We welcome the opportunity to respond to the comments of Tsamaloukas and the comments of Marcucci and colleagues. The primary aim of our study was to investigate the effect of vitamin supplementation on the incidence of recurrent venous thrombosis. Because factor V Leiden and prothrombin G20210A are not important risk factors for recurrent events, we did not control for these genetic factors.1  The same holds true for the MTHFR C677T polymorphism. which is not thought to influence recurrence risk independent of homocysteine levels. Moreover, in a randomized trial one assumes that confounding by extraneous risk factors is controlled by the randomization procedure. Nonetheless, we have measured these 3 polymorphisms2 ; their distribution is listed in Table 1. Adjustment for these polymorphisms did not change the hazard ratios for treatment effect substantially.

Table 1

Distribution of polymorphisms in MTHFR, FII G20210A, and factor V Leiden in the VITRO study

PolymorphismHyperhomocysteinemic group; n = 360
Normohomocysteinemic group; n = 341
Multivitamin; n = 177Placebo; n = 183Multivitamin; n = 176Placebo; n = 165
MTHFR     
    cc, no. (%) 68 (40) 80 (47) 92 (52) 80 (48) 
    ct, no. (%) 74 (44) 70 (41) 75 (43) 65 (39) 
    tt, no. (%) 28 (16) 22 (13) 9 (5) 20 (12) 
    Total no. 170 172 176 165 
FII     
    gg, no. (%) 161 (96) 160 (95) 166 (96) 155 (95) 
    ga, no. (%) 7 (4) 9 (5) 7 (4) 9 (5) 
    Total no. 168 169 173 164 
Factor V Leiden     
    gg, no. (%) 128 (81) 139 (87) 144 (84) 135 (84) 
    ga, no. (%) 30 (19) 19 (12) 28 (16) 23 (14) 
    aa, no. (%) 0 (0) 2 (1) 0 (0) 3 (2) 
    Total no. 158 160 172 161 
PolymorphismHyperhomocysteinemic group; n = 360
Normohomocysteinemic group; n = 341
Multivitamin; n = 177Placebo; n = 183Multivitamin; n = 176Placebo; n = 165
MTHFR     
    cc, no. (%) 68 (40) 80 (47) 92 (52) 80 (48) 
    ct, no. (%) 74 (44) 70 (41) 75 (43) 65 (39) 
    tt, no. (%) 28 (16) 22 (13) 9 (5) 20 (12) 
    Total no. 170 172 176 165 
FII     
    gg, no. (%) 161 (96) 160 (95) 166 (96) 155 (95) 
    ga, no. (%) 7 (4) 9 (5) 7 (4) 9 (5) 
    Total no. 168 169 173 164 
Factor V Leiden     
    gg, no. (%) 128 (81) 139 (87) 144 (84) 135 (84) 
    ga, no. (%) 30 (19) 19 (12) 28 (16) 23 (14) 
    aa, no. (%) 0 (0) 2 (1) 0 (0) 3 (2) 
    Total no. 158 160 172 161 

The prevalence of hormonal therapy use at time of the first event in hyperhomocysteinemic patients was 14% in the vitamin group and 15% in the placebo group; in the normohomocysteinemic group, it was 19% in the vitamin group and 12% in the placebo group. Adjustment for hormonal use did not change the hazard ratio for treatment effect substantially. We have no data on renal function (creatinine concentration), but there is no reason to assume any bias because of the randomized nature of the study.

We agree that the homocysteine levels in Table 1 of our paper might be somewhat puzzling. The explanation for this is that the distinction between the hyperhomocysteinemic and normohomocysteinemic groups was made on the basis of homocysteine measurements in acidic citrate tubes as stated in “Patients, materials, and methods.” If people met the other inclusion criteria, they were randomized to vitamin and placebo. Just before entering the study, another blood sample was taken in EDTA, which was used for the baseline value in the table. Because of variation in homocysteine levels, some patients in the hyperhomocysteinemic group had homocysteine levels lower than the cutoff level, and some patients in the normohomocysteinemic group had values higher than the cutoff level. In general, there is a clear contrast in homocysteine level between the hyper- and normohomocysteinemic groups (15.5 μM vs 9 μM). As for a trial in intermediate hyperhomocysteinemia, when we stratified patients in quartiles based on the homocysteine levels at baseline (just before they start using the study medication), there was no stronger effect in the highest quartile (with homocysteine levels between 14.7 and 108 μM).

Regarding the study medication, we intended to achieve a maximal homocysteine-lowering effect. The choice for the combination of folate (5 mg), vitamin B12 (0.4 mg), and vitamin B6(50 mg) was based on an earlier study3  in which this regimen showed the strongest effect on homocysteine levels; this dose is relatively high compared with several other homocysteine-lowering trials. We are aware that vitamin B2 is a coenzyme of MTHFR, and that riboflavin supplementation lowers homocysteine levels in vitamin B2–deficient patients.4  However, no data are available on the additional effect of riboflavin next to the combination of folate, vitamin B12, and vitamin B6. Only 1 study shows a small additional effect of riboflavin next to folate alone.5 

Correspondence: Martin den Heijer, Department of Endocrinology (471), Radboud University, Nijmegen Medical Centre, PO Box 9101, 6500 HB; e-mail: [email protected].

Conflict-of-interest disclosure: The authors declare no competing financial interests.

1
Christiansen
 
SC
Cannegieter
 
SC
Koster
 
T
Vandenbroucke
 
JP
Rosendaal
 
FR
Thrombophilia, clinical factors, and recurrent venous thrombotic events.
JAMA
2005
, vol. 
293
 (pg. 
2352
-
2361
)
2
Keijzer
 
MB
Borm
 
GF
Blom
 
HJ
Bos
 
GM
Rosendaal
 
FR
den Heijer
 
M
No interaction between factor V Leiden and hyperhomocysteinemia or MTHFR 677TT genotype in venous thrombosis: results of a meta-analysis of published studies and a large case-only study.
Thromb Haemost
2007
, vol. 
97
 (pg. 
32
-
37
)
3
den Heijer
 
M
Brouwer
 
IA
Bos
 
GM
et al. 
Vitamin supplementation reduces blood homocysteine levels: a controlled trial in patients with venous thrombosis and healthy volunteers.
Arterioscler Thromb Vasc Biol
1998
, vol. 
18
 (pg. 
356
-
361
)
4
McNulty
 
H
Dowey le
 
RC
Strain
 
JJ
et al. 
Riboflavin lowers homocysteine in individuals homozygous for the MTHFR 677C → T polymorphism.
Circulation
2006
, vol. 
113
 (pg. 
74
-
80
)
5
Chuang
 
CZ
Boyles
 
A
Legardeur
 
B
Su
 
J
Japa
 
S
Lopez-S
 
A
Effects of riboflavin and folic acid supplementation on plasma homocysteine levels in healthy subjects.
Am J Med Sci
2006
, vol. 
331
 (pg. 
65
-
71
)
Sign in via your Institution