Clin Rheumatol (2011) 30:1463–1469 DOI 10.1007/s10067-011-1826-y

ORIGINAL ARTICLE

Severe systemic lupus erythematosus in emergency department: a retrospective single-center study from China Yi Chen & Guang-liang Chen & Chang-qing Zhu & Xiaoye Lu & Shuang Ye & Cheng-de Yang

Received: 12 April 2011 / Revised: 23 July 2011 / Accepted: 28 July 2011 / Published online: 13 August 2011 # Clinical Rheumatology 2011

Abstract This study describes the characteristics of severe systemic lupus erythematosus (SLE) patients who visited the emergency department of the Shanghai Renji Hospital, and the aim is to identify the causes, the outcome, and the prognostic factors. SLE patients who visited the emergency department between January 2007 and August 2010, and who were subsequently hospitalized or who died in emergency department, were included in this retrospective study. Of the total 131 SLE emergency events, overall mortality was 16.8%. Older age (≥45 years), longer duration of disease, presence of pulmonary hypertension, renal insufficiency, and invasive infections are risk factors. Higher organ damage index (SLICC, 3.86±2.14 vs. 0.93± 1.14, p<0.001) but relatively lower disease activity (SLEDAI, 11.5±7.7 vs. 16.5±9.0, p=0.015) were the characteristics of the deceased when compared with the survivors. Recent-onset (duration of disease ≤3 months) SLE accounts for 32.1% of the patients in emergency, and this group showed a distinctive pattern of younger age with higher frequencies of neuropsychiatric and hematologic manifestations. A good short-term outcome with a 95.2% survival

Yi Chen and Guang-liang Chen contribute equally to this work. Y. Chen : C.-q. Zhu (*) : X. Lu Emergency Department, Renji Hospital, Shanghai JiaoTong University School of Medicine, 1630 Dongfang Rd, Shanghai 200127, China e-mail: [email protected] G.-l. Chen : S. Ye : C.-d. Yang (*) Department of Rheumatology, Renji Hospital, Shanghai JiaoTong University School of Medicine, 145 Shandong C Rd, Shanghai 200001, China e-mail: [email protected]

rate can be observed in such patients. Older age, longer disease duration and cumulative damage of vital organs determine the short-term outcome of severe SLE in the emergency department. Recognizing disease patterns and objective prognostic parameters may help physicians to provide better care, based on a risk-fitted approach, for the SLE patients in emergencies. Keywords Emergency . Prognosis . Systemic lupus erythematosus

Introduction Systemic lupus erythematosus (SLE) is a systemic autoimmune disease that may require visits to the emergency department (ED) as the disease can be acute and severe. Previous reports indicated that patients with older age and lower socioeconomic status were more likely to use ED when required, instead of scheduled outpatient visits with a rheumatologist or a primary care physician [1, 2]. Further analyses on a group of SLE patients making frequent ED visits revealed that depression, poor socioeconomic support, problems in medication adherence, and active disease are interdependent factors that lead to the “come-and-go” pattern among those patients [3, 4]. Notably, patients who frequently visit ED do not necessarily suffer from severe SLE or lifethreatening emergencies. Therefore, it is important for both the emergency physician and the on-call rheumatologist to identify those patients with poor prognosis well on time. On the other hand, the experience of the emergency physician regarding the acuity of diagnosis and the appropriateness of interventions has a strong impact on the prognosis. A review of the charts of the patients

1464

admitted via the ED revealed that more than 80% failed to include a reference to the musculoskeletal system. In the survey among junior doctors, 75% felt unable to diagnose a connective tissue disorder [5]. Another report based on interhospital transfer settings indicated that the mortality was 2.5 times higher when the physician at the receiving hospital was “not experienced” (by their definition, a physician who has treated three or fewer patients with SLE per year) [6]. In such circumstances, it is a challenge for ED to be the first place to make a diagnosis of SLE. The emergency physician should be vigilant to those with typical or atypical presentations and also develop the “know-how” to deal with such a condition. This study focuses on patients with severe SLE who seek medical attention in ED. It recognizes disease patterns and prognostic factors, and provides objective parameters to guide decision making for the SLE patients in ED.

Patients and methods Patient records between January 2007 and August 2010 were retrospectively derived from the emergency department of the Shanghai Renji Hospital, which is a 1,500-bed University Hospital and a tertiary referral center. The ED of the hospital had an average of 150,000 patient visits per year during the 3-year interval. The inclusion criteria for this study are as follows: (1) patients who fulfilled the American College of Rheumatology criteria for SLE [7] and (2) patients who were subsequently hospitalized or died in ED. In-patient charts were tracked and reviewed by a rheumatologist. The documented demographic data included gender and age. Educational status was classified as having completed less than high school, high school or vocational school, and college or higher. The following details were also recorded: whether the patient was from Shanghai or some other place, lived in an urban or a rural area, and had insurance coverage. Duration of disease was calculated as the number of months since the documented SLE diagnosis. Recent-onset SLE was defined as duration of disease <=3 months. The emergency presentations were classified according to the systems that were affected (1) Constitutional— fever, fatigue, and weight loss; (2) Mucocutaneous and musculoskeletal—skin rash, joint pain or swelling, hair loss, mouth ulcers, and Raynaud phenomenon; (3) Cardiopulmonary—cough and expectoration, hemoptysis, dyspnea, chest pain, palpitations, and syncope; (4) Digestive—abdominal pain and distension, nausea, vomiting, diarrhea, and jaundice; (5) Central nervous system—disturbance of consciousness or coma, seizure, paralysis, headache, and psychological symptoms; (6)

Clin Rheumatol (2011) 30:1463–1469

Urinary—oliguria, proteinuria or hematuria, and generalized edema; (7) Hematologic—bleeding tendency, thrombocytopenia, anemia, and leukopenia; (8) Miscellaneous—obstetric and gynecological problems. The clinical course and the outcome including duration of stay in the emergency department, subsequent hospitalization, and the direct causes of death were documented. SLE disease activity index (SLEDAI) [8] and systemic lupus international collaboration clinics damage index (SLICC/DI) [9] were available from the inpatient chart, and for those who died in ED, from the Emergency Resuscitation Unit chart. Specific diagnosis such as pulmonary hypertension (PH) was established when a Doppler echocardiograph showed systolic pulmonary artery pressure was greater than 40 mmHg [10], with a left ventricular ejection fraction >0.5. Pulmonary hypertension due to thromboembolism or interstitial lung disease had been ruled out. Neuropsychiatric SLE [11] and antiphospholipid syndrome [12] were diagnosed according to international criteria. The term “gastrointestinal (GI) vasculitis” was used to encompass the constellation of intestinal ischemia, ileus, with or without pancreatitis due to SLE [13]. Renal insufficiency was defined as serum creatinine>=133 μmol/L. Only invasive infections needing admission to hospital or intravenous antibiotics were analyzed, which excluded upper respiratory, lower urinary tract infections, superficial candidiasis, and herpes zoster infection. The causes of ED visits, specific diagnosis, and the causes of death were retrospectively re-evaluated by both an emergency physician and a rheumatologist with the agreement achieved. The study protocol was approved by the institutional review board of the Renji Hospital. Statistical analysis All numerical data were presented using means±SD. Oneway analysis of variance, chi-square test, or Fisher's exact test was employed to compare between the groups. COX regression and Kaplan–Meier curve were used for survival analysis. P<0.05 was considered as significant. All analyses were performed using the SPSS 13.0 software.

Results There were 125 SLE patients with 131 events in the emergency department that fulfilled the criteria mentioned earlier. Six patients had two visits during the investigation. The median interval between the two visits was 10 months (range 1–19 months). For patients with a shorter interval, for example, the patient who had two ED events 1 month apart, the first event was due to

Clin Rheumatol (2011) 30:1463–1469

active SLE and the second was septic meningitis. Under case-specific analysis, plus the small proportion of such cases, the authors believed that it is fair and safe to consider those as independent events. Two of the six patients died in their second ED events, and of note, both due to pulmonary hypertension. Major complaints presented in ED The major complaints presented in ED were classified according to the systems involved. Fever was the most frequent manifestation in 39.7% (52) of the patients, followed by cardiopulmonary in 35.9% (47), central nervous system in 29.8% (39), digestive in 26.0% (34), mucocutaneous and musculoskeletal in 16.0% (20), urinary in 6.9% (9), hematologic in 5.3% (7), obstetric and gynecological in 1.5% (2). As a comparison, the deceased patients had more cardiopulmonary manifestations (72.7%, 16 cases), followed by fever (22.7%, 5 cases), central nervous system and digestive system (both 18.2%, 4 cases for each), mucocutaneous and musculoskeletal (9.1%, 2 cases), and urinary (4.5%, 1 case) manifestations. Risk factors for short-term survival: univariate analysis The overall mortality was 16.8% (22/131). The deceased patients were older with a longer duration of disease (Table 1). There was no difference in the patients' place of residence and insurance coverage between the two groups. There were 44.3% (58/131) of patients from Shanghai who were covered by the government-sponsored insurance (Shanghai insurance). These SLE patients, although with better access to medical facilities and better financial reimbursements, showed no difference in terms of EDrelated survival in this study. The time of the ED stay before the in-patient hospitalization did not seem to be relevant to the outcome. Vital organ damage instead of disease activity is conceivably being related to mortality. The deceased had a higher SLICC index but a relatively lower SLEDAI when compared with the patients who survived. To be more specific, kidney is known to be one of the major target organs in SLE. Lupus nephritis (LN) presented in 39.7% (52/ 131) of the ED events. However, even if LN is absent, renal insufficiency becomes a prominent risk factor. Of the 21 patients (16.0%) with renal insufficiency, 3 showed acute kidney injury without a previous history of LN, and 2 of the 3 died. Another three had end-stage renal disease and underwent regular hemodialysis or peritoneal dialysis (>=3 months). One of them died from peritoneal infection and sepsis related to peritoneal dialysis. Overall, renal insufficiency contributed to mortality in 40.9% (9/22) of the SLE ED events.

1465

Strikingly, PH is the most important prognostic factor for short-term survival in this study. Of the 18 events (13.7%, 18/131) which were PH-related, 10 died, due to congestive heart failure, hemodynamic decompensation, or had cardiac sudden death. There were seven patients who died within 24 h in ED, and PH was responsible for five. Although PH accounted for 45.5% (10/22) of the overall mortality, the Doppler-estimated systolic pulmonary artery pressure among those deceased PH (83.04±30.72 mmHg) was not different from those who survived PH (79.83±24.26 mmHg, p=0.816). Infection was another common cause of the SLE ED visits with a prevalence of 32.8% (43/131) in this study. Pneumonia (79.1%, 34/43) was the most frequent infection, along with other less frequent syndromes, such as sepsis (9.3%, 4 cases), intra-abdominal (7.0%, 3 cases), and intracranial (4.7%, 2 cases) infections. Severe infection could be lethal and contribute, at least to some extent, to 59.1% (13/22) of the overall mortality. Of note, a relatively higher white blood cell count (WBC) was seen among the deceased. This may, in some cases, reflect an infection response, rather than active SLE, which typically presented as leukopenia. Indeed, invasive infection patients had a higher WBC when compared with those without invasive infections (8.24±4.50 vs. 6.65±4.78×109/L, p=0.029). As a comparison, NPSLE (18.3%, 24/131), although dramatically manifested as fever, seizures, confusion, or even coma, after appropriate differential diagnostic procedures, e.g., imaging study and CSF analysis, showed that its responsiveness to aggressive immunosuppressive therapy was generally good. A similar short-term outcome could be appreciated among those patients with GI vasculitis (11.5%, 15/131), which is consistent with a previous report from South Korea [13]. None of the 15 patients required surgical interventions. Hematologic involvements in the SLE ED patients were also relatively benign. There were no cases of neutropenia or internal bleeding due to thrombocytopenia, although extremely low platelet count (<10,000/mm3) did occur in 4.6% (6/131). The diagnosis of APS was only established among 4.6% (6/131) of the SLE ED patients and is seemingly irrelevant to the prognosis. Risk factors for short-term survival: COX regression and survival curves The parameters that formed the COX regression model reinforced the significance of the following prognostic predictors, i.e., the presence of PH (OR, 5.984; 95%CI, 2.545–14.072), invasive infections (OR, 2.709; 95%CI, 1.131–6.487), and age>=45 years (OR, 2.532; 95%CI, 1.052–6.094). Parameters significant in univariate analysis but not in the COX equation include renal insufficiency, duration of disease, recent-onset SLE, and WBC count.

1466

Clin Rheumatol (2011) 30:1463–1469

Table 1 Characteristics of severe SLE in emergency department and prognostic factors No. of patients (%)

Survived n=109

Age (years) ≥45 years Female gender Level of education Less than high school High school or vocational school College or higher Shanghai insurance Duration of disease (months) Recent-onset SLE (duration of disease ≤3 month) Duration of stay in ED (days) Mucocutaneous/musculoskeletal manifestations Cardiopulmonary manifestations Cytopeniaa

37.28±14.45 37 (33.9%) 97 (89.0%)

48.05±11.53 14 (63.6%) 18 (81.8%)

0.001 0.009 0.472

50 39 20 53

15 3 4 12

0.083

Lupus nephritis Renal insufficiency(sCr≥133umol/L) BUN (mmol/L) sCr (umol/L) Pulmonary hypertension Invasive infection White blood cell count (×109/L) Neuropsychiatric SLE GI vasculitis Antiphospholipid syndrome SLEDAI SLICC a

40 19 31 87

(45.9%) (35.8%) (18.3%) (48.6%) 54.25±61.47 (36.7%) 4.04±3.39 (17.4%) (28.4%) (79.8%)

41 (37.6%) 12 (11.0%) 8.98±7.10 99.93±139.51 8 (7.3%) 30 (27.5%) 6.67±4.2 22 (20.2%) 15 (13.8%) 5 (4.6%) 16.5±9.0 0.93±1.14

Deceased n=22

(68.2%) (13.6%) (18.2%) (54.5%) 93.23±93.06 2 (9.1%) 4.66±4.30 2 (9.1%) 16 (72.7%) 17 (77.3%)

11 (50%) 9 (40.9%) 24.93±14.47 230.14±248.75 10 (45.5%) 13 (59.1%) 9.65±6.3 2 (9.1%) 0 1 (4.5%) 11.5±7.7 3.86±2.14

p value

0.647 0.032 0.012 0.788 0.525 0.000 0.788 0.341 0.002 0.000 0.000 0.000 0.006 0.008 0.364 0.074 1.000 0.015 0.000

Either white cell count <4×109 /L, or hemoglobin <110 g/L, or platelet count <100×109 /L

Survival curves summarized these multivariate and univariate prognostic factors among the SLE ED patients in Fig. 1. The authors intended to use SLEDAI and SLICC as standardized instruments for research purpose, but avoided them as direct prognostic indicators in the analysis. The rationale is that the less time-consuming and operatordependent the instruments are, the better the parameters that are applicable in emergency practice.

and the overall short-term survival was excellent (95.2%). Among those of recent-onset SLE, 45.2% (19/42) only had their diagnosis made during the ED event. Comparing these with other recent-onset but diagnosis-established SLEs showed there was no difference in terms of all the aforementioned parameters.

Discussion Recent-onset SLE in ED Interestingly, 32.1% (42/131) of the ED patients were of recent-onset (≤3 month) SLE (Table 2); the patients tend to be younger, more likely to have come for treatment to Shanghai from other areas; have active disease (higher SLEDAI) but preserved vital system function (lower SLICC); have more NPSLE and hematologic involvement, but less cardiopulmonary complaints and fewer invasive infection complications. These patients generally showed immediate response to high-dose corticosteroid treatment,

SLE is a systemic autoimmune disorder characterized by multiorgan damage with heterogeneous clinical manifestations. Severe SLE could be acute and life-threatening, and such patients are most likely pursuing medical attention in the emergency department. Subsequent hospitalization, which may indicate the appropriateness of the ED visits, was used as a filter to exclude those frequent “come-andgo” ED users. The approach has apparently enrolled the severely ill SLE patients, as evidenced by a high mortality rate at 16.8% in this study.

Clin Rheumatol (2011) 30:1463–1469

A

1467

80 60

Age ≥ 45 ys (n=51)

40

without invasive infection (n=88)

100

Age < 45 ys (n=80)

80

Percent survival

Percent survival

100

p=0.0080

B

p=0.0074

20

60 40

with invasive infection (n=43)

20

0

0 0

10

20

30

40

50

60

0

10

20

Days

C

with pulmonary hypertension (n=18)

20

Percent survival

Percent survival

80

40

60

80 60 40

with renal insufficiency (n=21)

20 0

0 0

10

20

30

40

50

0

60

Days

E

10

20

30

40

50

60

Days

p=0.0060 Recent-onset SLE, disease duration ≤ 3 months (n=42)

100

Percent survival

50

without renal insufficiency (n=110)

100

without pulmonary hypertension (n=113)

60

40

p=0.0109

D

p< 0.0001 100

30

Days

80 60

SLE with disease duration >3 months (n=89)

40 20 0 0

10

20

30

40

50

60

Days Fig. 1 Predictors and survival curves of SLE patients in emergency department

The most frequent presentations in ED were fever (39.7%), followed by manifestations of cardiopulmonary, central nervous, and digestive systems. In a Mexican study of 180 SLE patients seen in the ED, the most frequent manifestations were fever, polyarthralgia, and abdominal pain [14]. The data probably represent relatively mild cases, as only 49 required in-patient hospitalizations. The different ED presentation patterns in this study are likely to be attributed to disease severity; however, racial, regional, and

referral bias cannot be excluded. In addition, the possibility of omission of documented reference to the musculoskeletal system by ED could also bias part of the results [5], although the authors believed that this could largely be overcome by an extensive inpatient chart review in the study. This retrospective study also found out the prognostic factors in the SLE ED patients. According to the data, two interdependent themes actually determined the short-term

1468

Clin Rheumatol (2011) 30:1463–1469

Table 2 Characteristics of recent-onset SLE (duration of disease ≤3 month) in ED No. of patients (%)

Duration of disease ≤3 month n=42

Outcome, survival Age <45(years) Female gender Educational level: less than high school Shanghai insurance Mucocutaneous/musculoskeletal manifestations Cardiopulmonary manifestations Lupus nephritis

40 35 40 25 11 10 7 13

Renal insufficiency Pulmonary hypertension Neuropsychiatric SLE GI vasculitis Cytopeniaa Invasive infection SLEDAI SLICC

4 (9.5%) 2 (4.8%) 14 (33.3%) 7 (16.7%) 38 (90.5%) 7 (16.7%) 20.48±7.41 0.40±0.91

a

Duration of disease >3 month n=89

(95.2%) (83.3%) (95.2%) (59.5%) (26.2%) (23.8%) (16.7%) (31.0%)

69 45 75 38 54 11 40 39

p value

(77.5%) (50.6%) (84.3%) (42.7%) (60.7%) (12.4%) (44.9%) (43.8%)

0.012 0.000 0.090 0.264 0.000 0.126 0.002 0.184

17 (19.1%) 16 (18.0%) 10 (11.2%) 8 (9.0%) 66 (74.2%) 36 (40.4%) 13.47±8.83 1.90±1.84

0.207 0.055 0.004 0.242 0.037 0.009 0.000 0.000

Either white cell count <4×109 /L, or hemoglobin <110 g/L, or platelet count <100×109 /L

outcome of the SLE ED patients, i.e., time effect and vital organ damage. The deceased patients were aged more than 10 years old and had 3 to 4 years of longer duration of disease. Older age is an established poor prognostic factor for SLE [15]. Likewise, 32.1% of the ED patients were of recentonset (≤3 month) SLE and a distinctive disease pattern could be appreciated, i.e., younger age, more CNS, and hematologic involvement. They tend to have active disease, but less target organ damage and infection, with a good short-term prognosis. On the other hand, vital system involvement is not surprisingly associated with prognosis [16]. To be more precise, renal insufficiency is more relevant to mortality, as opposed to LN, NPSLE, GI vasculitis, and hematologic involvement. This study's data underscored that PH is one of the leading risk factors of death, especially presented as sudden death, in the SLE ED patients. In addition, infection can often be the last straw, when the vital system function had already been compromised. The pattern recognition based on the aforementioned aspects may eventually help physicians in the process of time-effective decision making for the SLE patients regarding the diagnosis, triage, and therapy in an emergency setting. Other factors that could be relevant to the SLE ED outcome had been investigated. Lower educational level in the study showed a trend that was associated with mortality, which may reflect the effect of poor compliance. The patients' compliance to treatment had been reported as a crucial parameter linked to prognosis [17, 18]. However, nearly one third of the patients had recent-onset SLE, where compliance would not be a key factor. Among the rest, 37% were transferred from other cities or rural areas. Hence, their previous medical

records cannot be traced or precisely interpreted in terms of compliance in this retrospective study. Concerning the issue of the experience of the emergency physician and its impact on the prognosis, the authors investigated whether shorter exposure in ED and letting a specialist take over sooner would make a difference on outcome. However, this is not the case according to the data. The present study recognizes prognostic factors for severe SLE in ED. The presence of PH, invasive infection, older age (>=45 years), and renal insufficiency are highrisk factors for mortality. The results also suggest that recent-onset SLE in ED shares a distinctive pattern with a good short-term outcome. It is noteworthy that regional and ethnic biases exist in our retrospective observation and largely prohibited generalization of our results. Further studies are needed to validate these findings. Nevertheless, by appropriate training and collaboration, emergency physicians and rheumatologists should work together and provide better care based on a risk-fitted approach for those SLE patients in emergencies.

Disclosures None.

References 1. Yazdany J, Gillis JZ, Trupin L, Katz P, Panopalis P, Criswell LA et al (2007) Association of socioeconomic and demographic factors with utilization of rheumatology subspecialty care in systemic lupus erythematosus. Arthritis Rheum 57:593–600

Clin Rheumatol (2011) 30:1463–1469 2. Ward MM (2008) Avoidable hospitalizations in patients with systemic lupus erythematosus. Arthritis Rheum 59:162–168 3. Panopalis P, Gillis JZ, Yazdany J, Trupin L, Hersh A, Julian L et al (2010) Frequent use of the emergency department among persons with systemic lupus erythematosus. Arthritis Care Res 62:401–408 4. Julian LJ, Yelin E, Yazdany J, Panopalis P, Trupin L, Criswell LA, Katz P (2009) Depression, medication adherence, and service utilization in systemic lupus erythematosus. Arthritis Rheum 61:240–246 5. McCarthy EM, Sheane BJ, Cunnane G (2009) Greater focus on clinical rheumatology is required for training in internal medicine. Clin Rheumatol 28:139–143 6. Ward MM, Odutola JJ (2006) Inter-hospital transfers of patients with systemic lupus erythematosus: characteristics, predictors, and outcomes. J Rheumatol 33:1578–1585 7. Hochberg MC (1997) For the Diagnostic and Therapeutic Criteria Committee of the American College of Rheumatology. Updating the American College of Rheumatology revised criteria for the classification of systemic lupus erythematosus. Arthritis Rheum 40:1725 8. Bombardier C, Gladman DD, Urowitz MB, Caron D, Chang CH (1992) Derivation of the SLEDAI: a disease activity index for lupus patients, the Committee on Prognosis Studies in SLE. Arthritis Rheum 35:630–640 9. Stoll T, Seifert B, Isenberg DA (1996) SLICC/ACR Damage Index is valid, and renal and pulmonary organ scores are predictors of severe outcome in patients with systemic lupus erythematosus. Br J Rheumatol 35:248–254

1469 10. Rich S (ed) (1998) Executive summary from the world symposium on primary pulmonary hypertension, Evian, France, September 6–10, co-sponsored by The World Health Organization 11. No authors listed (1999) The American College of Rheumatology nomenclature and case definitions for neuropsychiatric lupus syndromes. Arthritis Rheum 42:599–608 12. Miyakis S, Lockshin MD, Atsumi T, Branch DW, Brey RL, Cervera R et al (2006) International consensus statement on an update of the classification criteria for definite antiphospholipid syndrome (APS). Thromb Haemost 4:295–306 13. Lee CK, Ahn MS, Lee EY, Shin JH, Cho YS, Ha HK et al (2002) Acute abdominal pain in systemic lupus erythematosus: focus on lupus enteritis (gastrointestinal vasculitis). Ann Rheum Dis 61:547–550 14. Rojas-Serrano J, Cardiel MH (2000) Lupus patients in an emergency unit: causes of consultation, hospitalization and outcome. A cohort study. Lupus 9:601–606 15. Kasitanon N, Magder LS, Petri M (2006) Predictors of survival in systemic lupus erythematosus. Medicine (Baltimore) 85:147–156 16. Bernatsky S, Boivin JF, Joseph L, Manzi S, Ginzler E, Gladman DD et al (2006) Mortality in systemic lupus erythematosus. Arthritis Rheum 54:2550–2557 17. Callahan LF, Pincus T (1990) Associations between clinical status questionnaire scores and formal education level in persons with systemic lupus erythematosus. Arthritis Rheum 33:407–411 18. Petri M, Perez-Gutthahn S, Longenecker JC, Hochberg M (1991) Morbidity of systemic lupus erythematosus: role of race and socioeconomic status. Am J Med 91:345