Introduction

Until the end of 1990s, laparoscopic resection for the treatment of colorectal cancer had gained limited acceptance because of concerns related to the oncological efficacy and safety of the approach. Since then, the short-term results of several randomised trials assessing the laparoscopic versus the open approach for the treatment of colon cancer have now been reported [1–6]. These trials show that the laparoscopic approach is associated with the same immediate morbidity and mortality, less pain and physical disability, less use of analgaesics, faster recovery and acquirement of a resection specimen of similar quality as compared to the open surgery. Furthermore, some of these trials [1–3, 6, 7] report similar 3- or 5-year local recurrence and survival rates between the two approaches, rendering laparoscopic colectomy an attractive alternative to conventional surgery.

Reports on short- and long-term results after laparoscopic resection of rectal cancer with or without anal sphincter preservation, deriving from trials comparing laparoscopic to open approach, are limited at present. This is because application of the standard total mesorectal excision (TME) as proposed by Heald et al. [8] poses severe technical problems when performed by laparoscopy. For example, the CLASICC trial [4] reports a higher rate of positive circumferential resection margins (CRM) and a conversion rate of more than 30% when attempting a TME by laparoscopy, although the same group [7] reports similar rates of local recurrence between laparoscopic and conventional approaches. According to a recent meta-analysis [8], which involved only three prospective randomised trials [4, 9, 10], laparoscopic surgery for rectal cancer results in faster recovery, less pain and use of analgaesia and a resected specimen oncologically comparable to conventional surgery [11]. At present, two prospective randomised studies [7, 12] report similar recurrence and survival rates between the two approaches.

There is still a great deal of issues, mostly technical, which have to be resolved in order for laparoscopy to become an attractive alternative to conventional surgery also for rectal cancer. Dissection deep into the pelvis to accomplish a TME and acquire a specimen with intact distal and circumferential margins and fashioning a safe anastomosis are technically demanding, particularly in obese and male subjects with bulky tumours, resulting in high conversion and anastomotic leak rates. These technical problems have resulted to thorough selection of patients with rectal cancer as candidates for laparoscopic surgery in most of the comparative studies, thus, violating the principle of the intention-to-treat. The aim of the present study was to assess the short-term outcomes after laparoscopic surgery for rectal cancer in an unselected group of consecutive patients and compare the results to those obtained form a series of matched patients operated upon conventionally by the same group of surgeons.

Patients and methods

After acquiring adequate experience and achieving standardisation in the laparoscopic surgery for rectal cancer, all patients presented with a rectal cancer (lower border of tumour <10 cm form anal verge, at rigid rectoscopcy) from March 2004 to March 2007 were offered the laparoscopic approach for surgical treatment (LAP group) with curative intent. These patients were matched for sex, age and preoperative stage of the disease with rectal cancer patients having conventional open surgery from January 1998 to October 2004 (OPEN group), by the same group of surgeons in different institutions. All operations were performed or supervised by the most experienced colorectal surgeon trained in TME surgery (E.X). Excluded were patients who refused to undergo laparoscopy, patients with distant metastasis detected preoperatively or identified intraoperatively and patients with severe co-morbidity or presenting as emergencies.

Preoperative management

Body mass index (BMI) was calculated, and American Society of Anaesthesiologists (ASA) score was assessed. Preoperative local staging of the disease included magnetic resonance imaging in all LAP and most of the OPEN cases. The rest of them had local staging with a computed tomography (CT) scan. Occasional local staging with an endorectal ultrasound was also performed. Upper abdominal and chest CT scan completed staging of the disease in all patients.

From 2001 onwards, patients fulfilling the criteria were given neo-adjuvant chemo-radiation. The criteria were: T4 tumours, T3 tumours threatening the CRM [13, 14], T2 tumours of the middle or lower rectum with an anterior rectal wall component, lymph node involvement and invasion of extramural mesorectal veins.

The clinical target volume to be irradiated included the primary tumour, the mesorectum, the presacral region, the pelvic side walls and the regional and internal ileac lymph nodes. With the patient at the prone position, the treatment was planned using a CT simulator and a three-dimensional conformal planning technique and delivered with a high energy 18MV linear accelerator. Either a three-field or a ‘box’ four-field technique was used to deliver a homogeneous dose to the tumour with minimum dose to small bowel and bladder. Preoperative fractionated radiotherapy was delivered in a period of 5 weeks usually with daily fractions of 1.8 Gy to a dose of 45 Gy to the entire pelvis followed by a boost to the primary tumour of 5.4 Gy (total dose 50.4 Gy in 28 fractions). As chemotherapy, 5-FU was given either as continuous infusion during the entire treatment period at doses of 200–220 mg/m2 with or without levamisole or as a two-course infusion (during the first and fifth week of radiotherapy). More recently, capecitabine has replaced 5-FU at doses of 1 g/m2 daily.

Surgical procedure

Patients to undergo conventional surgery had an epidural catheter prior to anaesthesia introduction. Duration of surgery and blood transfusion requirements were monitored.

Patients were placed on table in steep Trendelengourg position, with arms by the trunk, wearing inflatable leg boots and strapped at the chest. At the open approach, a mid-line abdominal incision from xyphoid process to symphysis pubis was made. At the laparoscopic approach five trocars were used. The 30° optic was introduced through the umbilicus trocar. Ligation and division of the inferior mesenteric artery was attempted 1–2 cm above the origin of the aorta, so to prevent damage to the hypogastric nerves. Splenic flexure was routinely mobilised. A total mesorectal excision was attempted in all patients as previously described [15, 16]. Either monopolar diathermy or harmonic scalpel (Ethicon Endosurgery, Inc, Cincinnati, OH, USA), or Ligasure–Atlas bipolar diathermy (Tyco Healthcare, Norwalk, CT, USA) were used for dissection. Transection of the rectum distal to tumour was attempted either by a linear stapling device (TA at open, Roticulator Endo-GIA at laparoscopy; Tyco Healthcare, Norwalk, CT, USA) or trans-anally with the use of scissors. At laparoscopy, the specimen was always excised through a 3–4 cm incision at the site of the left loin trocar, applying plastic wound protector. In case of staple-transection of rectum, anastomosis was performed with the use of a circular stapler (Tyco Healthcare, Norwalk, CT, USA). If transection was performed trans-anally, anastomosis was fashioned also trans-anally by hand-sawing. When anorectal stump was <4 cm, a transverse coloplasty at the proximal stump was made, prior to anastomosis. An ileostomy was added in case of (1) neo-adjuvant treatment, (2) anastomosis at <5 cm from anal verge, (3) hand-sewn anastomosis and (4) obese male patient. If preservation of anal sphincters did not guarantee adequate distal clearance, an abdominoperineal resection of the rectum (APRR) was undertaken. In this case, a wide resection of the lavator ani was performed, so to acquire a rather cylindrical specimen, without a waste configuration [14]. At laparoscopy, abdominal wall incision of any size not initially planed, in order to complete the operation, was registered as conversion-to-open.

Resected specimen were examined according to standard definitions [14, 16]. Macroscopic examination included assessment of posterior mesorectal fascia, Denonviller’s fascia or rectovaginal septum, anterior peritoneal fold and length of bowel wall distal to tumour. Standard microscopy included assessment of CRM and distal margin, grade of resection, and counting and examination of all retrieved lymph nodes. An at least 2-mm disease free circumferential margin was considered necessary for a complete resection [13].

Postoperative care

The concept of “fast-track” was not applied in the cases included in the study. The nasogastric tube was removed immediately after the end of the operation. Dose and type of analgaesia was given as needed and monitored. Low molecular weight heparin was given subcutaneously as prophylactic anticoagulant treatment during hospital stay. Per os liquid intake was initiated when bowel movements were detected by abdominal auscultation, and solid food when patient reported first gas passage, unless a locoregional was present. Daily until discharge and 1- and 3-month morbidity was recorded.

Adjuvant chemo-radiotherapy was given in patients with T4 tumours, T3 tumours threatening the CRM and involved lymph nodes and extramural nerves and veins at the histopathology report, who did not have neo-adjuvant treatment and apparently were missed at preoperative staging. If neo-adjuvant treatment was given, adjuvant chemotherapy was offered in patients with residual T3 tumour and node involvement. The ileostomy was closed 8 to 12 weeks postoperatively, after assessing integrity of the anastomosis.

Statistical analysis

Statistical analysis was performed with the use of the SPSS software package for Windows (SPSS 75.1, Chicago, IL, USA). Values are either expressed as median and range for non-parametric variables or as mean ± one standard deviation for parametric variables. The Chi-square (or Fisher’s exact test where appropriate) was applied to compare differences between groups in variables expressed in proportions. The Student’s t test and the Mann–Whitney U test for independent values were also applied to assess differences in parametric and non-parametric variables between groups, as appropriate. Comparisons between groups were made on an intention-to-treat basis. Differences were considered to be significant for P values of less than 0.05.

Results

There were 45 patients in the LAP and 43 matched cases in the OPEN group. No significant differences were accounted between groups in terms of age (p = 0.34), gender ratio (p = 0.83) and stage T of the disease (p = 0.4), as assessed by preoperative imaging. Although the overall difference in the BMI was statistically significant (27.12 vs. 29.07, p = 0.03), the proportion of patients with BMI > 30 was not (29% vs. 16%, P = 0.206). Threatened CRM ratio (21% vs. 31%, p = 0.336) and ratio of tumours with an anterior component (30% vs. 24.5%, p = 0.635) were similar between groups. Neo-adjuvant chemo-radiation was most often given to the LAP than the OPEN group (33% vs. 26%, p = 0.488), but the difference was not significant. That difference is attributed to the fact that neo-adjuvant treatment was initiated in late 2001, and at that time, there were OPEN cases already included in the study. Tumours in the LAP group were located significantly lower than those in the OPEN one (7.2 vs. 6 cm, p = 0.035; Table 1).

Table 1 Patients demographics and tumour characteristics
Full size table

On an intention-to-cure basis, a TME was attempted in all cases. Ten patients (22%) in the LAP and seven (17%) in the OPEN group had an APRR. Also, colorectal or colo-anal anastomosis was fashioned trans-anally with sutures more commonly in the LAP than in the OPEN (6% vs. 17%, p = 0.03). Increased ratio of APRR and hand-sawn anastomosis in the LAP group is attributed to the lower location of rectal tumours. Colo-rectal or colo-anal anastomosis were significantly closer to the anal verge in the LAP than in the OPEN group (3.6 vs. 5.5 cm, p < 0.001), apparently as a result of the aforementioned reason. There were 4 (9%) conversions to open: one case with a T4 tumour and suspected damage to the left ureter and three in whom a Pfannestiel incision was made to transect the rectum distal to the tumour. The latter three cases were male obese patients A covering ileostomy tended to be more commonly offered to the LAP than the OPEN group but the difference was not statistically significant (51% vs. 30%, p = 0.09). That was attributed to the larger proportion of patients with neo-adjuvant treatment and the lower location of the tumour in the former group. Operating time was significantly longer in the LAP than the OPEN group (166 vs. 143 min, p = 0.001). Operative duration values of the converted cases were within the range of the mean value ±1 SD of the duration of the laparoscopic group. Significantly more blood transfusions were required during or immediately after the operation in the OPEN group [2 (0–7) vs. 0 (0–4), p = 0.012; Table 2). Duration of the operation was significantly longer in patients with neo-adjuvant treatment as compared to those without, in the OPEN group (neo-adjuvant: 154 ± 45.74 min vs. no neo-adjuvant: 133 ± 22.1 min; p = 0.041). No such a difference was accounted in the LAP group (neo-djuvant: 158 ± 24.18 min vs. no neo-adjuvant: 160 ± 25.54 min; p = 0.78). Furthermore, more blood units were required to be transfused in patients with neo-adjuvant treatment as compared to those without, in both groups (LAP: neo-adjuvant: 2 (0–4) vs. no neo-adjuvant: 0 (0–3); p = 0.03, OPEN: neo-adjuvant: 3 (0–7) vs. no neo-adjuvant: 0 (0–4); p = 0.007).

Table 2 Operative details and histopathological characteristics
Full size table

In patients who had a LARR, the distance of the distal tumour margin to the distal resection margin was significantly longer in the OPEN than the LAP group (42.1 vs. 32.3 mm, p = 0.008), possibly as a result of the lower location of tumours in the latter one. However, it was in the OPEN group that the distal margin was involved in two cases (5%). A complete resected specimen without fasciae lacerations was achieved in more LARR cases in the LAP than the OPEN group, but the difference was not statistically significant even though it could be if the sample was larger (94% vs. 79%, p = 0.066). Also, the CRM was invaded not only in one case in the LAP but also in four cases in the OPEN group (2.2% vs. 9.3%, p = 0.198). Conceivably, there were more R0 resections in the LAP than the OPEN group, but this difference was not statistically significant as well (98% vs. 88%, p = 0.10). An enhanced specimen was acquired in all APRR cases in both groups. Numbers of retrieved and percentage of involved lymph nodes were similar between groups. As compared to preoperative imaging, pathological examination upstaged the disease in approximately one sixth to one fifth of the cases in both groups. Upstaging concerned lymph node status in the majority of the cases (Table 2).

Postoperative 1-month morbidity was significantly higher in the OPEN group (63% vs. 36%, p = 0.018). In detail, chest infection rate did not differ between the two groups (19% vs. 7%, p = 0.11), but more abdominal wound complications (31% vs. 9%, p = 0.008) and a higher rate of postoperative ileus (19% vs. 4%, p = 0.047) were observed in the OPEN group compared to the LAP group. Although not significantly, patients in the LAP group presented with a higher rate of anastomotic leak (16% vs. 10%, p = 0.522). In general, anastomotic leak was most commonly seen in male than in female patients (10/49 (20%) vs. 1/39 (3%), p = 0.02). In particular, patients with anastomotic leak had a significantly greater BMI as compared to those without anastomotic leak (31.18 ± 7.026 vs. 27.67 ± 3.693 kg/m2, p = 0.012). In all but two patients, one in each group, leak appeared in the presence of covering ileostomy and exhibited a subtle manifestation, consisting mostly of slight fever and fluid discharge per rectum. These patients were treated conservatively with success. In the remaining two cases, anastomotic leak was manifested with signs of peritonitis at the lower abdomen, and operative treatment included drainage and addition of ileostomy. The only death occurred in the OPEN group and was attributed to pulmonary embolism (Table 3).

Table 3 One-month postoperative morbidity
Full size table
Table 4 Postoperative in-hospital course
Full size table

Patients in the LAP group had a significantly faster recovery and rehabilitation. Namely, first passage of air per rectum, first bowel motion and resumption of oral intake occurred significantly earlier postoperatively in the LAP than in the OPEN group (p < 0.001). Similarly, postoperative hospital stay was significantly shorter in the former group of patients (p < 0.001). Furthermore, analgaesics, and in particular narcotic ones, were given for a significantly shorter period postoperatively in the LAP group (p < 0.001) (Table 4). At the 3-month postoperative follow-up, the rate of incisional hernias in the OPEN group was significantly higher than the one in the LAP group (32.6% vs. 6.7%, p = 0.003). On the other hand, any other kind of morbidity did not differ significantly between the two groups. Hospital readmission rate was also similar between groups (LAP group: three cases (7%) vs. OPEN group four cases (9%), p = 0.71). Two cases in the LAP and another two in the OPEN group were readmitted because of urine retention. The remaining three cases were readmitted because of either abdominal or perineal wound problems.

Satisfaction with the outcome of the operation was seen significantly more commonly in the LAP than in the OPEN group at the 1-month follow-up (LAP group: 31 patients well, 11 fairly and three not satisfied vs. OPEN group: nine well, 27 fairly and seven not satisfied; p = 0.035). That difference disappeared at the 3-month follow-up (respective figures in the LAP group: 25, 17 and three; and in the OPEN group: 19, 17 and six).

Discussion

Data from laparoscopic surgery for the treatment of rectal cancer are limited at present, and conclusions concerning safety and oncological efficacy of the approach cannot be drawn with certainty. In most of the published series, patients with either increased body mass index or bulky T4 tumours have been excluded, whilst the effect of neo-adjuvant treatment on the quality of the resected specimen and immediate postoperative morbidity have not been thoroughly investigated. In the present study, bias favouring selection of slim and in good status patients with early-stage disease has been avoided, as more than half of the cases included were of II or III ASA score, the median BMI was greater than 27 and more than two thirds of the patients had a stage II or III disease at preoperative imaging. Also, unlike other similar studies [2, 17–21], only mid and lower rectal cancer cases were included in the present one. This is because treatment modalities and surgical technique for these tumours differ substantially to those of the upper rectum, potentially resulting in different outcomes.

One third of the cases in the laparoscopic and one fourth of those of the open group had neo-adjuvant chemo-radiotherapy for down-staging or down-sizing of the disease, an independent factor of increased postoperative morbidity [22]. The increased rate of neo-adjuvant treatment in the laparoscopic as compared to open group of the present study is attributed to the fact that preoperative chemo-radiotherapy is being applied in our settings from 2001 onwards, and recruitment in the open group involved some cases having surgery before 2001. The non-significant difference in the incidence of APRRs in the laparoscopic group is not the result of selection but of the lower location of the tumours in this particular group of patients. In any case, the decision of the type of resection, LARR or APRR, was made at the time of the first examination of the patient, prior to image staging of the disease or the application of neo-adjuvant treatment. Neo-adjuvant chemo-radiotherapy has been shown to downstage or even downsize the disease and to reduce local recurrence [23–28]. Although there are no controlled studies comparing the short-course radiotherapy to the long-course chemo-radiotherapy preoperatively, the latter one seems to be associated with less morbidity and be more effective in reducing the size of the low rectal tumours, thus increasing the rate of sphincter preserving operations and reducing that of APRRs [26]. However, pathological response of rectal tumour to neo-adjuvant chemo-radiotherapy cannot be assessed with certainty in situ [29]. Therefore, the decision to offer a sphincter saving operation, taken some weeks after the completion of the treatment and on the clinical evidence of a diminished tumour, carries the potential risk of an oncologically compromised procedure and of an increased recurrence rate. For this reason, in the present study, the decision to offer a LARR or an APRR was taken after the first clinical assessment. However, the rate of APRR in either group is not excessive and is even less to that reported for middle and lower rectum tumours in the CLASICC trial [4]. The increased rate of APRRs in the laparoscopic group is again attributed to the lower location of tumours.

Similar to other reports [4, 8–10], data of the present study show that duration of the operation is significantly longer when performed by laparoscopy than by laparotomy, but blood loss as assessed by the need of blood transfusion is significantly less by the former approach. This might be of significance in the long-term outcomes, considering the concern of raised recurrence rate related to perioperative blood transfusion [30]. Neo-adjuvant chemo-radiation seemed to have a significant impact on blood loss in both groups but influenced the duration of the operation only in the open approach, possibly because more technical difficulties were accounted in that group of patients. Conversion rate in the present series was 9%, much lower than that reported in other series, where the respective figure varies from 10% to over 30% [4, 8, 17–21, 31, 32]. The independent predictors of conversion of laparoscopic to open surgery for rectal cancer are obesity, male sex, large T4 tumours of the mid rectum and surgeons experience [31,33]. According to some of these reports [4, 31, 32], conversion is associated with increased postoperative morbidity. The definition of conversion varies extensively in the literature: from small lower abdominal incisions to complete the colo-rectal anastomosis to large vertical abdominal incision for the mobilisation of the splenic flexure or because of inability to proceed laparoscopically at any stage of the procedure. In the present series any other incision on the abdominal wall than the originally planned for the extraction of the specimen was considered as conversion. The low rate of conversion in the present series is attributed to the fact that large rectal tumours had neo-adjuvant treatment and were down-sized, while the surgeons involved in laparoscopic surgery were adequately experienced having performed at least 100 laparoscopic colorectal cases before involved in the trial. The reason for conversion in the three out of the four converted cases of the present series was difficulty in distal transection of rectum and fashioning of the anastomosis in obese male subjects. Conversion was prompt resulting in duration of surgery similar to that observed in the non-converted cases.

Quality of resected specimen was similar between the two approaches. An adequate length of distal bowel margin was achieved and 18 lymph nodes at an average were retrieved by both approaches, while an intact CRM, an intact distal resection margin and a R0 resection were more likely to be achieved by laparoscopy than by laparotomy, although differences did not reach statistical significance possibly because of the limited number of cases included in the trial. This finding is different to the respective one in the CLASICC trial, where a positive CRM was more likely after the laparoscopic than the open approach, for reasons apparently inherited to the approach itself, according to the authors of the study [4]. However, other studies [18, 19, 21, 31, 32] do not confirm that finding and support the view that laparoscopy for rectal cancer yields an oncologically adequate specimen. An intact CRM is considered a more accurate indicator of oncological adequacy [13] than the number of lymph nodes retrieved because the latter depends mostly on the pathologist’s effort and on the neo-adjuvant chemo-radiotherapy if offered [25, 26]. The good quality of the resected specimen after laparoscopy is possibly the result of better visualisation deep in the lesser pelvis and the application of lesser traction to the tissues, thus, avoiding inadvertent lacerations of the mesorectal fascia. Also, the long scheme of neo-adjuvant treatment offers to the acquirement of a good specimen by either approach, in case of bulky T4 tumours.

Immediate overall postoperative morbidity was significantly less in the laparoscopic than the open approach, mostly because of more abdominal wound complications and postoperative ileus in the latter group. Also, although not significantly, respiratory complications were less common after the laparoscopic than the open approach, again unlike the CLASICC trial [4], according to which postoperative respiratory complications are more commonly seen after the laparoscopic approach, possibly because of the protracted duration of anaesthesia. Other authors [32] have also confirmed fewer medical complications after the laparoscopic approach. Anastomotic leak from the site of colo-rectal stapled anastomosis is reported to be constantly higher after laparoscopy than after open surgery for rectal cancer [4, 18, 19, 31, 34, 35], in rates varying between 10% and 20%. The respective rate of the present study was 16% in the laparoscopic and 10% in the open group. In the laparoscopic group, anastomotic leak was more likely to occur in male patients with BMI greater than 30. It has been suggested that the difficulty in achieving a distal transection line perpendicular to the axis of rectum in a narrow pelvis, the need for application of several cartridges that may result in a “zigzag”-ischaemic transection line at the distal stump and the tension at the anastomosis if the splenic flexure has not been taken down are the main predisposing technical factor for anastomotic leak [18, 19, 31, 32, 36]. In order to prevent anastomotic leak and until the optimisation of the endoscopic stapling devices, a colo-anal hand-sawn anastomosis, protected by a temporary ileostomy, is being suggested instead [31, 32, 36]. By increasing the rate of colo-anal hand-sawn anastomosis, Lelong et al. [32] observed a reduction of the anastomotic leak rate in their series of laparoscopic TMEs.

Similar to other reports [4, 32, 33], the findings of the present study showed that recovery after the laparoscopic approach for the surgical treatment of rectal cancer is significantly accelerated as compared to the open approach: bowel mobilisation and resuming of oral diet are faster, analgaesia in general and narcotic analgaesics in particular are given in lower doses and for fewer days, and hospitalisation is significantly shorter. Despite faster recovery and less morbidity, patients after laparoscopy expressed significantly more satisfaction with the outcome only by the first postoperative month, while at the 3-month follow-up, satisfaction rate was similar between groups. The only additional complain of patients after the open approach at the 3-month follow-up was the increased incidence of incisional hernia, however balanced by the promise of repair by the time of closing the ileostomy. Possibly, other, than faster recovery and less morbidity, factors, such as neo-adjuvant treatment, long-term urinary functional problems or presence of diverting stoma may influence further patient’s attitude toward a specific surgical approach. The application of quality-of-life scores in larger number of patients, where all different parameters are taken into account, may offer a definite answer to whether laparoscopic or open approach is associated with better biological, functional and social results.

In conclusion, laparoscopy for the surgical treatment of rectal cancer seems to be at least equally safe and effective to laparotomy, in terms of immediate postoperative recovery and morbidity. The laparoscopic approach may be even better in acquiring an oncologically adequate specimen, if performed by experienced surgeons. The 3-year results of the CLASICC trial [7] show that local recurrence and survival is similar between the two approaches. However, further studies are required to confirm those results or even show the potential superiority of the laparoscopic approach based on the observation that laparoscopy impairs immune function in lesser degree than laparotomy, at least in colon cancer surgery [37].