Volume 29, Issue 8, November–December 2015, Pages 1009–1014
Abstract
Aim
To explore the association between alcohol consumption and the severity of diabetic retinopathy (DR).
Methods
In
this cross-sectional study, patients with type 2 diabetes answered
questions on consumption of low and full-strength beer, white
wine/champagne, red wine, fortified wines, and spirits. Never, moderate
and high consumption of each alcoholic beverage, and overall alcoholic
beverage consumption, were defined as < 1, 1–14 and > 14 standard
drinks/week, respectively. DR was categorized into none; non
vision-threatening DR (VTDR) and VTDR. Multivariable logistic regression
determined the associations between alcohol consumption and DR.
Results
Of
the 395 participants (mean age ± SD [standard deviation]
65.9 ± 10.4 years; males = 253), 188 (47.6%) consumed alcohol and 235
(59.5%) had any DR. Compared to no alcohol consumption, moderate alcohol
consumption (overall) was significantly associated with reduced odds of
any DR (OR = 0.47, 95% CI [confidence interval] 0.26–0.85). Moderate
consumption of white wine/champagne or fortified wine was also
associated with reduced odds of any DR (OR = 0.48, 95% CI 0.25–0.91, and
OR = 0.15, 95% CI 0.04–0.62, respectively). Similar results were
observed for non-VTDR and VTDR.
Conclusions
The
amount and type of alcohol are associated with risk of DR in patients
with type 2 diabetes. A longitudinal study is needed to assess the
protective effect of alcohol consumption and DR.
Keywords
- Diabetic retinopathy;
- Alcohol consumption;
- Association;
- White wine;
- Fortified wine
1. Introduction
People
with type 2 diabetes who consume moderate amounts of alcohol have up to
36% lower relative risk of fatal coronary heart disease than
non-consumers (Howard et al., 2004 and Koppes et al., 2006).
However, it is less clear how alcohol consumption influences the risk
of microvascular complications, such as diabetic retinopathy (DR), a
serious complication of diabetes and a leading cause of blindness. There
are several plausible mechanisms through which alcohol may have a
protective effect for DR. For example, alcohol consumption increases
high-density lipoprotein (HDL), and decreases fibrinogen levels and
platelet aggregability, all of which have been reported to be inversely
related to development and progression of DR (Wang, Wang, & Wong, 2008).
Similarly, a recent meta-analysis has reported that moderate alcohol
consumption may decrease fasting insulin and HbA1c concentrations among
non-diabetic subjects (Schrieks et al., 2015).
Few studies have explored the association between alcohol consumption and DR, with inconsistent results (Wang et al., 2008). Two cross-sectional studies reported that moderate alcohol consumption (Beulens et al., 2008) and being a current or former alcohol drinker (Moss, Klein, & Klein, 1992) were protective for proliferative DR (PDR). In contrast, in a study over three decades ago, Young et al. (1984)
reported that heavy alcohol consumption (defined as > 10 pints of
beer or equivalent per week) was associated with twice the risk of
developing severe DR; however, confounding variables such as glycaemic
control and blood pressure levels were not controlled for in the
analysis. Other cross-sectional studies (Xu et al., 2009 and Yang et al., 2013) and several longitudinal studies (Del Canizo Gomez et al., 2011, Lee et al., 2010 and Moss et al., 1994)
have reported no association between any type of alcohol consumption
and risk of development or progression of DR, although the study by Lee et al. (2010) found that moderate and heavy alcohol consumption increased the risk of visual acuity loss in DR patients.
The
relationship between alcoholic beverage types and DR severity has also
not been comprehensively explored. In those studies which have assessed
the association between different alcoholic beverage types and DR (Beulens et al., 2008, Harjutsalo et al., 2014 and Lee et al., 2010),
low and high strength beers, and red and white wine, were not analyzed
separately. Given the well-known differential effects of red and white
wine on cardiovascular health (Sparwel et al., 2009), it is likely that the association between red and white wine consumption and DR may also differ.
Therefore,
in this study we explored the hypothesis that alcohol consumption is
associated with reduced risk of DR in a well-defined sample of
Australian adults with type 2 diabetes. We also explored the
relationship between low and high strength beer, red wine, white
wine/champagne, fortified wines (sherry or port), and spirits, and
severity of DR.
2. Subjects, materials and methods
Participants were recruited from the Diabetes Management Project (DMP), a clinical study conducted in Melbourne, Australia (Lamoureux et al., 2012).
In brief, English-speaking adults aged 18 years or older, with type 1
or type 2 diabetes, free of significant hearing and cognitive impairment
and living independently met the DMP inclusion criterion. Eligible
participants were screened using medical records and invited to
participate in the study by trained interviewers during their routine
clinical appointments at the Royal Victorian Eye and Ear Hospital
(RVEEH). The 6-item cognitive impairment test (6-CIT) (Brooke & Bullock, 1999) was used to assess patients' cognitive capacity, and those who failed were excluded from the main data analysis.
In the current study, only patients with type 2 diabetes were included (n = 510).
Thirteen and 102 participants were excluded from the analysis due to
missing DR severity data or alcohol consumption data, respectively,
resulting in a final sample size of 395. Comparison of baseline
characteristics of those included and those excluded due to missing
alcohol data revealed that those with alcohol data were more likely to
speak English as a first language (p < 0.001). There were no other significant differences between the two groups.
All
study procedures adhered to the tenets of the Declaration of Helsinki
and all privacy requirements were met. Written informed consent was
obtained from each participant prior to the study assessment. Ethical
approval for the study was provided by the Royal Victorian Eye and Ear
Hospital (RVEEH) Human Research and Ethics Committee (08/815H).
2.1. Testing protocol
All
examinations were conducted at the Centre for Eye Research Australia
(CERA) Melbourne, Australia. A trained interviewer administered a range
of psycho-social and behavioral questionnaires and data on
socio-demographic and health-related parameters, such as age, gender,
medical history, height, weight, lifestyle factors and duration of
diabetes were also collected.
2.2. Alcohol consumption
Alcohol intake was assessed using a self-administered, validated 145 item food frequency questionnaire (FFQ) (Smith et al., 1998).
This questionnaire has been shown to be valid, reliable and
reproducible, with a Spearman ranked correlation of 0.66 between the
alcohol data and three, four-day weighed food records spaced evenly over
one year in older Australian adults. Participants were asked how often
in the past 12 months they had drunk beer (high and low strength), red
wine, white wine/champagne, fortified wines (sherry or port), and
spirits (e.g. whisky/gin) (Supplementary Table 1).
Consumption was quantified according to the following categories: 1)
never; 2) less than 1 per month; 3) 1–3 per month; 4) 1 per week; 5) 2–4
per week; 6) 5–6 per week; 7) 1 per day; 8) 2–3 per day; and 9) 4 + per
day. To analyze our data, we first standardized the frequency
categories to units consumed per week, and second we converted the data
into standard drinks according to the definition of a ‘standard drink’
by the Australian Government Department of Health. For example, one
bottle/can of high strength beer (i.e. 4.8% alc. vol) was converted to
1.4 standard drinks. For white wine/champagne and red wine, we made the
assumption that people would drink a 150 ml average restaurant serving,
so one red wine glass was converted to 1.5 standard drinks (Supplementary Table 1).
For
data analysis, alcohol consumption was categorized as: (a) binary
variables (abstainer vs. alcohol consumer), (b) overall alcohol
consumption (number of standard drinks per week, any alcohol type), and
(c) consumption of each of the six types of alcoholic beverages. Never,
moderate and high consumption of each alcoholic beverage, as well as
overall alcohol consumption, were defined as < 1, 1–14 and > 14
standard drinks/week, respectively. High consumption of beer (low and
high strength), fortified wines, and spirits was not assessed due to
lack of data in those categories. Participants were also asked as part
of the FFQ if they had changed their eating habits in any way in the
last 5 years (yes/no).
2.3. Fundus photography and DR assessment
Two-field
(macula and optic disc) dilated fundus photos were captured using a
non-mydriatic retinal camera (Cannon CR6-45NM), Cannon Inc, Japan and
were graded using the Early Treatment Diabetic Retinopathy Study (ETDRS)
protocol (Dirani et al., 2011) and the American Academy of Ophthalmology classification (American Academy of Ophthalmology, 2002)
for the presence and severity of DR and DME, respectively: no
DR = 13–15, mild NPDR (non-proliferative DR) = 20; moderate
NPDR = 31–41; severe NPDR = 51; PDR = 60–80; and severe DME = 50.
Similar to previous studies by our group (Fenwick et al., 2012a and Fenwick et al., 2012b),
DR severity was categorized into none, non-vision threatening DR
(non-VTDR–mild/moderate NPDR) and VTDR (severe NPDR, PDR and/or severe
macular edema). Presence and severity of DR were the main outcome
variables.
2.4. Blood collection and blood pressure (BP) measurements
A total fasting blood sample of 34.5 ml was collected to assess glycosylated hemoglobin (HbA1c)
levels, fasting glucose and lipids (total cholesterol [TC],
triglyceride [TG], low-density lipoproteins [LDL] and high-density
lipoproteins [HDL]). All biochemical parameters were analyzed at
Melbourne Pathology, Melbourne, Australia.
A
BP assessment was completed on each individual using an automated BP
machine, model 5200-103Z (Welch Allyn, New Zealand). The average of two
separate measurements was recorded for systolic (SBP) and diastolic
(DBP). In cases where there was a difference of 10 mm Hg for SBP or 5 mm
Hg for DBP or greater, a third measurement was taken. The closest two
BP measurements were then averaged.
2.5. Anthropometric measurements, physical activity, and energy intake
All
individuals had their height and weight measured using a wall-mounted
adjustable measuring scale (Surgical and Medical products, China) and a
calibrated digital scientific weight scale (Oregon Scientific, PRC),
respectively. Individuals were instructed to remove any footwear and
heavy clothing prior to testing. BMI was calculated as weight (kg)
divided by height in meters squared (kg/m2). Physical
activity (PA) was assessed using the validated 7-day recall of habitual
physical activity (7-dPAR) which captures how often participants were
engaged in moderate, hard and very hard PA, and how much sleep they had,
and then infers time spent in light intensity activity (Sallis et al., 1985).
An average daily energy expenditure (kcal/day) is then calculated based
on time spent in these differing intensity activities (Richardson et al., 2001).
Total energy intake (kcal/day) was calculated from the FFQ data using
nutrient data from the electronic nutrient database for use in Australia
(NUTTAB, 2010) (NUTTAB, 2010).
2.6. Statistical analysis
Patient
demographics and baseline characteristics were summarized by mean and
standard deviation (SD) for normally distributed continuous data, or the
median and inter-quartile range for skewed distributed data, and counts
and percentages for categorical data. Key covariables included age
(years), gender, education (< 14 years/≥ 14 years), income
(<$30,000/≥$30,000), smoking status (non-smoker/current or past
smoker), insulin use (yes/no), change in dietary habit in the last five
years (yes/no), use of hypertensive medication (yes/no), use of
lipid-lowering medication (yes/no), poor diabetes control (HbA1c ≥ 7%,
yes/no), presence of comorbidity (none/at least one), presence of at
least one other diabetes complication (renal, peripheral vascular
disease, neuropathy: yes/no), country of birth (Australia/other),
physical activity (total daily energy expenditure, kcal/day); total
energy intake (kcal/day), main language spoken at home (English/other),
body mass index (BMI), duration of diabetes (years), systolic and
diastolic BP (SBP and DBP, mm Hg), HbA1c (%), fasting glucose (mmol/L),
TC (mmol/L), LDL (mmol/L); HDL (mmol/L), and TG (mmol/L). Multivariable
logistic regression analysis was used to examine the relationship
between alcohol consumption and DR, adjusted for age, gender and all
covariables that were significant in univariate analysis. When analysing
each of the specific beverage types (e.g. red wine), we adjusted for
all confounders as well as each of the other beverage types.
Associations were considered statistically significant if p < 0.05; all statistical analyses were undertaken using Stata version 12.0 (StataCorp, College Station, TX).
3. Results
A total of 395 people with type 2 diabetes participated in this study (alcohol consumers = 188, 47.6%; Table 1). The mean age ± SD [standard deviation] of the sample (males = 253) was 65.9 ± 10.4 years old and n = 235 (59.5%) had DR. Of those with DR (n = 235,
59.5%), 130 (55.3%) and 105 (44.7%) had non-VTDR and VTDR,
respectively. Those who consumed alcohol were more likely to be male,
older, a current or past smoker, and not using insulin, and were less
likely to have changed their dietary habits in the last five years
compared to abstainers (all p < 0.05, Table 1).
Those who consumed alcohol also had a higher energy intake and a lower
BMI compared to abstainers; however, given that both groups had a mean
BMI in the obese range, this result may not be clinically significant.
There was no significant difference between duration of diabetes between
abstainers and alcohol consumers. However, alcohol consumers had
significantly lower HbA1c levels (p < 0.05) compared to abstainers.
- Table 1. Sociodemographic and clinical characteristics of the participants (n = 395).
Characteristic Abstainers (n = 207)a
Alcohol consumers (n = 188)
p Categorical variables n % n % Gender (male) 106 51.2 147 78.2 < 0.0001 Current/past smoker (yes) 99 48.7 111 59.4 0.036 Income <$30,000 139 73.9 121 69.9 0.398 ≥$30,000 49 26.1 52 30.1 Education < 14 years 147 73.1 136 73.5 0.933 ≥ 14 years 54 26.9 49 26.5 Language spoken at home (English) 144 83.2 134 82.2 0.803 Country of birth (Australia) 94 45.4 71 37.8 0.124 Insulin use (yes) 85 41.3 52 27.8 0.005 Lipid-lowering medication (yes) 59 29.7 46 25.4 0.357 Hypertension medication (yes) 68 34.2 63 34.8 0.896 At least one diabetes complicationb 65 31.4 62 33.0 0.737 At least one comorbidityc (yes) 183 88.4 157 83.5 0.160 DR severity No DR 76 36.7 84 44.7 0.137 Non-VTDR 68 32.9 62 33.0 VTDR 63 30.4 42 22.3 Alcohol consumption (standard drinks) Beer (low strength) n/a n/a 18 11.3 Beer (high strength) n/a n/a 31 19.4 Red wine n/a n/a 56 35.0 White wine or champagne n/a n/a 38 23.8 Sherry or port n/a n/a 13 8.1 Spirits 28 17.5 Change in dietary habit in last 5 years (yes) 126 63.3 86 47.8 0.002 Continuous variablesd and e Mean/median SD/IQR Mean/median SD/IQR p-Value Age (years) 64.9 10.3 67.0 10.5 0.043 Systolic blood pressure (mm Hg) 138.9 19.9 141.9 17.8 0.120 Duration of diabetes (years)f 12.1 14.5 12.0 14.0 0.949 Body mass index (kg/m2) 31.9 6.4 30.5 6.1 0.036 HbA1c (%)f 7.6 1.9 7.2 1.4 0.015 Fasting glucose (mmol/L)f 8.1 3.5 7.4 2.7 0.145 Total cholesterol (mmol/L)f 4.4 1.5 4.4 1.7 0.695 HDL cholesterol (mmol/L)f 1.3 0.6 1.3 0.5 0.184 Triglycerides (mmol/L)f 1.6 1.1 1.5 1.0 0.214 LDL cholesterol (mmol/L)f 2.2 1.2 2.2 1.3 0.618 Total daily energy expenditure (kcal/day) 39.7 6.5 40.0 5.7 0.691 Total energy intake (kcal/day) 1730.6 641.1 1879.2 661.9 0.033 - Bolded values indicate significant results.
DR = diabetic retinopathy; HbA1c = hemoglobin A1c; IQR = interquartile range; SD = standard deviation. -
- a
- Chi-square test assessed difference in frequency distributions between abstainers and alcohol consumers.
- b
- Includes nephropathy, peripheral vascular disease, and neuropathy.
- c
- Includes hypertension, angina, irregular heartbeat, stroke, high cholesterol, asthma, anaemia, migraine, arthritis, and osteoporosis.
- d
- Student's unpaired t-test was used for the comparison of continuously distributed variables.
- e
- Wilcoxon rank-sum test was used for the comparison of nonparametric variables.
- f
- Characteristics were expressed as the median (interquartile range (IQR)) for non-normally distributed continuous variables.
In
multivariable models adjusted for age, gender, poor diabetes control
(HbA1c ≥ 7%), diabetes duration, smoking, BMI, SBP, insulin use, and
presence of at least one other diabetes complication, moderate alcohol
consumption (overall) was associated with reduced odds for any DR
compared to abstainers (OR = 0.47, 95% CI [confidence interval]
0.26–0.85, p = 0.013). In the beverage-specific analysis, those
who consumed moderate amounts of white wine/champagne and fortified
wine had reduced odds of any DR compared with abstainers (OR = 0.48, 95%
CI 0.25–0.91, p = 0.024 and OR = 0.15, 95% CI 0.04–0.62, p = 0.009, respectively) (Table 2). The plots in Fig. 1 and Fig. 2
display all significant risk and protective factors for DR in the
multivariate regression models for white wine/champagne and fortified
wine (exposures).
- Table 2. Association between overall alcohol consumption and consumption of six alcoholic beverage types, with any diabetic retinopathy.
Unadjusted OR (95% CI) p Adjusted OR (95% CI)a p Alcohol consumption (overall) None 1 Moderateb 0.69 (0.45, 1.07) 0.095 0.47 (0.26, 0.85) 0.013 High 0.83 (0.40, 1.73) 0.619 0.75 (0.25, 2.20) 0.599 Beer (low strength)c None 1 1 Moderate 1.29 (0.70, 2.38) 0.418 0.86 (0.38, 1.91) 0.703 Beer (high strength)c None 1 1 Moderate 0.64 (0.36, 1.12) 0.118 0.56 (0.27, 1.17) 0.123 Red wine None 1 1 Moderate 0.73 (0.47, 1.14) 0.166 0.62 (0.34, 1.14) 0.122 High 1.09 (0.31, 3.82) 0.893 1.43 (0.36, 5.70) 0.613 White wine or champagne None 1 1 Moderate 0.57 (0.33, 0.96) 0.036 0.48 (0.25, 0.91) 0.024 High 1.24 (0.30, 5.05) 0.765 1.07 (0.13, 8.82) 0.951 Sherry or Portc None 1 1 Moderate 0.25 (0.09, 0.70) 0.009 0.22 (0.05, 0.93) 0.039 Spiritsc None 1 1 Moderate 0.98 (0.57, 1.70) 0.955 1.28 (0.62, 2.65) 0.497 - Bolded values indicate statistically significant associations (p < 0.05) between amount and type of alcohol consumption, and any diabetic retinopathy.
-
- a
- Adjusted for: age, gender, smoking, body mass index, systolic blood pressure, diabetes control, insulin use, duration of diabetes, and presence of at least one other diabetes complication. We also simultaneously adjusted for each of the other beverage types.
- b
- None = < 1 standard drink/week; moderate consumption = 1–14 standard drinks/week; High consumption = > 14 standard drinks/week; CI = confidence interval; OR = odds ratio.
- c
- High consumption not assessed due to lack of data.
- Fig. 1.
Odds ratios (95% confidence intervals) of our multivariable regression model exploring the association between consumption of white wine/champagne (exposure) and DR (outcome). This plot shows that moderate consumption of white wine/champagne is an independent protective factor for DR. Notes: BMI = body mass index; SBP = systolic blood pressure.
- Fig. 2.
Odds ratios (95% confidence intervals) of our multivariable regression model exploring the association between moderate consumption of fortified wine (exposure) and DR (outcome). This plot shows that moderate consumption of fortified wine is an independent protective factor for DR. Notes: BMI = body mass index; SBP = systolic blood pressure.
Looking
at severity of DR, moderate alcohol consumption (overall) was also
associated with reduced odds for non-VTDR and VTDR compared to
abstainers (OR = 0.52, 95% CI 0.27–0.98, p = 0.044 and OR = 0.40, 95% CI 0.19–0.83, p = 0.015,
respectively). Moderate consumption of white wine/champagne was
associated with lower odds for VTDR (OR = 0.35, 95% CI 0.15–0.80, p = 0.013),
and moderate consumption of fortified wine was associated with lower
odds for both non-VTDR and VTDR (OR = 0.21, 95% CI 0.05–0.91, p = 0.038 and OR = 0.07, 95% CI 0.01–0.82, p = 0.034, respectively) (Table 3). Other alcoholic beverages were not associated with the presence or severity of DR.
- Table 3. Association between overall alcohol consumption and consumption of six alcoholic beverage types, with severity of diabetic retinopathy.
Non-VTDR
VTDR
Non-VTDR
VTDR
Unadjusted OR p Unadjusted OR p Adjusted ORa p Adjusted ORa p Alcohol consumption (overall) None 1 1 1 1 Moderateb 0.78 (0.48, 1.28) 0.327 0.60 (0.36, 1.02) 0.060 0.52 (0.27, 0.98) 0.044 0.40 (0.19, 0.83) 0.015 High 1.04 (0.46, 2.36) 0.930 0.60 (0.23, 1.58) 0.306 0.97 (0.33, 2.90) 0.963 0.39 (0.09, 1.65) 0.199 Beer (low strength)c None 1 1 1 1 Moderate 1.43 (0.72, 2.84) 0.302 1.11 (0.52, 2.38) 0.779 0.97 (0.42, 2.25) 0.941 0.69 (0.25, 1.91) 0.479 Beer (high strength)c None 1 1 1 1 Moderate 0.67 (0.34, 1.30) 0.239 0.59 (0.29, 1.23) 0.161 0.64 (0.29, 1.41) 0.271 0.44 (0.16, 1.21) 0.114 Red wine None 1 1 1 1 Moderate 0.79 (0.48, 1.32) 0.368 0.66 (0.38, 1.15) 0.144 0.67 (0.35, 1.27) 0.217 0.55 (0.26, 1.16) 0.117 High 0.86 (0.19, 3.93) 0.843 1.36 (0.33, 5.64) 0.665 1.12 (0.20, 6.29) 0.900 2.20 (0.52, 9.22) 0.282 White wine or champagne None 1 1 1 1 Moderate 0.65 (0.35, 1.20) 0.168 0.46 (0.23, 0.95) 0.034 0.56 (0.28, 1.12) 0.100 0.35 (0.15, 0.80) 0.013 High 1.14 (0.23, 5.78) 0.874 1.36 (0.27, 6.87) 0.714 1.16 (0.15, 8.83) 0.888 0.91 (0.06, 14.95) 0.948 Sherry or Portc None 1 1 1 1 Moderate 0.36 (0.11, 1.13) 0.080 0.11 (0.01, 0.85) 0.034 0.21 (0.05, 0.91) 0.038 0.07 (0.01, 0.82) 0.034 Spiritsc None 1 1 1 1 Moderate 1.33 (0.73, 2.41) 0.349 0.60 (0.28, 1.28) 0.185 1.64 (0.77, 3.49) 0.197 0.72 (0.29, 1.79) 0.475 - Bolded values indicate statistically significant associations (p < 0.05) between amount and type of alcohol consumption, and severity of diabetic retinopathy.
-
- a
- Adjusted for: age, gender, smoking, body mass index, systolic blood pressure, diabetes control, insulin use, duration of diabetes, and presence of at least one other diabetes complication. We also simultaneously adjusted for each of the other beverage types.
- b
- None = < 1 standard drink/week; moderate consumption = 1–14 standard drinks/week; high consumption = > 14 standard drinks/week; CI = confidence interval; OR = odds ratio.
- c
- High consumption not assessed due to lack of data.
4. Discussion
Our
study demonstrates that both the amount of alcohol and the type of
beverage are associated with the risk of DR in patients with Type 2
diabetes. Individuals who consumed moderate amounts of alcohol and, in
particular, white wine/champagne or fortified wines, had reduced the
odds of any DR, non-VTDR and VTDR compared to non-consumers. This is the
first study to specifically assess a range of alcoholic beverages
including fortified wine and to differentiate between red and white
wine. We are also the first study to find a significant association
between any type of alcohol consumption, and any DR and non-VTDR, with
all other studies reporting an association with VTDR only. Overall,
these findings suggest that a drinking pattern whereby alcohol
(particularly white or fortified wine) is consumed on several days of
the week in moderation may be a healthy pattern for the development and
progression of DR in people with type 2 diabetes.
Our findings support previous cross-sectional studies, which have reported a protective effect for overall alcohol consumption (Moss et al., 1992) and moderate wine consumption (defined as 30.0–69.0 grams of alcohol per week) (Beulens et al., 2008) on PDR, and for age-related macular degeneration (AMD) (Obisesan et al., 1998).
Similarly, abstainers and former drinkers have been shown to have
greater odds of DR compared to light alcohol consumers in another
cross-sectional study (Harjutsalo et al., 2014).
However, the same study also found that alcoholic spirit drinkers had
more than twice the odds of having severe DR than wine or beer drinkers,
which was not consistent with our results where consumption of spirits
was not significantly associated with severity of DR (Harjutsalo et al., 2014). Our results also differ from studies finding no association between any type of alcohol consumption and DR (Del Canizo Gomez et al., 2011, Lee et al., 2010, Moss et al., 1994, Xu et al., 2009 and Yang et al., 2013); however, in some of these studies, ‘alcohol consumption’ was either a dichotomous variable (yes/no) (Xu et al., 2009 and Yang et al., 2013), poorly defined (Del Canizo Gomez et al., 2011 and Yang et al., 2013),
or only heavy consumption was assessed, which may have limited their
capacity to find a significant association between alcohol consumption
and DR (Yang et al., 2013).
We found no association between high alcohol consumption and DR, unlike
other eye-related studies which have reported an association between
high alcohol consumption, and increased risk of AMD (Adams et al., 2012) and cataract (Gong et al., 2015).
It
has been hypothesized that wine may be protective for DR due to its
high polyphenolic content, which contains anti-oxidants that combat
sustained oxidation (Bola, Bartlett, & Eperjesi, 2014).
Indeed, resveratrol – the main active polyphenol in wine – inhibits
angiogenesis; prevents inflammation; and facilitates vaso-relaxation,
all of which result in increased blood flow in the retina and counteract
the reduced blood flow resulting from DR (Bola et al., 2014).
Interestingly, in our study, white wine/champagne, but not red wine,
was a protective factor for DR. Although red wine is more often
associated with reduced risk of cardiovascular and other diseases, some
studies have shown that moderate regular consumption of both red and
white wines has similar beneficial effects in reducing markers of
cardiovascular diseases (Williams et al., 2004), with white wine delivering even better results than red wine for some parameters (Lachtermann et al., 1999).
Moreover, a recent study has found that caffeic acid, a phenol found in
white wine, may exert a protective effect on endothelial cell function
which may limit cardiovascular and kidney disease progression associated
with oxidative stress-mediated endothelial injury (Migliori et al., 2015).
Similarly, another study has shown that polyphenols present in
champagne wine may induce a neuroprotective effect against oxidative
neuronal injury (Vauzour et al., 2007).
It is possible that similar mechanisms underlie the protective effect
of white wine and champagne on DR since both endothelial and neuronal
dysfunction have been implicated in DR pathogenesis (Cheung, Mitchell, & Wong, 2010).
It
is also interesting that fortified wine but not red wine was protective
for DR. Being derived from red wine, fortified wines also have high
polyphenolic content, which could explain the positive association,
although this does not explain why we did not find an association with
red wine. Data are scarce on the health benefits of fortified wines;
however, moderate consumption of fortified wine has been found to be
inversely associated with peripheral arterial disease (Vliegenthart et al., 2002).
However, it is possible that white wine or champagne drinkers, and
fortified wine drinkers may differ from drinkers of other beverages,
such as beer and spirits, in a number of sociodemographic
characteristics, such as healthier diet, more exercise, and higher
socioeconomic background (Barefoot et al., 2002),
all of which could be driving the protective effect of these beverages
on DR. To explore whether physical activity and energy intake had a
confounding effect on the association between alcohol consumption and
DR, we added them to our multivariable model (data not shown) but the
associations were not affected. However, there may still be a chance of
residual confounding.
Strengths
of our study include a well-characterized sample of Australian persons
with type 2 diabetes and DR. However, certain limitations have to be
addressed. First, as our study was cross-sectional, we are unable to
determine causality. Importantly, we do not know if current abstainers
were former drinkers who may have modified their lifestyle upon
diagnosis of DR. Indeed, evidence is emerging to suggest that the
frequently found benefit from moderate alcohol use is actually due to
confounding and mis-classification of former and occasional drinkers as
abstainers, rather than a true protective association (Chikritzhs et al., 2015).
We accounted for this issue to some extent by exploring whether change
of dietary habit in the last five years confounded the association
between alcohol and DR, but it had no effect (data not shown); however,
given that the mean duration of diabetes was over 10 years in our
sample, this variable would not capture change in drinking behavior at
diagnosis of DR when it may be most likely to occur. Therefore, current
consumption levels may underestimate the true relationship between
alcohol and DR and the findings of this study must be interpreted very
cautiously. Longitudinal studies are required to determine if our
findings could simply be cross-sectional data phenomena among people who
may have changed their alcohol consumption patterns following a
diagnosis of diabetes.
Second,
consumption of alcohol was self-reported, leading to potential recall
bias and the possibility for under-reporting due to the social stigma
associated with heavy drinking. However, at present there is no
objective means to quantify alcohol consumption. Therefore, to increase
measurement validity (Feunekes et al., 1999), we included a detailed assessment of alcohol consumption using a valid and reliable food frequency questionnaire (Smith et al., 1998),
which included overall alcohol consumption, frequency of consumption,
and specific beverage type. Future studies, particularly those
longitudinal in design, could consider using an ‘alcohol consumption’
diary with frequent text message reminders, as well as collecting data
on previous drinking habits, which would optimize data quality. In
addition, data could be collected on alcohol-related illnesses, which
would reveal if misclassification of heavy drinkers as abstainers has
occurred.
Third, we had
to exclude 20% of our sample from the analysis due to missing alcohol
data, which could have introduced selection bias. This likely occurred
because the alcohol questions were part of the very long (145-item) FFQ
and were usually conducted at the end of the assessment procedure, which
was approximately three hours in total and which included a battery of
clinical tests and other questionnaires. For these reasons, the FFQ was
often excluded if the patient was fatigued, the interview was running
overtime, or communication was difficult. This suggests that the
selection bias introduced by these missing data is not related to the
questions themselves (e.g. the topic of drinking habits), but rather
about time and convenience issues. Although there were no significant
differences between those with and without alcohol data on most
sociodemographic and clinical variables, we found that those who had
alcohol data were more likely to have English as a first language.
Similar results have been observed in a New Zealand study, which found
that non-responders to a survey measuring alcoholic drinking behavior
were more likely to be male, younger, of Maori descent and living in
deprived areas (Maclennan et al., 2012).
Therefore, the results of our study may not be generalizable to the
wider Australian population who have English as their second language.
Finally, the low frequency counts in some of our high consumption
categories (e.g. beer, fortified wine and spirits) meant that we were
unable to explore the association between high consumption of these
beverage types and DR.
In
summary, we found that moderate consumption of alcohol overall, and
white wine/champagne and fortified wines specifically, was associated
with reduced odds of any DR, non-VTDR and VTDR in this cross-sectional
study of people with type 2 diabetes. Our findings need further
exploration in a longitudinal study of people with diabetes to assess
whether the found association translates into a protective effect for
the incidence and progression of DR, and to better understand how
alcohol consumption might mediate the risk of DR.
The following is the supplementary data related to this article.
-
-
Classification of alcoholic beverage types.
Acknowledgments
This study was supported by the National Health and Medical Research Council Centre for Clinical Research Excellence (CCRE) (No. 529923)—Translational Clinical Research in Major Eye Diseases; CCRE Diabetes; Australian Research Council (ARC) Grant LP0884108; Royal Victorian Eye and Ear Hospital (RVEEH); and Operational Infrastructure Support from the Victorian Government. Dr Eva Fenwick is funded by a National Health and Medical Research Council (NHMRC) Early Career Fellowship (No. 1072987). The funding organizations had no role in the design or conduct of this research.
References
- Adams et al., 2012
- 20/20—Alcohol and age-related macular degeneration: The Melbourne Collaborative Cohort Study
- American Journal of Epidemiology, 176 (4) (2012), pp. 289–298
- | |
- Anonymous, 2002
- International Clinical Classification of Diabetic Retinopathy Severity of Macular Oedema
- American Academy of Ophthalmology (2002)
- Barefoot et al., 2002
- Alcoholic beverage preference, diet, and health habits in the UNC Alumni Heart Study
- The American Journal of Clinical Nutrition, 76 (2) (2002), pp. 466–472
- |
- Beulens et al., 2008
- Alcohol consumption and risk of microvascular complications in type 1 diabetes patients: The EURODIAB Prospective Complications Study
- Diabetologia, 51 (9) (2008), pp. 1631–1638
- | |
- Bola et al., 2014
- Resveratrol and the eye: Activity and molecular mechanisms
- Graefe's Archive for Clinical and Experimental Ophthalmology, 252 (5) (2014), pp. 699–713
- | |
- Brooke and Bullock, 1999
- Validation of a 6-item cognitive impairment test with a view to primary care usage
- International Journal of Geriatric Psychiatry, 14 (1999), pp. 936–940
- | |
- Cheung et al., 2010
- Diabetic retinopathy
- Lancet, 376 (9735) (2010), pp. 124–136
- | | |
- Chikritzhs et al., 2015
- Has the leaning tower of presumed health benefits from ‘moderate’ alcohol use finally collapsed?
- Addiction, 110 (5) (2015), pp. 726–727
- | |
- Del Canizo Gomez et al., 2011
- Microvascular complications and risk factors in patients with type 2 diabetes
- Endocrinología y Nutrición, 58 (4) (2011), pp. 163–168
- | | |
- Dirani et al., 2011
- Are obesity and anthropometry risk factors for diabetic retinopathy? The diabetes management project
- Investigative Ophthalmology & Visual Science, 52 (7) (2011), pp. 4416–4421
- | |
- Fenwick et al., 2012a
- Assessing disutility associated with diabetic retinopathy, diabetic macular oedema and associated visual impairment using the Vision and Quality of Life Index
- Clinical & Experimental Optometry, 95 (3) (2012), pp. 362–370
- | |
- Fenwick et al., 2012b
- The impact of diabetic retinopathy and diabetic macular edema on health-related quality of life in type 1 and type 2 diabetes
- Investigative Ophthalmology & Visual Science, 53 (2) (2012), pp. 677–684
- | |
- Feunekes et al., 1999
- Alcohol intake assessment: The sober facts
- American Journal of Epidemiology, 150 (1) (1999), pp. 105–112
- | |
- Gong et al., 2015
- Different amounts of alcohol consumption and cataract: A meta-analysis
- Optometry and Vision Science, 92 (4) (2015), pp. 471–479
- | |
- Harjutsalo et al., 2014
- Patients with Type 1 diabetes consuming alcoholic spirits have an increased risk of microvascular complications
- Diabetic Medicine, 31 (2) (2014), pp. 156–164
- | |
- Howard et al., 2004
- Effect of alcohol consumption on diabetes mellitus: A systematic review
- Annals of Internal Medicine, 140 (3) (2004), pp. 211–219
- Koppes et al., 2006
- Meta-analysis of the relationship between alcohol consumption and coronary heart disease and mortality in type 2 diabetic patients
- Diabetologia, 49 (4) (2006), pp. 648–652
- | |
- Lachtermann et al., 1999
- Moderate red and white wine consumption and the risk of cardiovascular disease
- Herz Kreislauf, 31 (1) (1999), pp. 25–31
- |
- Lamoureux et al., 2012
- Methodology and early findings of the Diabetes Management Project: A cohort study investigating the barriers to optimal diabetes care in diabetic patients with and without diabetic retinopathy
- Clinical & Experimental Ophthalmology, 40 (1) (2012), pp. 73–82
- | |
- Lee et al., 2010
- Association between alcohol consumption and diabetic retinopathy and visual acuity—The AdRem Study
- Diabetic Medicine, 27 (10) (2010), pp. 1130–1137
- | |
- Maclennan et al., 2012
- Non-response bias in a community survey of drinking, alcohol-related experiences and public opinion on alcohol policy
- Drug and Alcohol Dependence, 126 (1–2) (2012), pp. 189–194
- | | |
- Migliori et al., 2015
- Caffeic Acid, a phenol found in white wine, modulates endothelial nitric oxide production and protects from oxidative stress-associated endothelial cell injury
- PloS One, 10 (4) (2015), p. e0117530
- Moss et al., 1992
- Alcohol consumption and the prevalence of diabetic retinopathy
- Ophthalmology, 99 (6) (1992), pp. 926–932
- | | |
- Moss et al., 1994
- The association of alcohol consumption with the incidence and progression of diabetic retinopathy
- Ophthalmology, 101 (12) (1994), pp. 1962–1968
- NUTTAB, 2010
- Australian Food Composition: Food Standards Australia New Zealand
- (2010) (Canberra)
- Obisesan et al., 1998
- Moderate wine consumption is associated with decreased odds of developing age-related macular degeneration in NHANES-1
- Journal of the American Geriatrics Society, 46 (1) (1998), pp. 1–7
- Richardson et al., 2001
- Validation of the Stanford 7-day recall to assess habitual physical activity
- Annals of Epidemiology, 11 (2) (2001), pp. 145–153
- Sallis et al., 1985
- Physical activity assessment methodology in the Five-City Project
- American Journal of Epidemiology, 121 (1) (1985), pp. 91–106
- Schrieks et al., 2015
- The effect of alcohol consumption on insulin sensitivity and glycemic status: A systematic review and meta-analysis of intervention studies
- Diabetes Care, 38 (4) (2015), pp. 723–732
- Smith et al., 1998
- Validity and reproducibility of a self-administered food frequency questionnaire in older people
- Australian and New Zealand Journal of Public Health, 22 (1998), pp. 456–463
- Sparwel et al., 2009
- Differential effects of red and white wines on inhibition of the platelet-derived growth factor receptor: Impact of the mash fermentation
- Cardiovascular Research, 81 (4) (2009), pp. 758–770
- Vauzour et al., 2007
- Champagne wine polyphenols protect primary cortical neurons against peroxynitrite-induced injury
- Journal of Agricultural and Food Chemistry, 55 (8) (2007), pp. 2854–2860
- Vliegenthart et al., 2002
- Alcohol consumption and risk of peripheral arterial disease: The Rotterdam study
- American Journal of Epidemiology, 155 (4) (2002), pp. 332–338
- Wang et al., 2008
- Alcohol and eye diseases
- Survey of Ophthalmology, 53 (5) (2008), pp. 512–525
- Williams et al., 2004
- Acute effect of drinking red and white wines on circulating levels of inflammation-sensitive molecules in men with coronary artery disease
- Metabolism, 53 (3) (2004), pp. 318–323
- Xu et al., 2009
- Prevalence of alcohol consumption and risk of ocular diseases in a general population: The Beijing Eye Study
- Ophthalmology, 116 (10) (2009), pp. 1872–1879
- Yang et al., 2013
- Prevalence and factors associated with diabetic retinopathy in a Korean adult population: The 2008–2009 Korea National Health and Nutrition Examination Survey
- Diabetes Research and Clinical Practice, 102 (3) (2013), pp. 218–224
- Young et al., 1984
- Alcohol: Another risk factor for diabetic retinopathy?
- British Medical Journal (Clinical Research Ed.), 288 (6423) (1984), pp. 1035–1037
Copyright © 2015 Elsevier Inc. All rights reserved.
