High intake of dietary fibre from fruit and vegetables reduces the risk of hospitalisation for diverticular disease

This major prospective cohort study among middle-aged men and women demonstrates that high intake of dietary fibre, especially from fruits and vegetables, reduces the risk of hospitalisation for diverticular disease.

In general, fruits and vegetables contain higher level of cellulose than cereals. As an insoluble fibre, cellulose represents an average of 30% in fruits and 50% in vegetables. Insoluble fibres are metabolised by colonic bacteria less than (water-) soluble fibres. Insoluble fibres can increase faecal output by acting as a fermenter, by stimulating microbial growth resulting in the production of short chain fatty acids (SCFA). SCFA are known as an important fuel source for the colon in general and particularly in sigmoid colon [27]. This might partially explain the observed beneficial effect of fibre from fruits and vegetables in our study.

According to the results in the present study, the difference between women eating the highest and lowest amount of fibres from fruit and vegetables, was 8.5 g/day (12.6 − 4.1 g/day). This is how much extra fibre one must eat every day to achieve a risk reduction of 30%. For comparison, an ordinary sized apple or orange contains 3–3.5 g of fibre. For men, the corresponding difference was 7.4 g/day to achieve a risk reduction of 35%.

When discussing the role of fibres in preventing diverticular disease, it is important to define where in the progress of the disease an increased fibre intake might be beneficial. Risk factors for developing asymptomatic findings of diverticulosis might differ from risk factors affecting symptomatic disease. Firstly, Painter and Burkitt [8] and later others [28] found that a low fibre intake could promote the formation of colonic diverticulosis when comparing high fibre-consuming native Africans with populations in the Western world with a low fibre intake.

In contrast, Peery et al. [9] recently postulated that a high fibre diet could promote diverticulosis in a study of 2104 patients undergoing an extensive diet interview using validated food frequency questionnaires, 3 months after a colonoscopy-proven diagnosis of asymptomatic diverticulosis had been established and communicated to the patients.

Secondly, it has been shown that a high fibre intake could prevent symptomatic diverticular disease, mainly diverticulitis. Aldoori et al. found that the insoluble component of dietary fibre was inversely associated with risk of diverticular disease among male health professionals in the USA [6]. This finding is contradictory to the present study in which fibres from fruits and vegetables were shown to be most beneficial. Yet, the outcome of the study of Aldoori differed because the present study included only hospitalisations for DD, compared to self-reported symptomatic DD in the study of Aldoori et al. This might partly explain the diverging results.

The results in the present study confirm the recently reported findings by Crowe et al. [17], though they also reported that fruit could be a more beneficial fibre contributor compared to vegetables. Our data did not allow us to examine the effect of fruits and vegetables separately.

Thirdly, studies have also addressed the potential benefit from high fibre intake to prevent recurrences after an episode of medically managed diverticulitis. Brodribb et al. [29] found a reduction of symptoms after 3 months with fibre supplement in a placebo-controlled trial. However, other studies have been contradictory [9].

Other significant risk factors for symptomatic DD in the diet, high intake of red meat and fat, have been postulated, though the relationships are weak [11]. Also, previous data have stated smoking, low physical activity and obesity to increase the risk for DD [12, 14]. Nuts, seeds and corn do not seem to affect the risk [30]. Recently, two studies of mono- and dizygotic twins have found a hereditary component in the development of DD, with a genetic effect up to 40% [3, 4]. Medications might also affect the incidence of DD. Use of corticosteroids, both inhaled and orally, increase the risk of hospital admission because of DD, according to a recent report, and we therefore included use in the multivariate analysis [13]. Yet, the effect of non-steroid anti-inflammatory drugs (NSAIDs) and acetylsalicylic acid (ASA) is contradictory. A previous study in the SMC cohort could not explore any effect of these medications on the incidence of DD and was therefore omitted in our analysis [13].

The strengths of the present study are the large numbers in the cohorts, independent prospective data ascertainment, homogenous classification of exposures and outcomes. Potential confounders such as smoking, BMI, physical activity, use of steroid medication and co-morbidity were included in the multivariate analysis and almost complete follow-up was achieved through national registers.

There are limitations of our study, since there is information bias in our questionnaires and low response rate of answering among the men in the COSM cohort. Diet intake was assessed only at baseline and may not reflect the actual intake during the whole study period. Potential confounders such as smoking, BMI, educational level and physical activity were included in the multivariate analysis, but there is a possibility that unmeasured or residual confounders may have biased our findings. Also, disease classification could have been flawed. The final diagnosis was made by a clinician and stated in the discharge summary, and then subsequently reported to the National Patient Register. In the present study, men and women with mild symptoms were likely missed, [13], as only those with symptomatic disease requiring in-hospital care were included. Diverticular bleeding has no separate code in ICD 10 and such events are therefore most likely missed. Moreover, fluid intake might be a potential confounder. Unfortunately, our collected data regarding fluid intake were not complete. Water was available as a calculated variable from all dietary questions in the survey. However, there were no data of how much water they drank. Therefore, we could not control for this in the analysis.

For validation purposes, 528 consecutive patients discharged (at least once) from Danderyd Hospital in Stockholm, Sweden, with a primary discharge code of K572-9 were re-evaluated [12]. In all, 95.8% were correctly classified as having symptomatic diverticular disease and misclassification was found in 4.2% only. This is in line with a previous Swedish report [31].

To conclude, high intake of fruits and vegetables may reduce the risk of hospitalisation due to DD in this major cohort study of both men and women, while intake of cereals does not influence the risk.