Alcohol Consumption, Genetic Risk, and Intraocular Pressure and Glaucoma: The Canadian Longitudinal Study on Aging

Abstract

Purpose: The purpose of this study was to examine the association of alcohol consumption with intraocular pressure (IOP) and glaucoma and to assess whether any associations are modified by a glaucoma polygenic risk score (PRS).

Methods: Cross-sectional analysis of data from the Canadian Longitudinal Study on Aging Comprehensive Cohort, consisting of 30,097 adults ages 45 to 85 years, was done. Data were collected from 2012 to 2015. Alcohol consumption frequency (never, occasional, weekly, and daily) and type (red wine, white wine, beer, liquor, and other) were measured by an interviewer-administered questionnaire. Total alcohol intake (grams/week) was estimated. IOP was measured in mm Hg using the Reichert Ocular Response Analyzer. Participants reported a diagnosis of glaucoma from a doctor. Logistic and linear regression models were used to adjust for demographic, behavioral, and health variables.

Results: Daily drinkers had higher IOP compared to those who never drank (β = 0.45, 95% confidence interval (CI) = 0.05, 0.86). An increase in total weekly alcohol intake (per 5 drinks) was also associated with higher IOP (β = 0.20, 95% CI = 0.15, 0.26). The association between total alcohol intake and IOP was stronger in those with a higher genetic risk of glaucoma (P for interaction term = 0.041). There were 1525 people who reported being diagnosed with glaucoma. Alcohol consumption frequency and total alcohol intake were not associated with glaucoma.

Conclusions: Alcohol frequency and total alcohol intake were associated with elevated IOP but not with glaucoma. The PRS modified the association between total alcohol intake and IOP. Findings should be confirmed in longitudinal analyses.

Alcohol consumption is a significant public health concern and has been recognized as a major modifiable risk factor for all-cause mortality¹ and age-related disease, including cancer,² liver disease,³ and diabetes.³ Despite the disease burden, alcohol use remains highly prevalent. According to the 2017 Canadian Tobacco, Alcohol, and Drugs Survey, 21% of those who consumed alcohol exceeded Canada's Low-Risk Alcohol Drinking Guideline for chronic effects, which at the time was 10 drinks per week for women and 15 drinks per week for men.⁴ 

Excessive alcohol consumption has previously been found to be associated with neurodegenerative diseases,³ including Alzheimer's disease, which shares common pathophysiological mechanisms with diseases like glaucoma.⁵ Alcohol consumption is also associated with an increased risk of hypertension,⁶ itself a risk factor for high eye pressure. It was therefore hypothesized that alcohol may also be implicated in the development of glaucoma. Studies reporting the associations of habitual alcohol consumption with intraocular pressure (IOP) and glaucoma, however, have been inconsistent. A recent systematic review and meta-analysis found alcohol use to be associated with higher IOP and open-angle glaucoma in pooled analyses⁷; however, effect estimates were small and heterogeneity was considerable. Some of this heterogeneity may be explained by the lack of consideration of genetic factors in most previous studies. With heritability estimates ranging from 0.29 to 0.79, genetic factors may account for a large proportion of the variance in IOP^8–14 and, in recent years, a polygenic predisposition to elevated IOP/glaucoma has been identified.^15-17 For a disease like glaucoma, assessing genetic factors or environmental factors in isolation without considering their interaction could be misleading. In fact, interactions between genetic predisposition to IOP/glaucoma and lifestyle factors, such as caffeine intake and diet, have been reported previously.^18,19 To our knowledge, only one study has reported on the joint effects of polygenic risk and alcohol consumption in relation to IOP/glaucoma, and they found a stronger association between alcohol consumption and IOP in those with the most genetic risk.²⁰ Other studies reported sex-related differences in the association between alcohol and IOP.^21,22 

As such, we used cross-sectional data from a large population-based sample of Canadian adults to examine associations of alcohol consumption with IOP and glaucoma and assessed whether associations were modified by a polygenic risk score or sex. 

Methods

Study Population and Design

We performed a cross-sectional analysis using the first round of data collection from the Canadian Longitudinal Study on Aging (CLSA) Comprehensive Cohort, which consists of 30,097 Canadian adults aged 45 to 85 years, with data collected every 3 years.²³ We only used the first round of data collection because not enough people have developed incident glaucoma to perform a longitudinal analysis at this time. The sample of the Comprehensive Cohort was obtained by utilizing stratified random sampling of provincial healthcare registration databases and random digit dialing of landline telephones. Baseline data were obtained between 2012 and 2015 via in-home interviews and in-person physical examinations and biospecimen sample collections at CLSA data collection sites, which are located in Victoria, Vancouver, Surrey, Calgary, Winnipeg, Hamilton, Ottawa, Montreal, Sherbrooke, Halifax, and St. John's, Canada. To be included in the study, participants had to be aged 45 to 85 years, community dwelling, cognitively unimpaired, and speak English or French. Exclusion criteria included being a full-time member of the Canadian Armed Forces, residing on a federal First Nations reserve or settlement, living in a long-term care institution, or not being a permanent resident or Canadian citizen. 

Informed Consent and Ethics Approval

Written informed consent was obtained for all participants. Research ethics board approval was obtained for all CLSA affiliated sites in July 2010. Ethics approval for the present analysis was obtained from the University of Ottawa in October 2021. 

Ocular Data

Participants were asked to report if they have ever had a physician’s diagnosis of glaucoma. IOP was measured at the CLSA data collection sites using the Reichert Ocular Response Analyzer (Reichert Technologies, Depew, NY, USA). The average IOP of the right and left eyes was used to derive participant-level IOP values. If one eye had missing IOP data, then the IOP value of the other eye was used. To estimate the pretreatment IOP, the IOP of those participants taking medications with a Drug Identification Number indicative of an IOP-lowering eye drop was divided by 0.7, which is the mean estimated treatment effect.²⁴ We used corneal-compensated IOP in our analyses, which is adjusted for corneal mechanical properties. IOP values greater than 60 mm Hg were excluded, as these were considered probable measurement errors (n = 3 people) based on expert opinion. 

Alcohol

Alcohol use was measured by interviewers during the in-home visit by asking participants “Have you ever drunk alcohol” and “About how often during the past 12 months did you drink alcohol?” Participants were then categorized as never, occasional, weekly, or daily drinkers. We defined occasional drinking as occurring zero to three times per month and weekly drinking as one to five times per week. Daily drinking was defined as drinking six or more times per week. 

Data on the type of alcohol consumed (white and red wine, beer, liquor, and other alcohol) were obtained in responses to the following questions: “In a typical week during the past 12 months, how many drinks of each of the following do you drink on weekends, that is, on Fridays and Saturdays?” and “In a typical week during the past 12 months, how many drinks of each of the following do you drink on weekdays, that is, from Sundays through Thursdays?” A drink was defined as 12 ounces of beer, 5 ounces of wine, or 1.5 ounces of liquor. Weekly consumption was then calculated for each alcohol type by adding weekday and weekend consumptions. People not currently drinking were assigned zero drinks of each type of alcohol. The weekly consumptions of each type of alcohol were added together. Total alcohol intake (grams/week) was then calculated by multiplying the total number of drinks per week by 13.45 grams, the number of grams of pure alcohol in a standard drink in Canada. Total alcohol values greater than 500 g/week were considered probable reporting errors and were excluded, similar to other research.²⁰ For more interpretable regression results, total alcohol intake was divided by 70 grams (approximately 5-drink increase per week). 

Polygenic Risk Score

The Affymetrix Axiom array was used to perform genomewide genotyping of non-fasting blood samples from consenting participants of the CLSA Comprehensive Cohort, resulting in 794,409 single nucleotide polymorphisms (SNPs) from 26,622 participants.²⁵ Release 3 of the CLSA genomic data was used and followed the marker- and sample-based quality control checks performed by the CLSA according to standard procedures.²⁶ Marker-based checks included checks for genotype consistency across genotyping batch, chromosomally defined sex, Hardy-Weinberg equilibrium, and discordance of genotyping across control replicates, whereas sample-based checks included checks for relatedness, heterozygosity, and genotype missingness. We excluded 15 individuals with extreme values of heterozygosity and genotype missingness and 1666 related individuals. In addition, the CLSA genomic data release included genotype data imputed using the TOPMed reference panel at the University of Michigan Imputation Service, containing 97,256 reference samples at 308,107,085 genetic markers.²⁶ 

Both the genotyped and imputed SNP data were used to calculate a glaucoma polygenic risk score (PRS) for each CLSA participant with available genotype data that passed quality control checks. The PRS was developed by Craig et al.²⁷ based on 2673 independent SNPs associated with glaucoma from their recent multitrait analysis of genomewide association studies. After SNP alignment between the Craig et al. PRS SNPs and the CLSA data based on genome build GRCh38/hg38, there were 2652 SNPs that were available to calculate the PRS in the CLSA data (i.e. 0.8% SNPs were either not present in the CLSA data or were removed in the quality control steps). Given the small proportion of missing SNPs, proxy SNPs were not selected to replace the missing ones. The PRS was calculated for each CLSA participant using a weighted sum of the 2652 SNPs: \(\mathop \sum \nolimits_{i = 1}^{2652} {\hat \beta _i} \times {\rm{SN}}{{\rm{P}}_i}\), where \({\hat \beta _i}\) is the estimated effect size of SNP_i on glaucoma from Craig et al. and SNP_i is the number of copies of the effect allele in an individual genotype or the expected number of copies of the effect alleles for imputed genotypes (allelic dosage). 

Demographic, Health, and Lifestyle Data

Demographic data, including age, sex, race, education, and household income, were collected during the in-home visit using an interviewer-administered questionnaire. Participants were grouped into White and Nonwhite race to have sufficient sample size for analysis. Height and weight were measured using standardized procedures at data collection site visits. Body mass index (BMI) was calculated and classified according to the World Health Organization cutoff points (underweight <18.5 kg/m², normal weight 18.5–24.9 kg/m², overweight 25.0–29.9 kg/m², and obese ≥30.0 kg/m²).²⁸ 

Participants were asked to report whether they ever received a physician’s diagnosis of chronic conditions, including diabetes and high blood pressure. Blood pressure was measured six times using the BpTru BPM200 blood pressure monitor (Medaval, Dublin, Ireland). The first reading was discarded, and the average of the subsequent five readings was used. Hypertension was defined if a participant reported a physician’s diagnosis of hypertension or if the average systolic blood pressure was 130 mm Hg or higher or diastolic blood pressure was 80 mm Hg or higher.²⁹ 

Smoking status was classified as current, never, or former based on participant responses to the interview questions “Have you smoked at least 100 cigarettes in your life?” and “At the present time, do you smoke cigarettes daily, occasionally (at least once in the last 30 days), or not at all (not in the last 30 days)?” A current smoker was defined as a person who reported smoking at least 100 cigarettes and currently smokes daily or occasionally, whereas a former smoker was someone who reported smoking at least 100 cigarettes in their life but had not smoked in the last 30 days. 

Diet was assessed using a validated 36-item Short Diet Questionnaire (SDQ), designed to measure usual consumption (during the last 12 months) of total fat, fatty acids, cholesterol, trans fat, dietary fiber, calcium, vitamin D, and servings of fruits and vegetables.^30,31 Total caloric intake was calculated by methods previously described³⁰ using participants’ reported frequencies of consumption of each SDQ item, for standard portion sizes estimated from a full food frequency questionnaire³² administered in the NuAge study,³³ and a nutrient database based on the 2015 Canadian Nutrient File. 

Statistical Analysis

Demographic, health, and behavioral factors were compared by alcohol consumption frequency. In separate multivariable analyses, linear regression was used to determine the relationship among alcohol consumption frequency, total alcohol intake, and alcohol type with IOP, whereas logistic regression was used for glaucoma. We also ran regression models with all alcohol types entered together. Locally weighted scatterplot smoothing was used to graph IOP versus predictor variables and linearity was checked by visual inspection. Possible nonlinearity was examined by adding squared terms or spline terms and testing their statistical significance. Regression models were adjusted for potential confounding variables, including age, sex, education, income, race, smoking, diabetes, systemic hypertension, BMI, total caloric intake, and province. The variables were entered into the regression models either as continuous variables or categorical variables, as they are shown in Table 1. Potential effect modification by the PRS was examined in two ways: by stratifying the regression models by PRS quartile and by fitting an interaction term between total alcohol intake and PRS quartile, entered as a linear term. Sensitivity analyses were done (1) using current IOP instead of pretreatment IOP, and (2) limiting IOP analyses to those without glaucoma. As recommended by the CLSA, sampling weights and strata variables were incorporated into all analyses using the SVY commands in Stata SE 16 (StataCorp, College Station, TX, USA). 

Table 1.

 

Distribution of Participant Characteristics by Alcohol Consumption Frequency

 

Results

Descriptive Characteristics

Over 99% of participants in the CLSA Comprehensive Cohort (n = 30,084) had complete alcohol consumption frequency data. Of the sample, 16% were daily drinkers, 41% were weekly drinkers, 40% were occasional drinkers, and 2% never drank. Participant characteristics by alcohol consumption frequency are presented in Table 1. As compared to never, occasional, and weekly drinkers, daily drinkers were older, more likely to be current or former smokers, and to have higher IOP. Daily drinkers were less likely to make less than $20,000 per year and to be obese. Never drinkers were more likely to be women, Nonwhite, and to never smoke. 

In Table 2, we compared those with and without a report of glaucoma. People with glaucoma were older, had higher pretreatment and current IOP, were more likely to take ocular antihypertensive medications, and were more likely to be in the highest PRS quartile than those without glaucoma. 

Table 2.

 

Characteristics of People With and Without Glaucoma

 

Alcohol and IOP

Higher alcohol consumption frequency was associated with higher IOP after adjustment for demographic, lifestyle, and health variables, as shown in Table 3. As compared with never drinkers, daily (β = 0.45, 95% confidence interval [CI] = 0.05, 0.86) drinkers had higher IOP. Occasional and weekly drinkers did not have higher IOP compared to never drinkers (P > 0.05). The total amount of alcohol consumed was also associated with IOP. An increase in total alcohol intake of five drinks per week was associated with higher IOP after adjustment (β = 0.20, 95% CI = 0.15, 0.26). 

Table 3.

 

Linear Regression Analysis of the Association of Alcohol Consumption Frequency With IOP

 

Associations between alcohol type and IOP are presented in Supplementary Table S1. Among alcohol types, a five drink per week increase in daily red wine (β = 0.23, 95% CI = 0.15, 0.32), daily white wine (β = 0.18, 95% CI = 0.06, 0.31), and beer (β = 0.20, 95% CI = 0.10, 0.30) was associated with higher IOP, whereas liquor (β = 0.13, 95% CI = −0.01 to 0.27), and other (β = 0.10, 95% CI = −0.51 to 0.71) alcohol types were not statistically significantly associated with IOP. 

Alcohol and Glaucoma

In Table 4, no statistically significant associations were found between alcohol consumption frequency and glaucoma after adjustment for demographic, lifestyle, and health variables (P > 0.05). Glaucoma was also not statistically significantly associated with alcohol type (Supplementary Table S2) or total alcohol intake (odds ratio [OR] = 1.01, 95% CI = 0.95, 1.08; Supplementary Table S3) after adjustment for demographic, lifestyle, and health variables (P > 0.05). 

Table 4.

 

Logistic Regression Analysis of the Association of Alcohol Consumption Frequency With Glaucoma

 

Investigation of Interaction by PRS

Analyses including the PRS were limited to those with genetic data. The PRS was very strongly associated with IOP with β values for each ascending quartile being 0.79, 1.38, and 2.04 in those per quartile of increasing genetic risk, respectively (P < 0.001). These β values represent the difference in the mean IOP (in mm Hg) for a one-category difference in the PRS. Total alcohol intake was more strongly associated with IOP in people in higher quartiles of genetic risk (Table 5). For example, the β values between alcohol and IOP in the 4 PRS quartiles were 0.13, 0.17, 0.20, and 0.25 per quartile of increasing genetic risk, respectively. This interaction was statistically significant (interaction term P = 0.041). 

Table 5.

 

Investigation of Interaction Between PRS and Alcohol With IOP

 

By contrast, total alcohol intake was not associated with glaucoma in any of the PRS quartiles and there was no trend to indicate a stronger association in those with greater genetic risk of disease (interaction term P value = 0.654; Supplementary Table S4). 

Sensitivity Analyses

Given that we estimated pretreatment IOP for people taking IOP-lowering medication to account for treatment effects, we did a sensitivity analysis using the current IOP values. The results from the sensitivity analysis were consistent with our main results with the exception that frequency of liquor consumption was now statistically significantly related to IOP (β = 0.14, 95% CI = 0.01, 0.28). In addition, people with glaucoma were excluded to see if associations differed in those without glaucoma. After exclusion of those with glaucoma (n = 1103), daily drinking showed an attenuated relationship with IOP (β = 0.35, 95% CI = −0.06 to 0.76), P = 0.090). The association between total alcohol intake and IOP was unaffected (P < 0.001). The interaction between alcohol intake and the PRS was very similar as the β values between alcohol and IOP in the 4 PRS quartiles were 0.12, 0.18, 0.18, and 0.22 per quartile of increasing genetic risk, respectively. However, the loss of sample size led to an attenuated P value for the interaction term (P = 0.102). 

Discussion

Consuming an increased frequency and amount of alcohol, particularly red wine and beer, was associated with higher levels of IOP, whereas it was not associated with the prevalence of glaucoma. Daily drinkers had an IOP that was 0.45 mm Hg higher than people who never drank alcohol. Furthermore, the association between total alcohol intake and IOP was stronger in those with higher genetic risk of glaucoma. 

The acute effect of alcohol (within 1–3 hours) is to lower the IOP.³⁴ However, long-term, detrimental effects of chronic alcohol consumption on IOP are biologically plausible. Alcohol consumption can lead to dehydration by increasing urine production. Dehydration may cause increases in blood viscosity and flow resistance, which could impact IOP.³⁵ Chronic drinking also releases cortisol, which can increase blood pressure,³⁶ a risk factor for IOP. Furthermore, increased oxidative stress and DNA damage associated with chronic alcohol use may exacerbate and/or accelerate age-related changes³⁷ of the trabecular meshwork.³⁸ 

One would think that because alcohol was related to higher IOP, it would also be related to glaucoma. Perhaps explaining our finding that alcohol was associated with higher IOP but not glaucoma, certain types of alcohol, like red and white wine and beer, contain varying concentrations of polyphenols including flavonoids, which may exert neuroprotective effects on the retina.³⁹ In patients with glaucoma and ocular hypertension, a systematic review and meta-analysis of eight randomized controlled clinical trials found that dietary flavonoid interventions had statistically significant benefits on improving or maintaining visual field relative to a placebo but had no significant effects on IOP, systolic, or diastolic blood pressure suggesting that any potential mechanism of action of flavonoids may be IOP-independent.⁴⁰ In addition, the use of the self-report of glaucoma, which may lead to misclassification, could explain the null finding. Finally, the use of cross-sectional data may have led to reverse causality such that people with glaucoma might have had higher alcohol consumption when younger, but after being diagnosed with glaucoma, they may have reduced their consumption. This would dilute any potential association. 

Our findings coincide with several previous studies reporting positive associations between alcohol use and IOP.^{7,20-22,41,42} We confirmed the finding by Stuart et al. of an interaction between alcohol consumption and the same glaucoma PRS on IOP.²⁰ Stronger associations in those at higher genetic risk may indicate a reduced reserve to withstand elevations of IOP due to dietary exposures, as discussed in other studies that examined associations of habitual caffeine and alcohol consumption with IOP and glaucoma.^19,20 

To our knowledge, only one previous study examined alcohol type (beer, wine, liquor, and sherry) with IOP.⁴³ In contrast to our finding in which servings of red wine, white wine, and beer were positively associated with IOP, Ramdas et al. found no association between alcohol type and IOP. Similar to our finding, prior research has demonstrated that daily red wine and beer consumption are both associated with increases in systemic blood pressure.⁴⁴ 

Our results differ from the findings of a recent systematic review and meta-analysis which found a positive association of alcohol use and open-angle glaucoma (OAG)⁷; however, the pooled effect size was small and of borderline statistical significance (OR = 1.18, 95% CI = 1.02, 1.36). Further, we found no evidence of associations between any alcohol type and glaucoma and the glaucoma PRS did not significantly modify the associations of alcohol consumption frequency or total intake with glaucoma, which coincides with previous research findings.²⁰ 

A major strength of this study is the utilization of a large population-based sample that includes genetic data. Among the limitations, glaucoma was based on self-report and no information was available on severity or subtype. However, studies that did have information on glaucoma subtype found that alcohol was associated with POAG, which is the most common type of glaucoma in Canada.⁷ Alcohol consumption was also based on self-report, which may have led to an under-report of drinking in some people due to social desirability. We did not have data on caffeine or sodium intake, which may have led to residual confounding. In addition, data on retinal nerve fiber layer and macular thickness, which have been previously found to be adversely associated with alcohol consumption,⁴⁵ were unavailable in the CLSA. Further, due to the cross-sectional design, we are unable to delineate temporality of the alcohol consumption and the onset of glaucoma/high IOP. Finally, the use of a PRS created from European-derived index variants, which sometimes do not replicate in non-European samples,⁴⁶ may not have captured the genetic risk of glaucoma as well in non-European samples. However, our results in the European sample were similar to our overall results. 

Although the effect sizes in this research may seem small and not clinically significant, it is important to remember that our results compare average IOP between participants rather than within participants. It is possible that daily drinking in a particular individual, especially at high genetic risk, may lead to much higher elevations of IOP within that individual than what our study showed on average. It is possible that daily drinking may make it more difficult to achieve the target IOP with treatment. 

Our research suggests that greater alcohol use and certain alcohol types are associated with elevated IOP but not with glaucoma. The association between total alcohol intake and IOP was stronger in those at higher genetic risk of glaucoma. Longitudinal research is needed to further understand the interaction of dietary and genetic factors on their risk of disease. 

Acknowledgments

This research has been conducted using the CLSA Comprehensive Baseline Dataset version 6.0 under Application Number 180911. The CLSA is led by Drs. Parminder Raina, Christina Wolfson, and Susan Kirkland. The opinions expressed in this paper are the authors’ own and do not reflect the views of the Canadian Longitudinal Study on Aging. 

Supported by the CIHR under Grants LSA 94473 and PJT-180615. 

Sponsor's Role: Funding for the Canadian Longitudinal Study on Aging (CLSA) is provided by the Government of Canada through the Canadian Institutes of Health Research (CIHR) under grant reference: LSA 94473 and the Canada Foundation for Innovation, as well as the following provinces: Newfoundland, Nova Scotia, Quebec, Ontario, Manitoba, Alberta, and British Columbia, Canada. This research was funded by CIHR operating grant PJT-180615 to Drs Roy-Gagnon and Freeman. The funders had no role in the design, analysis, or the interpretation of results. 

Disclosure: A. Grant, None; M.-H. Roy-Gagnon, None; J. Bastasic, None; A. Talekar, None; M. Jessri, None; G. Li, None; R. Buhrmann, None; E.E. Freeman, None