Diverticulosis and neoplastic lesions in screening colonoscopy: a large, multicenter study

Introduction

During colonoscopies performed in routine practice, a concurrent presence of diverticula and adenomas is often detected. Indeed, diverticulosis is vastly prevalent in Caucasian populations, particularly in the elderly, and an increasing prevalence in subjects under 50 years old has also been reported [1,2]. On the other hand, colorectal cancer, which mainly develops from colonic adenomas, remains among the most frequent tumors in Western countries, and its incidence is growing in young subjects [3,4]. Based on these epidemiological observations, an association between diverticulosis and neoplastic lesions in the colon could be hypothesized. Data from 2 meta-analyses showed that diverticulosis significantly increases (by 47-67%) the risk of adenomas, but not that of advanced adenomas or colorectal cancer [5,6]. However, a large cohort study based on a healthcare registry in Sweden found that diverticulosis was associated with a 71% greater risk of developing colorectal cancer [7]. Similarly, a 76% greater risk of developing early (<50 years) colorectal cancer in patients with preexisting diverticulosis was found by a United States of America study [4]. Conversely, a Taiwanese study found that the risk of colorectal cancer was not greater during the first year after a diverticulosis diagnosis [8]. Thus, the available data on the potential association between diverticulosis and colonic neoplastic lesions are still controversial. The conflicting results are probably due to heterogeneity among the studies, whose setting was widely ranging, from population-based retrospective studies to endoscopic, radiological or surgical series [5,6]. Moreover, some relevant differences were present regarding the clinical scenario, with studies mixing data from patients who underwent diagnostic or screening colonoscopy, as well as patients with diverticulosis or diverticulitis [9,10]. We designed the present study to further investigate this clinically relevant issue in subjects participating in an organized cancer screening program, using the fecal immunological test (FIT) and high-quality colonoscopy [11].

Patients and methods

Results

Outcomes

On univariate analysis, sex distribution was significantly different between the cases with diverticula and controls (M/F: 1222/954 vs. 1413/1461, P<0.001), whereas the age was similar (mean age: 62.7±6.2 vs. 60.9±6.5 years, P=0.9). Overall, the prevalence of adenomas in patients with diverticula (46.7%, 95%CI 44.6-48.8%) was significantly higher than in controls (43.8%, 95%CI 42.1-45.7%), prevalences of both advanced adenomas (13.8%, 95%CI 16.3-19.5% vs. 15.7%, 95%CI 16.3-19.5%), and cancers (2.4%, 95%CI 1.7-3.1% vs. 3.7%, 95%CI 3.1-4.4%) were lower, while the frequency of multiple adenomas did not differ among groups (Table 1).

Table 1 Data from the univariate analysis

On multivariate analysis, sex and age were found to be independent variables associated with different neoplastic lesions in the colon, whereas diverticulosis, as well as all the other variables considered, were not (Table 2). Adenomas, advanced adenomas and cancers were less prevalent in female than in male patients, and their prevalence increased with age, whereas patients with diverticulosis showed a lower risk of cancer.

Table 2 Variables associated with different neoplastic lesions on multivariate analysis

Discussion

Some epidemiological evidence has highlighted a potential association between diverticulosis and neoplastic lesions in the colon [4-9]. However, the literature provides conflicting data on this issue. Indeed, the results of some meta-analyses showed that the presence of diverticulosis significantly increases the risk of adenomas, but not advanced adenomas or colorectal cancer [5,6]. Conversely, a significant association between diverticulosis and colorectal cancer emerged in some large nationwide studies based on healthcare registries [4,7], but not in others [8]. Furthermore, a recent meta-analysis, including data from more than 27,000 patients with diverticulosis, found a statistically significant association between diverticulosis and either adenomas or advanced adenomas, but not with colon cancer [13]. At least in part, these conflicting results may depend on the wide heterogeneity among these previous studies. Therefore, other data are needed to unravel the association; the findings of our large study certainly provide further information about this clinically relevant issue.

The potential association between diverticula and adenomas in the colon could be either causal or spurious: namely, the result of shared risk factors. For the former hypothesis, colonic microbiota might produce some toxic compounds that could be trapped in the diverticular tract, resulting in prolonged contact with the colonic mucosa and promoting adenoma development [14]. For the latter hypothesis, a dietary intake low in fiber and rich in saturated fat could play a role in both conditions. Indeed, a low-fiber diet predisposes to the disordered motility of the colonic wall responsible for the formation of diverticula, and a fat-rich diet has been shown to be strongly correlated with an increased risk of adenomas [14]. Irrespective of the process involved, the determination of an association between diverticula and adenoma may affect both preventive measures and follow-up procedures.

We performed this large study to search for a potential association between diverticula and colorectal neoplastic lesions in an organized screening program. Indeed, this setting would appear ideal for such a kind of investigation, when we consider that it represents a quite homogeneous population in term of age and sex distribution, and a sample rich in neoplastic lesions, given the positive FIT result. Moreover, a high-quality colonoscopy is performed by expert and retrained endoscopists [12]. The findings of this large study were that the significant association between diverticulosis and adenomas detected on univariate analysis was not confirmed by multivariate analysis. Indeed, we found an even lower risk of cancer in subjects with diverticulosis who underwent screening colonoscopy. This finding suggests that the results of those studies where multivariate analysis was lacking, and a positive association was observed only on univariate analysis, should be interpreted with caution [15-18]. Our observations are in contrast with data from a similar study performed on 970 subjects who underwent screening colonoscopy in a single center, where the presence of diverticulosis emerged as an independent risk factor for both adenoma detection rate (OR 1.58, 95% CI 1.14–2.18) and advanced adenoma (OR 1.57, 95% CI 1.10–2.24), but not for cancer, on multivariate analysis [19]. However, in the same study [19], male sex and age were found to be independent risk factor for adenomas and advanced adenomas, as emerged in our analysis. The potential “protective effect” of diverticulosis against colorectal cancer we detected remains unclear, when we take into account both data from epidemiological studies, showing similar trend, and the shared risk factors, such as low-fiber diet, obesity and aging [20]. However, under-reporting of diverticula when a cancer is detected during colonoscopy [21], and missed lesions within the diverticular segment could be proposed as potential explanations [22]. Indeed, diverticulosis was found to be an independent predictor of interval cancer development in patients who underwent colorectal screening colonoscopy with an OR of 4.25 (95% CI 2.58-7.00) and of 6.00 (95% CI 5.57-6.46) in two studies [22,23].

An association between diverticulitis, rather than diverticulosis, and colon cancer has been proposed [24]. However, data from a recent Danish nationwide study showed that the prevalence of colon cancer was higher in diverticulitis (2.1%) than in matched controls (1.5%), but that the difference remained statistically significant only in during the first 6 months after a diverticulitis bout, whereas thereafter the risk was even significantly reduced (OR 0.8, 95%CI 0.7-0.9) [25]. It is likely that this was due to a missed cancer diagnosis at radiological evaluation in the acute setting, particularly in those patients with a complicated diverticulitis. Indeed, in a recent study of 91,993 colonoscopies performed following diverticulitis onset, it was found that colon cancer prevalence was significantly greater in patients with complicated (1.43%) but not uncomplicated (0.31%) diverticulitis, as compared to screening colonoscopies (0.33%) [26], confirming data from a previous systematic review [27]. Thus, the association between diverticulitis and colon neoplasia remains controversial.

In conclusion, this large, multicenter study found no significant association between diverticulosis and neoplastic lesions in subjects who underwent screening colonoscopy.