Abstract
Previous studies have linked liver diseases to lung cancer (LC) risk; however, few studies evaluated the associations of circulating liver enzyme levels with LC risk. We conducted a study of 353 incident LC cases and 646 matched controls with baseline serum alanine aminotransferase (ALT) and of 548 cases and 1032 matched controls with baseline serum alkaline phosphatase (ALP) nested within the Southern Community Cohort Study. Conditional logistic regression and generalized linear models were used to estimate adjusted odds ratios (ORs) and 95% confidence intervals (CIs) among all study participants and by stratification of potential effect modifiers. Most participants had clinically normal liver enzyme levels. Higher serum ALT levels were associated with reduced LC risk. Compared with the lowest tertile, participants in the second and third tertiles had OR (95% CI) of 0.74 (0.48–1.14) and 0.47 (0.28–0.78) (Ptrend < .01), respectively. The inverse association was observed in African Americans (AAs) and European Americans, which was especially prominent among men, and was seen in both those diagnosed within [ORT3 versus T1 = 0.41 (0.19–0.88)] and beyond [ORT3 versus T1 = 0.35 (0.17–0.73)] a median follow-up time of 39 months. Higher serum ALP levels were associated with increased LC risk among AA men only [ORT3 versus T1 = 2.01 (1.19–3.39)] (Ptrend < .01). Our results indicate that in a predominantly low-income American population, higher serum ALT levels may be related to lower LC risk. Further studies are warranted to confirm our findings and elucidate the potential underlying biological mechanisms of the associations.
Keywords: lung cancer, liver enzymes, African Americans, low-income population, alanine aminotransferase, alkaline phosphatase
Graphical Abstract
Graphical Abstract.
1. Introduction
Lung cancer (LC) is one of the most common and lethal cancer types for both men and women in the USA and worldwide [1, 2]. Racial differences in LC incidence and survival are well-documented. African Americans (AAs) are more likely to develop LC and have a poor prognosis compared with European Americans (EAs) [3, 4]. Although cigarette smoking is the primary risk factor, AAs had lower smoking intensity and shorter smoking duration than EAs [5]. This suggests that other factors, such as genetic factors and socioeconomic status, also contribute to these disparities in LC risk [6–8]. The underlying reasons for LC disparities are still unclear.
The liver is involved mainly in metabolism, and also in nutrient storage, immune surveillance, and more. Previous studies reported that chronic liver diseases were linked to an elevated risk of LC [9–11]. It has been suggested that the liver plays a role in the development of lung diseases by gut–liver–lung and liver–lung axes, and through involvement in immune regulation [12, 13]. Thus, we hypothesized that liver function biomarkers, including alanine aminotransferase (ALT) and alkaline phosphatase (ALP), may be associated with LC risk. Until now, two studies have investigated the associations of these two markers with LC risk among only European populations. In a case-cohort study including 195 LC cases, elevated ALP levels were associated with an increased LC risk, whereas ALT was not associated with LC risk [14]. This study is limited by its small sample size. Another study observed an inverse association between ALT and LC risk among men but not women in the UK Biobank, a large population-based prospective cohort [15]. They also found that ALP was positively associated with LC risk in both men and women. To date, no study has been conducted among AAs or socioeconomically disadvantaged populations who were experiencing elevated LC incidence rates.
In this study, we evaluated the associations of prediagnostic serum ALT and ALP levels with LC risk in the Southern Community Cohort Study (SCCS), a cohort study with a large proportion of low-income AAs and EAs.
2. Materials and method
2.1. Study population
The SCCS is a large-scale prospective cohort study that has enrolled more than 85 000 participants ages 40 to 79 between 2002 and 2009 in 12 states in the southeastern USA [16]. Of them, ∼70% were AAs. Approximately 86% of participants were enrolled through community health centers (CHC) that mainly provide health care for medically and financially underserved populations, and ∼14% were recruited from the general population. At recruitment, trained staff collected baseline data regarding demographic, lifestyle, dietary, and medical information via computer-assisted personal interviews for CHC participants and via mailed questionnaires for general population participants. Blood samples were collected from CHC participants without known human immunodeficiency viruses/acquired immunodeficiency syndrome or hepatitis infections, with just under half providing 20 ml samples. All the participants signed informed consent, and the Institutional Review Board at Vanderbilt University and Meharry Medical College have approved this study.
A total of 552 incident LC cases and 1039 matched controls with stored serum samples were included in this study. The median (interquartile range [IQR]) follow-up time among the cases was 56.0 (30.0–81.5) months. Incident cases of LC were ascertained using the International Statistical Classification of Disease (ICD) codes (ICD10 code for LC: C34.0-C34.9) via linkage to state cancer registries and the National Death Index by November 2016. Controls were randomly selected from the cancer-free population and individually matched to cases at a 2:1 ratio on age (±2 years), sex, race, date (±6 months), and site (CHC) of study enrollment. However, some controls did not have ALT or ALP data and, thus, were excluded from the analysis. Thus, some of the cases only have one control. A total of 353 cases and 646 controls with available ALT data were included for the downstream analyses with median (IQR) follow-up time among the cases was 39.0 (22.0–65.0) months. For ALP, a total of 548 cases and 1032 controls were included for the downstream analyses with median (IQR) follow-up time among the cases was 56.0 (30.0–82.0) months.
2.2. Assessment of serum levels of liver enzymes
The details of the blood collection and storage methods and laboratory procedures have been previously described [16, 17]. Briefly, after blood draw, samples were immediately refrigerated and shipped cold overnight to the Vanderbilt University Medical Center Molecular Epidemiology Core Laboratory. Serum samples were isolated from blood and then stored at −80°C until biomarker analysis. Serum levels of liver enzymes were quantified as unit per liter (U/L) using the Beckman Coulter clinical chemistry analyser DXC 600 or Roche Modular Analytics System (Roche) following the manufacturers’ protocols. Each case-control matched pair was included in the same assay batch and adjacently in order to reduce the batch effects. To avoid measurement errors, laboratory staff was blinded to the case-control status and the identity of quality control samples (3%) which were included in the assays. The intra-assay and interassay coefficients of variation were 1.45% and 3.41% for ALP and 15.99% and 15.77% for ALT. The clinical threshold for the serum levels of both liver enzymes is shown in Supplementary Table 1. Due to insufficient serum volume, we only have serum ALT results for 353 cases and 646 controls.
2.3. Covariates assessment
All the covariates were collected at the baseline interview using well-established questionnaires by trained interviewers. The baseline questionnaire collects information about sociodemographic and lifestyle characteristics as well as medical history. Self-reported race/ethnicities were categorized into AAs, EAs, and Other. Education levels (<11 years, completed high school, vocational/technical school, and university degree or higher) and annual household income (<$15 000, $15 000–$24 999, and ≥$25 000) were also categorized. Body mass index (BMI) was calculated as the formula: . BMI was dichotomized into normal/over-weight (<30 kg/m2) and obese (≥30 kg/m2). Smoking status (current, former, and never smokers), pack-years (≥20 and <20 pack-years), and alcohol consumption (never, light, and heavy alcohol drinkers) were categorized. Light drinkers were defined as >0 to ≤2 drink/day for men and >0 to ≤1 drink/day for women, while heavy drinkers were defined as >2 drink/day for men and >1 drink/day for women. Self-reported history of diabetes, hypertension, and chronic obstructive pulmonary disease (COPD) were categorical variables (yes and no).
2.4. Statistical analysis
Missing data for covariates were observed among limited (1.3%–3.0%) proportions of participants, so we imputed these data using race-sex-specific mode (for categorical variables) or median (for continuous variables) values. We used mean and standard deviations (SDs), median and quartiles, or frequency and number to describe distributions of exposures and confounders across both cases and control groups. Paired student t-test, Wilcoxon rank, and McNemar tests were applied to examine whether the differences were statistically significant between the two groups. Spearman correlation between the serum levels of ALT and ALP was calculated.
Due to racial and gender differences in circulating liver enzyme levels [18], we categorized the levels of ALT and ALP into three groups using race- and sex-specific tertiles among all controls. We then built conditional logistic regression models (CLR) to calculate adjusted odds ratios (OR) and their 95% confidence intervals (CI). In Model 1, we adjusted for age at enrollment (continuous), smoking status and intensity (current with ≥20 pack-years, current with <20 pack-years, former with ≥20 pack-years, former with <20 pack-years, and never), alcohol consumption, and self-reported COPD history. In Model 2, we additionally controlled for education levels, annual household income, self-reported history of diabetes and hypertension, and BMI (continuous). We also conducted stratified analyses by race, sex, histological subtypes (adenocarcinoma [ADE]; squamous cell carcinoma [SQC]), and median time between blood draw and LC diagnosis using CLR, as well as by smoking status, and BMI categories using a generalized linear model with the equation estimation (GEE) approach and autoregressive (1) method. GEE outperforms CLR in terms of uncompleted matched pairs [19]. P-value for trend was calculated using the Wald test by treating the categories of the liver enzyme levels as an ordinal variable. We also estimated ORs on per SD increase logarithm values of liver enzyme levels. We additionally applied curve fitting using restricted cubic splines with three knots (5th, 50th, and 95th) in the models and found these P-values for nonlinear trend (results not shown) were similar to those in the linear models. We further assessed interactions between serum liver function markers and potential modifiers using Wald tests. No significant interactions were observed for sex, race, smoking status, or BMI (results not shown).
Sensitivity analyses were performed among participants without clinical abnormal liver enzyme levels (results not shown) and without a self-reported diabetes diagnosis. To avoid potential reverse causation, we restricted our analyses to those diagnosed with LC after 2 years of blood draw. All the statistical analyses were performed in SAS (version 9.4; SAS Institute). The significance level was defined as two-sided P < .05. We followed the Strengthening Reporting of Observational studies in Epidemiology guidelines to present this nested case-control study.
3. Results
Table 1 shows the baseline characteristics of our study population. Incident LC cases were more likely than controls to be less educated, heavy smokers, heavy alcohol drinkers, have low household income, have low BMI, and report a history of COPD but no diabetes (P < .05) among those included in ALT and ALP analyses. There were no obvious differences in the medical history of hypertension between cases and controls among those included in ALT (P = .72) and ALP (P = .79) analyses. Among the LC cases, the main histological subtypes were ADE followed by SQC and then other nonsmall cell LC (NSCLC).
Table 1.
Baseline characteristics of study population by liver enzyme, Southern Community Cohort Study.
| ALT (N = 999) |
ALP (N = 1580) |
|||||
|---|---|---|---|---|---|---|
| Cases (N = 353) | Controls (N = 646) | P-valuea | Cases (N = 548) | Controls (N = 1032) | P-valuea | |
| Age at enrollment (years), Mean ± SD | 56.5 ± 8.8 | 56.4 ± 9.0 | <.01 | 56.5 ± 9.0 | 56.4 ± 9.0 | <.01 |
| Race, N (%) | ||||||
| EAs | 114 (32.3) | 205 (31.7) | 188 (34.3) | 351 (34.0) | ||
| AAs | 231 (65.4) | 427 (66.1) | 343 (62.6) | 649 (62.9) | ||
| Other | 8 (2.3) | 14 (2.2) | 17 (3.1) | 32 (3.1) | ||
| Gender, N (%) | ||||||
| Female | 146 (41.4) | 270 (41.8) | 245 (44.7) | 463 (44.9) | ||
| Male | 207 (58.6) | 376 (58.2) | 303 (55.3) | 569 (55.1) | ||
| Education levels, N (%) | <.01 | <.01 | ||||
| Less than 11 years | 167 (47.3) | 275 (42.6) | 255 (46.5) | 404 (39.2) | ||
| Completed high school | 115 (32.6) | 211 (32.7) | 187 (34.1) | 368 (35.7) | ||
| Vocational/technical school | 62 (17.6) | 109 (16.9) | 92 (16.8) | 180 (17.4) | ||
| University degree or higher | 9 (2.6) | 51 (7.9) | 14 (2.6) | 81 (7.8) | ||
| Household income levels, N (%) | .03 | <.01 | ||||
| < $15 000 | 256 (72.5) | 424 (65.6) | 396 (72.3) | 655 (63.5) | ||
| $15 000–$24 999 | 65 (18.4) | 139 (21.5) | 101 (18.4) | 230 (22.3) | ||
| ≥ $25 000 | 32 (9.1) | 83 (12.9) | 51 (9.3) | 147 (14.2) | ||
| Smoking status, N (%) | <.01 | <.01 | ||||
| Current smokers | 257 (72.8) | 295 (45.7) | 407 (74.3) | 481 (46.6) | ||
| Former smokers | 73 (20.7) | 170 (26.3) | 110 (20.1) | 256 (24.8) | ||
| Never smokers | 23 (6.5) | 181 (28.0) | 31 (5.7) | 295 (28.6) | ||
| Pack-years, median (Q1, Q3)b | 33.0 (18.0, 49.5) | 18.6 (10.0, 35.0) | <.01 | 31.0 (18.0, 48.0) | 19.2 (10.0, 36.0) | <.01 |
| Alcohol consumption, N (%) | .04 | <.01 | ||||
| Nondrinkers | 149 (42.2) | 312 (48.3) | 226 (41.2) | 508 (49.2) | ||
| Light drinkers | 124 (35.1) | 222 (34.4) | 194 (35.4) | 354 (34.3) | ||
| Heavy drinkers | 80 (22.7) | 112 (17.3) | 128 (23.4) | 170 (16.5) | ||
| BMI (kg/m2), Mean ± SD | 26.9 ± 6.0 | 29.4 ± 6.9 | <.01 | 26.8 ± 6.2 | 29.5 ± 6.9 | <.01 |
| History of COPD, N (%) | <.01 | <.01 | ||||
| No | 295 (83.6) | 592 (91.6) | 451 (82.3) | 932 (90.3) | ||
| Yes | 58 (16.4) | 58 (8.4) | 97 (17.7) | 100 (9.7) | ||
| History of hypertension, N (%) | .72 | .79 | ||||
| No | 149 (42.2) | 273 (42.3) | 234 (42.7) | 428 (41.5) | ||
| Yes | 204 (57.8) | 373 (57.7) | 314 (57.3) | 604 (58.5) | ||
| History of diabetes, N (%) | .02 | .01 | ||||
| No | 286 (81.0) | 486 (75.2) | 450 (82.1) | 793 (76.8) | ||
| Yes | 67 (19.0) | 160 (24.8) | 98 (17.9) | 239 (23.2) | ||
| Serum ALT (U/L), median (Q1, Q3) | 16.0 (11.0, 24.0) | 17.0 (11.0, 25.0) | <.01 | |||
| Clinically abnormal, N (%) | 25 (7.0) | 57 (8.8) | .30 | |||
| Serum ALP (U/L), median (Q1, Q3) | 79.0 (66.0, 95.0) | 76.0 (63.0, 93.0) | .07 | |||
| Clinically abnormal, N (%) | 37 (6.8) | 70 (6.8) | .96 | |||
| Follow-up time (months), median (Q1, Q3) | 39.0 (22.0, 65.0) | 56.0 (30.0, 82.0) | ||||
| Histological subtypes, N (%) | ||||||
| ADE | 110 (31.2) | 178 (32.5) | ||||
| SQC | 81 (23.0) | 124 (22.6) | ||||
| Large cell carcinoma | 12 (3.4) | 16 (2.9) | ||||
| Other nonsmall cell LC | 63 (17.9) | 85 (15.5) | ||||
| Small cell LC | 43 (12.2) | 69 (12.6) | ||||
| Other/unknown | 44 (12.5) | 76 (13.9) | ||||
aPaired t-tests were used for age and BMI; paired Wilcoxon rank tests were used for circulating liver function markers and pack-years; simple conditional logistic regression models were used for categorical variables.
bPack-years were calculated among current and former smokers only.
The majority of participants had clinically normal serum ALT and ALP levels. Clinical abnormal levels of ALT and ALP were observed among 8.2% and 6.8% of participants, respectively. Serum ALT levels [median (IQR)] were significantly lower in cases [16.0 (11.0–24.0), U/L] than controls [17.0 (11.0–25.0), U/L] (P < .01), while serum ALP levels were higher in cases [79.0 (66.0–95.0), U/L] than in controls [76.0 (63.0–93.0), U/L] with the difference being marginally statistically significant (P = .07) (Table 1). The spearman correlation between these two liver enzymes is 0.01 (P = .73), suggesting no significant correlation. Older age and higher BMI were inversely associated with circulating ALT levels, whereas current smoking status and male sex were positively associated with ALT levels. Older age, higher BMI, and history of hypertension and diabetes were associated with increased ALP levels. Conversely, male sex, AA race, greater pack-years of smoking, and heavy alcohol consumption were associated with decreased ALP levels (data no shown).
The associations of serum liver function markers with LC risk are presented in Table 2. Overall, serum levels of ALT were inversely associated with the overall risk of LC (ORT2 versus T1 = 0.74, 95% CI: 0.48–1.14; ORT3 versus T1 = 0.47, 95% CI: 0.28–0.78; Ptrend = .003) in fully adjusted models. There were no major differences in OR estimates between minimally and fully adjusted models. In subgroup analyses, the inverse associations of serum ALT levels with LC risk were observed in both AAs (OR per SD: 0.70, 95% CI: 0.54–0.90) and EAs (OR per SD: 0.65, 95% CI: 0.43–0.98). No significant interactions on LC risk between ALT levels and sex, smoking status, or BMI were seen. A significant inverse association of ALT levels with LC risk was only seen among men, with ORs of 0.52 (95% CI: 0.29–0.92) and 0.35 (95% CI: 0.18–0.69) for the second and third tertiles, compared with the first tertiles (Ptrend = .002) (Table 2). In addition, the inverse association of serum ALT levels with LC risk were more evident among current smokers (ORT3 versus T1 = 0.66, 95% CI: 0.49–0.90), and participants with a normal-/over-weight [BMI < 30 kg/m2] (ORT3 versus T1 = 0.58, 95% CI: 0.41–0.82) (Table 3). In the analysis stratified by LC histological types, the inverse association was only significant between serum levels of ALT and SQC (OR per SD = 0.55, 95%CI: 0.31–0.98) (Table 4). We further evaluated the association by the time between blood draw and LC diagnosis. The inverse association of serum ALT levels with LC risk were significant in both those diagnosed within (ORT3 versus T1 = 0.41, 95% CI: 0.19–0.88) and beyond (ORT3 versus T1 = 0.35, 95% CI: 0.17–0.73) a median follow-up time of 39 months (Table 5). Excluding individuals with extremely low and high serum levels of liver function markers (data not shown), individuals with a self-reported history of diabetes (Table 3), or individuals diagnosed within 2 years of blood draw (Table 5) had little effects on OR estimates.
Table 2.
Associations of serum liver enzyme levels with overall LC risk as well as by race and sex.a
| ALT | ALP | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Median | Cases | Controls | OR (95% CI)b | OR (95% CI)c | Median | Cases | Controls | OR (95% CI)b | OR (95% CI)c | |
| All participants | ||||||||||
| Tertile 1 | 9.0 | 124 | 200 | Ref | Ref | 58.0 | 152 | 328 | Ref | Ref |
| Tertile 2 | 16.0 | 126 | 213 | 0.72 (0.47, 1.10) | 0.74 (0.48, 1.14) | 76.0 | 186 | 355 | 1.08 (0.80, 1.45) | 1.05 (0.77, 1.42) |
| Tertile 3 | 29.0 | 103 | 233 | 0.43 (0.26, 0.71) | 0.47 (0.28, 0.78) | 101.0 | 210 | 349 | 1.09 (0.81, 1.47) | 1.13 (0.83, 1.53) |
| Ptrendd | .001 | .003 | .562 | .447 | ||||||
| Per SD increasee | 0.68 (0.55, 0.83) | 0.68 (0.55, 0.84) | 1.00 (0.89, 1.13) | 1.03 (0.91, 1.16) | ||||||
| AAs | ||||||||||
| Tertile 1 | 9.0 | 78 | 131 | Ref | Ref | 56.5 | 88 | 208 | Ref | Ref |
| Tertile 2 | 16.0 | 80 | 145 | 0.74 (0.45, 1.21) | 0.73 (0.43, 1.22) | 73.0 | 110 | 222 | 1.13 (0.77, 1.66) | 1.14 (0.76, 1.71) |
| Tertile 3 | 28.5 | 73 | 151 | 0.51 (0.28, 0.92) | 0.54 (0.28, 1.01) | 99.0 | 145 | 219 | 1.36 (0.93, 1.97) | 1.44 (0.97, 2.14) |
| Ptrend | .004 | .055 | .106 | .069 | ||||||
| Per SD increase | 0.70 (0.54, 0.89) | 0.70 (0.54, 0.90) | 1.08 (0.93, 1.25) | 1.10 (0.94, 1.29) | ||||||
| EAs | ||||||||||
| Tertile 1 | 9.5 | 44 | 66 | Ref | Ref | 60.0 | 58 | 112 | Ref | Ref |
| Tertile 2 | 18.0 | 41 | 62 | 0.86 (0.37, 2.02) | 0.95 (0.39, 2.34) | 78.0 | 72 | 120 | 1.01 (0.61, 1.67) | 1.01 (0.60, 1.69) |
| Tertile 3 | 30.0 | 29 | 77 | 0.41 (0.15, 1.10) | 0.42 (0.15, 1.18) | 104.0 | 58 | 119 | 0.68 (0.39, 1.16) | 0.69 (0.39, 1.20) |
| Ptrend | .056 | .065 | .158 | .187 | ||||||
| Per SD increase | 0.67 (0.45, 0.99) | 0.65 (0.43, 0.98) | 0.83 (0.67, 1.04) | 0.85 (0.68, 1.07) | ||||||
| Females | ||||||||||
| Tertile 1 | 8.0 | 40 | 81 | Ref | Ref | 61.0 | 82 | 150 | Ref | Ref |
| Tertile 2 | 15.0 | 57 | 85 | 1.33 (0.64, 2.75) | 1.43 (0.67, 3.05) | 80.0 | 83 | 157 | 1.00 (0.64, 1.56) | 0.98 (0.61, 1.57) |
| Tertile 3 | 25.0 | 49 | 104 | 0.69 (0.30, 1.58) | 0.75 (0.32, 1.78) | 108.5 | 80 | 156 | 0.79 (0.50, 1.26) | 0.85 (0.52, 1.38) |
| Ptrend | .276 | .363 | .331 | .506 | ||||||
| Per SD increase | 0.75 (0.54, 1.04) | 0.72 (0.52, 1.01) | 0.94 (0.78, 1.15) | 0.99 (0.81, 1.22) | ||||||
| Males | ||||||||||
| Tertile 1 | 10.0 | 84 | 119 | Ref | Ref | 55.0 | 70 | 178 | Ref | Ref |
| Tertile 2 | 18.0 | 69 | 128 | 0.54 (0.31, 0.94) | 0.52 (0.29, 0.92) | 72.0 | 103 | 198 | 1.22 (0.81, 1.83) | 1.19 (0.77, 1.83) |
| Tertile 3 | 34.0 | 54 | 129 | 0.34 (0.18, 0.65) | 0.35 (0.18, 0.69) | 96.0 | 130 | 193 | 1.47 (0.98, 2.19) | 1.50 (0.98, 2.29) |
| Ptrend | .001 | .002 | .059 | .056 | ||||||
| Per SD increase | 0.62 (0.47, 0.81) | 0.63 (0.47, 0.83) | 1.07 (0.91, 1.24) | 1.07 (0.91, 1.25) | ||||||
| AA males | ||||||||||
| Tertile 1 | 10.0 | 66 | 95 | Ref | Ref | 55.0 | 50 | 132 | Ref | Ref |
| Tertile 2 | 18.0 | 49 | 98 | 0.60 (0.32, 1.09) | 0.55 (0.29, 1.04) | 71.0 | 69 | 143 | 1.20 (0.74, 1.97) | 1.32 (0.77, 2.25) |
| Tertile 3 | 32.0 | 46 | 102 | 0.40 (0.20, 0.80) | 0.38 (0.17, 0.82) | 95.0 | 102 | 140 | 1.83 (1.14, 2.94) | 2.01 (1.19, 3.39) |
| Ptrend | .010 | .013 | .010 | .007 | ||||||
| Per SD increase | 0.63 (0.47, 0.85) | 0.63 (0.46, 0.86) | 1.12 (0.93, 1.35) | 1.12 (0.92, 1.37) | ||||||
aConditional logistic regression models were used. Liver enzymes levels were categorized into race-sex-specific tertiles among controls.
bAdjusted for age at enrollment, alcohol consumption, smoking status and pack-years.
cAdditionally adjusted for BMI, education levels, household income levels, self-reported history of Chronic obstructive pulmonary disease, self-reported history of hypertension, and self-reported history of diabetes.
d P trend was calculated using Wald test by treating tertiles as an ordinal variable.
eLogarithmic liver enzyme level divided by its SD was included in the models. SD, standard deviation.
Table 3.
Associations of serum liver enzyme levels with overall LC risk by smoking status, BMI, and alcohol consumption.a
| ALT | ALP | |||||||
|---|---|---|---|---|---|---|---|---|
| Median | Cases | Controls | OR (95% CI) b | Median | Cases | Controls | OR (95% CI) b | |
| Current smokers | ||||||||
| Tertile 1 | 10.0 | 105 | 91 | Ref | 58.0 | 120 | 151 | Ref |
| Tertile 2 | 18.0 | 87 | 103 | 0.80 (0.59, 1.08) | 78.0 | 150 | 167 | 1.19 (0.91, 1.55) |
| Tertile 3 | 33.0 | 64 | 101 | 0.66 (0.49, 0.90) | 101.5 | 137 | 163 | 1.11 (0.83, 1.48) |
| Ptrendc | .007 | .513 | ||||||
| Per SD increased | 0.84 (0.74, 0.95) | 1.03 (0.92, 1.16) | ||||||
| Never and former smokers | ||||||||
| Tertile 1 | 9.0 | 31 | 107 | Ref | 58.0 | 46 | 174 | Ref |
| Tertile 2 | 16.0 | 41 | 123 | 1.45 (0.86, 2.42) | 74.0 | 38 | 189 | 0.81 (0.50, 1.31) |
| Tertile 3 | 27.0 | 24 | 121 | 0.75 (0.43, 1.33) | 100.0 | 57 | 188 | 1.06 (0.67, 1.68) |
| Ptrend | .308 | .768 | ||||||
| Per SD increase | 0.97 (0.79, 1.20) | 1.03 (0.87, 1.22) | ||||||
| BMI <30 kg/mc | ||||||||
| Tertile 1 | 10.0 | 108 | 121 | Ref | 58.0 | 117 | 194 | Ref |
| Tertile 2 | 17.0 | 95 | 131 | 1.01 (0.74, 1.38) | 76.0 | 146 | 211 | 1.12 (0.83, 1.51) |
| Tertile 3 | 30.0 | 66 | 137 | 0.58 (0.41, 0.82) | 98.0 | 146 | 204 | 0.94 (0.67, 1.31) |
| Ptrend | .002 | .709 | ||||||
| Per SD increase | 0.79 (0.69, 0.90) | 0.97 (0.85, 1.09) | ||||||
| BMI ≥30 kg/mc | ||||||||
| Tertile 1 | 9.0 | 19 | 75 | Ref | 58.0 | 41 | 136 | Ref |
| Tertile 2 | 15.0 | 30 | 91 | 1.26 (0.81, 1.95) | 76.0 | 48 | 143 | 1.01 (0.66, 1.53) |
| Tertile 3 | 28.0 | 35 | 91 | 1.21 (0.79, 1.86) | 104.0 | 50 | 144 | 1.25 (0.83, 1.89) |
| Ptrend | .412 | .265 | ||||||
| Per SD increase | 1.08 (0.91, 1.29) | 1.07 (0.91, 1.25) | ||||||
| Alcohol drinker | ||||||||
| Tertile 1 | 10.0 | 74 | 102 | Ref | 57.0 | 92 | 165 | Ref |
| Tertile 2 | 18.5 | 74 | 118 | 1.01 (0.69, 1.47) | 74.0 | 113 | 181 | 0.98 (0.73, 1.31) |
| Tertile 3 | 33.0 | 55 | 114 | 0.79 (0.57, 1.11) | 98.0 | 117 | 178 | 1.01 (0.74, 1.37) |
| Ptrend | .175 | .937 | ||||||
| Per SD increase | 0.89 (0.78, 1.01) | 0.97 (0.87, 1.08) | ||||||
| Nondrinker | ||||||||
| Tertile 1 | 9.0 | 46 | 79 | Ref | 58.0 | 57 | 160 | Ref |
| Tertile 2 | 15.0 | 63 | 119 | 1.06 (0.66, 1.70) | 77.0 | 75 | 173 | 1.12 (0.75, 1.69) |
| Tertile 3 | 26.0 | 40 | 114 | 0.77 (0.48, 1.24) | 104.0 | 94 | 175 | 1.25 (0.87, 1.80) |
| Ptrend | .264 | .229 | ||||||
| Per SD increase | 0.93 (0.77, 1.13) | 1.08 (0.94, 1.25) | ||||||
| Excluded diabetes | ||||||||
| Tertile 1 | 9.0 | 97 | 146 | Ref | 57.0 | 125 | 250 | Ref |
| Tertile 2 | 16.0 | 113 | 172 | 1.00 (0.76, 1.32) | 74.0 | 153 | 276 | 1.01 (0.79, 1.28) |
| Tertile 3 | 30.0 | 76 | 168 | 0.68 (0.52, 0.89) | 99.0 | 172 | 267 | 1.11 (0.87, 1.42) |
| Ptrend | .004 | .382 | ||||||
| Per SD increase | 0.87 (0.78, 0.96) | 1.03 (0.94, 1.13) | ||||||
aGeneralized estimating equation models were used. Liver enzymes levels were categorized into race-sex-specific tertiles among controls.
bAdjusted for age at enrollment, alcohol consumption, smoking status and pack-years, BMI, education levels, household income levels, self-reported history of chronic obstructive pulmonary disease, self-reported history of hypertension, and self-reported history of diabetes.
c P trend was calculated using Wald test by treating tertiles as an ordinal variable.
dLogarithmic liver enzyme level divided by its SD was included in the models. SD, standard deviation.
Table 4.
Associations of serum liver enzyme levels with LC risk by histological subtypes.a
| ALT | ALP | |||||||
|---|---|---|---|---|---|---|---|---|
| Median | Cases | Controls | OR (95% CI)b | Median | Cases | Controls | OR (95% CI) b | |
| ADE | ||||||||
| Tertile 1 | 10.0 | 40 | 62 | Ref | 57.0 | 41 | 106 | Ref |
| Tertile 2 | 16.0 | 44 | 70 | 1.23 (0.56, 2.70) | 75.0 | 54 | 111 | 1.12 (0.62, 2.05) |
| Tertile 3 | 30.5 | 26 | 72 | 0.41 (0.15, 1.08) | 100.0 | 83 | 120 | 1.62 (0.94, 2.81) |
| Ptrendc | .060 | .072 | ||||||
| Per SD increased | 0.86 (0.59, 1.26) | 1.21 (0.96, 1.51) | ||||||
| Squamous cell | ||||||||
| Tertile 1 | 9.0 | 25 | 44 | Ref | 56.0 | 33 | 74 | Ref |
| Tertile 2 | 15.0 | 32 | 50 | 0.99 (0.35, 2.82) | 76.0 | 48 | 83 | 0.80 (0.37, 1.72) |
| Tertile 3 | 29.0 | 24 | 55 | 0.32 (0.09, 1.29) | 102.5 | 43 | 81 | 0.54 (0.24, 1.24) |
| Ptrend | .138 | .143 | ||||||
| Per SD increase | 0.55 (0.31, 0.98) | 0.80 (0.58, 1.11) | ||||||
aConditional logistic regression models were used. Liver enzymes levels were categorized into race-sex-specific tertiles among controls. ADE and squamous cell LC was defined based on ICD-10 code.
bAdjusted for age at enrollment, alcohol consumption, smoking status and pack-years, body mass index, education levels, household income levels, self-reported history of chronic obstructive pulmonary disease, self-reported history of hypertension, and self-reported history of diabetes.
c P trend was calculated using Wald test by treating tertiles as an ordinal variable.
dLogarithmic liver enzyme level divided by its SD was included in the models. SD, standard deviation.
Table 5.
Associations of serum liver enzyme levels with LC risk by time between blood collection and LC diagnosis.a
| ALT | ALP | |||||||
|---|---|---|---|---|---|---|---|---|
| Median | Cases | Controls | OR (95% CI)b | Median | Cases | Controls | OR (95% CI)b | |
| ≤Median Diagnosis timec | ||||||||
| Tertile 1 | 8.0 | 61 | 86 | Ref | 59.0 | 86 | 164 | Ref |
| Tertile 2 | 14.0 | 67 | 129 | 0.58 (0.31, 1.06) | 77.0 | 91 | 175 | 0.95 (0.62, 1.46) |
| Tertile 3 | 27.5 | 51 | 111 | 0.41 (0.19, 0.88) | 101.0 | 98 | 174 | 0.97 (0.63, 1.50) |
| Ptrendd | .020 | .892 | ||||||
| Per SD increasee | 0.69 (0.51, 0.93) | 1.04 (0.88, 1.24) | ||||||
| >Median diagnosis time | ||||||||
| Tertile 1 | 11.0 | 69 | 94 | Ref | 57.0 | 75 | 168 | Ref |
| Tertile 2 | 19.0 | 54 | 114 | 0.50 (0.26, 0.98) | 75.0 | 89 | 174 | 0.96 (0.60, 1.54) |
| Tertile 3 | 31.0 | 51 | 112 | 0.35 (0.17, 0.73) | 100.0 | 109 | 177 | 1.08 (0.69, 1.70) |
| Ptrend | .006 | .707 | ||||||
| Per SD increase | 0.65 (0.48, 0.88) | 1.01 (0.84, 1.22) | ||||||
| Diagnosed after 2 years of enrollment | ||||||||
| Tertile 1 | 10.0 | 97 | 146 | Ref | 57.0 | 115 | 269 | Ref |
| Tertile 2 | 18.0 | 84 | 159 | 0.64 (0.38, 1.08) | 76.0 | 164 | 284 | 1.24 (0.88, 1.76) |
| Tertile 3 | 32.0 | 74 | 163 | 0.48 (0.27, 0.87) | 101.0 | 163 | 285 | 1.08 (0.76, 1.53) |
| Ptrend | .016 | .736 | ||||||
| Per SD increase | 0.74 (0.59, 0.94) | 1.00 (0.87, 1.15) | ||||||
aConditional Logistic regression models were used. Liver enzymes levels were categorized into race-sex-specific tertiles among controls.
bAdjusted for age at enrollment, alcohol consumption, smoking status and pack-years, BMI, education levels, household income levels, self-reported history of chronic obstructive pulmonary disease, self-reported history of hypertension, and self-reported history of diabetes.
cThe median diagnosis time after enrollment is 39 (for ALT) and 56 months (for ALP).
d P trend was calculated using Wald test by treating tertiles as an ordinal variable.
eLogarithmic liver enzyme level divided by its SD was included in the models. SD, standard deviation.
Trends in tertile serum ALP levels were not associated with LC risk overall (ORT2 versus T1 = 1.05, 95% CI: 0.77–1.42; ORT3 versus T1 = 1.13, 95% CI: 0.83–1.53; Ptrend = .447) or in the analyses stratified by race and gender (Table 2). Interestingly, serum levels of ALP were associated with increased risk of LC among AA men (ORT2 versus T1 = 1.32, 95% CI: 0.77–2.25; ORT3 versus T1 = 2.01, 95% CI: 1.19–3.39; Ptrend = .007) (Table 2). No associations between serum ALP levels and LC risk were significant in the strata of smoking status, BMI, alcohol consumption, time between blood draw and diagnosis, or histological type (Tables 3 to 5).
4. Discussion
In our study investigating the associations of liver enzymes with LC risk among predominantly low-income Americans, we observed inverse associations between serum ALT levels and LC risk among both AAs and EAs. The inverse association between ALT levels and LC risk was more prominent among men, and was seen in both those diagnosed within and beyond a median follow-up time. Serum ALP levels were not associated with LC risk overall, but a significant positive association was observed among AA men.
An inverse but statistically nonsignificant relationship between circulating ALT levels and LC risk has been reported in Germany [14]. In addition, a significantly negative association between ALT and COPD was recently observed in the UK Biobank [20]. In our study, the ALT levels largely laid in clinically normal ranges. Although it is not clear why low ALT, indicative of the absence of liver disease, would be linked to higher LC risk, there are several possible explanations of the observed inverse association. First, previous studies have found ALT is positively related to total cholesterol (TC) and low-density lipoprotein cholesterol (LDL-c) [21, 22]. A meta-analysis showed a negative association between TC and LC risk [23]. High LDL-c levels have also been reported to be associated with reduced risk of SQC in addition to overall risk [24]. This may help explain our observed inverse dose-response relationship for SQC. Second, it has been reported that increased ALT levels were associated with elevated levels of serum transferrin and selenium [25]. Higher levels of transferrin and selenium have been associated with lower LC risk in meta-analyses [26, 27]. It is also possible that ALT is a proxy for still other biologic factors involved in lung carcinogenesis. Further studies are warranted to validate our results and clarify the underlying mechanisms.
Although interactive effects were not significant, the inverse associations appeared more evident among men, current smokers, and nonobese participants. A previous study showed that low serum ALT levels were associated with increased risk of COPD among males [28]. COPD is an intermediate phenotype of LC. ALT may protect individuals from developing LC through reducing the risk of COPD. Recent study conducted in UK Biobank also found an inverse association between ALT and LC risk among men but not women [15]. Alternatively, obesity is inversely associated with LC risk [29, 30] and mounting evidence shows a positive association between ALT and obesity [31]. High ALT levels may indicate poor liver function and/or liver diseases such as nonalcoholic fatty liver disease, which may lead to abnormalities in energy metabolism and then induce obesity. Sex differences might exist in this process [32], such as estrogens negatively link to the risk of nonalcoholic fatty liver disease [33]. Future studies are needed to confirm our results and elucidate the underlying mechanisms.
The inverse association of ALT and overall LC risk was significant among current smokers. Smoking can lead to inflammation, which promotes the development of LC. Previous researchers found a positive relationship between IL10, an anti-inflammatory biomarker, and ALT [34]. High ALT may inspire the anti-inflammatory functions, thereby reducing the risk of LC. Future research is needed to explore the anti-inflammatory roles of ALT. Moreover, we found a significant inverse relationship between serum levels of ALT and overall LC risk among nonobese participants. Further studies are needed to investigate the roles of obesity in association with ALT and LC risk.
Serum ALP levels were not associated with overall LC risk in our study population. However, we observed a significant positive association between serum ALP levels and LC risk among AA males. Previous studies also showed that ALP was positively associated with lung and colorectal cancer [14, 15, 35]; these studies were based on European populations. In a recent UK biobank study, Du et al. found a positive association between ALP and COPD in men [20]. A study from the UK Biobank also found that the positive association between ALP and LC risk was stronger among men than women [15]. ALP is positively associated with higher C-reactive protein and leukocytes counts, indicating that high ALP levels are correlated with chronic inflammation, a putative risk factor for LC [36]. Moreover, there is no correlation between ALT and ALP in our study, consistent with earlier studies. ALT and ALP might characterize different patterns. Not only liver, but also bone generates ALP. A high level of ALP is an indicator of low bone mass density. The relationship between low bone mass density and increased risk of NSCLC was observed [37]. But the underlying mechanisms of why ALP is linked to LC risk are largely unclear. Further studies should be conducted to explore the underlying mechanisms.
Our study has several strengths; to the best of our knowledge, this is the first study evaluating prediagnostic serum levels of liver enzymes and subsequent LC risk among predominantly low-income AAs and EAs, populations at elevated risk of both LC and liver diseases. The nested case-control study design and relatively large number of cases provide more statistical power to assess relationships with the liver enzyme markers. Comprehensive covariate information from the SCCS allowed us to adjust for smoking and other major confounders. Our study was prospective in design and used blood samples collected before LC diagnosis and treatment, thus minimizing potential reverse causation. However, there are several limitations. We had only a single blood sample collected at cohort entry and thus could not evaluate any changes in ALT or ALP levels over time. Because most participants had clinically normal levels of ALT or ALP, we were not able to assess the associations of abnormalities with the risk of LC. We did not have measurements of some other blood biomarkers, such as TC and other lipid levels, to assess confounding by these factors. While the number of cases was sizeable, it was not adequate to assess moderate interactive effects. Larger sample sizes are required for well-powered stratified analyses. Another concern is that we cannot eliminate the residual confounding effects, although differences between our basic and fully adjusted models tended to be small. Finally, our study had a median follow-up time of 39 months for ALT analyses and 56 months for ALP analyses. Thus, we were not able to evaluate the associations of ALT and ALP levels with LC risk with a longer follow-up time.
In summary, we found that higher serum levels of ALT were associated with a reduced risk of overall LC among a predominantly low-income AA and EA population, while higher serum levels of ALP were associated with increased risk of LC among AA men. Further studies are warranted to confirm our findings and elucidate the potential underlying mechanisms of the associations.
Supplementary Material
Acknowledgement
The authors appreciate the contributions and efforts of the researchers and staff of SCCS. We also thank Ms. Kathleen Harmeyer for assistance with manuscript editing and preparation.
Contributor Information
Shuai Xu, Division of Epidemiology, Department of Medicine, Vanderbilt Epidemiology Center, Vanderbilt-Ingram Cancer Center, Vanderbilt University Medical Center, Nashville, TN 37203, United States.
Hui Cai, Division of Epidemiology, Department of Medicine, Vanderbilt Epidemiology Center, Vanderbilt-Ingram Cancer Center, Vanderbilt University Medical Center, Nashville, TN 37203, United States.
Jie Wu, Division of Epidemiology, Department of Medicine, Vanderbilt Epidemiology Center, Vanderbilt-Ingram Cancer Center, Vanderbilt University Medical Center, Nashville, TN 37203, United States.
Jiajun Shi, Division of Epidemiology, Department of Medicine, Vanderbilt Epidemiology Center, Vanderbilt-Ingram Cancer Center, Vanderbilt University Medical Center, Nashville, TN 37203, United States.
Regina Courtney, Division of Epidemiology, Department of Medicine, Vanderbilt Epidemiology Center, Vanderbilt-Ingram Cancer Center, Vanderbilt University Medical Center, Nashville, TN 37203, United States.
Hyung-Suk Yoon, Division of Epidemiology, Department of Medicine, Vanderbilt Epidemiology Center, Vanderbilt-Ingram Cancer Center, Vanderbilt University Medical Center, Nashville, TN 37203, United States; University of Florida Health Cancer Center, Gainesville, FL 32610, United States; Department of Surgery, College of Medicine, University of Florida, Gainesville, FL 32610, United States.
Xiao-Ou Shu, Division of Epidemiology, Department of Medicine, Vanderbilt Epidemiology Center, Vanderbilt-Ingram Cancer Center, Vanderbilt University Medical Center, Nashville, TN 37203, United States.
William J Blot, Division of Epidemiology, Department of Medicine, Vanderbilt Epidemiology Center, Vanderbilt-Ingram Cancer Center, Vanderbilt University Medical Center, Nashville, TN 37203, United States.
Wei Zheng, Division of Epidemiology, Department of Medicine, Vanderbilt Epidemiology Center, Vanderbilt-Ingram Cancer Center, Vanderbilt University Medical Center, Nashville, TN 37203, United States.
Qiuyin Cai, Division of Epidemiology, Department of Medicine, Vanderbilt Epidemiology Center, Vanderbilt-Ingram Cancer Center, Vanderbilt University Medical Center, Nashville, TN 37203, United States.
Author contributions
Q.C. conceived and supervised the study. S.X., H.C., J.W., H.-S.Y., J.S., and Q.C. analysed and interpreted data. J.W. and R.C. conducted laboratory analyses. S.X. and Q.C. drafted the original manuscript. All the authors reviewed and revised the manuscript.
Supplementary data
Supplementary data is available at Carcinogenesis online.
Funding
The SCCS was supported by the National Institutes of Health (U01CA202979). Q.C. was partially supported by a grant from the National Institutes of Health (R01MD015396). H.-S.Y. was partially supported by a grant from the National Institutes of Health (R03CA273625). Data collection and sample preparation were performed by the Survey and Biospecimen Shared Resource, which is supported in part by the Vanderbilt-Ingram Cancer Center (P30CA068485).
Data availability
Data used in the present study are available upon request from the Southern Community Cohort Study at http://ors.southerncommunitystudy.org.
Reference
- 1. Siegel RL, Giaquinto AN, Jemal A. Cancer statistics, 2024. CA Cancer J Clin 2024;74:12–49. 10.3322/caac.21830 [DOI] [PubMed] [Google Scholar]
- 2. Sung H, Ferlay J, Siegel RL et al. Global Cancer Statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2021;71:209–49. 10.3322/caac.21660 [DOI] [PubMed] [Google Scholar]
- 3. Stram DO, Park SL, Haiman CA et al. Racial/ethnic differences in lung cancer incidence in the Multiethnic Cohort Study: an update. J Natl Cancer Inst 2019;111:811–9. 10.1093/jnci/djy206 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Zeng H, Yuan Z, Zhang G et al. Racial disparities in histological subtype, stage, tumor grade and cancer-specific survival in lung cancer. Transl Lung Cancer Res 2022;11:1348–58. 10.21037/tlcr-21-794 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Arauz RF, Mayer M, Reyes-Guzman C et al. Racial disparities in cigarette smoking behaviors and differences stratified by metropolitan area of residence. Int J Environ Res Public Health 2022;19:2910. 10.3390/ijerph19052910 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Wang M, Liu Y, Ma Y et al. Association between cancer prevalence and different socioeconomic Strata in the US: the National Health and Nutrition Examination Survey, 1999–2018. Front Public Health 2022;10:873805. 10.3389/fpubh.2022.873805 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Byun J, Han Y, Li Y et al. Cross-ancestry genome-wide meta-analysis of 61,047 cases and 947,237 controls identifies new susceptibility loci contributing to lung cancer. Nat Genet 2022;54:1167–77. 10.1038/s41588-022-01115-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Zhang X, Jiang N, Wang L et al. Chronic obstructive pulmonary disease and risk of lung cancer: a meta-analysis of prospective cohort studies. Oncotarget 2017;8:78044–56. 10.18632/oncotarget.20351 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Mantovani A, Petracca G, Beatrice G et al. Non-alcoholic fatty liver disease and increased risk of incident extrahepatic cancers: a meta-analysis of observational cohort studies. Gut 2022;71:778–88. 10.1136/gutjnl-2021-324191 [DOI] [PubMed] [Google Scholar]
- 10. Kalaitzakis E, Gunnarsdottir SA, Josefsson A et al. Increased risk for malignant neoplasms among patients with cirrhosis. Clin Gastroenterol Hepatol 2011;9:168–74. 10.1016/j.cgh.2010.10.014 [DOI] [PubMed] [Google Scholar]
- 11. Ponvilawan B, Charoenngam N, Rujirachun P et al. Chronic Hepatitis C virus infection is associated with an increased risk of lung cancer: a systematic review and meta-analysis. Lung 2020;198:705–14. 10.1007/s00408-020-00365-y [DOI] [PubMed] [Google Scholar]
- 12. Hilliard KL, Traber AE, E K et al. The lung-liver axis: a requirement for maximal innate immunity and hepatoprotection during pneumonia. Am J Respir Cell Mol Biol 2015;53:378–90. 10.1165/rcmb.2014-0195OC [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Young RP, Hopkins RJ, Marsland B. The gut-liver-lung axis. Modulation of the innate immune response and its possible role in chronic obstructive pulmonary disease. Am J Respir Cell Mol Biol 2016;54:161–9. 10.1165/rcmb.2015-0250PS [DOI] [PubMed] [Google Scholar]
- 14. Katzke V, Johnson T, Sookthai D et al. Circulating liver enzymes and risks of chronic diseases and mortality in the prospective EPIC-Heidelberg case-cohort study. BMJ Open 2020;10:e033532. 10.1136/bmjopen-2019-033532 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Christakoudi S, Tsilidis KK, Gunter MJ et al. Allometric fat mass index and alanine aminotransferase attenuate the associations of platelet parameters with lung cancer risk. Sci Rep 2024;14:26318. 10.1038/s41598-024-78281-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Signorello LB, Hargreaves MK, Steinwandel MD et al. Southern community cohort study: establishing a cohort to investigate health disparities. J Natl Med Assoc 2005;97:972–9. https://pmc.ncbi.nlm.nih.gov/articles/PMC2569308 [PMC free article] [PubMed] [Google Scholar]
- 17. Yoon H-S, Shu X-O, Shidal C et al. Associations of Pre-diagnostic Serum levels of total bilirubin and albumin with lung cancer risk: results from the Southern Community Cohort Study. Front Oncol 2022;12:895479. 10.3389/fonc.2022.895479 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Schneider ALC, Lazo M, Ndumele CE et al. Liver enzymes, race, gender and diabetes risk: the Atherosclerosis Risk in Communities (ARIC) study. Diabet Med 2013;30:926–33. 10.1111/dme.12187 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Lin I-F, Lai M-Y, Chuang P-H. Analysis of matched case-control data with incomplete Strata: applying longitudinal approaches. Epidemiology 2007;18:446. 10.1097/EDE.0b013e318064630a [DOI] [PubMed] [Google Scholar]
- 20. Du W, Guan H, Wan X et al. Circulating liver function markers and the risk of COPD in the UK Biobank. Front Endocrinol 2023;14:1121900. 10.3389/fendo.2023.1121900 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Kathak RR, Sumon AH, Molla NH et al. The association between elevated lipid profile and liver enzymes: a study on Bangladeshi adults. Sci Rep 2022;12:1711. 10.1038/s41598-022-05766-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Liu C, Shao M, Lu L et al. Obesity, insulin resistance and their interaction on liver enzymes. PLoS One 2021;16:e0249299. 10.1371/journal.pone.0249299 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Lin X, Lu L, Liu L et al. Blood lipids profile and lung cancer risk in a meta-analysis of prospective cohort studies. J Clin Lipidol 2017;11:1073–81. 10.1016/j.jacl.2017.05.004 [DOI] [PubMed] [Google Scholar]
- 24. Carreras-Torres R, Johansson M, Haycock PC et al. Obesity, metabolic factors and risk of different histological types of lung cancer: a Mendelian randomization study. PLoS One 2017;12:e0177875. 10.1371/journal.pone.0177875 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Ruhl CE, Everhart JE. Relation of elevated serum alanine aminotransferase activity with iron and antioxidant levels in the United States. Gastroenterology 2003;124:1821–9. 10.1016/s0016-5085(03)00395-0 [DOI] [PubMed] [Google Scholar]
- 26. Wang Q, Cui Q, Gao J-P et al. Role of iron biomarkers and iron intakes in lung cancer risk: a systematic review and meta-analysis. J Trace Elem Med Biol 2022;74:127060. 10.1016/j.jtemb.2022.127060 [DOI] [PubMed] [Google Scholar]
- 27. Cai X, Wang C, Yu W et al. Selenium exposure and cancer risk: an updated meta-analysis and meta-regression. Sci Rep 2016;6:19213. 10.1038/srep19213 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Choi YJ, Kwon DS, Kim T et al. Low alanine aminotransferase as a risk factor for chronic obstructive pulmonary disease in males. Sci Rep 2021;11:14829. 10.1038/s41598-021-94385-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Zhou W, Liu G, Hung RJ et al. Causal relationships between body mass index, smoking and lung cancer: univariable and multivariable Mendelian randomization. Int J Cancer 2021;148:1077–86. 10.1002/ijc.33292 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Yu D, Zheng W, Johansson M et al. Overall and central obesity and risk of lung cancer: a pooled analysis. J Natl Cancer Inst 2018;110:831–42. 10.1093/jnci/djx286 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Ali N, Sumon AH, Fariha KA et al. Assessment of the relationship of serum liver enzymes activity with general and abdominal obesity in an urban Bangladeshi population. Sci Rep 2021;11:6640. 10.1038/s41598-021-86216-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Pasquali R. Obesity and androgens: facts and perspectives. Fertil Steril 2006;85:1319–40. 10.1016/j.fertnstert [DOI] [PubMed] [Google Scholar]
- 33. Ballestri S, Nascimbeni F, Baldelli E et al. NAFLD as a sexual dimorphic disease: role of gender and reproductive Status in the development and progression of nonalcoholic fatty liver disease and inherent cardiovascular risk. Adv Ther 2017;34:1291–326. 10.1007/s12325-017-0556-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. El-Emshaty HM, Nasif WA, Mohamed IE. Serum cytokine of IL-10 and IL-12 in chronic liver disease: the immune and inflammatory response. Dis Markers 2015;2015:707254. 10.1155/2015/707254 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. He M-M, Fang Z, Hang D et al. Circulating liver function markers and colorectal cancer risk: a prospective cohort study in the UK Biobank. Int J Cancer 2021;148:1867–78. 10.1002/ijc.33351 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Seo M-S, Shim J-Y, Lee Y-J. Relationship between serum alkaline phosphatase level, C-reactive protein and leukocyte counts in adults aged 60 years or older. Scand J Clin Lab Invest 2019;79:233–7. 10.1080/00365513.2019.1585567 [DOI] [PubMed] [Google Scholar]
- 37. Hatlen P, Langhammer A, Forsmo S et al. Bone mass density, fracture history, self-reported osteoporosis as proxy variables for estrogen and the risk of non-small-cell lung cancer–a population based cohort study, the HUNT study: are proxy variables friends or faults? Lung Cancer 2013;81:39–46. 10.1016/j.lungcan.2013.04.001 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data used in the present study are available upon request from the Southern Community Cohort Study at http://ors.southerncommunitystudy.org.