Abstract
Objectives: to understand epidemiological trends in severe COPD exacerbations through analyzes of hospitalizations and deaths during three consecutive years in a French administrative region area. Methods: Medico-administrative records of hospitalizations for COPD exacerbations were sorted from 2010 to 2012 using selected International Classification of Diseases (ICD10) codes. Four groups of hospitalization for COPD severe exacerbations were elicited leading to hospitalizations (general ward without respiratory failure, general ward with acute respiratory distress, ICU without mechanical ventilation, ICU with mechanical ventilation). Results: Data extraction identified 5007, 4986 and 5359 admissions related to 4136, 4155 and 4460 patients in 2010, 2011 and 2012, respectively. Marked seasonal variations were observed. Duration of stay (median (IQR), 7 (7) vs 9 (8) vs 10 (9) vs 14 (16) days, P < .001), death rates (3.6% vs 14.2% vs 14.4% vs 21.2%, P < .01), number of co-morbid conditions (median (IQR), 2 (2) vs 2 (2) vs 4 (5) vs 4 (4.5), P < .01), type of institution (64.9% in public institution vs 79.9% vs 87.8% vs 76.6%, P < .01) were significantly associated with the hospitalization group and more than 8% of admissions led to death (3% to 24%). Age, type of institution and past hospitalizations were independent risk factors for deaths. Readmissions were infrequent but mainly related to the worsening of the co-morbid conditions. Conclusion: COPD severe exacerbations are frequent and lead to an important numbers of deaths related to the severity of acute respiratory failure and the number of co-morbid conditions.
Introduction
Chronic Obstructive Pulmonary Disease (COPD) is a devastating disease characterized by an accelerated decline of lung function, dyspnea inducing handicap and quality of life impairment, recurrent exacerbations and frequent co-morbid conditions (Citation1). The disease burden is still increasing and COPD will be the third cause of death in the world by 2030. Some patients that exhibit repeated exacerbations are associated with poorer prognosis (Citation2, Citation3). Severe exacerbations lead to emergency room visits and frequent hospitalizations, which represent an important cost for the health care system and the society. A better knowledge of the characteristics of severe exacerbation hospitalizations and their risk factors for poor outcome should contribute to develop better health care programs and improve care for the most susceptible patients.
Hospital admissions are good indicators of the severity of exacerbations and represent an important outcome for developing innovative treatments and a valuable target to decrease the costs of the disease (Citation4, Citation5).
The impact of co-morbid conditions on morbidity and mortality through hospitalizations has been characterised as a major source of severity, cost and deaths in all non communicable diseases including COPD (Citation6, Citation7). A multidisciplinary approach is thus required to improve care and prevent fatalities and future risks.
The “Programme de Médicalisation des Systèmes d'Information” (PMSI) coding (the French equivalent of the UK Caspe Healthcare Knowledge Systems (CHKS)) is a wide and exhaustive medico-administrative database containing information on diagnosis, co-morbid conditions and in-hospital outcome of patients hospitalized in the public and private sectors in France. The PMSI is a law commitment database coding performed by the senior clinicians at the end of every stay which provides 100% of inpatient coverage in France, especially since 2007 (Citation8). In the present study, we aimed to describe COPD-related hospitalizations and risk factors for in-hospital deaths in a large French administrative region during the period 2010–2012. We designed a database sorting rule to identify and compare four different levels of hospitalization severity for exacerbation. Our ultimate purpose is to provide health care decision-makers with information that may help to develop efficient health policies.
Materials and methods
Study design and population
This was a retrospective observational database study using exhaustive standardized data from the regional PMSI, collected from 2007 to 2012. This data set contains medico-economic information about admissions based on the International Classification of Diseases (ICD-10) classification from both public and private hospitals in Languedoc-Roussillon (France). It includes all data collected by healthcare providers to the social welfare office. The study focuses on hospital admissions for severe COPD exacerbations from 2010 to 2012 (end of stay between January 1, 2010 and January 1, 2013), and yearly numbers of admissions since 2007, which were used to highlight the trends in COPD hospitalizations. Past admissions history was tracked with a unique identification number (called ‘MAGIC’ number) that allowed chaining from 2010 to 2012. Hospitalization was defined by at least one night stay in a public or private institution. All patients with a permanent address in the region were included.
We reviewed anonymous data for each patient including the chronological order of admissions to various medical units during the hospitalization (e.g., intensive care unit (ICU), high dependency unit (HDU), monitoring, regular ward, etc.). Patients with respiratory diseases other than COPD (e.g., including asthma or diffuse bronchiectasis) were excluded. Hospitalizations were categorized into four mutually exclusive levels of hospitalization severity (mild, moderate, severe and very severe). Note that our categorisation is only based on structural hospitalization severity and do not represent a COPD severity classification. Categories definitions are depicted in Supplemental File 1 and were based on ICU requirement, mechanical ventilation support, and clinical signs of acute respiratory failure (according to PMSI coding). The above-mentioned data were described in the whole study population and compared to different levels of hospitalization categories according to age, gender, duration of stay, vital status at discharge (allowing to derive in-hospital mortality), month of discharge, the primary diagnosis (reason for referral), secondary diagnoses (co-morbid conditions), patient home address (to compute distance and transportation time to medical facilities), past hospitalizations either related or unrelated to COPD and care centre characteristics (geographic area, private or public institution).
Ethics and consent
According to French Law (Law 88-1138 relative to Biomedical Research of December 20, 1988, modified on August 9, 2004), this anonymous retrospective observational database study did not require approval by an ethics committee or informed signed consent from patients. The dataset presented in this study was declared to the ATIH (Agence Technique de l'Information sur l'Hospitalization) and approved by the CNIL (national commission for data processing and civil liberties).
Statistical analysis
Quantitative variables were expressed as mean ± standard deviation or median (IQR interquartile range), according to the normality of the distribution, assessed with the Shapiro–Wilk test and compared using the parametric and nonparametric ANOVA. Qualitative variables were expressed as numbers (percentage) and compared using the chi-square test. Distances and transportation times to medical facilities were obtained using a specific application computing the distance in kilometres and in minutes between the residency address and the institution of admission. Cox regressions were performed to assess risk factors of in-hospital death. Variables were selected if P-value was <0.20 in the univariate analysis, and a stepwise backward procedure was used to select the final model.
Statistical significance was considered at P < 0.05. The statistical analysis was performed by the medical statistical department of the Montpellier University Hospital with the help of statistical software (sas, version 9.3; SAS Institute; Cary, NC, USA; and R, version 2.15.2).
Results
First, 5007 hospital inpatient stays of 4136 different patients have been recorded for 2010 in our region (2,636,350 inhabitants in 2010). It reached 0.552% of total admissions. The general trend showed a nearly constant number corresponding to a relative stable incidence (from 0.495% to 0.538% of total admissions) of COPD hospitalizations between 2007 and 2012 (Table ). Demographic characteristics during the three years of interest are presented in Table . In 2012, 3207, 1306, 312 and 534 admissions entered the different subsets. The mean age of admissions was 75.3 ± 12.5 years, among them, 63.8% were males and 71.1% were hospitalized in a public institution with a length of stay of 8 (Citation8) days, 28.1% with oxygen supplementation. The mean number of co-morbid conditions per patient was 2 (Citation3); 8.6% of stays led to death. Generally, the results were consistent during the three years but significantly different depending on the categories of hospitalization except for gender. More severe patients were younger (P < 0.01). The percentage of hospitalizations in a public institution was higher for moderate, severe or very severe than in mild patients (P < 0.01). Mortality, length of stay and the number of co-morbid conditions increased in connection with the category of hospitalization (P < 0.01).
Table 1. Trends in COPD admissions in the Languedoc-Roussillon region between 2007 and 2012
Table 2. Main characteristics of hospitalizations of COPD
Among patients admitted in 2011 or 2012 for COPD, 52% had been admitted at least once in the past 12 months (52%, 52%, 49% and 53% for mild, moderate, severe and very severe pateints respectively, P = 0.58). The annualized number of previous admissions for COPD was 0.24 (0.67). The number of previous admissions for other reasons than COPD was 1.14 (4.53). Details of co-morbid conditions were assessed in each category and the results are presented in Supplemental File 2. A higher prevalence of co-morbid conditions was generally observed in categories of highest hospitalization severity.
Seasonality of admissions and deaths are presented in Figure . Data distributions were similar from 2010 through 2012. A non-uniform distribution of annual admissions (P < 0.01) and deaths (P < 0.01) was observed with an increase in incidence during fall-winter season. This non uniform distribution was also observed for admissions in each category of hospitalization (P < 0.01, P = 0.02, P = 0.03 and P < 0.01, for mild, moderate, severe and very severe respectively).
Figure 1. Seasonality of admissions and in-hospital deaths from 2010 to 2012.
Variables associated with in-hospital deaths are presented in and . The multivariate model built to identify factors predicting death was dominated by the categories of hospitalization and death was also independately associated to age, number of past admissions and type of institution. Co-morbid conditions independently associated with an increased risk of death within each category are summarized in Supplemental File 3. Sepsis and lung cancer dominated the prognosis.
Table 3a. Multivariate model of factors predicting in-hospital death among the whole population of COPD admitted because of an exacerbation
Table 3b. Co-morbid conditions independently associated with an increased risk of in-hospital death
Patients were living at a distance of 43.1 (72.7) kms from the institutions. No differences were observed due to distances to medical facilities between categories (with a mean of 41.1 kms to 43.9 kms according to the category). The transportation time was 36.6 (42.6) minutes and was similar between the categories (with a mean of 34.4 minutes to 36.8 minutes according to the category).
Discussion
In this study, we report an overall in-hospital mortality rate overtaking 8% which followed seasonal variations and reached 24% in ICU. Categories of hospitalization severity, mainly based on severty of the signs of acute respiratory failure and the structure of admissions, were the highest predictors of death, in relation with age, number of previous admissions and type of institution.
We observed a high prevalence approaching 1% of total admissions in our region. This might still be underestimated despite recent efforts to reduce COPD under-diagnosis (Citation9). Smoking ban (Citation10) and other behavioral interventions are leading to a small decrease in COPD admission rates in some western countries (Citation11). The method we chose to define our groups demonstrated some robustness as we carefully removed all potential confounders (notably asthma and bronchiectasis and overlap syndrome of bronchial diseases) with a lowest sensitivity explaining potentially our relative small incidences.
Mortality related to COPD reached levels as high as 24% of patients admitted in ICU and 3% in conventional medical wards, even in the absence of any clinical sign of respiratory failure, with a mean rate of 8%. Mortality rates at day-30 (Citation12) or later (Citation13) could be more relevant to properly estimate the burden of COPD. The mortality rates we observed are in the range of previous reports in similar western countries (Citation14, Citation15).
We observed that admission rates and mortality followed a seasonal distribution whatever the category of hospitalization. Seasonality is a known factor affecting exacerbation incidence in chronic airway diseases such as asthma (Citation16) or COPD (Citation17–Citation19). COPD exacerbations are known to be frequently related to infections mainly viral (Citation20–23), thus following seasonal variations in viral loads in the atmosphere and carriers. This factor was still present within the categories and even in the most severe patients indicating that infections are factors of exacerbations leading to hospitalizations in COPD. Nonetheless, season was not an independent risk factor for death. This suggests that viral induced-COPD exacerbations are more frequent in winter but not necessarily more severe.
Moderate patients were admitted in the ward but with respiratory failure according to PMSI coding. Moderate patients were slightly older and had been previously admitted for COPD more often. We tried to better understand who these patients were since their mortality rate (nearly 14%) was similar to that of severe (i.e., admitted in ICU but who did not require mechanical ventilation). The most frequent recorded co-morbid conditions were dementia, cognitive impairment and bedridden. Malignancies were not over represented. The rates of dependence on respirator and/or oxygen supplementation were similar to those recorded in severe patients.
Independent risk factors for death were sepsis and lung cancer in the whole population. Conversely, for moderate patients, bedridden and lung cancer were independent risk factors for deaths that could in part explain why these patients were not admitted in ICU. It could be hypothesized that these patients were confined to ward because of medical decision of supportive palliative care, i.e., treatment limitations. However, this was not the case as we track this decision as it is coded in the PMSI system, especially because palliative care is over-refunded to the hospital by the payer. This fact suggests that the patients may have benefited from ICU, supporting the need for redefining health resources in the different types of institutions criteria for ICU admissions.
The risk factors for death were age, number of previous admissions and type of institution (public versus private). The total number of co-morbid conditions was not associated with the risk of death, but this has to be interpreted with great caution since it may be due to undereporting and/or underdiagnosis of co-morbidities. In addition, all co-morbidities do not have the same weight as determinants of survival, as shown by Divo et al. (Citation24). These authors identified 12 co-morbidites associated to death in COPD patients and proposed a specific co-morbidities index (COTE) to assess mortality risk in this population. Paradoxically, in our study obesity, diabetes mellitus, sleep apnea and current smoking were associated with less mortality. This must also be interpreted with caution since it may be due to underreporting of these co-morbidities specifically in the most severe patients, in whom their financial impact is limited, if any.
Readmissions were frequent in our study and frequently associated to other reasons than COPD exacerbations. The number of previous admissions is a minor risk factor for death with an OR of 1.01 (CI 95% 1.01–1.02). This contrasts with other reports that show a greater importance of readmissions as risk factors for death in COPD and the progressive shortening in the delay between episodes of COPD exacerbations (Citation25). We have analyzed readmissions in the present study but only during 12 months before the index of hospitalization, which was a real limitation of our research to address this question properly.
Many authors regard co-morbid conditions as an important component of the expression of a systemic component of COPD (Citation26). COPD has also been considered as a part of the metabolic syndrome (Citation27). Accordingly, there is an increased attention paid to a rigorous and intense screening for a better management of the co-morbid conditions. In our database, co-morbid condition reporting is strongly encouraged as it improves hospital payment. However, the prevalences of co-morbid conditions are lower than those reported in most studies. In our study, because they do not lead to any subsequent increase in reimbursement, some co-morbid conditions were probably underreported in PMSI, which underlies the limitations of using this administrative database.
Patients admitted in ICUs who had more co-morbid conditions, were predominantly males, median age 71 years, and stayed about 8 days. They had rarely been admitted before because of COPD exacerbations, and the profile of co-morbid condition in these patients is highly consistent with phenotypic characteristics known to be related with a poor prognosis: they received oxygen or ventilation support at home, frequently presented malnutrition, lung cancer, and cardiovascular disease. The diagnosis of lung cancer at admission or during ICU stay with MV requirement is the most important risk factor for death in this group; about 40% of patients with lung cancer are diagnosed following an emergency admission to hospital (Citation28). Poorer access to care could in part explain why some patients seek attention later, so that lung cancer is discovered only when the patient's condition becomes severe. The distance to medical facility (which we considered as a surrogate marker of some aspects of access to care) was not a predictor for death. Thus, other patients characteristics, which could not be captured in this database study, may be involved in late cancer diagnosis.
We recorded whether patients were admitted in public or private institutions. In fact, private institutions were more frequently managing patients in the less severe categories probably because fewer beds are dedicated to ICU. In France, private institutions are traditionally orientated toward surgery and “lighter” cares. This may have strongly affected our data and may explain why patients admitted in private institutions have had a nearly 50% reduction risk for death.
Therefore, we believed PMSI-derived data can be a worth-while source of information when using strict rules of sorting. Nevertheless, clinically, the overlapping conditions (COPD exacerbations and near-fatal asthma) are very hard to distinguish, especially in an emergency context where clinical signs and symptoms and therapeutic management are very similar. Optimisation of PMSI coding with the aim of improving institution funding may cause an other bias as different levels of funding are assigned to different codes.
However, PMSI is an exhaustive database, where coding by a clinician is enforced by law in France. So, all admissions were reported. Moreover, French social welfare inspectors are regularly checking cohesiveness between codes and clinical reports with the ability to impose penalties for mistakes.
Conclusion
In conclusion, chronic obstructive pulmonary disease exacerbations leading to hospital admission are responsible for a significant burden and high mortality rates. Several risk factors for death are easy to identify and should prompt more intensive care in the hospital and closer follow-up in the community, to decrease mortality. The results of this study may (i) inform healthcare policy-makers regarding the qualitative (types of units) and quantitative (number of beds) needs in terms of hospitalizations structures (medical wards, ICU, intermediate care units), and (ii) help healthcare providers regarding patients orientation.
Declaration of Interest Statement
Each author's contribution is as follows:
Molinari N: study concept, data analysis and interpretation, manuscript preparation and drafting; Briand C: statistical data analysis; Vachier I: manuscript preparation and review; Malafaye N: acquisition of the data; Aubas P :manuscript preparation and review; Georgescu V: statistical data analysis; Roche N: manuscript preparation and review; Chanez P: manuscript preparation and review; Bourdin A: study design, Statistical methods, statistical data analysis and manuscript review.
The authors have no conflicts of interest to report. The authors alone are responsible for the content and writing of the paper.
References
- GOLD: Global Initiative for Obstructive Lung Disease www.goldcopd.com 2013.
- Hurst JR, Vestbo J, Anzueto A, Locantore N, Mullerova H, Tal-Singer R, Miller B, Lomas DA, Agusti A, Macnee W et al.: Susceptibility to exacerbation in chronic obstructive pulmonary disease. N Engl J Med 2010; 363(12):1128–1138.
- Garcia-Aymerich J, Serra Pons I, Mannino DM, Maas AK, Miller DP, Davis KJ: Lung function impairment, COPD hospitalizations and subsequent mortality. Thorax 2011; 66(7):585–590.
- Wouters EF: The burden of COPD in The Netherlands: results from the Confronting COPD survey. Respir Med 2003; 97 Suppl C:S51–59.
- Wouters EF: Economic analysis of the Confronting COPD survey: an overview of results. Respir Med 2003; 97 Suppl C:S3–14.
- Decramer M, Janssens W, Miravitlles M: Chronic obstructive pulmonary disease. Lancet 2012; 379(9823):1341–1351.
- Agusti A, Faner R: Systemic inflammation and co-morbidities in chronic obstructive pulmonary disease. Proc Am Thorac Soc 2012; 9(2):43–46.
- Aubas C, Bourdin A, Aubas P, Gamez AS, Halimi L, Vachier I, Malafaye N, Chanez P, Molinari N: Role of co-morbid conditions in asthma hospitalizations in the south of France. Allergy 2013; 68(5):637–643.
- Langsetmo L, Platt RW, Ernst P, Bourbeau J: Underreporting exacerbation of chronic obstructive pulmonary disease in a longitudinal cohort. Am J RespirCrit Care Med 2008; 177(4):396–401.
- Gaudreau K, Sanford CJ, Cheverie C, McClure C: The effect of a smoking ban on hospitalization rates for cardiovascular and respiratory conditions in Prince Edward IslandCanada. PLoS One 2013; 8(3):e56102.
- Baillargeon J, Wang Y, Kuo YF, Holmes HM, Sharma G: Temporal trends in hospitalization rates for older adults with chronic obstructive pulmonary disease. Am J Med 2013; 126(7):607–614.
- Walker PP, Thompson E, Crone H, Flatt G, Holton K, Hill SL, Pearson MG: Use of mortality within 30 days of a COPD hospitalization as a measure of COPD care in UK hospitals. Thorax 2013; 68(10):968–970.
- Groenewegen KH, Schols AM, Wouters EF: Mortality and mortality-related factors after hospitalization for acute exacerbation of COPD. Chest 2003; 124(2):459–467.
- Ai-Ping C, Lee KH, Lim TK: In-hospital and 5-year mortality of patients treated in the ICU for acute exacerbation of COPD: a retrospective study. Chest 2005; 128(2):518–524.
- Seneff MG, Wagner DP, Wagner RP, Zimmerman JE, Knaus WA: Hospital and 1-year survival of patients admitted to intensive care units with acute exacerbation of chronic obstructive pulmonary disease. JAMA 1995; 274(23):1852–1857.
- Johnston NW, Sears MR: Asthma exacerbations. 1: epidemiology. Thorax 200661(8):722–728.
- Rabe KF, Fabbri LM, Vogelmeier C, Kogler H, Schmidt H, Beeh KM, Glaab T: Seasonal distribution of COPD exacerbations in the Prevention of Exacerbations with Tiotropium in COPD trial. Chest 2013; 143(3):711–719.
- Bhowmik A, Seemungal TA, Donaldson GC, Wedzicha JA: Effects of exacerbations and seasonality on exhaled nitric oxide in COPD. EurRespir J 2005; 26(6):1009–1015.
- Jenkins CR, Celli B, Anderson JA, Ferguson GT, Jones PW, Vestbo J, Yates JC, Calverley PM: Seasonality and determinants of moderate and severe COPD exacerbations in the TORCH study. EurRespir J 2012; 39(1):38–45.
- Bafadhel M, McKenna S, Terry S, Mistry V, Reid C, Haldar P, McCormick M, Haldar K, Kebadze T, Duvoix A et al.: Acute exacerbations of chronic obstructive pulmonary disease: identification of iologic clusters and their biomarkers. Am J RespirCrit Care Med 2011; 184(6):662–671.
- Mallia P, Johnston SL: How viral infections cause exacerbation of airway diseases. Chest 2006; 130(4):1203–1210.
- Mallia P, Message SD, Gielen V, Contoli M, Gray K, Kebadze T, Aniscenko J, Laza-Stanca V, Edwards MR, Slater L et al.: Experimental rhinovirus infection as a human model of chronic obstructive pulmonary disease exacerbation. Am J RespirCrit Care Med 2011; 183(6):734–742.
- Papi A, Bellettato CM, Braccioni F, Romagnoli M, Casolari P, Caramori G, Fabbri LM, Johnston SL: Infections and airway inflammation in chronic obstructive pulmonary disease severe exacerbations. Am J RespirCrit Care Med 2006; 173(10):1114–1121.
- Divo M, Cote C, de Torres JP, Casanova C, Marin JM, Pinto-Plata V, Zulueta J, Cabrera C, Zagaceta J, Hunninghake G, Celli B; BODE Collaborative Group: Co-morbidities and risk of mortality in patients with chronic obstructive pulmonary disease. Am J RespirCrit Care Med 2012; 186(2):155–161.
- Suissa S, Dell'Aniello S, Ernst P: Long-term natural history of chronic obstructive pulmonary disease: severe exacerbations and mortality. Thorax 2012; 67(11):957–963.
- Vanfleteren LE, Spruit MA, Groenen M, Gaffron S, van Empel VP, Bruijnzeel PL, Rutten EP, Op ‘t Roodt J, Wouters EF, Franssen FM: Clusters of co-morbidities based on validated objective measurements and systemic inflammation in patients with chronic obstructive pulmonary disease. Am J RespirCrit Care Med 2013; 187(7):728–735.
- Antonelli-Incalzi R, Pedone C: COPD: more than respiratory. Lancet 2008; 371(9606):26–27; author reply 28.
- Wilcock A, Crosby V, Freer S, Freemantle A, Caswell G, Seymour J: Lung cancer diagnosed following emergency admission: a mixed methods study protocol to improve understanding of patients’ characteristics, needs experiences and outcomes. BMC Palliat Care 2013; 12(1):24.