Introduction

Anastomotic leak rate in sigmoidectomy varies between 5% and 15% [1,2,3]. Ischemia of the anastomotic region counts to the most meaningful factors, leading to anastomotic leakage. Therefore, the preservation of the anatomical blood supply of the rectum was investigated concerning its impact on anastomotic healing by several authors, commonly focusing on the inferior mesenteric artery (IMA, [4, 5]). Valdoni et al. first described the technique of sigmoidectomy with preservation of the IMA in 1972 [6]. The objective of the present analysis was to figure out, whether preservation of the superior rectal artery (SRA) contributes to a decrease in anastomotic leakage rate in sigmoidectomy for diverticular disease.

Materials and methods

Data collection was conducted according to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) statement [7].

Inclusion criteria

Randomized and nonrandomized studies, reporting on the influence of preservation or dissection of the superior rectal artery on anastomotic leak rate in sigmoidectomy for diverticular disease were included.

Exclusion criteria

Studies, giving no information on defined endpoints as well as studies dealing with alternate pathologies such as inflammatory bowel disease or colorectal cancer were excluded.

Outcomes

Primary outcome: anastomotic leak rate; secondary outcome: overall postoperative morbidity.

Search strategy

A systematic database search for all studies, comparing preservation or dissection of the superior rectal artery in sigmoidectomy for diverticular disease was conducted in PubMed and in the Cochrane library. A combination of the following search terms was used: sigmoidectomy, diverticulitis, diverticular disease, superior rectal artery, vascular preservation, pedicle preservation, artery preservation. The “similar articles” function of PubMed was used, in order to extend potential hits. Moreover, references of relevant and eligible studies were screened for additional reports according to the objectives of the review.

Study selection

Two authors (SM and AA) independently assessed titles and/or abstracts of all identified reports and excluded those considered irrelevant. Afterwards, full-text papers were screened on their accordance to inclusion criteria. One author (SM) extracted data from the included studies and a second (AA) checked extracted data. Disagreements were discussed.

Data items

The following parameters were extracted from included studies: number of patients/intervention, year, study design, inclusion/exclusion criteria, indication for the operation, surgical technique, definition of anastomotic leakage.

Assessment of study quality and risk of bias

The Newcastle–Ottawa Scale (NOS) was used to estimate the risk of bias of nonrandomized studies ([8, 9]; Table 1). NOS is a simple instrument, recommended within the Cochrane handbook for systematic reviews [10]. Construct validity and interrater reliability of NOS are well established. Quality assessment was based on three domains: patient selection, comparability of case and control group and assessment of outcomes. A maximum of one star could be allocated for each item in the categories “Selection” and “Exposure/Outcome”. “Comparability” could be labeled with two stars. Maximum score is nine stars.

Table 1 Newcastle–Ottawa Scale for assessing methodological quality of nonrandomized studies
Full size table

Results

Using the described search strategy, 21 studies were selected from an overall number of 381 hits (Fig. 1). In all, 14 studies were excluded as duplicates after a review of title and abstract. The remaining eight studies were included into the detailed analysis. Therein, five more studies were excluded after full-text screening (Table 2). Thus, three studies were included into the qualitative analysis (Table 3). Two of them are retrospectively conducted and one has a prospective, but nonrandomized design. Overall, no randomized study could be identified. A total of 564 patients were investigated. Therein, SRA was preserved in 305 cases and sacrificed in 259.

Fig. 1
figure 1

Flow diagram for the different phases of the systematic review

Full size image
Table 2 Studies excluded from the qualitative analysis
Full size table
Table 3 Current studies
Full size table

Primary outcome

Exclusively, Borchert et al. demonstrate a significant reduction of anastomotic leak rate after a preservation of the SRA ([11]; Table 3). The same, but without significant difference, was stated by Sohn et al. in 2017 [12]. This study is a multicentric analysis, reporting on the highest number of patients with SRA preservation yet. Data were collected in three surgical centers. Overall, 267 patients were included; a preservation of the SRA was performed in 157 cases. The analysis demonstrated a tendential reduction of anastomotic leak rate (SRA preservation: 1.9%, SRA dissection: 7%, p = 0.053). In contrast to these results, Lehmann et al. found a nonsignificant increase of anastomotic leaks after a preservation of the SRA. Data of 92 patients were retrospectively analyzed; in 48 cases, SRA was preserved (SRA preservation: 6.5%, SRA dissection: 3.7%, p = 0.610, [5]).

Secondary outcome

In the series of Lehmann et al., solely two cases of a temporarily retrograde ejaculation occurred. Both cases are associated with an arterial dissection [5]. Borchert et al. found a significant increase of wound dehiscence in case of a SRA dissection. Apart from that, no other differences arose, comparing postoperative complications [11]. Postoperative complications are equal in both groups within the analysis of Sohn et al.

Inclusion and exclusion criteria of all analyzed studies are depicted in Table 4, technical aspects in Table 5, and definitions of anastomotic leakage in Table 6. Quality assessment of included studies according to the Newcastle–Ottawa scale was homogenous. All studies were rated with a maximum of six of possible nine stars. Deductions were found due to identical reasons in all studies. The parameter “Selection of controls” (Selection, No. 3) requires a control group out of the general population. Since controls are all hospitalized and had operations, requirements were not fulfilled. Within the category “comparability”, all studies were rated with one of two possible stars because a matched pair analysis is demanded for maximum appraisal. No included investigation could meet this requirement. A lacking statistic difference concerning the baseline characteristics does not qualify sufficiently. The parameter “non-response rate” in the category “assessment of outcome” (Assessment of outcome, No. 3) was achieved in no study.

Table 4 Inclusion criteria
Full size table
Table 5 Operative characteristics
Full size table
Table 6 Definition anastomotic leak
Full size table

Discussion

To investigate the influence of preservation of the SRA on anastomotic leak rate and postoperative morbidity after sigmoidectomy for diverticular disease, a systematic review of the current specific literature was conducted. Overall, availability of data is weak. Three studies were included into the qualitative analysis. Extracted results are inhomogeneous (Table 3). One investigation showed a significant reduction of anastomotic leak rate in case of a vascular preservation, a second study demonstrated a tendential reduction of leaks, and a third was in line with a nonsignificant increase of anastomotic leakages. Altogether, a preservation of the SRA seems to have a positive effect on anastomotic healing. No investigation showed a significant negative influence. In 2012, Cirocchi et al. published the first systematic review and meta-analysis on the preservation of the inferior mesenteric artery, according to requirements of the PRISMA statement. After the selection process, four studies were included into the qualitative analysis, therein only one randomized [4]. Meta-analysis failed to demonstrate a significant influence of preservation of the SRA on anastomotic leak rates. However, within an individual consideration of the different studies, leak rate of the only randomized trial could be shown to be significantly lower after SRA preservation (p = 0.03, [16]). The authors discussed different definitions of anastomotic leaks and the low number of included studies as potential reasons for inhomogeneous results. In 2013, Trencheva et al. published results of a prospectively conducted, nonrandomized observational study [15]. The objective of this investigation was to identify predictors for the development of anastomotic leakages after colon and rectum resections. A total of 616 patients were included. No selection of the underlying disease was performed. A subgroup analysis showed that there was no significant influence between preservation (n = 57) and dissection (n = 246, p = 0.219) of the SRA on anastomotic leak rate. Bergamaschi et al. reported a 0% leak rate in a consecutive series of 30 patients after resection rectopexy with SRA preservation [13]. However, the study was retrospectively conducted and neither controlled nor randomized and thereby of low quality.

Apart from anastomotic leak rate, Sohn et al. found a significant reduction of the procedural length, which needs to be emphasized for patient and economic reasons. Short- and long-term follow-up, investigating postoperative functional aspects is lacking in all studies. Due to the risk of a potential damage of the hypogastric nerves, this aspect is of principal interest. The decreased risk of a nerve damage, resulting from tubular dissection of the mesocolon in case of arterial preservation could relevantly influence long-term function and thereby associated quality of life. Concluding, to date, evidence on the influence of SRA preservation or dissection on anastomotic leak rate in sigmoidectomy for diverticular disease is insufficient. Studies from Borchert and Sohn showed a positive influence. Since preservation of the superior rectal artery improves aboral blood supply of the anastomosis, and appropriate perfusion is crucial for anastomotic healing, SRA preservation generally should be intended. Additional reasons are the decreased risk of nerve damage with consecutively improved functional outcomes and shorter length of surgery. Due to a lack of randomization, a risk of bias is inherent. Moreover, definitions of anastomotic leakage were inconsistent. For these reasons, the level of evidence of included studies is limited. For appropriate assessment of the influence of a SRA-preserving technique, randomized studies must be performed.

Conclusion

Preservation of the SRA seems to have a positive effect on anastomotic leak rate. Therefore, vascular preservation is generally recommended in sigmoidectomy for diverticular disease. Due to the low number of available studies and the absolute lack of randomized trials, to date the level of evidence is relatively low.