Abstract
Background
The association between total testosterone (T) and chronic obstructive pulmonary disease (COPD), remains poorly understood. We aim to investigate this association and how it varies by smoking status, body fatness, and race/ethnicity in a nationally representative sample of American men.
Methods
Data included a full sample (NHANES 1988–1991, 1999–2004, 2011–2012) and subset sample (excluding 2011–2012, no estradiol and SHBG levels available) of 2748 and 906 men (≥20 years), respectively. COPD was measured by self-report or spirometry test. Total T (ng/mL) was measured among men who participated in a morning examination session. Weighted multivariable-adjusted logistic regression models were conducted.
Results
Low T was positively associated with self-reported COPD in the full sample (OR = 2.10, 95% CI = 1.18–3.74, Ptrend = 0.010), and when stratified by current smokers and body fatness. When examined across race and ethnicity strata, this association persisted among White men (OR = 2.50, 95% CI = 1.30–4.79, Ptrend = 0.002) but not among Hispanic or Black men. In the subset sample, low T was positively associated with self-reported COPD (OR = 1.42, 95% CI, 0.57,3.55, Ptrend = 0.04), including among smokers and White men, but not body fatness. No significant associations were observed with COPD defined with spirometry plus self-report.
Conclusion
Low levels of T were associated with an increased prevalence of self-reported COPD in the full and subset samples. Similar associations were observed after stratifying by smoking status, body fatness, and race/ethnicity in the full sample and subset sample. Prospective studies are warranted to confirm these significant associations among understudied and underserved populations.
Keywords: Serum testosterone, COPD, Race, Ethnicity, Smoking
Introduction
Chronic obstructive pulmonary disease (COPD) affects an estimated 300 million people globally and is one of the leading causes of death in the USA [1, 2]. COPD is a progressive chronic disease of the lungs, and it is typically associated with an increased risk of poor quality of life, increased hospitalizations, and increased mortality [3, 4]. Definitive risk factors for COPD include smoking, exposure to environmental tobacco smoke, occupational exposure to pollutants and chemicals, and pneumonia or other childhood respiratory infection [5]. However, other potential risk factors have been reported, such as obesity, race and ethnicity, and low serum testosterone (T) levels [3, 6–9].
Serum T is reported to have beneficial effects on sexual function, muscle strength, and lean body mass [10, 11]. Moreover, low T or T deficiency (total T ≤ 300 ng/dL) [12] has been associated with cardiovascular disease (CVD) [13], prostate cancer (PCa), and all-cause mortality [14–17]. Over the past decade, it has been observed that around 38.7% of men above 45 years have low T [18, 19], while approximately 2.4 million US men in the 40–69-year age group have been reported to have low T [20]. The estimated projection of low T in the 30- to 80-year age group for men stands at an approximate 6.5 million by the year 2025. The above-described increasing prevalence of low T is attributed, in part, to the aging of the US population and the obesity epidemic [18, 21].
Although previous studies of the association between COPD and low T have provided valuable insights, they often included only small samples of understudied and underserved populations thus having limited generalizability to non-Hispanic (NH)-Black and Hispanic men [3, 8, 9]. Previous studies have demonstrated racial and ethnic differences (NH-White, NH-Black, and Hispanics) with total T levels among adult and adolescent men [22, 23]. Numerous studies have shown African Americans with COPD to have worse outcomes, as well as those with other pulmonary diseases, such as asthma, which lead to higher hospitalization and mortality rates compared to other racial groups [24–26].
Therefore, the objectives of this study were to investigate the association of total T with prevalent self-reported COPD in a full (1988–1991, 1999–2004, 2011–2012) and subset sample (excluding 2011–2012, no estradiol and SHBG levels available) of NHANES, and combined with diagnosed COPD (spirometry test only available 1988–1991 and 2011–2012). We further determined whether these associations vary by smoking status, body fatness, and ethnicity in a US nationally representative sample of NH-White, NH-Black, and Hispanic men.
Methods
Study population
The National Health and Nutrition Examination Survey (NHANES) is a program of studies undertaken by the National Center for Health Statistics (NCHS) of the US Centers for Disease Control and Prevention (CDC) to assess the health and nutritional status of adults and children in the USA [27]. NHANES is a cross-sectional study that uses a multistage, stratified, and clustered probability sampling strategy in which Hispanics, NH-Blacks, and the elderly are oversampled to ensure adequate sample size and to represent the total US civilian, non-institutionalized population [28]. Details of the survey design, methods, and data collections are available on the NHANES website (https://wwwn.cdc.gov/nchs/nhanes/Default.aspx. Accessed Jan. 2020).
This study included data from males aged ≥ 20 years in the 1988–1991 (NHANES III-Phase 1), 1999–2004, and 2011–2012 NHANES cycles. A subset sample was chosen to include estradiol and SHBG, as the 2011–2012 cycles did not include these variables.
The spirometry values of FEV1/FVC were available for the NHANES III as well as the 2011–2012 cycles, while self-reported physician diagnosis of COPD was available in the 1988–1991, 1999–2004, and 2011–2012 cycles. Participants with missing information from covariates and missing sex hormone measurements and having extreme hormone measurements based on identification of outliers from a graphical representation of the distribution of each hormone were excluded. The following threshold points were used to determine extreme measurements based on a visual inspection of the distribution: total T = 50 ng/mL (n = 5), estradiol = 360 pg/mL (n = 6), and sex hormone binding globulin (SHBG) > 170 nmol/L (n = 8), leaving a final full sample of 2748 men. The NHANES 2011–2012 cycle did not include estradiol and SHBG, and a subset sample of 1988–1991 and 1999–2004 cycles was created to be able to adjust for estradiol and SHBG (n = 906).
Assessment of COPD
Patient history of COPD was mainly evaluated by self-reported physician diagnosis of COPD, since the spirometry test was not available in the NHANES 1999–2004 cycles. Self-reported COPD was confirmed largely by NHANES questions that ask, “has a doctor or other health professional ever informed that you have the disease?” A positive response to either chronic bronchitis or emphysema and a negative response to current asthma were considered as physician-diagnosed COPD [29], also referred to as the self-report of COPD in our study. Spirometry data from the NHANES were used to diagnose persons with varying levels of lung obstruction. This was performed as per the recommendations of the American Thoracic Society [30], and the procedures undertaken have been published in previous studies [31]. Participants were identified as having COPD when the ratio between the forced expiratory volume in one second (FEV1) and the forced vital capacity (FVC) were less than 0.7, which indicates the presence of obstructive pulmonary disease [32].
Assessment of testosterone, estradiol, and SHBG
Endogenous sex hormones were measured from stored surplus serum samples of male participants in the morning examination sessions of each cycle, the subjects being stratified at random. Morning sample participants were chosen to reduce extraneous variation due to diurnal production of hormones.
Details on the blood draw, process, storage, and shipping methods have previously been reported [28]. In NHANES 1988–1991 and 1999–2004, total T, estradiol, and SHBG were measured using electrochemiluminescence immunoassays on the 2010 Elecsys system (Roche Diagnostics, Laval, QC, Canada; and Roche Diagnostics, Indianapolis, IN, USA). In the NHANES 2011–2012 cycle, total T determination was performed using HPLC tandem mass spectrophotometry. T levels were categorized into tertiles as low < 230 ng/dL, normal 230-400 ng/dL, and high ≥ 400 ng/dL.
The laboratory technicians were blinded to the participant characteristics of the samples. The lower limits of detection of the assays were 2 ng/dL for T, 5 pg/mL for estradiol, and 3 nmol/L for SHBG.
Assessment of covariates
Age, race/ethnicity, cigarette smoking, alcohol consumption, and physical activity during the past 30 days were self-reported during the NHANES interviews. Age was further grouped as 20–44, 44–64, and ≥65 years. BMI was calculated as weight in kilograms divided by height in meters squared. Overall obesity was defined by BMI ≥ 30 kg/m2. Education level was classified as high school, high school graduate or General Equivalency Diploma (GED), college degree, or above. The income-to-poverty ratio (PIR) was used as a measure of family income, with outcomes grouped as <1.30 (considered low income), 1.30–3.49, and ≥ 3.50. A current smoker was considered as someone who reportedly smoked at least 100 cigarettes in their lifetime and smoked every day or some days or has a serum cotinine value greater than 10 ng/mL currently [33]. A former smoker was considered as someone who smoked at least 100 cigarettes during their lifetime but currently does not smoke. A never smoker was considered as someone who reportedly smoked fewer than 100 cigarettes during their lifetime.
Statistical analysis
Sampling weights were applied to account for selection probabilities, over-sampling, non-response, and differences between the sample and the total US population. Quantitative variables are presented using mean (95% CI), and geometric means were included as serum testosterone does not follow a normal distribution: it was thus estimated for total T, calculated free T, estradiol, and SHBG concentrations in the full and subset samples. The assessment of covariates has previously been described [28]. For descriptive purposes, the comparison of the distribution of sociodemographic and lifestyle factors by full and subset samples was carried out using the t-test statistic for independent samples (Table 1; Supplemental Table 1).
Table 1.
Selected characteristics of the US population of adult men 20 years and older in the full sample in NHANES 1988–1991, 1999–2004, 2011–2012
| Characteristics | Low testosterone (<230 ng/dL) N = 311 |
Normal testosterone (230–<400 ng/dL) N = 962 |
High testosterone (≥400 ng/dL) N = 1475 |
|---|---|---|---|
| Mean (95% CI) | Mean (95% CI) | Mean (95% CI) | |
| Age (years) | 50.4 (48.6–52.1) | 47.7 (46.5–48.8) | 42.0 (41.2–42.8) |
| Testosterone (ng/dL)a | 167.8 (162.4–173.1) | 320.9 (318.5–323.3) | 597.4 (586.2–608.7) |
| *Estradiol (pg/mL)a | 28.7 (24.1–33.4) | 29.8(28.6–31.1) | 38.3 (37.2–39.4) |
| *SHBG (nmol/L)a | 33.3 (27.6–39.0) | 29.5(28.4–30.6) | 40.3 (39.1–41.5) |
| *Free Testosteronea | 0.036 (0.032–0.040) | 0.073 (0.071–0.075) | 0.124 (0.120–0.128) |
| **FEV1/FVC | 0.75 (0.74–0.77) | 0.77 (0.76–0.78) | 0.77 (0.77–0.78) |
| **FEV1/FVC <0.7 | 22.4 (18.2–26.6) | 18.1 (14.7–21.4) | 16.8 (14.1–19.5) |
| Self-report COPD | 12.0 (7.7–16.0) | 7.0 (4.9–8.7) | 5.0 (4.0–5.8) |
| **Self-Report or FEV1/FVC<0.7 COPD | 25.7 (20.9–30.5) | 20.9 (17.0–24.9) | 19.7 (17.0–22.4) |
| Smoking status | |||
| Never | 50.1 (43.5–56.7) | 44.7 (40.9–48.4) | 41.5 (39.2–43.9) |
| Past | 33.1 (27.3–38.8) | 37.3 (33.7–40.9) | 24.2 (21.9–26.5) |
| Current | 16.8 (13.0–20.6) | 18.0 (14.9–21.2) | 34.3 (32.0–36.5) |
| Body fatness | |||
| BMI (≥30) | 55.1 (49.9–60.3) | 42.6 (39.9–45.3) | 17.4 (15.5–19.3) |
| Waist circumference (≥102 cm) | 68.0 (62.6–73.4) | 55.8 (52.1–59.5) | 25.1 (23.4–26.9) |
| Race | |||
| Hispanic | 12.7 (7.9–17.4) | 15.6 (11.7–19.6) | 11.7 (9.5–14.0) |
| Non-Hispanic Whites | 76.7 (70.5–82.9) | 75.8 (70.9–80.8) | 77.3 (74.3–80.3) |
| Non-Hispanic Blacks | 10.6 (7.3–14.0) | 8.5 (6.3–10.8) | 11.0 (9.1–12.9) |
| Alcohol status | |||
| 0–1 drinks/day | 19.2 (13.9–24.5) | 12.1 (10.2–13.9) | 11.3 (9.6–13.0) |
| 1–2 drinks/day | 22.8 (18.3–27.2) | 17.3 (14.7–19.9) | 12.7 (11.1–14.3) |
| >2 drinks/day | 20.6 (14.9–26.3) | 31.1 (27.5–34.8) | 22.5 (20.5–24.5) |
| Self-reported health | |||
| Excellent/very good health | 36.2 (29.5–42.9) | 48.9 (45.6–52.2) | 57.5 (55.1–59.8) |
| Good health | 42.3 (35.7–48.8) | 34.8 (31.9–37.8) | 30.4 (28.2–32.6) |
| Fair/poor health | 21.6 (17.8–25.4) | 16.3 (14.0–18.6) | 12.1 (10.4–13.8) |
| Physical activity | |||
| Moderate exercise | 40.5 (33.3–47.6) | 46.9 (42.9–50.9) | 50.4 (47.2–53.6) |
| Vigorous exercise | 17.1 (11.6–22.5) | 29.2 (25.6–32.8) | 31.3 (28.7–33.8) |
| Family income (PIR) | |||
| <1.30 | 19.38 (14.37–24.39) | 19.01 (15.95–22.08) | 20.59 (17.92–23.25) |
| 1.30–3.49 | 29.61 (24.12–35.10) | 32.77 (29.67–35.87) | 32.40 (28.90–35.90) |
| ≥ 3.50 | 5.23 (2.61–7.85) | 6.51 (4.57–8.45) | 5.04 (3.67–6.42) |
| Educational level | |||
| High school | 1.89 (0.19–3.59) | 3.14 (1.96–4.32) | 3.74 (2.73–4.76) |
| High school graduate or GED | 4.56 (1.94–7.17) | 6.36 (4.47–8.24) | 9.86 (8.21–11.52) |
| Some college or above | 0.79 (0.11–1.68) | 3.74 (2.67–4.81) | 7.20 (5.52–8.87) |
| Employment status | |||
| Unemployed | 29.53 (23.77–35.28) | 24.54 (21.84–27.24) | 19.94 (17.73–22.16) |
| Employed | 59.53 (53.66–65.41) | 63.84 (60.32–67.36) | 64.88 (61.80–67.95) |
Abbreviations: NHANES, National Health and Nutrition Examination Survey
NHANES III (Phase I) and 1999–2004
NHANES III (Phase I) and 2011–2012
Geometric mean (95% confidence interval)
We used weighted logistic regression models to estimate multivariable-adjusted odd ratios (ORs) and 95% CIs for self-reported COPD or spirometry associated independently with total T. Multivariable analyses were conducted in two models, namely, full sample (1988–1991, 1999–2004, and 2011–2012) and subset sample (1988–1991 and 1999–2004). Full sample models were adjusted for age, race, BMI, waist circumference, self-reported health, smoking status, moderate exercise, vigorous exercise, and alcohol consumption. Similarly, subset sample models were adjusted for the risk factors mentioned above plus estradiol and SHBG (not included in the NHANES 2011–2012 wave). Since total T and estradiol have previously been correlated [34] and SHBG is a major carrier of circulating T, they were included in the multivariable models. To test for a linear trend across categories of total T, we modeled the categories of total T as continuous variables using the median for each category. A sensitivity analysis was performed with the outcome being either spirometry or self-reported COPD.
Stratified and multivariable analyses were conducted by race and ethnicity (NH-White, NH-Black, and Hispanic), since this factor has been observed to modify T levels [22, 23]. All p-values were two-sided; alpha = 0.05 was considered the cut-off for statistical significance. We incorporated multiplicative interaction terms into the models and tested them using the Wald test. All statistical analyses were performed using SAS (SAS Institute v.9.4, Cary, NC).
Results
In the full and subset samples, we identified 2748 and 906 men, respectively. A total of 2107 men were NHWs (76.7%), 291 NHBs (10.6%), and 348 Hispanics (12.7%) in the full sample with similar percentages in the subset sample. The distribution of baseline characteristics by categories of total T in the full sample and subset samples, after applying sample weights, is reported in Table 1 and Supplemental Table 1, respectively.
The mean age in the full sample and subset sample was 46.7 and 46.1 years, respectively. The average FEV1/FVC ratio in the NHANES III (Phase I) 2011–2012 cycles had lower values in the low T category compared to the normal and high T categories. Participants in the low T category had the highest percentage of FEV1/FVC < 0.7 of 22.4% compared to the normal and high T categories, which had 18.1% and 16.8 %, respectively (Table 1), as seen in the NHANES III (Phase I), 2011–2012 cycles. A self-report of COPD in the full sample was seen in 12% of low T participants, while the normal and high T participants had lesser self-report of COPD with 7 and 5%, respectively. Similar representations were seen with either self-report or FEV1/FVC < 0.7 in all three groups in the NHANES III (phase I) 2011–2012 cycles (Table 1). The baseline characteristics were further classified based on smoking status, body fatness, race, alcohol status, and physical activity: the participants in the low T category were seen to have greater body fatness (55% had BMI ≥ 30 and 68% had waist circumference ≥ 102 cm) and were less physically active, as shown in Table 1.
Figure 1 shows the multivariable associations of total T with self-reported COPD. In the overall stratified analysis, we found positive association of low T (OR = 2.10, 95% CI = 1.18–3.74, Ptrend = 0.010) with self-reported COPD in the full sample. We found similar positive associations between low T and self-reported COPD when we stratified by current smokers (OR = 3.68, 95% CI = 1.36–9.94, Ptrend = 0.008). Similar findings were identified when we stratified by body fatness among those with waist circumference ≥ 102 cm (OR = 2.87, 95% CI =1.31–6.31, Ptrend = 0.010) and BMI ≥ 30 (OR = 2.73, 95% CI =1.07–6.95, Ptrend = 0.050). The associations of total T with self-reported COPD stratified by race and ethnicity in the full sample are shown in Table 2. We found significant associations for NH-Whites (OR = 2.50, 95% CI = 1.30–4.79, Ptrend = 0.002) in the fully adjusted models, but not among Hispanics and NH- Blacks.
Fig. 1.
Multivariable associations of total T with self-reported COPD and stratified by smoking and body fatness — NHANES III Phase I, 1999–2004, and 2011–2012. Overall results adjusted for age, race, family income, educational level, employment status, BMI, waist circumference, self-reported health, smoking status, moderate exercise, vigorous exercise, and alcohol consumption. Results for stratified analysis were adjusted for all the same covariates as overall analysis, except the covariate used for stratification was removed from the model
Table 2.
Testosterone level association with self-reported COPD stratified by age group, NHANES III-PHASE 1, 1999–2004, 2011–2012 (full sample)
| Testosterone level–OR (95% CI) |
||||||||
|---|---|---|---|---|---|---|---|---|
| High (>400 mg/dL) |
Normal (230–400 mg/dL) |
Low (<230 mg/dL) |
P-TREND | P-Interaction | ||||
| OR | 95% CI | OR | 95% CI | OR | 95% CI | 0.83 | ||
| Hispanic | ||||||||
| Model 1 | 1.0 | REF | 0.39 | 0.12, 1.26 | 1.18 | 0.16, 8.83 | 0.7 | |
| Model 2 | 1.0 | REF | 0.44 | 0.15, 1.30 | 1.28 | 0.14, 11.74 | 0.73 | |
| Non-Hispanic White | ||||||||
| Model 1 | 1.0 | REF | 1.34 | 0.86, 2.10 | 2.32 | 1.31, 4.12 | 0.003 | |
| Model 2 | 1.0 | REF | 1.42 | 0.90, 4.79 | 2.50 | 1.30, 4.79 | 0.002 | |
| Non-Hispanic Black | ||||||||
| Model 1 | 1.0 | REF | 0.62 | 0.35, 1.11 | 0.69 | 0.28, 1.70 | 0.15 | |
| Model 2 | 1.0 | REF | 0.60 | 0.29, 1.25 | 0.57 | 0.21, 1.54 | 0.13 | |
Model 1: adjusted for age. Model 2: adjusted for age, family income, educational level, employment status, BMI, waist circumference, self-reported health, smoking status, moderate exercise, vigorous exercise, and alcohol consumption
The prevalence of self-reported COPD adjusted for estradiol and SHBG for the subset analysis of NHANES III Phase I, 1999–2004 cycles is depicted in Table 3. We found significant associations of low T (OR = 1.42, 95% CI, 0.57,3.55, Ptrend = 0.04) with self-reported COPD in the fully adjusted model.
Table 3.
Testosterone association with self-reported COPD in NHANES III-PHASE1, 1999–2004, after adjusting for estradiol and SHBG (2011–2012 not included-subset sample)
| Testosterone Level–OR (95% CI) | |||||||
|---|---|---|---|---|---|---|---|
| Model | High (>400 mg/dL) |
Normal (230–400 mg/dL) |
Low (<230 mg/dL) |
P-trend | |||
| OR | 95% CI | OR | 95% CI | OR | 95% CI | ||
| Model 1 | 1.0 | REF | 1.15 | 0.47, 2.81 | 1.96 | 1.13, 3.42 | 0.08 |
| Model 2 | 1.0 | REF | 2.11 | 1.16, 3.83 | 1.42 | 0.57, 3.55 | 0.04 |
Model 1: adjusted for age. Model 2: adjusted for age, family income, educational level, employment status, race, BMI, waist circumference, self-reported health, smoking status, moderate exercise, vigorous exercise, alcohol consumption, estradiol, and SHBG
Table 4 presents the association of T levels with self-reported COPD stratified by smoking status in the subset analysis. We found that among former/current smokers, there was a significant association of low T (OR = 2.69, 95% CI = 0.87–8.30, Ptrend = 0.01) with self-reported COPD. There was no significant association when we stratified by body fatness (waist circumference [OR = 1.86, 95% CI = 0.51–6.72, Ptrend = 0.2] and BMI [OR = 1.48, 95% CI = 0.22–9.91, Ptrend = 0.66]) in the fully adjusted models (Tables 5 and 6). When stratified by race and ethnicity in the subset analysis, we found significant associations of low T with self-reported COPD (OR = 1.88, 95% CI = 0.68–5.17, Ptrend = 0.008) in the NH-Whites (Table 7), but not in Hispanics and NH-Blacks, which was similar to the associations seen in the full sample (Table 2).
Table 4.
Testosterone level association with self-reported COPD stratified by smoking status, NHANESIII-PHASE 1 1999–2004 after adjusting for estradiol and SHBG
| Smoking status | Testosterone level–OR (95% CI) |
|||||||
|---|---|---|---|---|---|---|---|---|
| High (>400 mg/dL) |
Normal (230–400 mg/dL) |
Low (<230 mg/dL) |
P-trend | P-Interaction | ||||
| OR | 95% CI | OR | 95% CI | OR | 95% CI | 0.2 | ||
| No smoking | ||||||||
| Model 1 | 1.0 | REF | 1.18 | 0.29, 4.91 | 0.43 | 0.08, 2.27 | 0.78 | |
| Model 2 | 1.0 | REF | 0.60 | 0.21, 1.72 | 0.29 | 0.04, 2.04 | 0.16 | |
| Former/current smoke | ||||||||
| Model 1 | 1.0 | REF | 2.17 | 1.15, 4.10] | 1.80 | 0.66, 4.92 | 0.02 | |
| Model 2 | 1.0 | REF | 2.79 | 1.25, 6.21 | 2.69 | 0.87, 8.30 | 0.01 | |
Model 1: Adjusted for age. Model 2: Adjusted for age, family income, educational level, employment status, race, BMI, waist circumference, self-reported health, moderate exercise, vigorous exercise, alcohol consumption, estradiol, and SHBG
Table 5.
Testosterone level association with self-reported COPD stratified by waist circumference, NHANES III-PHASE 1, 1999–2004 after adjusting for estradiol and SHBG
| Testosterone level–OR (95% CI) |
||||||||
|---|---|---|---|---|---|---|---|---|
| High (>400 mg/dL) |
Normal (230–400 mg/dL) |
Low (<230 mg/dL) |
P-trend | P-Interaction | ||||
| OR | 95% CI | OR | 95% CI | OR | 95% CI | 0.51 | ||
| Waist circumference <102 cm | ||||||||
| Model 1 | 1.0 | REF | 1.66 | 0.74, 3.72 | 0.24 | 0.06, 1.06 | 0.68 | |
| Model 2 | 1.0 | REF | 2.12 | 0.84, 5.34 | 0.21 | 0.02, 2.21 | 0.41 | |
| Waist circumference ≥ 102 cm | ||||||||
| Model 1 | 1.0 | REF | 1.83 | 0.93, 3.59 | 1.47 | 0.51, 4.25 | 0.16 | |
| Model 2 | 1.0 | REF | 1.83 | 0.73, 4.60 | 1.86 | 0.51, 6.72 | 0.2 | |
Model 1: Adjusted for age. Model 2: Adjusted for age, family income, educational level, employment status, race, BMI self-reported health, smoking status, moderate exercise, vigorous exercise, alcohol consumption, estradiol, and SHBG
Table 6.
Testosterone level association with self-reported COPD stratified by BMI, NHANESIII-PHASE1, 1999–2004, adjustment for estradiol and SHBG
| BMI | Testosterone level–OR (95% CI) |
|||||||
|---|---|---|---|---|---|---|---|---|
| High (>400 mg/dL) |
Normal (230–400 mg/dL) |
Low (<230 mg/dL) |
P- rend | P-Interaction | ||||
| OR | 95% CI | OR | 95% CI | OR | 95% CI | 0.75 | ||
| BMI <30 | ||||||||
| Model 1 | 1.0 | REF | 2.03 | 1.12, 3.68 | 0.51 | 0.19, 1.38 | 0.11 | |
| Model 2 | 1.0 | REF | 2.46 | 1.45, 4.17 | 0.54 | 0.13, 2.33 | 0.01 | |
| BMI ≥30 | ||||||||
| Model 1 | 1.0 | REF | 1.68 | 0.57, 5.00 | 1.45 | 0.45, 4.66 | 0.34 | |
| Model 2 | 1.0 | REF | 1.28 | 0.27, 6.11 | 1.48 | 0.22, 9.91 | 0.66 | |
Model 1: Adjusted for age. Model 2: Adjusted for age, family income, educational level, employment status, race, waist circumference, self-reported health, smoking status, moderate exercise, vigorous exercise, alcohol consumption, estradiol, and SHBG
Table 7.
Testosterone level association with self-reported COPD stratified by race and ethnicity, NHANES III-PHASE 1, 1999–2004, after adjusting for estradiol and SHBG
| Race | Testosterone level–OR (95% CI) |
|||||||
|---|---|---|---|---|---|---|---|---|
| High (>400 mg/dL) |
Normal (230–400 mg/dL) |
Low (<230 mg/dL) |
P-trend | P-Interaction | ||||
| OR | 95% CI | OR | 95% CI | OR | 95% CI | 0.84 | ||
| Hispanic | ||||||||
| Model 1 | 1.0 | REF | 0.40 | 0.04, 4.00 | 0.09 | 0.01, 1.51 | 0.31 | |
| Model 2 | 1.0 | REF | ND† | ND† | ND† | ND† | ND† | |
| Non-Hispanic White | ||||||||
| Model 1 | 1.0 | REF | 2.22 | 1.20, 4.11 | 1.37 | 0.52, 3.62 | 0.04 | |
| Model 2 | 1.0 | REF | 2.42 | 1.25, 4.70 | 1.88 | 0.68, 5.17 | 0.008 | |
| Non-Hispanic Black | ||||||||
| Model 1 | 1.0 | REF | 1.21 | 0.59, 2.48 | 0.61 | 0.13, 2.84 | 0.84 | |
| Model 2 | 1.0 | REF | 1.33 | 0.52, 3.36 | 0.43 | 0.10, 1.81 | 0.73 | |
Model 1: adjusted for age. Model 2: adjusted for age, family income, educational level, employment status, BMI, waist circumference, self-reported health, smoking status, moderate exercise, vigorous exercise, alcohol consumption, estradiol, and SHBG
Not determine (ND) due to limited sample size
A sensitivity analysis was performed with the outcome being either spirometry or self-reported COPD in the NHANES III (Phase I) and 2011–2012 cycles, for which no significant associations were obtained (Supplemental Table 2). Similar non-significant results were observed with further stratification by smoking status, body fatness, and race and ethnicity (Supplemental Table 3–6). A stratified by age multivariate analysis (Supplemental Table 7) was performed, and there were no significant findings.
Discussion
The key results of this study were the positive association between low T and self-reported prevalent COPD in the overall population among men with positive smoking history, body fatness, and belonging to different racial and ethnic groups. Our findings showed that among current smokers, there was a significant association between low T and a self-reported COPD in the fully adjusted models of the full and subset samples (additionally adjusted for estradiol and SHBG). Similarly, among those with high waist circumference (≥ 102 cm) and BMI (≥ 30), there was a significant association between low T and prevalent self-report COPD in the overall population, although it was attenuated after additionally adjusting for estradiol and SHBG [35]. In general, the association of low T with COPD was significant among NH-White men, but not among NH-Black or Hispanics.
Previous studies have demonstrated higher T levels in Hispanic men and NH-Black adolescent males compared to NH-White men, including the adult population [22, 36]. A previous study reported that men with established COPD had lower levels of T: however, the sample size of the study was only 140 participants [3]. Earlier studies have also noted that NH-Whites have 4.1 times higher odds of developing COPD than Hispanics [37]. Our study thus adds to the body of literature reporting that ethnicity and testosterone levels influence expression of COPD.
A previous US study on Hispanic smokers identified a threefold increase of dyspnea compared to NH-Whites [38]. Given the social determinants of health of Hispanics who tend to have lower education, income status, and access to healthcare/insurance compared to NH-Whites [39], it is possible to suggest that the previously observed higher prevalence rates of COPD among Hispanics may be due, in part, to these social determinants of health.
Our study suggests that the prevalence of COPD may be influenced by serum T levels. Previous studies have shown an inverse association between body fatness and low T [21], while our study has found that low T is highly associated with prevalence of COPD [40]. Therefore, prospective studies are recommended to investigate this interplay between body fatness, low T, and COPD. If low T deficiency were able to exert a direct or indirect impact on the respiratory muscles by positively affecting overall strength and aerobic capacity, then testosterone replacement therapy could help to reduce the progression of COPD, as demonstrated in previous studies [41]. Even though we found an inverse association between low T levels and COPD, our study is limited as concerns data availability on exogenous testosterone supplementation. We can therefore not propose a clinical recommendation regarding the association between low T and COPD. In addition, since we conducted a cross-sectional analysis, this limits our ability to suggest a causal inference between low T and COPD.
Strengths and limitations
The strengths of our study include the fact that NHANES is a program of studies which is nationally representative of the civilian non-institutionalized US population and thus aids in the generalizability of these findings. Additionally, NHANES follows a protocol which is rigorous, which enabled us to utilize extensive quality control procedures for the collection of the exposures, outcome of interest, and potential confounding factors that were analyzed and adjusted in this study. We also mutually adjusted for total T, SHBG, and estradiol.
Despite these strengths, the current study has limitations that may influence interpretation of the results. First, we conducted a cross-sectional study, which precludes the investigation of a temporal investigation between T and COPD and because of which it is not possible to assess the direction of causality between T and COPD. Moreover, due to the cross-sectional nature of NHANES, limited data were available on medication use and genetic predispositions. It is advisable that future studies consider this information. Furthermore, we relied on a single measurement of sex steroid hormones, which limited our ability to make a causal inference between T levels and COPD outcomes, while it is also possible that pharmacotherapy of COPD disease can lower T levels. Similarly, NHANES does not collect data on the treatment of COPD, e.g., bronchodilators or inhaled corticosteroids. However, since no information was available on treatment for low T with testosterone replacement therapy, which can help normalize T levels, this could have influenced our findings. Future studies should take this into consideration. Second, the serum blood collection, collection in tubes, and centrifugation procedures have the potential to influence the reliability, accuracy, and precision of total testosterone, estradiol, and SHBG measurements [42]. Estradiol and SHBG were not available in all the NHANES cycles (2011–2012); therefore, a subset (1988–1991, 1999–2004) was created to adjust for estradiol and SHBG. The use of these two samples will allow comparison with future studies where estradiol and SHBG may or may not be available. Third, the latter NHANES waves (2011–2012) measured total T with HPLC tandem mass spectrophotometry, which has been recommended by some investigators, while the previous ones (1988–1991, 1999–2004) used electrochemiluminescence immunoassays. Fourth, the cut-off point to define low T may not fully capture its effect on the development of COPD, and there is also no clinical cut-off point for low T in women, while the sample size of serum testosterone among women is limited and not well-characterized. Future studies should thus pursue this investigation among women. Fifth, although we adjusted for several COPD risk factors, there remains the possibility of residual confounding by unmeasured factors and/or of lack of medication information.
Conclusion
In summary, the results of this study showed that low levels of T were associated with a higher prevalence of COPD, including among men with smoking history, body fatness, and from different racial and ethnic backgrounds. Prospective studies that enable assessment of incident COPD and studies with larger underserved and understudied populations and including such lifestyle factors as smoking and body fatness are required to confirm these findings.
Supplementary Material
Acknowledgements
The authors thanked Alexander Rader.
Funding
David S. Lopez was supported by the National Institutes of Health (NIH) and National Institute on Aging, Grant no.: P30 AG059301, and the Cancer Prevention and Research Institute of Texas (CPRIT), Grant no.: RP210130.
Footnotes
Supplementary Information The online version contains supplementary material available at https://doi.org/10.1007/s42000-023-00506-x.
Ethics approval The protocols for the conduct of the NHANES were approved by the Institutional Review Board.
Informed consent We conducted a secondary data analysis using data from NHANES. NHANES obtained informed consents from participants that also included consent for publication.
Conflict of interest The authors declare no competing interests.
Submission declaration and verification This work has not been published previously, and it is not under consideration for publication elsewhere. The publication is approved by all authors and, if accepted, it will not be published elsewhere in the same form, in English, or in any other language.
Data availability
All data generated or analyzed in this study were provided by the National Center for Health Statistics (NCHS) of the US Centers for Disease Control and Prevention (CDC) — https://www.cdc.gov/nchs/nhanes/index.htm. For further data inquiries, please contact the corresponding author of this manuscript (David S. Lopez).
References
- 1.Mannino DM (2022) The changing definition and perception of COPD. Respir Care 67(6):750–755 [DOI] [PubMed] [Google Scholar]
- 2.Xu J, Murphy SL, Kochanek KD, Arias E (2022) Mortality in the United States. NCHS Data Br 456:1–8 [PubMed] [Google Scholar]
- 3.Mousavi SA, Kouchari MR, Samdani-Fard SH, Gilvaee ZN, Arabi MJT (2012) Relationship between serum levels of testosterone and the severity of chronic obstructive pulmonary disease. Tanaffos 11(3):32–35 [PMC free article] [PubMed] [Google Scholar]
- 4.Karakou E, Glynos C, Samara KD, Msaouel P, Koutsilieris M, Vassilakopoulos T (2013) Profile of endocrinological derangements affecting PSA values in patients with COPD. In Vivo 27(5):641–649 [PubMed] [Google Scholar]
- 5.Scott TW (2023) Chronic obstructive pulmonary disease: risk factors and risk reduction. UpToDate. https://www.uptodate.com/contents/chronic-obstructive-pulmonary-disease-risk-factors-and-risk-reduction
- 6.Camiciottoli G, Bigazzi F, Magni C et al. (2016) Prevalence of comorbidities according to predominant phenotype and severity of chronic obstructive pulmonary disease. Int J Chron Obstruct Pulmon Dis 11:2229–2236 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Balasubramanian V, Naing S (2012) Hypogonadism in chronic obstructive pulmonary disease: incidence and effects. Curr Opin Pulm Med 18(2):112–117. 10.1097/MCP.0b013e32834feb37 [DOI] [PubMed] [Google Scholar]
- 8.Slim A, Hedhli A, Ouahchi Y et al. (2020) Testosterone and chronic obstructive pulmonary disease. Rev Mal Respir 37(10):790–799. Testosteronemie et bronchopneumopathie chronique obstructive. 10.1016/j.rmr.2020.08.013 [DOI] [PubMed] [Google Scholar]
- 9.Rubinsztajn R, Przybylowski T, Maskey-Warzechowska M, Karwat K, Chazan R (2019) Serum testosterone depression as a factor influencing the general condition in chronic obstructive pulmonary disease patients. Adv Clin Exp Med 28(6):783–788. 10.17219/acem/94153 [DOI] [PubMed] [Google Scholar]
- 10.Khera M, Adaikan G, Buvat J et al. (2016) Diagnosis and treatment of testosterone deficiency: recommendations from the Fourth International Consultation for Sexual Medicine (ICSM 2015). J Sex Med 13(12):1787–1804 [DOI] [PubMed] [Google Scholar]
- 11.Buvat J, Maggi M, Guay A, Torres LO (2013) Testosterone deficiency in men: systematic review and standard operating procedures for diagnosis and treatment. J Sex Med 10(1):245–284 [DOI] [PubMed] [Google Scholar]
- 12.Mulhall JP, Trost LW, Brannigan RE et al. (2018) Evaluation and management of testosterone deficiency: AUA guideline 200(2):423–432. 10.1016/j.juro.2018.03.115 [DOI] [PubMed] [Google Scholar]
- 13.Lopez DS, Canfield S, Wang R (2016) Testosterone replacement therapy and the heart: friend, foe or bystander? Transl Androl Urol 5(6):898–908 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Lopez DS, Advani S, Tsilidis KK, Wang R, Canfield S (2017) Endogenous and exogenous testosterone and prostate cancer: decreased-, increased-or null-risk? Trans Androl Urol 6(3):566–579 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Lopez D, Qiu X, Advani S et al. (2018) Double trouble: Co-occurrence of testosterone deficiency and body fatness associated with all-cause mortality in US men. Clin Endocrinol 88(1):58–65 [DOI] [PubMed] [Google Scholar]
- 16.Morgentaler A 3rd, Conners WP (2015) Testosterone therapy in men with prostate cancer: literature review, clinical experience, and recommendations. Asian J Androl 17(2):206–211 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Platz EA (2008) Low testosterone and risk of premature death in older men: analytical and preanalytical issues in measuring circulating testosterone. Clin Chem 54(7):1110–1112 [DOI] [PubMed] [Google Scholar]
- 18.Araujo AB, Esche GR, Kupelian V et al. (2007) Prevalence of symptomatic androgen deficiency in men. J Clin Endocrinol Metab 92(11):4241–4247 [DOI] [PubMed] [Google Scholar]
- 19.Travison TG, Araujo AB, O’Donnell AB, Kupelian V, McKinlay JB (2007) A population-level decline in serum testosterone levels in American men. J Clin Endocrinol Metab 92(1):196–202 [DOI] [PubMed] [Google Scholar]
- 20.Malik RD, Lapin B, Wang CE, Lakeman JC, Helfand BT (2015) Are we testing appropriately for low testosterone?: Characterization of tested men and compliance with current guidelines. J Sex Med 12(1):66–75 [DOI] [PubMed] [Google Scholar]
- 21.Rohrmann S, Shiels MS, Lopez DS et al. (2011) Body fatness and sex steroid hormone concentrations in US men: results from NHANES III. Cancer Causes Control 22(8):1141–1151 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Lopez DS, Peskoe SB, Joshu CE et al. (2013) Racial/ethnic differences in serum sex steroid hormone concentrations in US adolescent males. Cancer Causes Control 24(4):817–826. 10.1007/s10552-013-0154-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Rohrmann S, Nelson WG, Rifai N et al. (2007) Serum estrogen, but not testosterone, levels differ between black and white men in a nationally representative sample of Americans. J Clin Endocrinol Metab 92(7):2519–2525 [DOI] [PubMed] [Google Scholar]
- 24.Erickson SE, Iribarren C, Tolstykh IV, Blanc PD, Eisner MD (2007) Effect of race on asthma management and outcomes in a large, integrated managed care organization. Arch Intern Med 167(17):1846–1852 [DOI] [PubMed] [Google Scholar]
- 25.Lederer DJ, Benn EK, Barr RG et al. (2008) Racial differences in waiting list outcomes in chronic obstructive pulmonary disease. Am J Respir Crit Care Med 177(4):450–454 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Omachi TA, Iribarren C, Sarkar U et al. (2008) Risk factors for death in adults with severe asthma. Ann Allergy Asthma Immunol 101(2):130–136 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Centers for Disease Control and Prevention (CDC) National Center for Health Statistics. National Health and Nutrition Examination Survey: Overview. https://www.cdc.gov/nchs/nhanes/index.htm. Accessed Jun 2022
- 28.National Center for Health Statistics (1994) Plan and operation of the third national health and nutrition examination survey, 1988–94. series 1: programs and collection procedures. Vital Health Stat 1:1–407 [PubMed] [Google Scholar]
- 29.Schnell K, Weiss CO, Lee T et al. (2012) The prevalence of clinically-relevant comorbid conditions in patients with physician-diagnosed COPD: a cross-sectional study using data from NHANES 1999-2008. BMC Pulm Med 12:26. 10.1186/1471-2466-12-26 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Miller MR, Hankinson J, Brusasco V et al. (2005) Standardisation of spirometry. Eur Respir J 26(2):319–338 [DOI] [PubMed] [Google Scholar]
- 31.Centers for Disease Control and Prevention (CDC) (2011) National Center for Health Statistics (NCHS). National Health and Nutrition Examination Survey (NHANES): Respiratory Health Spirometry Procedures Manual. U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, Hyattsville, MD [Google Scholar]
- 32.Mohamed Hoesein FAA, Zanen P, Lammers JWJ (2001) Lower limit of normal or FEV1/FVC <0.70 in diagnosing COPD: An evidence-based review. Respir Med 105(6):907–915. 10.1016/j.rmed.2011.01.008 [DOI] [PubMed] [Google Scholar]
- 33.Hukkanen J, Jacob P, Benowitz NL (2005) Metabolism and disposition kinetics of nicotine. Pharmacol Rev 57(1):79–115 [DOI] [PubMed] [Google Scholar]
- 34.Menke A, Guallar E, Rohrmann S et al. (2010) Sex steroid hormone concentrations and risk of death in US men. Am J Epidemiol 171(5):583–592 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Wu TD, Ejike CO, Wise RA, McCormack MC, Brigham EP (2019) Investigation of the obesity paradox in chronic obstructive pulmonary disease, according to smoking status, in the United States. Am J Epidemiol 188(11):1977–1983. 10.1093/aje/kwz185 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Mazur A (2009) The age-testosterone relationship in black, white, and Mexican-American men, and reasons for ethnic differences. Aging Male 12(2–3):66–76. 10.1080/13685530903071802 [DOI] [PubMed] [Google Scholar]
- 37.Aiyer A, Surani S, Aguillar R, Sharma M, Ali M, Varon J (2020) Ethnic variance in prevalence of COPD among smokers in a real world setting. Open Respir Med J 14:93–98. 10.2174/1874306402014010093 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Hardie G, Liu R, Darden J, Gold WM (2010) Ethnic differences in methacholine responsiveness and word descriptors in African Americans, Hispanic-Mexican Americans, Asian-Pacific Islanders, and Whites with mild asthma. J Asthma 47(4):388–396. 10.3109/02770903.2010.481341 [DOI] [PubMed] [Google Scholar]
- 39.Boen CE, Hummer RA (2019) Longer-but Harder-Lives?: the Hispanic health paradox and the social determinants of racial, ethnic, and immigrant-Native health disparities from midlife through late life. J Health Soc Behav 60(4):434–452. 10.1177/0022146519884538 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Pelusi C, Pasquali R (2012) The significance of low testosterone levels in obese men. Curr Obes Rep 1(4):181–190. 10.1007/s13679-012-0029-4 [DOI] [Google Scholar]
- 41.Baillargeon J, Urban RJ, Zhang W et al. (2019) Testosterone replacement therapy and hospitalization rates in men with COPD. Chron Respir Dis 16:1479972318793004. 10.1177/1479972318793004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Paduch DA, Brannigan RE, Fuchs EF, Kim ED, Marmar JL, Sandlow JI (2014) The laboratory diagnosis of testosterone deficiency. Urology 83(5):980–988. 10.1016/j.urology.2013.12.024 [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
All data generated or analyzed in this study were provided by the National Center for Health Statistics (NCHS) of the US Centers for Disease Control and Prevention (CDC) — https://www.cdc.gov/nchs/nhanes/index.htm. For further data inquiries, please contact the corresponding author of this manuscript (David S. Lopez).