Learn more: PMC Disclaimer | PMC Copyright Notice
Association of high-sensitivity CRP and FEV1%pred: a study on non-pulmonary disease in a population in Beijing, China
3Abstract
Background
No studies have investigated whether high-sensitivity C reactive protein (hsCRP) can be used to predict the forced expiratory volume in 1 s (FEV1)/estimated value of FEV1 (FEV1%pred). This study aimed to assess the association between hsCRP and FEV1%pred in middle-aged and elderly individuals without underlying lung disease.
Methods
The data for this study were obtained from a prospective cohort study that included 1047 middle-aged and elderly citizens from Beijing aged 40–75 years without any evidence of underlying lung diseases with FEV1 >70% after receiving inhalational bronchodilators. The baseline analysis of the participants was performed from 30 May 2018 to 31 October 2018. Restricted cubic spline regression and multivariate linear regression models were used to assess the non-linear association and linear association between hsCRP and FEV1/FEV in 6 s (FEV6) and FEV1%pred, respectively.
Results
The hsCRP values of 851 participants were recorded; the values were normal in 713 (83.8%) participants. The remaining 196 participants (18.7%) had missing data. A non-linear association was observed between normal hsCRP values and FEV1/FEV6. hsCRP was linearly and negatively correlated with FEV1%pred, and each 1 SD increase in hsCRP was significantly associated with a 2.4% lower in FEV1%pred. Significantly higher FEV1/FEV6 differences were observed in the female subgroup than those in the male subgroup (p=0.011 for interaction).
Conclusions
hsCRP had a non-linear association with FEV1/FEV6 and a linear negative association with FEV1%pred in individuals with normal hsCRP values. hsCRP can be used to predict FEV1%pred, which can be used to predict the development of chronic obstructive pulmonary disease. hsCRP has a stronger association with lung function in women than that in men.
Trial registration number
Introduction
Lung function is influenced by various environmental and pathophysiological factors. Macrophages migrate to the alveoli when the lungs are exposed to harmful agents and produce many inflammatory mediators and cytokines.1 2 However, the secondary amplification of inflammation leads to lung injury and contributes to a decline in lung function. C reactive protein (CRP), a classical acute response protein, is a reliable and sensitive marker of acute inflammation. Previous studies have reported that CRP levels are correlated with lung function.3 Aronson et al reported a negative correlation between lung function and CRP levels.4 Several studies have revealed an association between CRP levels and the development of chronic obstructive pulmonary disease (COPD) at the genetic level.5 6 CRP level has also been recognised as a predictive factor for prognosis.7 However, it is unknown whether CRP levels can be used to predict the forced expiratory volume in 1 s (FEV1)/estimated value of FEV1 (FEV1%pred). Therefore, this study aimed to assess whether CRP can be used as a predictor of FEV1%pred in middle-aged and elderly individuals from Beijing with no underlying lung disease.
Methods
Study design and population
Community residents of Beijing were recruited via incidental sampling performed by the Community Health Service Centre in accordance with a previously published protocol.8 A cross-sectional questionnaire was completed by the participants, and lung function tests were performed. The laboratory data were collected by the medical staff of the Department of Respiratory Medicine at the Peking University First Hospital and the local Community Health Centre. Informed consent was obtained from the participants before enrolment.
Between 30 May 2018 and 31 October 2018, 1164 community residents aged 40–75 years were enrolled in the study. Among these 1164 individuals, 117 individuals who did not meet the inclusion criteria were excluded. Thus, the baseline data from 1047 participants were analysed (figure 1). To investigate the association between high-sensitivity CRP (hsCRP) and FEV1%pred and the FEV1/FEV in 6 s (FEV1/FEV6), we analysed different data sets, which included 851 individuals with hsCRP data, 546 individuals with normal hsCRP and FEV1/FEV6 data, and 529 individuals with normal hsCRP and FEV1%pred data. The results of the other subgroup and the interaction analyses processes are shown in figure 2.
Data collection
Assessment of respiratory symptoms
The Respiratory Symptom Assessment Questionnaire used in this study was the St. George’s Respiratory Questionnaire.
Physical examination
As part of the physical examination, the height (cm), weight (kg), HR (beats/min), systolic blood pressure (mm hg), neck circumference (cm), waist circumference (cm) and hip circumference (cm) of the participants were recorded by the medical staff of the Peking University First Hospital.
Lung function tests
Spirometry was performed using COPD-6, a portable spirometer from Ireland. The participants were instructed to hold their breath at the end of a deep inhalation while pinching their nose and rapidly exhale with the disposable nozzle fully in the mouth at maximum exhalation force until two beeps were heard, which indicated the completion of the 6 s breath test. The screen displaying ‘!’ at the end of the breath test indicated a failed test, and the test was performed again in such cases. After the completion of three passing tests, the project investigator pressed ‘Enter’ to upload the data to the computer. Quality control of the lung function test was performed by professionals familiar with lung function testing. Cases with only one lung function manoeuvre or no good blow did not qualify during the lung function quality control. Only qualified data were analysed. FEV1/FVC was replaced by FEVI/FEV6 in this study.
Lung function calculation formula
Males: FEV1%pred = (FEV1/{0.0430 × (100 × height) – 0.029 × age – 2.49}) × 100%.
Females: FEV1%pred = (FEV1/{0.0395 × (100 × height) – 0.025 × age – 2.60}) × 100%.
The FEV1%pred formula used in this study met the age and nationality requirements of the population enrolled in this study.9 10
Laboratory examinations
hsCRP is clinically known as hsCRP. In this study, hsCRP≤3 mg/L was considered normal.
The peripheral blood tests included routine blood tests as well as an assessment of the interleukin-6 (IL-6) and hsCRP levels. Routine urine tests were also performed. Venous blood and urine samples were collected from each participant and sent to the Laboratory Department of the Peking University First Hospital for testing.
Statistical analysis
The baseline characteristics are expressed as mean (SD) or median (Q1, Q3), and categorical variables are described as numbers (percentages). The participants were divided into two groups according to the normal range of hsCRP: hsCRP ≤3 mg/L was considered normal, and hsCRP >3 mg/L was considered abnormal. The statistical significance of the differences was assessed using the t-test for normally distributed continuous variables, the rank sum test for non-normally distributed continuous variables and χ² test for categorical variables.
In the hsCRP ≤3 mg/L dataset, hsCRP, FEV1/FEV6 and FEV1%pred were grouped according to quartiles and descriptive statistical analysis, univariate ordered logistic regression analysis, least absolute shrinkage and selection operator (LASSO) regression were used to determine the final effect of the optimal model variables, multifactorial ordered logistic regression was used to investigate the effects of multiple variables such as hsCRP, sex and age on FEV1/FEV6 or FEV1%pred, in addition to further subgroup and trend analyses in the two factors of presence of haematuria and sex. In this case, trend analysis was performed by dividing the variables into four groups based on the quartiles of hsCRP and coding them as 1, 2, 3 and 4, which were included in a multifactorial ordered logistic regression model, and p values and OR values were calculated.
In addition, non-linear relationships between hsCRP and FEV1/FEV6 or FEV1%pred were analysed in the hsCRPhsCRP ≤3 mg/L dataset using restricted cubic spline regression models. The linear relationship between the two variables was tested using linear regression and the interaction of sex and haematuria status with FEV1/FEV6 and FEV1%pred was further assessed using multivariate regression models with the addition of an interaction term test. All analyses were performed by using R (V.4.1.2) software. All analyses were performed by using R (V.4.1.2) software.
Results
Population characteristics
This study included 1047 subjects. Among these participants, 34.6% were males, and 65.4% were females, with a median age of 62 years (Q1, Q3: 57, 62). Among these 1047 participants, 196 had no hsCRP data available. Table 1 presents the demographic and clinical characteristics of the population and the hsCRP data of the groups with normal and elevated hsCRP levels. An hsCRP level ≤3 mg/L was defined as the normal value. Among the participants with the hsCRP data available, 718 had normal hsCRP levels with a median of 0.99 mg/L (Q1, Q3: 0.56, 1.61), and 138 had abnormal hsCRP levels with a median of 5.22 mg/L (Q1, Q3: 3.74, 7.87). The median age was 62 years (Q1, Q3: 57, 66) and 284 (33.5%) participants were male, among whom 242 (34.1%) had normal hsCRP, and 42 (30.4%) had abnormal hsCRP. In addition, 397 (46.7%) participants had cigarette exposure, among whom 328 (46.0%) had normal hsCRP, and 69 (50.0%) had abnormal hsCRP.
Table 1
Patient characteristics at baseline
| All participants | hsCRP ≤3 mg/L | hsCRP >3 mg/L | P value | |
| Characteristics | 851 | 713 | 138 | |
| Age, year | 62 (57, 66) | 62 (57, 66) | 62 (57, 66) | 0.99 |
| Sex | ||||
| Male | 284 (33.5) | 242 (34.1) | 42 (30.4) | |
| Female | 564 (66.5) | 467 (65.9) | 96 (69.6) | 0.46 |
| Comorbidities | 573 (67.3) | 468 (65.6) | 104 (75.4) | 0.033 |
| Obesity | 17 (2.0) | 9 (1.3) | 8 (5.8) | 0.002 |
| OSAHS | 26 (3.1) | 17 (2.4) | 9 (6.5) | 0.021 |
| Cigarette exposure | 397 (46.7) | 328 (46.0) | 69 (50.0) | 0.44 |
| Smoking status | ||||
| Never | 454 (53.3) | 385 (54.1) | 69 (50.0) | 0.06 |
| Current | 149 (17.5) | 114 (16.0) | 35 (25.4) | |
| Passive smoking | 178 (20.9) | 153 (21.5) | 24 (17.4) | |
| Former | 70 (8.2) | 60 (8.4) | 10 (7.2) | |
| Assessment of respiratory symptom | ||||
| 1. Do you snore loudly (louder than speaking or through a closed door) (%) | 380 (46.0) | 306 (44.2) | 74 (55.6) | 0.020 |
| 2. Do you feel tired or sleepy during the day (%) | 282 (34.2) | 221 (32.0) | 61 (45.9) | 0.003 |
| Physical examination | ||||
| BMI, kg/m2 | 24.4 (22.5, 26.9) | 24.1 (22.2, 26.6) | 26.7 (24.1, 28.8) | <0.001 |
| HR (times/min) | 72 (66, 80) | 72.00 (66, 79) | 74 (68, 82) | 0.018 |
| SBP | 131 (119, 144) | 131 (119,144) | 134 (123, 147) | 0.048 |
| Neck circumference, cm | 36 (34, 39) | 36 (34, 39) | 37 (35, 40) | <0.001 |
| Waist circumference, cm | 88 (81, 96) | 88 (80, 94) | 95 (87, 101) | <0.001 |
| Hip circumference, cm | 99 (94, 104) | 99 (94, 103) | 103 (97, 108) | <0.001 |
| Lab examination | ||||
| FEV1, L | 2.18 (1.88, 2.56) | 2.19 (1.91, 2.58) | 2.09 (1.80, 2.40) | 0.012 |
| FEV6, L | 2.59 (2.23, 3.13) | 2.62 (2.25, 3.16) | 2.44 (2.04, 2.86) | 0.001 |
| FEV1/FEV6, % | 0.85 (0.80, 0.88) | 0.85 (0.80, 0.88) | 0.86 (0.81, 0.90) | 0.043 |
| FEV1%pred, % | 0.90 (0.14) | 0.91 (0.14) | 0.86 (0.14) | <0.001 |
| IL-6, pg/mL | 1.76 (1.11, 2.63) | 1.59 (1.05, 2.38) | 2.75 (1.86, 4.37) | <0.001 |
| WCC, ×109 cell/L | 5.93 (4.95, 6.85) | 5.76 (4.85, 6.60) | 6.79 (5.99, 7.91) | <0.001 |
| PLT, ×109 cell/L | 223 (191, 261) | 218 (189, 254) | 240 (200,277) | <0.001 |
| NEU, % | 57.23 (7.96) | 56.90 (7.95) | 58.90 (7.88) | 0.007 |
| LYM, % | 32.39 (7.41) | 32.75 (7.35) | 30.53 (7.46) | 0.001 |
| EOS, % | 2.00 (1.20, 3.20) | 2.00 (1.20, 3.10) | 2.20 (1.50, 3.20) | 0.041 |
| Glucose, mmol/L | 4.76 (4.21, 5.56) | 4.73 (4.20, 5.49) | 4.91 (4.33, 5.97) | 0.040 |
| hsCRP, mg/L | 1.18 (0.62, 2.22) | 0.99 (0.56, 1.61) | 5.22 (3.74, 7.87) | <0.001 |
BMI, body mass index; EOS, eosinophil; HR, heart rate; hsCRP, high-sensitivity C reactive protein; lym, lymphocyte; OSAHS, Obstructive Sleep Apnea Hypoventilation Syndrome; plt, platelet; SBP, systolic blood pressure; WCC, white cell count.
Additional baseline characteristics of the population with the hsCRP data are presented in online supplemental appendix E1. Online supplemental appendix E2 presents the analysis of the baseline characteristics of the population with and without the hsCRP data. Online supplemental appendices E3–E5 present the analysis only for the hsCRP normal population grouped according to hsCRP, FEV1%pred and FEV1/FEV6 quartiles, respectively.
Supplementary data
Association of hsCRP with lung function
Univariate analyses of hsCRP, FEV1/FEV6 and FEV1%pred (across quartiles) were performed separately for the participants with normal hsCRP. The results are presented in online supplemental appendices E6–E8.
We included all variables with p<0.05 in the univariate analysis. All variables with p<0.05 in the baseline and <20% missing data in the corresponding LASSO regression analyses were included to screen for the variables with the greatest effect on the association between hsCRP and lung function (online supplemental appendices Efig1–Efig3). The best variables screened by LASSO were included in the multifactorial model along with the sex and age to determine the final model variables (table 2). The results of the multifactorial analysis suggested that the final variables affecting FEV1/FEV6 were age, sex, physical condition (self-evaluation of the degree of physical health), body mass index (BMI) and waist circumference. The final variables affecting FEV1%pred were sex, age, hypertension, alcohol consumption, haematuria, type of occupation, BMI, waist circumference and white cell count.
Table 2
Multifactorial analysis of the factors affecting lung function
| Variables | FEV1%pred | FEV1/FEV6 | ||||||
| OR (95% CI) | P value | Adjusted OR (95% CI) | P value | OR (95% CI) | P value | Adjusted OR (95% CI) | P value | |
| Age | 1.012 (1.000 to 1.024) | 0.06 | 0.997 (0.984 to 1.010) | 0.644 | 1.012 (1.001 to 1.024) | 0.028 | 1.004 (0.991 to 1.017) | 0.57 |
| Sex | ||||||||
| Female | Reference | 1 | Reference | 1 | Reference | 1 | Reference | 1 |
| Male | 2.368 (1.980 to 2.833) | <0.001 | 1.863 (1.498 to 2.316) | <0.001 | 2.988 (2.514 to 3.552) | <0.001 | 3.121 (2.547 to 3.826) | <0.001 |
| Hypertension | ||||||||
| No | Reference | 1 | Reference | 1 | ||||
| Yes | 1.627 (1.374 to 1.927) | <0.001 | 1.291 (1.080 to 1.543) | 0.005 | ||||
| Drinking | ||||||||
| No | Reference | 1 | Reference | 1 | ||||
| Yes | 1.363 (1.116 to 1.664) | 0.002 | 0.689 (0.544 to 0.873) | 0.002 | ||||
| Physical condition | ||||||||
| Very good | Reference | 1 | Reference | 1 | ||||
| Good | 0.618 (0.447 to 0.853) | 0.003 | 0.661 (0.462 to 0.947) | 0.024 | ||||
| Acceptable | 0.620 (0.455 to 0.843) | 0.002 | 0.581 (0.412 to 0.820) | 0.002 | ||||
| Not good | 0.386 (0.213 to 0.699) | 0.002 | 0.371 (0.196 to 0.701) | 0.002 | ||||
| Very bad | 0.426 (0.066 to 2.766) | 0.371 | 0.794 (0.093 to 6.750) | 0.83 | ||||
| Haematuria | ||||||||
| No | Reference | 1 | Reference | 1 | ||||
| Yes | 1.671 (1.386 to 2.015) | <0.001 | 1.629 (1.351 to 1.966) | <0.001 | ||||
| Current/previous occupation | ||||||||
| Worker | Reference | 1 | Reference | 1 | ||||
| Public officer | 0.961 (0.760 to 1.215) | 0.74 | 0.865 (0.681 to 1.098) | 0.23 | ||||
| Teacher | 0.254 (0.132 to 0.487) | <0.001 | 0.400 (0.215 to 0.745) | 0.004 | ||||
| Military personnel | 1.754 (0.247 to 12.439) | 0.57 | 0.249 (0.056 to 1.101) | 0.07 | ||||
| Farmer | 0.493 (0.155 to 1.568) | 0.23 | 0.194 (0.064 to 0.592) | 0.004 | ||||
| Doctor | 0.817 (0.426 to 1.568) | 0.54 | 1.824 (0.929 to 3.585) | 0.08 | ||||
| Other occupations | 0.783 (0.638 to 0.96) | 0.019 | 0.783 (0.64 to 0.958) | 0.017 | ||||
| BMI* | 1.065 (1.038 to 1.093) | <0.001 | 0.906 (0.869 to 0.945) | <0.001 | 0.97 (0.947 to 0.994) | 0.015 | 0.867 (0.829 to 0.907) | <0.001 |
| Waist circumference* | 1.048 (1.039 to 1.057) | <0.001 | 1.059 (1.043 to 1.075) | <0.001 | 1.022 (1.014 to 1.03) | <0.001 | 1.033 (1.017 to 1.049) | <0.001 |
| WCC* | 1.182 (1.115 to 1.253) | <0.001 | 1.101 (1.037 to 1.169) | <0.001 | ||||
| hsCRP* | 1.244 (1.104 to 1.402) | <0.001 | 1.039 (1.017 to 1.061) | <0.001 | 0.72 (0.643 to 0.806) | <0.001 | 0.983 (0.964 to 1.003) | 0.092 |
*Epresents the continuous variables.
BMI, body mass index; FEV1%pred, predict the forced expiratory volume in 1 s; hsCRP, high-sensitivity C reactive protein; WCC, white cell count.
A non-linear relationship was observed between hsCRP and FEV1/FEV6 in the participants with normal hsCRP, whereas after adjusting for other variables, no non-linear relationship was observed between hsCRP and FEV1%pred (figure 3). In addition, because of the linearity test yielded p<0.05, we used a linear relationship to explain the relationship as a simple model was preferred. According to the findings listed in table 3, CRP was negatively correlated with FEV1%pred, with a change of 1 SD unit in CRP resulting in a lower of 2.4% in FEV1%pred.
Non-linear relationship between hsCRP and lung function. FEV1, forced expiratory volume in 1 s; hsCRP, high-sensitivity C reactive protein.
Table 3
Linear association between hsCRP and lung function
| Variables | Lung function | Estimated changes (L) (95% CI) by per SD of hsCRP | P value | hsCRP ≤3 mg/L | |||||
| Estimated changes (L) (95% CI) by quartiles of hsCRP | |||||||||
| Q1 | Q2 | Q3 | Q4 | P trend | P interaction | ||||
| All participants | FEV1%pred, % | −2.4 (−4.4 to, –0.5) | 0.015 | 0 (reference） | 0.3 (-3.3 to 3.9) | −1.6 (−5.4 to 2.1) | −3.6 (−7.4 to 0.2) | 0.038 | |
| Female | FEV1%pred, % | −2.6 (−4.9 to, –0.4) | 0.022 | 0 (reference） | 4.1 (−0.3 to 8.7) | −0.9 (−5.6 to 3.8) | −1.8 (−6.5 to 2.9) | 0.14 | 0.87 |
| Male | FEV1%pred, % | −3.0 (−6.4 to 0.3) | 0.08 | 0 (reference） | −4.3 (−10.1 to 1.4) | −5.0 (−10.8 to 0.8) | −7.3 (−13.1 to –1.4) | 0.018 | |
| Haematuria (yes） | FEV1%pred, % | −2.8 (−6.6 to 0.9) | 0.14 | 0 (reference） | −0.8 (−7.8 to 6.2) | −4.8 (−11.9 to 2.2) | −6.8 (−1.4 to 0.3) | 0.034 | 0.76 |
| Haematuria (no) | FEV1%pred, % | −2.2 (−4.4 to, –0.1) | 0.046 | 0 (reference） | −0.7 (−5.0 to 3.6) | −0.9 (−5.2 to 3.4) | −2.7 (−7.1 to 1.6) | 0.23 | |
Among the participants with hsCRP data, 713 participants (83.8%) had hsCRP within the normal range. To reduce the uncertainty of the results due to the small sample size, multiple linear regression analysis was performed including only the participants with hsCRP ≤3 mg/L. FEV1%pred resulted in adjusted age, sex, BMI, waist circumference and haematuria.
BMI, body mass index; FEV%pred, predict the forced expiratory volume in; hsCRP, high-sensitivity C reactive protein.
The subgroup analysis included the variables reported to be correlated with FEV1%pred and FEV1/FEV6, as well as the variables that were statistically significant according to the statistical tests (table 2). A non-linear relationship was observed between hsCRP and FEV1/FEV6 in the subgroup without haematuria (figure 4). A linear relationship was observed between hsCRP and FEV1%pred in the subgroup without haematuria. A linear trend was observed between hsCRP and FEV1%pred in the male subgroup and the subgroup with haematuria (table 3).
Non-linear association between hsCRP and lung function (Subgroup analysis). FEV1, forced expiratory volume in 1 s; hsCRP, high-sensitivity C reactive protein.
A curvilinear relationship was observed between CRP and FEV1/pred in the male subgroup, and a gradually decreasing curvilinear relationship between was observed CRP and FEV1/pred in the female subgroup. The interaction test resulted in p>0.05, indicating that there was no interaction between the male and female subgroups (figure 5). According to the findings presented in table 3, CRP had a significant negative correlation (p<0.05) with FEV1%pred in the female subgroup with an effect value of −2.6. The effect value was −3.0 in the male subgroup (p=0.08>0.05), and the trend test yielded p=0.018<0.05, indicating a negative trend of correlation between CRP and FEV1%pred.
Non-linear association between hsCRP and lung function (Interaction analysis). FEV1, forced expiratory volume in 1 s; hsCRP, high-sensitivity C reactive protein.
A curve relationship was observed between CRP and FEV1%pred in the subgroup with haematuria, whereas a progressively decreasing curve relationship was observed between CRP and FEV1%pred in the subgroup without haematuria. The interaction test yielded p>0.05, indicating that there was no interaction between the subgroups with and without haematuria (figure 5). According to the findings presented in table 3, although the subgroup with haematuria had an overall p=0.14>0.05, the trend test with p<0.05 indicated a negative correlation between CRP and FEV1%pred, with an effect value of −2.6. The effect value was −2.2 for the subgroup without haematuria, and a significant negative correlation was observed between CRP and FEV1%pred (p<0.05). The interaction test yielded p>0.05, indicating no significant difference in the change of effect values between the two subgroups.
A curvilinear relationship was observed between CRP and FEV1/FEV6 the male and female subgroups. The interaction test yielded p=0.011<0.05, indicating an interaction between the male and female subgroups (figure 5).
A curve relationship was observed between CRP and FEV1/FEV6 in the subgroups with and without haematuria. The interaction test yielded p>0.05, indicating there was no interaction between the two subgroups (figure 5).
Discussion
This study addresses the blind spot regarding the association between hsCRP and lung function in Chinese individuals aged 40–75 years without underlying lung disease. We observed a linear negative association between hsCRP and FEV1%pred and a non-linear association between hsCRP and FEV1/FEV6, even though the hsCRP levels were in the normal range. hsCRP may serve as a predictor of FEV1%pred and could rationally predict the development of COPD. Sex was found to be an independent factor influencing the relationship between hsCRP and FEV1/FEV6. The interaction between haematuria and hsCRP on FEV%pred deserves further investigation.
FVC can be exhaled in 6 s in healthy adults. Previous studies have also reported that FEV6 is approximately equal to FVC.11 12 Thus, FEV1/FEV6 used in this study can represent FEV1/FVC. An FEV1/FVC value of <0.7 indicates airflow limitation and is commonly used to diagnose COPD; however, it does not correlate with disease severity. Most guidelines define disease severity based on FEV1% pred, especially in patients with COPD. FEV1%pred was found to be more valuable than other indicators of lung function in predicting the onset of chronic respiratory symptoms in the general population, and it is a reliable indicator of the occurrence and prognosis of respiratory diseases, such as COPD and asthma.13
Lower concentrations of hsCRP can be measured depending on the development of technology related to the assessment of CRP levels. Several previous studies have reported that CRP levels that were previously below normal levels could be defined with hsCRP to assess the cardiovascular risk. Therefore, we used hsCRP as a study variable to predict lung function to obtain an accurate account of the association of CRP with lung function. Since adults aged between 40 and 75 years have a high likelihood of impaired lung function and a significantly increased prevalence of respiratory disease, this population should be prioritised during the prevention and treatment of respiratory diseases.14 Therefore, 1047 adults aged 40–75 years from Beijing, China, with FEV1/FVC >70% after receiving inhalational bronchodilators without any underlying lung disease were included in this study. The non-linear relationship between hsCRP and lung function was assessed using a restricted cubic spline regression model, and the linear association between hsCRP level and lung function was assessed using a multivariate regression model.
We reviewed the literature related to hsCRP and lung function and summarised the characteristics of 17 relevant studies published from 2006 to the present3 14–29 (online supplemental appendix E9). The published studies on the association between CRP levels and lung function have certain limitations: (1) most studies did not use hsCRP, thereby failing to accurately address the association between low CRP levels and lung function; (2) most studies did not examine data from a population without underlying lung disease, thereby failing to exclude the influence of underlying lung disease biasing the study conclusions; (3) the age of the population in most studies was either young or broad, focusing on individuals aged 20–85 years without precise targeting of the key study population, which increased the bias due to age in the study results; (4) no Chinese studies have investigated the association between hsCRP and lung function and (5) most of these studies lacked clinical application value. Our study population compensates well for these deficiencies and targets two key indicators of lung function for predicting COPD, namely FEV1%pred and FEV/FVC, to clarify the feasibility of the application of the study. In addition, the findings of these 17 studies overwhelmingly support the existence of a negative correlation between hsCRP and lung function, which is consistent with our findings. Gimeno et al 19 fitted the association of inflammatory indicators, such as hsCRP, with FVC, FEV1, TLC (total lung volume), DLCO (carbon monoxide dispersion) and MEF25 using a restricted cubic spline regression method and reported a non-linear negative association between hsCRP and these indicators. However, this study did not completely exclude the effect of underlying lung disease on lung function in the population, and there were no relevant studies on the two indices, FEV1/FEV6 and FEV1%pred. Thus, the exploration of the clinical significance of hsCRP is lacking. Our study completely excluded individuals with underlying lung disease, which controls for confounding factors affecting lung function and hsCRP, and highlighted studies on FEV1/FVC and FEV1%pred, which are extremely important for predicting COPD and prognosis. Since 713 (83.8%) of the 851 participants (with hsCRP data) in this study had hsCRP levels within the normal range, we only performed restricted cubic spline regression and multiple linear regression analysis to determine the association between the normal levels of hsCRP and lung function to reduce the uncertainty of the results due to the small sample size. We observed two patterns of association between hsCRP and the FEV1/FEV6 and FEV1%pred indices in the population with hsCRP within the normal range: a non-linear association between hsCRP and FEV1/FEV6 and a negative linear association between hsCRP and FEV1%pred. Our results indicate the potential of hsCRP as a predictor of lung function and specify the effective predictive range of hsCRP.
We performed subgroup and interaction analyses for sex and the presence of haematuria and observed an interaction between FEV1/FEV6 and sex and hsCRP using restricted cubic spline regression analysis. The findings of our study suggest that the association between hsCRP and lung function is not independent and that it is also influenced by sex. In the male and female subgroups, the effect values of CRP on FEV1/FEV6 differences were 0.2 and 1.2, respectively, with significantly higher FEV1/FEV6 differences observed in females than that in males (p=0.011 for the interaction). In addition, our study found a curvilinear relationship between CRP and FEV1/pred in males and a progressively decreasing curvilinear relationship between CRP and FEV1/pred in females; however, there was no interaction between the male and female groups. A simpler linear model was chosen for the analysis, and a significant negative correlation was observed between CRP and FEV1%pred in the female subgroup and a negative trend between CRP and FEV1%pred in the male subgroup. Therefore, we inferred that hsCRP has a stronger association with lung function in females than that in males, and our findings are consistent with those of a Danish study.30
In the subgroup with haematuria, a curve relationship was observed between hsCRP and FEV1%pred, whereas a gradually decreasing curve relationship was observed between hsCRP and FEV1%pred in the subgroup without haematuria. The interaction test suggested that there was no interaction between the two subgroups. Linear correlation analysis revealed a negative trend between CRP and FEV1%pred in the subgroup with haematuria and a significant negative correlation between CRP and FEV1%pred in the subgroup without haematuria, with effect values of −2.8 and −2.2 in the two groups, respectively. There was no interaction between the two groups and no difference in the alteration effect. A curve relationship was also observed between CRP and FEV1/FEV6 in the subgroups with and without haematuria, and there was no interaction between the two subgroups.
Although haematuria was not a statistically significant factor influencing the hsCRP on FEV1%pred as an independent factor. However, the graphical trend shows that there is a negative correlation between hsCRP and FEV1%pred in the non-haematuric group, while there is no significant negative correlation in the haematuric group. Whether this means that hsCRP is a better predictor of FEV1%pred in patients with non-haematuria, there are no similar reports, and the interaction between haematuria and hsCRP on FEV%pred has to be further investigated. In this study, due to the small sample size of the subgroup with haematuria compared with that of the subgroup without haematuria, the results may be biased and need to be confirmed by cohort studies with larger sample sizes.
Our study has some limitations. First, the number of women enrolled in the study exceeded the number of men, which may have led to sex bias in the results. However, we corrected for sex bias by incorporating sex into the model and conducting subgroup and interaction analyses of sex during the analysis. Second, the sample size was insufficient for subgroup analysis; therefore, it is necessary to expand the sample size to further explore the effect of other variables on the final results. Third, this study was a cross-sectional analysis based on a cohort study, and whether hsCRP is a good predictor of lung function indicators needs to be verified in our subsequent data analysis. Lastly, the data for hsCRP levels >3 mg/L in our study are insufficient, and there is a need to design studies that investigate the association of hsCRP levels >3 mg/L with lung function to complement our findings.
In this study, hsCRP had a non-linear association with FEV1/FEV6 and a linear negative association with FEV1%pred, although it was in the normal range. This suggests the possibility that hsCRP may predict the occurrence of COPD in the future. Due to the high prevalence of COPD and the lack of predictors, searching for valuable predictors of the development of COPD remains a great challenge to pulmonologists. The present results suggest that hsCRP can be a feasible predictor of FEV1%pred that may be helpful for the early recognition and prevention of COPD.
Footnotes
Contributors: GW designed this study: XY, JL, CheZ, XM, ChuZ, KS, YW and XM collected data. XY analysed the data and SZ provided statistical advice. All authors contributed to the interpretation of the data. XY drafted the manuscript and the other authors revised it for important intellectual content. All authors approved the final draft of the manuscript for submission. All authors had full access to all data (including statistical reports and tables) in the study, and GW and JL had final responsibility for the decision to submit for publication. GW is the guarantor.
Funding: This work was supported by grants from the National Key Research and Development Plan (No.2017YFC1309500) and Beijing Health Technologies Promotion Program (BHTPP202053). All authors are independent of the funders.
Competing interests: None declared.
Patient and public involvement: Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Provenance and peer review: Not commissioned; externally peer reviewed.
Supplemental material: This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.
Data availability statement
The dataset analysed in the current study is not publicly available because of contracts with the hospitals that provide data to the database.
Ethics statements
Patient consent for publication
Consent obtained directly from patient(s).
Ethics approval
This study involves human participants and was approved by Biomedical Research Ethics Committee of the Peking University First Hospital (EC201831). Participants gave informed consent to participate in the study before taking part.