Hospitalizations for Coronary Artery Disease Among Patients With Systemic Lupus Erythematosus

Transcription

1 ARTHRITIS & RHEUMATISM Vol. 48, No. 9, September 2003, pp DOI /art , American College of Rheumatology Hospitalizations for Coronary Artery Disease Among Patients With Systemic Lupus Erythematosus Christine M. Thorburn 1 and Michael M. Ward 2 1 Christine M. Thorburn, MD: Stanford University School of Medicine, Stanford, California; 2 Michael M. Ward, MD, MPH: Stanford University School of Medicine, Stanford, California, and Veterans Affairs Palo Alto Health Care System, Palo Alto, California (current address: National Institute of Musculoskeletal and Skin Diseases, Bethesda, Maryland). Address correspondence and reprint requests to Michael M. Ward, MD, MPH, Intramural Research Program, National Institute of Musculoskeletal and Skin Diseases, NIH, Building 10, Room 9S205, 10 Center Drive, MSC 1828, Bethesda, MD wardm1@mail.nih.gov. Submitted for publication January 2, 2003; accepted in revised form May 22, Objective. Although patients with systemic lupus erythematosus (SLE) have an increased risk of coronary artery disease (CAD) compared with persons without SLE, the burden of CAD among SLE patients is unknown. This study was undertaken to estimate this burden. Methods. We used the Nationwide Inpatient Sample of the Healthcare Cost and Utilization Project to estimate the number of hospitalizations for CAD among patients with SLE in the US in CAD diagnoses included acute myocardial infarction (MI), unstable angina, cardiac catheterization, percutaneous transluminal coronary angioplasty, or coronary artery bypass grafting as the primary reason for hospitalization. We compared these estimates with the frequency of hospitalization for other reasons. Results. There were an estimated 98,217 hospitalizations among patients with SLE in Of these, 11,947 (12%) were among men, 43,674 (44%) were among women <50 years of age, and 42,596 (43%) were among women >50 years of age. There were 4,951 hospitalizations for CAD, with 1,763 of these for acute MI. In women <50 years old, there were an estimated 311 hospitalizations for MI. Hospitalizations for CAD were less common than hospitalizations for SLE itself or for infections, and in young women, were less common than hospitalizations for complications of chronic renal failure. Conclusion. CAD is an important comorbid condition in patients with SLE, but is not as common a reason for hospitalization as SLE itself, infections, and, in some patient subgroups, chronic renal failure. Premature morbidity from coronary artery disease (CAD) among young women with systemic lupus erythematosus (SLE) has been demonstrated in several observational studies (1 4). Manzi and colleagues reported that the relative risk of acute myocardial infarction (MI) in young women with SLE was 52 times higher than in women without SLE (2). In a study by Esdaile and colleagues (4) and in a previous study by one of us (3), the risk of acute MI was increased 8-fold in patients with SLE. While these studies demonstrate a markedly increased relative risk of CAD in patients with SLE, they do not provide information on the burden of CAD in these patients. Moreover, these studies raise the question of how morbidity from CAD compares with morbidity from other causes in SLE. This information would help us understand the relative impact of CAD on the health of SLE patients and help prioritize allocation of limited health care resources. Since most patients with suspected acute coronary syndromes are hospitalized, hospitalizations for CAD can be used as one measure of disease burden. We determined the number of hospitalizations for CAD in 1998 in a national population-based sample of patients with SLE and examined how often CAD was the principal reason for hospitalization, relative to other causes of morbidity. PATIENTS AND METHODS We used data from the Nationwide Inpatient Sample of the Healthcare Cost and Utilization Project, which is operated by the Agency for Healthcare Research and Quality, US Department of Health and Human Services (5). The National Inpatient Sample is a nationally representative sample of 20% of all community hospitals in the US. The sample is drawn from hospitals in 22 states: Arizona, California, Colorado, Connecticut, Florida, Georgia, Hawaii, Illinois, 2519

2 2520 THORBURN AND WARD Iowa, Kansas, Maryland, Massachusetts, Missouri, New Jersey, New York, Oregon, Pennsylvania, South Carolina, Tennessee, Utah, Washington, and Wisconsin. The project defines community hospitals as nonfederal, short-term care, general, and other specialty hospitals, excluding hospital units of psychiatric and long-term care institutions and rehabilitation hospitals. Hospitals were sampled in 5 strata, with the goal of providing a nationally representative sample of hospitalizations. The sampling strata were as follows: 1) geographic region (Northeast, Midwest, West, or South); 2) control (public, voluntary, or proprietary); 3) location (urban or rural); 4) teaching status (teaching or nonteaching); and 5) size (number of beds) (small, medium, or large). The hospitals were sorted by these variables prior to systematic random sampling. Once the universe of hospitals was identified, 20% of hospitals were randomly selected within each stratum. All hospitalizations at these hospitals were included in the National Inpatient Sample database. Information regarding diagnoses and demographic information for each patient, known as discharge abstracts, was available for each hospitalization. Data included the principal diagnosis (defined as the primary reason for hospitalization), up to 14 secondary diagnoses, the principal procedure, up to 14 secondary procedures, admission and discharge status, patient demographic data, expected payment source (i.e., private insurance, Medicare, Medicaid, self-pay, other), length of stay, and total charges. Data on ethnicity were not collected at hospitals in several states, and therefore could not be analyzed. Discharge diagnoses and procedures were coded by International Classification of Diseases Clinical Modification, Ninth Revision (ICD-9) codes (6). We selected hospitalizations for which a diagnosis of SLE (ICD-9 code 710.0) was included as any of the discharge diagnoses. We excluded hospitalizations of patients younger than 18 years old, and those for childbirth. No unique patient identifiers were included in the database, so it was not possible to determine how many hospitalizations were readmissions. We examined the principal diagnosis of each hospitalization for which SLE was included as one of the discharge diagnoses. In addition to hospitalizations for acute MI, unstable angina, and chest pain, we were interested in the number of hospitalizations for which SLE or its cardinal manifestations were listed as the principal diagnosis. These manifestations included seizures, autoimmune hemolytic anemia, thrombocytopenia, pericarditis, pleuritis, chronic renal failure, nephritis, and psychosis (including major depression). We also examined other specific diagnoses of concern as causes of morbidity in SLE, including osteoporotic fracture (pathologic fracture or hip fracture) and avascular necrosis, and common comorbid conditions included in the Charlson Index (7,8), as described previously (9). The 20 most common discharge diagnoses by individual ICD-9 codes were also examined. We also examined hospitalizations for which cardiac catheterization, percutaneous transluminal coronary angioplasty, coronary artery bypass grafting, and arteriovenous fistula placement or revision were listed as the principal procedures of the hospitalization. A list of the ICD-9 codes used for each condition or procedure is available from the authors. We combined related discharge diagnoses into groups of the most common comorbidities. We defined a category of CAD by combining the principal diagnoses or principal procedures acute MI, unstable angina, cardiac catheterization, percutaneous transluminal coronary angioplasty, and coronary artery bypass grafting. Hospitalizations that included both a principal diagnosis of acute MI or unstable angina and a principal procedure of cardiac catheterization, percutaneous transluminal coronary angioplasty, or coronary artery bypass grafting were counted only once in the CAD category. We did not include congestive heart failure in the CAD category because congestive heart failure may be due to renal insufficiency and subsequent fluid overload, valvular heart disease, or hypertensive heart disease, and not specifically due to ischemic heart disease. We created a category of infection by combining the principal diagnoses pneumonia, cellulitis, osteomyelitis, sepsis, and urinary tract infection. Finally, we defined a category of chronic renal failure and its complications by combining the principal diagnoses or principal procedures chronic renal failure, fistula placement/complication, and renal transplantation. Hospital discharge sampling weights were included in the data set to permit estimation of the total number of hospitalizations in the US by ICD-9 code. We used SAS programs (Statistical Analysis System, Cary, NC) to create frequencies of discharge diagnoses and Stata programs (Stata Corporation, College Station, TX) to perform the weighting function to obtain population estimates with standard deviations and 95% confidence intervals. RESULTS There were 6.8 million hospitalizations in the National Inpatient Sample database. Of these, 19,120 hospitalizations included a discharge diagnosis of SLE. After application of sampling weights to obtain national estimates, there were 98,217 hospitalizations among patients with SLE in the US in Of these, 11,947 (12%) were among men, 43,674 (44%) were among women 50 years of age, and 42,596 (43%) were among women 50 years of age or older. The mean age was 53 years for men and 51 years for women. Table 1 lists the 25 most common principal discharge diagnoses in the total sample and in the subgroups of men, women 50 years of age, and women 50 years of age. SLE was the most common principal diagnosis/procedure, followed by pneumonia, fistula placement/repair, and congestive heart failure. Acute MI was the ninth most common principal diagnosis in the group overall. There were an estimated 311 hospitalizations for MI among young women with SLE in the US in Among women, hospitalizations for cerebrovascular accidents and chronic obstructive pulmonary disease were more frequent than those for acute MI. Psychosis/major depression was the seventh most common reason for hospitalization overall and the fourth leading reason for hospitalization among young women.

3 HOSPITALIZATION FOR CAD AMONG SLE PATIENTS 2521 Table 1. Estimated number of hospitalizations among patients with SLE in the US in 1998, by principal diagnosis/procedure* Principal diagnosis/procedure Total (n 98,217), Men (n 11,947), (n 43,674), (n 42,596), SLE 11,507 (11,053 11,960) 1,459 (1,288 1,630) 7,448 (7,075 7,821) 2,600 (2,373 2,827) Pneumonia 4,925 (4,614 5,235) 746 ( ) 1,861 (1,667 2,055) 2,318 (2,104 2,531) Fistula placement/complication 2,900 (2,662 3,138) 427 ( ) 1,772 (1,584 1,959) 702 ( ) Congestive heart failure 2,872 (2,637 3,106) 430 ( ) 573 ( ) 1,869 (1,678 2,059) Cerebral vascular accident 2,619 (2,395 2,843) 304 ( ) 703 ( ) 1,612 (1,436 1,788) Chronic obstructive pulmonary disease 2,351 (2,138 2,564) 213 ( ) 720 ( ) 1,418 (1,252 1,584) Psychosis/major depression 2,305 (2,090 2,519) 191 ( ) 1,470 (1,297 1,642) 643 ( ) Sepsis 2,259 (2,051 2,466) 325 ( ) 850 ( ) 1,084 (940 1,227) Acute myocardial infarction 1,763 (1,578 1,948) 386 ( ) 311 ( ) 1,066 (921 1,211) Nephritis 1,756 (1,569 1,943) 337 ( ) 900 (766 1,034) 518 ( ) Cancer 1,705 (1,519 1,890) 240 ( ) 359 ( ) 1,106 (955 1,257) Urinary tract infection 1,658 (1,475 1,840) 82 (41 123) 850 ( ) 725 ( ) Cellulitis 1,588 (1,411 1,765) 181 ( ) 757 ( ) 650 ( ) Dehydration 1,349 (1,185 1,513) 120 (70 169) 567 ( ) 662 ( ) Chest pain NOS 1,226 (1,070 1,381) 94 (50 137) 494 ( ) 638 ( ) Deep vein thrombosis 1,212 (1,057 1,367) 221 ( ) 485 ( ) 505 ( ) Pancreatitis 1,080 (936 1,224) 123 (73 173) 671 ( ) 286 ( ) Osteoporotic fractures 898 (765 1,030) 51 (19 83) 105 (60 150) 742 ( ) Pleuritis 852 ( ) 127 (76 179) 383 ( ) 341 ( ) Avascular necrosis 788 ( ) 101 (56 147) 490 ( ) 196 ( ) Seizures 770 ( ) 58 (23 93) 450 ( ) 262 ( ) Respiratory failure 702 ( ) 128 (79 178) 226 ( ) 348 ( ) Atrial fibrillation 632 ( ) 139 (86 191) 71 (32 110) 423 ( ) Thrombocytopenia 620 ( ) 74 (35 112) 333 ( ) 213 ( ) Diabetes mellitus 601 ( ) 56 (23 89) 271 ( ) 275 ( ) * Values are rounded to the nearest integer; therefore, summation of values for men, women age 50 years, and women age 50 years may not equal the number in the Total column. SLE systemic lupus erythematosus; 95% CI 95% confidence interval; NOS not otherwise specified. Hospitalizations for nephritis were considered separately from complications of renal failure. Nephritis was the fifth most common principal diagnosis among young women (n 900) and sixth among men (n 337), but only the sixteenth most common among older women (n 518). Infection, osteoporotic fractures, and avascular necrosis can be complications of SLE treatment that necessitate hospitalization. Osteoporotic fractures and avascular necrosis were the eighteenth (n 898) and twentieth (n 788) most common principal diagnoses overall. The majority of hospitalizations for osteoporotic fractures occurred in older women (n 742), whereas hospitalizations for avascular necrosis were more frequent among young women (n 490). Table 2 lists the frequencies of principal diagnoses in the summary categories of CAD, infection, and chronic renal failure and its complications. Among all patients, hospitalizations for CAD (n 4,951) were half as common as hospitalizations for infection (n 10,715) or for SLE itself (n 11,507). In young women, hospitalizations for CAD (n 1,192) were also less common than hospitalizations for complications of chronic renal failure (n 2,557) and for psychosis/major depression (n 1,470). DISCUSSION Although CAD is an important comorbid condition in patients with SLE, it is not as common a reason for hospitalization as SLE itself, infections, and, in some patient subgroups, chronic renal failure and its complications. Previous studies have shown that the relative risk of CAD was substantially higher among SLE patients than among persons without SLE. The current data suggest that the burden of CAD, as represented by the number of hospitalizations, among patients with SLE is not as great as that due to SLE itself or to infections. The results of the previous studies and these current data can be reconciled by distinguishing the different questions being addressed by studies of relative risk and studies of prevalence. Studies of relative risk and studies of prevalence provide different information about atherosclerosis in SLE. Studies of relative risk provide information about causation. The increased relative risk of CAD in SLE

4 2522 THORBURN AND WARD Table 2. Estimated number of hospitalizations for coronary artery disease, infection, chronic renal failure and its complications, and SLE among patients with SLE in the US in 1998* Principal diagnosis/procedure Total (n 98,217), Men (n 11,947), (n 43,674), (n 42,596), Coronary artery disease 4,951 (4,643 5,258) 818 ( ) 1,192 (1,038 1,347) 2,940 (2,701 3,179) Myocardial infarction 1,763 (1,578 1,948) 386 ( ) 311 ( ) 1,066 (921 1,211) Unstable angina 479 ( ) 41 (12 69) 159 ( ) 279 ( ) Catheterization 2,139 (1,934 2,343) 282 ( ) 624 ( ) 1,233 (1,077 1,389) PTCA 1,063 (916 1,209) 247 ( ) 236 ( ) 579 ( ) CABG 555 ( ) 113 (65 161) 105 (58 151) 337 ( ) Infection 10,715 (10,276 11,155) 1,370 (1,204 1,537) 4,439 (4,145 4,733) 4,906 (4,601 5,211) Complications of chronic renal failure 4,058 (3,776 4,340) 582 ( ) 2,557 (2,331 2,782) 919 (782 1,057) SLE 11,507 (11,053 11,960) 1,459 (1,288 1,630) 7,448 (7,075 7,821) 2,600 (2,373 2,827) * Values are rounded to the nearest integer; therefore, summation of values for men, women age 50 years, and women age 50 years may not equal the number in the Total column. PTCA percutaneous transluminal coronary angioplasty; CABG coronary artery bypass grafting (see Table 1 for other definitions). Includes the principal diagnoses pneumonia, cellulitis, osteomyelitis, sepsis, or urinary tract infection. Includes the principal diagnoses or principal procedures chronic renal failure, fistula placement/complication, or renal transplantation. patients demonstrated in previous studies suggests that SLE, or factors associated with it, promote atherosclerosis and lead to the development of CAD. This possibility is supported by study results indicating that CAD is associated with the duration of SLE, independent of other cardiovascular risk factors (2,10). The finding that patients with SLE have an increased risk of coronary heart disease even after adjustment for the presence of traditional risk factors suggests either that SLE magnifies the risks associated with known cardiovascular risk factors or that there are risk factors unique to SLE (4,11). These findings, along with the reported increased risk of CAD in patients with rheumatoid arthritis, have led to speculation that immunologic and inflammatory mechanisms may lead to CAD (12,13). Studies demonstrating increased relative risk thus provide clues to new pathophysiologic mechanisms that may cause atherosclerosis, to which SLE patients may be distinctively prone. Exposures (such as SLE) associated with high relative risk may not be associated with high burdens of disease if the rate of disease in the comparison population is low. In contrast, studies of prevalence provide information on the magnitude of a health problem. The results of this study indicate that there were 5,000 hospitalizations for CAD among SLE patients in the US in For groups of patients, such information can be one factor to consider in assessing the importance of a problem and in prioritizing attention and resources. Other factors to consider include the severity of the condition and how amenable it may be to treatment, its impact on present or future quality of life, and its economic consequences. Since there are known effective interventions for CAD risk factors, risk factor identification and modification for the primary prevention of CAD should be aggressively pursued in patients with SLE. Available data suggest that in SLE patients, risk factor management by rheumatologists does not always follow recommended evidence-based guidelines (14). However, modification of traditional cardiovascular risk factors may be less effective in preventing CAD in SLE patients if atherosclerosis in SLE is primarily the result of unique, nontraditional risk factors. This concern suggests the need for a controlled trial of CAD prevention among patients with SLE (15). Our data on the frequency of hospitalizations for CAD also suggest that screening for asymptomatic CAD and treatment for secondary prevention of CAD in SLE patients should be carefully studied before being adopted. The strengths of this study include its national, population-based sample, large size, and examination of multiple comorbid conditions. However, there are limitations when a hospitalization database is used for analysis. We do not know if all patients with SLE were captured in the database, or the extent to which patients with remotely diagnosed or inactive SLE were not reported. Underdiagnosis for these reasons would be more likely to occur in older patients than in younger patients. We also do not know if those diagnosed as having SLE met current classification criteria for this disease; however, in all patients, the diagnosis of SLE was made by a physician. We do not have information about diagnoses of stable angina or other conditions that may be related to CAD for which patients were seen on an outpatient basis, which could have led to an under-

5 HOSPITALIZATION FOR CAD AMONG SLE PATIENTS 2523 estimation of the burden of CAD. However, this underestimation applies similarly to other comorbid conditions, such as infections that did not result in hospitalization. We did not include congestive heart failure as a manifestation of CAD because congestive heart failure in SLE could have other potential etiologies. The database was also limited in not having unique patient identifying codes. Therefore, we could estimate only the number of hospitalizations, rather than the number of individual patients with each diagnosis. Patients with conditions that result in frequently recurrent hospitalizations are thus overrepresented, but this is unlikely to have affected the estimates for CAD hospitalizations greatly. Finally, the database did not include complete information about patient ethnicity, so we could not assess differences in comorbid conditions among different ethnic populations. This study is the first to compare the burden of disease due to CAD, as represented by hospitalization frequencies, with other comorbid conditions in a national sample of patients with SLE. The data indicate that much morbidity among SLE patients is due to active SLE, infections, and renal failure, and less to CAD. From a clinical perspective, CAD risk factor identification and modification should be pursued in all patients, pending any acquisition of new information on risk factors that may be unique to SLE. From a health planning perspective, our data suggest that interventions to prevent infections or renal failure among patients with SLE should be addressed concomitantly with programs to prevent CAD. REFERENCES 1. Jonsson H, Nived O, Sturfelt G. Outcome in systemic lupus erythematosus: a prospective study of patients from a defined population. Medicine (Baltimore) 1989;68: Manzi S, Meilahn EN, Rairie JE, Conte CG, Medsger TA, Jansen-McWilliams L, et al. Age-specific incidence rates of myocardial infarction and angina in women with systemic lupus erythematosus: comparison with the Framingham study. Am J Epidemiol 1997;145: Ward MM. Premature morbidity from cardiovascular and cerebrovascular diseases in women with systemic lupus erythematosus. Arthritis Rheum 1999;42: Esdaile JM, Abrahamowicz M, Grodzicky T, Li Y, Panaritic C, du Berger R, et al. Traditional Framingham risk factors fail to fully account for accelerated atherosclerosis in systemic lupus erythematosus. Arthritis Rheum 2001;44: National Technical Information Service. Overview of the healthcare cost and utilization project nationwide inpatient sample (NIS). Springfield (VA): Agency for Healthcare Research and Quality; HCIA. The international classification of diseases, 9th revision, clinical modification: ICD-9-CM. 10th ed. Baltimore: HCIA; Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40: Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol 1992;45: Ward MM. Hospital experience and mortality in patients with systemic lupus erythematosus. Arthritis Rheum 1999;42: Petri M, Perez-Gutthann S, Spence D, Hochberg MC. Risk factors for coronary artery disease in patients with systemic lupus erythematosus. Am J Med 1992;93: Rahman P, Urowitz MB, Gladman DD, Bruce IN, Genest J Jr. Contribution of traditional risk factors to coronary artery disease in patients with systemic lupus erythematosus. J Rheumatol 1999; 26: Lopes-Virella MF, Virrella GT. Atherosclerosis and autoimmunity. Clin Immunol Immunopathol 1994;73: Svenungsson E, Jensen-Urstad K, Heimbürger M, Silveira A, Hamsten A, de Faire U, et al. Risk factors for cardiovascular disease in systemic lupus erythematosus. Circulation 2001;104: Bruce IN, Gladman DD, Urowitz MB. Detection and modification of risk factors for coronary artery disease in patients with SLE: a quality improvement study. Clin Exp Rheumatol 1998;16: Liang MJ, Mandl LA, Costenbader K, Fox E, Karlson E. Atherosclerotic vascular disease in systemic lupus erythematosus. J Natl Med Assoc 2002;94:813 9.