Critique 102: Possible prevention of myocardial infarction from alcohol’s effects on periodontal disease: An hypothesis — 6 February 2013

Håheim LL, Olsen I, Rønningen KS.  Oral infection, regular alcohol drinking pattern, and myocardial infarction.  Medical Hypotheses 2012;79:725–730.

Authors’ Abstract

Oral infections have been associated with an increased risk for myocardial infarction (MI) and other cardiovascular diseases (CVD).  Conversely, low, regular alcohol consumption is associated with a lower association of CVD.  The objective was to test the novel hypothesis that oral infections are modified by regular alcohol drinking which has the effect of lowering the incidence of MI’s.  

The effect has been observed where tooth extractions where carried out due to infections and compared with extractions unconnected to infections.  Oral infections and in particular periodontal infections impose an infectious load on the health in many people. In its advanced forms (periodontal pockets ≥ 6 mm), periodontitis affects ~ 10– 15% of adults.  The infection runs a chronic course with exacerbations.  The bacteria cause local infection destructive to the supporting tissues of the teeth and have been detected in systemic diseases through bacterial products and bacteria entering the circulation. The often persistent, long term history of chronic periodontal infection in individuals is a challenge to the immune system.  Over 700 oral bacteria and other microorganisms have been identified, many of which are virulent.  Control of the level of oral microbiota is through well known oral hygiene measures.  

Alcohol by being bactericidal is a factor that may reduce the bacterial level in the oral cavity.  If this effect truly exists, it should be observed through reduction of infections in the mouth.  Tooth extraction is the ultimate consequence of periodontal and dental infections and a reduction of tooth extraction due to infections should therefore be observed. The hypothesis was tested using the screening data of the Oslo II-study in a cross sectional analysis. The Oslo-study included men aged 48–67 years.

The main finding was that the effect of a drinking pattern of 2–7 times per week reduced the risk of MI among men who had a history of tooth extractions due to infections versus tooth extraction for other causes or no extractions.  This hypothesis supports an explanation as to why oral infection is a weaker independent risk factor for CVD in some studies.  It also gives an indication of the reason for an added benefit by a regular drinking pattern as part of the Mediterranean diet.  The important consequence of this hypothesis is the added importance of optimal oral hygiene for the prevention of CVD as well as for the benefit of good oral health.

Forum Comments

Background:  For decades, almost all prospective studies have shown a lower risk of coronary heart disease among moderate drinkers of alcohol, and alcohol is generally considered to be cardioprotective.  On the other hand, many studies have shown a positive association between the presence of periodontitis, which is estimated to affect 10-15% of the adult population, and the risk of coronary disease.  The most adverse sequel of periodontal infection is the extraction of teeth.

This cross-sectional analysis among Norwegian men was designed to determine if the frequency of alcohol consumption related to tooth extraction due to infectious disease.  The authors hypothesized that because it is bactericidal, frequent alcohol consumption would decrease periodontal infection and lower the risk of tooth extraction related to infection, whereas the effect on the risk of no tooth extraction or extraction related to other causes (e.g., trauma) would be less.

Comments on study:  The study was based on data from 5,900 subjects in the Oslo-Study cohort who attended an examination in 2000 and provided self-reported data on alcohol consumption, tooth extraction, and a history of myocardial infarction (MI).  The investigators did not have specific data on periodontal disease from oral examinations (the authors used tooth extractions in subjects with infections as a surrogate), and did not have validation of self-reported MI. 

There were 2,229 subjects with no tooth extraction or extraction due to non-infectious causes (combined into one comparison group due to limited numbers).  A total of 3,671 subjects had extractions related, at least partly, to infection (included in this group were those with extraction due to infectious causes and those whose extraction was due to both infectious and non-infectious causes).  There were more than 600 cases of MI reported. 

As previous research has indicated that the largest reduction in risk of MI is associated with frequent drinking, the authors compared the occurrence of MI according to whether the subjects reported that they consumed alcohol 2-7 days per week versus < 2 days/week.  In logistic regression analysis the investigators adjusted for common cardiovascular risk factors (age, smoking, lipids, blood pressure, level of education, and BMI)  

The authors report that subjects with tooth extraction due to infectious oral disease alone had 35% greater risk of MI (OR 1.35) than other subjects.  Conversely, when alcohol was considered alone, more frequent drinkers (2-7 days/week) had 18% lower risk (OR 0.82) of having a MI than those not drinking or drinking less frequently.  With both infections disease and alcohol in the model, the more frequent alcohol consumers were 19% less likely to have a MI (OR 0.81, 95% CI 0.67 – 0.98) than those drinking less frequently.  Subjects with infectious extractions were 36% more likely than others to have a MI (OR 1.36, CI 1.12 – 1.65). 

Statistical concerns about study design and results:  This was a rather complex design, as because of limited numbers the authors were able to compare only two groups of drinkers: (1) they combined those reporting no alcohol and those reporting the consumption on < 2 days/week, versus (2) those reporting drinking on 2-7 days/week.  Further, they combined subjects with no tooth extractions or extractions due to non-infectious periodontal disease and compared them with those who had extractions due to infectious disease or a combination of infectious disease and other causes. 

Unfortunately, the authors do not report the calculation of interaction terms, nor do they provide a statistical evaluation of the differences between MI risks for infectious tooth extractions among frequent drinkers and among others. There was no significant effect on whether the drinking subjects consumed ≤ 1 drink, 2-3, or ≥ 4 drinks/day.

Forum reviewer Zhang commented: “The authors stated that ‘there is little evidence that explains the risk of MI from oral infections being modified by alcohol;’ however, the current study did not directly test this hypothesis.  The authors state that the association between infectious and combined infectious and non-infectious extraction and MI was assessed according to alcohol drinking pattern, but no formal statistical testing was done to assess the effect measure of modification, i.e., if OR=1.27 (for frequent drinkers with infectious extractions) was significantly different from 1.42 (for less frequent drinkers with infectious extractions).”

Does the paper adequately test the hypothesis?  Forum reviewers were uniform in believing that this paper does not really help test the hypothesis that the effects of alcohol reduce the risk of MI among subjects with periodontal disease.  Several were concerned that drinking may not be an effective oral antiseptic agent, and that poor oral hygiene must be associated with many confounding factors.  Stated reviewer Orgogozo: “The pro-inflammatory pro-atherogenic effect of periodontal chronic infection is mostly at the teeth roots and adjacent mandibular bone, totally inaccessible to mouth antiseptics.”  Stated another reviewer: “The authors talk about alcohol and oral antisepsis, but then switch to discuss alcohol as a part of the Mediterranean diet.  The argumentation in the Discussion part is somewhat weak.”

Oral health and cardiovascular disease:  Forum reviewer Goldfinger was happy that this topic was being discussed.  “I believe it is a very plausible hypothesis in the prevention of cardiovascular disease.  Others have suggested that the Mediterranean diet may affect coronary disease through alcohol antisepsis in the mouth and also the process of mastication by more bulky foods that help to clean the mouth.  ‘An apple a day keeps the doctor away’ may be related to chewing and cleaning periodontal spaces, as much as the biologic benefits of the apple itself.” 

He continued, “However, this is far from a novel hypothesis, which is what the tone of the paper suggests.  Others have shown association of periodontal disease with increased carotid medial thickening, a known marker for coronary disease, and there is increased risk of hypertension among patients with dental disease.  In an excellent study by Tonetti et al (Tonetti MS et al. Treatment of periodontitis and endothelial function.  New Engl J Med 2007;356:911-920), periodontal disease was shown to be associated with endothelial dysfunction, and this could be reversed with intensive cleaning and dental care.  Thus, the association between endothelial dysfunction and cardiovascular disease is solid.”

Goldfinger added: “While there is a large sample size, the paper relies wholly on self reporting of alcohol intake, history of MI, as well as the pathologic cause for need for dental extraction.  In addition to problems with self reporting of alcohol, I am concerned that many people who are admitted to the hospital with non-cardiac chest pain syndromes, or who have non-ischemic arrhythmias, or misdiagnosed GI disorders, perpetuate the falsehood that they have had an MI.  Thus, I do not believe that this paper adds much to our knowledge about this topic.”

Stated Forum reviewer Skovenborg: “I agree with Goldfinger that accumulating evidence has associated severe periodontal disease with increased odds of future cardiovascular disease events.  An interesting, large population based survey from Scotland (Cesar de Oliveira et al. Toothbrushing, inflammation, and risk of cardiovascular disease: results from Scottish Health Survey. BMJ 2010;340:c2451) found a 70% increased risk (HR 1.70) of cardiovascular disease events associated with less frequent tooth brushing.  However, that study also identified other independent predictors of disease events: smoking (HR 2.4), hypertension (HR 1.7), and diabetes (HR 1.9).  Systemic inflammation could represent the underlying mechanism that links oral health and cardiovascular disease; however, future experimental studies will be needed to confirm whether the observed association between oral health behaviour and cardio-vascular disease is in fact causal or merely a risk marker.”

Skovenborg also had problems with the statement that “alcohol is a bactericidal agent.”  He stated: “The bactericidal effect of alcohol depends on the alcohol-water-ratio.  In general 10 and 20% solutions have little or no bactericidal effect in ten minutes or less at room temperature.  Solutions of 30, 40 and 50% alcohol show progressively greater germicidal power, and 60 to 90% solutions by weight are all strongly and rapidly bactericidal.  The antibacterial action of alcohol is neutralized by the presence of proteins.  Thus, alcoholic drinks with less than 30% alcohol by weight are not bactericidal, and intake of alcohol with food (protein) eliminates the bactericidal effect of alcohol, as has been known for more than 60 years (Price PB. Reevaluation of ethyl alcohol as a germicide. Arch Surg 1950;60:492-502.)

“The lack of beverage-specific data is also a large problem.  White and red wines have a significant bactericidal effect due to the combination of alcohol, polyphenols, organic acids and pH (Draczynski M, ûber die bactericide Wirkung von Wein und Traubenmost, mit Bacterium coli als Test. Wein Rebe Jahrb Weinbauwiss Önologie Deut Wein-Ztg 1950-51:25-42).  Enteric pathogens like Salmonella, Shigella and E. coli are killed by incubation in wine as opposed to beer and cola (Sheth NK, et al. Survival of enteric pathogens in common beverages: An in vitro study. Am J Gastroenterology 1988;83:658-60.)

“Besides the major flaws cited, the Norwegian paper is a catalogue of bias and confounding – especially confounding by social economic status (SES). A model with adjustment for education leaves plenty of room for residual SES confounding.  It appears that what the authors have managed to find is a group of sensible drinking inhabitants of Oslo, who are probably wine drinkers, who eat healthy food and brush their teeth twice a day.”

Reviewer Kiel also had real concerns about the paper: “We deal with a cross sectional analysis in
which the exposure variable alcohol, the mediating variable periodontal disease or oral infections, and the dependent variable MI have all been obtained via questionnaire data. While most studies on alcohol consumption and CHD are based on questionnaire assessed alcohol intake data [which once in a while are ‘validated’ with food frequency data obtained in the same subjects (showing good correlations)], the assessment of oral infections or oral health by questionnaire may not be accurate, and in my opinion the assessment of MI only by questionnaire is really outdated and hardly acceptable.  I fully agree with Skovenborg that we most likely are observing confounding by socioeconomic status.”

Keil concluded: “The propagation of good oral health is something we should do in its own right.  The poor data of this study do not justify the propagation of a causal pathway from alcohol intake via oral infections/oral health to myocardial infarction.”

Forum Summary

A cross-sectional analysis among more than 5,000 Norwegian men was designed to determine if the frequency of alcohol consumption related to the risk of myocardial infarction associated with infectious periodontal disease.  The authors hypothesized that because it is bactericidal, frequent alcohol consumption would decrease periodontal infection and lower the risk of tooth extraction related to periodontal infection.  The authors state that their main finding was that frequent drinking lowers the risk of MI associated with periodontal disease.

Forum reviewers had real concern about the paper and the conclusions of the authors.  Some stressed that the per cent of alcohol in most “drinks” is too low to be really bactericidal, that the key exposures (alcohol and oral infections) were based exclusively on self report and did not have information on type of beverage or pattern of drinking, that many confounders (especially related to socio-economic status) were not adequately taken into consideration, and that the authors did not really test the results of their study with appropriate statistical testing. 

Current scientific data show an increase in the risk of cardiovascular disease to be associated with periodontitis.  Further, data consistently show a decrease in the risk of cardiovascular disease from moderate drinking, and a very large number of mechanisms have been identified.   However, the present study does not confirm that alcohol consumption plays an important role in preventing heart disease through its effect on periodontitis.

                                                                         *                        *                      *

Comments on this critique by the International Scientific Forum on Alcohol Consumption were provided by the following members:

Tedd Goldfinger, DO, FACC, Desert Cardiology of Tucson Heart Center, Dept. of Cardiology, University of Arizona School of Medicine, Tucson, Arizona, USA

Ulrich Keil, MD, PhD, Institute of Epidemiology and Social Medicine,  University of Münster, Münster, Germany

Jean-Marc Orgogozo, MD, Professor of Neurology and Head of the Neurology Divisions, the University Hospital of Bordeaux, Pessac, France

Yuqing Zhang, MD, DSc, Epidemiology, Boston University School of Medicine, Boston, MA, USA

Erik Skovenborg, MD, Scandinavian Medical Alcohol Board, Practitioner, Aarhus, Denmark

Harvey Finkel, MD, Hematology/Oncology, Boston University Medical Center, Boston, MA, USA

Arne Svilaas, MD, PhD, general practice and lipidology, Oslo University Hospital, Oslo, Norway

R. Curtis Ellison, MD, Section of Preventive Medicine & Epidemiology, Boston University School of Medicine, Boston, MA, USA