Elevated serum bilirubin has been suggested to reduce the risk for mortality. Cardiorespiratory fitness (CRF) has also been reported to have inverse association with all-cause and cardiovascular disease (CVD) mortality. The association between serum bilirubin, all-cause and CVD mortality, and the effect of CRF on the observed association was investigated. A total of 1,279 men aged 30 to 82 years underwent baseline medical examinations from 1974 to 1997 at the Cooper Clinic in Dallas, Texas. During an average of 17 years of follow-up, 698 men died, with 253 deaths due to CVD (36%). Men in the highest bilirubin quartiles had significantly lower risk for all-cause mortality compared to men in the lowest quartiles (p for trend = 0.0043), after adjusting for age and examination year. This inverse association remained significant after further adjustment for known confounders (p for trend = 0.0018). Additional adjustment for treadmill time attenuated the association (p for trend = 0.0090). Similar patterns of association were observed between serum bilirubin quartiles and CVD mortality. CRF was inversely associated with all-cause mortality (p for trend <0.0001) after adjusting for age and examination year. This inverse association also was observed after further adjusting for known confounders (p for trend = 0.0004). After additional adjustment for serum bilirubin, the association between CRF and all-cause mortality remained significant (p for trend = 0.0012). All-cause mortality and CVD mortality were significantly lower in men in the moderate- to high-fitness quartiles in the low- and high-bilirubin groups. In conclusion, serum bilirubin level and CRF level were strongly and independently associated with all-cause and CVD mortality.
Serum bilirubin has been shown to have potent antioxidant properties whereas, cardiorespiratory fitness (CRF) has been suggested as a way to help maintain and promote the antioxidant defense capabilities of the body against oxidative stress. In recent years, mild hyperbilirubinemia has been reported to reduce the risk for mortality. The inverse association between CRF and all-cause and cardiovascular disease (CVD) mortality has also been well documented in the published research. The aims of this study were to investigate the association of elevated serum bilirubin levels with all-cause mortality and CVD mortality in men and to examine the impact of CRF on the observed association.
Methods
This study was based on data from a prospective nested case-control study obtained from the Aerobics Center Longitudinal Study (ACLS). In the original study, case subjects were those who died from any cause during the follow-up period (through the end of 1997), and control subjects for cases were selected randomly from among survivors who met the matching criteria for age (±5 years). For the present study, we used the original study sample and followed up for mortality through 2003 using an updated mortality database.
The present study population consisted of 1,279 adult men aged 30 to 82 years who underwent baseline medical examinations from 1974 to 1997 at the Cooper Clinic in Dallas, Texas. Medical evaluation was performed after an overnight fast of ≥12 hours. Participants came to the clinic for periodic health examinations and counseling regarding diet, exercise, and other lifestyle factors associated with increased risk for chronic diseases. Many of the study participants were referred by their employers for the examinations, some were referred by their personal physicians, and others were self-referred. Participants were excluded from the study if they had histories of CVD or cancer at baseline (n = 98), had liver problems (fatty food intolerance, gallbladder trouble, jaundice, hepatitis, or cirrhosis; n = 5), had body mass indexes (BMIs) ≤18 kg/m 2 (n = 1), were unable to achieve 85% of their age-predicted maximal heart rates (220 − age) on a graded exercise treadmill test (n = 74), or were followed for <1 year (n = 32). Most of the study population was white, well educated, and from middle- to upper-class socioeconomic strata. Participants provided written informed consent to participate in the study, and the study protocol was approved annually by the Cooper Clinic institutional review board.
The examination included fasting blood chemistry analysis, personal and family health history, anthropometry, blood pressure at rest, electrocardiography, and a maximal graded exercise test, performed during a single clinic visit. The baseline procedures have been described elsewhere. Briefly, body weight and height were measured using a standard physician scale and stadiometer, and BMI was computed as weight in kilograms divided by the square of height in meters. Blood pressure at rest was measured after a brief period of sitting quietly, using auscultatory methods with a mercury sphygmomanometer at the first and fifth Korotkoff sounds. Serum plasma and whole-blood samples were drawn from an antecubital vein and were analyzed for lipids and glucose using automated bioassays in accordance with the Centers for Disease Control and Prevention’s Lipid Standardization Program. Total serum bilirubin from nonhemolyzed serum was analyzed with a commercial calorimetric method using diazotized sulfanilic reagent. Diabetes mellitus was defined as fasting plasma glucose level ≥126 mg/dl, a history of physician-diagnosed diabetes mellitus, or current therapy with insulin. Hypertension was defined as blood pressure at rest ≥140/90 mm Hg or a history of physician-diagnosed hypertension. Information regarding smoking habits (never, former, or current), alcohol intake (number of drinks per week, correcting for alcohol content, where 1 U of alcohol was defined as 12 oz of beer, 5 oz of wine, or 1.5 oz of hard liquor), and parental history of CVD was obtained using a standardized questionnaire. Abnormal rest and exercise electrocardiographic (ECG) responses included rhythm and conduction disturbances and ischemic ST-T wave abnormalities, described in detail elsewhere. Previously, we found 90% agreement between the ECG interpretation recorded in our database and that of a group of 3 physicians who read a random sample of 357 patient records. CRF was determined on the basis of maximal treadmill time and was calculated using the modified Balke protocol. Patients were encouraged to give maximal effort, and the test end point was volitional exhaustion or termination by the physician for medical reasons. The total time (minutes) was used as an index of aerobic power. Time spent on the treadmill in this protocol correlates highly (r = 0.92) with maximal oxygen uptake in men. The speed and elevation of the final minute of the treadmill test were used to convert treadmill performance to METs. One MET is an oxygen uptake of 3.5 ml/kg/min, the energy expenditure at a rest state.
Participants were followed from the date of their baseline measurements until their dates of death or December 31, 2003. All-cause mortality and cardiovascular mortality were determined from the National Death Index or by review of death certificates in the decedents’ states of residence. CVD mortality was defined using International Classification of Diseases, Ninth Revision, codes 390 to 449.9 before 1999 and International Classification of Diseases, Tenth Revision, codes 100 to 178 from 1999 to 2003.
Baseline characteristics were summarized across bilirubin quartiles. Trends in covariates over bilirubin were assessed using F tests. Participants were classified into 4 groups on the basis of quartiles of bilirubin levels and age-specific treadmill time (20 to 39, 40 to 49, 50 to 59, and ≥60 years). For the joint analysis, participants were classified into low-bilirubin (quartiles 1 to 3) or high-bilirubin (quartile 4) and low-fitness (quartile 1) or high-fitness (quartiles 2 to 4) groups on the basis of our results showing lower mortality risk in the highest bilirubin quartile and lower all-cause mortality risk in quartiles 2 to 4 of treadmill time. The decision to use quartiles of CRF in this study was made to be consistent with the bilirubin classification. Cox regression analysis was used to estimate relative risk (RR) and 95% confidence intervals (CIs) of all-cause mortality and CVD mortality according to total serum bilirubin level (deaths per 10,000 person-years of follow up). Multivariate analysis included the following confounders: age, examination year, BMI, smoking status (never, past, or current smoker), alcohol intake (number of drinks per week), presence or absence of abnormal ECG findings, hypertension, diabetes, total cholesterol, family history of CVD, and maximal treadmill time or bilirubin levels. All p values are 2 sided, and p values <0.05 were regarded as statistically significant. All analyses were performed using SAS version 9.2 (SAS Institute Inc., Cary, North Carolina).
Results
In this study population, 698 men died during follow-up, with 253 deaths due to CVD (36%). The mean follow-up period was 17 ± 9 years, and the mean age was 52 ± 10 years. The baseline characteristics of the 1279 participants included in this study are listed in Table 1 . The mean serum bilirubin in each quartile was 0.36 ± 0.06, 0.55 ± 0.05, 0.75 ± 0.05, and 1.12 ± 0.27 mg/dl, respectively, and ranged from 0.10 to 2.60 mg/dl. The correlation between bilirubin level and treadmill time in this population was 0.07 (p = 0.009).
| Variable | Bilirubin Level (mg/dl) | p Value for Trend | |||
|---|---|---|---|---|---|
| Quartile 1 (0.10–0.49) | Quartile 2 (0.50–0.69) | Quartile 3 (0.70–0.89) | Quartile 4 (0.90–2.60) | ||
| Number of patients | 225 | 398 | 317 | 339 | |
| Age (years) | 51.97 ± 10.1 | 52.14 ± 10.41 | 52.12 ± 10.63 | 52.60 ± 10.60 | 0.89 |
| Bilirubin (mg/dl) | 0.36 ± 0.06 | 0.55 ± 0.05 | 0.75 ± 0.05 | 1.12 ± 0.27 | <0.0001 |
| Maximal treadmill time (min) | 14.18 ± 5.13 | 14.80 ± 5.40 | 15.40 ± 5.10 | 15.40 ± 4.78 | 0.02 |
| Maximal METs | 9.89 ± 2.42 | 10.19 ± 2.55 | 10.47 ± 2.42 | 10.45 ± 2.23 | 0.02 |
| Weight (kg) | 82.30 ± 12.48 | 82.53 ± 11.24 | 81.88 ± 13.01 | 83.34 ± 13.11 | 0.50 |
| Height (cm) | 177.81 ± 6.04 | 177.82 ± 6.03 | 177.91 ± 6.40 | 178.08 ± 6.49 | 0.94 |
| BMI (kg/m 2 ) | 25.97 ± 3.15 | 26.08 ± 3.18 | 25.82 ± 3.53 | 26.23 ± 3.55 | 0.47 |
| Smoking status | |||||
| Never | 163 (12.7%) | 293 (22.9%) | 240 (18.8%) | 268 (21%) | 0.24 |
| Former | 3 (0.2%) | 13 (1%) | 12 (0.9%) | 15 (1.2%) | 0.24 |
| Current | 59 (4.6%) | 92 (7.2%) | 65 (5.1%) | 56 (4.4%) | 0.03 |
| Alcohol intake (drinks/week) | 8.60 ± 12.52 | 8.75 ± 12.61 | 9.34 ± 11.24 | 9.94 ± 12.37 | 0.50 |
| Systolic blood pressure (mm Hg) | 125 ± 16 | 127 ± 17 | 126 ± 16 | 127 ± 17 | 0.15 |
| Diastolic blood pressure (mm Hg) | 82 ± 10 | 82 ± 10 | 82 ± 10 | 83 ± 11 | 0.37 |
| Fasting glucose (mg/dl) | 106.34 ± 24.57 | 107.24 ± 31.85 | 103.88 ± 22.13 | 106.45 ± 28.64 | 0.42 |
| Total cholesterol (mg/dl) | 218.12 ± 41.87 | 220.11 ± 38.32 | 219.41 ± 44.83 | 222.34 ± 37.74 | 0.65 |
| Hypertension ⁎ | 83 (6.5%) | 174 (13.6%) | 141 (11%) | 159 (12.4%) | 0.13 |
| Diabetes mellitus † | 32 (2.5%) | 49 (3.8%) | 31 (2.4%) | 41 (3.2%) | 0.46 |
| Abnormal ECG findings | 34 (2.7%) | 73 (5.7%) | 48 (3.8%) | 67 (5.2%) | 0.32 |
| Parental CVD | 99 (7.7%) | 167 (13.1%) | 108 (8.4%) | 134 (10.5%) | 0.08 |
⁎ Defined as a history of physician diagnosis or blood pressure at rest ≥140/90 mm Hg.
† Defined as a history of physician diagnosis, the use of insulin, or measured fasting glucose level ≥126 mg/dl (7.0 mmol/L).
Table 2 lists the RRs of mortality across quartiles of serum bilirubin. After adjusting for age and examination year (model 1), total serum bilirubin level was inversely associated with all-cause mortality (p for trend = 0.0043), and men in the fourth quartile had a significantly lower risk for all-cause mortality compared to men in the first and second quartiles. This inverse association remained significant after adjusting for smoking status, alcohol intake, BMI, total cholesterol, hypertension, diabetes, abnormal ECG findings, and parental CVD (model 2; p for trend = 0.0018). After additional adjustment for treadmill time, the inverse association between bilirubin quartile and all-cause mortality was attenuated (p for trend = 0.0090). However, men in the highest bilirubin quartiles still had a significantly lower risk for all-cause mortality. Similar patterns of association were observed between serum bilirubin quartiles and CVD mortality ( Table 2 ). After excluding men with total serum bilirubin levels >1.00 mg/dl, who were considered to have Gilbert syndrome, which causes elevation of bilirubin, the associations between total serum bilirubin level and all-cause mortality and CVD mortality in men were comparable (data not shown).
| Variable | Bilirubin (mg/dl) | p Value for Trend | |||
|---|---|---|---|---|---|
| Quartile 1 (0.10–0.49) | Quartile 2 (0.50–0.69) | Quartile 3 (0.70–0.89) | Quartile 4 (0.90–2.60) | ||
| All-cause mortality | |||||
| Number of deaths | 134 | 225 | 171 | 168 | |
| Death rate ⁎ | 374 | 339 | 310 | 274 | |
| RR (95% CI) | |||||
| Model 1 † | 1.00 (reference) | 0.91 (0.73–1.12) | 0.83 (0.66–1.04) | 0.74 (0.58–0.92) | 0.0043 |
| Model 2 ‡ | 1.00 (reference) | 0.86 (0.69–1.07) | 0.83 (0.66–1.04) | 0.69 (0.55–0.87) | 0.0018 |
| Model 3 § | 1.00 (reference) | 0.89 (0.71–1.10) | 0.86 (0.68–1.08) | 0.73 (0.58–0.92) | 0.0090 |
| CVD mortality | |||||
| Number of deaths | 46 | 84 | 67 | 56 | |
| Death rate ⁎ | 131 | 127 | 120 | 91 | |
| RR (95% CI) | |||||
| Model 1 † | 1.00 (reference) | 0.97 (0.68–1.39) | 0.92 (0.63–1.34) | 0.69 (0.47–1.02) | 0.0446 |
| Model 2 ‡ | 1.00 (reference) | 0.83 (0.58–1.20) | 0.88 (0.60–1.29) | 0.58 (0.39–0.86) | 0.0090 |
| Model 3 § | 1.00 (reference) | 0.85 (0.59–1.23) | 0.91 (0.62–1.34) | 0.61 (0.41–0.91) | 0.0217 |
⁎ Death rate per 10,000 person-years.
† Adjusted for age and examination year.
‡ Adjusted for age, examination year, smoking status, alcohol intake, BMI, total cholesterol, hypertension, diabetes, abnormal ECG findings, and parental CVD.
§ Adjusted for age, examination year, smoking status, alcohol intake, BMI, total cholesterol, hypertension, diabetes, abnormal ECG findings, parental CVD, and treadmill time.
Table 3 lists the results of the RR for mortality in the age-specific CRF quartiles. After adjusting for age and examination year (model 1), the age-specific CRF had a significant inverse association with all-cause mortality (p for trend <0.0001), with men in the quartiles 2 to 4 having significantly lower all-cause mortality compared to men in quartile 1. This association with all-cause mortality was observed after further adjusting for smoking status, alcohol intake, BMI, total cholesterol, hypertension, diabetes, abnormal ECG findings, and parental CVD (model 2; p for trend <0.0004). After additional adjustment for bilirubin level (model 3), the association between CRF and all-cause mortality remained significant (p for trend = 0.0012). A similar pattern of association was observed between CRF quartiles and CVD mortality (p for trend <0.05).
