We identified 991 breast cancer cases and 991 non-breast cancer controls, with a mean age of 50.0 years (SD: 8.7 years) and 49.5 years (SD: 8.7 years), respectively (Table 1). A total of 332 (33.5%), 481 (48.5%), 164 (16.5%), and 14 (1.4%) breast cancer cases had pathologically confirmed diagnosis stages of I, II, III, and IV, respectively. There were 55 (5.5%), 609 (61.5%), and 128 (12.9%) cases with tumor grades of I, II, and III, respectively; 199 cases (20.1%) did not have information on tumor grades. There were 143 (14.4%), 192 (19.3%), 155 (15.6%), and 471 (47.5%) cases with basal-like, HER2-enriched, Luminal A, and Luminal B molecular subtypes, respectively; 30 (3.0%) cases did not have completed tests for molecular subtypes. A majority of the controls (96.5%) had a benign breast lump, while other controls had either a mastitis (2.1%), a benign accessory breast lump (0.7%), hyperplasia of mammary glands (0.5%), a benign axillary lump (0.1%), or a lipoma of the breast (0.1%).
Table 1
Characteristics for cases and controls
Variable
|
Breast Cancer Cases
(N = 991)
|
Controls
(N = 991)
|
P
|
Age (years)
|
50.0 (8.7)
|
49.5 (8.7)
|
0.217
|
Body mass index (kg/m2)
|
24.3 (3.5)
|
24.2 (3.3)
|
0.553
|
Age at menarche (years)
|
15.2 (1.7)
|
15.1 (1.7)
|
0.340
|
Hypertension diagnosis (n, %)
|
110 (11.1)
|
130 (13.1)
|
0.169
|
Type 2 diabetes diagnosis (n, %)
|
38 (3.8)
|
26 (2.6)
|
0.127
|
History of cancer (n, %)
|
34 (3.4)
|
21 (2.1)
|
0.075
|
Smoker (n, %)
|
20 (2.0)
|
18 (1.8)
|
0.743
|
Alcohol consumption (n, %)
|
2 (0.2)
|
2 (0.2)
|
1.000
|
Family history of cancer (n, %)
|
44 (4.4)
|
49 (4.9)
|
0.595
|
Postmenopausal status (n, %)
|
407 (41.1)
|
355 (35.8)
|
0.016
|
Parity (n, %)
|
|
|
0.001
|
0
|
53 (5.4)
|
30 (3.0)
|
|
1
|
645 (65.1)
|
670 (67.6)
|
|
2
|
240 (24.2)
|
264 (26.6)
|
|
3+
|
53 (5.4)
|
27 (2.7)
|
|
Unless otherwise specified, variables are presented as the mean (standard deviation). |
Breast cancer cases were significantly associated with postmenopausal status and parity. Age, BMI, age at menarche, hypertension diagnosis, type 2 diabetes diagnosis, history of cancer, smoker, alcohol consumption, and family history of cancer were not significantly different between cases and controls.
Each SD increase in C3DC (OR 0.91; 95% CI 0.83-1.00), C10:1 (OR 0.87; 95% CI 0.79–0.96), and C10:2 (OR 0.90; 95% CI 0.82–0.99) level was associated with decreased odds of breast cancer (Table 2). However, higher C4 levels were associated increased risk of breast cancer (OR 1.12; 95% CI 1.02–1.23). No other carnitine compounds were associated with breast cancer. The FDRs for C3DC, C4, C10:1, and C10:2 were 0.172, 0.120, 0.064, and 0.139, respectively. The associations between C3DC, C4, C10:1, and C10:2 with breast cancer did not differ with stage of diagnosis or tumor grade (all P for trend > 0.1; Figs. 1–4). Subgroup analysis by molecular subtype suggests that the impact of C4 on breast cancer was higher for individuals with the luminal B subtype, but not for the other subtypes (P for interaction = 0.013; Fig. 1). However, we did not observe any evidence that molecular subtypes modified the associations of C4, C10:1, or C10:2 with breast cancer (Figs. 2–4).
Table 2
Multivariable logistic regression analysisa of the association between carnitine levels (per 1-SD increase) and breast cancer
Carnitine (Abbreviation)
|
Odds Ratio
(95% Confidence Interval)
|
P
|
False Discovery Rate
|
Free carnitine (C0)
|
1.05 (0.96, 1.15)
|
0.298
|
0.659
|
Acetylcarnitine (C2)
|
1.04 (0.94, 1.16)
|
0.435
|
0.659
|
Propionylcarnitine (C3)
|
1.03 (0.94, 1.14)
|
0.494
|
0.659
|
Malonylcarnitine (C3DC)
|
0.91 (0.83, 1.00)
|
0.043
|
0.172
|
Butyrylcarnitine (C4)
|
1.12 (1.02, 1.23)
|
0.015
|
0.120
|
Isovalerylcarnitine (C5)
|
0.98 (0.89, 1.08)
|
0.722
|
0.825
|
Tiglylcarnitine (C5:1)
|
0.98 (0.89, 1.08)
|
0.680
|
0.825
|
Hexanoylcarnitine (C6)
|
0.99 (0.91, 1.09)
|
0.880
|
0.880
|
Octanoylcarnitine (C8)
|
0.94 (0.86, 1.03)
|
0.196
|
0.627
|
Decanoylcarnitine (C10)
|
0.99 (0.90, 1.08)
|
0.810
|
0.864
|
Decenoylcarnitine (C10:1)
|
0.87 (0.79, 0.96)
|
0.004
|
0.064
|
Decadienoylcarnitine (C10:2)
|
0.90 (0.82, 0.99)
|
0.026
|
0.139
|
Dodecanoylcarnitine (C12)
|
1.03 (0.94, 1.13)
|
0.483
|
0.659
|
Myristoylcarnitine (C14)
|
1.04 (0.94, 1.14)
|
0.448
|
0.659
|
Palmitoylcarnitine (C16)
|
1.06 (0.95, 1.17)
|
0.288
|
0.659
|
Octadecanoylcarnitine (C18)
|
1.04 (0.95, 1.14)
|
0.372
|
0.659
|
aModels were adjusted for postmenopausal status and parity. |
In the multivariable model, C3DC and C10:1 were positively associated with greater number of parity, but not with other characteristics among controls (Additional file 1). There were no significant associations between C10:2 and baseline characteristics. C4 was only associated with age.