Long-term oncologic outcomes after laparoscopic vs. open colon cancer resection: a high-quality population-based analysis in a Southern German district

Laparoscopic procedures have been performed for oncologic diseases of the lower intestinal tract for many years. According to a systematic Cochrane review published in 2012 by Kuhry et al., “laparoscopic resection of carcinoma of the colon is associated with a long-term outcome no different from that of open colectomy” [3]. Recently, Deijen et al. reported the 10-year follow-up of the COLOR I trial and also came to the conclusion that “Laparoscopic surgery for non-metastatic colon cancer is associated with similar rates of disease-free survival, overall survival and recurrences as open surgery” [15]. Different representative trials from the Netherlands or the UK report laparoscopy rates exceeding 50%, whereas in other countries like Sweden or Germany, the proportion of laparoscopic procedures is often less than 20% [16‐19]. As far as the region observed in this study is concerned, medical underdevelopment does not seem to be the reason for low or even decreasing laparoscopy rates: three out of four resections were performed at specialized colorectal cancer centers. Obviously, doubt exists concerning whether the non-inferiority of laparoscopy proclaimed by many RCTs conducted in ambitious hospitals under ideal circumstances can be transferred to daily clinical practice. To answer this question, large observational studies are required. Unfortunately, truly population-based surveys on the topic are scarce. Although some large studies of this type are available from the United States, Canada, or the Netherlands, these focus exclusively on short-term outcomes. For example, Zheng et al. used the data of approximately 50,000 patients from the US National Cancer Database and reported a lower readmission rate after laparoscopic resections (laparoscopic 4.8% vs. open 5.5%; p = 0.003). These authors also observed a significantly lower perioperative mortality rate for laparoscopically treated patients (odds ratio (OR) 0.59; 95% CI 0.49–0.69 [20]). McKay et al. and Kolfschoten et al. used smaller but nevertheless representative study populations, and observed similar results [21, 22]. Apart from these studies, only a few trials deal with long-term survival. More than 10 years ago, Kube et al. conducted pioneering work and published results from a large German observational trial, reporting 5-year overall and tumor-free survival rates. The latter authors observed a substantial advantage for laparoscopy patients, with results quite close to the outcomes presented in this study (5-year overall survival rate open vs. laparoscopic: 66.9 vs. 82.8%; p = 0.005 [23]). It must, however, be mentioned that all hospitals participating in this study did so voluntarily and loss to follow-up seemed to play a substantial role. In 2016, Stormark et al. published their findings on long-term survival after laparoscopic colon surgery, with favorable results for the new approach [24]. These authors had at their disposal a considerable case number with an acceptable exclusion rate, used highly representative data from the Norwegian Colorectal Cancer Registry, and applied reasonably transparent statistical methods. In the same year, Benz et al. took data from 30 German cancer registries and compared long-term survival of open and laparoscopic colon cancer surgery up to a 10-year observation time [19]. Benz et al. also regarded laparoscopy to be a safe alternative to the open approach, although they excluded, for various reasons, a considerable number of patients who would generally have been eligible for inclusion; the external validity of these findings thus remains unclear to some degree.

Acknowledging the achievements and most common problems of previous publications, the current study aimed to minimize the risk of any form of bias. Before excluding any patient because of missing data, all possible measures to fill information gaps were undertaken. During this process, the four-eyes principle was always applied to match the high-quality standards of the Tumor Center Regensburg. If information on an important item was not available, exclusion of the patient was ultimately unavoidable. In order to quantify a potential bias caused by the exclusion process, a sensitivity analysis was performed. Interpreting the results of the sensitivity analysis, the conclusion can be reached that the presented outcomes are stable and not biased by missing data. Retrospective surveys such as the current analysis use data straight out of daily clinical routine and thus truly depict reality. On the downside, it has to be accepted that comparison groups are not created at random, which means that an indication bias always exists. Steele et al. could, for example, show that younger age and lower tumor stages are significant predictors for the selection of a laparoscopic approach [25], a result which is confirmed by the present study. Multivariate Cox regression analysis allows for correction of different risk profiles between groups. By adjusting for variables such as hospital status, tumor stage, or additional therapies, it was possible to cover a lot of inhomogeneities. One may question though, if it was justified to include the number of harvested lymph nodes as a covariable in the multivariate model. After all, there is a strong association between surgical quality and adequate lymphadenectomy [26, 27]. Adjusting for an associated survival benefit [28, 29] therefore disadvantages the superior surgical approach. On the other hand, if pathological examination tended to be less thorough with specimens of one approach, systematic understaging and worse survival rates would be the consequence. In order to avoid the latter problem, the harvested lymph nodes were finally included in the model. Unfortunately, there was no information available concerning non-oncologic comorbidities. This is probably the most important limitation of this survey, since weaker patients are more likely to die from strenuous treatment side effects or other conditions not directly associated with their tumor [30]. After all, cardiopulmonary problems are the most common cause of death, even in front of neoplastic diseases [31]. However, evidence exists that adjustment for age partially includes adjustment for comorbidities. The older a patient is, the more potentially life-threatening illnesses he or she suffers from [32]. According to survey of the Dutch Cancer Registry on colorectal cancer patients, there is a significant association between age and the number of a person’s comorbidities [33]. Notwithstanding this, systematic documentation of ASA or a different comorbidity score is required to be able to conduct an even more accurate risk adjustment.

Regardless of the statistical methods applied, the question of whether certain patient subgroups may benefit more from laparoscopy than others remains. Within the setting of this study, overall survival of patients with less to moderately invasive carcinomas was significantly positively influenced by laparoscopy. This supports the findings of the previously mentioned COLOR I trial, where the largest survival benefit after 5 years of observation time could be seen among UICC stage II patients, although the significance level was not reached [34]. Younger patients also seem to benefit from laparoscopic surgery, although the significance level in multivariate analysis is missed only narrowly. Whereas laparoscopy in low-risk situations can be recommended without restrictions, a closer look has to be taken in more advanced tumor stages and elderly patients. There is a certain suspiciousness concerning whether laparoscopy may be used with T4 patients. In the course of the COLOR I trial, half of the T4 patients undergoing laparoscopic surgery required conversion [34], which may be associated with worse survival [35, 36]. Thereafter, different publications addressed the topic with generally favorable outcomes for laparoscopy [37, 38]. The present study could also demonstrate that, even in high-risk situations like T4 or lymph node-positive tumors, laparoscopy is a non-inferior alternative. Moreover, it can add to the evidence that minimally invasive surgery should be considered for elderly patients, too [21, 39‐42]. Old people benefit from short-term advantages such as shorter hospital stays or lower complication rates like no other age group, without the need to fear a negative impact on their limited life expectancy.