Effect of alpha-tocopherol and beta-carotene supplementation on macrovascular complications and total mortality from diabetes: Results of the ATBC Study

Abstract

Aim. To determine whether alpha-tocopherol or beta-carotene supplementation affects diabetic macrovascular complications and total mortality.

Methods. This study was carried out as part of the Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study, a double-blind, randomized trial with a 2×2 factorial design. A total of 29,133 middle-aged male smokers received either vitamin E 50 mg/day or beta-carotene 20 mg/day, or both, or placebo for a median of 6.1 years. At base-line, 1700 men had type 2 diabetes. Of these men, 662 were diagnosed with first-ever macrovascular complication, and 1142 died during the 19-year follow-up.

Results. Neither supplementation affected the risk of macrovascular complication or total mortality during the intervention period. For the alpha-tocopherol-supplemented versus no alpha-tocopherol-supplemented, and beta-carotene-supplemented versus no beta-carotene-supplemented we found relative risk (RR) 0.84 (95% confidence interval (CI) 0.65–1.10) and RR 1.15 (95% CI 0.89–1.50) for macrovascular complication, respectively, and RR 1.00 (95% CI 0.80–1.25) and RR 1.06 (95% CI 0.85–1.33) for total mortality, respectively. No essential changes were found in these effects when the follow-up was extended up to 19 years.

Conclusion. Alpha-tocopherol or beta-carotene supplementation has no protective effect on macrovascular outcomes or total mortality of diabetic male smokers.

Key words::

• Antioxidants
• cohort studies
• complications
• diabetes
• epidemiology
• mortality

Key messages

• Supplementation with alpha-tocopherol or beta-carotene has no protective effect on macrovascular outcomes or total mortality of diabetic male smokers.

Abbreviations
ATBC Study	=	Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study
BMI	=	body mass index
CI	=	confidence interval
HDL	=	high-density lipoprotein
HOPE Study	=	Heart Outcomes Prevention Evaluation Study
HPLC	=	high-pressure liquid chromatography
ICD	=	International Classification of Diseases
RR	=	relative risk

Introduction

The World Health Organization estimated the worldwide prevalence of diabetes in adults to be 170 million in 2000, with the number predicted to grow to 366 million by 2030 (1). Diabetic persons have an equivalent risk of death from coronary heart disease to that of non-diabetic persons with prior coronary heart disease and are up to 3.7 times more likely to suffer from stroke than are non-diabetic persons (2,3).

Hyperglycaemia may induce diabetic macrovascular complications via several mechanisms (4), including advanced glycation end-products, oxidative and carbonyl stress, increased protein kinase C activity, altered growth factor or cytokine activities, reductive stress or pseudohypoxia, and mitochondrial dysfunction (5). Through oxidative stress, diabetes increases one’s risk for atherosclerotic cardiovascular outcomes (6). Antioxidants may protect against the detrimental effects of reactive oxygen radicals, thus retarding the development of atherosclerosis (7,8). A few large controlled trials have evaluated the effect of antioxidant supplementation on diabetic complications (9–14).

Within the Alpha-Tocopherol, Beta-Carotene Cancer Prevention (ATBC) Study, we investigated whether supplementation with alpha-tocopherol or beta-carotene affects the occurrence of macrovascular complications and total mortality in diabetic male smokers.

Patients and methods

The ATBC Study

The primary aim of the ATBC Study was to examine whether supplementation with alpha-tocopherol or beta-carotene reduces the incidence of lung and other cancers. A further aim was to evaluate the effect of these supplements on all-cause mortality and cardiovascular diseases.

The study was a randomized, double-blind, placebo-controlled clinical trial with a 2×2 factorial design (15). The participants of the ATBC Study, current smokers, were screened by a postal survey sent to the entire male population of south-western Finland aged 50–69 years (n = 290,406).

Exclusion criteria were prior malignancy (other than non-melanoma skin cancer or carcinoma in situ), severe angina on exertion, chronic renal failure, cirrhosis of the liver, alcoholism, anticoagulant therapy, other medical problems that might limit long-term participation, and current use of vitamin E (over 20 mg/day), vitamin A (over 20,000 IU/day), or beta-carotene (over 6 mg/day) supplements. A total of 29,133 eligible men in blocks of 8 within each of the 14 study sites were randomly assigned to four intervention groups: alpha-tocopherol (dl-alpha-tocopheryl acetate 50 mg/d), beta-carotene (20 mg/d), or both, or placebo. Half of the participants received alpha-tocopherol (n = 14,564), and half did not (n = 14,569). Similarly, half received beta-carotene (n = 14,560), and half did not (n = 14,573). Participants and all study staff remained blinded to the participants’ treatment assignments throughout the intervention.

Enrolment took place from 1985 through 1988, and the trial intervention continued until 30 April 1993 (median intervention time 6.1 years); thereafter, the participants were followed up through national registers. The institutional review boards of the National Public Health Institute, Finland, and the US National Cancer Institute approved the ATBC study protocol. All subjects provided their written, informed consent before randomization.

Study population

We defined as diabetic subjects those who reported a history of physician-diagnosed diabetes or had elevated fasting serum glucose (≥ 7.0 mmol/L) at base-line. Fifty-four men had insulin as the only diabetes medication in their case history and were excluded as likely cases of type 1 diabetes.

Accordingly, our study population included 1700 cases representing type 2 diabetes, with 699 receiving diabetes medication already at base-line.

The men had a follow-up visit at a local field centre three times annually. During each visit they returned their remaining supplement capsules and received a new supply. Adherence was assessed by counts of the remaining capsules at each visit and by measurement of alpha-tocopherol and beta-carotene concentration after 3 years of supplementation. Supplement adherence (median 99% of on-study capsules were taken) and drop-out rate (37% during intervention phase) were similar across the four treatment groups.

Data collection

At base-line, data on subjects’ background characteristics, such as smoking, alcohol consumption, physical activity, and medical and dietary histories, were collected with questionnaires and verified by specially trained staff. We assessed the use of alcohol during the previous year by inquiries about the frequency, amount and type consumed, and calculated the average daily ethanol intake in grams. We also assessed leisure time physical activity during the past year as sedentary, moderate, or heavy, measured height and weight, and calculated body mass index (BMI; kg/m²). Blood pressure was measured with a mercury sphygmomanometer on the right arm with the participant remaining seated. We recorded the lower of two measurements taken at least 1 minute apart. A blood sample was drawn and serum stored at –70°C. Serum glucose was determined with the enzymatic hexokinase method using an Optima analyser (ThermoFischer, Vantaa, Finland). Serum total and high-density lipoprotein (HDL) cholesterol concentrations were determined enzymatically (CHODPAP method; Boehringer Mannheim, Mannheim, Germany). HDL cholesterol was measured after precipitation with dextrane sulphate and magnesium chloride. Serum concentrations of alpha-tocopherol and beta-carotene were determined using high-pressure liquid chromatography (HPLC).

End-points

The end-points of this study, macrovascular outcomes and deaths, were identified through linkage with the National Hospital Discharge Register and the Register of Causes of Death; both registers use the codes of the International Classification of Diseases (ICD).

The macrovascular study outcome included major coronary event (non-fatal acute myocardial infarction and fatal coronary heart disease, n = 302), total stroke (cerebral infarction, intracerebral haemorrhage, and subarachnoid haemorrhage, n = 204), and peripheral arterial disease (atherosclerosis of lower extremities and claudication, n = 156). Major coronary event was searched for with ICD-8 and ICD-9 code 410 (until 1996) and ICD-10 codes I21–I23 (since 1997) from the Hospital Discharge Register, and with ICD-8 and ICD-9 codes 410–414 and ICD-10 codes I20–I25 from the Register of Causes of Death. In a validity study, 94% of the major coronary event diagnoses in the registers were reviewed according to strict criteria as true major coronary events (16).

Stroke was defined based on ICD-8 or ICD-9 codes 430, 431, 433, 434, and 436 (43101 or 43191, 4330X, 4331X, 4339X, and 4349X excluded) or on ICD-10 codes I60, I61, I63, and I64. About 90% of the stroke cases were cerebral infarction. A validation study of the register diagnoses showed that, according to standard diagnostic criteria, 90% of cerebral infarctions, 79% of subarachnoid haemorrhages, and 82% of intracerebral haemorrhages were valid (17). Peripheral arterial disease was defined based on ICD-8 codes 44020 and 44390, ICD-9 code 4402A, or ICD-10 code I702.

Statistics

Participants reporting a history of myocardial infarction, angina pectoris, stroke, or claudication at base-line (n = 462) were excluded, leaving 1238 men for analysis of macrovascular outcomes. The follow-up continued from the date of randomization until the first occurrence of a macrovascular outcome, death, or until the end of the follow-up period. We analysed the macrovascular outcomes and mortality for both the intervention period (i.e. through April 1993) and the 19-year follow-up period (i.e. through December 2004). All analyses were based upon intention-to-treat. Crude rates per 1000 person-years were calculated for the outcomes in each of the four intervention groups (alpha-tocopherol, beta-carotene, alpha-tocopherol plus beta-carotene, placebo) and according to the 2×2 factorial design after testing for no interaction between alpha-tocopherol and beta-carotene (participants who received alpha-tocopherol supplements versus those who did not, and participants who received beta-carotene supplements versus those who did not). The effect of intervention was estimated with Cox proportional hazards regression and expressed as relative risk (RR) with a 95% confidence interval (CI). We tested the proportional hazards assumption with no evidence of non-proportional hazards (18). Two models were examined: crude and multivariate adjusted for age, BMI, number of cigarettes/day, smoking years, serum total and HDL cholesterol, systolic and diastolic blood pressure, alcohol consumption, and leisure time physical activity. The crude estimates are reported throughout the paper, with the multivariate estimate given only when it differs notably from the crude estimate. The modification effect was analysed by base-line body mass index (BMI ≥ 30 kg/m² versus BMI, 30 kg/m²), physical activity (moderate or heavy physical activity versus light physical activity), and hypertension (systolic blood pressure, 160 mmHg versus systolic blood pressure ≥ 160 mmHg).

All analyses were carried out with the R statistical program, and differences between intervention groups were tested using ANOVA. All P-values were two-sided; values of less than 0.05 indicated a statistical significance.

Results

The base-line characteristics of participants did not differ among the four intervention groups, except for BMI, the median of which varied from 28.2 to 29.0 (P = 0.02) (Table I). The serum concentrations of alpha-tocopherol increased from a median of 28.8 μmol/L at base-line to 42.3 μmol/L at 3 years in individuals who received alpha-tocopherol supplementation.

Table I. Base-line characteristics of diabetic participants by intervention group in the Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study.

Characteristics^a	Alpha-tocopherol	Beta-carotene	Alpha-tocopherol and beta-carotene	Placebo
Number of subject	417	434	443	406
Age (years)	57.8	58.1	57.8	57.9
Number of cigarettes/day	20	20	20	20
Smoking (years)	38	39	38	37
Total cholesterol (mmol/L)	6.13	5.93	5.97	6.11
HDL cholesterol (mmol/L)	0.98	0.99	1.00	1.01
Body mass index (kg/m^2)	28.2	28.2	28.8	29.0
Systolic blood pressure (mmHg)	146	146	148	146
Diastolic blood pressure (mmHg)	90	90	88	90
Alcohol consumption (g/day)	10	11	12	10
Physical activity, moderate or heavy (%)	50	53	53	52

^aData are medians or proportions.

Similarly, beta-carotene supplementation increased serum beta-carotene concentrations from a median of 0.24 μmol/L at base-line to 4.36 μmol/L at 3 years. In the participants receiving placebo we observed no change in serum concentrations of alpha-tocopherol and beta-carotene in 3 years.

Macrovascular complications

The incidence of total macrovascular complications in the intervention groups varied non-significantly, from 31.1 to 42.2/1000 person-years (P = 0.22) during the intervention period and from 47.5 to 53.0/1000 person years (P = 0.65) during the 19-year follow-up (Table II). When participants receiving alpha-tocopherol supplements were compared with those receiving no alpha-tocopherol supplement, and participants receiving beta-carotene supplements were compared with those receiving no beta-carotene supplement, no significant effect on the risk of total macrovascular complications was evident for the two supplementation arms during the intervention period (RR 0.84, 95% CI 0.65–1.10, and RR 1.15, 95% CI 0.89–1.50, respectively). No essential changes were found in these effects when the follow-up was extended to 19 years.

Table II. Crude relative risk (RR) with 95% confidence intervals (CIs) for macrovascular outcomes and total mortality by intervention group among diabetic participants of the Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study during the intervention period and the extended 19-year follow-up (n for macrovascular outcomes = 1238 and for total mortality = 1700).

		Intervention period				19-year follow-up
		Placebo	Alpha-tocopherol	Beta-carotene	Alpha-tocopherol and beta-carotene	Placebo	Alpha-tocopherol	Beta-carotene	Alpha-tocopherol and beta-carotene
Total macrovascular outcomes	Number	52	51	74	54	158	159	184	161
	Rate^a	31.8	31.8	42.2	31.1	47.5	52.1	53.0	48.1
	RR	1.00	1.00	1.34	0.98	1.00	1.12	1.14	1.03
	95% CI		0.68–1.49	0.93–1.92	0.66–1.44		0.90–1.41	0.91–1.41	0.82–1.29
Major coronary event	Number	30	21	41	28	85	66	85	66
	Rate^a	18.3	13.1	23.4	16.1	25.5	21.6	24.5	19.7
	RR	1.00	0.71	1.27	0.88	1.00	0.86	0.97	0.78
	95% CI		0.40–1.26	0.79–2.06	0.52–1.48		0.62–1.20	0.71–1.32	0.56–1.08
Total stroke	Number	13	16	25	18	42	50	54	58
	Rate^a	7.9	10.0	14.2	10.4	12.6	16.4	15.6	17.3
	RR	1.00	1.27	1.82	1.31	1.00	1.33	1.25	1.39
	95% CI		0.60–2.68	0.92–3.61	0.63–2.71		0.88–2.03	0.83–1.89	0.93–2.09
Peripheral arterial disease	Number	9	14	8	8	31	43	45	37
	Rate^a	5.5	8.7	4.6	4.6	9.3	14.1	13.0	11.0
	RR	1.00	1.59	0.84	0.83	1.00	1.56	1.43	1.22
	95% CI		0.67–3.73	0.32–2.21	0.31–2.20		0.98–2.50	0.89–2.27	0.75–1.99
Total mortality	Number	72	78	84	84	255	292	296	299
	Rate^a	31.1	32.6	34.5	33.3	49.8	57.4	56.6	55.4
	RR	1.00	1.04	1.11	1.06	1.00	1.17	1.16	1.13
	95% CI		0.75–1.45	0.81–1.53	0.77–1.47		0.99–1.39	0.98–1.37	0.95–1.34

^aRate per 1000 person-years.

RR = relative risk; CI = confidence interval.

The incidence of major coronary event, total stroke, and peripheral arterial disease showed no significant difference between the four intervention groups during either the intervention period (P = 0.14, 0.34, and 0.35, respectively) or the 19-year follow-up (P = 0.36, 0.44, and 0.30, respectively) (Table II). The risk of major coronary event was marginally decreased during the intervention period in those who received alpha-tocopherol compared with those who did not (RR 0.70, 95% CI 0.48–1.01). In the multivariate model, the risk decreased slightly and became significant (RR 0.64, 95% CI 0.43–0.95). During the 19-year follow-up both the crude and multivariate estimates were non-significant (RR 0.83). Beta-carotene supplementation had no effect on the risk of major coronary event. When subjects receiving alpha-tocopherol supplements were compared with those receiving no alpha-tocopherol supplement, and subjects receiving beta-carotene supplements were compared with those receiving no beta-carotene supplement, no significant effect on the risks for total stroke and peripheral arterial disease was observed (Table III).

Table III. Crude relative risk (RR) with 95% confidence intervals (CIs) for macrovascular outcomes and total mortality by alpha-tocopherol and beta-carotene supplementation among diabetic participants of the Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study during the intervention period and the extended 19-year follow-up (n for macrovascular outcomes = 1238 and for total mortality = 1700).

		Intervention period				19-year follow-up
		Alpha-tocopherol	No alpha-tocopherol	Beta-carotene	No beta-carotene	Alpha-tocopherol	No alpha-tocopherol	Beta-carotene	No beta-carotene
Macrovascular outcomes, total	Number	105	126	128	103	320	342	345	317
	Rate^a	31.4	37.2	36.7	31.8	50.0	50.3	50.6	49.7
	RR	0.84	1.00	1.15	1.00	1.00	1.00	1.02	1.00
	95% CI	0.65–1.10		0.89–1.50		0.86–1.17		0.88–1.19
Major coronary event	Number	49	71	69	51	132	170	151	151
	Rate^a	14.7	20.9	19.8	15.7	20.6	25.0	22.1	23.7
	RR	0.70	1.00	1.26	1.00	0.83	1.00	0.94	1.00
	95% CI	0.48–1.01		0.87–1.82		0.66–1.05		0.74–1.18
Total stroke	Number	34	38	43	29	108	96	112	92
	Rate^a	10.2	11.2	12.3	9.0	16.9	14.1	16.4	14.4
	RR	0.91	1.00	1.38	1.00	1.21	1.00	1.14	1.00
	95% CI	0.57–1.45		0.85–2.23		0.91–1.60		0.86–1.51
Peripheral arterial disease	Number	22	17	16	23	80	76	82	74
	Rate^a	6.6	5.0	4.6	7.1	12.5	11.2	12.0	11.6
	RR	1.30	1.00	0.65	1.00	1.14	1.00	1.05	1.00
	95% CI	0.68–2.49		0.34–1.24		0.83–1.57		0.76–1.44
Total mortality	Number	162	156	168	150	591	551	595	547
	Rate^a	32.9	32.8	33.9	31.8	56.3	53.2	56.0	53.5
	RR	1.00	1.00	1.06	1.00	1.06	1.00	1.05
	95% CI	0.80–1.25		0.85–1.33		0.95–1.20		0.93–1.18	1.00

^aRate per 1000 person-years.

RR = relative risk; CI = confidence interval.

Beta-carotene supplementation increased the risk of stroke 2-fold during the intervention period compared with beta-carotene non-supplemented participants who already at base-line had diabetes medication (RR 2.06, 95% CI 1.00–4.27) (P for interaction = 0.13). This risk was slightly attenuated when the follow-up was extended to 19 years (RR 1.89, 95% CI 1.19–3.01). Otherwise, medical treatment for diabetes at base-line did not modify the effect of beta-carotene supplementation on major coronary event or peripheral arterial disease, nor of alpha-tocopherol supplementation on any of the macrovascular complications. We observed no effect modification by BMI, physical activity, or hypertension of the impact of alpha-tocopherol or beta-carotene on macrovascular complications.

Mortality

Among the diabetic participants, 318 deaths occurred during the intervention period and 1142 deaths during the 19-year follow-up (Table II). Mortality showed no significant difference between the four intervention groups during either the intervention or the 19-year follow-up (P = 0.93 and 0.32, respectively). No significant difference in mortality during the intervention period was observed between subjects receiving alpha-tocopherol supplements and those receiving no alpha-tocopherol supplement, or between subjects receiving beta-carotene supplements and those receiving no beta-carotene supplement (RR 1.00, 95% CI 0.80–1.25, and RR 1.06, 95% CI 0.85–1.33, respectively) (Table III). Nor was any effect of alpha-tocopherol or beta-carotene on mortality found when participants with a history of cardiovascular disease at base-line were excluded. There was no essential change in these estimates when the follow-up period was extended to 19 years. No effect modification by BMI, physical activity, or hypertension was observed.

Discussion

Supplementation with alpha-tocopherol or beta-carotene had no significant protective effect on macrovascular complications or total mortality of male smokers with diabetes. The decreased risk of major coronary event in men who received alpha-tocopherol and the increased risk of stroke in men with diabetes medication who received beta-carotene may be due to chance because of many comparisons.

Our findings of no effect are in line with previous trials (Table IV). In a subgroup analysis of The Heart Outcomes Prevention Evaluation (HOPE) Study, vitamin E supplementation significantly decreased the risks of non-fatal myocardial infarction (55%) and cardiovascular death (43%) in diabetic subjects homozygous for the haptoglobin 2 allele, an inferior antioxidant compared with the haptoglobin 1 allele (19). This finding was further supported by a randomized, double-blinded trial where the hazard ratio of the primary study outcome, a composite of myocardial infarction, stroke, and cardiovascular death, was 0.47 (95% CI 0.27–0.82) for subjects randomized to vitamin E 400 U/d for up to 18 months compared with those receiving placebo (20). However, the number of study outcomes was small, 16 and 33, respectively, and thus, a substantially larger trial is needed to confirm the effect of vitamin E in diabetic patients homozygous for the haptoglobin 2 allele.

Table IV. Intervention trials that reported the effects of antioxidant supplementation on cardiovascular complications in individuals with diabetes.

Study	Country	Study type	Study population	Duration of treatment	Daily dose	Results
The Heart Outcomes Prevention Evaluation (HOPE) Study (9)	19 countries from Europe, North and South America	Randomized, double-blind, placebo-controlled	3654 men and women, aged ≥ 55 years	Mean 4.5 years	400 IU vitamin E	No effect
MRC/BHF Heart Protection Study (11)	United Kingdom	Randomized, double-blind, placebo-controlled	3982 men and women, aged 40–80 years	Mean 5 years	20 mg beta-carotene, 600 mg alphatocopherol, 250 mg vitamin C	No significant differences in outcome
Primary Prevention Project Trial (10)	Italy	Randomized, open trial	1031 men and women, aged ≥ 50 years	Median 3.7 years	300 mg vitamin E	No effect
Women's Health Study (13)	United States	Randomized, double-blind, placebo-controlled	1027 women aged ≥ 45 years	Average 10.1 years	600 IU alpha-tocopherol every other day	No effect
Women's Antioxidant Cardiovascular Study (12)	United States	Randomized, double-blind, placebo-controlled, factorial	1564 women aged ≥ 40 years	Mean 9.4 years	500 mg vitamin C, or 600 IU alphatocopherol every other day, or 50 mg beta-carotene every other day	No benefit of any of the antioxidants
Physicians Health Study II (14)	United States	Randomized, double-blind, placebo-controlled, factorial	905 men, aged ≥50 years	Mean 8 years	400 IU alpha-tocopherol (alternate days), or 500 mg vitamin C	No effect

We found that supplementation with alpha-tocopherol had no effect on total mortality, which is in line with the results of the HOPE Study, where no significant difference in total mortality was shown between the active vitamin E group and the placebo group (RR 0.93, 95% CI 0.77–1.12) (9).

In the Primary Prevention Project trial, the RR of total mortality was 1.07 (95% CI 0.61–0.90) in diabetic patients supplemented with vitamin E compared with those receiving placebo (10).

We found that beta-carotene supplementation increased its serum concentration many-fold. It is, however, unclear whether high serum beta-carotene concentration per se has untoward effects.

Intervention trials with large doses of beta-carotene have found increased risk of lung cancer in smokers and workers exposed to asbestos (21,22). However, the risk of lung cancer was not related to serum beta-carotene response to supplementation (21). Though beta-carotene has provitamin A activity, even high beta-carotene intake has not been shown to cause hypervitaminosis A (23).

Our study has several strengths. Randomization yielded groups with an even distribution of base-line characteristics, with the exception of BMI. The estimates remained unchanged, however, when the risk models were adjusted for BMI. Successful supplementation was evidenced by high capsule compliance and substantial increases in serum alpha-tocopherol and beta-carotene concentrations in the respective supplementation groups. The study had a large number of outcomes identified from national registers. In addition, the long follow-up period enabled potential late effects of alpha-tocopherol and beta-carotene on diabetic complications to be investigated.

Our study also has some limitations. The ATBC Study was designed primarily to examine the effects of vitamin E and beta-carotene on lung cancer. Moreover, it enrolled only male smokers. Our findings are therefore not generalizable to women and non-smokers. The dose, timing, and duration of the supplementation may have been less than optimal. Our study included only two antioxidant agents, and thus the possibility that other antioxidants may protect against diabetic complications cannot be excluded.

The use of antioxidants would offer an appealing solution for preventing diabetic macrovascular diseases. Our findings do not, however, support the use of alpha-tocopherol or beta-carotene supplements in the prevention of diabetic macrovascular complications.

Acknowledgements

This ATBC Study was supported by US Public Health Service contracts N01-CN-45165, N01- RC-45035, and N01-RC-37004 from the National Cancer Institute, National Institutes of Health, and the Department of Health and Human Services.

Declaration of interest: Nothing to declare.