Trends, characteristics, in-hospital outcomes and mortality in surgical mitral valve replacement among patients with and without COPD in Spain (2001-2015)

Purpose We examined trends, characteristics and in-hospital outcomes in mechanical and bioprosthetic surgical mitral valve replacement (SMVR) among patients with and without chronic obstructive pulmonary disease (COPD) in Spain from 2001 to 2015. We also identified factors associated with in-hospital mortality (IHM) in both groups of patients according to the implanted valve type. Methods We analyzed data from the Spanish National Hospital Discharge Database for patients aged 40 years or over. We selected admissions of patients whose medical procedures included SMVR. We grouped hospitalizations by COPD status. Results Over 43,024 patients identified, 83.63% underwent mechanical mitral valve replacement and 16.37% bioprosthetic valve (6.71% and 7.78% with COPD, respectively). The incidence of SMVR decreased for mechanical valves and increased for bioprosthetic valves over time in both groups of patients. The incidence of SMVR admissions was lower among COPD patients than in those without COPD, both for mechanical and bioprosthetic valves. IHM decreased significantly over time, regardless of the type of valve, in both groups of patients. COPD was associated with a significant increase in IHM, but only among patients who underwent bioprosthetic SMVR (OR 1.32, 95% CI 1.01–1.73). Conclusions The incidence of mechanical SMVR decreased while that of bioprosthetic SMVR increased over time in both groups of patients. COPD patients were less surgically operated than non-COPD patients for both valve types. In COPD patients, bioprosthetic SMVR was proportionally more used than mechanical SMVR. Mortality decreased over time for both valve types in patients with and without COPD. COPD increased in-hospital mortality among patients undergoing a biological SMVR.


Introduction
Mitral valve disorder is one of the most common heart valve diseases in the developed countries [1]. About 2% of the general population has significant mitral valve disease [2]. Incidence increases in the elderly, affecting nearly 10% of people above the age of 75 years [3]. Surgery remains the gold standard treatment [4]. Valve replacement prolongs survival among patients with severe involvement [5]. Valve prostheses can be mechanical or biologic, and each one has risks and benefits. Bioprosthetic valves are subjected to structural degeneration and may require future reoperation, while mechanical valves typically require lifelong anticoagulation with associated risks of thromboembolic and bleeding [5,6]. Despite it, the use of biologic prostheses has increased substantially for surgical mitral valve replacement (SMVR) in recent years [5,7].
Among the factors associated with poor SMVR outcomes are advanced valvular cardiomyopathy, age, and associated comorbidities [6]. In fact, a substantial part of symptomatic patients with severe mitral valve disease are excluded from surgery by institutional heart teams for these reasons [8,9]. Chronic obstructive pulmonary disease (COPD) is one of the comorbidities that has been associated with the decision to not operate [9]. However, in a recent study, Mkalaluh et al [10] analyzed a contemporary series of octogenarians undergoing mitral valve surgery at their institution and found that COPD did not influence 30-day mortality.
Understanding the changing characteristics and the prognosis of these patients is important to interpret previous studies and design future trials [11]. However, nationwide data on trends in SMVR are scarce [12]. The aim of the present study was to examine nationwide trends, characteristics and in-hospital outcomes in mechanical and bioprosthetic SMVR among patients with or without COPD from 2001 to 2015 using the Spanish National Hospital Discharge Database (SNHDD). Also we identified factors associated with in-hospital mortality (IHM) among patients with and without COPD according to the implanted valve type.

Data source
The SNHDD covers over 95% of hospital admissions in Spain and contains nationwide information of up to 14 discharge diagnoses and up to 20 procedures performed during the hospital stay including admission and discharge dates [13]. International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) is used for coding in the SNHDD. For this retrospective study, we used data from years 2001 through 2015.

Study population
We selected admissions of patients (aged �40 years) whose medical procedures included mitral valve replacement (ICD-9-CM codes: 35.23 and 35.24) in the SNHDD database.

Covariates
Clinical characteristics included information on overall comorbidity at the time of discharge, which was assessed by calculating the Charlson comorbidity index (CCI) [14]. Calculation of the CCI was done excluding COPD as defined before as a disease.
Among the procedures we specifically analysed if Coronary Artery Bypass Graft (CABG), Surgical Aortic Valve Replacement (SAVR), surgical procedures on pulmonary and/or tricuspid valves, intra-aortic balloon counter-pulsation, pacemaker implantation and blood transfusions were done during the hospitalization for SMVR.
ICD-9-CM codes for conditions included in the CCI as well as other diseases and procedures performed during the hospital stay and analyzed in this investigation are shown in Table 1.
We evaluated the mean of length of hospital stay (LOHS) and we estimated the incidence of peri-operative Major Adverse Cardiovascular and Cerebrovascular Events (MACCE) as Table 1. Diagnosis and procedures with corresponding ICD-9-CM codes. described by Newman et al [15], MACCE includes acute myocardial infarction, ischemic stroke or death during the hospitalization.

End points
The main end points in our investigation were trends, characteristics, in-hospital outcomes, and IHM in SMVR among patients with and without COPD. IHM was defined by the proportion of patients who died during admission for each year of study.

Statistical analysis
The analyses were stratified according to the type valve use (mechanical and bioprosthetic) in the valve replacement and for each group of patients (COPD and non-COPD patients). We considered three time-periods that included five consecutive years each (2001-05; 2006-10; 2011-15).
In order to assess time trends we estimated the incidence rates of admission for SMVR in COPD patients and non-COPD patients per 100,000 inhabitants applying the methods described in previous studies and included Poisson regression analysis adjusted by age and sex [16].
A descriptive statistical analysis was performed for all continuous variables and categories. Variables are expressed as proportions as means with standard deviations. A bivariable analysis according to year was performed using the χ 2 test for linear trend and ANOVA, as appropriate.
To To identify predictors of IHM we performed multivariable logistic regression analyses, one for each type of mitral valve replacement (mechanical and bioprosthetic) and for each group of patients (COPD patients, non-COPD patients and both). The variables included in the multivariable models were those with significant results in the bivariable analysis and those considered relevant in other investigations. Estimates were Odds Ratios (ORs) with their 95%CI.

Sensitivity analysis
To avoid the confounding effect of clinical characteristics on our main study outcome variables (MACCE and IHM) we conducted a sensitivity analysis. To do so we used a propensity score matching (PSM) to select a matched non-COPD patient for each COPD patient. PSM was done using multivariable logistic regression for mechanical and bioprosthetic SMVR patients separately. The variables included to estimate the PSM were year of surgery, sociodemographic and clinical characteristics shown in Table 1.
For data management we used Access. All statistical analyses were performed with Stata version 10.1 (Stata, College Station, Texas, USA). Statistical significance was set at p<0.05 (2-tailed).

Ethical aspects
The study maintains data confidentiality at all times. Given the anonymous and mandatory nature of the database, it was not necessary to obtain informed consent or approval by an ethics committee in accordance with Spanish legislation.

Results
In our study we identified a total of 43,024 hospitalizations of patients aged 40 years or more who underwent SMVR in Spain (2001-2015) of them 2,965 suffered COPD (6.89%).

Trends in mechanical mitral valve replacement hospitalizations
The proportion of patients receiving a mechanical valve decreased from 90.22% in 2001-05 to 77.61% in 2011-15 (p<0.001). Equivalent figures for COPD patients were 88.59% to 74.18% and for non-COPD patients from 90.09% to 77.86% (p<0.001, in both cases).
The incidence of mechanical mitral valve replacement among COPD patients decreased from 11.95 cases per 100,000 individuals in the COPD population in 2001 to 11.56 in 2015. Among, non-COPD patients, the incidence also decreased significantly from 12.77 in 2001 to 9.87 in 2015 (p<0.001) (S1 Fig).
The Poisson regression models yielded an adjusted IRR for patients with COPD who underwent mechanical valve replacement of 0.76 (95%CI 0.73-0.80). Therefore, compared to patients without COPD the incidence of mechanical mitral valve replacement hospitalizations was lower among COPD patients.
In patients who underwent mechanical mitral valve replacement, there was a significant male predominance in COPD patients (67.4%), however in patients without COPD there was female predominance (63.85%). Overall, patients with COPD were older (66.5; SD = 8.43 years) than patients without COPD (64.49; SD = 9.43 years; p<0.05) and were more likely to undergo simultaneous CABG (16.26% vs. 10.78%). However, patients without COPD were more likely to undergo other valves procedures on pulmonary or tricuspid valves (25.97% vs. 20.94%) ( Table 2) We found a significant increase in the frequency of SAVR, other valves procedures on pulmonary or tricuspid valves, and pacemaker implantation in patients with and without COPD over time (Table 2).
Mean LOHS for patients with and without COPD undergoing mechanical mitral valve replacement was 24.87 and 22.88 days, respectively, in the period from 2001-05, decreasing to 19.79 and 20.42 days, respectively, in 2013-15 (all p<0.001). The overall mean LOHS was significantly higher in patients with COPD (23.05 vs. 22.02 days).
In COPD patients, IHM and MACCE after mechanical mitral valve replacement decreased significantly (all p<0.001) over time from 14.06% and 18.59%, respectively, in 2001-05 to 9.02% and 9.64% in 2011-15. Among patients without COPD, also IHM and MACCE decreased significantly over time. No differences were found between the two groups of patients ( Table 2).
We detected a significant increase in comorbidity according to the mean CCI over time in both patients with and without COPD (p<0.001). The frequency of type 2 diabetes mellitus, peripheral vascular disease, acute renal disease, congestive heart failure, coronary artery disease, obesity, renal disease and liver disease also increased significantly in both groups over time. In non-COPD patients we found a significant increase in frequency of cerebrovascular disease (from 4.5% in 2001-05 to 5.45% in 2011-15), atrial fibrillation (from 57.18% to 59.33%), pulmonary hypertension (from 26.94% to 28.34%), cardiogenic shock (from 3.53% to 4.33%), endocarditis (from 7.19% to 13.04) and cancer (from 0.66% to 1.07%) ( Table 3).

Trends in bioprosthetic mitral valve replacement hospitalizations
Bioprosthetic valves represented 9.78% of all valve replacements in period 2001-05 and increased to 22.39% in 2011-15 (p<0.001). Significant increments were observed for COPD Compared to patients without COPD the incidence of bioprosthetic mitral valve replacement hospitalizations was lower among COPD patients (IRR 0.89;95%CI 0.82-0.98). In patients who underwent bioprosthetic mitral valve replacement, there was a significant male predominance (76.82% for COPD and 38.27% for non-COPD patients). In non-COPD patients mean age increased significantly (p<0.001) from 72.12 years in 2001-05 to 74.42 years in 2011-15. We found a significant increase in the frequency of SAVR, other valves procedures on pulmonary or tricuspid valves and blood transfusion in patients without COPD patients over time. No differences were found among COPD patients. (Table 4).
In patients with and without COPD, mean LOHS decreased significantly over time (30.95 and 26. 56 days, respectively, in the first period vs. 24.65 and 22.73 days in the last one) ( Table 4).
For the total time period, crude IHM was significantly higher in COPD patients than in non-COPD patients (17.88% vs. 14.74%; p = 0.048). In non-COPD patients, IHM decreased over time, however no significant changes were found in COPD patients. MACCE was 21.74% for COPD patients and 19.19% for non-COPD individuals. MACCE decreased significantly in both groups over time (Table 4).
Atrial fibrillation (55.47%), coronary artery disease (34.12%) and congestive heart failure (30.66%) were the most common associated comorbidities for COPD patients who underwent  In patients with and without COPD the prevalence of obesity and renal disease increased significantly over time.
In non-COPD patients, the prevalence of type 2 diabetes mellitus, acute renal disease, congestive heart failure, atrial fibrillation, pulmonary hypertension and endocarditis also increased significantly over time. However the prevalence of gastrointestinal hemorrhage decreased from 1.28% in 2001-05 to 0.39% in 2011-15 (p = 0.002) in patients without COPD. No differences were found among COPD patients (Table 5).

Factors associated with IHM
The factors independently associated with IHM according to the type of mitral valve are shown in Tables 6 and 7.
Regardless of the type of valve implanted and for both patients with and without COPD, factors that increased IHM included older age, intra-aortic balloon counterpulsation during admission, acute renal disease, cardiogenic shock, endocarditis and liver disease.
Blood transfusion during admission, congestive heart failure and pneumonia increase the risk of IHM in patients with and without COPD with mechanical mitral valve replacement. Among COPD patients, IHM was significantly higher in those who underwent SAVR (OR 1. 16 Among COPD patients who underwent bioprosthetic valve, IHM was significantly higher in female sex, those who underwent SAVR, with blood transfusion during admission, with peripheral vascular disease, with cerebrovascular disease, with congestive heart failure, with coronary artery disease, with pneumonia and with renal disease. Atrial fibrillation was associated with a lower risk of dying for both COPD patients and non-COPD patients who underwent mechanical valve replacement (OR 0.49;95%CI 0.46-0.54 and OR 0.75;95%CI 0.56-0.99, respectively). Among COPD patients underwent mechanical valve replacement pacemaker implantation was associated with lower IHM (OR 0.57;95%CI 0.46-0.71).
Pacemaker implantation was associated with a lower risk of IHM in patients with and without COPD with bioprosthetic mitral valve replacement (OR 0.46; 95%CI 0.31-0.69 and OR 0.22;95%CI 0.05-0.98, respectively). Furthermore, atrial fibrillation was also associated with a lower risk of dying in patients with COPD (OR 0.59; 95%CI 0.5-0.7).
Over time, the IHM decreased significantly regardless of the type of valve. After multivariable adjustment, COPD was associated with a significantly higher IHM (OR 1.32;95%CI 1.01-1.73) among patients who underwent bioprosthetic mitral valve replacement in our study. However, COPD did not predict IHM among those patients who underwent mechanical mitral valve replacement (OR 0.99; 95%CI 0.86-1.16).
Finally, the results of the sensitivity analysis using PSM (S1 Table) confirmed the results of the multivariable logistic regression with significantly higher IHM among COPD patients who received a bioprosthetic valve than among non-COPD patients.

Discussion
In this large analysis of patients undergoing SMVR, we demonstrated that COPD was associated with a significantly higher IHM among patients who received a bioprosthetic valve, but this disease did not predict IHM among those receiving a mechanical valve. These results contrast with those obtained by other authors who have not found that COPD is a predictor of mortality in patients undergoing SMVR with biological or mechanical substitutes [17]. Goldstone et al [5] also found that mortality was lower among patients up to 70 years of age who received a mechanical mitral valve than among those who received a biologic prosthesis, although they did not stratify the results by the existence or not of COPD.
Compared to patients without COPD, the incidence of mitral valve replacement hospitalizations was lower among COPD patients in our study, regardless the type of valve used. The presence of comorbidities, as attested by a higher CCI, has been associated with a more frequent decision not to operate in patients with severe symptomatic mitral regurgitation [9]. This may because an increased CCI has a negative impact on survival [14] and most comorbidities may also increase operative risk [18], including chronic airway disease [19].
Despite the limited durability of biological valves, increasing numbers of younger patients are choosing to receive them, due mainly to the lack of a need for permanent anticoagulation [20]. In this way, we observed an increase of the incidence of bioprosthetic mitral valve replacement among COPD and non-COPD patients from 2001 through 2015. By contrast, the incidence of mechanical mitral valve replacement decreased in both groups of patients during the study period. Goldstone et al [5] also reported that the use of biologic prostheses increased substantially for mitral-valve replacement from 1996 through 2013 in California. Justification for this approach may be related to the improved durability of the current generation of tissue valves and the technical feasibility of a subsequent valve-to-valve procedure, in the event that occurs a degeneration of the tissue valve, to avoid having to resort an open-heart surgery [7]. Although this approach may be reasonable in high-risk and elderly patients, as often it occurs in COPD patients, it is not supported by scientific evidence in younger patients, especially in those without comorbidities or contraindications for anticoagulation [7]. In any case, the choice of replacement mitral valve type is a decision that should be taken by physician and patient in conjunction and must be individualized. For this, it should be accounting the risks of reoperation and of chronic anticoagulation and their consequences, in addition to characteristics of the patient, lifestyles, comorbidities and life expectancy, in order to increase quality of life and survival [17].
In our study, COPD patients underwent CABG more frequently than those without COPD, both for mechanical and bioprosthetic valves. However, they underwent other valves procedures on pulmonary or tricuspid valves less frequently than non-COPD patients. In this way, Sá et al [21] found that mitral valve replacement may be performed safely, concomitantly with CABG, in patients with moderate-to-severe ischemic mitral regurgitation. In such patients, the combined procedure resulted in lower rates of postoperative atrial fibrillation and low cardiac output than CABG only. In addition, although both surgical approaches resulted in significant improvement in postoperative left ventricular ejection fraction, there was greater improvement in the combined surgery group, Thereby, despite being a more aggressive approach, the combined surgical procedure did not increase morbidity or mortality. IHM following mitral valve replacement is high [12]. We found a significant decrease in IHM over time among COPD and non-COPD patients who underwent a mechanical or bioprosthetic mitral valve replacement. Berzingi et al [12] also demonstrated a significant decrease in IHM between 2003-2014 for isolated bioprosthetic mitral valve replacement. On the other hand, as it has been described in previous studies [22], being older and having additional concomitant procedures, were associated with a higher mortality risk in ours. Furthermore, several comorbidities were found to be associated with increased mortality risk in our multivariable analysis. These data might be useful when interpreting the emerging literature of transcatheter mitral valve replacement [12].

Limitations
The findings of this study must be viewed in light of limitations inherent to retrospective observational studies that rely on administrative data. We used the SNHDD, an administrative database rather than a clinical database. Therefore, it was subject to coding error. Our definition of COPD was not based on lung function test because SNHDD does not include the results of diagnostic tests. Therefore, clinical measures of severity and type of COPD were not available. We also did not have data on COPD treatment duration and medication dosing requirements, (e.g., dose and duration of steroid use). Implied in this information deficit is the inability to determine whether COPD symptoms were caused by chronic heart (CHF) failure and distinguishing symptoms caused by CHF but attributed to COPD. However, we have followed similar designs of other descriptive studies on COPD hospitalizations [23][24][25][26][27][28]. Additionally, in several studies it has been assessed the validity of healthcare databases for COPD [29][30][31][32][33][34]. In this way, administrative databases seem beneficial for planning health care interventions, including the COPD field. They constitute robust information systems subjected to periodic data quality controls [13] remaining the prevalent data source, reliable because of the amount of data and the population coverage, especially in countries with a National Health System [35], like ours. Thereby, this dataset reflects the outcome of the world-real population, which is different from that selected in randomized controlled trials. Thus, the results could be extrapolated to the general population.

Conclusions
In conclusion, the incidence of mechanical SMVR decreased while that of bioprosthetic SMVR increased in COPD and non-COPD patients from 2001 to 2015 in Spain. Patients with COPD were less surgically operated than those without COPD for both valve types. In COPD patients, bioprosthetic SMVR was proportionally more used than mechanical SMVR. Mortality decreased over time for both valve types in patients with and without COPD. COPD was associated with IHM in patients undergoing a biological SMVR, but not in those that underwent a mechanical SMVR.