Causes of death in Korean patients with systemic lupus erythematosus: A single center retrospective study

Transcription

1 Hypothalamus-pituitary-adrenocortical and -gonadal axis in RA / M. Cutolo LongoVital in Sjögren's syndrome / A. EDITORIAL Pedersen et al. Causes of death in Korean patients with systemic lupus erythematosus: A single center retrospective study W.-U. Kim, J.-K. Min, S.-H. Lee, S.-H. Park, C.-S. Cho, H.-Y. Kim Center for Rheumatic Diseases, Research Center in Catholic Medical Center, Kang-Nam St. Mary s Hospital, and the Department of Internal Medicine, Catholic University of Korea, Seoul, Korea. Abstract Objective To determine the causes of death in Korean patients with systemic lupus erythematosus (SLE). Methods We evaluated retrospectively Korean SLE patients who were monitored in the Center for Rheumatic Disease in Kang-Nam St. Mary s Hospital from 1993 to 1997 and who died. Results Forty-three (7.9%) of 544 patients died. Comparison of demographics and disease activity indices between the deceased and the survivors showed that the age was older and C3 at presentation was lower in the deceased (n = 40) than the survivors (n = 453) (age: 33.8 ± 13.6 versus 28.3 ± 10.6 years, p = 0.02, C3: 36.8 ± 21.4 versus 49.7 ± 20.8 mg/dl, p = 0.03). Among 40 patients who died, the frequency and causes of death were as follows: 13 from infection (32.5%), 10 SLE-related factors (25.0%), 6 pulmonary hypertension (15.0%), 4 cerebrovascular accidents (10.0%), and 3 thrombotic thrombocytopenic purpura (7.5%). The majority of the SLE-related deaths were non-renal in origin, including 3 cerebral nervous system disease, 2 TTP, 2 acute pulmonary hemorrhage syndrome, 1 acute myocarditis, and 1 multi-system illness. SLE-related renal causes were responsible for only death. The major organisms of infection were gram negative bacilli (69.2%), primarily manifesting as sepsis or bacteremia (76.9%). The patients (n = 13) who died from infection had lower levels of complement and higher levels of anti-ds DNA antibody at presentation than those (n = 27) who died from the other causes (C3: 24.7 ± 17.8 versus 41.7 ± 21.5 mg/dl, p = 0.02, anti-dsdna antibody: 68.0 ± 73.5 versus 27.0 ± 35.3 IU, p = 0.04). The mean steroid dose being administered one month before death was also higher in the patients who died of infection (30.5 ± 15.2 versus 15.2 ± 7.7 mg/day, p = 0.03). Patients who died of pulmonary hypertension, the third most common cause of mortality, showed extremely high pulmonary pressures at the initial diagnosis, with a short interval to death, and had less major organ involvement at death. There were no deaths due to coronary heart disease or neoplasm in this cohort. Conclusion The most common cause of death in 544 Korean lupus patients was infection, mainly manifesting as gram negative bacterial sepsis. SLE-related factors (mostly non-renal) were the next most frequent cause. Death from infection was associated with higher disease activity at presentation and a higher dose of steroid used previously. Death due to pulmonary hypertension was common, whereas death due to coronary heart disease was absent. Key words Systemic lupus erythematosus, Korean, cause of death. Clinical and Experimental Rheumatology 1999; 17:

2 Causes of death in Korean patients with SLE / W.-U. Kim et al. Supported by grants from the Catholic University of Korea. Wan-Uk Kim, MD, Instructor of Medicine; Jun-Ki Min, Instructor of Medicine; Sang- Heon Lee, MD, PhD, Assistant Professor of Medicine; Sung-Hwan Park, MD, Instructor of Medicine; Chul-Soo Cho, MD, PhD, Associate Professor of Medicine; Ho-Youn Kim, MD, PhD, Professor of Medicine. Please address correspondence and reprint requests to: Dr. Ho-Youn Kim, Division of Rheumatology, Department of Internal Medicine, The School of Medicine, The Catholic University of Korea, Center for Rheumatic Disease in Kang-Nam St. Mary s Hospital, # 505 Banpo-Dong, Seocho-Ku, Seoul , Korea. rheuma@cmc.cuk.ac.kr Received on January 7, 1999; accepted in revised form on April 13, Copyright CLINICAL AND EXPERIMENTAL RHEUMATOLOGY Introduction The overall mortality rate from systemic lupus erythematosus (SLE) has improved significantly over the past 50 years due to various treatment strategies including immunosuppressive agents (1-6). Most fatalities in SLE are attributable to active SLE, especially renal and central nervous system (CNS) involvement, and to infections. Several studies have demonstrated that the improvement in survival has been accompanied by a shift in the predominant causes of death, with fewer being due to SLE itself and a greater number being attributable to infection (2, 4, 5, 7-9). There are several recent reports showing that lupus patients have a higher prevalence of coronary heart disease associated with a specific mortality rate (7, 10, 11). In addition, an increased incidence of cancer, particularly non-hodgkin s lymphoma, in SLE has also been observed (12, 13). Collectively, during the past decades there have been considerable changes in the patterns of the causes of death associated with the improvement of survival in SLE. Race has also been documented to be related to mortality in SLE. Blacks show a higher prevalence of SLE than whites and have been reported to be more likely than white patients to die of infection (14, 15). A higher prevalence of SLE has also been noted among Asians compared with whites (16). In one study, Asians in England had a worse prognosis than English whites (17). The reasons for the differing morbidity and mortality among Asian SLE patients are unknown, but may be related to differences in exposure to environmental factors (18). Different environmental or genetic factors may also influence on the cause of death in SLE, but few attempts had been made to analyze the causes of death in Asian SLE patients. The aim of the present study was therefore to determine the causes of death in Korean patients with SLE over a recent 5-year period. Patients and methods This study included 544 SLE patients who were treated at the Lupus Clinic in the Center for Rheumatic Disease in Kang-Nam St. Mary s Hospital from January 1993 to December All patients fulfilled the revised American College of Rheumatology (ACR) criteria (19). A computer search for patients with a hospital disease code for SLE who had died was also carried out. The survival of patients was determined from the date of diagnosis of SLE through January 1, The current status of survivors was established by contacting the patients, their families or their physicians at the end of the study. Disease duration was calculated from the onset of the initial symptoms of SLE to death. Organ involvement was defined on the basis of both clinical and laboratory measures established by the revised ACR criteria (19). Demographic data, clinical manifestations, treatment modalities, and laboratory variables, including a full blood count, serum creatinine, urine microscopy, 24-hour urine protein, serum complements, anti-nuclear antibody (immunofluorescence method), and anti-dna antibody (radioimmunoassay), were recorded at presentation, and at death in all patients who died during the study period. Information regarding the cause of death was available for all 40 patients. To determine the cause of death in these SLE patients, detailed clinical records were investigated. A death was considered to be due to SLE if it was associated with active SLE or SLE-related organ damage. Active SLE was determined on the basis of the clinical manifestations, laboratory findings including a decrease in complement levels, an increase in antids DNA antibodies titers, and an increase in the SLE disease activity index (SLE- DAI) (20) compared to previous results. The causative microorganisms of infection were examined thoroughly by culturing and special stainings of biopsy material. The diagnosis of pulmonary hypertension was based on findings from echocardiograms and cardiac catheterization (systolic/diastolic pressure of the pulmonary artery 30/15 mmhg) after the exclusion of cases of pulmonary embolism, left side heart failure, and secondary pulmonary hypertension due to interstitial lung disease on high resolution CT and pulmonary function tests. Acute pulmonary hemorrhage syndrome was defined as pulmonary infiltrates together with an 540

3 Hypothalamus-pituitary-adrenocortical and -gonadal axis in RA / M. Cutolo Causes of death in Korean patients with SLE / W.-U. EDITORIAL Kim et al. acute fall in the hemoglobin level of 1.5 g/dl without any apparent blood loss or hemolysis, and the presence of more than three of the following; hemoptysis, dyspnea, bleeding on bronchoscopy, hemosiderin-laden macrophages on bronchoalveolar lavage, elevated diffusion capacity (DLCO) on the pulmonary function test, or post mortem evidence of pulmonary hemorrhage. Acute myocarditis was defined as elevated cardiac enzyme levels, diffuse decrease in wall motion on echocardiogram, and normal coronary artery on angiogram. Thrombotic thrombocytopenic purpura (TTP) was diagnosed from a combination of clinical, laboratory, and gum biopsy findings. These findings included fever, thrombocytopenic purpura, microangiopathic hemolytic anemia, fluctuating neurologic findings, renal dysfunction, elevated LDH and indirect bilirubin, reticulocytosis, normal fibrin degradation product and fibrinogen, and intravascular microthrombi on biopsy. Data was analyzed using the SPSS program (version 7.0). The chi-square test and the T-test were used for the comparisons between the two groups. Survival probabilities were determined by Kaplan-Meier life table analysis. In the case of data sets that were not normally distributed, comparisons of numerical data between groups were performed by the Mann-Whitney U test. P values less than 0.05 were considered significant. Results As of January 1, 1997, 453 (83.3%) of the 544 SLE patients who were monitored in this study were still alive, 43 (7.9%) had died, and 48 (8.8%) were lost to follow-up. Forty died at our hospital and three died at home. Among the patients who died, 3 cases were excluded because the SLE had not been diagnosed within the last 2 years or the medical records at diagnosis were incomplete. Therefore the medical records of 40 patients were analyzed for our study. Comparisons of the demographics and disease activity indices between the deceased and the survivors showed that the age was higher and C3 at presentation was lower in the deceased than the survivors (age: 33.8 ± 13.6 versus 28.3 ± 10.6 years, p = 0.02; C3: 36.8 ± 21.4 versus 49.7 ± 20.8 mg/dl, p = 0.03). However, no differences were found in sex, disease duration, C4 and anti-ds DNA titers at presentation between the two groups (Table I). A life table analysis demonstrated a survival rate of 97.9% at one year, 97.2% at 2 year, and 93.7% at 5 years. The frequency and causes of death in this cohort were as follows: 13 from infection (32.5%), 10 SLE-related deaths (25.0%), 6 pulmonary hypertension (15.0%), 4 cerebrovascular accidents (10.0%), 3 TTP (7.5%), one peptic ulcer perforation after kidney transplantation (2.5%), one ARDS associated with massive hemorrhage (2.5%) and one cor pulmonale due to silicosis (2.5%). One patient had a severe multi-system illness, and death could not be attributed to predominant involvement of any one organ system (Table II). Infection was the most common cause of death. Most infections manifested as either bacteremia or sepsis syndrome (76.9%). The major causative microorganisms were gram negative bacilli (69.2%) including Pseudomonas (1 case), Klebsiella (2), Serratia (1), Salmonella (1), Escherichia coli (1), Enterobacter (1), and fungi (23.1%) such as Candida (2) and Aspergillus (1). The main focus of infection was the lung (53.8%) (Table III). The patients (n = 13) who died from infection had lower levels of complement 3 (C3) and higher levels of anti-ds DNA antibody at presentation than those (n = 27) who died from other causes (C3: 24.7 ± 17.8 versus 41.7 ± 21.5 mg/dl, p = 0.02, anti-ds DNA an- Table I. Demographics and disease activity indices at presentation between the survivors and the deceased. Deceased Survivors P (n = 40) (n = 453) Sex (F/M) (number) 39/1 432/21 NS Disease duration (months) 46.6 ± ± 23.6 NS Age (years) 33.8 ± ± C3 (mg/dl) 36.8 ± ± C4 (mg/dl) 14.0 ± ± 12.3 NS Anti-ds DNA antibody (IU) 54.2 ± ± 59.6 NS Table II. Causes of death in 40 SLE patients who died during the study period. Causes No. of deaths % Infection SLE related death Cerebral nervous system disease 3 Thrombotic thrombocytopenic purpura 2 Acute pulmonary hemorrhage syndrome 2 Nephritis 1 Acute myocarditis 1 Multi-system* 1 Pulmonary hypertension Cerebrovascular accident Thrombotic thrombocytopenic purpura Peptic ulcer perforation after kidney transplantation ARDS associated with massive hemorrhage Cor pulmonale due to silicosis Unknown *One patient had a severe multi-system illness, so death could not be attributed to predominant involvement in any one organ system. One patient died of unknown causes while being operated on for avascular necrosis in the femoral head. 541

4 Causes of death in Korean patients with SLE / W.-U. Kim et al. tibody: 68.0 ± 73.5 versus 27.0 ± 35.3 IU, p = 0.04). Mean steroid dose used one month before death was higher in patients who died of infection (30.5 ± 15.2 versus 15.2 ± 7.7 mg/day, p = 0.03) (Table IV). SLE-related factors were the next most important cause of death in the present study. Among SLE-related causes, cerebral nervous system disease was responsible for the death of 3 patients, who died of status epilepticus (1 case) and cerebrovascular accident attributed to active SLE (2 cases), TTP (2 patients), acute pulmonary hemorrhage syndrome (2 patients), and nephritis (1 patient). One patient died of SLE-related cardiac death, specifically congestive heart failure following acute myocarditis. Pulmonary hypertension was the third most common cause of death (n = 6) in the present study. The duration from the diagnosis of SLE to pulmonary hypertension was significant longer than that from pulmonary hypertension to death (median [range]: 53.5 [12-78] versus 2.5 [0.5-15] months, p = 0.01). The median pulmonary systolic and diastolic pressures at initial diagnosis were 84 mmhg (69-100) and 34 mmhg (25-40), respectively (Table V). None of these six patients had any respiratory symptoms until the diagnosis of pulmonary hypertension. The patients who died of pulmonary hypertension were characterised by less nephritis (50% versus 88.2%, p = 0.023) and CNS disease (16.7% versus 61.8%, p = 0.041) at death, than those who died of other causes (n = 32). In addition, they showed normal complement levels and anti-ds DNA antibody titers and had no clinical evidence of a disease flare except for the respiratory symptoms at death (5 cases) Table III. Causative microorganisms in 13 SLE patients who died from infection. Type or site No. (%) Microorganism (no.) Focus of sepsis Bacteremia/sepsis 10 (76.9%) Pseudomonas aeruginosa (2) Lung (1) Enteritis (1) Klebsiella oxytica (1) Peritonitis (1) Klebsiella pneumoniae (1) Lung (1) Serratia marcescens (1) Lung (1) Salmonella typhi (1) Enteritis (1) Escherichia coli (1) Pyelonephritis (1) Enterobacter cloacae (1) Enteritis (1) Staphylococcus aureus (1) Skin abscess (1) Mycobacterium tuberculosis (1) Lung (1) Pneumonia 3 (23.1%) Aspergillus fumigatus (2) Candida albicans (1) Table IV. Comparison of clinical and laboratory variables between patients who died from infection and those who died from other causes. Death Death from infection from other causes P (n = 13) (n = 27) Disease duration (months) 37.9 ± ± 26.3 NS Age (years) 38.2 ± ± 13.2 NS C3, mg/dl) 24.7 ± ± C4 (mg/dl) 14.3 ± ± 11.1 NS Anti-ds DNA antibody (IU) 68.0 ± ± Previous dose of prednisone* (mg/day) 30.5 ± ± % of cytotoxic agents used previously NS * Mean prednisone dose one month before death. Cytotoxic drugs comprised azathioprine and cyclophosphamide. NS = not significant. or 1 month before death (one case). These reasons led us to categorize death due to pulmonary hypertension as non- SLE related. Five patients (12.5%) died from TTP despite aggressive treatment with systemic corticosteroid, plasma infusion, and plasmapheresis. TTP was associated with active SLE in two cases. Even though many of the clinical features in SLE are similar to those seen in TTP, SLE could be differentially diagnosed in all cases by the presence of anti-dna and/or anti-sm antibody in addition to the aforementioned revised ARA criteria. Actually, it was not difficult for us to discriminate SLE from TTP, since the TTP was found after the SLE had been diagnosed in all cases. Four patients died of cerebrovascular accidents (2 from massive cerebral infart and 2 cerebral hemorrhage) which were not associated with active SLE. Another patient died of peptic ulcer perforation one month after kidney transplantation, one died of adult respiratory distress syndrome after massive hemorrhage due to placenta previa. One patient died of unknown causes during an operation for avascular necrosis of the femoral head. Discussion The survival rate in SLE has recently increased to up to 90% at 5 years and 80% at 10 years compared to the 51% cumulative survival rate at 4 years in the 1950s (1-6). The presence of nephritis, seizure, thrombocytopenia, an older age and high disease activity at presentation are known to be negative prognostic factors for mortality in SLE (6, 21). Race, on the other hand, appears to be a significant predictor of survival. Blacks have a poorer prognosis than whites and a higher frequency of renal involvement (3, 5, 22). In 544 Korean lupus patients monitored in a single center, the 5-year survival rate was 93.7%, comparable to that in previous reports. The deceased patients tended to be older, with lower complement levels and higher anti-ds DNA titers, which confirms previous reports that older age and higher disease activity at presentation are associated with a poorer outcome. The relative proportions of infection versus SLE-related deaths varied consider- 542

5 Hypothalamus-pituitary-adrenocortical and -gonadal axis in RA / M. Cutolo Causes of death in Korean patients with SLE / W.-U. EDITORIAL Kim et al. Table V. Clinical and physiologic characteristics of patients who died of pulmonary hypertension. Onset age Duration Duration NYH of SLE 1 * 2 class PAPsys PAPdia DLCo Final event before death Patient % Right side heart failure Patient % Sudden death at home Patient % Arrhythmia Patient % Right side heart failure Patient % Right side heart failure Patient % Sudden death at home * Months from the diagnosis of SLE to of pulmonary hypertension; months from the diagnosis of pulmonary hypertension to death; pulmonary arterial systolic pressure (mmhg) by cardiac catheterization; pulmonary arterial diastolic pressure (mmhg) by cardiac catheterization; diffusion capacity on pulmonary function test. ably among the studies. Reports show death to be attributable to active SLE in 20-52% of patients, the primary organ involvement leading to death being renal and CNS (3, 10, 11, 23, 24), and to infection in 21-32% (2, 7, 25). The diversity of these results is thought to be due to variations in patient selection, the definition of variables, disease duration, sex, race, and socio-economic status. Recently, death due to infection has increased, being the most common cause of death in some series (25). Blacks are reported to be more likely than white patients to die of infection (14, 15). In one study in an Oriental population, death due to infection was relatively common compared to previous reports, accounting for 40.3% of the total mortality (26). Previous findings suggest that racial differences might have an influence on the causes of death in SLE. In the present study, infection was the most common cause of death, accounting for 34.2%, and SLE-related factors was the next most common cause. Death due to nephritis, the main cause of SLErelated death in Caucasians (10, 11, 23, 25), was rarely found in this study, whereas non-renal causes such as CNS disease, pulmonary disease, and TTP were leading causes of death related to active SLE. It is not clear in the present study whether these discrepancies are due to racial difference or not. Considering that the deceased patients showed higher disease activity at presentation compared to the survivors, it is feasible that more aggressive treatment with immunosuppressive agents was being applied in the deceased compared to the survivors. With the availability of dialysis and aggressive therapies, it is to be expected that infection might be more common, whereas nephritis would no longer be the leading cause of mortality in SLE. The most frequent sites for bacterial infection in patients with SLE are similar to those in individuals without lupus and, in general, gram-positive cocci and gram negative bacilli have been most often implicated in the infectious causes of death (27-29). Bacteremia is common, especially in hospitalized patients with SLE (30). Infections in SLE are usually associated with disease exacerbation. The ESR and the overall disease activity, as measured by SLEDAI, correlate well with the incidence of infection (31, 32). Under such circumstances, it may be difficult to separate the effects of active lupus from the effects of treatment with steroid or immune suppressive agents. In addition, it has been reported that treatment with steroid and cytotoxic agents as well as active lupus itself may play an important role in the susceptibility of lupus patients with infection (26, 31, 33). In the present study, sepsis and bacteremia were common (76.9%) and gram negative bacilli were the main (69.2%) causative microorganism, which contrasts with the primary organisms of infection in patients with SLE in Western series (10, 27-29). Our findings that deceased patients had lower complement levels and higher anti-ds DNA titers at presentation and took higher doses of steroid compared to those who died from other causes suggest that infectious death was closely linked to both disease activity and the dose of steroid used in this cohort. These findings, together with previous reports, provide persuasive evidence for the predominant proportion of bacteremia with gram negative bacilli in our Asian patients compared to the gram positive cocci in Western patients. Since active SLE often cannot be distinguished from infection and immunosuppressive treatment may lead to infectious death, immunosuppressive agents should not be excessively administered if the possibility of infection, especially gram negative infection, cannot be thoroughly ruled out. Pulmonary hypertension in SLE is reported to be not uncommon ( 10% in SLE), but is usually mild in degree (34, 35). Death due to pulmonary hypertension was very rare in several Western studies, accounting for less than 1% of the total (2, 5, 10, 11, 23, 25). In our study, death due to pulmonary hypertension was relatively high (15.7%). We do not know whether this higher proportion in our series was due to a change in the disease pattern of SLE or to racial differences. The duration from the diagnosis of pulmonary hypertension to death was much shorter compared to that from the diagnosis of SLE to pulmonary hypertension. The severity of the pulmonary hypertension at the initial diagnosis was very high. Theses findings suggest that pulmonary hypertension may remain unrecognized for a long period of time in those patients who eventually die from it. This could explain the relatively higher proportion of deaths due to pulmonary hypertension in our study, even 543

6 Causes of death in Korean patients with SLE / W.-U. Kim et al. though we cannot exclude the possibility of a higher prevalence of pulmonary hypertension in Korean lupus patients. It should be noted here that pulmonary hypertension can often go undetected, so that its presence should be monitored periodically, especially in Asian SLE patients. Death due to coronary heart disease was reported to underlie 6-23% of the deaths in several Western studies of SLE (7, 10, 11, 36), whereas it was absent among our patients. The prevalence of coronary heart disease in general is low in Korea, which may be explained by the low cholesterol diet associated with socioeconomic factors (37, 38). While we did not measure the levels of cholesterol in our patients, the low cholesterol diet of the average Korean could be one of the reasons for the absence of death due to coronary heart disease in this study. It has been documented that lupus patients have an increased risk of malignancy (10, 12, 13). Breast, lung, and gynecological tumors are the most common malignancies and breast cancer is significantly increased in Caucasian women (13). In contrast, the standardized incidence ratio for all cancer in SLE is only 1.4 compared with that expected in one series (39). In the present study, we did not find any deaths due to neoplasm among Korean lupus patients. Death due to neoplasm has been reported more frequently in studies with longer follow up periods. Since the follow-up period was relatively short in the present study, we can not conclude that the rarity of death due to neoplasm is a characteristic finding in Korean lupus patients. Further long-term study is needed to clarify this issue. We have demonstrated several important findings which contrast with the observations in Western studies. Compared to the largest recent study in Caucasian lupus patients (n = 144) by Ward et al. (10), the frequencies of death due to nephritis, infection, or CNS disease were not different, whereas the frequencies of death due to pulmonary hypertension and TTP were significantly higher in our lupus patients (pulmonary HBP: 15.0% (6/ 40) versus 0.7% (1/144), p < 0.001, TTP: 12.5% (5/40) versus 0.7% (1/144), p < 0.001). However, we did not directly compare the causes of death between Korean and Western lupus patients within one population and did not consider the influence of socioeconomic status, which may be associated with racial differences (10, 22). In addition, this study is limited by its retrospective nature. Therefore, we cannot conclude here that racial differences in the cause of death are evident between Asian and Western lupus patients. A prospective, comparative study of an inception cohort would be required in this respect. In summary, the most common cause of death in our cohort of 544 Korean lupus patients was infection, mainly manifesting as gram negative bacterial sepsis. SLE-related factors (mostly non-renal) were the next most common cause. Infectious death was associated with a higher disease activity at presentation and higher steroid dose used just before death. Death due to pulmonary hypertension was common, whereas death due to coronary heart disease and neoplasm was absent. Acknowledgements The authors wish to thank to Teta and Bill for correcting this manuscript. The authors are also so grateful to Dr. Jeong- Ah Shin for typing this manuscript. References 1. MERRILL M, SCHULMAN LE: Determination of prognosis in chronic disease, illustrated by systemic lupus erythematosus. J Chron Dis 1955; 1: UROWITZ MB, BOOKMAN AA, KOEHLER BE, GORDON DA, SMYTHE HA, OGRYZLO MA: The bimodal mortality pattern of systemic lupus erythematosus. Am J Med 1976; 60: GINZLER EM, DIAMOND HS, WEINER M, et al.: A multicenter study of outcome in systemic lupus erythematosus. 1. Entry variables as predictors of prognosis. Arthritis Rheum 1982; 25: JONSSON H, NIVED O, STURFELT G: Outcome in systemic lupus erythematosus: A prospective study of patients of patients from a defined population. Medicine 1989; 68: REVEILLE JD, BARTOLUCCI A, ALARCON GS: Prognosis in systemic lupus erythematosus: Negative impact of increasing age at onset, black race, and thrombocytopenia, as well as causes of death. Arthritis Rheum 1990; 33: WARD MM, PYUN E, STUDENSKI S: Mortality risk associated with specific clinical manifestations of systemic lupus erythematosus. Arch Int Med 1996; 156: WALLACE DJ, PODELL T, WEINER J, KLINENBERG JR, FOROUZESH S, DUBOIS EL: Systemic lupus erythematosus-survival pattern: Experience with 609 patients. JAMA 1981; 245: GUDMUNDSSON S, STEINSSON K: Systemic lupus erythematosus in Iceland : A nationwide epidemiological study in an unselected population. J Rheumatol 1990; 17: HELLMANN DB, PETRI M, WHITING-O KEEFE Q: Fetal infections in systemic lupus erythematosus: The role of opportunistic organism. Medicine (Baltimore) 1987; 66: WARD MM, PYUN E, STUDENSKI S: Causes of death in systemic lupus erythematosus. Arthritis Rheum 1995; 38: ROSNERS S, GINZLER EM, DIAMOND HS: A multicenter study of outcome in systemic lupus erythematosus: II. Causes of death. Arthritis Rheum 1982; 25: PETTERSSON T, PUKKALA E, TEPPO L, FRIMAN C: Increased risk of cancer in patients with systemic lupus erythematosus. Ann Rheum Dis 1992;51: RAMSEY-GOLDMAN R, MATTAI SA, SCHILL- ING E, et al.: Increased risk of malignancy in patients with systemic lupus erythematosus. J Invest Med 1998; 46: OTTEN MW, TEUTSCH SM, WILLIAMSON DF, MARKS JS: The effect of known risk factors on the excess mortality of black adults in the United States. JAMA 1990; 263: MANTON KG, PATRICK CH, JOHNSON KW: Health differentials between blacks and whites: Recent trends in mortality and morbidity. Milbank Q 1987; 65 (Suppl. 1): SERDULA MK, RHOADS GG: Frequency of systemic lupus erythematosus (SLE) in different ethnic groups of Hawaii. Arthritis Rheum 1979; 22: SAMANTA A, FEEHALLY J, ROY S, NICHOL FE, SHELDON PJ, WALLS J: High prevalence of systemic disease and Asian subjects with systemic lupus erythematosus. Ann Rheum Dis 1991; 50: LIANG M, PARTRIDGE AJ, DALTROY LJ, STRATTON KV, GALPER SR, HOLMAN HR: Strategies for reducing excess morbidity and mortality in blacks with systemic lupus erythematosus. Arthritis Rheum 1991; 34: TAN EM, COHEN AS, FRIES JF: The 1982 revised criteria for the classification of systemic lupus erythematosus. Arthritis Rheum 1982; 25: COHEN AS, CANOSO JJ: Criteria for classification of systemic lupus erythematosus. Arthritis Rheum 1972; 15: ABU-SHAKRA M, UROWITZ MB, GLADMAN DD, GOUGH J: Mortality studies in systemic lupus erythematosus. Results from a single center. Il. Predictor variables for mortality. J Rheumatol 1995; 22: STUDENSKI S, ALLEN NB, CALDWELL DS, RICE JR, POLISSON RP: Survival in systemic lupus erythematosus. A multivariate analysis of demographic factors. Arthritis Rheum 1987; 30: HELVE T: Prevalence and mortality rates of systemic lupus erythematosus and causes of 544

7 Hypothalamus-pituitary-adrenocortical and -gonadal axis in RA / M. Cutolo Causes of death in Korean patients with SLE / W.-U. EDITORIAL Kim et al. death in SLE patients in Finland. Scand J Rheumatol 1985; 14: SWAAK AJ, NOSSENT JC, BRONSVELD W et al.: Systemic lupus erythematosus: 1. Outcome and survival: Dutch experience with 110 patients studied prospectively. Ann Rheum Dis 1989; 48: ABU-SHAKRA M, UROWITZ MB, GLADMAN DD, GOUGH J: Mortality studies in systemic lupus erythematosus. Results from a single center. I. Causes of death. J Rheumatol 1995; 22: KOH ET, SEOW A, LEONG KH, CHNG HH: SLE mortality in an oriental population. Lupus 1997; 6: STAPLES PJ, GERDING DN, DECKER JL, GORDON RS JR: Incidence of infection in systemic lupus erythematosus. Arthritis Rheum 1974; 17: NIVED O, STURFELT G, WOLLHEIM F: Systemic lupus erythematosus and infection: A controlled and prospective study including an epidemiological group. Q J Med 1985; 55: ABRAMSON S, KRAMER SB, RADIN A, HOLZMAN R: Salmonella bacteremia in systemic lupus erythematosus. Eight-year experience at a municipal hospital. Arthritis Rheum 1985; 28: LEE P et al.: Systemic lupus erythematosus. A review of 110 cases with reference to nephritis, the nervous system, infections, asceptic necrosis and prognosis. Q J Med 1977; 66: PETRI M, GENOVESE M: Incidence of and risk factors for hospitalizations in systemic lupus erythematosus: A prospective study of the Hopkins Lupus Cohort. J Rheumatol 1992; 19: DUFFY KN, DUFFY CM, GLADMAN DD: Infection and disease activity in systemic lupus erythematosus: A review of hospitalized patients. J Rheumatol 1991; 18: KATTWINKEL N, COOK L, AGNELLO V: Overwhelming fatal infection in a young woman after intravenous cyclophosphamide therapy for lupus nephritis. J Rheumatol 1991; 18: SIMONSON JS, SCHILLER NB, PETRI M, HELLMANN DB: Pulmonary hypertension in systemic lupus erythematosus. J Rheumatol 1989; 16: BADUI E, GARCIA-RUBI D, ROBLES E et al.: Cardiovascular manifestations in systemic lupus erythematosus. Prospective study of 100 patients. Angiol 1985; 36: PETRI M, SPENCE D, BONE LR, HOCHBERG MC: Coronary artery disease risk factors in the Johns Hopkins lupus cohorts: Prevalence, recognition by patients, and preventive practices. Medicine 1992; 71: TCHAI BS: Changes of plasma lipids and biochemical markers for coronary artery disease in Korea. Ann NY Acad Sci 1993; 676: KESTELOOT H, LEE CS, PARK HM et al.: A comparative study of serum lipids between Belgium and Korea. Circulation 1982; 65: SWEENEY DM, MANZI S, JANOSKY J, et al.: Risk of malignancy in women with systemic lupus erythematosus. J Rheumatol 1995; 22: