High-sensitivity C-reactive protein levels and treadmill exercise test responses in men and women without overt heart disease

Abstract

BACKGROUND:

C-reactive protein (CRP) is a marker of systemic inflammatory activity and may be modulated by physical fitness. Treadmill exercise testing is used to evaluate cardiovascular health through different variables including exercise capacity, heart rate and blood pressure responses. It was hypothesized that CRP levels are associated with these variables in men and women without overt heart disease.

METHODS:

A total of 584 asymptomatic subjects (317 [54.3%] women and 267 [45.7%] men) were enrolled in the present study and underwent clinical evaluation. CRP levels in men and women were examined relative to clinical characteristics and to variables of treadmill exercise testing: peak heart rate, exercise systolic blood pressure, exercise time, chronotropic reserve and heart rate recovery at the first and second minutes after exercise. Multivariate analysis was performed using a log-linear regression model.

RESULTS:

In women, exercise time on the treadmill exercise test (P=0.009) and high-density lipoprotein cholesterol levels (P=0.002) were inversely associated with CRP levels. Body mass index (P<0.001) and total cholesterol levels (P=0.005) were positively associated with CRP levels. In men, exercise time on the treadmill exercise test was inversely associated with CRP levels (P=0.015). Body mass index (P=0.001) and leukocyte count (P=0.002) were positively associated with CRP levels. CRP levels were not associated with peak heart rate, chronotropic reserve, heart rate recovery at the first and second minutes, or exercise systolic blood pressure.

CONCLUSIONS:

These findings contribute to the evidence that CRP is lower in individuals with better exercise capacity and demonstrate that this relationship is also apparent in individuals without overt heart disease undergoing cardiovascular evaluation through the treadmill exercise test. Lowering inflammatory markers may be an additional reason to stimulate sedentary individuals with low exercise capacity in the treadmill exercise test to improve physical conditioning through regular exercise.

C-reactive protein (CRP) levels are influenced by sex (1–3), smoking (1,4,5), obesity (3,6), alcohol intake (7) and regular exercise (8). CRP levels measured using the high-sensitivity assay may be useful in the assessment of cardiovascular health (9–16).

Treadmill exercise testing is routinely performed in clinical practice to evaluate cardiovascular health. Exercise capacity, heart rate and blood pressure during and after exercise are usually evaluated and are associated with clinical outcomes in individuals without overt heart disease (17–22). These variables have been demonstrated to be different in men and women (23–25).

Physical fitness has been reported to be associated with CRP levels, independent of other traditional cardiovascular risk factors (25–29). These findings suggest that the beneficial effects of exercise on cardiovascular health may be mediated partially by a decrease in inflammatory status in individuals with or without cardiovascular disease. However, the relationship between blood pressure and heart rate responses during and after exercise, together with physical fitness and inflammatory markers, is not well defined.

CRP has been studied in different populations (15). Individuals without overt heart disease after clinical examination undergoing treadmill exercise testing as part of a clinical protocol are a specific subset of patients in clinical practice.

To test the hypothesis that CRP, as a marker of chronic inflammation, is influenced by cardiovascular responses during exercise, we studied the association between CRP levels and physiological responses during the treadmill exercise test including exercise capacity, heart rate and blood pressure, in a sample of men and women without overt heart disease.

METHODS

Study sample

A total of 584 asymptomatic subjects living in the metropolitan area of São Paulo City, Brazil, were enrolled in the present study. Participants volunteered for clinical evaluation in an outpatient clinic and underwent clinical and laboratory evaluation. The laboratory work-up included electrocardiography, chest x-rays, echocardiography, blood cell count and measurements of serum glucose, total cholesterol, high-density lipoprotein (HDL) cholesterol, low-density lipoprotein (LDL) cholesterol, triglyceride, serum creatinine, thyroid stimulating hormone and high-sensitivity CRP levels. Patients without any evidence of overt heart disease were enrolled in the study. A total of 317 (54.3%) women and 267 (45.7%) men were selected for the present cross-sectional study. The mean (± SD) age of the participants was 43.9±13.2 years; 115 (19.7%) were current smokers, 474 (81.2%) had a low Framingham risk score and 110 (18.8%) had an intermediate Framingham risk score.

Exclusion criteria

Participants with evidence of heart disease based on initial clinical evaluation or arterial blood pressure ≥160/100 mmHg were excluded. Patients with a medical history of diabetes mellitus, cerebrovascular disease, cancer, lung disease, thyroid disease or other significant systemic diseases were also excluded, as were patients with an inability to perform treadmill exercise testing, an ischemic electrocardiographic response and/or complex ventricular arrhythmia during exercise testing.

Treadmill exercise test

The participants underwent a symptom-limited treadmill exercise test according to the Ellestad protocol (30). The criteria for interruption of the exercise were physical exhaustion or exceeded the predicted maximum heart rate for the patient’s age. Individuals were encouraged to exercise until they experienced limiting symptoms, even if 85% of maximum predicted heart rate was achieved. Peak exercise capacity was estimated from the exercise time.

During each exercise stage and recovery stage, symptoms, blood pressure and heart rate were recorded. Predicted peak heart rate was calculated as the patient’s age (in years) subtracted from 220. Peak heart rate achieved, and maximum systolic and diastolic blood pressure achieved were recorded at the end of the exercise stage. Chronotropic reserve was estimated according to the following formula:

$$\begin{array}{c}\left(\text{peak}\hspace{0.17em}\text{heart}\hspace{0.17em}\text{rate}\hspace{0.17em}\text{achieved}-\text{baseline}\hspace{0.17em}\text{heart}\hspace{0.17em}\text{rate}\right)/\\ \left(\text{predicted}\hspace{0.17em}\text{peak}\hspace{0.17em}\text{heart}\hspace{0.17em}\text{rate}-\text{baseline}\hspace{0.17em}\text{heart}\hspace{0.17em}\text{rate}\right)\end{array}$$

The recovery stage followed peak exercise, in which individuals walked for a 3 min cool-down period at 1.5 mph (2.4 km/h) without inclination. Heart rate recovery was defined by the difference between heart rate measured at the peak exercise and during the first and second minutes of cool down period, with the patients walking at a speed of 1.5 mph (2.4 km/h) without inclination, as previously described (31).

The ST segment was measured 0.08 s after the J point in three consecutive QRS complexes with a flat baseline and R wave of equal amplitude.

High-sensitivity CRP assay

Peripheral venous blood samples were drawn into test tubes containing EDTA. Plasma samples were separated and stored at −80°C. CRP was analyzed using a high-sensitivity analyzer (Behring Nephelometer Analyzer II, Dade Behring Incorporated, USA).

Variables studied

CRP levels were examined relative to: demographic and clinical data (age, sex, smoking status, body mass index [BMI]); laboratory variables (serum glucose, total cholesterol, HDL cholesterol, LDL cholesterol, triglycerides, uric acid and leukocyte count); and exercise testing variables (baseline heart rate, peak heart rate, chronotropic reserve, exercise time, baseline systolic blood pressure and peak exercise systolic blood pressure and heart rate recovery at the first and second minutes of recovery).

Statistical analysis

All data were expressed as mean ± SD. Correlation coefficients were computed to assess the correlation among the studied variables. Because the distribution of CRP levels is asymmetric, the natural logarithm of CRP transform was used as the dependent variable. A stepwise log-linear regression model of CRP levels was used for the multivariate analysis. Different models were constructed for men and women due to differences in CRP levels (1) and in exercise performance between sexes in previous studies (28). A final model with independent variables (P<0.05) was obtained for each sex. Statistical calculations were performed using SAS (SAS Institute, USA) and R software.

Ethics

The protocol was approved by the Ethics Committee Human Research of the Heart Institute (InCor), University of Sao Paolo Medical School, Sao Paolo, Brazil, and all participants signed an informed consent form.

RESULTS

The baseline characteristics and exercise testing responses of the study subjects are presented in Table 1. The mean CRP level was higher in women than in men. The mean BMI and age were not significantly different between men and women. The maximum predicted heart rate during exercise testing was reached in 136 (42.9%) women and 141 (52.8%) men. Exercise capacity, heart rate and blood pressure during exercise testing were higher in men than in women. Heart rate recovery after exercise was not significantly different according to sex. Peak heart rate during the treadmill exercise test was not significantly different between participants with an exercise time lower or higher than 270 s. The distributions of BMI, age and peak heart rate of the study population are presented in Table 2. The distribution (scatter plot) of CRP in men and women relative to exercise time is presented in Figure 1.

TABLE 1.

Characteristics of the study participants

Characteristic	Men (n=267)	Women (n=317)
Age, years	43.3±12.9	44.4±13.4
Body mass index, kg/m2	26.6±4	26.4±4.6
Serum glucose, mmol/L	5.3±0.7	5.1±0.5
Triglycerides, mmol/L	7.9±5.5	5.7±3.1
Total cholesterol, mmol/L	10.7±2.1	10.9±2.2
HDL cholesterol, mmol/L	2.5±0.6	2.9±0.8
LDL cholesterol, mmol/L	6.7±1.8	6.8±1.9
C-reactive protein, mg/L	2.1±2.5	3±3.5
Exercise testing responses
Exercise time, s	492.1±104.1	402.1±101.2
Peak heart rate, beat/min	166.6±15.2	161.8±17.3
Chronotropic reserve	0.90±0.13	0.85±0.15
Heart rate recovery, first minute, beats/min	40.8±16.9	41.21±14.2
Heart rate recovery, second minute, beats/min	55.7±15	57.1±13.9
Baseline systolic blood pressure, mmHg	129.2±13.9	123.6±14.4
Baseline diastolic blood pressure, mmHg	84.6±9.9	80.7±8.6
Exercise systolic blood pressure, mmHg	181.1±20.4	161.9±20.3

Data presented as mean ± SD. HDL High-density lipoprotein; LDL Low-density lipoprotein

TABLE 2.

Distribution of body mass index, age and peak heart rate during exercise in the study participants

	Women	Men
Body mass index, kg/m2
Lower limit	17.1	17.9
First quartile	23.0	23.7
Median	25.7	26.1
Third quartile	29.1	29.1
Upper limit	45.8	42.1
Age, years
Lower limit	18	18
First quartile	34	34
Median	46	43
Third quartile	53	53
Upper limit	79	77
Peak heart rate, beats/min
Lower limit	103	114
First quartile	153	158
Median	163	169
Third quartile	172	177
Upper limit	201	200

Figure 1).

Distribution of C-reactive protein levels according to exercise time in men and women

In women, exercise capacity and HDL cholesterol levels were inversely associated with CRP levels (Table 3). CRP decreased by a factor of 0.99 when exercise time increased by 1 s, ie, lowering exercise time in 1 s increased expected CRP by 0.17% (all other factors held constant). A decrease of 1 mg/dL (0.0259 mmol/L) in HDL cholesterol level increased the expected level of CRP by 14.6%. BMI and serum total cholesterol levels were positively associated with CRP (Table 3). CRP levels increased by a factor of 1.086 (8.6%) when BMI increased by 1 kg/m² (all other factors held constant). An increase in total cholesterol of 1 mg/dL (0.0259 mmol/L) caused CRP levels to increase by 0.47%.

TABLE 3.

Variables associated with the logarithm of C-reactive protein levels in women (stepwise log-linear regression model)

Variable	Estimate	Exp (estimate)	SE	P
Body mass index	0.0822	1.0860	0.0139	<0.001
Exercise time	−0.0017	0.9983	0.0006	0.009
HDL cholesterol	−0.0145	0.9856	0.0045	0.002
Total cholesterol	0.0047	1.0047	0.0017	0.005

HDL High-density lipoprotein

In men, exercise capacity was inversely associated with CRP levels (Table 4). CRP decreased by a factor of 0.99 when exercise time increased 1 s (ie, decreasing exercise time in 1 s increased expected CRP levels by 0.17% [all other factors held constant]). BMI and leukocyte count were positively associated with CRP levels (Table 4). An increase in BMI of 1 kg/m² increased CRP levels by 6.48%. An increase of 1×10³ leukocytes/mm³ increased CRP levels by 4.58%.

TABLE 4.

Variables associated with the logarithm of C-reactive protein levels in men (stepwise log-linear regression model)

Variable	Estimate	Exp (estimate)	SE	P
Body mass index	0.0628	1.0648	0.0180	0.001
Exercise time	−0.0017	0.9983	0.0007	0.015
Leukocyte count	0.0458	1.0469	0.0313	0.002

Baseline heart rate, maximal heart rate, chronotropic reserve, heart rate recovery at the first and second minutes, and exercise systolic blood pressure were not significantly associated with CRP levels.

DISCUSSION

The main finding of the present study was that CRP levels were inversely associated with exercise capacity even after adjustment for blood pressure and heart rate variables during treadmill exercise testing in individuals undergoing clinical evaluation through the treadmill exercise test.

The mean CRP level was higher in women, consistent with previous investigations (1,2). Therefore, we performed additional analyses on men and women separately.

Our findings in individuals undergoing clinical evaluation with treadmill exercise testing support previous studies that demonstrated an association between CRP levels and exercise capacity. Lower CRP levels were reported in Caucasian and Native American women with higher physical fitness (32). In a study involving 892 middle-age subjects, CRP levels were inversely correlated with exercise capacity (18) as well as in 1660 subjects with metabolic syndrome (19). Exercise contributes to cardiovascular health by influencing several pathways including endothelial function, insulin resistance, platelet aggregation, among others (33). Chronic inflammation may be modulated by metabolic variables, such as BMI (34,35) and lipid profile (36), which may also be influenced by regular exercise (37). In the present study sample, CRP was associated with exercise capacity in both sexes independently of metabolic variables, suggesting that other factors may be relevant to this association. Exercise can reduce the release of cytokines such as IL-6 and TNF-α by adipocyte tissue, downregulating the sympathetic tonus (38). Exercise training and physical fitness have been associated with endothelial function by increased release of endothelium-derived nitric oxide (39). The increased production of nitric oxide by endothelial cells may reduce the effects of proinflammatory cytokines and decrease CRP levels. Our results are consistent with a beneficial influence of improved exercise capacity in relation to cardiovascular health by reducing the levels of chronic inflammatory markers.

Heart rate recovery as a measure of vagal tonus may be associated with inflammatory status. However, in contrast with the results of two previous studies (40,41), we did not detect an association between CRP levels and heart rate recovery. The absence of a significant association between CRP levels and heart rate recovery reported in our study sample may be a consequence of the inclusion of other variables of the treadmill exercise test in the statistical modelling. In addition, in our study, the influence of the exercise capacity on CRP levels was shown to be stronger than other exercise testing variables. This finding suggests that other pathways beyond the autonomic nervous system may be important to the association between exercise capacity and inflammation in our study sample.

Regarding exercise blood pressure, our study is consistent with a previous study that evaluated the relationship between maximum exercise systolic blood pressure and CRP levels in 43 men with elevated blood pressure during exercise and 42 men with normal blood pressure responses (42). In this study, exercise blood pressure was correlated with leukocyte count, but not with CRP levels. Our finding suggests that blood pressure response during exercise may not be a predictor of inflammatory status in apparently healthy individuals.

Previous reports have demonstrated an association of leukocyte count and serum cholesterol with outcome and markers of inflammatory activity in population studies (43,44) or in specific clinical settings (45). Investigating CRP levels prompted us to include leukocyte count and total cholesterol levels in the clinical variables, to test this previous finding in a study sample of individuals without overt heart disease. The exploratory characteristic of the study includes the possibility of detecting associations that may require further testing in future studies.

We recognize several limitations to our study. Its cross-sectional design limits inferences about causality between CRP levels and exercise testing performance. The stratification of the models according to sex decreased the number of participants in each group, which may have limited the power of the study to detect possible associations between CRP levels and other exercise testing variables beyond exercise capacity. However, the strength of the present study was the evaluation of the association between CRP levels and exercise testing variables commonly used in clinical practice. Exercise test variables, such as exercise capacity, heart rate and blood pressure during and after exercise, have been associated with cardiovascular health. These variables may interact with one another. The inclusion of these variables in the statistic modelling may be useful to investigate which variable would demonstrate a higher association with CRP levels.

CONCLUSION

Our findings add to the concept that CRP levels are lower in individuals with higher exercise capacity and this relationship may also be apparent in individuals without overt heart disease who undergo cardiovascular evaluation with the treadmill exercise test. Lowering levels of inflammatory markers may be an additional incentive to stimulate sedentary individuals with low exercise capacity to improve physical conditioning through regular exercise.