ABSTRACT
Background and Aims
Bilirubin, an antioxidant and anti‐inflammatory agent, has been shown to negatively correlate with adiposity. However, the extent to which adiposity level influences total bilirubin (TBili) concentration remains unclear. We sought to evaluate the associations of body mass index (BMI; kg/m²) and waist circumference (WC) with low TBili concentration.
Methods
We included adults ≥ 20 years of age without elevated liver enzymes, excess alcohol use, iron overload, anemia, TBili ≥ 1.2 mg/dL, low albumin, or positive hepatitis serology from the 1999–2018 National Health and Nutrition Examination Survey. The primary outcome was low TBili, defined as < 0.4 mg/dL. We used multivariable logistic regression to determine the odds ratios of low TBili by BMI and WC categories, stratified by gender. We additionally explored associations between past 1‐year, maximum lifetime, and non‐linear continuous BMI on low TBili.
Results
Among 13,953 participants (mean age 47.3 years, 49% men, 69% non‐Hispanic White), 17% were ≥ 65 years, 32% had BMI ≥ 30, and 12% had low TBili. Women with a BMI ≥ 27.5 had at least a 42% increased odds of low TBili compared to women with a BMI ≥ 23 and < 25. Women with a normal BMI but elevated WC of ≥ 88 cm had increased odds of low TBili (OR: 2.74 [1.61–4.66]) than those with a normal WC. A BMI 1 year prior to survey ≥ 35 (OR: 1.59 [1.05–2.40]) and lifetime maximum BMI ≥ 25 (OR: 1.60 [1.13–2.28]) were independently associated with low TBili in women. Similar point estimates, but no significant associations were found in men.
Conclusion
Elevated BMI and an elevated WC but normal BMI were associated with higher likelihood of low TBili in women, potentially reflecting a consequence or indirect biomarker of clinical obesity. Further studies should determine the long‐term impact of low TBili on cardiometabolic disease risk and mortality.
Keywords: adiposity, bilirubin, metabolic syndrome, National Health and Nutrition Examination Survey, obesity
Abbreviations
- ACC
American College of Cardiology
- AHA
American Heart Association
- BMI
body mass index (indices)
- BRI
body roundness index
- HbA1c
Hemoglobin A1c
- NHANES
National Health and Nutrition Examination Survey
- TBili
total bilirubin
- UGT1A1
uridine diphosphate glucuronosyltransferase 1A1
- WC
waist circumference
1. Introduction
Obesity remains a public health challenge and affects nearly 40% of U.S. adults using traditional body mass index (BMI) thresholds, and up to nearly 70% [1] per the recent Lancet Commission on Obesity definition [2]. Obesity is characterized by chronic low grade inflammation and oxidative stress [3]. While obesity is well established to have a causal link with metabolic illness, little is known about its effects on other biomarkers. One such routinely measured marker is total bilirubin (TBili), a byproduct of heme catabolism [4], that acts as an endogenous antioxidant and anti‐inflammatory agent [5, 6]. It acts through peroxisome proliferator‐activated receptor α nuclear receptors [7, 8, 9, 10], triggering enhanced mitochondrial function and β‐oxidation [7]. Through its downstream effects, TBili has been associated with metabolic function [8, 10].
While hyperbilirubinemia is often pathologic, mildly elevated TBili, especially its unconjugated form, has been associated with a lower prevalence of oxidative stress‐mediated diseases [5]. For example, TBili has been negatively associated with cardiovascular disease [5, 11, 12, 13, 14], metabolic syndrome [15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26], and diabetes mellitus [23, 27].
Although TBili concentration was negatively correlated with adiposity in patients with obesity [28], it is unclear whether there is an association between adiposity and low TBili across all adiposity levels. Accordingly, we aimed to determine whether BMI was associated with low TBili concentration, whether waist circumference (WC) modified this association, and explore effects of past and lifetime maximum BMIs.
2. Methods
2.1. Dataset
Every 2 years, the National Center for Health Statistics conducts the National Health and Nutrition Examination Survey (NHANES), a nationally representative survey of the civilian, non‐institutionalized United States population using a complex, stratified, multistage probability design. The survey consists of interviews, physical examinations at home or at a mobile examination center, and laboratory testing. We used ten data cycles from 1999 to 2018. Additional details regarding the dataset and the sampling design can be found elsewhere [29].
2.2. Ethical Approval and Consent to Participate
Given that the NHANES dataset is publicly available and all data have been anonymized, ethical approval was not required and was considered non‐human subjects research [30].
2.3. Study Population
We included non‐pregnant adults ≥ 20 years of age with data on BMI, WC, TBili concentration, and the covariates of interest described below. Participants with the following were excluded: elevated liver enzymes (ALT > 80 IU/L and AST > 112 IU/L) [31], excess alcohol consumption (> 2 drinks per day for men and > 1 drink per day for women) [32], iron overload (transferrin > 20%) [33], anemia (hemoglobin < 9 g/dl), TBili ≥ 1.2 mg/dL [34], albumin < 4.0 g/dL [35], or positive hepatitis B and/or C serology. These exclusion criteria were adopted to restrict to participants without liver disease, as liver disease could pathologically affect bilirubin levels and confound results.
2.4. BMI and WC Classification
Despite its limitations [36], we used BMI due to its pervasiveness in clinical settings. To mitigate some of the epidemiologic biases inherent to relying solely on current BMI, particularly illness‐related weight loss, reverse causality, and measurement error, we incorporated three distinct BMI measures: (1) current BMI, measured at the time of survey; (2) past BMI, defined as the BMI 1 year prior using self‐reported weight; and (3) lifetime maximum BMI, based on self‐reported highest weight attained in adulthood. The latter two measures have been shown to better capture adiposity exposure and latent metabolic risk profiles.
Incorporating past BMI introduces a limited temporal dimension into the cross‐sectional NHANES framework and reduces bias from weight loss due to undiagnosed or subclinical disease. Similarly, lifetime maximum BMI provides a marker of cumulative adiposity burden and has been associated with future chronic disease risk. By leveraging past and maximum BMI values, we aimed to better characterize the long‐term adiposity–bilirubin relationship and reduce misclassification bias that may arise when relying on cross‐sectional exposure–outcome associations [37].
BMI (kg/m2) was separated into seven categories: < 18.5, ≥ 18.5 and < 23, ≥ 23 and < 25, ≥ 25 and < 27.5, ≥ 27.5 and < 30, ≥ 30 and < 35, and ≥ 35. These categorizations were chosen based on other observational BMI studies [38] and ethnicity‐specific World Health Organization BMI categorizations [39]. A BMI ≥ 23 and < 25 kg/m2 was used as the reference as has been done in prior studies to avoid reverse causality in lower BMI categories that can occur with illness‐related weight loss. WC was considered elevated based on the National Cholesterol Education Program Adult Treatment Panel III recommended standard cutoffs of > 88 cm for women and > 102 cm for men [40]. Five BMI categories (≥ 23 and < 25, ≥ 25 and < 27.5, ≥ 27.5 and < 30, ≥ 30 and < 35, and ≥ 35 kg/m2) were stratified by WC to assess the impact of abdominal obesity on TBili concentration. The reference was BMI ≥ 23 and < 25 kg/m2 with a normal WC. Further references to BMI in the text are presented without the corresponding units.
2.5. Covariates of Interest and Outcome Measures
Covariates of interest were selected a priori based on prior studies and clinical knowledge and are listed in Tables 1 and 2. Gender (self‐reported, used as a proxy for sex assigned at birth) and primary race/ethnicity (Mexican American, Other Hispanic, non‐Hispanic White, non‐Hispanic Black, or Other [Asian, American Indian, Alaskan Native, Multiracial, Other]) were self‐reported. Smoking history was divided into ever‐smoker and non‐smoker. Self‐reported vigorous physical activity was dichotomized as any or none. Diabetes mellitus was defined as a hemoglobin A1c (HbA1c) ≥ 6.5%, fasting blood glucose > 125 mg/dL, self‐reported diabetes, or use of antiglycemic medications for this indication. Prognostic nutritional index was determined from the following formula: 10 x albumin (g/dL) + 0.005 x total lymphocyte count (per mm3) [41, 42]. Functional limitation was based on reported difficulty in six domains: seeing, hearing, mobility, communication, cognition, and self‐care [43]. Each covariate's questionnaire protocols and quality control procedures are described elsewhere [44]. The primary outcome measure was low TBili concentration, defined as < 0.4 mg/dL, which has been associated with a higher risk of certain diseases despite being in the physiological range [4].
Table 1.
Baseline characteristics for men according to body mass index.
| Body mass index (kg/m2) | Total | |||||||
|---|---|---|---|---|---|---|---|---|
| < 18.5 | ≥ 18.5 and < 23 | ≥ 23 and < 25 | ≥ 25 and < 27.5 | ≥ 27.5 and < 30 | ≥ 30 and < 35 | ≥ 35 | ||
| N | 72 | 954 | 933 | 1401 | 1281 | 1418 | 779 | 6838 |
| Age | ||||||||
| 20−34 | 30 (47.3) | 379 (44.1) | 281 (34.7) | 306 (23.1) | 261 (23.8) | 303 (23.2) | 202 (26.1) | 1762 (28.3) |
| 35−49 | 10 (19.3) | 201 (23.4) | 263 (31.1) | 411 (34.1) | 369 (35.5) | 433 (34.2) | 251 (35.4) | 1938 (32.5) |
| 50−64 | 16 (23.5) | 184 (20.4) | 185 (20.1) | 367 (26.7) | 349 (26.6) | 391 (28.7) | 201 (28.0) | 1693 (25.5) |
| ≥ 65 | 16 (9.9) | 190 (12.1) | 204 (14.2) | 317 (16.0) | 302 (14.1) | 291 (13.9) | 125 (10.5) | 1445 (13.7) |
| Age at screening | 40.5 (2.6) | 41.6 (0.8) | 43.8 (0.6) | 47.6 (0.5) | 47.2 (0.5) | 47.0 (0.5) | 45.3 (0.6) | — |
| Race/ethnicity | ||||||||
| Mexican American | 3 (2.0) | 98 (5.7) | 108 (5.0) | 228 (8.4) | 230 (8.2) | 243 (8.8) | 98 (6.6) | 1008 (7.3) |
| Other Hispanic | 4 (2.6) | 46 (3.6) | 72 (4.6) | 107 (5.0) | 123 (6.4) | 126 (5.3) | 65 (5.2) | 543 (5.1) |
| Non‐Hispanic White | 38 (74.4) | 422 (66.2) | 430 (70.3) | 626 (68.6) | 587 (70.7) | 670 (72.0) | 358 (68.3) | 3131 (69.7) |
| Non‐Hispanic Black | 16 (10.0) | 233 (13.3) | 177 (9.0) | 252 (8.8) | 221 (8.5) | 268 (8.6) | 213 (13.8) | 1380 (9.9) |
| Other | 11 (11.0) | 155 (11.2) | 146 (11.0) | 188 (9.2) | 120 (6.2) | 111 (5.4) | 45 (6.0) | 776 (8.0) |
| Income (tertile) | ||||||||
| Lowest | 32 (34.8) | 375 (28.8) | 319 (24.7) | 440 (21.4) | 431 (23.0) | 453 (20.6) | 246 (23.6) | 2296 (23.4) |
| Middle | 29 (43.5) | 302 (30.6) | 308 (29.9) | 469 (29.1) | 434 (31.8) | 512 (34.9) | 295 (35.4) | 2349 (32.0) |
| Highest | 11 (21.7) | 277 (40.6) | 306 (45.3) | 492 (49.5) | 416 (45.2) | 453 (44.4) | 238 (41.1) | 2193 (44.6) |
| Education | ||||||||
| More than high school | 32 (54.9) | 504 (62.5) | 540 (66.8) | 767 (65.9) | 686 (62.0) | 734 (60.3) | 411 (58.8) | 3674 (62.7) |
| High school | 21 (24.0) | 218 (21.2) | 176 (18.7) | 275 (18.9) | 278 (22.4) | 326 (24.6) | 197 (25.3) | 1491 (21.9) |
| Less than high school | 19 (21.1) | 232 (16.3) | 217 (14.5) | 359 (15.2) | 317 (15.6) | 358 (15.1) | 171 (15.9) | 1673 (15.5) |
| Marital status | ||||||||
| Married | 23 (37.4) | 444 (48.3) | 551 (61.7) | 881 (68.2) | 827 (70.1) | 901 (68.1) | 476 (65.3) | 4103 (64.3) |
| Never married | 21 (27.8) | 289 (32.1) | 189 (22.7) | 210 (14.2) | 169 (13.3) | 204 (14.2) | 143 (18.4) | 1225 (18.2) |
| Other | 11 (17.1) | 80 (8.9) | 62 (5.6) | 113 (8.1) | 90 (6.8) | 110 (5.9) | 57 (6.6) | 523 (7.1) |
| Separated | 15 (17.7) | 132 (10.6) | 113 (9.9) | 180 (9.5) | 166 (9.8) | 188 (11.8) | 93 (9.7) | 887 (10.4) |
| Ever smoker | 45 (57.4) | 502 (48.8) | 423 (40.8) | 642 (43.3) | 617 (42.4) | 710 (48.4) | 339 (42.9) | 3278 (44.7) |
| Vigorous physical activity | 17 (40.8) | 326 (49.9) | 347 (47.0) | 481 (48.5) | 384 (42.0) | 393 (36.9) | 183 (28.9) | 2131 (42.6) |
| Insurance | 55 (81.9) | 715 (78.4) | 707 (79.7) | 1084 (83.2) | 986 (82.0) | 1114 (82.1) | 606 (81.2) | 5267 (81.4) |
| Citizenship | ||||||||
| Citizen | 63 (92.4) | 794 (88.8) | 769 (88.6) | 1124 (86.4) | 1 058 (88.5) | 1216 (91.2) | 724 (95.5) | 5748 (89.5) |
| Not a citizen | 9 (7.6) | 157 (11.1) | 163 (11.3) | 274 (13.4) | 221 (11.5) | 200 (8.6) | 54 (4.4) | 1078 (10.4) |
| Body composition | ||||||||
| Body mass index (kg/m2) | 17.6 (0.1) | 21.3 (0.1) | 24.1 (0.02) | 26.2 (0.02) | 28.7 (0.03) | 32.1 (0.1) | 39.8 (0.2) | — |
| Waist circumference (cm) | 71.9 (0.5) | 81.8 (0.3) | 89.3 (0.2) | 95.4 (0.2) | 102.0 (0.2) | 110.5 (0.3) | 127.8 (0.6) | — |
| Vital signs | ||||||||
| Heart rate (bpm) | 69.0 (2.2) | 69.1 (0.5) | 68.9 (0.6) | 68.5 (0.4) | 69.1 (0.4) | 71.7 (0.4) | 75.8 (0.6) | — |
| Systolic blood pressure (mmHg) | 121.2 (2.5) | 118.8 (0.6) | 120.3 (0.6) | 121.6 (0.5) | 123.2 (0.4) | 124.2 (0.5) | 128.1 (0.7) | — |
| Diastolic blood pressure (mmHg) | 70.1 (1.5) | 69.9 (0.5) | 71.2 (0.5) | 72.4 (0.4) | 73.8 (0.3) | 75.0 (0.4) | 77.1 (0.6) | — |
| Comorbidities | ||||||||
| Hypertension (ACC/AHA stage 2) | 5 (2.5) | 26 (1.5) | 31 (2.4) | 67 (2.9) | 69 (2.7) | 63 (3.1) | 50 (5.8) | 311 (3.0) |
| Diabetes mellitus | 2 (2.9) | 57 (3.8) | 63 (4.2) | 139 (5.5) | 170 (8.9) | 218 (12.1) | 173 (18.6) | 822 (8.6) |
| HbA1c (%) | 5.3 (0.1) | 5.4 (0.02) | 5.4 (0.03) | 5.4 (0.02) | 5.5 (0.02) | 5.7 (0.04) | 5.8 (0.04) | — |
| Low HDL cholesterol | 8 (17.1) | 151 (16.4) | 215 (22.4) | 426 (29.6) | 492 (39.9) | 615 (44.0) | 399 (54.4) | 2306 (34.5) |
| Direct HDL cholesterol (mg/dL) | 55.7 (2.3) | 55.0 (0.7) | 51.2 (0.6) | 48.0 (0.4) | 45.2 (0.5) | 42.7 (0.4) | 40.7 (0.4) | — |
| Elevated triglycerides | 7 (13.8) | 184 (19.6) | 301 (30.5) | 559 (38.9) | 569 (43.7) | 721 (52.5) | 404 (54.6) | 2745 (40.5) |
| Microalbuminuria | 5 (4.3) | 31 (4.0) | 39 (4.0) | 51 (4.1) | 57 (4.2) | 92 (8.5) | 63 (13.1) | 338 (6.1) |
| Chronic kidney disease | 2 (1.5) | 41 (2.9) | 58 (4.4) | 107 (5.8) | 107 (6.3) | 105 (5.6) | 43 (5.0) | 463 (5.1) |
| eGFR | 99.4 (2.8) | 93.4 (0.8) | 89.8 (0.8) | 85.9 (0.7) | 85.6 (0.6) | 86.3 (0.6) | 91.4 (0.9) | — |
| Cardiovascular disease | 7 (4.0) | 56 (3.5) | 45 (2.9) | 88 (4.0) | 91 (5.1) | 124 (5.9) | 65 (6.7) | 476 (4.7) |
| Heart failure | 2 (1.5) | 13 (0.9) | 16 (1.1) | 18 (0.7) | 24 (1.4) | 41 (1.6) | 24 (2.0) | 138 (1.3) |
| Coronary artery disease | 3 (1.9) | 22 (1.3) | 16 (1.3) | 27 (1.4) | 37 (2.2) | 44 (2.2) | 18 (1.5) | 167 (1.7) |
| Myocardial infarction | 4 (2.9) | 28 (1.7) | 23 (1.6) | 42 (2.5) | 36 (1.8) | 59 (3.1) | 25 (2.6) | 217 (2.3) |
| Stroke | 2 (0.7) | 25 (1.3) | 18 (1.4) | 33 (1.1) | 25 (1.6) | 44 (1.8) | 24 (2.3) | 171 (1.6) |
| Thyroid disease | 0 | 14 (0.7) | 21 (2.4) | 30 (1.8) | 32 (2.5) | 44 (3.3) | 28 (3.5) | 169 (2.4) |
| Cancer | 3 (3.4) | 73 (7.1) | 71 (6.9) | 122 (8.0) | 108 (8.1) | 110 (7.4) | 51 (6.8) | 538 (7.4) |
| Chronic obstructive pulmonary disease | 6 (4.5) | 49 (4.0) | 42 (4.1) | 63 (4.7) | 61 (4.9) | 76 (5.0) | 64 (7.4) | 361 (4.9) |
| Asthma | 6 (9.9) | 46 (4.5) | 34 (3.6) | 54 (3.0) | 55 (4.8) | 77 (6.1) | 64 (7.2) | 336 (4.9) |
| Liver function and enzymes | ||||||||
| ALT (IU/L) | 18.1 (1.1) | 21.5 (0.4) | 23.8 (0.4) | 26.6 (0.4) | 27.2 (0.4) | 31.6 (0.5) | 34.5 (0.6) | — |
| AST (IU/L) | 22.3 (0.7) | 24.2 (0.4) | 24.4 (0.3) | 25.2 (0.3) | 24.7 (0.2) | 26.0 (0.3) | 27.1 (0.4) | — |
| Albumin (g/dL) | 4.6 (0.1) | 4.5 (0.01) | 4.5 (0.02) | 4.4 (0.01) | 4.4 (0.01) | 4.3 (0.01) | 4.2 (0.01) | — |
| Total serum bilirubin (mg/dL) | 0.7 (0.04) | 0.7 (0.01) | 0.69 (0.01) | 0.71 (0.01) | 0.68 (0.01) | 0.66 (0.01) | 0.63 (0.01) | — |
| Hemoglobin (g/dL) | 15.1 (0.1) | 15.0 (0.04) | 15.1 (0.04) | 15.2 (0.04) | 15.2 (0.03) | 15.3 (0.03) | 15.3 (0.1) | — |
| Prognostic nutritional index | 20 (17.4) | 244 (20.8) | 262 (23.2) | 436 (28.4) | 448 (30.0) | 582 (35.1) | 438 (53.3) | 2 430 (31.1) |
| Functional limitation | 11 (13.4) | 98 (7.8) | 79 (5.7) | 117 (5.9) | 114 (5.7) | 162 (7.9) | 111 (11.5) | 692 (7.2) |
Note: Baseline characteristics for men categorized by body mass index. Categorical variables are presented as n (weighted percent) while continuous variables are given as mean (standard error of the mean). Percentages were obtained from taking the n for each characteristic in each BMI category and dividing it by the total number of participants in each BMI category.
Abbreviations: ACC, American College of Cardiology; AHA, American Heart Association; bpm, beats per minute; eGFR, estimated glomerular filtration rate; HDL, high‐density lipoprotein; HbA1c, hemoglobin A1c.
Table 2.
Baseline characteristics for women according to body mass index.
| Body mass index (kg/m2) | Total | |||||||
|---|---|---|---|---|---|---|---|---|
| < 18.5 | ≥ 18.5 and < 23 | ≥ 23 and < 25 | ≥ 25 and < 27.5 | ≥ 27.5 and < 30 | ≥ 30 and < 35 | ≥ 35 | ||
| N | 148 | 1261 | 842 | 1099 | 1002 | 1448 | 1315 | 7115 |
| Age | ||||||||
| 20–34 | 63 (38.7) | 399 (29.4) | 206 (22.9) | 243 (21.6) | 187 (19.5) | 295 (20.0) | 281 (22.2) | 1674 (23.3) |
| 35–49 | 37 (28.3) | 355 (31.6) | 254 (33.2) | 288 (27.7) | 248 (27.0) | 381 (27.9) | 400 (32.8) | 1963 (30.0) |
| 50–64 | 21 (15.9) | 264 (23.4) | 204 (26.7) | 297 (28.4) | 288 (28.0) | 406 (28.1) | 404 (31.2) | 1884 (27.3) |
| ≥ 65 | 27 (17.1) | 243 (15.6) | 178 (17.1) | 271 (22.4) | 279 (25.6) | 366 (23.9) | 230 (13.8) | 1594 (19.4) |
| Age at screening | 43.8 (1.7) | 45.8 (0.5) | 48.3 (0.7) | 49.9 (0.7) | 51.6 (0.7) | 50.7 (0.6) | 47.9 (0.5) | — |
| Race/ethnicity | ||||||||
| Mexican American | 8 (1.8) | 127 (4.3) | 107 (5.7) | 223 (8.6) | 229 (9.5) | 316 (10.1) | 253 (9.6) | 1263 (7.8) |
| Other Hispanic | 2 (0.8) | 95 (4.2) | 93 (5.8) | 148 (8.2) | 115 (7.0) | 167 (6.4) | 131 (5.8) | 751 (6.0) |
| Non‐Hispanic White | 82 (76.7) | 639 (72.8) | 397 (71.4) | 456 (68.2) | 427 (69.1) | 556 (65.3) | 516 (64.9) | 3073 (68.8) |
| Non‐Hispanic Black | 18 (6.2) | 118 (4.6) | 112 (7.2) | 169 (8.1) | 155 (9.0) | 314 (12.9) | 359 (15.7) | 1245 (9.5) |
| Other | 38 (14.5) | 282 (14.1) | 133 (9.9) | 103 (6.9) | 76 (5.4) | 95 (5.3) | 56 (3.9) | 783 (7.9) |
| Income (tertile) | ||||||||
| Lowest | 55 (29.8) | 399 (23.9) | 287 (26.3) | 423 (27.1) | 413 (29.6) | 589 (29.9) | 572 (32.0) | 2738 (28.1) |
| Middle | 50 (35.3) | 401 (29.1) | 267 (28.0) | 364 (32.0) | 347 (34.1) | 532 (38.9) | 472 (37.1) | 2433 (33.4) |
| Highest | 43 (34.9) | 461 (47.0) | 288 (45.7) | 312 (40.9) | 242 (36.3) | 327 (31.2) | 271 (30.8) | 1944 (38.5) |
| Education | ||||||||
| More than high school | 84 (58.2) | 837 (72.8) | 502 (67.3) | 575 (61.9) | 478 (56.9) | 679 (54.1) | 633 (54.9) | 3788 (61.4) |
| High school | 33 (22.0) | 228 (16.7) | 172 (20.2) | 226 (21.0) | 220 (23.4) | 327 (26.5) | 312 (27.8) | 1518 (22.6) |
| Less than high school | 31 (19.8) | 196 (10.5) | 168 (12.5) | 298 (17.1) | 304 (19.7) | 442 (19.4) | 370 (17.3) | 1809 (16.0) |
| Marital status | ||||||||
| Married | 72 (56.2) | 685 (61.1) | 494 (64.8) | 569 (59.2) | 537 (59.9) | 712 (56.1) | 641 (56.1) | 3710 (59.2) |
| Never married | 35 (20.3) | 228 (16.7) | 91 (9.5) | 137 (10.6) | 104 (10.0) | 192 (12.6) | 216 (15.0) | 1003 (13.0) |
| Other | 6 (3.4) | 73 (4.9) | 51 (6.6) | 82 (6.3) | 69 (6.0) | 98 (6.1) | 87 (6.1) | 466 (5.90) |
| Separated | 32 (20.1) | 259 (17.3) | 194 (19.1) | 300 (23.9) | 284 (24.0) | 414 (25.2) | 353 (22.7) | 1836 (21.9) |
| Ever smoker | 61 (41.7) | 347 (30.8) | 229 (28.9) | 302 (29.6) | 271 (31.3) | 424 (30.9) | 453 (36.0) | 2087 (31.6) |
| Vigorous physical activity | 32 (34.2) | 383 (47.1) | 222 (39.0) | 204 (28.6) | 174 (25.8) | 229 (23.7) | 173 (20.8) | 1417 (31.7) |
| Insurance | 125 (86.8) | 1048 (86.6) | 710 (87.3) | 861 (85.2) | 789 (85.0) | 1139 (83.9) | 1041 (83.4) | 5713 (85.2) |
| Citizenship | ||||||||
| Citizen | 126 (90.4) | 1043 (90.1) | 685 (88.5) | 887 (89.4) | 814 (88.1) | 1202 (90.2) | 1167 (94.2) | 5924 (90.3) |
| Not a citizen | 22 (9.6) | 215 (9.8) | 153 (11.2) | 208 (10.4) | 186 (11.8) | 240 (9.6) | 147 (5.7) | 1171 (9.6) |
| Body composition | ||||||||
| Body mass index (kg/m2) | 17.5 (0.1) | 21.1 (0.04) | 24.0 (0.02) | 26.2 (0.03) | 28.7 (0.03) | 32.3 (0.05) | 40.8 (0.2) | — |
| Waist circumference (cm) | 69.6 (0.5) | 77.7 (0.2) | 85.2 (0.3) | 90.5 (0.3) | 96.7 (0.3) | 104.3 (0.3) | 119.9 (0.4) | — |
| Vital signs | ||||||||
| Heart rate (bpm) | 73.4 (1.1) | 72.4 (0.4) | 72.5 (0.5) | 73.0 (0.4) | 74.1 (0.6) | 74.0 (0.4) | 76.5 (0.4) | — |
| Systolic blood pressure (mmHg) | 112.4 (1.6) | 115.1 (0.6) | 118.7 (0.8) | 120.7 (0.8) | 122.5 (0.7) | 123.4 (0.6) | 124.8 (0.6) | — |
| Diastolic blood pressure (mmHg) | 68.0 (1.1) | 68.7 (0.3) | 69.4 (0.4) | 70.2 (0.4) | 69.8 (0.5) | 71.3 (0.4) | 72.4 (0.4) | — |
| Comorbidities | ||||||||
| Hypertension (ACC/AHA stage 2) | 8 (4.2) | 57 (3.5) | 41 (3.6) | 73 (5.4) | 59 (5.0) | 66 (3.8) | 76 (4.7) | 380 (4.3) |
| Diabetes mellitus | 1 (0.9) | 36 (1.4) | 45 (3.6) | 78 (5.4) | 122 (8.9) | 240 (11.3) | 290 (18.2) | 812 (8.0) |
| HbA1c (%) | 5.3 (0.04) | 5.3 (0.02) | 5.4 (0.02) | 5.5 (0.02) | 5.6 (0.03) | 5.7 (0.03) | 5.8 (0.03) | — |
| Low HDL cholesterol | 18 (10.5) | 251 (17.6) | 244 (26.3) | 423 (35.8) | 420 (39.8) | 712 (46.9) | 734 (57.0) | 2802 (36.5) |
| Direct HDL cholesterol (mg/dL) | 66.8 (1.4) | 65.4 (0.6) | 61.2 (0.6) | 59.2 (0.6) | 55.5 (0.6) | 53.2 (0.6) | 50.0 (0.5) | — |
| Elevated triglycerides | 10 (6.5) | 207 (16.8) | 195 (20.9) | 323 (26.6) | 389 (37.1) | 577 (39.8) | 536 (42.3) | 2237 (29.9) |
| Microalbuminuria | 10 (9.1) | 71 (8.2) | 38 (5.7) | 70 (8.6) | 60 (7.7) | 92 (8.9) | 104 (11.2) | 445 (8.6) |
| Chronic kidney disease | 12 (8.5) | 89 (7.7) | 69 (8.9) | 110 (9.8) | 116 (13.0) | 158 (10.7) | 115 (9.0) | 669 (9.7) |
| eGFR | 90.7 (2.0) | 88.2 (0.7) | 85.9 (0.8) | 87.3 (0.9) | 84.7 (0.9) | 87.5 (0.8) | 88.1 (0.8) | — |
| Cardiovascular disease | 8 (5.1) | 46 (2.7) | 44 (3.6) | 61 (4.7) | 56 (5.2) | 82 (5.1) | 101 (6.7) | 398 (4.6) |
| Heart failure | 2 (1.1) | 10 (0.7) | 13 (1.2) | 12 (1.0) | 14 (1.1) | 26 (1.6) | 41 (2.9) | 118 (1.4) |
| Coronary artery disease | 2 (1.1) | 8 (0.5) | 13 (1.0) | 11 (0.9) | 11 (1.0) | 14 (1.0) | 18 (1.1) | 77 (0.9) |
| Myocardial infarction | 3 (1.4) | 15 (0.7) | 17 (1.7) | 19 (1.4) | 13 (1.2) | 29 (1.9) | 40 (2.7) | 136 (1.6) |
| Stroke | 4 (3.3) | 24 (1.3) | 19 (1.6) | 34 (2.5) | 28 (2.5) | 39 (2.5) | 36 (2.4) | 184 (2.1) |
| Thyroid disease | 11 (9.0) | 100 (9.2) | 65 (7.6) | 104 (10.8) | 118 (13.7) | 142 (12.3) | 158 (14.1) | 698 (11.3) |
| Cancer | 18 (16.4) | 103 (10.0) | 78 (10.3) | 111 (11.0) | 98 (11.5) | 140 (10.3) | 127 (10.1) | 675 (10.6) |
| Chronic obstructive pulmonary disease | 16 (12.8) | 79 (6.2) | 46 (5.0) | 76 (8.0) | 69 (8.4) | 101 (9.3) | 155 (12.4) | 542 (8.4) |
| Asthma | 9 (8.3) | 70 (5.7) | 61 (7.6) | 74 (8.5) | 70 (6.7) | 128 (10.6) | 174 (12.7) | 586 (8.7) |
| Liver function and enzymes | ||||||||
| ALT (IU/L) | 17.8 (0.8) | 18.4 (0.3) | 19.0 (0.4) | 19.7 (0.4) | 20.6 (0.4) | 21.9 (0.3) | 22.7 (0.4) | — |
| AST (IU/L) | 22.9 (0.6) | 22.5 (0.2) | 22.1 (0.3) | 22.3 (0.2) | 22.2 (0.2) | 22.3 (0.2) | 22.1 (0.3) | — |
| Albumin (g/dL) | 4.4 (0.03) | 4.3 (0.01) | 4.3 (0.01) | 4.3 (0.01) | 4.2 (0.01) | 4.2 (0.01) | 4.1 (0.01) | — |
| Total serum bilirubin (mg/dL) | 0.59 (0.03) | 0.61 (0.01) | 0.59 (0.01) | 0.57 (0.01) | 0.55 (0.01) | 0.53 (0.01) | 0.51 (0.01) | — |
| Hemoglobin (g/dL) | 13.6 (0.08) | 13.6 (0.03) | 13.6 (0.04) | 13.6 (0.04) | 13.7 (0.04) | 13.6 (0.03) | 13.7 (0.03) | — |
| Prognostic nutritional index | 43 (22.4) | 523 (40.1) | 372 (40.7) | 575 (48.9) | 589 (56.5) | 933 (62.7) | 1 027 (74.2) | 4062 (53.4) |
| Functional limitation | 25 (14.8) | 100 (5.8) | 62 (7.3) | 91 (7.1) | 111 (11.1) | 184 (10.9) | 236 (15.7) | 809 (9.8) |
Note: Baseline characteristics for women categorized by body mass index. Categorical variables are presented as n (weighted percent) while continuous variables are given as mean (standard error of the mean).
Abbreviations: ACC, American College of Cardiology; AHA, American Heart Association; bpm, beats per minute; eGFR, estimated glomerular filtration rate; HDL, high‐density lipoprotein; HbA1c, hemoglobin A1c.
2.6. Statistical Analysis
Baseline characteristics according to BMI and gender are described. Analyses were stratified by gender due to known differences in TBili based on sex [4, 45]. Categorical variables are presented as n (weighted percent), and continuous variables are presented as weighted mean (standard error). Using univariate and multivariable Cox proportional hazard regression, we assessed the association between current, past, and lifetime maximum BMI with low TBili. The proportionality assumption was verified using Schoenfeld residuals. Models were stratified by gender and adjusted for age, income‐poverty ratio, education, smoking status, race/ethnicity, low high‐density lipoprotein cholesterol, HbA1c, mean systolic blood pressure, presence of cardiovascular disease, ALT, AST, albumin, and hemoglobin. We used similar analyses to determine the association between current BMI and WC with low TBili. These data are presented as adjusted odds ratios (95% confidence interval). As a sensitivity analysis, we assessed the non‐linear continuous association between current BMI and low TBili concentration using natural cubic splines with five knots (5th, 25th, 50th, 75th, and 95th percentiles). Using the same knots, we additionally examined non‐linear associations between body roundness index (BRI), a mathematical calculation assessing sphericity of body shape and a comparatively better indicator of poor body composition than BMI, and low TBili concentration. BRI was calculated as 364.2−365.5 × {1 − [(WC/2π)/(0.5*height)]2}0.5 [46]. Analyses accounted for the complex survey design, including post‐stratification weights, strata, and clustering. SAS 9.4 (Cary, NC) was used for the analyses. A p‐value of < 0.05 indicated statistical significance.
3. Results
3.1. Baseline Characteristics
There were 13,953 participants (mean age [standard deviation]: 47.3 [17] years, 49% men, 69% non‐Hispanic White). Among men (Table 1), the mean BMI ranged from 17.6 (standard error: 0.1) to 39.8 (0.2) in the lowest to highest BMI categories, respectively. There were 311 (3.0%) participants with American College of Cardiology (ACC)/American Heart Association (AHA) stage 2 hypertension, and participants with the highest BMI had the highest mean systolic and diastolic blood pressures. Nearly 19% of men with BMI ≥ 35 had diabetes. The proportion of adults with low high‐density lipoprotein cholesterol and high triglycerides increased from approximately 14% to 55% in the lowest to highest BMI groups. Mean albumin was ≥ 4.2 g/dL in all categories while mean TBili ranged from 0.63 (0.01) mg/dL in participants with a BMI ≥ 35 to 0.71 (0.01) mg/dL in participants with a BMI ≥ 25 and < 27.5.
Among women (Table 2), BMI ranged from a mean of 17.5 (0.1) in the BMI < 18.5 group to a mean of 40.8 (0.2) in the BMI ≥ 35 group. There were 380 (4.3%) participants with ACC/AHA stage 2 hypertension. Diabetes prevalence increased from 1.4% to nearly 19% with increasing BMI group. Low high‐density lipoprotein cholesterol and high triglycerides was prevalent in 55% of women with BMI ≥ 35. The mean albumin in each BMI category was ≥ 4.1 g/dL. TBili was lower in women than in men: the mean TBili concentration ranged from 0.51 (0.01) mg/dL at BMI ≥ 35 to 0.61 (0.01) mg/dL at BMI ≥ 18.5 and < 23.
3.2. Current BMI and Low Tbili Concentration
For men, univariate analysis showed a significant association between BMI and TBili when BMI was ≥ 30 (Table 3) with at least a 64% increased odds of low TBili concentration; this association became insignificant with multivariable analysis (Table 4). Among women, BMI and low TBili were associated when BMI was ≥ 27.5 with univariate analysis (Table 3). This association remained significant on multivariable analysis: BMI of ≥ 27.5 and < 30 (OR: 1.42 [1.01–2.01]), BMI of ≥ 30 and < 35 (OR: 1.74 [1.20–2.51]), and BMI ≥ 35 (OR: 1.64 [1.09–2.47]) (Table 4 and Figure 1).
Table 3.
Univariate analysis for odds of low total bilirubin based on past, current, and lifetime maximum body mass index, stratified by gender.
| Body mass index | ||||||
|---|---|---|---|---|---|---|
| Past BMI | Current BMI | Lifetime maximum BMI | ||||
| Men | Women | Men | Women | Men | Women | |
| BMI (kg/m2) | ||||||
| < 18.5 | 1.65 (0.90–3.05) | 0.94 (0.53–1.66) | 2.12 (0.97–4.63) | 1.44 (0.62–3.31) | 0.88 (0.11–7.39) | 0.97 (0.17–5.41) |
| ≥ 18.5 and < 23 | 1.14 (0.80–1.62) | 0.95 (0.64–1.41) | 1.13 (0.83–1.55) | 0.92 (0.63–1.33) | 1.27 (0.80–1.99) | 1.19 (0.74–1.93) |
| ≥ 23 and < 25 | Reference | Reference | Reference | Reference | Reference | Reference |
| ≥ 25 and < 27.5 | 1.03 (0.75–1.42) | 1.33 (0.87–2.03) | 1.11 (0.78–1.59) | 1.39 (0.96–2.00) | 1.22 (0.82–1.81) | 1.58 (1.13–2.19) |
| ≥ 27.5 and < 30 | 1.25 (0.87–1.80) | 1.54 (1.08–2.18) | 1.16 (0.82–1.65) | 1.58 (1.13–2.23) | 1.35 (0.90–2.01) | 1.53 (1.02–2.29) |
| ≥ 30 and < 35 | 1.47 (1.09–1.98) | 1.74 (1.20–2.52) | 1.64 (1.13–2.39) | 2.11 (1.51–2.96) | 1.58 (1.09–2.28) | 2.01 (1.45–2.79) |
| ≥ 35 | 1.87 (1.26–2.77) | 2.23 (1.53–3.26) | 1.92 (1.36–2.72) | 2.38 (1.64–3.46) | 1.90 (1.30–2.77) | 2.66 (1.89–3.76) |
Note: Univariate analysis of odds (95% confidence interval) of low total bilirubin according to past (one year prior based on self‐reported weight), current, and lifetime maximum body mass index, further stratified according to gender.
Abbreviation: BMI, body mass index.
Table 4.
Multivariable analysis for odds of low total bilirubin based on past, current, and lifetime maximum body mass index, stratified by gender.
| Body mass index | ||||||
|---|---|---|---|---|---|---|
| Past BMI | Current BMI | Lifetime maximum BMI | ||||
| Men | Women | Men | Women | Men | Women | |
| BMI (kg/m2) | ||||||
| < 18.5 | 1.36 (0.66–2.80) | 0.91 (0.51–1.61) | 1.98 (0.87–4.48) | 1.76 (0.72–4.32) | 0.80 (0.08–7.85) | 1.11 (0.21–5.87) |
| ≥ 18.5 and < 23 | 1.10 (0.78–1.56) | 1.01 (0.68–1.50) | 1.10 (0.82–1.49) | 0.96 (0.66–1.41) | 1.22 (0.76–1.96) | 1.33 (0.80–2.22) |
| ≥ 23 and < 25 | Reference | Reference | Reference | Reference | Reference | Reference |
| ≥ 25 and < 27.5 | 1.01 (0.73–1.40) | 1.30 (0.83–2.04) | 1.13 (0.79–1.62) | 1.33 (0.91–1.93) | 1.16 (0.77–1.76) | 1.60 (1.13–2.28) |
| ≥ 27.5 and < 30 | 1.15 (0.81–1.66) | 1.36 (0.95–1.94) | 1.07 (0.76–1.51) | 1.42 (1.01–2.01) | 1.25 (0.83–1.91) | 1.47 (0.98–2.22) |
| ≥ 30 and < 35 | 1.26 (0.93–1.71) | 1.40 (0.94–2.09) | 1.45 (0.99–2.12) | 1.74 (1.20–2.51) | 1.35 (0.92–1.97) | 1.70 (1.22–2.37) |
| ≥ 35 | 1.25 (0.81–1.94) | 1.59 (1.05–2.40) | 1.35 (0.89–2.06) | 1.64 (1.09–2.47) | 1.30 (0.85–1.99) | 1.94 (1.36–2.79) |
Note: Multivariable analysis of odds (95% confidence interval) of low total bilirubin according to past (one year prior based on self‐reported weight), current, and lifetime maximum body mass index, further stratified according to gender. Analysis was adjusted for the following: age, income‐poverty ratio, education, smoking status, race/ethnicity, low high‐density lipoprotein cholesterol, hemoglobin A1c, mean systolic blood pressure, presence of cardiovascular disease, ALT, AST, albumin, and hemoglobin.
Abbreviations: BMI: body mass index.
Figure 1.
Odds of low total bilirubin concentration in women. Forest plot of adjusted odds ratios of low total bilirubin concentration according to body mass index (BMI). The reference was BMI of ≥ 23 and < 25 kg/m2. All BMIs are in kg/m2 and presented without their units. Squares indicate past BMI. Triangles indicate current BMI, and diamonds represent lifetime maximum BMI.
Given changes in sex hormone and synthetic liver function with age, we explored associations by age subgroups (< 55 years and ≥ 55 years). There was no significant interaction between BMI category and age on odds of low TBili (p = 0.25). Non‐linear continuous associations are illustrated in Supplemental Figure 1.
3.3. Low Tbili Concentration According to BMI and WC
Men with a BMI ≥ 30 had increased odds of having low TBili regardless of WC on univariate analysis, but no associations were found on multivariable analysis (Table 5). In women, multivariable analysis showed increased odds of low TBili concentration with an elevated WC at normal or low overweight BMIs. An elevated WC conferred 2.74 times increased odds of low TBili (OR: 2.74 [1.61–4.66]) at a BMI of ≥ 23 and < 25. At a BMI of ≥ 25 and < 27.5, the odds ratios were 1.90 (1.00–3.60) for normal WC and 2.02 (1.27–3.24) for elevated WC compared to the reference (Table 5 and Figure 2).
Table 5.
Odds of low total bilirubin according to body mass index and waist circumference, stratified by gender.
| Body mass index (kg/m2) | Waist circumference | Univariate analysis | Multivariable analysis | ||
|---|---|---|---|---|---|
| Men | Women | Men | Women | ||
| < 23 | Normal | 1.18 (0.87–1.60) | 1.39 (0.89–2.19) | 1.15 (0.86–1.54) | 1.49 (0.94–2.35) |
| Elevated | — | 2.22 (0.87‐5.68) | — | 2.30 (0.87‐6.09) | |
| ≥ 23 and < 25 | Normal | Reference | Reference | Reference | Reference |
| Elevated | 0.50 (0.09–2.88) | 2.73 (1.67–4.46) | 0.52 (0.09–2.96) | 2.74 (1.61–4.66) | |
| ≥ 25 and < 27.5 | Normal | 1.15 (0.80–1.66) | 1.97 (1.05–3.70) | 1.16 (0.80–1.68) | 1.90 (1.00–3.60) |
| Elevated | 0.79 (0.44–1.43) | 2.08 (1.32–3.26) | 0.83 (0.44–1.56) | 2.02 (1.27–3.24) | |
| ≥ 27.5 and < 30 | Normal | 1.21 (0.80–1.82) | 3.56 (1.90–6.66) | 1.11 (0.74–1.67) | 3.01 (1.61–5.63) |
| Elevated | 1.10 (0.70–1.72) | 2.25 (1.49–3.40) | 1.01 (0.64–1.58) | 2.07 (1.35–3.19) | |
| ≥ 30 and < 35 | Normal | 1.87 (1.06–3.30) | — | 1.78 (0.95–3.33) | — |
| Elevated | 1.60 (1.07–2.38) | 3.13 (2.07–4.75) | 1.39 (0.92–2.09) | 2.64 (1.68–4.15) | |
| ≥ 35 | — | 1.90 (1.34–2.69) | 3.51 (2.28–5.40) | 1.33 (0.87–2.03) | 2.48 (1.54–3.99) |
Note: Odds (95% confidence interval) of low total bilirubin according to body mass index and waist circumference, stratified by gender. Waist circumference was defined as elevated using the standard National Cholesterol Education Program Adult Treatment Panel III cutoff of > 88 cm for women and > 102 cm for men [40]. Multivariable analysis was performed adjusting for age, income‐poverty ratio, education, smoking status, race/ethnicity, low high‐density lipoprotein cholesterol, hemoglobin A1c, mean systolic blood pressure, presence of cardiovascular disease, ALT, AST, albumin, and hemoglobin. Odds ratios indicated by—were due to low sample sizes.
Figure 2.
Odds of low total bilirubin concentration in women according to body mass index and waist circumference. Forest plot of adjusted odds ratio of low total bilirubin concentration according to body mass index (BMI) and waist circumference (WC). The key odds ratios are presented in the figure. BMI ≥ 23 and < 25 kg/m2 with a normal WC was the reference. An elevated WC was defined according to the National Cholesterol Education Program Adult Treatment Panel III recommended standard cutoffs of > 88 cm for women [40]. All BMIs are in kg/m2 and presented without their units.
3.4. Past and Lifetime Maximum BMI and Low Tbili
The majority of participants either maintained (stayed in the same BMI category) or had a modest increase in weight (moved 1 BMI category higher) in the past 1 year (Supplemental Tables 1 and 2). For men, a past and lifetime maximum BMI of ≥ 30 was associated with at least a 47% increased odds of having low TBili concentration on univariate analysis. Similar to findings with current BMI, no significant associations were found when adjusted for confounders. For women, there were progressively increased odds of having low TBili when past BMI was ≥ 27.5. However, only a past BMI ≥ 35 was associated with a low TBili concentration (OR: 1.59 [1.05–2.40]) on multivariable analysis. A lifetime maximum BMI ≥ 25 was associated with low TBili in women in univariate analysis: BMI of ≥ 25 and < 27.5 (OR: 1.58 [1.13–2.19]), BMI of ≥ 27.5 and < 30 (OR: 1.53 [1.02–2.29]), BMI of ≥ 30 and < 35 (OR: 2.01 [1.45–2.79]), and BMI ≥ 35 (OR: 2.66 [1.89–3.76]). From these categories, only a lifetime maximum BMI of ≥ 27.5 and < 30 was insignificant on multivariable analysis.
3.5. Sensitivity Analysis
To examine the robustness of the associations between BMI and low TBili, we treated BMI as a non‐linear continuous variable. Using BMI of 23 as a reference, women were noted to have significantly increased odds of low TBili concentration with increased BMI, whereas men had no significant increase in odds of low TBili until BMI > 35 (Supplemental Figure 2). Similar patterns were observed for BRI (Supplemental Figure 3).
4. Discussion
We found that a BMI ≥ 27.5 was associated with prevalent low TBili concentration in women. Despite a normal or low overweight BMI, an elevated WC conferred increased odds of low TBili in women, a reflection of the impact of central adiposity on bilirubin metabolism. A past BMI ≥ 35 and lifetime maximum BMI ≥ 25, not including BMI ≥ 27.5 and < 30, were also associated with low TBili concentrations in women. No significant associations were found in men. Sensitivity analysis confirmed the robustness of our findings.
While previous studies have also demonstrated an inverse association between bilirubin and metabolic disease, our study robustly examines the converse association between adiposity and low TBili and corroborates those reports. However, a positive relationship between adiposity, its distribution, and TBili concentration found only in women contrasts with previous studies. A cross‐sectional study with over 460,000 participants showed more favorable anthropometric measures, including fat mass and body fat distribution, in both men and women with higher bilirubin concentrations [12]. Another study found a negative correlation between serum bilirubin concentration and BMI and trunk fat only in men [47]. Among participants with obesity, bilirubin concentration was negatively correlated with BMI and adiposity compared to lean participants with stronger associations encountered in males [28]. Our findings may be due to the higher prevalence of low TBili concentration in women [4], driven by polymorphisms in uridine diphosphate gluconosyltransferase 1A1 (UGT1A1), the enzyme responsible for converting bilirubin to conjugated bilirubin, and the impact of sex hormones on its expression [45]. We speculate it may also be the result of possible outcome misclassification as men tend to accumulate android/visceral adiposity earlier, develop hepatic steatosis, and thus may have pathologic elevations in TBili that are treated as non‐cases. Future studies should identify sex‐specific cutoffs of TBili for increased risk of cardiometabolic disease [48].
An elevated WC conferred increased odds of low TBili in women despite a normal BMI. Several studies have found higher TBili concentration to be negatively correlated or associated with WC and/or central obesity [12, 15, 16, 18, 19, 20, 24, 25]. Additionally, BRI, which may more accurately reflect visceral adiposity [49], was associated with increased odds of low TBili concentration in our study. These findings reaffirm the limitations of BMI in metabolic disease prediction and suggest an important role of body composition, independent of BMI, on low TBili concentration [23, 50].
While most patients maintained the same BMI category, there was a trend towards an increase in BMI from past to current as well as lifetime maximum to current BMI of 0.4 and 2 kg/m2, respectively (data not shown). Including past and lifetime maximum BMI provides additional methodologic rigor to account for epidemiological biases such as reverse causality due to misclassification from illness‐related weight loss. Similar associations seen amongst all three BMI types, but with stronger associations seen with maximum lifetime BMI, suggests a robust association likely in the direction of high BMI/adiposity driving changes in TBili. Small prior studies have indirectly shown that weight loss [48, 51], including gastric bypass [52], and aerobic exercise [53] can increase TBili concentration. It remains to be seen whether TBili changes precede or are a consequence of development of clinical manifestations of obesity (e.g. development of hepatic steatosis or diabetes). Given that these findings were only found in women in our study, the differential impact of internal and external factors on bilirubin metabolism based on sex need further exploration.
Further studies are needed to elucidate whether the relationships found clinically are merely associative or causal. UGT1A1 levels have been shown to increase in obesity [54, 55], ultimately leading to the conversion of TBili into urobilin through multiple steps [55]. TBili and urobilin have negative and positive correlations with adiposity measures, respectively [55]. UGT1A1 suppression increases TBili and lowers urobilin [56], which can be further modified by exercise, as seen with experimental high aerobic exercise capacity mice [55]. Experimental studies have shown that TBili reduces body weight through actions on peroxisome proliferator‐activated receptor α [7, 9, 57] and decreases visceral obesity and insulin resistance by suppressing inflammatory cytokines [58]. In fact, treating mice fed a high‐fat diet with biliverdin stabilized adipocyte size [58].
While these underlying mechanisms suggest a causal process and multiple observational studies show a strong link between bilirubin concentration and BMI, Mendelian randomization studies have failed to suggest causality. Although mild hyperbilirubinemia in Gilbert syndrome was associated with a lower BMI compared to healthy controls in one observational study [59], no association between BMI or WC and the UGT1A1 locus was found with Mendelian randomization [60]. The relationship may be more nuanced than initially anticipated. A study in the pediatric population showed that the UGT1A1 polymorphism and body weight were significantly associated with bilirubin concentration in the control group, whereas among children and adolescents with obesity, the UGT1A1 polymorphism and body fat percentage were the primary determinants of bilirubin concentration [61].
This extends beyond BMIs and bilirubin concentrations to clinical implications. Serum bilirubin concentration has been associated with cardiovascular disease [11, 51], yet Mendelian randomization studies have not found a causal link between genetically elevated bilirubin and decreased risk of ischemic heart disease or myocardial infarction [62]. However, a causal link has been suggested between bilirubin and type 2 diabetes mellitus [63]. Thus, the dynamic interplay between obesity, bilirubin, and cardiometabolic diseases is yet to be determined, especially in the context of the obesity paradox [3]. Further studies should focus on the long‐term impacts of low TBili on outcomes such as major adverse cardiovascular events and cardiovascular and all‐cause mortality while determining whether there are causal effects or whether the associations represent underlying cardiometabolic changes.
We used a population‐based, nationally representative sample to examine the associations between excess adiposity and its distribution with low TBili concentration, stratified by gender. However, there are limitations. The cross‐sectional nature precluded longitudinal assessment, though we attempted to explore temporality by analyzing the relationship using past and lifetime maximum BMIs. We used BMI to assess the degree of adiposity, but an inability to distinguish between muscle and fat mass makes it an imperfect measure. Nonetheless, we provided a more detailed breakdown of BMI and included WC and BRI. There is a potential for reporting bias given that past BMI was based on self‐reported weight from 1 year prior. The low number of participants in the low BMI category limits interpretation in this category. Additionally, the smaller sample of participants who identified as other than Non‐Hispanic White limits generalizability. As with all observational studies, there were likely unmeasured confounders, which could have influenced the results. However, we excluded patients with TBili ≥ 1.2 mg/dL to minimize other etiologies of elevated TBili. Finally, we did not use fractionated TBili, a relevant distinction given that most of TBili's antioxidant effects come from unconjugated bilirubin [18, 19].
A current BMI ≥ 27.5, past BMI ≥ 35, and lifetime maximum BMI ≥ 25 were associated with low current TBili concentration in women but not men. Central adiposity was associated with low TBili in women independent of BMI, highlighting the role of fat distribution in metabolic processes. Future studies should focus on 1) understanding the temporality of this association, especially as it relates to development of metabolic consequences of obesity; 2) the long‐term impacts of low TBili, and 3) the differential effects of obesity on TBili in men and women. Low TBili may potentially serve as a marker reflecting poor cardiometabolic health in patients with overweight, obesity, or central adiposity, particularly in women.
Author Contributions
Trishna Parikh: writing – original draft, writing – review and editing, visualization. Fariha Hameed: writing – original draft, writing – review and editing. Sumaiya Islam: writing – original draft, writing – review and editing, visualization. Shivani Reddy: writing – original draft, writing – review and editing, visualization. Eshani Goradia: writing – review and editing. Bhargav Vemulapalli: writing – review and editing. Aayush Visaria: methodology, validation, investigation, formal analysis, writing – review and editing, visualization, supervision.
Funding
The authors received no specific funding for this work.
Conflicts of Interest
The authors declare no conflicts of interest.
1. Transparency Statement
Dr. Aayush Visaria affirms that this manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned have been explained.
Supporting information
Supplemental Figure 1: Association between body mass index as a continuous variable with low current total bilirubin according to age groups. Panel A describes the association in the overall population for those aged < 55 years while panel B shows the associations in those ≥ 55 years. The reference point was a body mass index of 23 kg/m2. BMI: body mass index.
Supplemental Figure 2: Association between body mass index as a continuous non‐linear variable and odds of low current total bilirubin concentration. Panel A describes the association in the overall population while panels B and C describe the associations in men and women, respectively. Natural cubic splines with five traditional knots were used. The reference point was a body mass index of 23 kg/m2. BMI: body mass index.
Supplemental Figure 3: Association between body roundness index with low current total bilirubin concentration. Panel A describes the association in the overall population, while panels B and C describe the association in men and women, respectively. Natural cubic splines with five traditional knots were used. The reference used was a body roundness index of 3.5.
Supplemental Table 1: Cross‐tabulation of past and current body mass index in men. Supplemental Table 2: Cross‐tabulation of past and current body mass index in women.
Parikh T., Hameed F., Islam S., et al., “Association between adiposity and low total serum bilirubin concentration: A retrospective, cross‐sectional study,” Health Science Reports 9 (2025): 1‐13, 10.1002/hsr2.71700.
Trishna Parikh and Fariha Hameed co‐first authorship.
Data Availability Statement
The data supporting the findings of this study are available upon reasonable request from the corresponding author.
References
- 1. Fourman L. T., Awwad A., Gutiérrez‐Sacristán A., et al., “Implications of a New Obesity Definition Among the All of us Cohort,” JAMA Network Open 8, no. 10 (2025): e2537619. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Rubino F., Cummings D. E., Eckel R. H., et al., “Definition and Diagnostic Criteria of Clinical Obesity,” Lancet Diabetes & Endocrinology 13, no. 3 (2025): 221–262. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Powell‐Wiley T. M., Poirier P., Burke L. E., et al., “Obesity and Cardiovascular Disease: A Scientific Statement From the American Heart Association,” Circulation 143, no. 21 (2021): e984–e1010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Vítek L. and Tiribelli C., “Bilirubin: The Yellow Hormone?,” Journal of Hepatology 75, no. 6 (2021): 1485–1490. [DOI] [PubMed] [Google Scholar]
- 5. Vítek L., “The Role of Bilirubin in Diabetes, Metabolic Syndrome, and Cardiovascular Diseases,” Frontiers in Pharmacology 3 (2012): 55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Vitek L., Bellarosa C., and Tiribelli C., “Induction of Mild Hyperbilirubinemia: Hype or Real Therapeutic Opportunity?,” Clinical Pharmacology & Therapeutics 106, no. 3 (2019): 568–575. [DOI] [PubMed] [Google Scholar]
- 7. Gordon D. M., Blomquist T. M., Miruzzi S. A., McCullumsmith R., Stec D. E., and Hinds T. D., “RNA Sequencing in Human HepG2 Hepatocytes Reveals PPAR‐α Mediates Transcriptome Responsiveness of Bilirubin,” Physiological Genomics 51, no. 6 (2019): 234–240. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Gordon D. M., Neifer K. L., Hamoud A. R. A., et al., “Bilirubin Remodels Murine White Adipose Tissue by Reshaping Mitochondrial Activity and the Coregulator Profile of Peroxisome Proliferator‐Activated Receptor Α,” Journal of Biological Chemistry 295, no. 29 (2020): 9804–9822. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Gordon D. M., Hong S. H., Kipp Z. A., and Hinds T. D., “Identification of Binding Regions of Bilirubin in the Ligand‐Binding Pocket of the Peroxisome Proliferator‐Activated Receptor‐A (PPARalpha),” Molecules 26, no. 10 (2021): 2975. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Stec D. E., John K., Trabbic C. J., et al., “Bilirubin Binding to PPARα Inhibits Lipid Accumulation,” PLoS One 11, no. 4 (2016): e0153427. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Kang S. J., Kim D., Park H. E., et al., “Elevated Serum Bilirubin Levels Are Inversely Associated With Coronary Artery Atherosclerosis,” Atherosclerosis 230, no. 2 (2013): 242–248. [DOI] [PubMed] [Google Scholar]
- 12. Seyed Khoei N., Wagner K. H., Sedlmeier A. M., Gunter M. J., Murphy N., and Freisling H., “Bilirubin as an Indicator of Cardiometabolic Health: A Cross‐Sectional Analysis in the UK Biobank,” Cardiovascular Diabetology 21, no. 1 (2022): 54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Song Y. S., Koo B. K., Cho N. H., and Moon M. K., “Effect of Low Serum Total Bilirubin Levels (≤0.32 mg/dl) on Risk of Coronary Artery Disease in Patients With Metabolic Syndrome,” American Journal of Cardiology 114, no. 11 (2014): 1695–1700. [DOI] [PubMed] [Google Scholar]
- 14. Vitek L., Hinds T. D., Stec D. E., and Tiribelli C., “The Physiology of Bilirubin: Health and Disease Equilibrium,” Trends in Molecular Medicine 29, no. 4 (2023): 315–328. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Choi S. H., Yun K. E., and Choi H. J., “Relationships Between Serum Total Bilirubin Levels and Metabolic Syndrome in Korean Adults,” Nutrition, Metabolism, and Cardiovascular Diseases 23, no. 1 (2013): 31–37. [DOI] [PubMed] [Google Scholar]
- 16. Giral P., Ratziu V., Couvert P., et al., “Plasma Bilirubin and Gamma‐Glutamyltransferase Activity Are Inversely Related in Dyslipidemic Patients With Metabolic Syndrome: Relevance to Oxidative Stress,” Atherosclerosis 210, no. 2 (2010): 607–613. [DOI] [PubMed] [Google Scholar]
- 17. Huang S. S., Chan W. L., Leu H. B., Huang P. H., Lin S. J., and Chen J. W., “Serum Bilirubin Levels Predict Future Development of Metabolic Syndrome in Healthy Middle‐Aged Nonsmoking Men,” American Journal of Medicine 128, no. 10 (2015): 1138.e35–1138.e41. [DOI] [PubMed] [Google Scholar]
- 18. Hwang H. J. and Kim S. H., “Inverse Relationship Between Fasting Direct Bilirubin and Metabolic Syndrome in Korean Adults,” Clinica Chimica Acta 411, no. 19–20 (2010): 1496–1501. [DOI] [PubMed] [Google Scholar]
- 19. Jo J., Yun J. E., Lee H., Kimm H., and Jee S. H., “Total, Direct, and Indirect Serum Bilirubin Concentrations and Metabolic Syndrome Among the Korean Population,” Endocrine 39, no. 2 (2011): 182–189. [DOI] [PubMed] [Google Scholar]
- 20. Kwon K. M., Kam J. H., Kim M. Y., et al., “Inverse Association Between Total Bilirubin and Metabolic Syndrome in Rural Korean Women,” Journal of Women's Health 20, no. 6 (2011): 963–969. [DOI] [PubMed] [Google Scholar]
- 21. Lee M. J., Jung C. H., Kang Y. M., et al., “Serum Bilirubin as a Predictor of Incident Metabolic Syndrome: A 4‐year Retrospective Longitudinal Study of 6205 Initially Healthy Korean Men,” Diabetes & Metabolism 40, no. 4 (2014): 305–309. [DOI] [PubMed] [Google Scholar]
- 22. Lee Y. B., Lee S. E., Jun J. E., et al., “Change in Serum Bilirubin Level as a Predictor of Incident Metabolic Syndrome,” PLoS One 11, no. 12 (2016): e0168253. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Nikouei M., Cheraghi M., Ghaempanah F., et al., “The Association Between Bilirubin Levels, and the Incidence of Metabolic Syndrome and Diabetes Mellitus: A Systematic Review and Meta‐Analysis of Cohort Studies,” Clinical Diabetes and Endocrinology 10, no. 1 (2024): 1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Oda E. and Aizawa Y., “Total Bilirubin Is Inversely Associated With Metabolic Syndrome but Not a Risk Factor for Metabolic Syndrome in Japanese Men and Women,” Acta Diabetologica 50, no. 3 (2013): 417–422. [DOI] [PubMed] [Google Scholar]
- 25. Wu Y., Li M., Xu M., et al., “Low Serum Total Bilirubin Concentrations Are Associated With Increased Prevalence of Metabolic Syndrome in Chinese,” Journal of Diabetes 3, no. 3 (2011): 217–224. [DOI] [PubMed] [Google Scholar]
- 26. Zhong P., Sun D. M., Wu D. H., Li T. M., Liu X. Y., and Liu H. Y., “Serum Total Bilirubin Levels Are Negatively Correlated With Metabolic Syndrome in Aged Chinese Women: A Community‐Based Study,” Brazilian Journal of Medical and Biological Research 50, no. 2 (2017): e5252. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Cheriyath P., “High Total Bilirubin as a Protective Factor for Diabetes Mellitus: An Analysis of NHANES Data From 1999–2006,” Journal of Clinical Medicine Research 2, no. 5 (2010): 201–206. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Kipp Z. A., Xu M., Bates E. A., Lee W. H., Kern P. A., and Hinds T. D., “Bilirubin Levels Are Negatively Correlated With Adiposity in Obese Men and Women, and Its Catabolized Product, Urobilin, Is Positively Associated With Insulin Resistance,” Antioxidants 12, no. 1 (2023): 170. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Johnson C. L., Paulose‐Ram R., Ogden C. L., et al., “National Health and Nutrition Examination Survey: Analytic Guidelines, 1999–2010,” Vital and Health Statistics. Series 2, Data Evaluation and Methods Research 2, no. 161 (2013): 1–24. [PubMed] [Google Scholar]
- 30. Guidance: Data sets not requiring IRB review . National Bureau of Economic Research, accessed November 30, 2025, https://www.nber.org/programs-projects/projects-and-centers/human-subjects-protection-and-institutional-review-board-irb/guidance-data-sets-not-requiring-irb-review.
- 31. McGill M. R., “The Past and Present of Serum Aminotransferases and the Future of Liver Injury Biomarkers,” EXCLI Journal 15 (2016): 817–828. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Ferreira M. P. and Willoughby D., “Alcohol Consumption: The Good, the Bad, and the Indifferent,” Applied Physiology, Nutrition, and Metabolism 33, no. 1 (2008): 12–20. [DOI] [PubMed] [Google Scholar]
- 33. Reja D., Makar M., Visaria A., Karanfilian B., and Rustgi V., “Blood Lead Level is Associated With Advanced Liver Fibrosis in Patients With Non‐Alcoholic Fatty Liver Disease: A Nationwide Survey (NHANES 2011–2016),” Annals of Hepatology 19, no. 4 (2020): 404–410. [DOI] [PubMed] [Google Scholar]
- 34. Van Wagner L. B. and Green R. M., “Evaluating Elevated Bilirubin Levels in Asymptomatic Adults,” Journal of the American Medical Association 313, no. 5 (2015): 516–517. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Sheinenzon A., Shehadeh M., Michelis R., Shaoul E., and Ronen O., “Serum Albumin Levels and Inflammation,” International Journal of Biological Macromolecules 184 (2021): 857–862. [DOI] [PubMed] [Google Scholar]
- 36. Goossens G. H., “The Metabolic Phenotype in Obesity: Fat Mass, Body Fat Distribution, and Adipose Tissue Function,” Obesity Facts 10, no. 3 (2017): 207–215. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Stokes A., “Using Maximum Weight to Redefine Body Mass Index Categories in Studies of the Mortality Risks of Obesity,” Population Health Metrics 12, no. 1 (2014): 6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Calle E. E., Thun M. J., Petrelli J. M., Rodriguez C., and Heath J. C. W., “Body‐Mass Index and Mortality in a Prospective Cohort of US Adults,” The New England Journal of Medicine 341, no. 15 (1999): 1097–1105. [DOI] [PubMed] [Google Scholar]
- 39. Chooi Y. C., Ding C., and Magkos F., “The Epidemiology of Obesity,” Metabolism: Clinical and Experimental 92 (2019): 6–10. [DOI] [PubMed] [Google Scholar]
- 40. Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults ., “Executive Summary of the Third Report of The National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III),” JAMA: The Journal of the American Medical Association 285, no. 19 (2001): 2486–2497. [DOI] [PubMed] [Google Scholar]
- 41. Mohri Y., Inoue Y., Tanaka K., Hiro J., Uchida K., and Kusunoki M., “Prognostic Nutritional Index Predicts Postoperative Outcome in Colorectal Cancer,” World Journal of Surgery 37, no. 11 (2013): 2688–2692. [DOI] [PubMed] [Google Scholar]
- 42. Onodera T., Goseki N., and Kosaki G., “[Prognostic Nutritional Index in Gastrointestinal Surgery of Malnourished Cancer Patients],” Nihon Geka Gakkai Zasshi 85, no. 9 (1984): 1001–1005. [PubMed] [Google Scholar]
- 43. Functional Limitation. National Center for Health Statistics . U.S. Centers for Disease Control and Prevention, accessed October 29, 2025, https://www.cdc.gov/nchs/hus/topics/functional-limitation.htm.
- 44. Zipf G., Chiappa M., Porter K. S., Ostchega Y., Lewis B. G., and Dostal J., “National Health and Nutrition Examination Survey: Plan and Operations, 1999–2010,” Vital and Health Statistics. Ser. 1, Programs and Collection Procedures 1, no. 56 (2013): 1–37. [PubMed] [Google Scholar]
- 45. Meech R., Hu D. G., Hulin J. A., and Mackenzie P. I., “Sex‐Specific UGT Expression and Function: Prevalence, Potential Mechanisms and Significance,” Expert Opinion on Drug Metabolism & Toxicology 21, no. 5 (2025): 511–518. [DOI] [PubMed] [Google Scholar]
- 46. Thomas D. M., Bredlau C., Bosy‐Westphal A., et al., “Relationships Between Body Roundness With Body Fat and Visceral Adipose Tissue Emerging From a New Geometrical Model,” Obesity 21, no. 11 (2013): 2264–2271. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47. Jenko‐Pražnikar Z., Petelin A., Jurdana M., and Žiberna L., “Serum Bilirubin Levels Are Lower in Overweight Asymptomatic Middle‐Aged Adults: An Early Indicator of Metabolic Syndrome?,” Metabolism: Clinical and Experimental 62, no. 7 (2013): 976–985. [DOI] [PubMed] [Google Scholar]
- 48. Andersson C., Weeke P., Fosbøl E. L., et al., “Acute Effect of Weight Loss on Levels of Total Bilirubin in Obese, Cardiovascular High‐Risk Patients: An Analysis From the Lead‐In Period of the Sibutramine Cardiovascular Outcome Trial,” Metabolism: Clinical and Experimental 58, no. 8 (2009): 1109–1115. [DOI] [PubMed] [Google Scholar]
- 49. Zhang X., Ma N., Lin Q., et al., “Body Roundness Index and All‐Cause Mortality Among US Adults,” JAMA Network Open 7, no. 6 (2024): e2415051. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50. Mohammadian Khonsari N., Khashayar P., Shahrestanaki E., et al., “Normal Weight Obesity and Cardiometabolic Risk Factors: A Systematic Review and Meta‐Analysis,” Frontiers in Endocrinology 13 (2022): 857930. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51. Jørgensen M. E., Torp‐Pedersen C., Finer N., et al., “Association Between Serum Bilirubin and Cardiovascular Disease in an Overweight High Risk Population From the SCOUT Trial,” Nutrition, Metabolism, and Cardiovascular Diseases 24, no. 6 (2014): 656–662. [DOI] [PubMed] [Google Scholar]
- 52. Silverman E. M., Sapala J. A., and Appelman H. D., “Regression of Hepatic Steatosis in Morbidly Obese Persons After Gastric Bypass,” American Journal of Clinical Pathology 104, no. 1 (1995): 23–31. [DOI] [PubMed] [Google Scholar]
- 53. Swift D. L., Johannsen N. M., Earnest C. P., Blair S. N., and Church T. S., “Effect of Different Doses of Aerobic Exercise Training on Total Bilirubin Levels,” Medicine & Science in Sports & Exercise 44, no. 4 (2012): 569–574. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 54. Xu J., Kulkarni S. R., Li L., and Slitt A. L., “UDP‐Glucuronosyltransferase Expression in Mouse Liver Is Increased in Obesity‐ and Fasting‐Induced Steatosis,” Drug Metabolism and Disposition 40, no. 2 (2012): 259–266. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 55. Hinds T. D., Creeden J. F., Gordon D. M., et al., “Rats Genetically Selected for High Aerobic Exercise Capacity Have Elevated Plasma Bilirubin by Upregulation of Hepatic Biliverdin Reductase‐A (BVRA) and Suppression of UGT1A1,” Antioxidants 9, no. 9 (2020): 889. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 56. Bates E. A., Kipp Z. A., Martinez G. J., et al., “Suppressing Hepatic UGT1A1 Increases Plasma Bilirubin, Lowers Plasma Urobilin, Reorganizes Kinase Signaling Pathways and Lipid Species and Improves Fatty Liver Disease,” Biomolecules 13, no. 2 (2023): 252. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 57. Hinds T. D., Burns K. A., Hosick P. A., et al., “Biliverdin Reductase A Attenuates Hepatic Steatosis by Inhibition of Glycogen Synthase Kinase (GSK) 3β Phosphorylation of Serine 73 of Peroxisome Proliferator‐Activated Receptor (PPAR) Α,” Journal of Biological Chemistry 291, no. 48 (2016): 25179–25191. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 58. Takei R., Inoue T., Sonoda N., et al., “Bilirubin Reduces Visceral Obesity and Insulin Resistance by Suppression of Inflammatory Cytokines,” PLoS One 14, no. 10 (2019): e0223302. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 59. Seyed Khoei N., Grindel A., Wallner M., et al., “Mild Hyperbilirubinaemia as an Endogenous Mitigator of Overweight and Obesity: Implications for Improved Metabolic Health,” Atherosclerosis 269 (2018): 306–311. [DOI] [PubMed] [Google Scholar]
- 60. McArdle P. F., Whitcomb B. W., Tanner K., Mitchell B. D., Shuldiner A. R., and Parsa A., “Association Between Bilirubin and Cardiovascular Disease Risk Factors: Using Mendelian Randomization to Assess Causal Inference,” BMC Cardiovascular Disorders 12 (2012): 16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 61. Belo L., Nascimento H., Kohlova M., et al., “Body Fat Percentage Is a Major Determinant of Total Bilirubin Independently of UGT1A1*28 Polymorphism in Young Obese,” PLoS One 9, no. 6 (2014): e98467. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 62. Stender S., Frikke‐Schmidt R., Nordestgaard B. G., Grande P., and Tybjærg‐Hansen A., “Genetically Elevated Bilirubin and Risk of Ischaemic Heart Disease: Threemendelian Randomization Studies and a Meta‐Analysis,” Journal of Internal Medicine 273, no. 1 (2013): 59–68. [DOI] [PubMed] [Google Scholar]
- 63. Abbasi A., Deetman P. E., Corpeleijn E., et al., “Bilirubin as a Potential Causal Factor in Type 2 Diabetes Risk: A Mendelian Randomization Study,” Diabetes 64, no. 4 (2015): 1459–1469. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplemental Figure 1: Association between body mass index as a continuous variable with low current total bilirubin according to age groups. Panel A describes the association in the overall population for those aged < 55 years while panel B shows the associations in those ≥ 55 years. The reference point was a body mass index of 23 kg/m2. BMI: body mass index.
Supplemental Figure 2: Association between body mass index as a continuous non‐linear variable and odds of low current total bilirubin concentration. Panel A describes the association in the overall population while panels B and C describe the associations in men and women, respectively. Natural cubic splines with five traditional knots were used. The reference point was a body mass index of 23 kg/m2. BMI: body mass index.
Supplemental Figure 3: Association between body roundness index with low current total bilirubin concentration. Panel A describes the association in the overall population, while panels B and C describe the association in men and women, respectively. Natural cubic splines with five traditional knots were used. The reference used was a body roundness index of 3.5.
Supplemental Table 1: Cross‐tabulation of past and current body mass index in men. Supplemental Table 2: Cross‐tabulation of past and current body mass index in women.
Data Availability Statement
The data supporting the findings of this study are available upon reasonable request from the corresponding author.