Statin use is associated with reduced mortality in COPD

Chronic obstructive pulmonary disease (COPD) is characterised by chronic airflow limitation that is usually progressive 1. Since the mid 1990s mortality from COPD has increased, and COPD is expected to become the third greatest cause of mortality in Western countries by the year 2020 2. The majority of the patients are current or former smokers 1, 3. Thus, patients with COPD are likely to develop other smoking-related diseases as well, predominantly symptomatic or asymptomatic ischaemic heart disease (IHD) 4, 5.

Chest pain and dyspnoea are the most common symptoms in patients with IHD 6. The likelihood of these symptoms increases with increasing intensity of exercise. Likewise, patients with COPD frequently experience dyspnoea or even chest pain on exertion or during an exacerbation. Thus, in COPD patients these symptoms may be misinterpreted as COPD-related symptoms even if the origin is cardiac. Moreover, exercise capacity may be limited by impaired lung function concealing the coronary symptoms. Unfortunately, exercise testing and pharmacological stress testing for detection of myocardial ischaemia is often poorly suited for patients with COPD. Hence, IHD among COPD patients may persist undiagnosed even though cardiovascular disease is increasingly being recognised as a leading cause of death in COPD 7.

After an exacerbation of COPD the median survival is only 3–5 yrs 8, 9. The prognosis is particularly serious in patients with hypercapnia and the presence of several comorbidities. Apart from long-term oxygen treatment, which is restricted to patients with chronic respiratory failure, no current therapies for COPD are known to alter long-term prognosis 3. Whether inhaled corticosteroids (ICS) improve the prognosis is still a matter of controversy 10. By contrast, several drugs have been shown to improve prognosis after acute coronary events during the last 20 yrs, in particular hydroxymethylglutaryl CoA reductase inhibitors (statins) 11.

The current authors hypothesised that many COPD patients might have subclinical IHD and that use of statins might be associated with improved survival. Moreover, whether concurrent use of ICS would modify outcome was also thought to warrant exploration.

METHODS

DATA ANALYSES

All-cause mortality after discharge from the hospital was used as the outcome measure. First, the age-adjusted relative mortality was assessed using the Mantel–Haenszel test for incidence density data. Secondly, a multivariate analysis was performed using Cox proportional hazards regression. The covariates listed in table 2⇓ were included in the initial multivariate model. If the age-adjusted association between statin use and mortality deviated meaningfully from the age-adjusted mortality of statin, a product-term between the corresponding covariate and statin was included in the initial multivariate model. Backward elimination of covariates was performed if the covariate was not significant and the removal of the covariate did not cause a meaningful change of the association between mortality and treatment with statins. The time after discharge from the hospital was used as the underlying time variable. The proportional hazard assumption was tested using Schoenfeld′s residuals on partial likelihood 12. Results are reported as rate ratio (RR) for stratified analyses or hazard ratio (HR) for Cox analyses, with 95% confidence intervals (CI) in parentheses.

RESULTS

The study consisted of 854 patients. The median follow-up was 1.9 yrs. Cumulative survival was 75.4, 65.0, 59.6 and 55.6% at 1, 2, 3 and 4 yrs after discharge, respectively. The crude relative mortality ratio among statin users versus nonusers was 0.58 (0.39–0.84), whereas the age-adjusted relative mortality ratio was 0.66 (0.45–0.96). The age-adjusted relative mortality for separate covariates is shown in table 2⇑. Bivariate analyses indicated that the adjusted mortality was lower in statin users than nonusers in almost all the subgroups (table 2⇑). The following covariates were included as product terms with statin in the initial multivariate Cox model: IHD; congestive heart failure; cerebrovascular disease; and treatment with acetyl salicylate. A product term with ICS was also added.

The results of the Cox regression analysis are presented in table 3⇓. The presence of IHD, congestive heart failure, atrial fibrillation, diabetes, venous thromboembolism, lung cancer, and cancers other than lung cancer, were all associated with increased mortality (table 3⇓). The HR for pneumonia was, however, not higher than in the non-pneumonia group (HR 0.98; p = 0.99). Moreover, the HR for females was not significantly different from that of males (HR 0.89; p = 0.38), and no significant association between current smoking and mortality was found (HR 1.2; p = 0.32).

The HR for patients treated with statins was 0.57 (0.38–0.87) compared with statin non users (table 3⇑). None of the product terms between statins and IHD, congestive heart failure, cerebrovascular disease, acetyl salicylate, or ICS were significant (p = 0.20, 0.16, 0.33, 0.57 and 0.48, respectively), suggesting no effect modification on statins by any of these covariates. Concomitant treatment with angiotensin converting enzyme inhibitors or angiotensin receptor blockers (HR 1.0 (0.73–1.4)), aspirin (HR 0.94 (0.68–1.3)), β-blockers (HR 0.92 (0.61–1.4)), or warfarin (HR 1.0 (0.67–1.6)) did not affect the mortality risk. However, treatment with ICS was associated with improved survival (HR 0.73 (0.57–0.94)).

Finally, statin users and statin nonusers were stratified into ICS users and ICS nonusers. Cox regression analysis, employing the same covariates as in table 3⇑, revealed that the relative mortality among ICS users only, statin users only, and those using a combination of statins and ICS was 0.75 (95% CI 0.58–0.98), 0.69 (95% CI 0.36–1.3), and 0.39 (95% CI 0.22–0.67), respectively, compared with patients receiving none of these treatments (p-trend <0.001). For the trend test, the coefficients from the corresponding Cox analysis were used as scores.

The final model did not violate the proportional hazard assumption for any of the covariates or the global test (p = 0.63).

DISCUSSION

It has been found that the mortality among COPD patients after discharge from hospital was lower among those who were taking a statin compared with those who did not. Furthermore, combined use of statins and ICS was associated with a more favourable prognosis than use of statin alone.

In the present study, nearly 30% of the patients had diagnosed IHD. Additionally, it is likely that several of the patients had unrecognised IHD. In an autopsy study at the hospital, 106 out of 144 (74%) of the patients with COPD diagnosis had coexisting heart disease 13. Moreover, in a pilot study of 30 COPD patients hospitalised for exacerbations, 6 patients had an elevated level of troponin T. Three (50%) of the patients with elevated troponin T died 5 months after discharge from the hospital, whereas only one (4%) patient died among the remaining 24 patients (p = 0.03; age-adjusted log-rank) 14. High prevalence of troponin elevation during exacerbation of COPD has also been described previously 15. Nevertheless, only 12% of these patients received a diagnosis of acute coronary syndrome. Thus, it appears a considerable proportion of COPD patients may have undiagnosed IHD.

Treatment with statins is now a cornerstone therapy for patients with IHD 11. A meta-analysis of 90,056 patients in 14 randomised trials showed a 12% proportional reduction in all-cause mortality, and 19% for coronary and 18% for respiratory mortality 16. The latter relationship was, however, not significant. As symptoms of cardiac ischaemia may be misinterpreted as pulmonary symptoms and many COPD patients have too low an exercise capacity to provoke coronary symptoms, it is difficult to identify COPD patients with concomitant IHD. Thus, it appears likely that COPD patients should benefit from statin treatment.

Recently, Mancini et al. 17 reported that statin treatment reduces mortality among COPD patients in a register-based retrospective cohort study. Interestingly, a decreased number of hospitalisations due to COPD exacerbation was also found. In an experimental study of rat lungs it was found that simvastatin had a suppressing effect on the inflammatory process induced by cigarette smoking 18. Thus, the beneficial effect of statins may be mediated by an anti-inflammatory effect in the lungs and the airways. In this regard, statin treatment may increase the anti-inflammatory effect on the airways that has been observed with ICS treatment 19. In the present study there was no effect modification by ICS on the association between statins and mortality, in agreement with Mancini et al. 17.

There is growing evidence that COPD is associated with a systemic inflammatory component, the intensity of which relates to the severity of the underlying disease 20. It is conceivable that this systemic component may exacerbate the inflammatory process that is associated with atherosclerosis and atherothrombosis. An association between airways inflammation and cardiovascular events has previously been suggested in studies of the effects of air pollution 21.

There is still controversy concerning a possible survival benefit from long-term use of ICS in COPD. A recent systematic review and meta-analysis found a pooled RR of 0.81 (95% CI 0.60–1.08) for mortality 10. The present study found that ICS use was associated with reduced mortality, and an additive benefit from concurrent statin use. Even though ICS do not seem to alter the rate of decline in lung function, there is strong evidence that they reduce the frequency of exacerbations, both of which are related to mortality in COPD 22, 23. A recent observational study by Huiart et al. 24 showed that use of ICS was associated with significant decrease in the risk of myocardial infarction among patients with COPD. Whether ICS can suppress systemic inflammation in COPD is not clear at present, but there are studies suggesting such an effect 25.

The overall mortality rate in the present study was comparable with the long-term mortality rate in previous studies of patients hospitalised with COPD exacerbation 26, 27. A physician specialised in internal medicine or pulmonary medicine verified the diagnoses. Thus, the study population was well defined. Mortality data were gathered from the Central National Register, which is based on a unique personal identification number for all Norwegian inhabitants. This ensured complete follow-up of the cohort, and makes misclassification of the outcome highly unlikely.

The main limitation of the present study is its nonrandomised design. Thus, statin treatment could have been restricted to patients with inherently better prognosis than patients who were not offered this treatment. Even though many important determinants of mortality were adjusted for in multivariate analyses, the possibility that important prognostic factors might have been imbalanced between the groups cannot be completely ruled out. Moreover, the results should be interpreted with caution due to the limited number of patients. Consequently, the present results should be considered to be hypothesis-generating and not definitive.

In conclusion, the present study shows that use of statins is associated with improved survival after chronic obstructive pulmonary disease exacerbation, regardless of whether or not the patients have a diagnosis of ischaemic heart disease. Concurrent medication with inhaled corticosteroids appears to further enhance this beneficial effect. If these findings can be confirmed in prospective randomised clinical trials, which are now warranted, it may lead to a paradigm shift in the treatment of this large patient population.