Critique 115: Potential association of alcohol consumption with diverticulosis — 25 June 2013
Sharara AI, El-Halabi MM, Mansour NM, Malli A, Ghaith OA, Hashash JG, Maasri K, Soweid A, Barada K, Mourad FH, El Zahabi L. Alcohol consumption is a risk factor for colonic diverticulosis. J Clin Gastroenterol 2013;47:420–425.
Background and Aim: The exact factors predisposing to colonic diverticulosis other than age are unknown.
Methods: Cross-sectional study of asymptomatic subjects undergoing screening colonoscopy. A detailed dietary and social questionnaire was completed on all participants. A worldwide review of the literature was performed to further investigate any association between identified risk factors and diverticulosis.
Results: Seven hundred forty-six consecutive individuals were enrolled (mean age, 61.1±8.3 y; female: male=0.98). Overall, the prevalence of diverticulosis was 32.8% (95% CI, 29.5-36.2). Diverticula were left-sided, right-sided, or both in 71.5%, 5.8%, and 22.7% of affected subjects, respectively. On univariate analysis, age, sex, adenomatous polyps, advanced neoplasia (adenoma ≥1 cm, villous histology, or cancer), aspirin, and alcohol use were significantly associated with diverticulosis. Diet, body mass index, physical activity, and bowel habits were not associated with the disease. On multivariate analysis, increasing age (P<0.001), advanced neoplasia (P=0.021), and alcohol consumption (P<0.001) were significantly associated with diverticulosis. The adjusted odds ratio for diverticulosis in alcohol users was 1.91 (1.36 to 2.69), with increasing prevalence with higher alcohol consumption (P-value for trend=0.001). When the prevalence of diverticulosis reported from 18 countries was analyzed against alcohol use, there was a strong correlation with national per-capita alcohol consumption rates (Pearson correlation coefficient r=0.68; P=0.002).
Conclusions: Alcohol use is a significant risk factor for colonic diverticulosis and may offer a partial explanation for the existing East-West paradox in disease prevalence and phenotype. Further studies are needed to investigate this association and its putative pathophysiological mechanisms.
The etiology of diverticulosis of the colon is poorly understood. Many, but not all, studies suggest that low fiber intake and obesity increase the risk. The large Health Professional Follow-up Study of almost 50,000 subjects found a lower risk to be associated with increased physical exercise and the intake of soluble fiber, and no increase from alcohol consumption.
This paper tests the cross-sectional association of alcohol consumption with diverticulosis among 746 asymptomatic consecutive subjects undergoing screening colonoscopies in Lebanon. In a pre-screening questionnaire, subjects reported the intake of several dietary constituents, as well as alcohol, with about one-half of subjects being lifetime abstainers. Most of the “drinkers” stated that they consumed alcohol occasionally or < 1 drink/day; only about 10% reported one or more drinks/day. No data on type of beverage or pattern of drinking were available. From their cross-sectional analyses, the authors conclude that alcohol consumption raises the risk of diverticulosis.
Specific comments on paper by Forum members: This study found a fairly high prevalence of diverticulosis (32.8%) among their subjects, a rate that is similar to many studies in Europe and North America, but higher than the prevalence found in Africa and Asia. There were some deficiencies in the study, such as a relatively small number of subjects, with low prevalence of drinking more than occasionally or < 1 drink/day; this limits its applicability to most western populations. Further, the investigators had very limited data on diet, and no specific data on fiber intake; it was a cross-sectional comparison between alcohol and diverticulosis, so causation cannot be determined.
Reviewer Ellison noted that there were no data on the ethnicity of subjects: “Given that Lebanon has a mixture of ethnic groups (including Muslims, who generally do not drink), it would have been interesting to see a break-down of effects by such groups; diet and other environmental factors, as well as alcohol intake, surely varied by ethnic group. Also, the authors’ use of an inter-country comparison of per-capita alcohol intake and reported diverticulosis has no relevance: there are large cultural differences between nations with high and those with low per-capita alcohol intake, and alcohol use may well have been just a marker for a western-type culture.”
Despite what the authors state, no dose-dependent effect is seen in the multivariate analysis (OR 1.96 for occasional/< 1 drink/day, 1.91 for ≥ 1 drink/day). Further, no description is given for the distribution of alcohol intake within the two drinker categories, so the effects of the pattern of drinking or binge drinking, both of which strongly affect health, cannot be estimated.
With a high percentage of abstainers in this population, the lack of data on the ethnicity of subjects, and no data on drinking patterns or type of alcohol, Forum members considered that this paper should be interpreted as raising a question about alcohol and diverticulosis, but certainly provides no answers. Most Forum members suggested that the association reported could well relate to uncontrolled confounding by other lifestyle factors.
Potential mechanisms by which alcohol could affect risk of diverticulosis: Forum member Skovenborg commented on potential mechanisms by which alcohol might relate to diverticulosis. “The authors quoted an early study from Salt Lake City to the effect that alcohol inhibits recto-sigmoid motility in humans (Berenson MM, Avner DL. Alcohol inhibition of rectosigmoid motility in humans. Digestion 1981;22:210–215). However, the clinical relevance of an acute intravenous alcohol infusion with a blood alcohol level of 140 mg/100 ml is questionable for a study where the association of alcohol consumption and colonic diverticulosis was similar in people with occasional alcohol use and people drinking ≥ 1 drink per day.”
Skovenborg continued: “An Italian study showed a significant prolongation of the oro-cecal transit time, both in patients with alcoholism and in social drinkers, as compared with teetotaler subjects (Addolorato G et al. Influence of alcohol on gastrointestinal motility: lactulose breath hydrogen testing in orocecal transit time in chronic alcoholics, social drinkers and teetotaler subjects. Hepatogastroenterology 1997;44:1076-1081). In contrast, another Italian study found no statistically significant difference of the oro-cecal transit time between moderate alcohol drinkers and teetotallers (Papa A et al. Effect of moderate and heavy alcohol consumption on intestinal transit time. Panminerva Medica 1998;40:183-185). According to a German study, all alcoholic beverages are not equal regarding their effect on gastrointestinal motility. Pfeiffer et al investigated the comparative effect of beer (7.0% v/v), with wine (7.5% v/v), ethanol (7.5% v/v), and water on the gastric emptying of a liquid test meal and on the gastro-cecal transit time of lactulose added to the test meal. The gastro-cecal transit time was significantly shorter when the liquid meal was administered with beer or wine compared with ethanol and water (Pfeiffer A, Kaess H. Effect of ethanol and commonly ingested alcoholic beverages on gastric emptying and gastrointestinal transit. Clin Investig 1992;70:487-491).” All of these factors illustrate the lack of a plausible pathophysiological mechanism for the reported association between and diverticulosis, and raise the question of confounding by other factors.
Reviewer Ursini questioned how many of the Hill criteria of causality are met. “When does an association become a cause? (I’ll never forget the extremely low P value for the relation of the number of storks in the Alps to birth rate in Germany!)” He added that the strength of the association, biological gradient, consistency with other studies, temporality, and several other of the Hill criteria are not met in determining causality in these analyses. Stated Forum member Thelle, “The association observed in a cross- sectional study cannot be considered a direct cause — at the best, it can be used to generate a hypothesis.” Adds Forum member Goldfinger: “An interesting association, but in and of itself, lackluster with respect to significant conclusions. In the absence of robust plausible biologic mechanisms, any suggestion that there may be a causal relationship will require more work.”
The etiology of diverticulosis of the colon is poorly understood. Many, but not all, studies suggest that low fiber intake and obesity increase the risk; few studies have evaluated the relation of alcohol consumption and diverticulosis, with inconclusive results. This paper tests the cross-sectional association of alcohol consumption with diverticulosis among 746 asymptomatic consecutive subjects undergoing screening colonoscopies in Lebanon.
One-half of the subjects in this study were abstainers. Most of the “drinkers” stated that they consumed alcohol occasionally or < 1 drink/day; only about 10% reported 1 or more drinks/day, and no data on type of beverage or pattern of drinking were available. Nevertheless, from their analyses the authors conclude that alcohol consumption raises the risk of diverticulosis.
Forum members considered that this cross-sectional analysis provides little support for a causative effect of alcohol on the risk of diverticulosis. Most larger observational studies have not shown such an effect. In the present study, about 90% of subjects were either abstainers or consumed alcohol only occasionally or averaged < 1 drink/day. Further, the investigators did not evaluate the pattern of drinking (including binge drinking) and the results do not support an increase in risk with greater drinking (lack of a dose-response relation). Mechanisms for such an effect are largely unknown. Thus, few of the Hill criteria for “causality” are met.
The authors’ use of an inter-country comparison of per-capita alcohol intake and reported diverticulosis has no relevance: there are large cultural differences between nations with high and those with low per-capita alcohol intake, and alcohol use may well have been just a marker for a western-type culture. While the present study could be used to generate an hypothesis of a relation of alcohol intake to diverticulosis, attempts to validate such an association will require much additional research.
* * *
Comments on this critique by the International Scientific Forum on Alcohol Research were provided by the following members:
Fulvio Ursini, MD, Dept. of Biological Chemistry, University of Padova, Padova, Italy
R. Curtis Ellison, MD, Section of Preventive Medicine & Epidemiology, Boston University School of Medicine, Boston, MA, USA
Tedd Goldfinger, DO, FACC, Desert Cardiology of Tucson Heart Center, Dept. of Cardiology, University of Arizona School of Medicine, Tucson, Arizona, USA
Erik Skovenborg, MD, Scandinavian Medical Alcohol Board, Practitioner, Aarhus, Denmark
Harvey Finkel, MD, Hematology/Oncology, Boston University Medical Center, Boston, MA, USA
Dag S. Thelle, MD, PhD, Senior Professor of Cardiovascular Epidemiology and Prevention, University of Gothenburg, Sweden; Senior Professor of Quantitative Medicine at the University of Oslo, Norway
Andrew L. Waterhouse, PhD, Marvin Sands Professor, Department of Viticulture and Enology, University of California, Davis; Davis, CA, USA
Arne Svilaas, MD, PhD, general practice and lipidology, Oslo University Hospital, Oslo, Norway