Page 145 - 2018_11-Haematologica-web

Dietary patterns and chronic lymphocytic leukemia
Table 2. Association between adherence to dietary patterns and chronic lymphocytic leukemia by severity of the disease, in the multicase-control (MCC-Spain) study.
Rai 0
Cases N(%) OR1 (95% CI)
(n=199)
45(23) 1
Rai I-IV
Controls N(%) (n=1605)
402(25) 401 (25) 401 (25) 401 (25)
402(25) 401 (25) 401 (25) 401 (25)
402(25) 401 (25) 401 (25) 401 (25)
Cases (n=150)
26(17) 39 (26) 37 (25) 48 (32)
30(20) 37 (25) 32 (21) 51 (34)
29(19) 29 (19) 46 (31) 46 (31)
OR1 (95% CI)
P-het2
Western Q1
Q2
Q3
Q4
P-trend
1-SD increase
Prudent Q1
Q2
Q3
Q4
P-trend
1-SD increase
Mediterranean Q1
Q2
Q3
Q4
P-trend
1-SD increase
47 (24) 47 (24) 60 (30)
1.11 (0.71;1.74) 1.17 (0.73;1.87) 1.60 (0.97;2.65) 0.07
1
1.40 (0.82;2.38) 1.32 (0.76;2.30) 1.71 (0.95;3.06) 0.11
1.15 (0.95;1.39) 42(21) 1
1.26 (1.02;1.56) 0.50
1
1.16 (0.69;1.93) 0.97 (0.56;1.66) 1.47 (0.86;2.51) 0.23
1.18 (0.96;1.45) 0.17
1
0.94 (0.55;1.62) 1.39 (0.83;2.30) 1.32 (0.76;2.27) 0.17
1.15 (0.93;1.41) 0.04
53 (27) 52 (26) 52 (26)
1.09 (0.70;1.70) 1.07 (0.68;1.69) 1.01 (0.62;1.64) 0.98
0.99 (0.83;1.18) 48(24) 1
51 (26) 55 (28) 45 (23)
0.88 (0.57;1.36) 0.88 (0.57;1.37) 0.65 (0.40;1.06) 0.11
0.88 (0.74;1.04)
OR: Odds Ratio; 95% CI: 95% Confidence Interval; Q: quartile; SD: Standard Deviation. 1Multinomial logistical regression models adjusted for age, sex, education, energy intake (kcal/day) with province of residence as a random effect. 2P-value for the heterogeneity of effects. In bold: P<0.05.
more highly educated, less prone to have worked in farm- ing or agriculture, and with a higher energy intake and lower alcohol consumption. Finally, controls with a greater adherence to a Mediterranean pattern were more likely to be men, physically active, showing a lower pro- portion of ever smokers and having worked in farming or agriculture, and a higher energy intake.
Figure 1 summarizes the adjusted ORs for the associa- tion between CLL and level of adherence to the Western, Prudent and Mediterranean dietary patterns. Individuals in the highest quartile of the Western score had an OR for CLL of 1.63 (95%CI: 1.11; 2.39) compared with individu- als with low adherence (P for trend 0.02). Each SD incre- ment in the score was associated with a 19% higher OR of having CLL (95%CI: 1.03; 1.37). No associations were observed for Mediterranean and Prudent diet patterns. The impact of each individual covariate (region, age, sex, education, and energy intake) in the association of the three dietary patterns and CLL is provided in Online Supplementary Figure S2.
Since CLL is more prevalent in men, who are also more likely to adhere to a Western dietary pattern (Table 1 and Online Supplementary Figure S1), all analyses were strati- fied according to sex. No differences across sexes were observed for any of the dietary patterns [P-heterogeneity (P-het): Western (0.79), Prudent (0.11) and Mediterranean (0.17); data not shown]. In addition, no differences were observed according to BMI, energy intake, tobacco, phys- ical activity, working on a farm, and family history of
hematologic malignancies (all P for interaction >0.05; data not shown).
Analyses according to Rai-stage did not show significant heterogeneity of effects for the Western or Prudent dietary patterns (P-het=0.50 and 0.17, respectively). However, weak opposite trends in relation to a Mediterranean diet pattern were observed; it was inversely associated (although not statistically significant) with Rai 0 CLL [OR 1-SD increase= 0.88 (95%CI: 0.74; 1.04)] and positively related with Rai I-IV CLL [OR 1-SD increase= 1.15 (95%CI: 0.93; 1.41)] (P-het=0.04) (Table 2).
Sensitivity analyses according to time from diagnosis to recruitment yielded similar results for the three dietary patterns (Online Supplementary Table S2). Similarly, exclud- ing cases treated prior to consent (n=79) did not materially modify the results [P-het in trends: Western (0.25), Prudent (0.32) and Mediterranean (0.33)], but higher ORs for a Western dietary pattern were observed in cases treat- ed prior to consent in comparison to those not treated (Online Supplementary Table S3).
Discussion
This study provides, for the first time, evidence of an association between adherence to a Western dietary pat- tern and CLL. By contrast, no associations were found for a Prudent or Mediterranean pattern.
There is limited evidence linking extrinsic-risk factors,
haematologica | 2018; 103(11)
1885