Impact of chronic obstructive pulmonary disease on morbidity and mortality after myocardial infarction

Study sample

Consecutive patients with MI admitted to Swedish coronary care units and entered in the nationwide Swedish Web-system for Enhancement and Development of Evidence-based care in Heart disease Evaluated According to Recommended Therapies (SWEDEHEART)23 registry between 2005 and 2010 were available for analyses. The study population consisted of a total of 81 191 patients with MI, including ST elevation myocardial infarction (STEMI) and non-STEMI. Of these patients, 4867 (6%) had a previous COPD hospital discharge diagnosis while 76 324 did not. The COPD diagnoses were based on International Classification of Diseases (ICD) codes that can be found in online supplementary table S1.

The SWEDEHEART registry enrols consecutive patients admitted to a coronary care unit because of symptoms suggestive of an acute coronary syndrome. On admission, patients receive written information about SWEDEHEART and other quality-of-care registries; patients are permitted to deny participation in the registry, although few of them exercise this right. According to Swedish law, written consent is not required because quality control is an inherent element of hospital healthcare. Research based on the registry is approved by an institutional ethics committee and all personal identifiers are removed from the SWEDEHEART data file when used for research purposes. Information is collected prospectively regarding baseline characteristics such as age and smoking status as well as ECG findings, examinations, interventions, in-hospital complications, discharge medication and diagnoses.23 Information on time of death was obtained from the Swedish National Cause of Death Registry. Information regarding medical history, including previous COPD diagnoses, and re-admissions for reinfarction, stroke or bleeding was obtained from the Swedish National Patient Registry,24 which includes diagnoses for all patients hospitalised in Sweden from 1987 and forward. Since 2001, the specialised outpatient care is also included. The validity of COPD diagnoses in the Swedish National Patient Registry has recently been reported to be good, with a diagnosis likely to be misclassified in less than 10%.25

Endpoints

The primary analysis tested the relationship between a prior COPD hospital discharge diagnosis and the primary endpoint of all-cause mortality during 1 year of follow-up after the initial coronary care unit hospitalisation. Secondary endpoints included 1-year re-admission for reinfarction, defined as a new hospitalisation with an MI diagnosis, new-onset admission for stroke, new-onset bleeding and new-onset heart failure. The corresponding ICD codes that the secondary endpoints are based on can be found in online supplementary table S1.

Statistical analyses

Rates of predefined endpoints in patients with and without a prior COPD hospital discharge diagnosis were calculated with the Kaplan-Meier estimator. Univariate and multivariate HRs were estimated using the Cox proportional hazards models. Covariates were tested for proportionality by visual inspection. Adjustments for potential confounders were performed stepwise in two models, the first including age, sex, smoking status and comorbidities (previous MI, previous stroke, heart failure, renal failure, hypertension, diabetes, peripheral artery disease, cancer and previous bleeding). The second model also included treatments during hospitalisation and at discharge (heparin, fondaparinux, dalteparin, enoxaparin, GPIIbIIIa-inhibitors, β-blockers, balloon angioplasty, coronary stenting, as well as discharge medications including ACE inhibitors, angiotensin II receptor blockers, aspirin, clopidogrel, prasugrel, β-blockers, calcium channel blockers, digoxin, diuretics, statins, nitrates and warfarin). The selection of covariates included in these models was performed with the use of a direct acyclical graph26 via a web-based tool (http://www.dagitty.net), as illustrated in online supplementary figure S1. Differences between continuous variables were evaluated using the Student t test. Differences between categorical variables were analysed with the Pearson χ² test. All tests were two-sided with a p value for significance <0.05. All analyses were performed in SPSS (SPSS V.20, IBM SPSS statistics).

Patient characteristics

Baseline characteristics for patients with MI and without COPD are outlined in table 1. Many variables differ between the two groups. The mean age was 5 years higher in patients with COPD as well as a threefold higher prevalence of prior heart failure. Furthermore, there was a more than twice as high proportion of renal failure, peripheral artery disease and cancer in patients with COPD. Patients with COPD were also more likely to have suffered from previous MI and stroke as well as being treated with more cardiovascular medications than the patients with non-COPD at baseline, with the exception of β-blockers.

Clinical presentation, laboratory findings and ECG changes

The pattern of symptoms differed between the two groups, as shown in table 2. Patients in the COPD group presented more frequently with dyspnoea and less frequently with chest pain as the main presenting symptom compared to patients without COPD. For the patients with COPD, the mean heart rate was higher, while the lab findings revealed lower mean total cholesterol and low-density lipoprotein levels with a higher mean CRP value. More often, the presenting ECG showed atrial fibrillation or flutter in the COPD group and the QRS complex revealed higher percentages of left bundle branch block and right bundle branch block. In contrast, ST elevation was more frequent in the non-COPD group.

Treatments, angiographic findings, complications during hospitalisation and discharge medications

The in-hospital characteristics for patients with MI with and without COPD are outlined in table 3. Invasive investigation and treatments in the form of coronary angiography, balloon angioplasty and stenting were less frequent among patients with COPD while the rate of coronary arterial bypass graft surgery did not differ. The indications for PCI differed between the groups, with STEMI being more prevalent among patients in the non-COPD group while the extent of coronary disease was similar.

Continuous positive airway pressure usage was more common for the COPD group as well as a bleeding requiring transfusion and/or surgery. Patients in the COPD group were also more likely to be discharged with atrial flutter or atrial fibrillation as well as with a lower left ventricular ejection fraction.

Patients with COPD were discharged with fewer medications that have been proven to reduce mortality such as aspirin and other platelet inhibitors as well as β-blockers, statins and ACE inhibitors but more of angiotensin receptor blockers. In contrast, patients in the COPD group were more often discharged with calcium channel blockers, digoxin, diuretics, nitrates and warfarin.

Outcomes

The crude 1-year mortality was significantly higher in the COPD group compared to the non-COPD group, 24.6% vs 13.8% (HR 1.86, 95% CI 1.76 to 1.98), as shown in figure 1 and table 4. After adjusting for differences in baseline characteristics, the mortality remained higher in the COPD group but the HR was significantly lowered (HR 1.32, 95% CI 1.24 to 1.40). After additional adjustment for treatments during hospitalisation and discharge medications, the difference in mortality was further decreased but remained statistically significant (HR 1.14, 95% CI 1.07 to 1.21).

The results of the secondary endpoint analyses are shown in table 4. Patients with COPD had a higher rate of reinfarction 16.6% vs 14.2% (HR 1.17, 95% CI 1.09 to 1.26), new-onset bleeding 4.1% vs 2.8% (HR 1.45, 95% CI 1.25 to 1.69) and new-onset heart failure 17.2% vs 9.7% (HR 1.84, 95% CI 1.70 to 1.99) compared to the non-COPD group in univariate analyses, while there was no difference in the rate of new-onset strokes. However, after adjusting for differences in baseline characteristics, treatment during hospitalisation and discharge medications, no differences in reinfarction rates or new-onset bleeding rates were noted. In contrast, the rate of new-onset heart failure remained higher for patients with COPD (HR 1.35, 95% CI 1.24 to 1.47).

Patient characteristics

In our study, 6% of the study population had a COPD hospital discharge diagnosis, lower than the estimated prevalence of COPD in the general population (9–10%), a finding in accordance with the previously reported problems of underdiagnosis.3 The increased age in the COPD group probably reflects that COPD is a late effect of lifelong smoking, but it could also be explained by underdiagnosis since COPD is relatively silent in early stages, and therefore the diagnosis does not surface until the manifestations are severe. As table 1 outlines, many of these patients also have previous cardiovascular events, in part due to a heavy smoking history but perhaps also due to reduced lung function and chronic inflammation of the lungs.

With regard to baseline characteristics and clinical presentation, several findings in our study are supported by previous studies. We found that patients with COPD had a larger burden of comorbidity and more atypical MI symptoms at presentation, in accordance with the findings of a previous study.20 However, we did not observe any differences in time delays from symptom onset to PCI or to the emergency room, as previously reported.20 Our data also did not support a previous study showing higher rates of cardiogenic shock in patients with COPD.27

Treatments

After an MI, many patients with COPD have previously not been prescribed β-blockers28 because clinicians fear that β-blockers will provoke bronchospasm and induce respiratory failure, even though cardioselective β-blockers have been proven to be safe and should not be routinely withheld from patients with COPD.29 Other types of standard post-MI treatment such as aspirin may also be used less often.22

As table 3 outlines, our findings are in line with these previous studies and show that standard post-MI treatment is withheld from patients with COPD more often than patients without COPD, especially with respect to β-blockers and surprisingly also statins which previously in observational studies have shown a dual cardiopulmonary protective effect.30 A systematic review including nine previous studies suggests that statins may also have a beneficial role in the treatment of COPD in itself.31

Outcomes

Previous studies have reported conflicting findings. Bursi et al21 reported a worse 5-year survival rate in patients with COPD (46%) compared to those without COPD (68%), and the association between COPD and death was independent of age and risk factors (HR 1.30, 95% CI 1.10 to 1.54). In another study by Salisbury et al22, patients with COPD had a twofold higher 1-year mortality rate after adjustment for baseline differences (HR 2.00, 95% CI 1.44 to 2.79) and higher rehospitalisation rates (HR 1.22, 95% CI 1.01 to 1.48). On the other hand, the older study by Behar et al32 did not find an independent association between COPD and a higher risk of early death or long-term mortality among survivors of acute MI.

Our study showed that patients with a prior COPD hospital discharge diagnosis had a considerably higher crude 1-year mortality after an MI (HR 1.86, 95% CI 1.76 to 1.98) compared to patients without COPD with an MI. However, we could show that this association was greatly lowered after adjusting for baseline characteristics and comorbidities (HR 1.32, 95% CI 1.24 to 1.40) and, perhaps most importantly, after also adjusting for different treatment patterns, only a modest increase in adjusted mortality remained (HR 1.14, 95% CI 1.07 to 1.21).

Therefore, our results indicate that patients with COPD with an MI constitute a very high risk group with a nearly doubled unadjusted mortality rate compared to patients without COPD with an MI and that the excess mortality could perhaps be lowered with more aggressive evidence-based treatments for both the MI as well as concomitant diseases. However, our results are only suggestive and it would require a prospective, interventional study to confirm our findings. COPD was not independently associated with a higher 1-year re-admission for reinfarction, new-onset stroke or new-onset bleeding rate but was independently associated with an increased new-onset heart failure rate.

The observed association between a prior COPD diagnosis before MI and a higher frequency of subsequent new onset of heart failure even after multivariate adjustment raises several questions. Not much is known about the association of heart failure and COPD. Previous authors have suggested a common inflammatory background between the conditions.33 The actual prevalence of decreased left ventricular function in patients with COPD is largely unknown and clinically poorly defined.33 ,34 Shared signs, symptoms and pulmonary function test findings between heart failure and COPD further complicate the relationship. Patients with COPD may suffer from pulmonary hypertension,35 which could lead to right ventricular dysfunction,36 and because of a similar symptomatology between cor pulmonale and true left ventricular failure, it is hard to discern and distinguish the exact aetiology of the heart failure diagnosis. Dyspnoea and exercise intolerance are cardinal symptoms for COPD and heart failure resulting in diagnostic difficulties, and misclassification in the National Patient Registry cannot be ruled out. However, in the present study, patients with COPD did have a higher rate of decreased left ventricular ejection fraction at discharge and they were also undertreated post-MI with guideline recommended secondary prevention medications. This could lead to a higher frequency of new-onset heart failure.

Limitations

Our study design was observational, and thus a certain degree of residual confounding cannot be ruled out. Since COPD is an underdiagnosed disease,4 ,5 a number of patients in the non-COPD group could have met the criteria for COPD if they had been thoroughly investigated with spirometry, an inherent limitation of any retrospective COPD study.37 The underdiagnosis of COPD could potentially result in an under-representation in the registry that could underestimate our findings. Furthermore, we did not have optimal data regarding the patients’ smoking history, as pack-years, date of smoking cessation and information about smoking post-MI were lacking. We also did not have information on pulmonary function testing, such as FEV1/forced vital capacity ratios, and therefore the severity of COPD diagnoses cannot be evaluated in our patient population. Moreover, a wide range of physicians has diagnosed the COPD cases, and thus the criteria for COPD may differ between patients in the population. The same problem is applicable to the heart failure diagnoses. However, the validity of COPD and heart failure diagnoses in national Swedish registers has recently been reported to be good.25 ,38