The impact of bariatric surgery on colorectal cancer risk

A literature search using PubMed was conducted to identify articles relevant to the review topic. Search terms used were “colorectal cancer,” “bariatric surgery” or “Roux-en-Y gastric bypass” or “sleeve gastrectomy,” “colorectal cancer risk,” “biomarkers of colorectal cancer,” “gut microbiota” or “microbiome,” “bile acid metabolism,” “short-chain fatty acids,” “butyrate.” Additional articles were identified from the references of the included studies.

Case reports, comments, and non-English publications were excluded from the review. In total, 95 papers (published between 1984 and 2021) were included in the review; of these, 66 were research articles, 27 were literature reviews, 1 was a systematic review, and 1 was a meta-analysis.

Bariatric procedures are classified into 3 main groups according to their mechanism of action in promoting weight loss. Malabsorptive bariatric surgery limits the absorption of nutrients by bypassing part of the small intestine to some degree []. In restrictive procedures, the size of the stomach is considerably reduced in order to induce an early sense of satiety in patients during food intake [,]. A further major category of bariatric procedures combines malabsorptive and restrictive components; these techniques involve significant reduction of available gastric capacity in conjunction with the bypassing of a section of the proximal small intestine [].

Laparoscopic sleeve gastrectomy (LSG), a restrictive procedure, is currently the most commonly performed bariatric technique worldwide []. LSG involves the transection and removal of the greater curvature portion of the stomach, leaving only a narrow tube along the lesser curvature. The popularity of this procedure is likely due to its promotion of significant weight loss and improvement in metabolic conditions, alongside a reduced risk of complications, compared with malabsorptive procedures [].

Roux-en-Y gastric bypass (RYGB), currently the second most common bariatric technique, is a combined restrictive and malabsorptive procedure []. This involves the creation of a gastric pouch from the upper stomach that is then anastomosed to the jejunum, forming the Roux or alimentary limb through which ingested nutrients flow. Anastomosis of the biliopancreatic limb with the jejunum allows the interaction of bile acids (BAs) and pancreatic secretions with nutrients in the common limb []. One anastomosis gastric bypass (OAGB) is a variation of gastric bypass, consisting of a single anastomosis of the gastric pouch to a loop of jejunum 150–200 cm from the ligament of Treitz [,].

Although bariatric surgery has proven to reduce overall cancer incidence, cohort studies have reported conflicting results on the impact of bariatric surgery on CRC. For instance, in 2010, Ostlund et al. conducted a population-based cohort study of bariatric surgery patients to test whether the risk of obesity-related cancer decreased with time after obesity surgery. The results of the study showed that CRC risk gradually and significantly increased with follow-up time after bariatric surgery. Patients followed up for more than 10 years showed a 2-fold increased risk. Such increases were not observed for the other main obesity-related malignancies, including breast, prostate, and endometrial cancer [].

Another population-based study, conducted by Derogar et al., identified an increased risk of CRC in patients with obesity undergoing surgery. An overall standardized incidence ratio (SIR) of 1.60 for CRC was reported for the obese surgery cohort, increasing to 2.00 after 10 years. In contrast, the SIR in the obese no-surgery cohort was 1.26, remaining stable over time []. This increased risk of CRC more than 10 years after surgery is consistent with the long natural history of colorectal carcinogenesis, from normal mucosa to malignant lesions. In both studies, bariatric surgeries consisted of restrictive procedures such as vertical banded gastroplasty and adjustable gastric banding, as well as malabsorptive procedures such as gastric bypass [,].

In addition, a large cohort study performed by Mackenzie et al. revealed an association between higher CRC risk and RYGB, although this association was not reported for sleeve gastrectomy or gastric banding [].

Likewise, a cohort study conducted across 5 Nordic countries demonstrated a higher risk of colon cancer in individuals who had undergone bariatric surgery, increasing further after ≥ 10 years. In contrast, the risk of rectal cancer was not significantly increased following bariatric surgery, although it appeared to increase with longer follow-up periods [].

These findings are in parallel with the results of a retrospective study conducted by Hussan et al., who examined the risk of colorectal polyp formation following RYGB. To focus on the long-term impact of RYGB, the authors compared colonoscopies performed 5 years or more after surgery with presurgery colonoscopies. Results revealed a higher percentage of serrated polyps (precursors of CRC) 5 years or more after surgery, suggesting an increased risk of precancerous lesion formation following RYGB [].

The impact of bariatric surgery is not limited to the incidence of CRC but also appears to affect prognosis. In 2016, a cohort study conducted by Tao et al. revealed that patients with CRC who had previously undergone bariatric surgery experienced higher mortality rates and a poorer prognosis than patients with CRC with obesity who had not undergone such surgery. When analyzed separately, the mortality rate was increased more than three-fold in patients with rectal cancer, while no statistically significant increase in mortality rate was found in patients with colon cancer [].

The negative impact of bariatric surgery found in these studies conflicted with the results of a meta-analysis conducted by Afshar et al. This revealed bariatric surgery to be associated with a 27% lower risk of CRC, although few studies were included in this meta-analysis, and follow-up was limited. Thus, lower CRC risk may have been related to weight loss alone []. Similarly, an English retrospective observational study found that obesity surgery was not associated with a higher incidence of CRC. However, the limitations of this study included follow-up period (3 years) and a small obese surgery cohort, compared with the obese no-surgery population []. In addition, a nationwide survey of CRC incidence following LSG or RYGB (conducted in Italy by the Italian Society of Obesity Surgery) revealed a low incidence of CRC (0.10%) 10 years after surgery. While this suggested that bariatric surgery had little impact on CRC development, the absence of no-surgery patients with obesity control group and the small number of cases observed (22 CRC cases in 20,571 bariatric patients) meant that the evidence was relatively weak, especially in terms of comparison of LSG and RYGB [].

The characteristics and findings of the investigations into CRC risk following bariatric surgery are summarized in Table 1. Their contradictory outcomes highlight the need to delve deeper into the mechanisms underlying this association.

A number of studies have investigated changes in CRC biomarkers following bariatric surgery. A study carried out by Sainsbury et al. in 2008 investigated mucosal biomarkers—considered indicators of future CRC risk [,]—in normal-weight individuals (n = 21) and patients with obesity undergoing RYGB (n = 24) []. Mucosal biomarkers of proliferation such as rectal epithelial cell mitosis, crypt area, and crypt branching were quantified by carrying out whole crypt microdissection; apoptosis was also evaluated by immunohistochemical staining for neo-cytokeratin 18. Before surgery, patients with obesity exhibited higher levels of rectal epithelial cell proliferation than normal-weight individuals, consistent with the link between obesity and CRC risk. Six months after RYGB, analysis of the same patients revealed a two-fold increase in the number of mitoses per crypt, compared with presurgery values. This further increase in rectal epithelial cell mitosis post-RYGB was accompanied by a decrease in the number of apoptotic cells. Systemic and mucosal markers of inflammation were also analyzed, including serum levels of C-reactive protein (CRP), cytokines, and mucosal pro-inflammatory gene expression. As expected, serum levels of CRP, interleukin (IL) 6, tumor necrosis factor α (TNFα), and macrophage migration inhibitory factor (MIF) were higher in patients with obesity than in normal-weight controls. After surgery, the patients with obesity had lower levels of CRP and IL-6 but significantly increased levels of TNFα and MIF, both of which are implicated in colorectal mucosal inflammation and carcinogenesis [, , ]. Furthermore, the expression of pro-inflammatory genes cyclooxygenase-1 (COX-1) and COX-2 increased significantly following RYGB surgery []. COX-2, in particular, is known to be overexpressed in CRC and appears to play a major role in its development and progression [,].

Interestingly, these mucosal biomarkers were quantified in the same group of patients (n = 19) 3 years after RYGB to determine their long-term persistence. Rectal epithelial cell mitosis and crypt size remained abnormally increased 3 years after RYGB, compared with preoperative values. COX-1 and COX-2 messenger RNA (mRNA) levels also remained elevated, similar to the values observed 6 months post-RYGB. A remarkable outcome of the study was the marked upregulation of MIF in the rectal mucosa 3 years after surgery. The authors observed a 40.5-fold increase in mucosal MIF mRNA levels compared with pre-RYGB values, as well as higher MIF protein levels in colorectal epithelial cells. However, serum MIF was reduced, compared with levels observed 6 months after surgery, indicating that MIF upregulation was a local pro-inflammatory phenomenon rather than a systemic response [].

A similar study conducted by Afshar et al. investigated the impact of RYGB on biomarkers of CRC risk. This study included 22 post-RYGB patients and 20 normal-weight controls. Notably, in this case, the results revealed lower rectal crypt cell proliferation and reduced systemic and mucosal markers of inflammation 6 months after RYGB []. However, the contrasting outcomes of the 2 studies may have been due to the different characteristics of the populations and differences in the surgical procedures [,]. For example, the decreased levels of inflammatory markers identified by Afshar et al. may have been due to the lower mean BMI of patients in this study, compared with the patients studied by Sainsbury et al. Furthermore, a remarkable difference between the 2 investigations was the length of the bypassed tract: the biliopancreatic limb was 150 cm long in the study of Sainsbury et al., whereas it was 60–75 cm long in the study of Afshar et al. It has been suggested, therefore, that the length of the biliopancreatic limb may be a key factor in the hyperproliferation and inflammation of the rectal mucosa [,]. This is supported by the results of a study conducted by Appleton et al. on patients with obesity who had undergone jejunoileal bypass (JIB). In these patients, the bypassing of a long portion of the small intestine led to a marked and persistent increase in rectal crypt cell proliferation [].

According to this hypothesis, these effects should not occur after restrictive bariatric procedures, where intestinal flow remains largely unaltered. Kant et al. tested this assumption by comparing mucosal biomarkers of CRC risk in patients undergoing LSG (n = 21) and normal-weight controls (n = 20). Unlike the results observed following RYGB, there was no significant change in rectal epithelial cell proliferation 6 months after LSG. Furthermore, mucosal MIF mRNA levels increased by 39% following LSG, but there were no changes in the expression of TNFα, IL-1, IL-6, COX-1, or COX-2 over the same period [].

Levels of serum biomarkers and expression of pro-inflammatory genes prior to bariatric surgery, 6 months after RYGB, 3 years after RYGB, and 6 months after LSG (reported by different authors) are summarized in Table 2.

The mechanisms leading to hyperproliferation of the rectal mucosa following RYGB and malabsorptive procedures (but not LSG) are still unclear. Nonetheless, several possible pro-carcinogenic pathways have been proposed, as outlined below.

Alteration of the gut microbiota has been suggested to be one of the main potential mechanisms linking bariatric surgery to an increased risk of CRC. This hypothesis arises from the intriguing finding that some of the gut bacteria found to be increased after bariatric surgery have been associated with CRC due to their pro-carcinogenic properties.

Bariatric surgery (especially RYGB and malabsorptive procedures) induces major anatomical rearrangements in the gastrointestinal tract, altering nutrient transit and affecting gut physiology. Malabsorption, changes in the metabolism of BAs, alteration of pH, and modulation of enteric and adipose hormones all affect the composition of the gut microbiota [, , ]. Accordingly, changes in the gut microbiota following RYGB have been widely demonstrated. The most commonly reported alterations in post-RYGB patients are a marked increase in the phylum Fusobacteria, higher levels of the class Gammaproteobacteria and the genera Bacteroides and Escherichia, and a decrease in the abundance of the phylum Firmicutes [, , , , , ].

The impact of LSG on gut microbiota composition has also been investigated. Three studies on human patients found that LSG led to changes in the fecal microbial community similar to those observed after RYGB; however, the alterations were more moderate than those seen following RYGB, demonstrating the lower impact of LSG [, , ]. These 2 bariatric procedures were further compared in terms of microbiota alteration by Shao et al. in a rat model. Analysis revealed a striking shift in the gut microbiota following RYGB. However, while RYGB was associated with a significantly higher proportion of Proteobacteria, the post-LSG microbiota was comparable to that of sham-operated rats [].

Overall, these findings imply that bariatric surgery results in procedure-dependent alterations in gut microbiota. In comparison with LSG, RYGB results in greater modification of the gastrointestinal tract and intestinal environment, resulting in more severe and profound effects on microbiota composition [,]. This may be due to the effect of the shortened small intestine in RYGB, as increased oxygen concentration in the normally anaerobic distal lumen favors the presence of facultative anaerobes (such as Gammaproteobacteria) over obligate anaerobes (e.g., Firmicutes) [,]. Furthermore, the bypassing of the upper small intestine in this procedure might result in the relocation of some of the typical small intestine microbiota, such as Enterobacteriaceae, to the large intestine [].

Another important modification related to RYGB is the reduction in stomach size and subsequent alteration in pH. Indeed, reduced gastric volume is associated with significantly decreased acid secretion in the gastric pouch, leading to higher pH levels []. Evidence regarding the effect of LSG on gastric acid secretion is somewhat limited. However, it has been hypothesized that the anatomical rearrangements associated with this bariatric procedure could also contribute to reduced gastric acid secretion, albeit to a lesser extent than RYGB, as some acid-producing cells are still present []. Increasing the pH in the gastrointestinal tract induces major changes in gut bacterial populations since it produces a more favorable environment for Bacteroides and inhibits the growth of butyrate-producing bacteria [,]. This evidence is confirmed by the study of Walker et al. who tested, in vitro, the response of microbial communities to pH increase and found higher levels of the genus Bacteroides and a decrease of Roseburia spp. and Faecalibacterium prausnitzii abundance []. Roseburia spp. and F. prausnitzii are considered the main butyrate-producing bacteria in the human gut [] and appear to have a protective role because of the beneficial and anti-inflammatory effects of butyrate [,]. Decreased levels of both bacterial groups (belonging to the Firmicutes phylum) have been observed in postsurgery patients. [,,,].

An indirect effect of RYGB is altered microbial catabolism, which produces changes in metabolite levels. Gastric acid secretion, which normally ensures the activation of proteolytic enzymes and protein digestion, is reduced after RYGB []. As a consequence, more incompletely digested proteins reach the colon, leading to increased microbial protein catabolism. This, in turn, results in elevated production of toxic polyamines such as putrescine [,]. Fecal water from post-RYGB rats has been observed to be highly cytotoxic, likely correlating with a postoperative increase in the levels of fecal putrescine, given the deleterious effects of this polyamine on cell survival [,]. Importantly, it has been reported that the levels of putrescine and other polyamines are elevated in patients with cancer and are correlated with tumor stage and progression [,]. Relevant to this finding, putrescine appears capable of stimulating cell proliferation in a dose-dependent manner. Furthermore, the accumulation of putrescine has been associated with colon inflammation, high levels of pro-inflammatory cytokines, and increased gut permeability [,].

Various human studies have reported, in the gut, the increased presence of oral microbes including Streptococcus spp., Veillonella spp., and Fusobacterium nucleatum following RYGB [,,]. Intestinal colonization by the oral microbiota could be related to increased pH and decreased acid secretion, which reduce the efficiency of the gastric barrier, facilitating the orofecal transit of bacteria [,].

Interestingly, there are several similarities between the fecal microbial profiles observed following bariatric surgery (particularly RYGB) and those of patients diagnosed with CRC. Indeed, several studies have analyzed and compared gut microbiota composition in patients with CRC and healthy individuals, revealing enrichment of the phylum Fusobacteria and the genera Bacteroides and Escherichia in the CRC group [, , , , , ]. Furthermore, the results of these studies indicate significantly decreased abundance of the butyrate-producing genera Faecalibacterium and Roseburia in patients with cancer, compared with controls [,,]. Zhu et al. found similar results using a murine model of CRC, observing increases in Proteobacteria, Fusobacterium, and Bacteroides populations and significant reductions in Roseburia and Eubacterium abundance []. Therefore, the microbial dysbiosis associated with CRC leads to both enrichment of pro-inflammatory microbes and depletion of butyrate-producing bacteria, resembling the effects of RYGB [,].

Alteration of microbiota composition is emerging as a major factor associated with inflammation and tumorigenesis that appears to modulate CRC risk and development []. Zackular et al. used a murine model of inflammation-associated CRC to determine the role of gut microbiota in the development of CRC. This study demonstrated that microbial alterations contribute to CRC through the initiation of inflammation and have a causal role in exacerbating tumor formation []. The authors hypothesized that enrichment of the genus Bacteroides, observed in the fecal samples of their murine model and in patients with CRC [,], may contribute to tumorigenesis []. In support of this hypothesis, Wu et al. found a positive correlation between Bacteroides abundance and tumor stage in patients with cancer []. Furthermore, according to a study conducted by Sobhani et al., Bacteroides populations were linked with high proportions of pro-inflammatory IL-17-producing cells in tumor samples and normal mucosa in patients with CRC []. Escherichia coli, a member of the Gammaproteobacteria class, is also enriched in post-RYGB patients [, , ] and has been found to be associated with colorectal carcinogenesis. Interestingly, the intestinal mucosa of patients with more advanced colorectal neoplasia is reportedly colonized by pathogenic and virulent E. coli strains with genotoxic properties [,].

Several studies have focused on the colorectal carcinogenic effects and tumor-promoting mechanisms of F. nucleatum, a pro-inflammatory bacterium. The increased presence of F. nucleatum in stool samples represents a further microbial alteration common to bariatric and CRC patients [,]. F. nucleatum has also been detected in premalignant colorectal lesions, suggesting its involvement in neoplastic initiation and the early stages of tumorigenesis [,]. Moreover, increased abundance of this species in patients with CRC has been associated with shorter survival time [].

The similarity of the microbial changes observed in postbariatric patients and patients with CRC (Table 3) is striking. Such changes highlight a potential mechanism by which the alteration of gut microbiota associated with bariatric surgery could lead to an increased risk of CRC.