Colorectal cancer (CRC) is the second-leading cause of cancer-related death in developed countries and has the third-highest incidence worldwide (1). After age, a family history of the disease is the most important risk factor for CRC. As many as one-third of CRC cases occur in first-degree relatives (FDRs) of patients with CRC, and ~10% of the general population comprises people with at least one FDR affected by the disease (2). The term “non-syndromic familial-CRC” refers to patients with suspected inherited CRC in whom a genetic mutation is not detected.
The FDR of patients with CRC have a two-fold or higher risk of developing CRC than that of the general population (3). In fact, having a FDR with CRC diagnosed before the age of 60 years, being a sibling of a patient with CRC, or having ≥2 affected FDR is associated with a 10–15% lifetime risk of CRC. In these individuals, the current guidelines recommend screening with colonoscopy every 5 years (4), starting at the age of 40 or 10 years before the youngest case (5). In individuals with one FDR diagnosed with CRC after the age of 60 years (lifetime CRC risk <10%), several guidelines recommend a less intensive approach based on annual or biennial fecal immunochemical testing (FIT) (5).
Overall, these recommendations are empirical in nature and are primarily based on retrospective studies with considerable methodological flaws. Apart from memory bias, inadequate risk stratification, and a lack of colonoscopy quality assessment, the most relevant limitation is that these studies have focused on the risk of detecting CRC or advanced adenoma, not on the benefit of screening colonoscopy to decrease CRC mortality or incidence in this population. Indeed, there are no data demonstrating that the natural history of non-syndromic familial CRC differs from that of sporadic CRC. This finding is supported by the observation that the risk of adenoma recurrence is more closely related to age, sex or the histological features of the lesion at the index colonoscopy than to family history (6, 7).
In Spain, as in many other European countries, population-based screening programs with biennial FIT have been progressively implemented for average-risk individuals. Individuals with higher risk (non-syndromic or hereditary CRC) are directed to high-risk CRC clinics, as they require colonoscopy. Two problems arise. The first is that FDRs of CRC patients are not systematically identified, and patients are therefore subject to unsystematic screening. The second is that there is poor adherence to colonoscopy in Spain. Two studies assessing the participation rate for colonoscopy in the FDRs of patients with CRC (8, 9) showed that only 38% (8) and 42.5% (9) of invitees underwent colonoscopy. To overcome these problems, a plausible solution would be to integrate the FDR of patients with non-syndromic familial CRC into organized FIT-based screening programs. This would also help to identify patients with syndromic-CRC, particularly individuals with Lynch syndrome. Supporting this hypothesis, recent evidence has shown that adding familial-risk assessment to a FIT-based screening program may help to identify otherwise undetected at-risk individuals, thus enabling them and their relatives to undergo screening (10).
Several studies in Spain, as well as in other European countries, support the use of FIT to screen the FDR of patients with non-syndromic familial CRC (Table 1)(11, 12, 13, 14, 15, 16). First, a large study comparing the prevalence of advanced colorectal neoplasia (ACN) between average and familial-risk individuals undergoing their first screening colonoscopy has shown that people with a single FDR diagnosed with CRC have a prevalence of ACN similar to that of average-risk individuals, regardless of the age at which CRC was diagnosed (Table 2)(17). However, individuals with two FDR diagnosed with CRC at any age are nearly twice as likely as average-risk individuals to have ACN. In the multivariate analysis, elderly age, male sex, and ≥2 FDR affected have been independently associated with a higher risk of ACN. In addition, results from a longitudinal study performed within the randomized controlled prostate, lung, colorectal, and ovarian cancer-screening trial of flexible sigmoidoscopy vs. usual care are in line with these findings (18). After a follow-up period of 13 years, individuals with two FDR with CRC have been found to have a two-fold higher risk of incident CRC, whereas those with one affected FDR do not have a significantly increased risk of CRC incidence or mortality, regardless of the age at the time of diagnosis. Together, these prospective studies suggest that most individuals in the familial-risk population (those with one affected FDR) could follow the same screening strategy as the average-risk population. Second, the reasonable performance of FIT for CRC detection in the average-risk population and the suboptimal adherence to colonoscopy in the FDR of patients with CRC has been noted by European authors. Therefore, FIT has been considered a potential screening alternative to colonoscopy in familial-risk individuals (11, 12, 13, 14, 15, 16).
Although there is currently no evidence of the benefit of FIT with regard to major outcomes, such as the incidence or mortality in FDR, several prospective studies have assessed the diagnostic performance of this test (11, 12, 13, 14, 15, 16). A recent meta-analysis of 12 studies (n=6204 participants) (19) has assessed the accuracy of FIT in detecting colorectal neoplasia, using colonoscopy as the gold standard. Eleven of the studies were cross-sectional studies, and one was an RCT. Eight were European, five of which were from Spain, including the only RCT published to date. Overall, the diagnostic accuracy of one-time FIT for CRC and ACN detection in the FDR of patients with CRC was sensitivity 86% (95% confidence interval (31–99%)) and 46% (37–56%), respectively; specificity 91% (89–93%) and 93% (91–95%), respectively; likelihood ratio +10.0% (5.8–17.5%) and +6.9% (4.9–9.7%), respectively; likelihood ratio −0.16% (0.02–1.48%) and −0.56 (0.46–0.69%), respectively. The results remained unchanged when only studies that used a quantitative FIT with a cut-off set at <25 μg/g feces were included. Although the authors have concluded that FIT offers a reasonable accuracy for CRC detection similar to that observed in average-risk individuals, the high heterogeneity of the studies, the large confidence intervals, small sample sizes, and low prevalence of CRC call the reliability of the conclusions into question.
The only RCT published to date has compared the efficacy of yearly FIT during 3 consecutive years vs. one-time colonoscopy in FDR of patients with CRC (15). In that study, the FIT-based screening strategy has been found to be equivalent to colonoscopy for detecting ACN in both the per-protocol (odds ratio 1.56 (0.95–2.56)) or the intention-to-screen (odds ratio 1.41 (0.88–2.26)) analyses. However, the study had some methodological flaws. First, because two thirds of the FDR included had only one index case >60 years, the study was underpowered to discern whether the detection rate of ACN observed with FIT in the entire FDR group could be extrapolated to other subpopulations (i.e., FDRs with index cases <60 years, with two index cases, or with at least one affected sibling, regardless of the age at diagnosis). Second, the study was not designed to compare the acceptance of both strategies, because FDR of the same index case were not systematically invited to participate. To elucidate the adherence issue, a multicenter RCT is underway in Spain (clinicaltrials.gov NCT02567045). The hypothesis of that study is that the participation rate of annual FIT, followed by colonoscopy if there is a positive test, is significantly higher than that of straightforward colonoscopy.
We believe that the holy grail for CRC screening will be in offering different alternatives beyond colonoscopy to increase the participation rate. In this regard, and taking into account the low adherence rate to colonoscopy screening in the familial-risk population (8, 9), we have recently carried out an RCT to assess whether colon capsule endoscopy improves the colonoscopy rate in this at-risk population (20). Unexpectedly, the efficiency of colon capsule endoscopy was compromised by a poor participation rate because of the need for another bowel preparation in the event of significant findings in the colon capsule endoscopy examination.
In summary, the Spanish perspective is that the recommendations regarding CRC screening in the FDR of patients with CRC should be revised, especially for individuals with only one index-case affected in the family. In our view, the current evidence does not support that these individuals should follow a different screening strategy from that of the average-risk population. Therefore, FIT could be offered as an alternative screening strategy to this subgroup if forthcoming studies demonstrate a higher uptake for FIT than colonoscopy. However, colonoscopy is still the best option in those FDRs with ≥2 index cases. It should be kept in mind that the main goal of screening in the non-syndromic familial-risk population should be achieving a high adherence rate, regardless of the screening test performed.