http://orcid.org/0000-0002-4465-1951
Figures
Abstract
There are different phenotypes of obstructive sleep apnoea (OSA), many of which have not been characterised. Identification of these different phenotypes is important in defining prognosis and guiding the therapeutic strategy. The aim of this study was to characterise the entire population of continuous positive airway pressure (CPAP)-treated patients in Catalonia and identify specific patient profiles using cluster analysis.
A total of 72,217 CPAP-treated patients who contacted the Catalan Health System (CatSalut) during the years 2012 and 2013 were included. Six clusters were identified, classified as “Neoplastic patients” (Cluster 1, 10.4%), “Metabolic syndrome patients” (Cluster 2, 27.7%), “Asthmatic patients” (Cluster 3, 5.8%), “Musculoskeletal and joint disorder patients” (Cluster 4, 10.3%), “Patients with few comorbidities” (Cluster 5, 35.6%) and “Oldest and cardiac disease patients” (Cluster 6, 10.2%). Healthcare facility use and mortality were highest in patients from Cluster 1 and 6. Conversely, patients in Clusters 2 and 4 had low morbidity, mortality and healthcare resource use.
Our findings highlight the heterogeneity of CPAP-treated patients, and suggest that OSA is associated with a different prognosis in the clusters identified. These results suggest the need for a comprehensive and individualised approach to CPAP treatment of OSA.
Citation: Turino C, Bertran S, Gavaldá R, Teixidó I, Woehrle H, Rué M, et al. (2017) Characterization of the CPAP-treated patient population in Catalonia. PLoS ONE 12(9): e0185191. https://doi.org/10.1371/journal.pone.0185191
Editor: Clemens Heiser, Technische Universitat Munchen, GERMANY
Received: June 16, 2017; Accepted: September 7, 2017; Published: September 21, 2017
Copyright: © 2017 Turino et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper.
Funding: This study was supported by the Spanish Respiratory Society (SEPAR), Associació Lleidatana de Respiratori (ALLER) and ResMed, a company that provides diagnostic services and treatment for sleep apnoea. The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. ResMed paid the salaries of the author HW, but did not have any additional role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: ResMed paid the salaries of the author HW, but did not have any additional role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. This does not alter our adherence to PLOS ONE policies on sharing data and materials.
Introduction
Obstructive sleep apnoea (OSA) is a chronic disorder characterised by recurrent episodes of upper airway collapse during sleep, and affects 5–14% of adults aged 30–70 years [1]. OSA has been linked with increased rates of morbidity and mortality due to its strong association with hypertension, metabolic, cardiovascular and cerebrovascular diseases, and cancer [2,3]. OSA has a negative impact on quality of life, increases the risk of traffic accidents and has an important socioeconomic impact [4,5]. Given these multiple medical and social consequences, OSA could be considered a complex and heterogeneous disorder, deserving of a multidisciplinary approach and personalised treatment. However, there is really only one standard approach to the management of OSA–the application of nocturnal continuous positive airway pressure (CPAP) to splint the upper airways open. CPAP has been shown to improve quality of life and to decrease arterial blood pressure in patients with resistant hypertension [6, 7].
Using distinct types and sources of data, some authors have recently used cluster analysis to identify different phenotypes of OSA patients [8–11]. Cluster analysis allows patients to be grouped according to similar characteristics while maximising differences among different patient groups. Applied to OSA patients, cluster analysis could help to improve knowledge about the condition, confirm known associations with comorbidities, and potentially identify currently unknown associations. In Catalonia, approximately 1% of the general population is currently estimated to be using CPAP. However, there are no clear data on the profiles of OSA patients treated with CPAP, whether there is heterogeneity within this population, and which pathologies or comorbidities might be associated with different patient profiles.
This study characterised the entire CPAP-treated population of Catalonia using cluster analysis in order to define specific profiles based on age, sex, associated comorbidities, mortality and the use of healthcare resources.
Methods
Design, setting and study population
This cross-sectional study was conducted in Catalonia (Spain) based on data from the Agency for Health Quality and Assessment of Catalonia (AQuAS). AQuAS is a public entity working under the auspices of Catalonia’s Health Services Department promoting the quality, safety and sustainability of the public Catalan healthcare system. All OSA patients in the Catalan Health Service who were treated with CPAP and had any use of healthcare resources during 2012 and/or 2013 were included in the analysis. Patients receiving CPAP via private health services were excluded because full data were not available. Since all data were anonymised, neither individual patient consent nor ethical approval were required.
Coding and selection of diseases
The International Classification of Disease version 9 (ICD-9) was used for disease coding at each contact with the Catalan Public Health Service (in primary care, hospital or nursing home).
For this study, we selected a combination of the most frequent diagnoses in our dataset and made a list of the most clinically relevant diagnoses (Table 1) [9]. Several diagnoses were grouped in disease categories in order to facilitate information management. To obtain consistent and clinically relevant patterns of association, and to avoid spurious relationships that could bias the results, we considered only diagnoses with a prevalence of > 1%.
Statistical analysis
Descriptive statistics of mean (standard deviation) or median [interquartile range (IQR)] were estimated for quantitative variables with a normal or non-normal distribution, respectively, while absolute and relative frequencies were used for qualitative variables. Normal distribution was analysed using the Shapiro-Wilks test.
A Multiple Correspondence Analysis (MCA) was used to reduce dimensionality on the studied diagnoses. The number of dimensions obtained by MCA was identified using the Kaiser's criterion and the Scree test. The individual coordinates obtained with MCA were introduced in a k-means cluster analysis. Cluster analysis has been used to describe homogeneous subgroups (clusters) with similar characteristics (intra-cluster distance minimised), but different from other groups (inter-cluster distance maximised). The final number of clusters was defined on the basis of maximising the ratio between intra-cluster and inter-cluster variance, more specifically Calinsky-Harabasz's criterion. The R statistical software, version 3.3.1, was used for all the analyses.
Results
Patient characteristics
Of the 7,478,968 population in Catalonia (2013), 71,217 patients (0.95%) were being treated with CPAP; 70,469 of these used healthcare services (2012–2013) and were included in the analysis (Fig 1). Median age was 64.5 years [IQR57.0; 72.0], 74.9% were men, and 5.29% died during the study period. Median time on CPAP was 34.9 months [IQR 14.8; 58.5]. The most frequent diagnoses were hypertension (61.2%), dyslipidaemia (29.9%), diabetes (29.6%) and obesity (18.3%).
Abbreviations: OSA, obstructive sleep apnoea; CPAP, continuous positive airway pressure.
Cluster analysis
Six clusters of CPAP-treated OSA patients were identified. Hypertension and diabetes were present in almost all the clusters among the most frequent comorbidities (Fig 2). The main characteristics of each cluster are summarized in Table 2.
Cluster 1: Neoplastic patients. Cluster 2: Metabolic syndrome patients. Cluster 3: Asthmatic patients. Cluster 4: Musculoskeletal and joint disorders patients. Cluster 5: Patients with few comorbidities. Cluster 6: Oldest and cardiac disease patients. Abbreviations: DLP, dyslipidaemia; OB, obesity; BD, back disorders; OA, osteoarthrosis; HF, heart failure; CD, cardiac dysrhythmia; ADT, addiction; ANXTY, anxiety; OHD, other heart disease; OJD, other joint disease; PT, peripheral tendinitis; CKD, chronic kidney disease; ASTH, asthma; STD, soft tissue disease; MN, malignant neoplasm; HBP, hypertension; CVD, cerebrovascular; COPD, chronic obstructive pulmonary disease; ANM, anaemia; DM, diabetes mellitus; OP, osteoporosis.
Cluster 1 (Neoplastic patients) included 7,340 patients (10.4%), a high proportion of whom had malignant neoplasm (88.5%), and the mortality rate was high (15.0%). Cluster 2 (Metabolic syndrome patients) included 19,535 patients (27.7%). High proportions of patients in this group had hypertension (84.1%), dyslipidaemia (57.1%), obesity (35.9%) and diabetes (53.9%); mortality was low (1.9%). Cluster 3 (Asthmatic patients) included 4,082 patients (5.8%). This cluster of patients all had asthma, included a high proportion of women (53.0%) and had a low mortality rate (4.8%). Cluster 4 (Musculoskeletal and joint disorders patients) included 7,234 patients (10.3%), who had peripheral tendinitis (58.2%), joint diseases (15.2%) and muscular disease (51.0%); this group had the lowest mortality rate (1.4%). Cluster 5 (Patients with few comorbidities) grouped the patients with few comorbidities (n = 25,088 patients, 35.6%). Use of healthcare resources by this group was low, as was the mortality rate (3.6%). Cluster 6 (Oldest and cardiac disease patients, n = 7,190, 10.2%) had a median age of 72.0 years [IQR 64.5;79.5], and patients had dysrhythmia (67.5%) and heart failure (57.1%). This group had one of the highest mortality rates (14.6%).
Mortality rates in both Cluster 1 and 6 were high, but the patient groups differed with respect to the primary comorbidity diagnosis and healthcare resource use. Specifically, patients in Cluster 6 used a wide range of healthcare resources whereas those in Cluster 1 had more hospital and nursing home visits (Fig 3). Mortality rates in Cluster 2 and 4 were also had similar (low), but patients in Cluster 4 used more primary care resources than those in Cluster 2. In addition, mortality rates were similar in Clusters 3 and 5, but healthcare resource use was again different, being very low for Cluster 5 and higher for Cluster 3.
Cluster 1: Neoplastic patients. Cluster 2: Metabolic syndrome patients. Cluster 3: Asthmatic patients. Cluster 4: Musculoskeletal and joint disorders patients. Cluster 5: Patients with few comorbidities. Cluster 6: Oldest and cardiac disease patients. (% Mortality).
Discussion
This study is the first cluster analysis involving the entire CPAP-treated population with OSA of Catalonia. Using data from the Catalan Health System, we defined a general profile of OSA patients treated with CPAP, largely characterised by middle-aged men, with a high prevalence of hypertension, dyslipidaemia, diabetes and obesity. At the same time, cluster analysis identified six patient groups that showed different patterns of comorbidities, mortality, and healthcare resource use.
Similar to previous literature, we found that hypertension, diabetes and dyslipidaemia were highly prevalent among OSA patients [12–15]. In a sample of more than 18,000 patients from a prospective national registry, Bailey et al. also confirmed the high burden of comorbidities in OSA patients, identifying six clusters [8]. However, similar to other previous cluster studies, they characterised OSA patients using data from a national registry, clinical practice and sleep registry analysis [8–11], while we exclusively used the coded diseases and discharge data from hospitals, nursing homes and primary care institutions, the number of visits to the emergency room or primary care, and medication use. The observation of a cluster comprised entirely of patients with asthma, mostly women, confirmed previous observations of an OSA-asthma overlap syndrome [16] and suggests a need for more specific studies in this field. Asthma is, in fact, a recognised risk factor for developing OSA [17] and women with OSA are more likely to be diagnosed with asthma [18].
Over and above the characterisation of six clusters, the CPAP-treated population from Catalonia could be divided into two major groups. Almost 20% of the overall population were allocated to clusters 1 and 6, and showed the most advanced age and the highest mortality. The majority of patients (88.5%) in Cluster 1 (Neoplastic patients) had a malignant neoplasm, most frequently of genitourinary or gastrointestinal origin. This is of interest given some current literature reports of a higher prevalence of cancer (particularly pancreatic or renal tumours) in patients with OSA [19]. Cluster 6 (Oldest and cardiac patients) included the oldest individuals (median age 72.0 years), with the highest prevalence of cardiac failure and cardiac arrhythmias. Patients in Clusters 1 and 6 showed the highest mortality and rate of hospitalisation, almost certainly due to the underlying comorbidities rather than as a result of OSA. The presence of cardiovascular diseases has been associated with a worse prognosis in patients using CPAP [20] and CPAP treatment of sleep apnoea has not been shown to improve survival in patients with these comorbidities [21]. Furthermore, in patients over 65 years of age, the presence of OSA seems to have only a slight impact on quality of life, which is determined to a greater extent by the presence of comorbidities [22]. In contrast, more than a half of our population was male, had few comorbidities, and low mortality and healthcare resource use (Clusters 2 and 5). In these groups, OSA appears to be the most important determinant of patient prognosis [23], and patients with these characteristics could be more likely to benefit from CPAP treatment.
Given the different phenotypes identified, the results of our study could have an important impact on Catalan health policies. The Catalan Health System provides free healthcare services to more than 7 million people [24] with annual spending of around €8,000 million [25]. CPAP treatment is provided at no cost to approximately 70,000 people, and cost effectiveness is only achieved after the second year of treatment and exclusively in patients who are compliance with therapy [26]. In Catalonia CPAP therapy is typically prescribed to patients with severe OSA or for more mild disease that is accompanied by daytime hypersomnolence or other symptoms attributable to OSA. Daily CPAP compliance is closely monitored because CPAP treatment is completely free of charge only for patients who used their device for more than 3 hours per night; if this is not the case, CPAP is withdrawn [27]. Even in the absence of specific data about the compliance of our population, it would be reasonable to assume device usage of at least 3 h/night given the treatment criteria and the fact that median duration of CPAP use was 34 months.
Understanding inter-patient differences in clinical presentations of OSA could facilitate more efficient resource management and provision of care. In the light of cluster analysis results, one-third of all CPAP-treated OSA patients in Catalonia (Clusters 1 & 6) are in fact receiving a treatment that probably will not markedly influence their life expectancy or quality of life. Medical resources could be better spent for the remaining population of patients with few comorbidities, for whom a clinical benefit of CPAP treatment would be expected.
This study has several strengths, including the large and comprehensive study population and the statistical method used. In addition, data about patient characteristics, comorbidities and resource use were provided by AQuAS, ensuring high quality information. Furthermore, data were collected from different public health settings (primary care, nursing homes and hospitals) increasing the generalisability of our findings. However, there are some limitations to be considered. Firstly, the absence of clinical information about the study population (e.g. symptoms, quality of life, sleep records, CPAP compliance) limits the ability to fully characterise each cluster. Secondly, use of the ICD-9 classification system reduces the specificity of disease definitions. The ICD-9 sometimes groups similar diseases together, reducing the ability to differentiate between them. However, it does ensure the homogeneity and accuracy of the disease classification. Finally, the absence of a control group also limits our capacity to define clusters and assess whether the features identified are specific to OSA patients.
Conclusions
This study used cluster analysis based on diagnostic profile to characterise the entire CPAP-treated population of Catalonia for the first time. Six clusters were identified, but the majority of patients could be distributed into two broad groups: one older with high mortality and healthcare resource use, and the other with few comorbidities, low mortality and lower healthcare resource use. Our study highlights the heterogeneity of OSA patients on CPAP treatment, emphasises the importance of identifying the indication and expected benefits of CPAP in specific OSA phenotypes, and offers the opportunity to tailor interventions for specific patient groups.
Acknowledgments
The authors would like to thank the Agency for Health Quality and Assessment of Catalonia (AQuAS) for providing the datasets.
References
- 1. Peppard PE, Young T, Barnet JH, Palta M, Hagen EW, Hla KM. Increased prevalence of sleep-disordered breathing in adults. Am J Epidemiol. 2013; 177: 1006–14. pmid:23589584
- 2. Campos-Rodriguez F, Martinez-Garcia MA, Martinez M, Duran-Cantolla J, Peña Mde L, Masdeu MJ, et al. Spanish Sleep Network. Association between obstructive sleep apnea and cancer incidence in a large multicenter Spanish cohort. Am JRespirCritCare Med. 2013; 187;99–105.
- 3. Sánchez-de-la-Torre M, Campos-Rodriguez F, Barbé F. Obstructive sleep apnoea and cardiovascular disease. Lancet Respir Med. 2013;1:61–72. pmid:24321805
- 4. Terán-Santos J, Jiménez-Gómez A, Cordero-Guevara J. The association between sleep apnea and the risk of traffic accidents. Cooperative Group Burgos-Santander. N Engl J Med. 1999;340:847–851. pmid:10080847
- 5. Barbé F, Pericás J, Muñoz A, Findley L, Antó JM, Agustí AG. Automobile accidents in patients with sleep apnea syndrome. An epidemiological and mechanistic study. Am J Respir Crit Care Med. 1998;158:18–22. pmid:9655701
- 6. Barbé F, Durán-Cantolla J, Sánchez-de-la-Torre M, Martínez-Alonso M, Carmona C, Barceló A, et al. Spanish Sleep And Breathing Network. Effect of continuous positive airway pressure on the incidence of hypertension and cardiovascular events in non sleepy patients with obstructive sleep apnea: a randomized controlled trial. JAMA 2012;307:2161–8. pmid:22618923
- 7. Martínez-García MA, Capote F, Campos-Rodríguez F, Lloberes P, Díaz de Atauri MJ, Somoza M, et al. Spanish Sleep Network. Effect of CPAP on blood pressure in patients with obstructive sleep apnea and resistant hypertension: the HIPARCO randomized clinical trial. JAMA. 2013;310:2407–15. pmid:24327037
- 8. Bailly S, Destors M, Grillet Y, Richard P, Stach B, Vivodtzev I, et al. Obstructive sleep apnea: acluster analysis at time ofdiagnosis. PLoS One. 2016;11(6):e0157318. pmid:27314230
- 9. Howell S, Coory M, Martin J, Duckett S.Using routine inpatient data to identify patients at risk of hospital readmission. BMC Health Serv Res 2009;9:96. pmid:19505342
- 10. Ye L, Pien GW, Ratcliffe SJ, Björnsdottir E, Arnardottir ES, Pack AI, et al. The different clinical faces of obstructive sleep apnoea: a cluster analysis. Eur Respir J. 2014;44:1600–7. pmid:25186268
- 11. Lacedonia D, Carpagnano GE, Sabato R, Storto MM, Palmiotti GA, Capozzi V, et al. Characterization of obstructive sleep apnea-hypopnea sydrome (OSA) population by means of cluster analysis. J Sleep Res. 2016;25:724–30. pmid:27191534
- 12. Somers VK, White DP, Amin R, Abraham WT, Costa F, Culebras A, et al. Sleep apnea and cardiovascular disease: an American Heart Association/American College of Cardiology Foundation Scientific Statement from the American Heart Association Council for High Blood Pressure Research Professional Education Committee, Council on Clinical Cardiology, Stroke Council, and Council on Cardiovascular Nursing. J Am Coll Cardiol. 2008;52:686–717. pmid:18702977
- 13. Kent BD, McNicholas WT, Ryan S. Insulin resistance, glucose intolerance and diabetes mellitus in obstructive sleep apnoea. J Thorac Dis. 2015;7:1343–57. pmid:26380761
- 14. Trzepizur W, Le Vaillant M, Meslier N, Masson P, Humeau MP, et al. Independent association between nocturnal intermittent hypoxemia and metabolic dyslipidemia. Chest. 2013;143:1584–9. pmid:23699812
- 15. Pamidi S, Tasali E. Obstructive sleep apnea and type 2 diabetes: is there a link? Front Neurol. 2012; 3: 126. pmid:23015803
- 16. Madama D, Silva A, Matos MJ. Overlap syndrome- asthma and obstructive sleep apnea. Rev Port Pneumol. 2016;22:6–10. pmid:26603307
- 17. Teodorescu M, Barnet JH, Hagen EW, Palta M, Young TB, Peppard PE. Association between asthma and risk of developing obstructive sleep apnea. JAMA. 2015;313:156–64. pmid:25585327
- 18. Greenberg-Dotan S, Reuveni H, Simon-Tuval T, Oksenberg A,Tarasiuk A. Gender differences in morbidity and health care utilization among adult obstructive sleep apnea patients. Sleep. 2007; 30: 1173–80. pmid:17910389
- 19. Gozal D, Ham SA, Mokhlesi B. Sleep apnea and cancer: Analysis of a nationwide population sample. Sleep. 2016;39:1493–500. pmid:27166241
- 20. Jennum P, Tønnesen P, Ibsen R, Kjellberg Jm. Obstructive sleep apnea: effect of comorbidities and positive airway pressure on all-cause mortality. Sleep Med. 2017 Aug;36:62–66. pmid:28735924
- 21. McEvoy RD, Antic NA, Heeley E, Luo Y, Ou Q, Zhang X, Mediano O, et al. CPAP for prevention of cardiovascular events in obstructive sleep apnea. N Engl J Med. 2016;375:919–31. pmid:27571048
- 22. Martınez-Garcia MA, Soler-Cataluña JJ, Roman-Sanchez P, Gonzalez V, Amoros C, Montserrat JM. Obstructive sleep apnea has little impact on quality of life in the elderly. Sleep Med. 2009 Jan;10(1):104–11 pmid:18207454
- 23. Gottlieb DJ, Yenokyan G, Newman AB, O’Connor GT, Punjabi NM, et al. A prospective study of obstructive sleep apnea and incident coronary heart disease and heart failure: the Sleep Heart Health study. Circulation. 2010; 122: 352–60. pmid:20625114
- 24.
Generalitat de Catalunya. Institut d’Estadistica de Catalunya. http://www.idescat.cat/pub/?id=aec&n=245&lang=es Date last accessed: March 28 2017.
- 25.
Generalitat de Catalunya. Institut d’Estadistica de Catalunya. http://www.idescat.cat/pub/?id=aec&n=535&lang=es Date last accessed: March 28 2017.
- 26. Català R, Villoro R, Merino M, Sangenís S, Colomés L, Hernández Flix S, Pérez de Llanod LA. Cost-effectiveness of Continuous Positive Airway Pressure Treatmentin Moderate-Severe Obstructive Sleep Apnea Syndrome. Arch Bronconeumol. 2016;52(9):461–469. pmid:26993090
- 27.
http://aquas.gencat.cat/web/.content/minisite/aquas/projectes/atles/atles_variacions/Pressio_positiva_continua_respiratories/PressioPositivaContinua_CPAP_Atles_Variacions_AQuAS.pdf