Comparative analysis of linear- and circular-stapled gastrojejunostomies in Roux-en-Y gastric bypass: a focus on postoperative morbidity using the comprehensive complication index

The laparoscopic Roux-en-Y gastric bypass (RYGB) is one of the most frequently performed metabolic-bariatric surgeries (MBS) worldwide and leads to both sufficient body weight loss and resolution of obesity-associated medical problems [1‐3]. After RYGB became laparoscopically feasible in the late 1990s [4], three techniques for the gastrojejunostomy (GJ) were regularly applied: entirely hand-sewn, hand-sewn combined with linear-stapled (LSA) or entirely circular-stapled anastomoses (CSA). A period of technical refinements followed resulting in significantly shorter operation times for the stapled anastomoses turning them into the gold standard GJ-technique in RYGBs [5‐9]. Comparing the two stapled techniques revealed not only a shorter operation time, but also less anastomotic strictures at the GJ and less wound infections for the LSA [7, 9‐11]. Consequently, a shift from CSA to LSA has been observed over the last few years [12].

However, the reporting of postoperative complications still varies among clinical studies making a comparison of results difficult [13‐16]. The already well-established Clavien-Dindo classification (CDC) offers a simple tabular form to rate postoperative complications according to the invasiveness of treatment needed to correct a complication [17]. Usually, only the most severe complication is reported with the CDC. Adverse events of lesser severity are however “neglected” and may involve an underestimation of the true overall morbidity of a surgical procedure. To overcome this limitation, the idea of the Comprehensive Complication index (CCI) was developed, which incorporates all postoperative complications graded with CDC in a score from 0 (no complication) to 100 (death). It therefore gives a representative picture of the cumulative postoperative morbidity correlating with overall costs [18] and the patient’s individual burden of several adverse events. We herein present our single-center experience with CSA- and LSA-technique and quantify postoperative morbidity with the CCI, which has not yet been published before.

Overall, 588 patients after primary, laparoscopic RYGB were analyzed and divided in two cohorts according to the anastomotic technique for the gastrojejunostomy: 298 patients with CSA and 290 patients with LSA. All included patients completed a follow-up of at least 2 years. Patients in the LSA-group were significantly older (41.4 vs 37.1 years, p < 0.001), more frequently female (75.9% vs 48.7%, p < 0.001) and underwent more previous abdominal surgery (45.5% vs 27.2%, p < 0.001). The mean preoperative BMI was comparable with 44.8 ± 6.3 kg/m² in the CSA- and 44 ± 7 kg/m² in the LSA-group (p = 0.061) (Table 1).

A shift from circular- to linear-stapled gastrojejunostomies in RYGB has been observed over the last few years. This is the first study, which defines postoperative outcome of the two surgical techniques with the CCI. We found that the cumulative CCI for the first 90 days after RYGB was significantly higher in the CSA- than in the LSA-group, despite more, but not significantly, high-risk patients in the latter cohort. In particular, the incidence of SSI and GJ-stenosis in the early postoperative phase as well as of internal hernias during the first 2 years after MBS were significantly higher in the CSA-group. Remarkably, changes in wound management and closure of all mesenteric spaces equalized the incidence of SSI and internal hernias in the two cohorts. Finally, patients with CSA showed a greater inflammatory response after surgery with higher postoperative CRP-levels and higher postoperative need for opioid analgesics compared to the LSA-group.

The RYGB is a well-established MBS-procedure, and its surgical steps are predominantly standardized with a view to reduce morbidity in patients often with high-risk profiles. Still, routine quality assessment is key to guarantee best achievable outcome for surgeons and patients [28, 29]. The CCI offers a numeric value representing cumulatively all negative events during 90 days after surgery and is routinely applied in our clinic to compare surgical outcome with for example benchmark outcome [25, 30]. The aim of assessing postoperative results is to stimulate surgical improvement. The two surgical techniques for the GJ, which were analyzed in this study, are technically feasible and reach similar weight loss. However, previous analysis demonstrated an advantage for LSA in terms of a lower complication rate. Peterli et al. showed a significant lower GJ-stricture rate for LSA, while Muller et al. demonstrated a better quality of life and less complications overall, less re-hospitalizations and reoperations for the LSA cohort after a 5-year follow-up [11, 31]. Here, we present the data of our high-volume center reflecting the postoperative outcome with the CCI. Overall, the incidence of early complications within the first 90 days after both surgical techniques may seem comparably [32, 33] high in this study with 38.3% in the CSA and 19.3% in the LSA group. This is especially true taking into consideration our own analysis comparing the outcome of benchmark patients demonstrating that 90% of low-risk patients should have an uneventful course during the first 90 days after RYGB [25]. However, there are two factors that need to be considered when interpreting these results. First, the included patient-cohort had a rather high-risk profile at baseline, indicated by more than 80% of the included patients fulfilling the criteria of non-benchmark cases. The incidence of complications could be rather due to patient selection than poor surgical technique. Of note, the mean 90-day CCI of all patients with complications in our analysis is below the published 90-days cut-off value of 33.73—despite the high-risk profile of the included patients [25, 30].

Second, due to the fact that the CDC score was developed by members of our clinic [17, 26], our clinical staff is sensitized to document postoperative complications as accurate as possible. CDC grades of all patients are routinely defined and documented at the moment of discharge by our residents and regularly discussed at our weekly “morbidity and mortality conference” with all senior consultants being present [27]. This may lead to a comparably high overall incidence of postoperative complications in our institution. Nevertheless, CDC grade I complications are even in our clinic underreported and justify a mean CCI calculation with and without CDC grade I complications (Table 2). Especially from the patient’s perspective, CDC grad I complications are usually less impairing than more severe complications and therefore may “dilute” the true CCI. However, the difference of the mean CCI between the two cohorts was surprisingly little. This can be explained by the fact that—for example—SSIs (CDC grade I) have no or almost no impact on CCI calculation and that repetitive interventions (such as gastroscopies, for example) were counted only once if they were planned [27].

In our study, the CSA-group presented significantly more early complications compared to the LSA-group, especially CDC grades higher than II. This fact may be explained by the higher incidence of stenosis at the CSA requiring more diagnostic and therapeutic gastroscopies [12, 34]. By comparison, there was only one stenosis diagnosed during the first 90 days postoperatively in patients of the LSA-group compared to 41 (13.8%) in the CSA-group. GJ-stenosis is a well-known soft spot for circular-stapled GJ [35, 36]. However, in a recent meta-analysis including 22 articles, LSA was not significantly superior to CSA in terms of GJ strictures [37]. This lack of significant difference has already been documented in a previous meta-analysis by Edholm [7]. In. contrast, a nationwide population-based cohort study from the Netherlands including 12,400 patients showed a significantly higher postoperative overall complication rate in the CSA-group, but especially for major bleeding [38]. Several studies including a systematic review confirm the finding of a significant higher postoperative complication rate for CSA—even up to a complication rate of 50% [7, 11, 30, 38]. The wide range and contradictory findings represent the challenge of comparing complication rates due to different reporting mentality and/or study designs.

As previously published by various authors, our data also showed a comparably high incidence of wound infections in the CSA-group [5, 7, 10, 35, 39]. However, this is not entirely unexpected due to the fact, once the CSA is created, the circular stapler device is removed from the inside of the alimentary limb and abdominal cavity without using a specific wound protection. This finding is not related to the CSA technique per se, but connected to the fact that circular staplers must be inserted intraabdominally without skin protection and potentially contaminate the subcutis. Interestingly, after changing our strategy from primary to secondary closure, the wound infection rate in the CSA-group decreased significantly [23]. Thus, adequate wound management seems to equalize the incidence of SSI between the two techniques.

There was also a significant difference of the occurrence of late complications between the two groups, especially for the incidence of internal hernia, which was significantly higher in the CSA group. This observation can be explained by the different strategies in closing the mesenteric defects in both groups. In the CSA-group, only the mesenteric defect at the level of the jejunojejunostomy was closed, while the Petersen’s space remained open. Unsurprisingly, the majority of internal hernias in the CSA-group were herniations through the Petersen’s space, only a few through the jejunojejunostomy’s mesenteric space and/or through both spaces. Accordingly, Stenberg et al. demonstrated 2016 the importance of closing both mesenteric defects to avoid internal herniations [13]. After the introduction of the LSA technique in our clinic—almost simultaneously to Stenberg et al.’s publication—both mesenteric defects were routinely closed during RYGB leading to a reduced incidence of internal hernias thereafter. Our data are in line with a recent meta-analysis and underline the importance of routine closure of all mesenteric defects during RYGB [40, 41]. Another aspect, which may explain why recurrent Petersen-hernias were only seen in the CSA-group, is the location of the biliary limb. For the CSA, the BL was placed on the right side of the abdominal cavity making it undercrossing the antecolic AL, which leads to a “physologically” incomplete closure of the Petersen space. The other way around in the LSA group: the BL is placed on the left side making it easy to close the Petersen space completely (Fig. 3). However, the discussion is still controversial. A recent long-term analysis showed significantly more internal hernia in the circular-stapled gastrojejunostomy group compared to linear-stapled GJ [12]. Whereas the propensity scored matched analysis from 2016 could not show a significant difference [11].

Despite these two technical learning-by-doing improvements (wound management and closure of both mesenteric spaces), early complication rates were not affected very much in our study. Most of SSIs are CDC complication grade I, which will not be counted for the relevant CCI calculation. The inner hernia instead is a late complication and was not part of the CCI calculation.

Further, the CSA-group presented higher CRP-levels postoperatively and needed higher doses of morphine. Both facts suggest a higher surgical stress for the CSA-group when compared to the LSA-group as postoperative CRP levels may be indicative for the amount of tissue damage during an operation [42]. On the other hand, a higher CRP-level postoperative may also go along with the higher incidence of postoperative complications [43, 44]. In addition, more postoperative pain in the CSA-group can be explained by the larger soft tissue trauma due to the insertion of the circular stapler which adds to the higher tissue damage for the CSA technique [22].

Finally, a higher incidence of postoperative complications due to higher surgical stress and/or longer operation time after RYGB with circular-stapled GJs is not only relevant for each individual patient, but also for health systems worldwide when it comes to related costs. This aspect was analyzed in a recent prospective longitudinal cohort study including 385 patients, where the CSA technique was demonstrated to be associated with higher material costs [8]. Additional data from Switzerland show similar results with improved financial outcomes after transition from a circular to a linear stapling protocol in RYGB surgery [12]. The longer operation time in RYGB with CSA, which seems to be attributed to the technique of the procedure, also leads to higher costs [45, 46].

Our study has several limitations. First, there is the retrospective study design including data over a time period of nearly ten years. Second, the CSA- and LSA-techniques were used in a chronological order suggesting that the evolution of perioperative care during the use of the CSA-technique may have contributed to the better outcomes of the LSA-group. Third, there is some heterogeneity regarding the involved surgeons. Further biases may have been generated by changing protocols for wound management and by different strategies for closing the mesenteric defects.

Our data show a higher mean CCI score for the CSA-group 90 days after surgery compared to the LSA-group indicating that the CSA technique is associated with a higher incidence of early postoperative complications when compared to the LSA technique for the GJ in RYGB surgery. Further, the CSA technique seems to provoke more surgical stress as indicated by higher CRP levels and use of opioid analgesia during the first postoperative days.