenal surgery (RIRS) as first-line treatments for kidney stones < 2 cm in diameter, and PCNL is recommended as the first-line treatment for stones > 2 cm [7]. In the case of a lower pole stone of 1-2 cm, endourologic procedures, including RIRS and PCNL, are recommended for patients with unfavorable factors for use of ESWL. Compared with PCNL and RIRS, ESWL is the only interventional treatment with non-invasive properties and plays a pivotal role in the treatment of urinary stones [8,9]. Compared with PCNL and mPCNL, RIRS has the advantage of being less invasive due to the use of a natural orifice [10]. PCNL is a standard treatment for renal stones >2 cm and can be considered a treatment option for large stones with resistance to shock waves [11]. mPCNL shows reduced complications compared with PCNL when performed in selective patients [12]. Prospective studies and meta-analyses comparing the three surgical treatments (PCNL, mPCNL, and RIRS) and discussing their advantages and disadvantages have been reported, but no network meta-analysis has been conducted to simultaneously compare the three treatments. Network meta-analysis is a research method that can compare multiple treatments using direct and indirect comparisons [13][14][15]. We performed a systematic review and network meta-analysis comparing the success and stone-free rates of PCNL, mPCNL, and RIRS.

Inclusion Criteria
Publicly available RCTs were included according to the following criteria: (1) The study design evaluated 2 or 3 arms, including PCNL, mPCNL, and RIRS for the treatment of kidney stones. (2) Baseline data of the 2 or 3 groups of patients were matched, including the entire number of patients and the value of each index. (3) The results of each treatment were analyzed in stone-free (or success) rate. (4) Standard surgical indications of each treatment for kidney stones were applied. (5) Endpoint outcomes included complication rates. (6) Only English-language articles were allowed. This study has been performed based on the Preferred Reporting Item for System Review and Meta-Analysis (PRISMA) guidelines (S1 Table) [16].

Searches
All literature searches before January 1, 2017 were conducted at PubMed and EMBASE. In addition, cross-reference searches of available studies were carried out to recognize articles that were not searched during the computerized literature search. The progress of the relevant meeting was also retrieved. Percutaneous nephrolithotomy, nephrolithotomy, percutaneous, flexible ureteroscopy, flexible, ureterorenoscopy, retrograde intrarenal surgery, renal stone, urolithiasis, success rate, miniature, mini, and stone-free of Medical Subject Headings (MeSH) and keywords were applied.

Data Extraction and Quality Assessment
An author (JYL) screened all titles and abstracts found by the search strategy. 2 other authors (DHK and HDJ) independently analyzed the all details of each article to ensure that they met the inclusion criteria. If there is any inconsistency between the two researchers, it was resolved through discussion until consensus was reached, or via third party adjudication performed by another author (WSJ). Once the final article group was agreed, 2 authors independently investigated the quality of each study in accordance with the Downs and Black checklist [17]. High scores were considered an indicator of a finequality research.

Heterogeneity Investigation
We derived the heterogeneity of the enrolled studies using the Q statistic and Higgins' I 2 statistic [18]. Higgins' I 2 represents the percentage of the variability in effect estimates that is due to heterogeneity rather than chance. Higgins' I 2 = 100% x (Q-df)/Q in which 'Q' is the Cochran's heterogeneity statistic, and 'df' is the degrees of freedom. An I 2 ≥ 50% means that it exhibits substantial heterogeneity [19]. Heterogeneity of the Q statistic was considered significant if p < .10 [20]. If evidence of heterogeneity is determined, we analyzed the data using a random-effects model. We also created L 'Abbe plots and Galbraith's radial plots to assess heterogeneity [21,22].

Statistical Analysis
We compared the outcome variables by odd ratios (ORs) and their 95% confidence intervals (CIs). The analysis was performed based on non-informative priors for effect sizes and precision. Convergence and lack of autocorrelation were identified after 4 chains and a 50,000 samples burn-in steps; At last, direct probability was derived from an additional 100,000 samples stage. We evaluated the probability of the lowest incidence of clinical events in each group using Bayesian Markov chain Monte Carlo methods. Sensitivity analysis was conducted by repeating the main calculation using fixed-effects models. We calculated and compared estimates for the deviation and deviation information criteria to assess model's goodness of fit. R (R Foundation for Statistical Computing, Vienna, Austria), version 3.3.2, was used for all statistical analyses and network plots were created by netmeta, pcnetmeta, and gemtc packages.

Results
L i t e r a t u r e S e a r c h A total of 259 studies were identified. After screening, 41 articles were assessed for eligibility. 16 of these studies were excluded for the following reasons: 3 articles did not have any data on stone-free rates, 10 articles were review articles, and 3 articles were case report series. Finally, the remaining 25 studies were included in the meta-analysis.

Quality, Heterogeneity and Inconsistency
The mean of the quality scores was 15.12 (Table 1). Briefly, the result of quality assessment shows the quality scores within subscales were relatively low-moderate. In particular, the external validity was not satisfactory for both significant and nonsignificant groups in most studies. There was no evidence of heterogeneity between PCNL and RIRS (Fig. 3); however, heterogeneity was demonstrated between mPCNL and PCNL (I 2 = 51.0%, p = .05; Fig. 4), and between mPCNL and RIRS (I 2 = 50.9%, p = .06; Fig. 5). Thus, the random-effects model was applied using the Mantel-Haenszel method to compare mPCNL with PCNL (Fig. 4) and mPCNL with RIRS (Fig. 5). After applying the effect model, very little heterogeneity was observed in L'Abbe ( Fig. 6) and radial plots (Fig. 7). Node split analysis has not demonstrated inconsistencies in direct, indirect and network comparisons (Fig. 8).

Risk of Bias Assessment Results
The Begg-Mazumdar rank correlation tests revealed no evidence of publication bias between PCNL and RIRS (p = .79), between PCNL and mPCNL (p = .81), or between mPCNL and RIRS (p = .45). Egger regression tests also did not show any publication bias between PCNL and RIRS (p = .87), between PCNL and mPCNL (p = .99), or between mPCNL and RIRS (p = .56). In addition, little publication bias was detected in the funnel plots for each comparison (Fig. 9).

Network Meta-analysis of the Stone-free Rate
No difference in the stone-free rate was found between PCNL and mPCNL (OR: 0.95; 95% CI: 0.51-1.9) in network mata-analysis. Between RIRS and mPCNL, the stone-free rate of RIRS was lower than mPCNL (OR: 0.41; 95% CI: 0.021-0.82). Comparison of RIRS and PCNL revealed that RIRS was also lower than PCNL in terms of stone-free rate (OR: 0.43; 95% CI: 0.22-0.82) (Fig. 8). In the ranking analysis, mPCNL was ranked first and RIRS was ranked third (Fig. 10). The P-score test showed that mPCNL (P-score: 0.820) was better than PCNL (P-score: 0.680) and RIRS (P-score: 0) in terms of the stone-free rate [47].

Discussion
Renal stones are one of the most common urological diseases and are characterized by high recurrence rates [48]. In the case of asymptomatic, tiny renal stones, observation can be performed without any treatment. However, if the stone causes obstruction or infection, is associated with symptoms such as pain or hematuria, or has a high possibility of size increase, treatment is recommended. Interventional treatment for renal stones may be considered if the size is greater than 1.5 cm or if removal of the stone is necessary because of the patient's social situation. The EAU guideline recommends ESWL and RIRS as first-line treatments for kidney stones < 2 cm in diameter and PCNL as the first-line treatment for stones > 2 cm [7]. As surgical procedures, PCNL and RIRS have an anesthetic burden and their invasiveness is a disadvantage, but the stone-free rates of PCNL and RIRS are higher than those of ESWL [49]. The development of surgical techniques and instruments continues to play a major role in the popularization of PCNL and RIRS [48,50]. mPCNL is defined as PCNL performed using Amplantz sheaths with a diameter of 14-20 Fr [3,51]. mPCNL has the advantage of reducing complications that may arise from larger instruments and sheaths [4].
The evaluation of perioperative and postoperative outcomes in the surgical treatment of renal stones is very important. The stone-free rate, operative time, and complications may be appropriate indicators for the perioperative and postoperative outcomes. Among these indicators, the stone-free rate can be one of the most important outcomes to avoid the need for auxiliary treatment as well as complications related to residual fragments. The stone-free rate is mainly correlated with the stone burden; however, most importantly, differences have been reported between different procedures [52]. The stone-free rate is the most important parameter to estimate the efficacy of all the approaches [53].
According to previous reports, PCNL and mPCNL have a higher stone-free rate than RIRS, although various imaging modalities were used.
In 2017, Kang et al. reported a systematic review and meta-analysis, in which updated evidence of stone-free rates of RIRS and PCNL in > 2-cm renal stones were compared with a previous report [54]. In their meta-analysis comparing the success (stone-free) rates between PCNL and RIRS, the forest plot using the random-effects model showed a risk ratio of 1.11 (95% CI:  [57]. They concluded that mPCNL was a safe and effective procedure with a stone-free rate comparable to that of PCNL. In addition, they found that mPCNL resulted in less bleeding, fewer transfusions, less pain, and shorter hospitalization. Another recent systematic review demonstrated that smaller tracts used in mPCNL tended to be associated with significantly lower blood loss or the need for blood transfusion, at the cost of a significantly longer procedure than standard PCNL [58]. The results of our study support these previous study findings. In network meta-analysis of the  [6,50]. In addition, mPCNL often offers even higher stone-free rates than conventional PCNL perhaps because of the vacuum-cleaner effect [60].
The American Urological Association and EAU have not yet presented specific recommendations for the use of mPCNL to treat renal stones. However, previous reported evidence shows that mPCNL can achieve similar outcomes as standard PCNL in the treatment of renal stones > 2 cm [23]. RIRS, which uses a flexible ureteroscope through a natural orifice, can be a competitor of mPCNL for treatment of renal stones, not Staghorn stones [58]. However, it is clear the stone-free rate of mPCNL is superior to that of RIRS and that stones can be removed easily through the vacuum-cleaner effect. However, the long operation time of mPCNL can be a major disadvantage. In our network meta-analysis, the operation time of mPCNL was longer than that of PCNL and RIRS. No significant difference was seen between the complication rates of these three surgeries; however, the total number of complications of PCNL was higher than that of mPCNL and RIRS. Finally, a well-designed prospective study is needed to better understand the use of mPCNL, and to explore its potential to replace PCNL.

Conclusions
PCNL and mPCNL showed the highest success or stone-free rate in the surgical treatment of renal stones. RIRS has the lowest success or stone-free rate and the lowest rank.
Patient selection should be performed based on the complexity of individual patients, and a well-designed prospective study is needed to better understand the use of mPCNL. preparation of the manuscript, or decision to publish. Any opinions and findings expressed in this study are those of the authors alone, and do not necessarily reflect the views of funders.

Availability of data and materials
All relevant data are within the manuscript and its Supporting Information files.

Ethics approval and consent to participate
The study was exempt from requiring the participants' written informed consent because this is systematic review and network meta-analysis. The approval of the Institutional Review Board was also exempted.

Consent for publication
Not applicable. Network plots for the included studies. Eight studies compared PCNL versus mPCNL treatments. Seven trials reported the outcomes between mPCNL and RIRS. Ten studies compared the outcomes between PCNL and RIRS.

Figure 3
Pairwise meta-analysis of the success rate in PCNL and RIRS. Pooled data showed a significantly higher stone-free rate with PCNL compared with RIRS (OR: 2.31; 95% CI: 1.45-3.67; p < .001).

Figure 4
Pairwise meta-analysis of the success rate in PCNL and mPCNL. In PCNL and mPCNL, there was no difference in the stone-free rate (OR: 0.89; 95% CI: 0.46-1.71; p = .73).  Network meta-analysis of the success rate of mPCNL, PCNL, and RIRS and node-splitting analyses of inconsistency.  Rank-probability test of network meta-analyses. In the rank-probability test, mPCNL was ranked as No. 1 and RIRS was ranked as No. 3.

Supplementary Files
This is a list of supplementary files associated with the primary manuscript. Click to download.