On 2014 Nov 25, Harri Hemila commented:

Virtamo J, 2014 analyzed post-trial mortality in the Alpha-Tocopherol, Beta-Carotene Cancer Prevention (ATBC) Study, which included male smokers aged 50 to 69 years at baseline. In this new post-trial analysis, they calculated that post-trial relative mortality was 1.02 (95% CI, 0.98 to 1.05) for vitamin E recipients compared with nonrecipients. Unfortunately, Virtamo et al. ignored the strong evidence that the effects of vitamin E are not uniform over the ATBC Study participants.

A previous analysis of the ATBC Study found that the effect of vitamin E on pneumonia incidence was significantly modified by three different measures of cigarette smoking exposure Hemilä H, 2004, DOI. Another analysis focusing on common cold incidence found significant modification of vitamin E effect by age, smoking and residential neighbourhood Hemilä H, 2006, DOI. The modifying factors on respiratory infections cannot be directly extrapolated to mortality, yet those findings led to an investigation on whether vitamin E effects on mortality might also be heterogeneous in the ATBC Study. During the 5-8 year intervention period of the ATBC Study, overall mortality was 2% higher in the vitamin E participants than in the no-vitamin E participants. However, if the effect of vitamin E is heterogeneous, then the 2% overall estimate might not be valid for a substantial proportion of the ATBC Study participants.

In our analysis of all-cause mortality during the intervention period of the ATBC Study, we found that the combination of age and dietary vitamin C intake modified the effect of vitamin E supplementation to the extent that the heterogeneity over 6 subgroups was highly significant (P = 0.0005) Hemilä H, 2009, DOI. In 11,448 ATBC Study participants aged 50-62 years at the baseline who had dietary vitamin C intake above the median, vitamin E increased all-cause mortality by 19%. In contrast, in 872 ATBC participants aged 66-69 years who had vitamin C intake above the median, vitamin E reduced mortality by 41%. Vitamin E had no influence on mortality among those who had dietary vitamin C intake less than the median. The modifying effect of vitamin C was not explained by other substances in fruit and vegetables. The interaction between vitamins C and E is well documented Bruno RS, 2006, DOI, which may explain the role of vitamin C as a modifying factor for some vitamin E effects.

Furthermore, in the younger ATBC Study participants (50-62 y at the baseline) who had vitamin C intake above average, vitamin E started to increase mortality only after a lag period of 3.3 years. There was no effect on mortality during the first 3.3 years of supplementation, but thereafter mortality increased in the vitamin E participants by 38% Hemilä H, 2009. The effect modification by supplementation-time is not easily explained by chance since the addition of the second vitamin E effect to start at 3.3 years improved the regression model significantly (P = 0.007). Thus, in addition to age and dietary vitamin C intake, the duration of vitamin E supplementation also modified the effect in this subgroup. The lag period may be explained by the fat solubility of vitamin E, since the body stores are changed slowly with changes in intake levels Handelman GJ, 1994, AJCN.

Finally, the decrease in mortality in participants aged 66 and over by vitamin E implies that the survival time might also be influenced. A subsequent analysis found that among ATBC participants with dietary vitamin C intakes above the median who smoked less than a pack of cigarettes per day, vitamin E extended lifespan by 2 years at the upper limit of the follow-up age span Hemilä H, 2011, DOI.

These subgroup findings invite the following questions. Given that mortality in the older ATBC Study participants was decreased by vitamin E, did the discontinuation of vitamin E at the end of the trial period lead to a corresponding increase in post-trial mortality in the oldest vitamin E recipients as compared with nonrecipients? Furthermore, given that after 3.3 years vitamin E started to increase mortality in the younger participants, did discontinuation of vitamin E lead to a corresponding decrease in post-trial mortality after a lag period? Because of the abovementioned subgroup findings, these questions are of obvious relevance and should have been investigated in the post-trial analysis of ATBC Study mortality. Virtamo et al. do not give any justification for dismissing the strong evidence of heterogeneity in the vitamin E effects on mortality.

On 2014 Nov 25, Harri Hemila commented:

Virtamo J, 2014 analyzed post-trial mortality in the Alpha-Tocopherol, Beta-Carotene Cancer Prevention (ATBC) Study, which included male smokers aged 50 to 69 years at baseline. In this new post-trial analysis, they calculated that post-trial relative mortality was 1.02 (95% CI, 0.98 to 1.05) for vitamin E recipients compared with nonrecipients. Unfortunately, Virtamo et al. ignored the strong evidence that the effects of vitamin E are not uniform over the ATBC Study participants.

A previous analysis of the ATBC Study found that the effect of vitamin E on pneumonia incidence was significantly modified by three different measures of cigarette smoking exposure Hemilä H, 2004, DOI. Another analysis focusing on common cold incidence found significant modification of vitamin E effect by age, smoking and residential neighbourhood Hemilä H, 2006, DOI. The modifying factors on respiratory infections cannot be directly extrapolated to mortality, yet those findings led to an investigation on whether vitamin E effects on mortality might also be heterogeneous in the ATBC Study. During the 5-8 year intervention period of the ATBC Study, overall mortality was 2% higher in the vitamin E participants than in the no-vitamin E participants. However, if the effect of vitamin E is heterogeneous, then the 2% overall estimate might not be valid for a substantial proportion of the ATBC Study participants.

In our analysis of all-cause mortality during the intervention period of the ATBC Study, we found that the combination of age and dietary vitamin C intake modified the effect of vitamin E supplementation to the extent that the heterogeneity over 6 subgroups was highly significant (P = 0.0005) Hemilä H, 2009, DOI. In 11,448 ATBC Study participants aged 50-62 years at the baseline who had dietary vitamin C intake above the median, vitamin E increased all-cause mortality by 19%. In contrast, in 872 ATBC participants aged 66-69 years who had vitamin C intake above the median, vitamin E reduced mortality by 41%. Vitamin E had no influence on mortality among those who had dietary vitamin C intake less than the median. The modifying effect of vitamin C was not explained by other substances in fruit and vegetables. The interaction between vitamins C and E is well documented Bruno RS, 2006, DOI, which may explain the role of vitamin C as a modifying factor for some vitamin E effects.

Furthermore, in the younger ATBC Study participants (50-62 y at the baseline) who had vitamin C intake above average, vitamin E started to increase mortality only after a lag period of 3.3 years. There was no effect on mortality during the first 3.3 years of supplementation, but thereafter mortality increased in the vitamin E participants by 38% Hemilä H, 2009. The effect modification by supplementation-time is not easily explained by chance since the addition of the second vitamin E effect to start at 3.3 years improved the regression model significantly (P = 0.007). Thus, in addition to age and dietary vitamin C intake, the duration of vitamin E supplementation also modified the effect in this subgroup. The lag period may be explained by the fat solubility of vitamin E, since the body stores are changed slowly with changes in intake levels Handelman GJ, 1994, AJCN.

Finally, the decrease in mortality in participants aged 66 and over by vitamin E implies that the survival time might also be influenced. A subsequent analysis found that among ATBC participants with dietary vitamin C intakes above the median who smoked less than a pack of cigarettes per day, vitamin E extended lifespan by 2 years at the upper limit of the follow-up age span Hemilä H, 2011, DOI.

These subgroup findings invite the following questions. Given that mortality in the older ATBC Study participants was decreased by vitamin E, did the discontinuation of vitamin E at the end of the trial period lead to a corresponding increase in post-trial mortality in the oldest vitamin E recipients as compared with nonrecipients? Furthermore, given that after 3.3 years vitamin E started to increase mortality in the younger participants, did discontinuation of vitamin E lead to a corresponding decrease in post-trial mortality after a lag period? Because of the abovementioned subgroup findings, these questions are of obvious relevance and should have been investigated in the post-trial analysis of ATBC Study mortality. Virtamo et al. do not give any justification for dismissing the strong evidence of heterogeneity in the vitamin E effects on mortality.