Diabetologia

, Volume 48, Issue 6, pp 1051–1054

Alcohol consumption and type 2 diabetes

Meta-analysis of epidemiological studies indicates a U-shaped relationship
Mini Review

Abstract

To clarify the relationship between alcohol consumption and type 2 diabetes we conducted a meta-analysis of published epidemiological studies. Data from 13 cohorts were included in the analysis. The results of these studies are consistent with regard to moderate alcohol consumption, indicating a protective effect in the order of 30% (relative risk [RR]meta=0.72, 95% CI=0.67–0.77). The reduced risk is seen in men as well as in women, although few studies investigated women. No protective effect of high alcohol consumption was seen and one cannot rule out that large intakes of alcohol may increase the risk of type 2 diabetes. Results from published studies suggest a U-shaped relationship between alcohol and type 2 diabetes, but this is based on rather few studies with heterogeneous design and definitions. It seems important to further investigate if, and to what extent, high alcohol consumption increases the risk of type 2 diabetes. Aspects of moderate alcohol consumption also need further investigation; these include type of drink, frequency of drinking, sex and ethnic differences.

Keywords

Alcohol Epidemiology Meta-analysis Review Type 2 diabetes 

Abbreviations

RR

relative risk

Introduction

It is becoming increasingly clear that alcohol consumption is a factor of importance for the risk of type 2 diabetes. Data from epidemiological studies suggest a protective effect of moderate alcohol consumption (reviewed in [1]). However, the strength of these data with regard to different aspects of alcohol exposure has, to our knowledge, not been evaluated in detail. Important aspects include the influence of high versus moderate alcohol consumption, different susceptibilities for men and women and the magnitude of any protective effect of alcohol consumption. The purpose of this commentary is to provide a better understanding of the epidemiological literature, including the aspects mentioned, by using meta-analysis.

Material and methods

By Medline search we identified 25 epidemiological studies on the relationship between alcohol and type 2 diabetes (search strategy and keywords available from authors) [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26]. These were published between 1988 and March 2004. All studies except three [5, 6, 14] were prospective cohort studies with a follow-up time of up to 20 years [24]. The majority of studies concern men, but some included women [2, 4, 5, 7, 8, 11, 16, 17, 19, 20, 22, 24, 25, 26]. Nine studies were not included in the meta-analyses since the authors analysed alcohol consumption continuously [5, 8] or dichotomised [3, 11, 19], did not provide confidence intervals [4] or analysed men and women together [7, 20, 22] (complete information on the strategies behind the pooling is available from the authors). Furthermore, the British Regional Heart study [9, 21], the US Health Professionals Follow-up Study [10, 18] and the Nurses Health Study [2, 17] were reported twice. Of these papers, only the more recent of the two were included in the meta-analysis [17, 18, 21]. The meta-analysis was performed with Episheet [27] and the pooled estimate was calculated using methods described by Fleiss [28].

Results

Moderate alcohol consumption

Moderate alcohol intake, corresponding to about 5–30 g of alcohol per day, was associated with a reduced risk of type 2 diabetes in the order of 30% (relative risk [RR]meta=0.72, 95% CI=0.67–0.77) compared to low consumption or abstention (Fig. 1). The magnitude of this risk reduction corresponds to the reduced risk of cardiovascular disease reported in moderate drinkers (reviewed in [29]). There are few studies in women, but available data seem to indicate that moderate alcohol consumption is equally beneficial in women as in men (RRmeta in women=0.68, 95% CI=0. 61–0.75). Most of these studies were based on Caucasians, but the results were consistent across other ethnic groups such as Japanese [6, 12, 23], African-Americans [16] and Nauruans [5]. Similar results were seen in most [3, 5, 7, 8, 11, 19, 20, 22], but not all [4] of the studies excluded from the meta-analysis due to differences in methodology.
Fig. 1

Moderate alcohol consumption and type 2 diabetes. Moderate consumption is defined as approximately 5–30 g of alcohol per day. This corresponds to about 0.5–2.5 drinks per day. Abstainers (Todoroki, Tsumura, Ajani, Wei, Wannamethee 2003, Saremi, Hu) or low consumers (approximately <5 g of alcohol per day) (Carlsson 2000, Conigrave, Kao, Nakanishi, Wannamethee 2002, Carlsson 2003) are used as reference category

Moderate consumers of alcohol have a reduced risk compared to low consumers [14, 16, 18, 21, 23, 24] and to abstainers [6, 13, 15, 17, 25]. Thus it does not seem likely that inclusion of people who did not drink because of disease or previous high alcohol consumption as reference can explain a protective effect in moderate consumers. Moreover, the reduced risk persisted after adjustment for a number of potential confounders such as age, body mass index, smoking, physical activity, family history of diabetes, socioeconomic group and diet (see for example [18] or [21]).

High alcohol consumption

The relationship between high alcohol consumption (corresponding to about 30 g of alcohol per day or more) and type 2 diabetes is inconsistent across studies (Fig. 2). In all studies except two [6, 18], high consumers have an increased risk of type 2 diabetes compared to moderate consumers, but whether they have an increased risk compared to low consumers or abstainers is less clear. This was suggested in two studies [14, 16] but not confirmed by others. Notably, in women an increased risk of diabetes has been found in subgroups of high consumers such as lean women, (RR=2.8, 95% CI=1.1–7.3) [24], and women with high consumption of spirits (RR=2.5, 95% CI=1.0–6.2) [25].
Fig. 2

High alcohol consumption and type 2 diabetes. The definition of high consumption varies from ≥20 g of alcohol per day [Carlsson 2003 (in women)] to >69 g of alcohol per day (Nakanishi), but in the majority of studies a cut-off of ≥30 g per day (corresponding to about three drinks per day or more) was used. Abstainers (Todoroki, Tsumura, Wei, Wannamethee 2003) or low consumers (approximately <5 g of alcohol per day) (Carlsson 2000, Conigrave, Kao, Nakanishi, Wannamethee 2002, Carlsson 2003) are used as reference category

Discussion

Since the effects of long-term alcohol intake can hardly be investigated in an experimental setting, research largely has to rely on epidemiological methods. Two major methodological problems are evident in all epidemiological studies on alcohol and type 2 diabetes: how to obtain accurate exposure information and how to identify all cases of diabetes. Bias resulting from misclassification of exposure and/or disease may be one reason for the inconsistent results on high alcohol consumption and type 2 diabetes. Questionnaire information on alcohol was used in all but one of the studies [4] we investigated. Such information is often under-reported [30] and as a consequence, some high consumers may have been classified as moderate or low consumers. This will not produce a spurious risk reduction in moderate consumers, but rather it may make it difficult to observe a relationship between high alcohol intake and type 2 diabetes. In addition, under-reporting implies that the cut-off levels used to define low, moderate and high alcohol intake cannot be interpreted as a measure of absolute consumption. With regard to identification of cases, self-reported diabetes information is often used in epidemiological studies and will be sufficient for many purposes. However, in studies of alcohol, such information may introduce bias since the likelihood of having your diabetes diagnosed may depend on your level of alcohol consumption. Some of the larger studies included in our meta-analysis were based on self-reported diabetes and the results may hence be afflicted with such bias [13, 17, 18, 21, 25]. Another potential reason for the inconsistent results on high alcohol intake and diabetes could be chance, since the number of cases among high consumers of alcohol was very small in most studies.

The results of any meta-analysis are limited since they involve pooling of risk estimates from studies that may have more-or-less-pronounced differences in methodology and, probably, setting. Some of the differences that need to be kept in mind while considering the pooled risk estimates include (1) different reference categories were used to calculate the relative risks, some have used abstainers for comparison whereas other have used low consumers; (2) the definition of low, moderate and high alcohol consumption varies between the studies; (3) the relative risk estimates are not adjusted for the same confounders; (4) different questions are used to measure alcohol intake; and (5) it was not possible to include all studies of alcohol and diabetes in the pooled analyses, due to differences in methodology and publication bias.

Conclusions

The results of published studies on alcohol consumption and type 2 diabetes consistently indicate that moderate alcohol consumption reduces the risk of type 2 diabetes. The risk reduction is in the order of 30%. This is comparable to what has been reported for cardiovascular disease [29]. The reduced risk is seen both in men and in women, although it should be noted that few studies investigated this issue in women. From a mechanistic point of view, a protective effect of moderate alcohol consumption is compatible with findings that alcohol can enhance insulin sensitivity [31, 32].

As to the effect of high alcohol intake on the risk of type 2 diabetes, it is difficult to draw any conclusions. The most consistent finding is that no beneficial effect can be attributed to high alcohol consumption. Whether high alcohol intake actually increases the risk of type 2 diabetes is unclear. Methodological problems such as difficulties in measuring alcohol consumption, underdiagnosing of diabetes cases and small numbers may contribute to partially inconsistent results.

Should one then recommend alcohol to the general population in order to reduce the risk of type 2 diabetes? It seems premature to formulate any such recommendations. Further investigations are needed to decide to what extent high alcohol consumption may increase the risk of diabetes and what moderate and high alcohol consumption means in absolute terms. The possible protective effect of moderate alcohol consumption also needs further attention, considering influence from type of beverage, frequency of drinking and interaction with other risk factors such as overweight and family history of diabetes. More studies in women and different ethnic groups are also warranted.

Notes

Acknowledgements

The authors wish to thank their colleagues A. Ahlbom and T. Andersson for valuable advice regarding the meta-analysis.

References

  1. 1.
    Howard AA, Arnsten JH, Gourevitch MN (2004) Effect of alcohol consumption on diabetes mellitus: a systematic review. Ann Intern Med 140:211–219PubMedGoogle Scholar
  2. 2.
    Stampfer MJ, Colditz GA, WiIlett WC et al (1988) A prospective study of alcohol drinking and risk of diabetes in women. Am J Epidemiol 128:549–558PubMedGoogle Scholar
  3. 3.
    Feskens EJ, Kromhout D (1989) Cardiovascular risk factors and the 25-year incidence of diabetes mellitus in middle-aged men. The Zutphen Study. Am J Epidemiol 130:1101–1108Google Scholar
  4. 4.
    Holbrook TL, Barrett-Connor E, Wingard DL (1990) A prospective population-based study of alcohol use and non-insulin-dependent diabetes mellitus. Am J Epidemiol 132:902–909PubMedGoogle Scholar
  5. 5.
    Hodge AM, Dowse GK, Collins VR, Zimmet PZ (1993) Abnormal glucose tolerance and alcohol consumption in three populations at high risk of non-insulin-dependent diabetes mellitus. Am J Epidemiol 137:178–189PubMedGoogle Scholar
  6. 6.
    Todoroki I, Shinchi K, Kono S, Imanishi K (1994) Lifestyle and glucose tolerance: a cross-sectional study of Japanese men. Ann Epidemiol 4:363–368PubMedGoogle Scholar
  7. 7.
    Gurwitz JH, Field TS, Glynn RJ et al (1994) Risk factors for non-insulin-dependent diabetes mellitus requiring treatment in the elderly. J Am Geriatr Soc 42:1235–1240PubMedGoogle Scholar
  8. 8.
    Monterossa AE, Haffner SM, Stern MP, Hazuda HP (1995) Sex difference in lifestyle factors predictive of diabetes in Mexican-Americans. Diabetes Care 18:448–456PubMedGoogle Scholar
  9. 9.
    Perry IJ, Wannamethee SG, Walker MK, Thomson AG, Whincup PH, Shaper AG (1995) Prospective study of risk factors for development of non-insulin dependent diabetes in middle aged British men. BMJ 310:560–564PubMedGoogle Scholar
  10. 10.
    Rimm EB, Chan J, Stampfer MJ, Colditz GA, Willett WC (1995) Prospective study of cigarette smoking, alcohol use, and the risk of diabetes in men. BMJ 310:555–559PubMedGoogle Scholar
  11. 11.
    Sugimori H, Miyakawa M, Yoshida K et al (1998) Health risk assessment for diabetes mellitus based on longitudinal analysis of MHTS database. J Med Syst 22:27–32CrossRefPubMedGoogle Scholar
  12. 12.
    Tsmura K, Hayashi T, Suematsu C, Endo G, Fujii S, Okada K (1999) Daily alcohol consumption and the risk of type 2 diabetes in Japanese men. Diabetes Care 22:1432–1437PubMedGoogle Scholar
  13. 13.
    Ajani UA, Hennekens CH, Spelsberg A, Manson JE (2000) Alcohol consumption and risk of type 2 diabetes mellitus among US male physicians. Arch Intern Med 7:1025–1030CrossRefGoogle Scholar
  14. 14.
    Carlsson S, Hammar N, Persson P-G, Efendic S, Östenson C-G, Grill V (2000) Alcohol consumption, type 2 diabetes and impaired glucose tolerance in middle-aged Swedish men. Diabet Med 17:776–781CrossRefPubMedGoogle Scholar
  15. 15.
    Wei M, Gibbons LW, Mitchell TL, Kampert JB, Blair SN (2000) Alcohol intake and incidence of type 2 diabetes in men. Diabetes Care 1:18–22Google Scholar
  16. 16.
    Kao LHW, Puddy IB, Boland LL, Watson RL, Brancati FL (2001) Alcohol consumption and the risk of type 2 diabetes mellitus. Am J Epidemiol 154:748–757CrossRefPubMedGoogle Scholar
  17. 17.
    Hu FB, Manson JE, Stampfer MJ et al (2001) Diet, lifestyle, and the risk of type 2 diabetes mellitus in women. N Engl J Med 345:790–797CrossRefPubMedGoogle Scholar
  18. 18.
    Conigrave KM, Hu B F, Camargo CA, Stampfer MJ, Willett WC, Rimm EB (2001) A prospective study of drinking patterns in relation to risk of type 2 diabetes among men. Diabetes 50:2390–2395PubMedGoogle Scholar
  19. 19.
    Watanabe M, Barzi F, Neal B et al (2002) Alcohol consumption and the risk of diabetes by body mass index levels in a cohort of 5,636 Japanese. Diabetes Res Clin Pract 57:191–197CrossRefPubMedGoogle Scholar
  20. 20.
    de Vegt F, Dekker JM, Groeneveld WJ et al (2002) Moderate alcohol consumption is associated with lower risk for incident diabetes and mortality: the Hoorn Study. Diabetes Res Clin Pract 57:53–60CrossRefPubMedGoogle Scholar
  21. 21.
    Wannamethee SG, Shaper AG, Perry IJ, Alberti KG (2002) Alcohol consumption and the incidence of type II diabetes. J Epidemiol Community Health 7:542–548CrossRefGoogle Scholar
  22. 22.
    Lu W, Jablonski KA, Resnick HE et al (2003) Alcohol intake and glycemia in American Indians: the Strong Heart Study. Metabolism 52:129–135CrossRefPubMedGoogle Scholar
  23. 23.
    Nakanishi N, Suzuki K, Tatara K (2003) Alcohol consumption and risk for development of impaired fasting glucose or type 2 diabetes in middle-aged Japanese men. Diabetes Care 26:48–54PubMedGoogle Scholar
  24. 24.
    Carlsson S, Hammar N, Grill V, Kaprio J (2003) Alcohol consumption and the incidence of type 2 diabetes: a 20-year follow-up of the Finnish twin cohort study. Diabetes Care 26:2785–2790PubMedGoogle Scholar
  25. 25.
    Wannamethee SG, Camargo CA Jr, Manson JE, Willett WC, Rimm EB (2003) Alcohol drinking patterns and risk of type 2 diabetes mellitus among younger women. Arch Intern Med 163:1329–1336CrossRefPubMedGoogle Scholar
  26. 26.
    Saremi A, Hanson RL, Tulloch-Reid M, Williams DE, Knowler WC (2004) Alcohol consumption predicts hypertension but not diabetes. J Stud Alcohol 65:184–190PubMedGoogle Scholar
  27. 27.
  28. 28.
    Fleiss JL (1993) The statistical basis of meta-analysis. Stat Methods Med Res 2:121–145PubMedGoogle Scholar
  29. 29.
    Fagrell B, De Faire U, Bondy S et al (1999) The effects of light to moderate drinking on cardiovascular diseases. J Intern Med 246:331–340CrossRefPubMedGoogle Scholar
  30. 30.
    Feunekes GIJ, Van’t Veer P, van Staveren WA, Kok FJ (1999) Alcohol intake assessment: the sober facts. Am J Epidemiol 150:105–112PubMedGoogle Scholar
  31. 31.
    Facchini FS, Chen Y-DI, Reaven GM (1994) Light-to-moderate alcohol intake is associated with enhanced insulin sensitivity. Diabetes Care 17:115–119PubMedGoogle Scholar
  32. 32.
    Mayer EJ, Quesenberry CP Jr, Friedman GD, Selby JV (1993) Alcohol consumption and insulin concentrations: role for insulin in associations of alcohol intake with high-density lipoprotein cholesterol and triglycerides. Circulation 88:219Google Scholar

Copyright information

© Springer-Verlag 2005

Authors and Affiliations

  1. 1.Department of Epidemiology, Institute of Environmental MedicineKarolinska InstitutetStockholmSweden
  2. 2.Department of Molecular MedicineKarolinska InstitutetStockholmSweden
  3. 3.Division of Epidemiology, Stockholm Centre of Public HealthKarolinska University HospitalStockholmSweden

Personalised recommendations