Radical hysterectomy with bilateral pelvic lymph node dissection is the standard treatment for early uterine cervical cancer. Most of urologic complications that developed during surgery were associated with gynecologic field. The incidence of perioperative urologic complication is relatively higher in radical hysterectomy than that in other gynecologic surgeries. Intraoperative injuries of the bladder and the ureter as well as vesicovaginal and ureterovaginal fistulas in the postoperative period are important complications of radical hysterectomy [1]. Bladder injuries can occur during bladder dissection to obtain adequate vaginal resection margin. Bladder injuries are frequent urologic complications in radical hysterectomy. Incidence of bladder injuries during radical hysterectomy ranges from 0.4 to 3.7% [2]. Clinically, ureter injury is more important because of its general tendency to be unrecognized. On the other hand, bladder injuries can be easily recognized and repaired immediately. Ureter injuries occur between uterine arteries and bladder. Changes in surgical techniques that limit extreme dissection of ureter injury have reduced ureteral injuries for many years. However, ureteral injury is still a major complication of radical hysterectomy.

Radical hysterectomy can be performed using laparoscopic radical hysterectomy (LRH) or abdominal radical hysterectomy (ARH). LRH is an alternative to ARH because laparoscopic surgery is associated with less postoperative pain, shorter length of hospital stay, better cosmetic outcome, and faster recovery of bowel function compared to open surgery [3]. Although LRS has many advantages, it has not been widely used by gynecologic oncologists due to technical difficulties, long learning curve, and concerns of surgical complications.

Many studies have compared perioperative complications including urologic complications of patients treated by LRH with those of patients treated by ARH. Most of these studies have reported that there is no statistical significance in surgical complications including urologic complications between the two groups. Several studies have reported that LRH is associated with a low rate of perioperative complications [4,5,6,7]. However, the population examined was usually too small to definitively evaluate perioperative complications. A few meta-analysis [8, 9] have evaluated oncologic safety and effectiveness of LRH compared to those of LRH. However, they did not focus on urologic complications. The important step of technical skill in radical hysterectomy is unroofing of ureteral tunnel and wide dissection of periureteral tissue and bladder which is associated with perioperative urologic complication.

Thus, the objective of this study was to determine the risk of intraoperative and postoperative urologic complications in LRH compared to that in ARH through a meta-analysis.

Methods

Literature search

A comprehensive, systemic search for published studies was conducted using Pubmed, EMBASE, and Cochrane library up to December 2018. Predefined keywords used for the search were ‘laparoscopic radical hysterectomy’, ‘abdominal radical hysterectomy’ in combination with ‘urologic complication’, and ‘uterine cervical cancer or carcinoma’. We also scanned bibliographies of relevant articles to locate additional publications. Only articles written in English were included.

Selection criteria

We included comparative studies designed to assess intraoperative and postoperative complications of LRH and ARH in uterine cervical cancer. We also included laparoscopic radical parametrectomy [10] and laparoscopic radical trachelectomy [11] in addition to LRH. Robotic radical laparoscopy (RRH) was excluded. Review, editorials, letters, and meeting abstracts were excluded. Publications including fewer than 10 patients were also excluded. Three randomized controlled studies have been reported. Two of them were excluded since their patient numbers were less than ten. We excluded those studies without available data for perioperative complications. If data were duplicated, the larger study was selected in meta-analysis.

Data extraction

Two researchers independently evaluated the eligibility of all studies retrieved from the database based on predetermined criteria. Disagreements between evaluators were resolved by discussion. Of articles found in the three databases, duplicate articles and those that did not meet the selection criteria were excluded. We extracted the following data from the remaining studies: first author, journal name, year of publication, year enrolled, country where research was conducted, study design, study population, operation type, body mass index, and intraoperative and postoperative urologic complications. Intraoperative urologic complications included bladder injury, ureteral injury while postoperative urologic complications included ureterovaginal fistula, vesicovaginal fistula, ureteral stenosis, and urinary tract infection. Postoperative urinary retention and urinary incontinence were excluded.

Main and subgroup analyses

We investigated the association between laparoscopic radical hysterectomy and risk of urologic complication as a main analysis. We also performed subgroup analyses for published year, quality of study design, body mass index (BMI, kg/m2), country, and operation method [laparoscopic assisted radical vaginal hysterectomy (LARVH) vs. total laparoscopic radical hysterectomy (TLRH)]. Published years of articles were divided into two groups: before 2012 and after 2012. Matched, prospective cohort studies were categorized as high quality while retrospective studies were categorized as low quality in the current meta-analysis. We considered BMI > 24 as obese group and BMI < 24 as standard group. Countries were categorized as Asia, America, and Europe. LARVH involved opening the paravesical and pararectal space, resecting a vaginal cuff, ligating the cardinal ligament and the uterosacral ligament, and unroofing of the ureter.

Statistical analyses

Data for dichotomous variables were analyzed using odds ratio (OR). To compute pooled odds ratio with 95% CI, we used adjusted OR and 95% CI reported in each article whenever possible. We also carried out subgroup meta-regression analysis to assess the effect of subgroups and covariates such as published year, quality of study methodology, BMI, country, and operation method. We evaluated the possible heterogeneity in results across studies using Higgins I [12] to measure the percentage of total variation across studies. We considered an I2 value > 30% as indicative of substantial heterogeneity. We estimated a pooled OR with 95% CI using both fixed-effects [13] and random-effects models [14]. Fixed-effects model was utilized in the absence of significant heterogeneity while a random-effect model was used in the presence of significant heterogeneity. We evaluated publication bias of studies included in the final analysis using Begg’s funnel plot and Egger’s test. Meat-analyses were conducted using Stata MP version 15.0 software package (StatCorp, College Station, TX, USA).

Results

Literature search

We identified a total of 38 relevant studies published between 2001 and 2018 in the final analysis. Figure 1 shows a flow diagram of the process used to identify relevant studies. A total of 2705 articles were searched. We identified 322 articles from the three databases and bibliographies of relevant articles. After exclusion of 132 duplicate articles, the remaining 190 articles were screened by review according to their titles, abstracts, and publication types. Of these articles, 111 articles that did not meet the selection criteria were excluded. After reviewing the full text for the remaining 79 articles, we included 38 articles in the final analysis. The main reason for excluding 41 studies from the final review was due to no available data for urologic complication (N = 27). Other reasons included less than 10 patients (N = 4), comparison between robotic radical hysterectomy (RRH) vs. LRH or comparison between RRH and ARH (N = 5). Five articles were published in the same institution (N = 5). The largest study was selected for meta-analysis.

Fig. 1
figure 1

Flow diagram of the identification of relevant studies for meta-analysis of this study

Full size image

Intraoperative and postoperative complications were not described in four [6, 7, 11, 15] and three articles [4, 16, 17], respectively. Intraoperative and postoperative urologic complications were analyzed in 34 studies and 35 studies, respectively. We excluded one article in postoperative urinary complication due to publication bias and heterogeneity (I-squared 35.9%, p = 0.02) during meta-analysis.

Characteristics of the 38 studies included in the final analysis

Table 1 shows main characteristics of the 38 studies [4,5,6,7, 10, 11, 15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46] included in the final analysis. Their study design types were retrospective study (n = 25), matched case–control study (n = 10), prospective cohort study (n = 2), and RCT (n = 1). Locations of these studies were America (n = 9), Europe (n = 12), and Asia (n = 13). The enrollment period (year) of participants across studies ranged from 1991 to 2016.

Table 1 Characteristics of 38 studies included in the current meta-analysis
Full size table

We identified 2720 patients in LRH and 4084 patients in ARH (2582 patients in LRH vs. 3989 patients in ARH in intraoperative analysis, 2398 patients in LRH vs. 2367 patients in ARH in postoperative analysis). LARVH and TLRH were performed in 7 and 26 studies, respectively. Median or mean BMI was described in 24 studies. If BMD was greater than 24, the group was classified as obese (n = 10). If BMD was less than 24, then the group was classified as standard (n = 10). BMI was reported in 14 studies.

LRH and risk of urologic complications

Figures 2 and 3 show the effect of LRH on the risk of intraoperative and postoperative urologic complications in meta-analysis of 34 studies, respectively. Intraoperative complications were detected in 98 of 2,582 patients (bladder injury, n = 65; ureter injury, n = 33) who underwent LRH. The incidence of bladder injury was significantly higher than that of ureter injury (p = 0.001). In a fixed-effects meta-analysis of 34 studies including intraoperative urologic complications, the overall risk was increased in the laparoscopic group (OR 1.40; 95% CI 1.05–1.87, p = 0.022). In a fixed-effects meta-analysis of 34 studies including postoperative urologic complications, overall risk was increased (OR: 1.35; 95% CI 1.01–1.80, p = 0.039). There was no heterogeneity among studies of intraoperative urologic complications (I2 = 0.0%, p = 0.850) or among studies of postoperative complications (I2 = 0.0%, p = 0.763). No publication bias was observed in these selected studies. Results of assessment of publication bias showed symmetric distribution. All scattered hollow round points were symmetrical on both sides, indicating that the bias of data applied in this study was small (Fig. 4). Cumulative meta-analysis for intraoperative urologic complication of laparoscopic radical hysterectomy was performed. As shown in the cumulative meta-analysis plot (Fig. 5), the summary estimate for intraoperative complications began to be significantly higher in the laparoscopy group than that in the laparotomy group after inclusion of the study (2001) reported by Malur et al. [18]. The OR remained significantly higher in the laparoscopy group up to the study reported by Van De Lande et al. [33] in 2012. As time went on, the OR tended to decrease. The OR remained statistically insignificant after the study reported by Wright et al. [17] in 2012. The meta-regression showed a strong evidence of an association between published year and intraoperative urologic complication (p = 0.036, coefficient: − 0.084).

Fig. 2
figure 2

Intraoperative urologic complications of laparoscopic radical hysterectomy compared to abdominal radical hysterectomy in a fixed effect model of case–control studies. Forest plot showed intraoperative complication in 34 studies. No significant between-study heterogeneity was detected (I2 = 0%, p = 0.850) (OR 1.40; 95% CI 1.05–1.87, p = 0.022)

Full size image
Fig. 3
figure 3

Postoperative urologic complications of laparoscopic radical hysterectomy compared to abdominal radical hysterectomy in a fixed effect model of case–control studies. Forest plot showed intraoperative complication in 34 studies. No significant between-study heterogeneity was detected (I2 = 0%, p = 0.763) (OR 1.35; 95% CI 1.01–1.80, p = 0.039)

Full size image
Fig. 4
figure 4

Funnel plots of intraoperative (A) and postoperative (B) urologic complications of laparoscopic radical hysterectomy compared to abdominal radical hysterectomy

Full size image
Fig. 5
figure 5

Cumulative meta-analysis intraoperative urologic complications of laparoscopic radical hysterectomy compared to open radical hysterectomy

Full size image

Subgroup meta-analysis

Table 2 shows the effect of laparoscopy on intraoperative urologic complication in subgroup meta-analyses by published year, quality of study, BMI, country, and laparoscopic type. Significant harmful effects of intraoperative urologic complications in LRH were observed among articles published before 2012 (OR 2.0, 95% CI 1.24–3.23), but not in articles published after 2012 (OR 1.12, 95% CI 0.77–1.62) (Fig. 6). There was no significant association between postoperative urologic complications and published year (articles before 2012: OR 1.35; 95% CI 0.829–1.352; articles after 2012: OR 1.352; 95% CI 0.948–1.928).

Table 2 Association between laparoscopic radical hysterectomy and intraoperative urologic complication in subgroup analysis by published year, quality of study methodlogy, body mass index, country and laparoscopic type (N = 34)
Full size table
Fig. 6
figure 6

Intraoperative urologic complications of laparoscopic radical hysterectomy compared to abdominal radical hysterectomy in a fixed effect model of case–control. Subgroup meta-analysis results for articles published before 2012 (A) and articles published after 2012 (B) are shown

Full size image

There was no evidence for an association between the quality of study design and intraoperative urologic complications. Intraoperative urologic complications in LRH showed significant association with obese patients (OR 2.12, 95% CI 1.14–3.96), but not with patients having standard BMI (OR 1.08, 95% CI 0.61–1.92). In a fixed-effects meta-analysis of 26 studies including TLRH, the overall risk of intraoperative urologic complications was not increased (OR 1.28, 95% CI 0.90–1.83). However, the overall risk of intraoperative urologic complications was increased in 7 studies including LARVH (OR 1.67, 95% CI 1.01–2.77). Subgroup meta-analysis of 13 Asian studies (OR 1.12, 95% CI 0.72–1.74) and 9 American studies (OR 1.39, 95% CI 0.83–2.33) showed that the overall risk of intraoperative urologic complications was not increased. However, the overall risk of intraoperative urologic complications was increased in a subgroup meta-analysis of 12 European studies (OR 2.05, 95% CI 1.13–3.69). No significant association was observed for study quality, BMI, country, or laparoscopic type in a meta-regression analysis.

Discussion

LRH is thought to be associated with better recovery, smaller scar, and faster back to normal life than ARH [3]. Many comparative studies have reported that survival outcome and perioperative complications after LRH are comparable to those after ARH in patients with uterine cervical cancer. Most of these studies comparing LRH with ARH were retrospective studies or small sized RCT with limited data. Recently published large-scale RCT (LACC trial) [47] showed that minimally invasive radical hysterectomy including LRH was associated with lower rates of overall survival and disease-free survival compared to ARH. Unfortunately, perioperative complications were not described in LACC trial. The incidence of urologic complications during LRH is thought to be higher than that of any other gynecologic surgical procedure due to wide dissection of periureteral tissue and bladder, unroofing the ureter in the cardinal ligament, and the distorted pelvic anatomy caused by mass such as cervical cancer. Most of previous comparative studies revealed no statistical significance in urologic complications between LRH and ARH.

In 2012, we performed a meta-analysis of perioperative urologic complications related to LRH and reported that the risk of intraoperative urologic complications was higher in LRH group than that in ARH group based on data released through 2011 [48]. Although the incidence of intraoperative urologic complications was significantly higher in LRH, studies analyzed were performed at different time. The risk of intraoperative urologic complications in LRH had a tendency to decrease as time went by in cumulative meta-analysis. Laparoscopy continue to evolve. It could reduce complications due to technical changes and the development of instruments. The learning curve over time could also affect complication rate. The accumulation of surgical experience, the improvement of surgical skill, and the development of laparoscopic instruments such as advanced bipolar devices, Endo-GIA staplers, and Endo-Clips could reduce complications, thus affecting surgical outcomes as time went by. Thus, we performed meta-analysis again by including data released between 2012 and 2017. Our results showed that there was still an association between LRH and risk for intraoperative urologic complications. However, subgroup analysis showed that LRH among published articles after 2012 (OR 1.12, 95% CI 0.77–1.62) was not associated with a significant increased risk of intraoperative complications. There was a strong evidence for an association between published year and intraoperative urologic complications in meta-regression. Postoperative urologic complications were also higher in the LRH group, inconsistent with our previous study [48]. In subgroup analysis by published year, articles published after 2012 or before 2012 were not associated with high incidence of postoperative urologic complication in LRH. Difference became evident as the number of studies and patients included in the meta-analysis became larger.

There were two large comparative studies [49, 50] of perioperative complications about LRH versus ARH. Uppal et al. [49] reported perioperative complications of LRH and robotic radical hysterectomy (RRH) compared with ARH for 7180 cases of uterine cervical cancer from 2012 to 2015 using the National Inpatient Sample (NIS) from the Healthcare Cost and Utilization Project in USA. Perioperative complications were significantly lower for LRH and RRH. However, intraoperative urologic complication including ureter and bladder injury was higher in LRH and RRH than that in ARH (p = 0.027). Ratios of LRH and RRH in all study groups were 17.2% and 82.2%, respectively. They did not distinguish LRH from RRH. Thus, we excluded that study from the meta-analysis. Kim et al. [50] have reported that intraoperative complications of LRH have lower (OR 0.73; 95% CI 0.63–0.86) than ARH for 6,335 cases of uterine cervical cancer from 2011 to 2014 using the Korea Health Insurance Review and Assessment Service database. Intraoperative complications included vascular and intestinal injury in addition to ureter and bladder injury. Unfortunately, accurate assessment of urologic complications was difficult because the article [50] did not describe urinary complications that could be distinguished from other intraoperative complications.

Risks of bladder injuries are statistically higher than those of ureter injuries. In radical hysterectomy, ureter injuries usually occur during dissection of periureteral tissue to confirm ureter passage and find the uterine artery. The process of unroofing the ureter in the cardinal ligament, an important step in radical hysterectomy, plays a crucial role in inducing ureteral injury.

In the subgroup meta-analysis by country, results of European studies showed that the risk of urinary complications during surgery was significantly higher, although there was no significant difference between results of Asian and American studies. It is not easy to predict basic laparoscopic surgery technique. The habit of using chopsticks for Asian people and different surgical education systems in different countries might affect the outcome [51]. With regard to surgical approach, the incidence of intraoperative urologic complication rate was increased during LARVH. Vaginal route allows a precise dissection of vaginal cuff. In LARVH, after the uterine ligament was cut, the bladder and ureter are seen by traction of the uterus through vaginal route. This provides a theoretical background in which complications may be different due to differences in surgical methods. Obesity is a predisposing factor of perioperative urinary track complications in patients undergoing cancer surgery through laparoscopy [52, 53]. The current meta-analysis showed that intraoperative urologic complications were higher in patients who were relative overweight with LARVH, consistent with our previous study [48]. Our meta-analysis has several limitations. First, most studies involved were retrospective. They failed to provide the highest level of evidence. As a result, selective deflection and missing data could degrade the quality of study. A randomized clinical trial comparing LRH with ARH in patients with early stage uterine cervical cancer is needed. Results of such study will help determine the risk of perioperative urologic complications in LRH. Recently, RRH has been replacing LRH. One meta-analysis showed that RRH was superior to LRH in perioperative complication [54]. However, it focused on all intraoperative and postoperative complications related to surgery, not urologic complications. In the future, a study is needed to assess the risk of perioperative urologic complications of RRH and LRH. Second, we included only English language articles which might be a bias of our study. Exclusion of studies written in language other than English might have affected our outcomes. However, there have been few studies on this topic written in non-English languages. Third, we did not classify the incidence of urologic complications according to the type of radical hysterectomy. Although most patients underwent type III radical hysterectomy, some patients underwent type II radical hysterectomy. The incidence of urologic complications might be higher in advanced uterine cervical cancer and type III radical hysterectomy. Bladder dysfunction and urinary incontinence were excluded from the analysis of postoperative complication because many articles did not describe them as it was uncommon to evaluate these complications objectively in postoperative care. This might have led to biased results in analysis of postoperative urologic complications. Stage of the disease and diabetes that could act as a predisposing factor of the urinary track injury during radical hysterectomy [55] was not evaluated either.

Despite these limitations, our meta-analysis demonstrated that there was an association between LRH and risk for intraoperative and postoperative urologic complications. We also confirmed that intraoperative urologic complication decreased after 2012 in subgroup analysis and meta-regression. Of the 38 studies included in this meta-analysis, two studies [21, 38] reported that intraoperative urologic complications were significantly higher in the LRH group. Thus, publication bias is minimal. It was confirmed by funnel plot. Although this meta-analysis was based on results of retrospective studies, it included more than 4,000 patients with a comprehensive review of the available literature. Our findings should be verified by large-scale prospective cohort studies or randomized controlled trials in the future.