There were 4 (8%) conversions in group I. In 1 patient (2%) introduction of an additional 5 mm trocar in the right subcostal area was necessary. It was a result of abnormal anatomy of the hepato-duodenal ligament and consequent difficulties to identify structures of Calot's triangle. In 3 patients (6%) conversion to multiport laparoscopic cholecystectomy was necessary due to extensive peritoneal adhesions. In none of the patients in this group was conversion to open surgery necessary.

There was 1 (2%) conversion to open surgery in group II. It was due to duodenal injury during the dissection of pericystic adhesions.

Five patients initially allocated to group I were excluded due to surgical technique modification that no longer fulfilled the inclusion criteria (conversion – 4 patients; laparotomy on the 4^th postoperative day due to bile leak – 1 patient). Similarly, 1 patient was excluded from group II due to conversion to an open cholecystectomy. Table II shows a comparison of outcomes in groups I and II after the exclusions described above.

Postoperative complications were divided into two groups: complications that occurred as a consequence of the surgery and complications associated with wound healing. All complications with the time of occurrence are shown in Table III.

One of the very important factors assessed during the comparative analysis of surgical techniques is total operative time. In the present study the mean operative time of SILS cholecystectomy was 66 min (range: 35-110 min). It was very close to the results of Solomon et al. [8]. When they excluded the first 10 cases (mean operative time of 80 min) and performed a sub-analysis of operative times for consecutive groups consisting of 10-11 patients, they could see that the mean operative time shortened successively, being 73, 71, 58 and 65 min in the consecutive groups. These results were similar to other published series [9–15]. It significantly differed from the mean operative time of 47 min for multiport cholecystectomy. This difference would be much greater if we compared our results with the best achieved operative times for laparoscopic cholecystectomy published by Stephenson et al. In his group of patients the mean operative time was 39 min (range: 25-60 min) [16]. Certainly, the significantly longer operative time of SILS cholecystectomy in comparison to multiport laparoscopy is a weakness of the technique. However, one should remember that this technique is novel. Historically, operative times of the first 100 laparoscopic cholecystectomies published by great laparoscopic surgeons were 98 min (Zucker et al.) and 85 min (Peters et al.) [17, 18]. These operative times are still longer when one compares these results with our and some other authors’ initial outcomes of SILS cholecystectomy. One can expect that this scenario will be repeated soon with significantly improved published outcomes.

The post-operative pain is the next key parameter allowing a comparative analysis of both techniques. Post-operative pain is particularly important from the patient's point of view. It is a subjective measure of suffering that influences the post-operative quality of life. It is unquestionable that the post-operative pain after laparoscopic procedures lasts significantly shorter and its intensity is reduced when compared to open techniques [19–22]. The introduction of SILS and further reduction of operative trauma allows us to expect further improvement of post-operative pain. The main difference between multiport and SILS cholecystectomy is the number of abdominal wall incisions; thus one would expect a reduction of pain mainly in the abdominal wall and not in its cavity. The specific time intervals for pain measurement were at 6 h and 24 h post-operatively, when the patient passed urine for the first time and was fully mobilized. The intensity of post-operative pain 6 h after surgery was significantly smaller after SILS cholecystectomy, with a mean VAS score of 3.49, when compared to multiport surgery, with a mean VAS score of 4.53 (p = 0.00096). Similar results were published by Tsimoyiannis et al. [23]. Importantly, in their study most of the patients also declared that they were ready to leave the hospital 6 h after surgery. Similar results were published by Bresadola et al. with assessment performed at 4 h and 8 h after surgery [24]. Similarly, a significant but smaller difference in VAS scores was seen at 24 h, with VAS = 1.18 and 1.55 (p = 0.0187) for SILS and multiport cholecystectomy respectively. Reduction of the difference in VAS score, seen also by Tsimoyiannis et al. [23], is a result of the fact that the main difference between the two techniques is the cumulative length of surgical incision. The remaining factors such as the mode of creation of pneumoperitoneum, intra-abdominal pressure, temperature and type of gases used during both techniques were identical. The larger the wound, the greater is the extent of operative trauma. Therefore, the difference in the VAS scores in favour of the SILS procedure was most clearly visible during the first 24 h after surgery. Reduced pain severity after SILS cholecystectomy was reflected in a decreased analgesic requirement during the post-operative period. Although the statistical analysis showed that the difference in analgesic requirement between groups was not significant, the p value of 0.0737 was very close to the significance level. Similar results were published by Bresadola et al. [24].

Hospitalization time might be measured in several ways. We decided to measure only the time that the patient stayed in the hospital after surgery, as the pre-operative period varies significantly depending on the activities required to prepare the patient for surgery in Polish hospital settings. The mean post-operative hospitalization time in group I was 1.33 days (range: 1-2 days) and in group II was 1.96 days (range: 1-3 days) (p < 0.001). Therefore, in terms of post-operative hospitalization time SILS cholecystectomy is more beneficial. Similar conclusions come from most of the cited publications, although direct comparisons are difficult to make as we assessed only post-operative hospitalization time, not the total hospitalization time as did other authors [10, 11, 13, 25–27].

Since the introduction of laparoscopy, one of the major factors allowing assessment of its value and usefulness has been the rate of conversions. It is obvious that SILS cholecystectomy, like all previous laparoscopic procedures, had to be assessed against this parameter. In the present study there were 4 conversions (8%) in group I. Our conversion rate seems to be good when compared with previously published reports, especially since the number of patients in our group is not that small. Most of the studies with conversion rates of 0 reported on outcomes in groups of 10 to 20 patients [13, 14, 27–31]. Studies with a similar or greater number of patients than our study reported similar conversion rates [8, 10–12, 25, 32]. Edwards et al. reported the necessity to introduce an additional 5 mm trocar in 6 cases and in 3 cases they had to convert to classic 4-port cholecystectomy in a study group of 76 patients [12].

Only three studies reported on single cases that required conversion to an open procedure [8, 11, 25]. All of these conversions were due to solid adhesions or inflammatory infiltration that was not amenable to laparoscopic manoeuvres. There was only 1 (2%) conversion to an open procedure in group II, which is in favour of SILS cholecystectomy that did not require such conversions. Nevertheless, one should not draw conclusions from this result as the assessment of conversions should be performed on significantly larger groups of patients.

Despite being minimally invasive, laparoscopic techniques are not free of post-operative complications. As SILS cholecystectomy is a very specific procedure, we decided to divide the complications into two groups. The first group was composed of complications that were a direct result of the course of operation and the second group was composed of complications of wound healing. There were 2 cases of right pneumothorax that occurred directly after putting the suspension suture into the gallbladder fundus via the 7^th intercostal space (7^th and 16^th operations). In these cases we decided to reduce the pneumoperitoneum to 9-10 mm Hg and we pulled the gallbladder towards the abdominal wall with more strength. At the same time the anaesthesiologist increased the O₂ concentration in the inhaled gases to 50% and corrected the ventilation to keep EtCO₂ between 30% and 35%. In both cases the undertaken measures were sufficient and allowed the procedure to be finished as planned. Neither of these 2 patients required any thoracic drainage as the control bed-side chest X-ray showed no pneumothorax at the end of the procedure. Only one study has reported such a complication so far [33], although it seems it might be more frequent than reported. When we planned the procedure with the use of a transabdominal suspension suture we based it on the experience of Endo et al., who placed such a suture in the intersection of the 7^th intercostal space and midclavicular line with no pneumothorax in his group of 132 patients [34]. Based on our experience of pneumothorax in these 2 patients, in all further patients we decided to place the suspension suture in the 8^th intercostal space and slightly lateral to the midclavicular line.

Another complication seen in our set of patients was the omission of the opened Luschka duct with subsequent choleperitonitis. In our opinion the bile leak was due to the gallbladder punctures caused by the suspension suture and our relatively small experience in this type of procedure (17 operations). That misled the surgeon about the origin of the bile leak and prevented the search for the leakage. Similar problems were encountered by the teams of Solomon et al. and Edwards et al. [8, 12], but in their cases endoscopic exploration, bile duct stenting and percutaneous drainage of the peritoneal cavity was sufficient. We did not encounter such a complication in group II. Available literature shows that the rate of bile leaks after classic multiport cholecystectomy varies between 1% and 3% [35–41]; thus our 2% leak rate after SILS cholecystectomy is comparable with other centres.

We observed 4 cases (8%) of wound complications in group I. The first patient with a wound complication had marginal necrosis of the umbilical skin (4^th patient). On the 7^th postoperative day the patient had a necrectomy done and the wound was left to heal via granulation. Two other patients had seroma formations in the umbilical wounds. In 1 patient, drainage on the 4^th postoperative day was sufficient to heal the wound. In the other case, the exudate was present for 3 weeks after drainage and formed a granuloma the size of a pea. It was later excised together with one of the fascial sutures under local anaesthesia with good results. One patient had a wound infection that required opening of the wound and delayed healing via granulation.

In group II there was 1 (2%) umbilical wound infection, which corresponds to the rates published by other authors [42]. Subanalysis of wound complications showed that although the complication rate in group I was 8%, which seemed much higher than 2% for group II, it was only true for the first subset of patients (n = 25 in each group). In the remaining half of study patients the frequencies of wound complications were equal (Table III). We think that this phenomenon was caused by the effect of the learning curve associated with the SILS technique, wound closure and recreation of the umbilicus. Most studies did not report on wound complications after SILS cholecystectomy [8, 10–15, 26, 27, 30, 43]. Just a few authors had no complications [29, 44, 45] and only Tacchino et al. reported a single case of periumbilical haematoma that was evacuated on the 7^th postoperative day, which allowed for wound healing [28].