The Association between Prebiotic Fiber Supplement Use and Colorectal Cancer Risk and Mortality in the Women's Health Initiative

Colorectal cancer is the third most common cancer among women in the United States, accounting for approximately 8% of new female cancer cases annually (1). Fiber is often promoted as a dietary colorectal cancer prevention strategy, with daily intakes of ≥25 grams (2) recommended as part of a healthy diet (3). Prebiotics are defined as carbohydrates that are resistant to digestion and are instead fermented by gut microbiota in the colon; these can be both soluble and insoluble fibers in foods or supplements (4). Soluble prebiotic fibers form viscous gels when dissolved in water, whereas insoluble prebiotic fibers do not dissolve in or gel in water (5). Such isolated prebiotic fibers may have a protective effect in colorectal carcinogenesis as indicated in some studies (6–8). Prebiotics are widely available as over-the-counter dietary supplements, also labeled as fiber-containing bulk laxatives, and commonly are used to regulate bowel movements and alter digestive symptoms.

Fiber maintains a critical physiologic role in digestion where it promotes healthy bowel functioning including reduced transit time and bulking of stool (9). Other potential protective mechanisms of fiber include fermentation of fiber in the colon by microbiota producing butyrate, a substrate that inhibits carcinogenesis and promotes normal homeostasis of the colonic epithelium (10, 11). Metabolic syndrome (MetS), a clinical condition characterized by higher waist circumference, dyslipidemia, insulin resistance, and hypertension, is related to colorectal cancer risk (12). MetS, independent of body mass index (BMI), is associated with an increased risk of colorectal cancer in women (13). Fiber may mitigate risk related to MetS through regulating systemic glucose and insulin responses (14). Calcium supplementation remains a common recommendation for the prevention of colorectal cancer (2, 15) and may interact with supplemental fiber and impact the effect on recurrent adenomas in high-risk individuals (16). In addition, calcium supplementation, while not associated with colorectal cancer risk in the Women's Health Initiative (WHI; ref. 17) has been previously inversely associated with colorectal cancer risk in observational studies and recent meta-analyses (18). Thus, both of these exposures warrant investigation as effect modifiers of any potential relationship between prebiotic supplements and colorectal cancer.

Epidemiologic observational evidence generally supports the potential role for dietary fiber in colorectal cancer risk reduction (19, 20). In a pooled analysis of two clinical trials, high dietary fiber intake in men, but not women, was associated with a lower odds of adenoma recurrence, a precursor for colorectal cancer (21). Recent analyses within WHI indicated a modest trend toward lower colorectal cancer risk with increasing intake of total dietary fiber, as well as soluble and insoluble dietary fiber (22). In addition, higher dietary fiber intake has been associated with higher colorectal cancer survival (23). Yet, this evidence has largely overlooked the role of supplemental prebiotic fiber. Nonetheless, increasing evidence is linking individual prebiotics to the regulation of inflammation and gut microbiota (24), as well as metabolic biomarkers, suggesting potential to favorably modulate colorectal cancer risk and potentially colorectal cancer outcomes (25).

Although efficacy studies remain limited in humans, approximately 2.0%–3.3% of the U.S. adult population report use of fiber-based prebiotic supplements (26). The majority of older adults who take fiber supplements report using the supplement to promote good colon health with many taking such supplements daily for five years or more (27). To date, few epidemiologic studies have investigated use of prebiotic supplements and colorectal cancer risk (28, 29), suggesting the need to more robustly evaluate the role these supplements in cancer outcomes. Despite prevalence data indicating frequent prebiotic use in women, current prebiotic marketing shifts and provider recommendations for selected use of prebiotic supplements (30), there is a distinct gap in the literature related to gender-specific colorectal cancer risk and colorectal cancer–specific mortality and prebiotic use. Our objective was to examine the associations between prebiotic supplement use and colorectal cancer risk as well as colorectal cancer–specific mortality in the Women's Health Initiative cohort. We hypothesized that supplemental prebiotics would be associated with lower risk of colorectal cancer and colorectal cancer–specific mortality in postmenopausal women. We further hypothesized that this protective association would be attenuated in the presence of MetS.

The WHI is a multicenter longitudinal study of postmenopausal women in the United States. Recruitment and enrollment methods have been published previously (31, 32). In brief, 161,808 women were enrolled from 40 clinic sites throughout the United States between 1993 and 1998 and were followed for events including cancer and mortality. There are two major components of the WHI: an observational cohort study (OS, n = 93,676) and three clinical trials (CT, n = 68,132) including randomized trials of hormone therapy, a low-fat diet, and calcium plus vitamin D supplementation. Women in OS and CT components completed comprehensive baseline examinations and were followed over time and were included in this analysis. Women with a history of colorectal cancer at enrollment or who developed colorectal cancer within the first year of follow-up (n = 923) or who were missing follow-up data (n = 690) were excluded from the analysis, leaving 160,195 women in the analytic cohort.

Prebiotic use was captured at baseline on the current medications and dietary supplement forms. During the in-person clinic visit, the following questions were asked on the study questionnaires regarding use of prebiotic fiber supplements: (i) “do you take bulk laxatives or fiber-containing medications?” and (ii) “do you use any medications to help you with digestion?” Women were able to write-in “other” on either form. Prebiotic use was categorized as “user” if they reported use of any medication, digestive aid, or other dietary supplement containing fiber, determined by generic product name (e.g. psyllium, methylcellulose, polycarbophil, pectin, etc.). On the basis of the primary generic ingredient, the prebiotic supplement fiber type was categorized as soluble or insoluble as established in published literature. Absence of any reported use at baseline was categorized as “nonuser.”

WHI participants were followed for outcomes through March 2018. Information on new colorectal cancer diagnoses was collected by self-report annually and adjudicated through medical chart review and centralized review by trained WHI physician adjudicators. Stage data for colorectal cancer diagnoses were matched to the Surveillance, Epidemiology, and End Results Program (SEER) database when available. Mortality and cause of death were determined by medical record and death certificate review (33). The primary outcomes for these analyses were colorectal cancer diagnosis and colorectal cancer–specific mortality. Time to colorectal cancer diagnosis was calculated from the date of study enrollment until recorded date of colorectal cancer diagnosis. Time to colorectal cancer mortality was calculated as the time of study enrollment until the date of recorded death due to colorectal cancer. If a woman did not experience either a colorectal cancer or mortality outcome, her follow-up time was censored as last documented follow-up day.

Other reported medications, including pharmaceutical-based laxatives, nonsteroidal anti-inflammatory drugs, and aspirin, were coded separately. Dietary intake, including total dietary fiber was estimated through the administration of the previously validated WHI food frequency questionnaire for use as a covariate in the model (34). Neighborhood socioeconomic status (NSES; range: 0–100) was calculated as a summary measure of several dimensions of wealth and income using previously described methods (35). Additional information regarding demographics, physical activity, other medication use, medical history, and use of supplemental calcium and/or vitamin D were completed through self-report questionnaires at baseline.

Descriptive statistics of participants' baseline characteristics by current prebiotic users and nonusers were obtained: comparisons of the baseline characteristics between the two groups were conducted using χ² test for categorical variables and t tests for continuous variables. Cox proportional hazards models were used to estimate HRs and 95% confidence intervals (95% CI) for the association of prebiotic use and colorectal cancer risk and separately, colorectal cancer mortality, with and without adjusting for potential confounders. The exposure for each model was overall prebiotic use (user vs. nonuser) and categorical prebiotic supplement fiber type (insoluble vs. soluble). Analyses were performed for colorectal cancer risk, and separate colorectal cancer mortality, within each prebiotic use category.

Likelihood ratio tests were utilized to compare models with and without interaction terms between prebiotic use and other factors associated with colorectal cancer in prior published work (36) including: baseline calcium supplementation, meeting dietary fiber recommendations, history of a colonoscopy, and age to determine potential effect modification. Concurrent use of calcium supplements was evaluated as binary (use and nonuse). Meeting recommendations for dietary fiber intake was categorized into fiber intake <25 g (below recommendations) and ≥25 g (at or above recommendations).

All models were adjusted for demographic and lifestyle characteristics including: study assignment (OS or CT and randomization arm), age, BMI, ethnicity and race, smoking status, alcohol intake, dietary fiber intake, total calcium and vitamin D intake, physical activity, NSES, aspirin use, family history of colorectal cancer, and history of: colonoscopy, hemoccult test, and adenoma removal. Other risk factors including red and processed meat intake, use of nonsteroidal anti-inflammatory medication and other laxatives as well as colorectal cancer stage at diagnosis did not change crude HR estimates >10% and therefore were not included in final adjusted models. The proportional hazards assumption was tested by evaluating the residuals versus time in all models with Schoenfeld residuals and no violations were observed in any model. All statistical analyses were conducted using STATA 15.1 (StataCorp LLC). All tests were two sided and a P value <0.05 was considered statistically significant.

Earlier work from the WHI suggested that metabolic disruption was associated with the development of colorectal cancer (37). To evaluate the association of prebiotic supplementation in presence of MetS on colorectal cancer risk, a sensitivity analysis was completed using adjusted Cox proportional hazard models in a subsample of participants. The presence of MetS was calculated for the sample using available laboratory values collected on 6% of the sample (n = 5,370) at baseline. Clinical MetS was defined using the Third Report of the National Cholesterol Education Program's Adult Treatment Panel (ATP III; ref. 38). Presence of MetS was determined if any individual met any three or more of the following: (i) waist circumference >88.0 cm; (ii) serum triglycerides ≥150 mg/dL; (iii) blood pressure ≥130/85 mm Hg; (iv) HDL cholesterol <50 mg/dL; and (v) serum glucose ≥100 mg/dL.

A total of 3,032 colorectal cancer cases were diagnosed during an average 15.4 years of follow-up. Overall, 35,746 women died during follow up, of which, 841 were due to colorectal cancer. Overall, prebiotic users (n = 5,944, 3.7%) were predominately non-Hispanic whites, nonsmokers, with a normal BMI. The majority of prebiotic supplement used was soluble fiber (85.4% for colorectal cancer cases and 88.2% for noncases). The majority (87.8%) of prebiotic users did not meet recommendations to consume ≥25 g/day of dietary fiber and the majority of prebiotic users had a history of a colonoscopy (72.4%). In comparison, 89.7% of nonusers did not meet dietary fiber recommendations and 45.7% previously had a colonoscopy (Table 1).

Stage data obtained via the SEER database was available for 87.5% (n = 2,653) of colorectal cancer cases (Supplementary Table S1). The majority of women (53.1%; n = 1,408) were diagnosed with stage III–IV colorectal cancer with the remaining women diagnosed at stage I–II colorectal cancer (44.9%, n = 1,192). Of the colorectal cancer cases that used any prebiotic (n = 107), over half of the women who used soluble fiber prebiotic supplements were diagnosed with stage I–II colorectal cancer (51%, n = 48), whereas the majority of women who used insoluble fiber prebiotic supplements were diagnosed with stage III–IV colorectal cancer (57%, n = 8). Among all women, psyllium was the most common prebiotic fiber reported (74% of users), followed by polycarbophil (13%) and methylcellulose (9%; Table 2).

Unadjusted models indicated no statistically significant association between prebiotic supplement use and colorectal cancer risk. Estimates were similar in the adjusted models (HR, 1.13; 95% CI, 0.93–1.39; Table 3). Prebiotic supplement fiber type (none vs. insoluble or soluble) was not associated with colorectal cancer risk. There were no statistically significant interactions between prebiotic use and other colorectal cancer risk factors including calcium supplementation (P = 0.14), meeting dietary fiber recommendations (P = 0.13), previous history of a colonoscopy (P = 0.28) or age (P = 0.16).

There was no overall association between the use of any prebiotic supplements and colorectal cancer mortality in unadjusted and adjusted models (Table 3). In unadjusted models of prebiotic supplement fiber type, insoluble fiber had a significant positive association with colorectal cancer mortality compared with soluble fiber (HR, 2.56; 95%CI: 1.22–5.40, P = 0.013). After adjustment, this relationship remained (HR, 2.79; 95% CI: 1.32–5.90; P = 0.0007). Likelihood ratio tests indicated no statistically significant interactions present between colorectal cancer mortality and prebiotic use and calcium supplementation (P = 0.86), meeting dietary fiber recommendations (P = 0.17), previous history of a colonoscopy (P = 0.39), or age (P = 0.81).

There was no association between prebiotic supplement use and colorectal cancer risk in the subsample of individuals with laboratory data available to evaluate the presence or absence of clinical MetS (Table 4). In women with MetS and available covariate data (n = 796), there was a nonsignificant higher risk for colorectal cancer in prebiotic users (HR, 2.98; 95% CI: 0.73–12.15). Likelihood ratio tests indicated no interaction between colorectal cancer risk and prebiotic use and MetS (P = 0.08).

Overall, when assessing both soluble and insoluble prebiotic use, there was no association with colorectal cancer risk in postmenopausal women within the WHI. Dietary fiber intake, calcium supplementation, previous history of a colonoscopy, age, nor MetS did not significantly modulate the associations observed between prebiotic use and increased colorectal cancer risk or mortality. Importantly, use of insoluble fiber prebiotics compared with soluble fiber prebiotics was strongly associated with colorectal cancer–specific mortality.

Low prevalence of prebiotic use in the WHI cohort may have influenced observed associations, but reported use was comparable to nationwide prevalence (26). At baseline, prebiotics were reported primarily as a “fiber-containing bulk laxatives” within the WHI. This reflects product marketing trends of the era. Recent rebranding and marketing currently characterize these products as prebiotics. While previous estimates have indicated that prebiotic use has remained stable over the past decade (26), recent analyses indicate that the use of prebiotics increased 4-fold from 2007 to 2012 (39). Indications for prebiotic use include the management of digestive symptoms as well as increasing overall dietary fiber intake. Furthermore, gastrointestinal symptom burden may influence use. The most commonly reported prebiotic supplements (psyllium, polycarbophil, and methylcellulose), have previously demonstrated prebiotic activity, evidenced by modulation of the gut microbiota and production of butyrate (40–42). The estimated average daily dose of supplemental fiber from available psyllium labeling instructions in this sample was 5 grams (43), increasing daily total fiber intake to 22 grams in psyllium users. This level of total fiber intake is just below current fiber recommendations (44).

Given these estimates of average exposure in WHI women, we would have expected, contrary to what we found, that supplementation was sufficient to demonstrate a biological effect associated with healthy gut microbiota and potentially a lower colorectal cancer risk and colorectal cancer mortality. However, our data do not provide specifics on regularity of use or indication. The findings presented here could be due to reverse causality wherein women who used prebiotics were already at increased risk for colorectal cancer because of gastrointestinal symptoms that may have indicated the need for a colonoscopy. In fact, women who reported using prebiotics in our sample were more likely to previously have had a colonoscopy or an adenoma; therefore, it is possible that women were taking such prebiotic supplements to manage gastrointestinal symptoms. These same women also may be predisposed to colorectal cancer and subsequently, prone to poorer colorectal cancer outcomes.

The finding related to insoluble fiber prebiotics and colorectal cancer–specific mortality is interesting and begs further investigation. The broader literature-based evidence related specifically to insoluble fiber and colorectal cancer remains inconsistent and is limited for supplemental insoluble prebiotics. Previous evaluation in the Nutrition Examination Survey (NHANES) III showed a protective effect for colorectal cancer mortality comparing high dietary insoluble fiber intake to low intake (45), which is contrary to our findings regarding supplemental insoluble fiber prebiotics. One large wheat bran fiber (an insoluble fiber) supplement trial in older (> 50 years) adults with a previous history of colonic polyps did not show any protective effect from insoluble fiber against recurrent colorectal adenomas (46). Mechanistically, in vitro findings suggest that insoluble fiber supplements may downregulate colorectal cancer–promoting genes and upregulate colorectal cancer–inhibiting genes (47), while other studies indicated that some insoluble fibers may stimulate colonic cell proliferation and carcinogenesis when combined with carcinogen exposure (48). Furthermore, intake of insoluble fiber may impact colorectal cancer outcomes through drastically changing the primary site of microbiota fermentation in the colon, which may shift butyrate concentrations and colonic microbiota composition (49, 50).

Contrary to our hypothesis, use of a prebiotic supplement did not show any statistically significant association with colorectal cancer risk in women with or without MetS. Similarly, calcium supplementation and dietary fiber intake, previous history of a colonoscopy, and age were not effect modifiers in relation to prebiotic use and colorectal cancer risk or mortality.

This study is among the first to provide evidence of supplemental prebiotics and colorectal cancer risk and mortality. The WHI cohort provided rich data on lifestyle, medication use, medical history, and dietary intake to adjust for confounding of known colorectal cancer risk factors. Effect modification of MetS on prebiotic supplement use and colorectal cancer -risk was able to be evaluated with available biomarker data for clinical MetS.

We were not able to further investigate duration of exposure and indication for use due to the limited number of total prebiotic users who were recruited in the mid-1990s. However, dietary supplement use is commonly a habitual regimen where users take their chosen products daily for many years (51). Therefore, using available self-report data from a single time point is a common approach in evaluating association between dietary supplements and colorectal cancer risk (18). The significant finding of insoluble fiber prebiotic supplement use with colorectal cancer mortality is limited in its generalizability as our data were based on nine colorectal cancer deaths among insoluble prebiotic users. Furthermore, we were limited in our analyses regarding the role of MetS on prebiotic supplement use and colorectal cancer mortality due to six colorectal cancer–specific deaths in the MetS subsample of which one reported insoluble prebiotic use. In addition, the available data was limited with consideration to colorectal cancer site and subtype and therefore, we were not able to evaluate this marker of clinical importance in the presented analyses. Women diagnosed with stage III–IV colorectal cancer more commonly used insoluble prebiotic fiber supplements; however, colorectal cancer stage did not change the magnitude of any estimates in any model. Thus, the estimates seen in our analysis may be the result of residual confounding that could not be accounted for with available data. While this association is statistically significant, this may not necessarily be clinically significant.

Common primary treatment for symptoms related to colonic diverticulitis, chronic idiopathic constipation, and irritable bowel syndrome is fiber therapy, either through dietary modifications or prebiotic supplements (52–54). These conditions demonstrate similar symptoms and underlying etiology and are potential precursors to colorectal cancer (55). Practitioners may decide to recommend prebiotic supplements to patients due to the potential digestive benefits and to mediate bowel symptoms (30). Caution about prescribing insoluble fiber prebiotics for colorectal cancer prevention specifically should be exercised given our current findings which will require replication. Our findings do not support use of prebiotic supplements to reduce risk of colorectal cancer or colorectal cancer–specific mortality among postmenopausal women.

No potential conflicts of interest were disclosed.

Conception and design: M.B. Skiba, T.E. Crane, C.A. Thomson

Development of methodology: M.B. Skiba, L.N. Kohler, C.A. Thomson

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M.B. Skiba, L.N. Kohler, T.E. Crane, E.T. Jacobs, A.H. Shadyab, I. Kato, L. Qi

Writing, review, and/or revision of the manuscript: M.B. Skiba, L.N. Kohler, T.E. Crane, E.T. Jacobs, A.H. Shadyab, I. Kato, L. Snetselaar, L. Qi, C.A. Thomson

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): L.N. Kohler

Study supervision: C.A. Thomson

The authors acknowledge the important role the women participating in WHI over the past 20+ years in providing data for this work. The WHI program is funded by the National Heart, Lung, and Blood Institute, NIH, U.S. Department of Health and Human Services through contracts HHSN268201600018C, HHSN268201600001C, HHSN268201600002C, HHSN268201600003C, and HHSN268201600004C. This work was supported by NCI Cancer Center support grant P30 CA023074.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.