In this population-based case-control study, we found no association of folate, vitamin B2
, vitamin B6
, vitamin B12
, or methionine intake with overall classical Hodgkin lymphoma risk. Although no prior studies of these factors in relation to Hodgkin lymphoma exist, 3 case-control (36
) and 2 prospective (39
) studies have examined these nutrients in relation to non-Hodgkin lymphoma risk with mixed findings. Overall, the current evidence does not support a strong association between non-Hodgkin lymphoma and one-carbon nutrients, which is in agreement with our largely null findings for Hodgkin lymphoma risk.
Because the etiology and risk factor profile of Hodgkin lymphoma may differ by EBV status, we examined the risk of each subtype separately (22
). Total vitamin B6
intakes and multivitamin use were positively associated with EBV-negative disease, but not EBV-positive disease, in our study population. These results may, however, be explained through multivitamins, because there was no association for vitamin B6
from food sources alone, and adjustment of total vitamin B6
intake for multivitamin use attenuated the associations. We also observed a slight positive association between total vitamin B2
and risk of EBV-positive disease, although this finding is based on a small number of cases and, like other observed associations, may be due to chance.
The positive association between multivitamin use and EBV-negative disease may be biased, because diet was assessed postdiagnostically, and a proportion of cases may have initiated supplement use as a result of their diagnosis (42
). However, a small number of prospective studies have also found positive associations between multivitamin use and cancer risk: breast cancer (43
), prostate cancer (44
), and non-Hodgkin lymphoma (45
). Zhang et al. (45
) reported a 48% increased risk of non-Hodgkin lymphoma among women who used multivitamins for >10 years but found no association for men. The lack of an association between multivitamins and EBV-positive disease in our study may be due to low power. Alternatively, the pathogenesis of EBV-positive Hodgkin lymphoma may be largely determined by the oncogenic properties of EBV (2
), not by dietary factors. Overall, the role of multivitamins in lymphomagenesis merits further investigation.
We did not find a link between genetic variants in key folate-metabolizing genes, including the MTHFR
677C>T and MTHFR
1298A>C polymorphisms, and Hodgkin lymphoma risk. Our null findings for the MTHFR
polymorphisms are consistent with results from 3 prior case-control studies of 25–50 cases (17
). However, Deligezer et al. (18
) reported an increased risk of Hodgkin lymphoma (n
= 51 cases) associated with the MTHFR
677 CC versus the CT/TT genotype, which contradicts the hypothesis that reduced enzymatic activity increases Hodgkin lymphoma risk (46
Inconsistencies across prior genetic studies may be due in part to inadequate statistical power or effect modification by nutrient status. In fact, we found suggestive evidence of gene-diet interactions in our study. In general, high nutrient intake or multivitamin use was associated with increased risk of Hodgkin lymphoma among carriers of the wildtype genotype, whereas the association was null or nonsignificantly inverse among variant carriers. Because the variant alleles in MTHFR
reduce enzymatic activity (46
), these results suggest that individuals with unimpaired enzyme activity and high intakes of certain one-carbon metabolism-related nutrients are at highest risk of Hodgkin lymphoma. This conclusion contradicts the hypothesized mechanism that a diet high in one-carbon metabolism-related nutrients would provide a more stable environment for DNA synthesis and methylation, thereby reducing cancer risk (1
). An alternative explanation is that reduced MTHFR
activity favors the supply of 5,10-methylenetetrahydrafolate toward DNA synthesis and repair, away from the methylation cycle of the pathway, thereby reducing the likelihood of aberrant hypermethylation of tumor suppressor genes. This pattern has been seen in colorectal cancer, where individuals carrying the MTHFR
677C>T variant allele and with adequate folate status had the lowest cancer risk (48
). We cannot, however, rule out the possibility that the gene-diet interactions we observed are chance findings, especially because none remained significant after Bonferroni correction for multiple comparisons.
Although our study population of Hodgkin lymphoma cases was one of the largest to date, statistical power was limited, especially for analyses within subgroups and tests for interactions. Response rates among controls were rather low, which could have led to selection bias, particularly if participation was related to quality of nutrition. We attempted to limit this potential selection bias by replacing nonparticipating Massachusetts controls with individuals drawn from the same residential area, and Chang et al. (29
) further showed that the income distribution across census tracts of our consented Massachusetts controls was representative of the source population. However, data to evaluate this potential source of bias were not available among Connecticut controls. Recall bias is another potential limitation, because the estimate of nutrient intake relied on the participant's ability to recall usual dietary habits in the previous year, and diet was assessed after diagnosis of Hodgkin lymphoma. Finally, our study mostly includes young adult cases and may not be representative of older-onset (>55 years) disease etiology.
Our study has several strengths. We are the first to examine the link between Hodgkin lymphoma and dietary intake of nutrients involved in one-carbon metabolism. The assessment of nutrient intake was of high quality, as measured by a validated food frequency questionnaire, and the extensive baseline questionnaire allowed us to evaluate several potential confounders. Another unique aspect of our study was the collection of both dietary and genetic information, allowing for exploratory assessment of gene-diet interactions with respect to Hodgkin lymphoma risk. Consideration of the joint role of folate-metabolizing genes and related nutrients is biologically relevant and can aid in understanding the etiology of lymphomagenesis.
Overall, our study does not support a strong link between Hodgkin lymphoma risk and dietary intake of nutrients involved in one-carbon metabolism or genetic variation in this pathway. The potential increased risk of EBV-negative Hodgkin lymphoma among multivitamin users should be explored in other studies, ideally with prospective data. Additionally, larger studies with sufficient power to detect moderate gene-diet interactions are warranted to clarify the role of one-carbon metabolism in Hodgkin lymphoma risk.