Is a fusobacterium nucleatum infection in the colon a risk factor for colorectal cancer?: a systematic review and meta-analysis protocol

In 2012, over a million new cases of colorectal cancer (CRC) and more than half a million deaths due to CRC were estimated to occur globally [1]. Of cancers that affect both men and women, CRC is the third most commonly diagnosed malignancy and the fourth most fatal in the world [2]. The numbers of new cases of CRC and CRC deaths are expected to increase to 60% by 2030 [2].

The etiology of CRC is widely recognized as being multifactorial [3, 4], and previous research has suggested that modification of environmental and lifestyle factors can lead to important changes in cancer risk [5, 6]. Still, according to comprehensive reviews of the available evidence conducted by expert panels from the American Institute for Cancer Research and the World Cancer Research Fund, the overall evidence for the causal nature of the association with CRC is considered convincing for only some of the previously suggested factors, namely, excess body fat, consuming processed meats and red meat, physical inactivity, cigarette smoking, and alcohol consumption [7]. For the majority of the putative risk factors, the level of evidence is considered either fair or inadequate [7‐10]. Thus, identification of modifiable risk factors that could serve as targets for preventive interventions is a current public-health priority.

In the past few years, advances in high-throughput sequencing technologies have led to important discoveries on the role of gut microbial dysbiosis and specifically, of Fusobacterium nucleatum (F. nucleatum) in colorectal carcinogenesis [11‐21]. F. nucleatum is a Gram-negative, non-spore-forming anaerobic bacterium commonly found in saliva and oral biofilm [18, 22, 23] . It is one of the dominant species of more than 500 organisms of the oral cavity and has five subspecies with different specific genome sequences [24‐31]. This invasive proinflammatory agent is involved in the pathogenesis of periodontal diseases [22] as well as of other oral [32] and extra-oral infections [33, 34]. F. nucleatum can independently invade host cells via surface adhesins and invasion molecules such as FadA [21, 35]. Importantly, once disseminated outside the oral cavity, FadA activates proinflammatory and oncogenic signals and stimulates the growth of epithelial cells. Human studies have demonstrated that the FadA gene level in CRC tissue is higher than in normal tissue and is correlated with expression of inflammatory genes [21]. Furthermore, a recent study found a strong correlation between F. nucleatum and proinflammatory markers such as COX-2, IL-8, IL-6, IL1ß, and TNF-α in CRC [15]. This evidence suggests that colonization resistance of the healthy gut can be disrupted by bacterial species that trigger a systematic inflammatory response, such as seen in periodontal disease. In a study by Dejea et al. [36], the rate of CRC occurrence was more than five times as high in individuals with gut bacterial biofilms as in those without them [36]. Interestingly, the gut bacterial biofilm composition and invasiveness were similar to those found in oral biofilm in periodontal disease, with Fusobacteria being a dominant species [36].

F. nucleatum is now considered to be a pathogenic bacterium of the gut that can invade the colorectal submucosa and epithelium. Various studies have shown an overabundance of F. nucleatum in tumors and fecal samples [37] of CRC patients [15, 17, 19‐21, 38] . Additionally, some studies have demonstrated that levels of F. nucleatum increased in parallel with the transition from healthy colorectal tissue to adenomas and finally to CRC [39‐41]. F. nucleatum levels in cancerous colorectal tissue have also been shown to serve as a prognostic indicator in CRC [11, 39, 42]. In vitro and in vivo studies showed that F. nucleatum interrupts oncogene signaling and cell–cell adhesion and inhibits the anti-tumor activities of natural killer and cytotoxic T cells as well as anti-tumor immunity [38, 43]. Increased levels of F. nucleatum have been shown to be associated with microsatellite instability and molecular subsets of CRCs such as the CpG island methylator phenotype [11, 44]. Decreased expression of MLH1, a primary cause of microsatellite instability, was found in samples abundant in F. nucleatum [13, 42]. Other markers of poor prognosis such as KRAS and BRAF are also overexpressed in samples rich in F. nucleatum [13, 45, 46]. Moreover, CRC patients have been found to have an increased level of serum anti-F. nucleatum antibodies [47].

The literature on the association between F. nucleatum and CRC is growing but has not yet been systematically reviewed to date. We aim to conduct a systematic review of observational studies on the association between F. nucleatum and CRC.

The aim of this review is to systematically identify, review, and assess the quality of available literature on the association between F. nucleatum and CRC. The findings of this systematic review will help answer the following question: does F. nucleatum play a role in the etiology of CRC? If feasible, a meta-analysis will be conducted to estimate pooled measures of association between F. nucleatum and CRC.

Studies will be selected according to the following criteria:

An electronic literature search will be conducted in the following databases: MEDLINE (OVID interface, 1946 onwards), EMBASE (OVID interface, 1974 onwards), Web of Science, and Cochrane Database of Systematic Reviews (CDSR). Since the vast majority of research conducted on this topic has been carried out in the past 10 years, the start date for the literature search will not be limited in order to maximize the number of publications considered. The electronic literature search will be complemented by hand searching the list of references in the identified publications or relevant reviews. NICE Evidence and TRIP database will be searched for gray literature using subject keywords. Ongoing studies (which have not resulted in publications on the topic at issue) will not be considered in the review.

Using medical subject headings (MeSH), EMTREEs and text words related to the field of the study, the research team has developed a draft version of literature search strategies with the help of an expert librarian at Université de Montréal (Additional file 2). This draft of the MEDLINE search strategy will be finalized and adapted to the other databases using the proper syntax, subject headings, and controlled vocabulary considering maximized sensitivity of the search. In order to maximize the yield of the search strategy, no language restrictions in the search strategy will be used. Hand searching the list of references in the identified publications, NICE Evidence and TRIP database will be done to identify the relevant articles in the gray literature. The PROSPERO will be searched for recently completed systematic reviews on the topic. The references of relevant studies will be verified for relevant publications.

As stated in the outcomes section, colorectal cancer will be the study outcome.

Two reviewers will independently assess the risk of bias of the eligible studies using the Newcastle-Ottawa scale (NOS) [51]. This 8-item scale allows to assess the quality of the articles based on the selection of the study groups; the comparability of the groups; and the ascertainment of either the exposure or outcome of interest for case-control or cohort studies, respectively. Disagreement will be resolved by consultation with a third reviewer.

We will also evaluate the level of evidence of all studies according to the Oxford Level of Evidence [52].

The potential publication bias will be assessed by funnel plots [54]. Tests for funnel plot asymmetry will be conducted if the number of studies included in the meta-analysis is more than 10 [55].

The Oxford Level of Evidence [56, 57] and The Grading of Recommendations Assessment, Development, and Evaluation (GRADE) approach will be used to evaluate the level of evidence of all studies.

Any deviations from the protocol due to unanticipated issues will be reported in the final review [58].