- Click here for explanation of classification scheme.

To assess the effect of vitamin E and B carotene on long-term cardiovascular mortality and morbidity.

Randomized control trials of antioxidant vitamins in primary and secondary prevention.

Only studies that had 1000 or more patients.

Studies done on populations from developed countries without overt evidence of vitamin deficiencies.

Trials without all-cause morality data.

Studies were identified through MEDLINE searches and searches by hand of the bibliographies of primary studies identified in the initial search. Additional searches were done of known trial acronyms cited in review articles.

Studies were included based on quality of the study’s methods, trial size, randomization scheme, and use of an intention-to-treat analysis.

Outcomes of interest prospectively identified were:

• All-cause mortality
• Cardiovascular death
• All-causeCVA
• Non-fatal MI

Findings were compared in an intention-to-treat analysis.

The frequency of each cardiovascular endpoint for patients receiving vitamin treatment and those not receiving treatment was assessed from the primary pub source.

Missing data from the original pub was obtained from the author of the pub.

Within each study a X² test was used to assess the effect of vitamin treatment, and odds ratios and 95% CI were calculated.

P,0.05 was significant.

A Breslow-Day test was used to test homogeneity of the odds ratio across trials with p,0.05 being sig.

Cochran-Mantel Haenszel tests were used to examine the effect of vitamin therapy on each endpoint across the B carotene studies and vitamin E studies separately, and pooled odds ratio and 95% CI were calculated.

12 studies were identified and included in this MA and were:

• ATBC
• PHS      
• CARET
• WHS
• HPS 
• CHAOS
• SCP
• GISSI
• AREDS
• HOPE
• NSCP 
• PPP

7 RTCs of vit E treatment were analyzed. The dose range was 50-800 IU.

n=81,788

8 RTCs of B carotene as the treatment were analyzed. The dose range was 15-50 mg.

n=138,113

Follow-up ranged from 1.4 to 12 years.

See Table 1 for sample size and characteristics of participants in each trial.

Table 1.  Sample size and characteristics of participants in each of the 12 trials analyzed .

Trial	Characteristics	B carotene	Control
ATBC	100% male smokes ages 50-69 n=29,133	14,560	14,573
CARET	Former/active smokers or asbestos exposure Ages 45-69 66% male n=18,314	9420	8894
HPS	Ages 40-80 Known vascular dz or risk for. 75% male n=20,536	10,269	10,267
SCP	Age <85 but most <65 Previous non-melanoma skin cancer 69% male n=1805	913	892
AREDS	Ages 55-80 At risk of cataracts or vision loss 44% male n=4757	2370	2387
NSCP	Ages 20-69 Risk for basal cell or squamous cell ca. 44% male n=1621	820	801
PHS	Ages 40-84 No hx ca. or vascular dz 100% male physicians n=22071	11,036	1035
WHS	Age >45 No hx cancer or vascular dz 100% female health professionals n=39,876	19,939	19,937
		Vitamin E	Control
ATBC	X age = 57 Male smokers without known lung cancer n=29,133	14,564	14,569
CHAOS	Median age 62 CAD proven w/angiography 84% male n=2002	1035	967
GISSI	Survivors of MI <3mos 85% male n=11,324	5660	5664
HOPE	X age 66 Known CV dz or diabetes 73% male n=9541	4761	4780
HPS	Ages 40-80 Known or risk for vascular dz 75% male n=20,536	10,269	10,267
AREDS	Ages 55-80 At risk of cataracts or vision loss 44% male n=4757	2370	2387
PPP	X age 64.4 Primary prevention in patients w/at least 1 risk factor 57% male n=4495	2231	2264

Vit E did not provide benefit in mortality or significantly decrease risk of cardiovascular death or CVA compared to controls.

B carotene led to a significant increase in all-cause mortality and cardiovascular death.

No sig heterogeneity was noted for any analysis.

See Table 2 for summary of statistics.

Table 2. Results Summary.

Endpoint	Odds ratio (95% CI) Pooled	Absolute event rates Treatment         Control	P value        Breslow-                               Day
All-cause mortality rate, B carotene	1.07 (1.02-1.11) n=138,113	7.4%               7.0%	0.003             0.32
Cardiovascular death rate, B carotene	1.1 (1.03-1.17) n=131,551	3.4%             3.1%	0.003              0.12
All-cause CVA,          B carotene	1.0 (0.91-1.09) n=82,483	2.3%              2.3%	0.92                0.18
All-cause mortality rate, vitamin E	1.02 (0.98-1.06) n=81,788	11.3%             11.1%	0.42               0.63
Cardiovascular death rate, vitamin E	1.0 (0.94-1.06) n=77,031	6.0%                 6.0%	0.94               0.73
All-cause CVA, vitamin E	1.02 (0.92-1.12) n=45,896	3.6%                3.5%	0.70               0.31
Cardiovascular death or non-fatal MI, vitamin E	1.0 (0.94-1.07) n=38,357(error in text which reports 38,367)	9.8%                9.8%	0.93                 0.053

Given the results of this meta analysis the use of vitamin supplements containing B carotene and vitamin A should be actively discouraged as this family of agents is associated with a small but significant excess of all-cause mortality and cardiovascular death.

Author’s recommend that clinical trials of B carotene should be discontinued because of its risks.

When used as a secondary prevention vitamin E did not reduce the risk of cardiovascular endpoints. 

Given the results of this meta analysis and the lack of mechanistic data supporting efficacy of vitamin E as a potent antioxidant in vivo, the authors do not support the continued use of vitamin E treatment or in future primary or secondary prevention trials in patients with high risk of CAD.

Not-for-profit

0 Foundation associated with industry:

Authors did excellent job discussing the limitations and reasons for their findings.

Did an excellent job pointing out potential harms and benefits of vit E and B carotene use.