4005 articles were retrieved from the initial search (Fig. 1). Following the exclusion of duplicate papers (n = 1238) and those that did not meet the inclusion criteria (n = 2737), 30 potentially relevant full-text articles were identified. Of the 30 papers, a further 7 articles were excluded. These studies used a case-control design (n = 4) [40,41,42,43], reporting an overall risk estimate of gastrointestinal cancer, but not CRC (n = 1) [44], and reporting colorectal adenoma as an outcome (n = 2) [45, 46]. Five duplicate articles were identified, of which two were related to the Netherlands cohort study [31, 47], three to the Nurses’ Health Study and Health Professionals Follow-up Study [30, 48, 49]. As these articles evaluated similar exposure variables, we included only the one with higher quality, with the greatest number of cases, or higher follow-up period [30, 31], and excluded others [47,48,49]. Two other duplicate articles were also identified, and all were included, given that different outcome variables (CRC, colon cancer, and rectal cancer) were investigated [13, 19]. Accordingly, 20 prospective cohort studies were included in the current systematic review and meta-analysis [13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32], of which fourteen evaluated the association between n-6 PUFA and CRC (10 on dietary n-6 PUFA intake [13, 15, 18, 21, 22, 29, 30], 3 on tissue levels of n-6 PUFA [16, 20, 23], and one on both [17]), eleven assessed the link between LA and CRC (7 on dietary LA intake [22, 24, 26, 27, 29, 30, 32], 3 on tissue levels of LA [14, 20, 23], and one on both [17]), nine assessed the relation between AA and CRC (5 on dietary AA intake [22, 24, 26, 29, 30], 3 on tissue levels of AA [14, 20, 23], and one on both [17]). In addition, three studies investigated the relationship between dietary or tissue levels of n-6 PUFA and colon or rectal cancers without considering CRC in relation to dietary or tissue levels of n-6 PUFA [19, 28, 31].

Flow diagram of study selection.

The characteristics of included studies, published between 1999 and 2021, are summarized in Supplementary Table 2. The number of participants ranged from 460 to 134,017, totaling 787,490, with an age range between 27 and 84 years. Duration of the follow-up ranged from 6 to 26 years, with a total of 10,694 CRC, 5417 colon cancer, and 2533 rectal cancer cases were recorded. Three studies recruited men only [16, 26, 27], five were conducted solely on women [22, 24, 25, 29, 32], and the remaining studies enrolled both genders [23, 28, 30, 31], of which, only four reported sex-stratified effect sizes [15, 20, 28, 30]. Five studies were from the United States (US) [15, 16, 22, 25, 30], one from Australia [17], and seven from European populations [18, 21, 23, 27, 29, 31, 32], and Asian [13, 14, 19, 20, 24, 26, 28] countries. In addition to the assessment of dietary intake at the study baseline, 5 studies repeated this assessment during the follow-up period [15, 24, 26, 28, 30].

Of sixteen studies on dietary intake of n-6 PUFAs, fifteen used food frequency questionnaire [13, 15, 17, 19, 21, 22, 24,25,26,27,28,29,30,31,32] and one used food diaries [18] for dietary assessment. Two studies collected dietary data through a face-to-face interview [19, 26], and others used self-reported data in their analysis [13, 15,