Introduction

Chronic obstructive pulmonary disease (COPD) is a chronic, progressive lung disease and an increasing cause of global morbidity and mortality [1]. In contrast to the trend for cancer and cardiovascular diseases, death rates from COPD have been progressively rising over the last decade, [2] and, according to estimates of the World Health Organization (WHO), by 2030 this condition may become the third leading cause of death worldwide [3].

Over the last few years, there has been an increased awareness about the extrapulmonary manifestations of COPD, [4] and the frequent and important chronic comorbidities that may contribute significantly to its severity and mortality [5]. In particular, there is a growing body of literature indicating that cardiovascular disease is more frequent in COPD patients than in the general population, representing a burden even greater than that of the lung disease itself in this clinical setting [6]. It has been estimated that for every 10% decrease in lung function [as expressed by forced expiratory volume in 1 s (FEV1)], cardiovascular mortality increases by nearly 30% [7].

COPD and cardiovascular disease share some major risk factors such as smoking habit, diabetes mellitus, hypertension and obesity [8, 9]. However, some evidence confirms that COPD is associated with cardiovascular manifestations independently from concomitant risk factors [7, 10], thus suggesting that COPD itself may be considered an independent cardiovascular risk factor. To further address this issue, a growing attention has been given to the assessment of the association between COPD and subclinical atherosclerosis, a recognized marker of cardiovascular disease [11].

Endothelial dysfunction is the earliest stage of the atherosclerotic process and even a trigger of cardiovascular events [12]. Flow-mediated dilatation (FMD) and nitrate-mediated dilatation (NMD) are widely accepted as accurate and non-invasive methods to assess endothelial function in humans [13], and in turn, as surrogate markers of subclinical atherosclerosis [14]. Moreover, FMD is currently considered an independent predictor of cardiovascular events [15], thus providing important prognostic data over and above traditional cardiovascular risk factors.

Some prospective studies [16, 17] document impaired endothelial function in COPD patients. However, these data have been challenged in recent studies, [18, 19] and no meta-analytical data providing an overall information about this issue are currently available.

The aim of the present study is to perform a systematic review with meta-analysis of all studies evaluating the impact of COPD on FMD and NMD. Moreover, we implemented some meta-regression models to evaluate the effect of some clinical and demographic variables on these outcomes.

Methods

Search strategy

To identify all available studies, a detailed search pertaining to COPD and the markers of cardiovascular risk (i.e., FMD and NMD) was conducted according to PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analyses) guidelines [20]. A systematic search was performed in the electronic databases (PubMed, Web of Science, Scopus, EMBASE), using the following search terms in all possible combinations: chronic obstructive pulmonary disease, flow-mediated dilatation, nitrate-mediated dilatation, endothelium-dependent dilatation, endothelium-independent dilatation, endothelial function, and endothelial dysfunction. The last search was performed on 20th of December 2016. The search strategy was developed without any language restriction.

In addition, the reference lists of all retrieved articles were manually reviewed. In case of missing data, study Authors were contacted by e-mail to try to retrieve original data. Two independent Authors (PA and MNDDM) analyzed each article, and performed the data extraction independently. In case of disagreement, a third investigator was consulted (RL). Discrepancies were resolved by consensus. Selection results showed a high inter-reader agreement (κ = 1.00), and have been reported according to PRISMA flowchart (Fig. 1).

Fig. 1
figure 1

Prisma flow diagram. Asterisk indicates three studies without control groups, one study on acute exacerbation of chronic obstructive pulmonary disease

Full size image

Data extraction and quality assessment

All studies evaluating the impact of COPD on the markers of cardiovascular risk were included. Case reports, case series without a control group, reviews and animal studies were excluded. To be included in the analysis, a study had to provide values (means with standard deviation or standard error) of brachial artery FMD or NMD among COPD patients and controls. The included studies were classified as having a prospective or a retrospective design.

In each study, data regarding sample size, major clinical and demographic variables, values of FMD and NMD in COPD patients and controls were extracted. Data on patients with acute exacerbation of COPD were excluded.

Given the characteristics of the included studies, the evaluation of methodological quality of each study was performed with the Newcastle–Ottawa Scale (NOS), which is specifically developed to assess quality of non-randomized observational studies [21]. The scoring system encompasses three major domains (selection, comparability, exposure) and a resulting score range between 0 and 8, a higher score representing a better methodological quality. Results of the NOS quality assessment are reported in Supplemental Table 1.

Statistical analysis and risk of bias assessment

Statistical analysis was carried out using the Review Manager software (Version 5.2, The Cochrane Collaboration, Copenhagen, Denmark).

Differences among cases and controls are expressed as mean difference (MD) with pertinent 95% confidence intervals (95% CI).

Endothelial function has been expressed as the percentage change (%) in the brachial artery diameter from baseline to maximal dilatation during reactive hyperemia (FMD) or after sublingual nitrate administration (NMD).

The overall effect is tested using Z scores and significance is set at P < 0.05. Statistical heterogeneity between studies is assessed with Chi-square Cochran’s Q test and with I 2 statistic, which measures the inconsistency across study results, and describes the proportion of total variation in study estimates, that is due to heterogeneity rather than sampling error. In detail, I 2 values of 0% indicate no heterogeneity, 25% low, 25–50% moderate, and 50% high heterogeneity [22].

Publication bias is represented graphically by funnel plots of the effect size (mean difference) versus precision (1/standard error of the mean difference) for studies evaluating flow-mediated dilatation and nitrate-mediated dilatation in patients with COPD, and in control subjects using both fixed and random effect. Visual inspection of funnel plot asymmetry is performed to address the possible small-study effect, and Egger’s and Begg and Mazumdar tests are used to assess publication bias, over and above any subjective evaluation. A P < 0.10 is considered statistically significant [23]. Moreover, the Duval and Tweedie’s trim and fill analysis is used to allow for the estimation of an adjusted effect size after trimming and imputing studies [24].

To be as conservative as possible, the random-effect method is used to take into account the variability among included studies.

Sensitivity analyses

We repeated analyses by including only the studies judged as “high quality” according to NOS (i.e., NOS ≥ to the median value found among included studies). Moreover, to avoid the risk of data overlap, a sensitivity analysis was performed after excluding studies involving the same recruitment Centres and enrolling patients in the same period time as other included studies. We also plan to perform a further sensitivity analysis after stratifying studies according to design (prospective and retrospective).

Meta-regression analyses

We hypothesize that differences among included studies may be affected by demographic variables [mean age, male gender] and clinical data related to disease severity [FEV1, forced expiratory volume in 1 s/forced vital capacity (FEV1/FVC), Global Initiative for Chronic Obstructive Lung Disease (GOLD) class, Modified Medical Research Council (MMRC) class], disease activity [C-reactive protein (CRP), interleukin-6 (IL6) levels], COPD treatment [therapy with β-agonists, anticholinergic agents, corticosteroids (CCS), supplemental oxygen], concomitant use of cardiovascular medications [aspirin, statins, angiotensin converting enzyme-inhibitors (ACE-I), angiotensin receptor blockers (ARBs), β-blockers, calcium channel blockers (CCB)], and coexistence of traditional cardiovascular risk factors (hypertension, smoking habit, smoking pack-years, diabetes mellitus, obesity, hyperlipidemia). To assess the possible effect of such variables in explaining different results observed across studies, we planned to perform meta-regression analyses after implementing a regression model with changes in FMD or NMD as dependent variables (y) and the above mentioned covariates as independent variables (x). This analysis was performed with Comprehensive Meta-analysis [Version 2, Biostat, Englewood NJ (2005)].

Results

After excluding duplicate results, the search retrieved 512 articles. Of these studies, 385 were excluded because they were off the topic after scanning the title or the abstract, 115 because they were reviews/comments/case reports or they lacked data of interest. Another four studies were excluded after full-length paper evaluation.

Thus, eight articles were included in the final analysis [16,17,18,19, 25,26,27,28] (Fig. 1). In detail, eight studies (334 COPD patients and 281 controls) report data on FMD and two studies (104 COPD patients and 88 controls) evaluate NMD.

Study characteristics

All included studies are observational and have a prospective design. Major characteristics of study populations are shown in Table 1 and further data about disease activity, disease severity and treatment are reported in Supplemental Tables 2, 3.

Table 1 Demographic and clinical data of patients with chronic obstructive pulmonary disease (COPD) and controls in included studies
Full size table

The number of patients varies from 10 to 96, the mean age from 62 to 76.7 years, and the prevalence of male gender from 40 to 100%.

The presence of diabetes mellitus is reported by 0–43.3% of patients, obesity by 0%, hypertension by 0–86.7%, smoking history by 70.5–100%, and hyperlipidemia by 0–56.7%.

Mean body mass index (BMI) varies from 25 to 29.1 kg/m2 and mean number of smoking pack-years from 24.7 to 66. Mean values of total cholesterol (TC) ranged from 4.83 to 5.04 mmol/l, of LDL-cholesterol (LDLc) from 2.76 to 3.13 mmol/l, of HDL-cholesterol (HDLc) from 1.30 to 1.99 mmol/l, and of triglycerides (TGs) from 1.16 to 1.52 mmol/l.

Mean values of FEV1 vary from 41 to 61.9% predicted, of FEV1/FVC from 43 to 59, of CRP from 2.4 to 4 mg/l, and of IL6 from 2.5 to 4.1 pg/ml. Only one study [27] provides information on GOLD class (class I in 10%, class II in 43.3%, class III in 26.7%, class IV in 20%) and no study reports data on MMRC class.

An ongoing treatment with aspirin is reported by 31.3–52.3%, with statins by 16.7–20.4%, with ACE-I by 33.3–54%, with ARBs by 6.7–21.9%, with β-blockers by 11.4–18.8%, and with CCB by 28.1–30%.

The use of β-agonists is documented in 46.6–68% of patients, of anticholinergics in 23.3–44%, of CCS in 36.6–75%, and of supplemental oxygen in 0–66.7%.

Endothelial function assessment

In eight studies [16,17,18,19, 25,26,27,28], we find that the 334 COPD patients show a significantly lower FMD than the 281 controls (MD −3.15%; 95% CI −4.89, −1.40; P < 0.001, Fig. 2). Heterogeneity among these studies is statistically significant (I 2 = 96%; P < 0.001). Moreover, a significantly lower NMD is found in two studies [25, 28] evaluating a total of 104 COPD patients and 88 controls (MD −3.53%; 95% CI −7.04, −0.02; P = 0.049, I 2 = 83.1%; P = 0.015).

Fig. 2
figure 2

Forest plot of the difference in flow-mediated dilatation (FMD) among chronic obstructive pulmonary disease (COPD) patients and controls. 95% CI 95% confidence interval, COPD chronic obstructive pulmonary disease. Gray squares represent the mean difference in FMD values among COPD patients and controls in each study. Lines are the standard deviation from the mean difference. The black diamond represents the cumulative mean difference for all included studies

Full size image

Sensitivity analyses

The median value of NOS quality assessment is 7. Thus, the analyses were repeated by including only the five studies classified as “high quality” (NOS ≥7) [19, 25,26,27,28] (Supplemental Table 1). Of interest, after excluding studies classified as “low quality,” [16,17,18] our results are substantially confirmed both for FMD and NMD (Table 2).

Table 2 Sensitivity analyses
Full size table

Similarly, after excluding studies [28] potentially reporting on the same population as other included studies [18], the significant difference in FMD between COPD patients and controls is confirmed while results on NMD are unchanged (Table 2).

A sensitivity analysis according to study design (prospective or retrospective) was not performed because all included studies were prospective.

Publication bias

Because it is recognized that publication bias can affect the results of meta-analyses, we attempted to assess this potential bias using funnel plots analysis (Supplemental Fig. 1).

Funnel plots of effect size versus precision for studies evaluating FMD are rather symmetrical, suggesting the absence of publication bias and small-study effect. Egger’s and Begg and Mazumdar tests consistently confirm the absence of significant publication bias. Moreover, the Duval and Tweedie’s trim and fill analysis shows that all results are confirmed after trimming and imputing studies. Because of the low number of studies on NMD, publication bias was not assessed for this outcome.

Meta-regression analyses

Regression models for studies comparing COPD patients and healthy controls showed that age significantly impacts upon FMD (Z = 3.51; P < 0.001, Supplemental Fig. 2A).

Moreover, we find that a more severe disease (as expressed by lower FEV1 values) is associated with a more severe FMD impairment in COPD patients than in controls (Z = 3.76; P < 0.001, Supplemental Fig. 2B).

All the other demographic and clinical variables do not impact upon the observed results (Supplemental Table 4).

Because of the low number of studies, no meta-regression analysis was performed for NMD.

Discussion

Results of the present meta-analysis consistently show that COPD is associated with endothelial dysfunction. In particular, we demonstrate impaired FMD and NMD in COPD patients, confirmed by appropriate sensitivity analyses. Moreover, regression models are able to further refine results showing that a more severe disease (as expressed by lower FEV1 values) is associated with a higher difference in FMD between COPD patients and controls.

Endothelial dysfunction is the earliest stage of the atherosclerotic process, and a trigger of cardiovascular events [12]. Our findings extend results of previous studies [29], showing an increased carotid intima-media thickness (IMT) accompanied by higher prevalence of carotid plaques in patients with COPD.

Thus, the assessment of these markers consistently suggests an increased cardiovascular risk in COPD patients, which is widely confirmed by some epidemiological studies showing an increased incidence of major cardiovascular events in this clinical setting [6]. This supports the need for a strict monitoring of cardiovascular risk factors and signs of subclinical atherosclerosis in this clinical setting.

Many cardiovascular risk factors are thought to have a causal role in the atherosclerotic process [30]. Although most COPD patients have a smoking history, the relationship between subclinical atherosclerosis and COPD seems to be more complex, and the smoking habit might not entirely explain the accelerated atherosclerotic process in this clinical setting [10]. Accordingly, our regression models show that the association between COPD and endothelial dysfunction is independent of baseline smoking status and most traditional cardiovascular risk factors. Thus, other mechanisms should be considered to explain such an association.

Inflammatory and immunological mechanisms have been suggested to independently impact upon atherosclerosis progression in COPD patients [31]. Immune-mediated inflammation seems to play a crucial role in the pathogenesis of the atherosclerotic process [32,33,34], being involved in endothelial dysfunction, plaque rupture and thrombosis [35, 36]. COPD patients have elevated levels of CRP [37], a marker of inflammation associated with the atherosclerotic burden in the general population, [38, 39] and poor prognosis in COPD patients [40]. The reduction in the rate of decline of lung function with inhaled corticosteroid further confirms the role of inflammation in the pathogenesis of COPD [41]. More recently, the presence of systemic inflammation has also been documented in moderate COPD [31]. This may explain why even a modest decline in FEV1 (from a mean of 109% of predicted to 88% of predicted) is associated with a fivefold increase in cardiovascular morbidity and mortality [6]. However, a pooled analysis of large cohort studies (>500 patients) reporting on the relationship between FEV1 and cardiovascular mortality demonstrates that individuals in the lowest FEV1 quintile have a 75% increase in the risk for cardiovascular mortality compared with those in the highest FEV1 quintile [31]. Overall, these data suggest that cardiovascular risk progressively increases as the lung function decreases. This is consistent with the results of our regression analyses, showing that a more severe disease (as expressed by lower FEV1 values) is associated with a higher difference in FMD between COPD patients and controls.

Overall, our findings extend some previous evidence supporting the hypothesis that premature atherosclerosis may be one of the main features of COPD. The clinical relevance of our results can be better understood when we consider that each 1% decrease in FMD is associated with a 12% increase of cardiovascular events [42]. In contrast, few literature data are currently available on NMD, and its role as a predictor of cardiovascular events has not been properly studied [43]. In some studies on FMD, NMD is also assessed as a measure of smooth muscle function (endothelium-independent dilation). Interestingly, in the present meta-analysis, we find an impairment of both FMD and NMD, thus suggesting an impairment of both NO-dependent and NO-independent endothelial function in COPD patients. However, the very limited number of studies evaluating NMD (n = 2) should be taken into account. In addition, with the less used technique of forearm plethysmography, no significant impairment in endothelium-dependent or -independent vasomotor function has been previously documented in COPD patients [44].

Our results suggest the need for a strict monitoring of cardiovascular risk factors in COPD patients, with particular regard to those with severe disease. Careful risk stratification may orientate the choice on which COPD patients might benefit from a very strict monitoring of cardiovascular risk factors and of subclinical signs of atherosclerosis. Another issue is the possibility that COPD treatment may somehow have an impact upon the cardiovascular risk profile of these patients. The SUMMIT study [41] recently demonstrates that treatment with a combination of an inhaled corticosteroid and long-acting β-agonists reduces the rate of decline of lung function in COPD patients, but these benefits do not lead to reduction in overall mortality or cardiovascular events. However, this trial was limited to patients with moderate COPD. In the previous TORCH trial [45], inhaled combination therapy with salmeterol and fluticasone propionate seem to be associated with a 2.6% absolute reduction (17.5% relative risk reduction) in all-cause mortality in patients with moderate-to-severe COPD, and a post hoc analysis suggests that the combination treatment is also able to reduce cardiovascular mortality [46]. In keeping with this, a pilot study demonstrates that salmeterol has positive effects on alveolar fluid clearance in COPD patients, thus potentially reducing both cardiogenic and non-cardiogenic pulmonary oedema [47].

Some potential limitations of our study need to be discussed. First, studies included in our meta-analysis have different inclusion and exclusion criteria, although most of patients included in the analysis have concomitant cardiovascular risk factors (hypertension, smoking, obesity, diabetes mellitus, hyperlipidemia) and different disease activity status. Since meta-analysis is performed on aggregate data and some missing information is present in each study, the multivariate approach allows for the adjustment for some (but not all) potential confounders. Thus, although results of meta-regression analyses are able to refine analyses by assessing the influence of most clinical and demographic variables on the observed results, caution is necessary in overall results interpretation.

As a further limitation of our meta-analysis, heterogeneity among the studies is generally significant, and it is not possible to conclusively ascertain sources of heterogeneity. However, the impact of clinical and demographic data on the outcome has been evaluated by means of meta-regression models, thus partially offsetting the above limitation.

Finally, we have to consider that FMD measurement may be influenced by many confounding factors [48], significantly limiting reproducibility of FMD assessment, and in turn, the relevance of results. In particular, some studies indicate that cuff placement [49] or duration of cuff occlusion [50] may affect results.

In conclusion, COPD is significantly associated with endothelial dysfunction, and in turn, with cardiovascular morbidity and mortality. Thus, patients with severe COPD and lower FEV1 values may benefit from a more meticulous screening for cardiovascular risk factors and more specific cardiovascular prevention strategies.