Aggregate Level Association between Alcohol and the Peptic Ulcer Mortality Rate
Razvodovsky, Yuri E., Alcoholism and Psychiatry Research
Peptic ulcer disease (PUD) is one of the common disorders affecting the digestive system. The lifetime risk for peptic ulcer is 5-10% in developed countries.1 The pathology of peptic ulcer is complex and may involve the overproduction of acid or pepsin, inadequate mucosal defense, reflux of bile and pancreatic juice into stomach.2 Peptic ulcer is a classical example of the biopsychosocial model of disease. Although Helicobacter pylori is considered causal for peptic ulcer, a variety of host factors besides infection may play an important contributing role for this disease: smoking, alcohol consumption, non-steroid anti-inflammatory drug use and psychosocial stress.3
Studies on the relationship between alcohol consumption and risk for PUD, that have been published during the last decade have reported the conflicting results. There is a large body of clinical and research evidence that the excessive alcohol consumption increases the risk of PUD by impairing the mucosal defense mechanisms.4'5 In one longitudinal prospective study, it was shown that, when compared with those who consumed less than 1.5 units of alcohol per week, the individuals who consumed 63 units of alcohol per week are 4 times more likely to develop a bleeding ulcer.6 On the other hand, several large-scale prospective studies have suggested a protective effect of moderate alcohol consumption (one or two drink a day) on the development of gastric ulcer.3·7 In the prospective cohort study, a tendency towards an increase in the PUD incidence proportion with the number of consumed drinks was observed resulting in a U-shaped relationship.8
Several researchers had reported a protective effect of wine and beer consumption against the active Helicobacter pylori infection.9,7 There are several possible biological mechanisms that could explain the protective effects of moderate alcohol consumption: 1) moderate alcohol intake might invigorate the mucosal defense by its effects on prostaglandins; 2) the increased gastrin secretion and the resulting rise in gastric acid production enhance the antibacterial activity; 3) wine has a strong antibacterial effect.9·10 There is also evidence of a beverage-specific effect of alcohol on PUD risk: wine drinking showed a protective effect against ulceration, whereas intake of spirits increased ulcer risk.8 A study that addressed the role of drinking pattern in relation to risk of PUD suggests that the binge drinking substantially increases the risk. For example, the Danish study provides evidence that wine drinking is associated with a much lower risk than beer drinking.6 This might reflect the fact that the beer drinkers are more likely to take alcohol in binge session than wine drinkers. Thus, the research evidence suggests that alcohol consumption may be related to positive or negative risk factors of PUD, depending on dose and type of drinking. In line with these findings, we assume that the combination of the higher level of alcohol consumption per capita and binge drinking strong spirits would result in close association between alcohol and PUD at the aggregate level. To test this hypothesis, trends in PUD mortality and beverage-specific level of alcohol sale per capita in Belarus from 1970 to 2005 were analyzed, applying the time series analyses.
All data provided in the article were taken from the Ministry of Statistics of Belarus annual report for the period from 1970 to 2005. The statistical analysis (correlation, regression, ARIMA, distributed lags analysis) was dose using the package "Statistics 7". It is generally agreed that the bivariate correlations between two row time-series are spurious due to common sources of trends and autocorrelation.11 Therefore, in order to reduce the risk of obtaining a spurious relation between two variables that have common trends, the trends should be removed by means of a differentiating procedure: ∇x^sub t^= x^sub t^- x^sub t-1^ This means analyzing the annual changes rather that row data. …