Comprehensive Assessment of Diet Quality and Risk of Precursors of Early-Onset Colorectal Cancer

Xiaobin Zheng, MD, PhD; Jinhee Hur, PhD; Long H. Nguyen, MD, MS; Jie Liu, MD; Mingyang Song, MD, ScD; Kana Wu, MD, PhD; Stephanie A. Smith-Warner, PhD; Shuji Ogino, MD, MS, PhD; Walter C. Willett, MD, DrPH; Andrew T. Chan, MD, MPH; Edward Giovannucci, MD, ScD; Yin Cao, MPH, ScD


J Natl Cancer Inst. 2021;113(5):543-552. 

In This Article


In this large prospective cohort study of young women, Western diet was associated with an increased risk of early-onset adenoma, whereas intake more consistent with the prudent pattern and healthy DASH, AMED, and AHEI-2010 scores was associated with a lower risk. The associations were largely confined to high-risk adenomas, especially those in the distal colon and rectum. Coupled with secular trends toward low diet quality that is more prevalent among young adults in the United States,[16] our findings suggest poor diet may partially contribute to the rapid increase in early-onset CRC.

Human nutrition has changed dramatically over the past century, represented by increased intake of meat, fats, oils, and added sugars and sweeteners, as well as reduced consumption of vegetables and whole grains.[51] In addition to exploring the role of specific dietary elements in human health, researchers are increasingly focused on overall diet quality.[52] Poor diet quality has been considered a putative CRC risk factor in older individuals. A meta-analysis of 40 studies showed that Western diet was associated with an increased risk of CRC, whereas a prudent pattern was associated with a lower risk.[53] Findings have been mixed for recommendation-based indexes,[52,54] together with no association observed in our parallel cohort of older women.[52] Studies of early-onset CRC have been limited to case-control studies[55,56] from Pakistan and Italy. Our analyses are thus among the first prospective investigations of diet in early-onset colorectal neoplasia. The consistent findings across a posteriori dietary patterns and recommendation-based indexes, driven by high-risk adenomas, lend substantial support to the important and potentially stronger role of diet in early-onset than late-onset colorectal carcinogenesis.

Increases in early-onset CRC were primarily driven by distal colon and rectal tumors between 2004 and 2013, with incidence having increased by approximately 2% annually.[57] Intriguingly, we observed stronger associations between diet quality and early-onset advanced adenomas in the distal colon and rectum, compared to those in the proximal colon. These results were in line with prior findings on diet quality demonstrating a prominent association for distal colon and rectal tumors in older adults.[54,58] Proximal CRC is more likely to progress through the serrated neoplasia pathway,[59] whereas the vast majority of distal cancers originate from conventional adenomas associated with molecular alterations such as APC and TP53 mutations.[60] Indeed, prior analyses in 2 older cohorts showed that Western dietary pattern was more strongly associated with tumors that were MSS or microsatellite instability low, non-CpG island methylator phenotype, BRAF and KRAS wild type,[58] a molecular subtype common in early-onset CRC.[21] Taken together, our findings also indirectly support that diet may exert a stronger influence on neoplasia originating from the traditional adenoma-carcinoma sequence. As incidence for proximal tumors among younger individuals appears to be increasing with similar rates to distal colon and rectal tumors since 2012,[2] studies investigating the role of diet in recent years and other factors are warranted.

One of the potential mechanisms may relate to the summation of individual dietary constituents previously associated with CRC risk. For instance, a Western dietary pattern is high, whereas a prudent pattern and DASH, AMED, and AHEI-2010 scores are low in red and processed meats, which are known to be associated with increased risk of CRC.[61] The DASH diet is rich in low-fat dairy products, a good source of dietary calcium. Higher calcium intake has been inversely associated with CRC risk, especially for distal colon cancer,[62] which could be attributable to its known functions of reducing cellular proliferation and promoting cell differentiation and apoptosis.[63] Secondly, diets may influence risk of adenoma by regulating levels of intestinal inflammation[64] and altering gut microbial composition and diversity.[12,13] For advanced adenomas, microbial population shifts similar to CRC have been observed.[11] For instance, gut commensals such as Bifidobacterium animalis and Streptococcus thermophilus that could inhibit potential pathogens in the colon were found to be relatively depleted in both advanced adenomas and CRC tissue.[11] Last, unhealthy diets can lead to obesity,[65] which is associated with increased risk of early-onset CRC,[14] although our results did not change appreciably when body mass index was adjusted for.

Our study has several strengths. First, NHSII, with nearly 30 000 women younger than 50 years with at least 1 prior lower endoscopy and detailed data on indications, provided a unique opportunity to study early-onset adenomas, the most common precursors to early-onset CRC. Prior studies have reported comparable prevalence of adenomas in individuals aged 40–49 years vs 50–59 years who underwent employer-based screening colonoscopy.[35,36,66,67] These studies also in part indicated the emerging lower endoscopy practice in the younger individuals. Our similar advanced adenoma detection rate for those aged 40–49 years with the prior report[36] lent additional support for the reliability and generalizability of our data. Second, our dietary data, assessed by FFQ, captured long-term intake and have been validated and collected in a prospective manner, limiting recall or ascertainment bias.[68] Regular updates on dietary habits, accrued over 20 years of follow-up, also allowed us to evaluate long-term intake. Third, investigators masked to exposure information reviewed all the retrieved medical records and extracted data on histological subtype, which enabled us to perform subanalysis according to different adenoma subtypes as well as the malignant potential.

Several limitations need to be considered. First, as an observational study, the possibility of residual confounding could not be ruled out. Nevertheless, minimal changes after adjustment for a wide variety of putative risk factors of CRC indicated the robustness of our findings. Second, the dietary data were assessed using FFQs and subject to measurement errors. Using factor analysis to derive dietary patterns requires some decisions, such as the way to group individual food items into food groups, making it subjective to some extent for defining a posteriori dietary pattern. However, dietary measurement errors are expected to be nondifferential for CRC risk,[52] and it has been well established that repeated FFQs can accurately capture long-term dietary intake.[34] Finally, the generalizability of our findings to other populations, particularly men or other racial and ethnic groups, remains unknown.

In conclusion, higher scores for a Western diet were associated with an increased risk of early-onset adenoma overall, whereas intake of healthier dietary patterns (prudent diet and DASH, AMED, and AHEI-2010 scores) was associated with a lower risk, largely driven by associations for adenomas in the distal colon and rectum that were of high malignant potential. The slightly different associations based on dietary index classification system might inspire future work exploring the specific mechanisms involved. More detailed studies of differences in dietary index adherence and CRC risk by anatomic site in youth are also warranted.