Colorectal cancer risk difference between dietary patterns identified by posteriori and hybrid methods and its implications for dietary data analysis

Zegeye Abebe^1*2, Molla Mesele Wassie², Amy Reynolds², Yohannes Adama Melaku²

¹Department of Human Nutrition, Institute of Public Health, College of Medicine and Health Sciences, University of Gondar, Gondar Ethiopia

²Flinders Health and Medical Research Institute, College of Medicine and Public Health, Flinders University, Adelaide, South Australia

*Corresponding author: zegeye24@gmail.com

Abstract

Background: To estimate cancer risk associated with dietary patterns, we must summarize the existing evidence and determine which methods of a dietary pattern analysis approach better explain the variation in diet intake among study participants as well as the mechanism of the diet-disease relationship. As a result, this systematic review and meta-analysis was conducted to assess the association of principal component (PCA) and reduced rank regression (RRR) derived dietary patterns and risk of colorectal cancer (CRC) by including prospective observational studies.

Methods: Using Medline, SCOPUS, Cochrane Library, PsycINFO, ProQuest, and Web of Sciences electronic databases, all prospective studies published have been searched. Random-effect meta-analyses were used to combine the multivariable-adjusted relative risks (RRs) and 95% confidence intervals (CIs) comparing the highest and lowest categories of western, healthy, and drinker dietary patterns. We evaluated publication bias using Egger's test and funnel plot symmetry, as well as heterogeneity using I².

Results: During the initial search, 2146 articles were found, and 15 articles with a sample size of 975,993 participants were included in this systematic review and meta-analysis. PCA derived three dietary patterns (healthy, western, and drinkers) and RRR-derived two dietary patterns (healthy and western) were used to estimate CRC risk. The results indicated that a healthy dietary pattern was associated with a 10% reduced risk of CRC (RR pooled=0.90, 95%CI: 0.84, 0.96; I²=37.4%, P=0.249). However, western dietary pattern was associated with a 13% increased risk (RR=1.13, 95%CI: 1.04, 1.24; I²=42.4%, P=0.03), while the drinker dietary pattern (RR=1.16, 95%CI: 0.95, 1.40; I²=0%, P=0.88) had no association with CRC. Similarly, RRR-derived healthy (RR= 0.83; 95%CI: 0.61, 1.12; I²=41.4%, P=0.18) and western (RR=0.93; 95%CI: 0.57, 1.52; I²=84.3%, P=0.002) dietary patterns had no risk associated with CRC, however, the studies included are very limited.

Conclusion: PCA-derived dietary patterns are associated with the risk of CRC. This might apply to prioritizing nutritional interventions and educating people about nutrition to lower their risk of CRC. Considering the availability of fewer studies, RRR-derived dietary patterns did not appear to be associated with the risk of CRC. But it is crucial for establishing a link between dietary intake and health outcomes. Therefore, prospective studies are required to further support the findings of dietary patterns derived from RRR and cancer risks.

Keywords: Colorectal Cancer, Dietary Patterns, Principal Component Analysis, Reduced Rank Regression