Background

Scholars agree that the partial nephrectomy for small renal mass is advantageous over radical nephrectomy in terms of renal function [1, 2]. The conventional partial nephrectomy technique includes the clamping of the renal artery (on-PN); this method allows tumor resection and renal reconstruction in a relatively bloodless field [3,4,5]. However, occluding the renal artery places the remaining nephrons at risk of ischemia−reperfusion injury and mitigates the renoprotective purpose of surgery [3, 6]. Shorter ischemia intervals have been correlated with better renal functional preservation [7, 8]. Off-clamp partial nephrectomy (off-PN) has been a common strategy to avoid ischemia in small renal tumor. Although a consensus has been reached on off-PN risking more blood loss, the impact of the two methods on the change of postoperative short- and long-term renal function remains unclear [1, 9]. Therefore, we systematically searched and analyzed the clinical studies comparing off-PN with on-PN for small renal masses published until January 2018 to assess the methods’ impact on short- and long-term renal function.

Methods

The literature search methods, inclusion and exclusion criteria, outcome measures, and statistical analysis methods were well defined in a prospective protocol.

Literature-search strategy

A literature search was performed in January 2018 with no restriction to region or language. The primary sources were the electronic databases of PubMed, EMBASE, and the Cochrane Library. The following terms and their combinations were searched as follows: (“partial nephrectomy” OR “nephron sparing surgery” [Title/Abstract]) and (“clamp*” OR “ischemia” [Title/Abstract]). Our computer search was supplemented with manual searches of reference lists of all retrieved review articles. When multiple studies were reported by the same institution and/or authors, the most complete report was included in our analysis.

Inclusion and exclusion criteria

All retrospective or prospective comparative studies (cohort or case−control studies) containing a comparison of off-PN and on-PN with or without a third group, such as cold-ischemia partial nephrectomy, and those providing available data to assess postoperative renal function, were included. Animal experimental studies, editorials, letters to the editor, review articles, case reports, conference abstracts, studies without available data about postoperative renal function, and non-comparative studies were excluded.

Data extraction and outcome measures

Two of the authors extracted and summarized data from the included studies independently. Any disagreement was resolved by mutual discussion with another two adjudicating senior authors.

The primary outcomes were short-term change of estimated glomerular filtration rate (eGFR), short-term change of Cr level, long-term change of eGFR, long-term % change of eGFR, and long-term change of Cr level.

The secondary outcomes were postoperative long-term eGFR, short-term eGFR, short-term Cr level, postoperative increase in CKD (stage≥3), and postoperative acute renal failure (ARF).

Quality assessment and statistical analysis

The modified Newcastle−Ottawa scale [10, 11] was used to assess the methodological quality of nonrandomized studies comprising three factors: patient selection, comparability of study groups, and assessment of outcome. Every study was scored from 0 to 9.

The level of evidence was assessed on the basis of the criteria enacted by the Center for Evidence-Based Medicine in Oxford used to rate the included studies [12].

Review Manager 5.3 (Cochrane Collaboration, Oxford, UK) was used to perform all meta-analysis. The odds ratio (OR) with 95% confidence interval (CI) was used to compare the dichotomous variables consisting of ARF and postoperative increase in chronic kidney disease (CKD; stage≥3), and the weighted mean difference (WMD) with 95%CI was used to compare the remaining continuous variables. The corresponding authors were contacted when the data were missing or incomplete. The technique summarized by Hozo et al. was used to convert medians to means [13].

Statistical heterogeneity was considered significant when the Cochrane Q test p value was < 0.10. The standard heterogeneity test, I2 statistic, was used to assess the consistency of the effect sizes. The fixed-effects model was used when no significant heterogeneity exists between the studies; otherwise, the random-effects model is used [14].

Subgroup analysis was performed to verify the impact of two surgical procedures and to assess the efficacy of different studies in more homogeneous subsets in accordance with the sample size in the studies.

Sensitivity analysis was performed by repeating the primary analysis without including the highest scored study or studies when some papers achieved the highest scores at the same time.

STATA SE 12.0 was then utilized to evaluate potential publication bias, which was screened on funnel plots and assessed statistically using the Begg’s test and Egger’s test. The tests were two sided, and the p values of < 0.05 were considered significant statistically.

Results

A total of 3766 patients (off-PN 1197; on-PN 2569) from 25 studies fulfilled the predefined inclusion criteria and were included in this analysis. The detailed process of research screening and selection is shown in Fig. 1. Most eligible studies were designed retrospectively; among these works, 7 included a third group besides the off-PN and on-PN groups. The comparison of preoperative renal function between off-PN and on-PN was not significantly different in 13 studies, whereas that in others differed significantly or did not use statistics between the two groups.

Fig. 1
figure 1

Flowchart of study selection

Full size image

Characteristics of included studies

Table 1 summarizes the characteristics of the included studies [9, 15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38]. Among eligible studies, two were prospective cohort studies (level of evidence: 2b) [28, 35], 19 were retrospective studies comparing contemporary series of patients (level of evidence: 3b) [9, 15, 16, 18,19,20,21, 23, 25, 27, 29,30,31,32,33,34, 36,37,38], and the remaining 4 studies used a historical series as controls (level of evidence: 4) [17, 22, 24, 26]. A total of 14 studies containing a total number of ≤110 patients were considered as small-sample studies [9, 15,16,17,18,19,20, 23, 28, 29, 34,35,36, 38], whereas those containing over 110 patients in all were considered as large-sample studies [21, 22, 24,25,26,27, 30,31,32,33, 37].

Table 1 Characteristics of included studies
Full size table

Quality of included studies

No appropriate protocol was employed in advance to allocate patients and assign treatment in retrospective studies but only based on physician’s discretion. Information on allocation concealment or blinding method was not mentioned in studies. Patients were sequentially distributed individually for off-PN or on-PN by the same surgeon in two prospective studies [28, 35]. A modified Newcastle−Ottawa scale was used to evaluate the risks of bias (Additional file 1). Two prospective studies achieved a proper protocol for the design. The matching about preoperative characteristics was performed in terms of age, tumor size, preoperative renal function, American Society of Anesthesiologists score, gender, tumor side, body mass index, nephrectomy (R.E.N.A.L.) score, and tumor location. Outcomes include the assessment of renal function in the postoperative period. Methods for handling missing data and intention-to-treat analysis were not adequately described in some studies.

Primary outcomes

Within the postoperative 3 months, the short-term decrease of eGFR and short-term increase of Cr level in an off-PN group were significantly less ([WMD: 4.81; 95%CI: 3.53 to 6.08; p < 0.00001] and [WMD: − 0.05; 95%CI: − 0.09 to − 0.00; p = 0.04], respectively) (Figs. 2 and 3). After 6 months or more of surgery, the long-term loss of eGFR and percent decrease of eGFR in the off-PN group remained significantly less than that of the on-PN group ([WMD: 1.26; 95%CI: 0.04 to 2.48; p = 0.04] and [WMD: 2.52; 95%CI: 1.53 to 3.50; p < 0.00001], respectively) (Figs. 4 and 5). However, the long-term increase of Cr level was insignificant in an on-PN group (WMD: − 0.04; 95%CI: − 0.13 to 0.05; p = 0.40) (Fig. 6).

Fig. 2
figure 2

Forest plot and meta-analysis of postoperative short-term eGFR change

Full size image
Fig. 3
figure 3

Forest plot and meta-analysis of postoperative short-term Cr level change

Full size image
Fig. 4
figure 4

Forest plot and meta-analysis of postoperative long-term eGFR change

Full size image
Fig. 5
figure 5

Forest plot and meta-analysis of percent change of eGFR

Full size image
Fig. 6
figure 6

Forest plot and meta-analysis of postoperative long-term Cr level change

Full size image

Secondary outcomes

Within postoperative 3 months, the short-term eGFR was significantly higher in an off-PN group (WMD: 9.72; 95%CI: 6.13 to 13.30; p < 0.00001) (Additional file 2). On the contrary, the short-term Cr level was significantly lower in the off-PN group (WMD: − 0.08; 95%CI: − 0.13 to − 0.02; p = 0.007) (Additional file 3). After 6 months or more of surgery, the long-term eGFR was not significantly different between the two groups (WMD: − 0.18; 95%CI: − 3.89 to 3.53; p = 0.92) (Additional file 4). Postoperative ARF was more likely to occur in the on-PN group (OR: 0.25; 95%CI: 0.10 to 0.63; p = 0.003) (Additional file 5) than in the off-PN group. The morbidity of CKD (stage≥3) in the postoperative period was not significantly different between the two groups (OR: 0.73; 95%CI: 0.39 to 1.39; p = 0.34) (Additional file 6) (Table 2).

Table 2 Results of meta-analysis comparison of off-clamp partial nephrectomy and on-clamp partial nephrectomy
Full size table

Subgroup analysis

Subgroup analysis was performed to evaluate whether the primary outcomes were different in accordance with the sample size (Additional files 7, 8, 9, 10 and 11) (Table 3).

Table 3 Subgroup analysis and sensitivity analysis of primary outcomes
Full size table

In the large sample size subgroup, a significant difference from the original analysis was obtained in all the primary outcomes except for the short-term change of Cr level.

In the small sample size subgroup, the degree of between-study heterogeneity for all primary outcomes decreased except for the long-term change of Cr level, but the significant difference was no longer found in all primary outcomes except for short-term change of eGFR.

Sensitivity analysis

Sensitivity analysis was performed by excluding the highest scored study or studies when given the highest score at the same time on the basis of the modified Newcastle−Ottawa Scale. No change in the significance of any of the primary outcomes was noted in the sensitivity analysis. The degree of between-study heterogeneity dropped down to zero for all the primary outcomes except for the short-term eGFR change, of which the degree of between-study heterogeneity slightly increased from 7 to 17% (Table 3).

Publication bias

A funnel plot of the studies that reported the short-term change of Cr level is shown in Fig. 7. All studies lie inside the 95%CI. Begg’s test and Egger’s test were used to evaluate publication bias. There was no significant bias (Begg’s test: p = 0.655; Egger’s test: p = 0.521).

Fig. 7
figure 7

funnel plot of the studies that reported short-term change of Cr level

Full size image

Discussion

Nephron-sparing surgery has been the recommended treatment of clinical T1a renal masses and favored over radical nephrectomy in patients with T1b tumors when technically feasible [39]. Every minute counts to preserve renal function when the renal hilum is clamped during partial nephrectomy [7, 40, 41]. For minimizing ischemic injury as much as possible to preserve the functional renal nephron, avoiding ischemia altogether by performing surgery entirely off-clamp is a good strategy. Many studies concluded that partial nephrectomy for small renal masses can be performed without hilar clamping, although considered as a procedure of potentially increased risk of augmented blood loss [2, 3, 5, 9, 15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38]. However, the impact of off-PN on short- and long-term renal functional residual capacity compared with conventional on-PN remains unclear [27]. Thus, we reviewed the published studies and conducted a standard meta-analysis to evaluate the impact of off-PN on short- and long-term renal function compared with conventional on-PN.

Within the postoperative 3 months, the short-term decrease in eGFR was found in both groups in almost all the eligible studies [19, 24, 26, 28, 32, 34, 35, 38], except for Taweemonkongsap et al. [9] and Kaczmarek et al. [27]. The postoperative short-term eGFR of the off-PN group reported by Taweemonkongsap et al. [9] and Kaczmarek et al. [27] achieved an increase compared with preoperative eGFR (mean: 2.225 ml/min/1.73 m2 and 1.4 ml/min/1.73 m2, respectively). In the two studies, preoperative eGFR records in the off-PN group were higher, and all surgeries were robot assisted, which promoted an enhanced and hastened postoperative renal functional recovery. In all included studies, the short-term decrease of eGFR in the on-PN group was more than that in the off-PN group.

Although no significant difference was found for the postoperative short-term eGFR between two groups in the two studies [24, 35], the pooled analysis of postoperative short-term eGFR showed a significantly higher eGFR in the off-PN group (WMD: 9.72; 95%CI: 6.13 to 13.30; p < 0.00001). This result was consistent with a significantly higher decrease of the pooled short-term eGFR in the on-PN group.

The pooled analysis for short-term change of Cr level indicated a significantly higher Cr level increase in the on-PN group compared with that in the off-PN group. While most of the included studies showed a higher increase in the on-PN group [15, 17, 20, 23, 24, 26, 28, 34], the remaining study reported by Kane et al. [16] showed a higher increase, which was probably driven more by a 38% larger tumor size than by the effect of renal artery occlusion in the off-PN group than in the on-PN group. All included studies showed a higher postoperative short-term Cr level in the on-PN group than in the off-PN group in accordance with a pooled meta-analysis for the postoperative short-term Cr level; this result also indicates the better renal functional outcome of off-PN.

After 6 months or more of surgery, although long-term decrease of eGFR and percent eGFR in both groups was found in the included studies, the pooled meta-analysis for postoperative long-term eGFR change and percent eGFR change revealed a significantly greater decrease in the on-PN group than in the off-PN group. Between-study heterogeneity was significant for long-term eGFR change, the random-effects model was then utilized to reduce the effect of heterogeneity, whereas the long-term percent eGFR change was only the opposite. Subgroup analysis for both long-term eGFR change and percent eGFR change showed that the included large sample studies [22, 25, 30,31,32,33] and small sample studies [9, 19, 29, 36, 38] was pooled without significant between-study heterogeneity. The same conclusions were drawn in a large-sample subgroup with original pooled analysis, whereas no significant difference was found between two groups in the small sample subgroup. Sample size was considered a reason for heterogeneity, and a large sample subgroup is believed to be close to the truth. Sensitivity analysis was performed by exclusion of the highest scored study or studies in the modified Newcastle−Ottawa Scale. No change in the significance of the outcome was noted.

No significant difference was found between two groups for postoperative long-term Cr change. The random-effects model was used to pool the included studies owing to the significant difference of between-study heterogeneity. We attributed this result to the limited number of included studies. Sensitivity analysis with exclusion of the highest scored study [38] found a significant difference between the two groups about long-term Cr change and a very low degree of between-study heterogeneity. Additional studies are needed to confirm the conclusion.

The pooled data of postoperative increase in CKD (stage≥3) suggests the lack of significant difference between the two groups, and the degree of between-study heterogeneity was moderate. In fact, eGFR after on-PN may not decrease to 60 ml/min/1.73 m2 or lower even when renal ischemia−reperfusion injury truly exists.

The ARF rate is significantly lower for the off-PN group than for the on-PN group. This result may be explained by the injury caused by renal ischemia during operation.

To assess any impact of the highest scored study or studies on the effect on primary outcomes, we performed a sensitivity analysis with exclusion of the highest scored study or studies. Given the moderate degree of between-study heterogeneity, all results performed with the fixed-effects model were similar to those of the original analysis except for the long-term Cr change, which was significantly increased in the on-PN group than in the off-PN group (WMD: − 0.09; 95%CI: − 0.13 to − 0.04; p = 0.0002).

Limitations and strength

The primary limitation of this systematic review and meta-analysis was that no RCTs were included for evaluation and subsequent analysis; hence, sufficient data are difficult to acquire for meaningful results. Moreover, the studies that provide data on the change in postoperative long-term Cr level were exceedingly few to offer a more convincing result than currently attained. In addition, the operations were performed by surgeons with different levels of surgical expertise and different choices of surgical approaches. Finally, patient allocation and treatment assignment were usually based on the physician’s attitude instead of randomized allocation; this aspect led to a significant selection bias.

Although a small number of papers have compared the two surgical procedures, no paper has discussed their long-term impact on renal function; this topic is particularly important to effectively choose the proper treatment when possible. An increasing number of T1-T2a peripheral renal tumors have been detected at a young age, and an improved outcome is meaningful. This meta-analysis was conducted at an appropriate time. A sufficient number of studies have been accumulated for inspection by meta-analytical methods. Studies were identified using multiple strategies; the methodological quality of the studies was evaluated on the basis of strict inclusion and exclusion. Subgroup and sensitivity analyses were performed to analyze the source of heterogeneity. The MOOSE guidelines were used to report our systematic review. Publication bias was not significant.

Conclusions

This meta-analysis demonstrated that off-PN positively impacts short- and long-term renal function relative to that of conventional on-PN. Given the inherent limitations of included studies, large-volume and well-designed RCTS with extensive follow up are wanted to confirm and update the conclusion of this analysis in future.