Adrenal Insufficiency in the Elderly: A Nationwide Study of Hospitalizations in Taiwan

Transcription

1 Tohoku J. Exp. Med., 2010, 221, Adrenal Insufficiency in the Elderly 281 Adrenal Insufficiency in the Elderly: A Nationwide Study of Hospitalizations in Taiwan Yi-Chun Chen, 1 Yu-Chun Chen, 2 Li-Fang Chou, 3 Tzeng-Ji Chen 1,4 and Shinn-Jang Hwang 1,5 1 Department of Family Medicine, Taipei Veterans General Hospital, Taipei, Taiwan 2 Department of Medical Informatics, Institute for Medical Biometry and Informatics, University of Heidelberg, Heidelberg, Germany 3 Department of Public Finance, National Chengchi University, Taipei, Taiwan 4 Institute of Hospital and Health Care Administration, National Yang-Ming University, Taipei, Taiwan 5 Department of Family Medicine, School of Medicine, National Yang-Ming University, Taipei, Taiwan Keywords: adrenal insufficiency; comorbidity; elderly; epidemiology; mortality Tohoku J. Exp. Med., 2010, 221 (4), Tohoku University Medical Press Adrenal insufficiency (AI) is a life-threatening disorder but its early diagnosis is not easy. It usually starts with variable nonspecific symptoms and signs of poor appetite, fatigue, fever, and gastrointestinal discomfort, and may progress to adrenal crisis with complications of electrolyte imbalance, change of consciousness or even shock. Common etiologies of AI include autoimmune disease, infection, tumor, hemorrhage, and hypothalamic-pituitaryadrenal (HPA) axis suppression related to critical illness or long term use of steroid (Arlt and Allolio 2003). The diagnosis of AI is more difficult in the elderly people because the patients tend to have several morbidities and the symptoms might be mistaken as aging process (Tayal el at. 1994). Most researches in AI are focused on treatment regimen. Epidemiological studies about AI in the elderly are uncommon and often limited in few hospitals or districts (Kong and Jeffcoate 1994; Løvås and Husebye 2002). The aim of our current study is to analyze the incidence of AI among the elderly people by means of the complete nationwide insurance claims of hospitalizations in Taiwan. Special attention is also paid to co-morbidity at AI diagnosis and the consequences within one year after AI diagnosis. Our study would give a better picture of adrenal insufficiency in the elderly for countries with aging population. Methods In Taiwan, the National Health Insurance (NHI) program started in 1995 and covered nearly all inhabitants (22,484,427 beneficiaries at the end of 2006, equivalent to 98.3% of the inhabitants). The contracted medical care facilities in 2006 included 24 academic medical centers, 70 metropolitan regional hospitals, 414 local community hospitals. Since 1999, the Bureau of NHI has released all claims data in electronic form to the National Health Research Institutes under the Received April 7, 2010; revision accepted for publication June 29, doi: /tjem Correspondence: Tzeng-Ji Chen, M.D., Department of Family Medicine, Taipei Veterans General Hospital, No. 201, Sec. 2, Shih-Pai Road, Taipei, Taiwan. tjchen@vghtpe.gov.tw 281

2 282 Y.C. Chen et al. National Health Insurance Research Database (NHIRD) project. The data are available on request to researchers. We obtained the complete datasets of hospitalizations from 1996 to 2007 with a total of 2,912,099 records. One record of these datasets contains the patient s identification number, gender, date of birth, dates of admission and discharge, up to five discharge diagnoses coded in the International Classification of Disease, Ninth Revision, Clinical Modification (ICD-9-CM) format, and the identification number of the medical care facility. Original identification numbers of patients and medical care facilities are encrypted for the reason of privacy protection, but the uniform encrypting procedure allows the linking of data belonging to the same patient or facility within the whole datasets. For example, a medical care facility in the hospitalization record can be linked to the registry file of medical care facilities (HOSB2006.DAT) to know its accreditation status: academic medical centre, metropolitan regional hospital, local community hospital or primary care clinic. Besides, we obtained the registry files of beneficiaries in 2006 to calculate the denominators. In the current study, we firstly identified the patients with AI according to the discharge diagnoses (255.4 in ICD-9-CM). A new case of AI in 2006 was defined as the patient who was diagnosed with AI in 2006 but not in the past ten years from 1996 to Because AI is more frequently seen in the elderly population, our main analysis was limited to patients aged 60 years and over. To analyze the co-morbidities of new AI cases, the first discharge diagnosis was extracted from the hospitalizations with newly diagnosed AI. If AI was the primary diagnosis, the second discharge diagnosis would be extracted instead. The Clinical Classification Software (CCS) 2006 version was used to classify diagnoses into a manageable number of clinically meaningful categories. From our complete hospitalization datasets, these new AI cases were further followed for one year after their discharge. The percentage of the patients with another hospitalization was calculated. The discharge diagnosis and status, especially death, were analyzed. Again, only the primary diagnosis was examined and the CCS used. Statistical analysis The database software of Microsoft SQL Server 2008 (Microsoft Corp., Redmond, WA, USA) was used for data linkage and processing. From our non-sampled population datasets, descriptive statistics were presented. Results Patient characteristics From the nationwide hospitalization datasets, the prevalence of AI increased from 6.5/10 5 (n = 1,293) in 1996 to 20.8/10 5 (n = 4,681) in In 2006, AI was newly diagnosed in 3,494 patients (1,701 women and 1,793 men), which represented an incidence of 15.5/10 5 in the whole population. Nearly four fifths (1,349 women and 1,429 men) of the patients were 60 years old and over, corresponding to an incidence of 92.4/10 5 in the elderly population. The incidence of AI in the elderly men (96.8/10 5 ) was higher than that in the elderly women (88.2/10 5 ). In both sexes, the incidences increased with age and reached 190.6/10 5 and 238.8/10 5 in women and men aged 90 years and over respectively (Table 1). Specialty and accreditation level of hospital in which AI was newly diagnosed The majority (84.6%, n = 2,350) of elderly patients with newly diagnosed AI were treated in wards of internal medicine, followed by those of neurology and surgery (Table 2). Only one ninth (n = 323) of the patients were firstly identified in the local community hospitals. Co-morbidity Among hospitalizations with newly diagnosed AI, only one elderly patient had AI as the sole diagnosis at discharge. The most common co-morbidity was pneumonia (8.6%, n = 238), followed by urinary tract infections (6.9%, n = 193), electrolyte imbalance (6.6%, n = 184), and septicemia (5.7%, n = 159) (Table 3). Cause of hospitalizations and in-hospital death within one year A total of 40 elderly patients died at the hospitalization in which AI was firstly found. During the one-year observation after discharge of the other 2,738 patients, 65% (n = 1,803) of patients had been hospitalized 4,520 times. Pneumonia as the most frequent diagnosis was found in 367 patients with 512 hospitalizations. The relative risk of Table 1. Age- and sex-specific incidence of adrenal insufficiency (AI) in the elderly aged over 60 in Age Population New AI cases Incidence (1/10 5 ) Female Male Female Male Female Male , , , , , , , , , , ,328 58, ,756 18, Total 1,530,342 1,476,981 1,349 1,

3 Adrenal Insufficiency in the Elderly 283 Table 2. Type of hospital and specialty in which adrenal insufficiency was treated. Specialty Academic medical center Metropolitan hospital Local community hospital Total Internal medicine 1,027 1, ,350 Surgery Orthopedics Neurology Others Total 1,261 1, ,778 Table 3. Most common co-morbidities at adrenal insufficiency firstly diagnosed. Disease Grouping (according to CCS*) with co-morbidity No. of in-hospital death hospitalized within 1 year after discharge (%) Code Definition 122 Pneumonia (except that caused by tuberculosis or sexually transmitted disease) (71.0) 159 Urinary tract infections (65.8) 55 Fluid and electrolyte disorders (60.3) 2 Septicemia (except in labor) (60.5) 50 Diabetes mellitus with complications (70.1) 127 Chronic obstructive pulmonary disease and bronchiectasis (84.1) 197 Skin and subcutaneous tissue infections (54.3) 131 Respiratory failure; insufficiency; arrest (adult) (64.5) 109 Acute cerebrovascular disease (65.7) 153 Gastrointestinal hemorrhage (61.2) Others 1, Total 2, ,803 * CCS, Clinical Classifications Software. ( Table 4. Causes of hospitalizations within one year after discharge of newly diagnosed adrenal insufficiency. Disease Grouping (according to CCS*) hospitalized No. of hospitalizations Code Definition 122 Pneumonia (except that caused by tuberculosis or sexually transmitted disease) Urinary tract infections Septicemia (except in labor) Chronic obstructive pulmonary disease and bronchiectasis Respiratory failure; insufficiency; arrest (adult) Diabetes mellitus with complications Congestive heart failure; nonhypertensive Skin and subcutaneous tissue infections Gastrointestinal hemorrhage Coronary atherosclerosis and other heart disease * CCS, Clinical Classifications Software. ( developing pneumonia in patients with newly diagnosed AI was 6 times as high as the risk in general elderly population (13.2% vs. 2.2%). The top ten diagnostic groups accounted for almost a half (n = 2,241) of all hospitalizations (Table 4). During the year, 178 (6.5%) patients died in the hospital and the common causes of death were respiratory failure (n =

4 284 Y.C. Chen et al. 34), septicemia (23), pneumonia (16), and COPD (9). Discussion The causes of AI are heterogeneous. While epidemiological studies usually focused on AI of primary causes or secondary cause related to brain tumor, few had published data of AI in the elderly group. Previous studies about Caucasian population showed that primary AI was relatively rare, with an incidence of /10 5 people (Kong and Jeffcoate 1994; Løvås and Husebye 2002). Although secondary AI was thought to be more prevalent than primary AI, the definite data of incidence is still lacking (Arlt and Allolio 2003). Our study showed the incidence of AI of all causes was high in Taiwan and even six-fold higher among the elderly population. The disparity was also observed by a study in which the incidence of AI among the patients in surgical intensive care unit increased 2.5-fold in patients older than 55 years (Barquist and Kirton 1997). It might be explained by the fact that older people are proved to have decreased sensitivity of HPA axis to cortisol feedback (Wilkinson et al. 1997; Beale et al. 2002). Secondary AI is usually caused by abrupt discontinuation of glucocorticoid therapy or resultant HPA axis suppression due to long term steroid use and subsequent lack of cortisol production in response to a physiologic stress (Schuetz et al. 2008), especially in the elderly group with long term use of steroids due to COPD. In our study, only one-ninth of elderly patients with newly diagnosed AI were found in local community hospitals in The total number of beds in local community hospitals nearly equals that in metropolitan hospitals in Taiwan (40,855 vs. 39,447). ( The lower number of AI patients in local community hospitals might result from the fact that critically ill patients are usually treated in the higher level of hospitals. It remained unknown whether underdiagnosis existed for lack of alertness on AI. Our study also found most elderly AI patients were treated in the internal medicine wards. Although medical illnesses such as autoimmune disorder and infection play a predominant role in AI, some patients might have been transferred from wards of other specialties during hospitalization. In our study, AI was rarely the sole diagnosis the patients had during hospitalizations. Among various co-morbidities, infectious and pulmonary diseases, especially pneumonia, were most common at initial AI diagnosis, during subsequent hospitalizations, and in in-hospital deaths. One epidemiological study in Japan also found the infectious disease was the major inducement of adrenal crisis (Takayanagi et al. 2000). Another study in Japan revealed the close relationship between adrenal function and the severity of community-acquired pneumonia, in-hospital death, and length of stay (Gotoh et al. 2008). Other studies found a high proportion of patients with severe communityacquired pneumonia or sepsis had adrenal insufficiency (Salluh et al. 2006; Annane et al. 2006). A study in Sweden also found that the mortality rate in the AI group due to infection was 5 times more prevalent than that expected in the age-adjusted background population (Bergthorsdottir et al. 2006). All of these studies implied the role of infectious stress in AI. Our study had some limitations. Only the datasets of hospitalizations were available for analysis. The diagnosis of asymptomatic and mild AI might have been made earlier or only at ambulatory care. The incidence was underestimated because only treated and hospitalized cases were included in our study. Besides, we did not differentiate the primary and secondary causes of AI because ICD-9-CM does not provide a clear classification. Moreover, our datasets did not provide any laboratory results or drug prescriptions so that the severity of AI could not be known. The claims diagnosis at in-hospital death might not be identical to that filled in the official certificate of death. To ascertain the cause of death in elderly patients with AI required scrutiny of paper records. Because the anonymized claims data does not contain the residency of each beneficiary, geographic disparities of AI could also not be studied. The types and causes of admissions were unknown because there were only discharge diagnoses available in claims. Finally, the AI diagnosis in our current study could not be validated as in most studies utilizing the anonymized datasets. The NHIRD does not permit any attempt to identify the patients to link with their original medical records. In conclusion, our large nationwide observational study in Taiwan revealed that AI was more prevalent in the elderly population and the incidence increased with age. Infectious and pulmonary diseases are the most common co-morbidities and play a major role in subsequent hospitalizations and in-hospital deaths. The disparity of incidences among different accreditation levels of hospitals needs further study to elucidate the existence of underdiagnosis. Further study could be directed to the risk and causal association between AI and infectious diseases in the elderly. Acknowledgments This study is based in part on data from the National Health Insurance Research Database provided by the Bureau of National Health Insurance, Department of Health and managed by National Health Research Institutes in Taiwan. The interpretation and conclusions contained herein do not represent those of Bureau of National Health Insurance, Department of Health or National Health Research Institutes. This study was supported by a grant from the Taipei Veterans General Hospital (V98S2-006 and V99S2-006). References Annane, D., Maxime, V., Ibrahim, F., Alvarez, J.C., Abe, E. & Boudou, P. (2006) Diagnosis of adrenal insufficiency in severe sepsis and septic shock. Am. J. Respir. Crit. Care Med., 174, Arlt, W. & Allolio, B. (2003) Adrenal insufficiency. Lancet, 361, Barquist, E. & Kirton, O. (1997) Adrenal insufficiency in the surgical intensive care unit patient. J. Trauma, 42, Beale, E., Zhu, J. & Belzberg, H. (2002) Changes in serum cortisol

5 Adrenal Insufficiency in the Elderly 285 with age in critically ill patients. Gerontology, 48, Bergthorsdottir, R., Leonsson-Zachrisson, M., Odén, A. & Johannsson, G. (2006) Premature mortality in patients with Addison s disease: a population-based study. J. Clin. Endocrinol. Metab., 91, Gotoh, S., Nishimura, N., Takahashi, O., Shiratsuka, H., Horinouchi, H., Ono, H., Uchiyama, N. & Chohnabayashi, N. (2008) Adrenal function in patients with community-acquired pneumonia. Eur. Respir. J., 31, Kong, M.F. & Jeffcoate, W. (1994) Eighty-six cases of Addison s disease. Clin. Endocrinol. (Oxf.), 41, Løvås, K. & Husebye, E.S. (2002) High prevalence and increasing incidence of Addison s disease in western Norway. Clin. Endocrinol. (Oxf.), 56, Salluh, J.I., Verdeal, J.C., Mello, G.W., Araújo, L.V., Martins, G.A., de Sousa Santino, M. & Soares, M. (2006) Cortisol levels in patients with severe community-acquired pneumonia. Intensive Care Med., 32, Schuetz, P., Christ-Crain, M., Schild, U., Süess, E., Facompre, M., Baty, F., Nusbaumer, C., Brutsche, M. & Müller, B. (2008) Effect of a 14-day course of systemic corticosteroids on the hypothalamic-pituitary-adrenal-axis in patients with acute exacerbation of chronic obstructive pulmonary disease. BMC Pulm. Med., 8, 1. Takayanagi, R., Miura, K., Nakagawa, H. & Nawata, H. (2000) Epidemiologic study of adrenal gland disorders in Japan. Biomed. Pharmacother., 54, Tayal, S.C., Bansal, S.K. & Chadha, D.K. (1994) Hypopituitarism: a difficult diagnosis in elderly people but worth a search. Age Ageing, 23, Wilkinson, C.W., Peskind, E.R. & Raskind, M.A. (1997) Decreased hypothalamic-pituitary-adrenal axis sensitivity to cortisol feedback inhibition in human aging. Neuroendocrinology, 65,