Temporal trends in characteristics and outcome of intensive care unit patients with systemic lupus erythematosus in Taiwan: a national population-based study

Abstract

This study assessed temporal trends in characteristics and outcome of the intensive care unit (ICU) patients with systemic lupus erythematosus (SLE). We analyzed 2870 SLE patients (≥15 years) with first-time ICU admission between 1999 and 2008. Patient data were retrieved from the Taiwan National Health Insurance Research Database. Median age of the patients increased from 38.5 to 44.0 years during the study period. Most patients were women (overall 85.5%); the proportion varied insignificantly over time. Median duration from SLE diagnosis to ICU admission (disease duration) increased by four years. The proportion of patients with disease duration of more than three years increased from 42.6% to 61.4%, whereas the proportion of patients with a primary rheumatic diagnosis decreased from 41.7% to 25.1%. The incidence of infection increased from 39.1% to 47.2%. The incidence of organ dysfunction increased from 64.1% to 69.2%, especially in neurological, cardiovascular, and respiratory systems. Conversely, the incidence of renal dysfunction decreased from 36.2% to 28.8%. Meanwhile, there were decreasing uses of hemodialysis (from 37.3% to 28.7%) and vasopressors (from 58.3% to 47.9%). Hospital mortality also decreased from 42.6% to 31.2% during the 10 years. After the patient and hospital characteristics were adjusted, patients admitted in 2007 and 2008 were 50% less likely to die than those admitted in 1999 and 2000 (adjusted odds ratio 0.5, 95% confidence interval 0.38–0.67). This study demonstrates a temporal improvement of the short-term survival of the ICU patients with SLE in Taiwan despite changing characteristics and increasing severity of the acute critical illnesses.

Keywords

Systemic lupus erythematosus intensive care outcome

Introduction

Systemic lupus erythematosus (SLE) is an autoimmune disease of unknown cause, featuring episodic flares and thrombophilic and immuno-compromised states,^1–3 which can predispose patients to multi-organ damage and serious infection requiring intensive care.⁴ The five-year survival rate of SLE patients has increased from 50% in the 1950s to over 90% in the 2000s.⁵ The improved survival is achieved at the costs of immunosuppression and accrual of organ damage, which has led to changing morbidity and mortality of these patients.^6,7 Whether similar trends are also present in critically ill SLE patients remains poorly understood. In the last two decades, more than a dozen studies have investigated the characteristics and outcome of critically ill patients with SLE^8–12 or with systemic rheumatic diseases that included SLE.^13–20 The short-term mortality rates in these studies varied widely, ranging from 20% to 53%.^13–20 These studies are limited by experiences from a single or a few institutions and by the inclusion of only a few patients, which is prone to selection bias.

In the present study, we retrieved inpatient data of adult SLE patients having an intensive care unit (ICU) stay from a national administrative database,²¹ used in several studies on SLE^22–29 and some critical illnesses.^30–32 We aimed to investigate temporal trends in characteristics and outcome of these patients over a 10-year period. We hypothesized that there is a temporal improvement in the outcome of ICU patients with SLE.

Methods

Database

Data in this retrospective population-based cross-sectional study were retrieved from a national administrative database. In Taiwan, a compulsory and universal National Health Insurance (NHI) program has been implemented by the government since 1995. There are two major ethnic groups in Taiwan, namely Han Chinese (98%) and Taiwanese aborigines (2%). During the study period, the per capita income in Taiwan increased from US$13,609 (USD) in 1999 to US$17,399 in 2008. Patients in this study were drawn from the National Health Insurance Research Database (NHIRD),²¹ released for research purposes by the National Health Research Institute (NHRI), Taipei, Taiwan. The NHIRD provides information on inpatient and outpatient claims of nearly all the population (>22 million) and has been used in relevant studies on SLE^22–29 and some critical illnesses.^30–32 The NHIRD information includes encrypted patient identification numbers, sex, birthday, dates of admission and discharge, medical institutions providing the services, the ICD-9-CM (International Classification of Diseases, Ninth Revision, Clinical Modification) diagnosis codes (up to five) and procedure codes (up to five), outcome at hospital discharge (recovered, died or transferred), and hospital charges.

Ethics statement

The review board of the Medical Research Committee in Chi Mei Medical Center approved the study (grant No. CMFHR10081) and waived the need for formal ethical approval and written informed consent from the participants due to the use of a reimbursement database. Researchers using the database must sign an agreement based on the Computer-Processed Personal Data Protection Law and related regulations of the Bureau of NHI and the NHRI.

Study sample and definitions

SLE was defined by ICD-9-CM code 710.0. Patients with SLE were initially identified from a catastrophic illness registry, which includes selected major injuries or illnesses (including rheumatic diseases) and is published by the Department of Health, Executive Yuan. To register as SLE, the diagnosis must be made by board-certified rheumatologists and the application is further reviewed and approved by the Bureau of NHI, which ensures the accuracy and reliability of the diagnosis.²⁷ The study period spanned from 1999 through 2008 (both inclusive). We then linked patients with SLE in the registry to inpatient claims data and identified 14,668 patients with hospitalization; among them, 2994 (20.4%) had at least one ICU admission during the study period. After excluding patients less than 15 years old, the rest (n = 2870) were enrolled in the analysis. To ensure the independence of observations, only the first ICU admission during the study period was analyzed. The definitions of infection, acute organ dysfunction and hospital mortality were described previously.³²

Measurements

Measurements of the study subjects included age, sex, year of admission, hospital levels (medical center [>500 beds], regional [250–500 beds] and district hospitals [20–249 beds]), disease duration (measured by the time period from SLE diagnosis to the first ICU admission), major surgical condition (defined by the use of spinal, epidural or general anesthesia), Charlson Comorbidity Index,^33,34 selected comorbid conditions, principal diagnoses, infection, acute organ dysfunction, corticosteroid and immunosuppressive therapy for SLE, uses of life-support measures (including hemodialysis, vasopressors, and mechanical ventilation [MV]), lengths of ICU and hospital stay, and hospital mortality. Hospitals are periodically accredited by the Taiwan Joint Commission on Hospital Accreditation and Quality Improvement (authorized by the government) and classified into the three levels, representing different hospital sizes and service capabilities. The Charlson index, adapted for use with ICD-9-CM coded administrative databases, is a weighted summary measure of clinically important concomitant diseases including rheumatic disease.^33,34 For the purpose of this study, rheumatic disease or SLE was excluded from the Charlson index.

Statistics

Continuous variables are presented as median (interquartile range) owing to a skewed distribution; discrete ones as counts or percentages. To investigate temporal trends, analysis of variance was performed for linearity of scaled variables, and linear-by-linear association chi-square analysis was performed for categorical data. To assess the effect of calendar year on hospital mortality, we performed univariable and multivariable logistic regression models with generalized estimating equations methods,³⁵ which was used to account for hospital clustering effect³⁶ and was done by specifying an exchangeable structure of a working correlation matrix. Covariates in the model included age (<65 vs. ≥65 years), sex, disease duration (<1, 1–3, >3 years), non-SLE comorbidity (y/n), major surgical condition (y/n), rheumatic vs. non-rheumatic principal diagnosis, infection (y/n), acute organ dysfunction (y/n), and hospital level. As we simultaneously included a number of covariates in the multivariable regression model, a potential numerical problem concerned collinearity between covariates, rendering estimated regression coefficients invalid. We assessed this potential problem by examining the estimated slope coefficients and standard errors of the mean, and found no indication of collinearity. Model performance was assessed using c-statistics. Data analyses were performed using SPSS for Windows, version 17.0 (SPSS Inc., Illinois, US). Statistical significance was set at p < .05 (two-tailed).

Results

Demographics and clinical characteristics

Characteristics of the study subjects by subperiods are shown in Table 1. Two-thirds of the patients were treated in medical centers; the proportion did not change significantly during the study period. Most patients were women (overall 85.5%). The proportion of men slightly increased from 14.9% in 1999 and 2000 to 18.6% in 2007 and 2008 (p value for trend = .063). Median age of the patients increased from 38.5 to 44.0 years. The proportion of older patients significantly increased from 11.4% to 18.4%. Median disease duration was average 4.0 (interquartile range 1.0–7.0) years, rising from a median of two years in 1999–2000 to six years in 2007–2008. The proportion of patients with disease duration of more than three years increased steadily from 42.6% to 61.4% in 10 years.

Table 1

Characteristics of the intensive care unit (ICU) patients with systemic lupus erythematosus (SLE) by subperiod

Variables	1999–2000 (n = 448)	2001–2002 (n = 514)	2003–2004 (n = 575)	2005–2006 (n = 628)	2007–2008 (n = 705)	p Values for trend
Hospital level						0.813
Medical center	69.6	68.9	65.9	65.9	64.3
Regional hospital	19.9	21.8	28.6	25.2	30.9
District hospital	10.5	9.3	5.6	8.9	4.8
Age, years	38.5 (27.0–52.0)	43.0 (31.0–61.0)	41.0 (29.0–56.0)	43.5 (31.0–57.8)	44.0 (31.0–59.0)	<0.001
Male sex	14.9	15.7	15.9	15.1	18.6	0.063
Disease duration, year	2.0 (0–4.0)	3.5 (0–6.0)	4.0 (1.0–8.0)	5.0 (1.0–10.0)	6.0 (1.0–11.0)	<0.001
Comorbidity
Charlson/Deyo index ^a	1 (0–2)	1 (0–2)	1 (0–2)	1 (0–2)	1 (0–2)	0.733
≥1 non-SLE comorbidity	54.2	60.5	51.3	54.8	53.6	0.275
Neurological	12.1	10.3	11.0	10.4	10.6	0.552
Cardiovascular	13.8	14.4	11.3	12.9	11.8	0.216
Pulmonary	2.2	5.4	4.7	5.4	4.1	0.301
Renal	21.4	26.8	20.3	20.5	19.6	0.051
Hepatic	5.1	6.6	5.4	3.5	5.1	0.267
Cancer	4.2	3.5	3.7	6.2	4.4	0.272
Diabetes	4.7	6.2	7.8	4.9	7.0	0.375
Major surgical condition	22.3	21.0	25.4	24.2	26.2	0.050
Principal diagnosis						<0.001
Rheumatic	41.7	36.2	29.6	26.0	25.1
Circulatory	16.1	17.5	16.9	16.1	18.2
Respiratory	10.3	13.2	13.7	13.5	13.6
Infectious	6.5	3.5	6.6	8.3	9.8
Genitourinary	4.7	7.8	5.7	5.4	4.0
Injury/poisoning	4.2	3.5	6.4	4.9	5.4
Neoplasm	3.8	2.5	3.1	4.6	4.1
Digestive	3.1	6.6	5.0	4.8	5.5
Neurological	1.3	2.7	1.7	2.4	2.1
Others	8.3	6.4	11.1	14.0	12.2
Infection site ^b
Any	39.1	39.7	44.0	46.8	47.2	<0.001
Central nervous system	2.2	0.6	2.3	1.8	1.6	0.958
Cardiovascular	0.7	1.4	0.9	0.3	0.9	0.546
Respiratory	20.5	25.1	25.7	24.7	24.3	0.320
Intra-abdominal	4.2	4.3	5.0	6.5	5.0	0.247
Genitourinary	8.3	11.7	9.9	12.1	9.6	0.576
Skin/soft tissue	2.2	2.3	4.7	4.3	3.7	0.075
Sepsis/septicemia	4.5	1.6	3.1	6.8	6.0	0.001
Tuberculosis	3.1	1.9	3.1	2.5	2.7	0.944
Others	7.1	6.4	6.8	6.1	8.4	0.432
Organ dysfunction
Any	64.1	70.2	72.2	72.8	69.2	0.033
Neurological	2.5	3.1	3.5	5.4	5.1	.004
Cardiovascular	9.6	13.2	15.3	15.1	14.5	.024
Respiratory	40.6	50.6	56.7	53.7	55.0	<.001
Renal	36.2	35.6	34.3	33.1	28.8	.004
Hepatic	0.4	2.3	2.6	0.8	1.8	.620
Hematological	6.3	5.4	6.3	6.5	6.8	.462
Metabolic	0.4	0.2	0.3	0.8	0.6	.345

Data are presented as median (interquartile range) or as percentages.

Rheumatic diseases including SLE were not counted.

Patients may have more than one infection site.

Comorbidity other than SLE was common (overall 54.7%); the percentage varied insignificantly over time. The most common comorbid conditions were renal (overall 21.5%), cardiovascular (12.7%) and neurological diseases (10.8%). Approximately one-fourth of the patients had a major surgical condition; and the percentage slightly increased in recent years. The proportion of patients with a principal diagnosis of rheumatic disease (mainly SLE) fell significantly from 41.7% to 25.1%. Nevertheless, the most common principal diagnosis was still rheumatic disease, which was followed by circulatory and respiratory diagnoses. The incidence of infection increased gradually from 39.1% to 47.2%; however, the increase was significant only in sepsis/septicemia when individual infection sites were examined separately. The severity of patient conditions was increasing, as reflected by the high and increasing incidence of organ dysfunction, which increased from 64.1% in 1999 and 2000 to 69.2% in 2007 and 2008. The increase of organ dysfunction was mainly seen in neurological (from 2.5% to 5.1%), cardiovascular (from 9.6% to 14.5%), and respiratory systems (from 40.6% to 55.0%). Conversely, the incidence of renal dysfunction decreased from 36.2% to 28.8% during the same period.

Treatment, resource use and outcome

Table 2 shows the treatment and resource use of the study subjects by subperiods. Immunosuppressive or steroid therapy was prescribed in more than 80% of the patients during hospitalization; the percentage changed little over time. However, the percentages receiving hydroxychloroquine (from 27.4% to 38.6%), mycophenolate mofetil (from 1.1% to 3.5%) and cyclosporine (from 0.9% to 3.0%) significantly increased. Little change was found in the uses of corticosteroids, azathioprine, cyclophosphamide, tacrolimus, methotrexate, intravenous immunoglobulin, and plasmapheresis. More than three-fourths of the patients required some form of life-support measures, but the proportion decreased over time, especially of hemodialysis (from 37.3% to 28.7%) and vasopressors (from 58.3% to 47.9%). The use and duration of MV, as well as lengths of ICU and hospital stay changed little during the study period.

Table 2

Treatment and resource use of the intensive care unit (ICU) patients with systemic lupus erythematosus by subperiod

Variables	1999–2000 (n = 448)	2001–2002 (n = 514)	2003–2004 (n = 575)	2005–2006 (n = 628)	2007–2008 (n = 705)	p Values for trend
Immunosuppressive and steroid therapy
Corticosteroids a	80.8	75.9	76.3	76.8	76.2	0.187
Prednisolone <10 mg/d	38.4	40.9	43.8	42.5	41.4
Prednisolone 11–20 mg/d	12.9	14.4	8.3	11.1	10.2
Prednisolone 21–30 mg/d	4.7	3.5	4.0	2.9	4.7
Prednisolone >30 mg/d	24.8	17.1	20.2	20.2	19.9
Hydroxy-chloroquine	27.4	30.2	37.6	35.5	38.6	<0.001
Azathioprine	9.2	12.5	16.2	11.8	12.9	0.220
Cyclo-phosphamide	8.0	9.5	13.6	12.4	10.5	0.124
Mycophenolate mofetil	1.1	1.0	1.2	2.2	3.5	<0.001
Cyclosporine	0.9	1.2	3.3	2.2	3.0	0.011
Tacrolimus	0.7	0	0.3	0.8	1.1	0.062
Immunoglobulin	0.2	0.8	1.0	0.2	1.0	0.426
Methotrexate	0	1.8	0.7	1.0	0	0.239
Plasmapheresis	6.9	6.0	6.4	4.9	7.9	0.614
Hemodialysis	37.3	35.8	32.0	27.5	28.7	<0.001
Vasopressor use	58.3	52.7	50.6	47.0	47.9	<0.001
Mechanical ventilation (MV)	67.9	67.5	66.8	62.7	65.4	0.139
Duration of MV, d	5 (2–11)	4 (2–10)	4 (2–12)	3 (1–8)	4 (2–8)	0.271
ICU length of stay (LOS), d	4 (2–8)	4 (1–8)	3 (2–8)	3 (2–7)	4 (2–9)	0.565
Hospital LOS, d	24 (11–43)	21 (10–44)	23 (9–46)	19 (10–41)	21 (10–40)	0.204

Data are presented as median (interquartile range) or as percentages.

Mean daily dose of prednisolone. Note: Patients who received parenteral corticosteroids were classified as the group of >30 mg/d.

Hospital mortality decreased from 42.6% in 1999 and 2000 to 31.2% in 2007 and 2008 (Table 3). After the patient and hospital characteristics were adjusted, SLE patients admitted to the ICU in the last two years were 50% less likely to die in the hospital than those admitted in the first two years (adjusted odds ratio [aOR] 0.5, 95% confidence interval [CI] 0.38–0.67) (Table 3). Other independent predictors of mortality were older age (aOR 1.74), major surgical condition (aOR 0.61), infection (aOR 1.59), acute organ dysfunction (aOR 6.80), and hospital level (regional hospital vs. medical center, aOR 0.69) (Table 3). The performance of the model was acceptable (c-statistic 0.73, 95% CI 0.72–0.75).

Table 3

Predictors of hospital mortality in the intensive care unit patients with systemic lupus erythematosus (SLE)

Variables	n	Death %	Crude OR^a	Adjusted OR^a
Year of admission b
1999–2000	191	42.6	1.00	1.00
2001–2002	191	37.2	0.80 (0.63–1.02)	0.66 (0.50–0.87)
2003–2004	211	36.7	0.80 (0.63–1.03)	0.65 (0.50–0.84)
2005–2006	185	29.5	0.57 (0.47–0.70)	0.43 (0.35–0.53)
2007–2008	220	31.2	0.62 (0.49–0.80)	0.50 (0.38–0.67)
Age, years
<65	796	33.3	1.00	1.00
≥65	202	42.3	1.56 (1.25–1.96)	1.74 (1.29–2.34)
Sex
Women	853	34.8	1.00	1.00
Men	145	34.9	1.00 (0.77–1.31)	0.91 (0.70–1.18)
Disease duration, years
<1	263	39.4	1.00	1.00
1–3	223	35.2	0.83 (0.66–1.04)	0.98 (0.78–1.24)
>3	512	32.6	0.72 (0.61–0.85)	0.91 (0.78–1.07)
Non-SLE comorbidity
No	441	33.9	1.00	1.00
Yes	557	35.5	1.09 (0.93–1.27)	0.99 (0.82–1.21)
Major surgical condition
No	825	37.9	1.00	1.00
Yes	173	25.0	0.53 (0.41–0.70)	0.61 (0.47–0.80)
Principal diagnosis
Rheumatic	340	38.5	1.00	1.00
Non-rheumatic	658	34.1	0.80 (0.64–1.02)	0.86 (0.67–1.12)
Infection
No	452	28.1	1.00	1.00
Yes	546	43.4	2.04 (1.71–2.44)	1.59 (1.29–1.97)
Acute organ failure
No	93	10.8	1.00	1.00
Yes	905	45.1	7.15 (4.99–10.23)	6.80 (4.81–9.59)
Hospital level
Medical center	710	37.2	1.00	1.00
Regional hospital	217	29.3	0.75 (0.61–0.91)	0.69 (0.57–0.85)
District hospital	71	32.7	0.87 (0.63–1.19)	0.77 (0.54–1.09)

CI: confidence interval; OR: odds ratio.

Logistic regression models with generalized estimating equations methods.

p values for trend were less than .001 for both univariable and multivariable analyses.

Discussion

In this study, we found that ICU patients with SLE in Taiwan were getting older, were increasingly admitted later after SLE diagnosis and more commonly for non-rheumatic disease, and had more overall incidence of infection and organ dysfunction over the recent 10 years. Conversely, the incidence of renal dysfunction and the percentages of patients receiving hemodialysis and vasopressors decreased over time. More importantly, the risk of death was halved during the study period when changes in the patient and hospital characteristics were adjusted.

The characteristics of the ICU patients with SLE are expected to change over time because of an improving survival of SLE patients over the decades. The five-year survival rate of SLE patients increases from 50% in the 1950s to over 90% in the 2000s.⁵ To the best of our knowledge, this is the first and largest population-based study that enrolled ICU patients with SLE and showed the temporal change of the patient characteristics and the short-term mortality over a long period of time. Unlike previous reports on ICU patients with SLE,^8–20 our study is strengthened by its generalizability through an unbiased selection of all treated patients in Taiwan. Most of the previous studies are done earlier (spanned from 1978 to 2000), enroll fewer (usually less than 30) patients from single institutions, combine with other rheumatic diseases, or do not provide separate analyses of SLE.^13–18,20 Nevertheless, these studies, taken together, can provide some useful information. For example, these reports reveal a trend of an increasing age of ICU patients with SLE over time (from average 27 years in earlier reports^8–11 to 32 years in recent ones),^12,19 which is consistent with our results. Furthermore, we found that patients were admitted to the ICU increasingly later after diagnosis of SLE and more commonly for non-rheumatic reasons. These findings likely reflect an improved treatment and survival of SLE patients over the years.^5,37

The previous largest study on ICU patients with SLE is reported by Ñamendys-Silva and coworkers in 2009.¹² They investigated 104 SLE patients, with a mean age of 32 years and mean disease duration of 70 months, in a single ICU in Mexico from 1994 to 2004.¹² Among these patients, 96.2% were women, 61.5% had infection, 46.2% had acute renal dysfunction, 54.2% received hemodialysis, 67.3% received vasopressors/inotropics, and 74% received MV.¹² They found an overall ICU mortality rate of 32.7%; which was significantly lower in the later years than in the first few years (24.2% in 2000–2004 vs. 47.6% in 1994–1999).¹² The corresponding figures in our study were mean (median) age of 42 (44) years, mean (median) disease duration of 56 (50) months, women 85.5% (overall), infection 43.9%, acute renal dysfunction 33.2%, hemodialysis 31.7%, vasopressors 50.7%, MV 65.9%, and hospital mortality 34.8%, respectively. Among studies enrolling fewer ICU patients with SLE, the ICU mortality rates varied more widely.^8–11,19 The variation in mortality among these studies is likely due to differences in time periods, patient populations, and study designs.

The occurrence of acute organ dysfunction is the strongest predictor of death in the ICU patients with SLE (aOR 6.80, see Table 3), which is not unexpected because organ dysfunction is usually an intermediate pathway leading to death.³² Findings on the temporal change of organ dysfunction may also provide further evidence for the changing pattern of morbidity and mortality found in patients with SLE.^6,7 Previous studies on all patients with SLE have shown that deaths due to active SLE, lupus nephritis, and infection are decreasing, whereas deaths due to thrombosis and circulatory diseases are increasing.^6,7 The decreasing incidence of renal dysfunction and increasing incidence of cardiovascular dysfunction found in the present study agree with previous observations in all patients with SLE.^6,7 However, the increasing incidence of infection and neurological dysfunction in our ICU patients with SLE somewhat differs from these previous observations.^6,7 Moreover, we also found an increasing incidence of acute respiratory dysfunction in the ICU patients with SLE. Because only 20% of the Taiwanese patients with SLE had an ICU stay during the study period, the characteristics of these patients may significantly differ from those of without an ICU stay. Further research is needed to find these differences and factors that may be associated with ICU admission.

Treatment of SLE is changing,³⁷ as was also found in the ICU patients. The finding of increasing prescriptions of non-steroid medications, especially hydroxychloroquine (an antimalarial drug), for the ICU patients with SLE is consistent with a previous longitudinal study on patients with newly diagnosed SLE in Canada.³⁷ However, hydroxychloroquine was less frequently prescribed for patients in our study than for those in the Canadian Study (<40% vs. 46% in the 2000s).³⁷ One study has reported even more frequent use of hydroxychloroquine, especially in SLE patients managed by rheumatologists (88%).³⁸ The less frequent use of this medication in our study cohort may be due to different patient characteristics, physician specialties, and/or discontinuity of care during ICU stay. When patients are admitted to the ICU, discontinuity of care is likely to occur because both open and closed ICU systems exist in Taiwan. Hydroxychloroquine has been advocated as the cornerstone of SLE treatment due to its good safety profile and numerous beneficial effects.^39,40 Treatment of the ICU patients with SLE may be further improved by education and/or through a multidisciplinary team approach.^38,41

Reasons for the observed mortality decline in the critically ill patients with SLE cannot be determined in this study. However, it is likely that the mortality decline was somewhat attributed to increased use of early, effective lupus medication such as mycophenolate mofetil and hydroxychloroquine along with reduced side effects from lower doses of corticosteroids for SLE patients. Another reason for the mortality decline is likely attributed to the general improvement of critical care because the decline persisted after changes in patient and hospital characteristics were adjusted and there are still no treatment guidelines specific for these patients. Beginning in the early 2000s, there has been an increasing worldwide application of some important concepts of critical care, such as improving patient safety through quality improvement initiatives and implementing care bundles or protocols through evidence-based medicine. These approaches may help reduce adverse events and provide early and effective treatment for critically ill patients with SLE, leading to the improved outcome.

Three poor prognostic factors (i.e., older age, infection, and organ dysfunction) found in the present study may potentially hamper the temporal improvement of the short-term survival in the ICU patients with SLE. Improved treatment of SLE reduces and postpones the occurrence of acute critical illnesses, as was shown in our study: the median time from SLE diagnosis to ICU admission had increased by four years during the 10-year study period. However, the improved survival is achieved at the costs of immunosuppression and accrual of organ damage.^42,43 Moreover, the increasing patient age and incidence of infection and organ dysfunction in the ICU patients with SLE will pose greater challenges to caring physicians in the future.

Comorbidity, measured by the Charlson Comorbidity Index, is associated with a decreased long-term survival in SLE patients,⁴⁴ independent of the Systemic Lupus International Collaborating Clinics/American College of Rheumatology Damage Index.⁴³ In the present study, non-SLE comorbidity was not associated with the short-term survival in the multivariable model. Because acute organ dysfunction is closer to death than comorbidity, we speculated that the inclusion of acute organ dysfunction in the model may attenuate the effect of comorbidity on mortality. When acute organ dysfunction was excluded from the model, we confirmed the speculation and found that non-SLE comorbidity was associated with increased risk of hospital mortality (aOR 1.21, 95% CI 1.02–1.44).

Several limitations deserve comments. First, the status of anticardiolipin antibodies, the disease activities of SLE, the renal status related to lupus activity and the acute severity and organ dysfunction scores of the critical illnesses were not available in the database. Despite these, we provided data on patients with chronic renal disease and acute renal dysfunction as well as those undergoing hemodialysis. Information on acute organ dysfunction shown in Table 1 may also provide a surrogate measure of the severity of the acute illnesses. Second, the disease duration is likely underestimated because the onset of SLE usually predates the registry as a catastrophic illness in the database. Third, the prevalence of some comorbidities and the incidence of infection and organ dysfunction may be underestimated due to the limited diagnosis space in the database. Fourth, we did not measure the use of antibiotics for the patients because the information is complex and the issue is not the primary purpose of this study. Further research on this topic is needed. Finally, hospital mortality may also be underestimated because of inability to verify through death certificate. Despite these limitations, using the same ICD-9-CM codes over the study period would retain the internal validity of the temporal trend analyses.

In conclusion, the characteristics and outcomes of the ICU patients with SLE are changing over time. The short-term survival of these patients is improving in Taiwan despite an increasing severity of the acute critical illnesses.

Footnotes

Acknowledgements

The study was 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. The interpretation and conclusions contained herein do not represent those of the Bureau of National Health Insurance, Department of Health or National Health Research Institutes.

Funding

The study was performed in Chi Mei Medical Center and supported by grant CMFHR10081 from the hospital. The funding source had no role in the design, execution, analysis and interpretation of data, or writing of the study.

Conflict of interest

None declared.

References

Palatinus

Adams

. Thrombosis in systemic lupus erythematosus. Semin Thromb Hemost 2009; 35: 621–629.

Navarra

Leynes

. Infections in systemic lupus erythematosus. Lupus 2010; 19: 1419–1424.

Ruiz-Irastorza

Khamashta

Castellino

Hughes

. Systemic lupus erythematosus. Lancet 2001; 357: 1027–1032.

Raj

Murin

Matthay

Wiedemann

. Systemic lupus erythematosus in the intensive care unit. Crit Care Clin 2002; 18: 781–803.

Pons-Estel

Alarcon

Scofield

Reinlib

Cooper

. Understanding the epidemiology and progression of systemic lupus erythematosus. Semin Arthritis Rheum 2010; 39: 257–268.

Cervera

Khamashta

Font

. Morbidity and mortality in systemic lupus erythematosus during a 10-year period: A comparison of early and late manifestations in a cohort of 1000 patients. Medicine (Baltimore) 2003; 82: 299–308.

Bernatsky

Boivin

Joseph

. Mortality in systemic lupus erythematosus. Arthritis Rheum 2006; 54: 2550–2557.

Ansell

Bedhesi

Ruff

. Study of critically ill patients with systemic lupus erythematosus. Crit Care Med 1996; 24: 981–984.

Williams

Chinn

Hughes

Leach

. Critical illness in systemic lupus erythematosus and the antiphospholipid syndrome. Ann Rheum Dis 2002; 61: 414–421.

10.

Alzeer

Al-Arfaj

Basha

. Outcome of patients with systemic lupus erythematosus in intensive care unit. Lupus 2004; 13: 537–542.

11.

Hsu

Chen

Yeh

. Outcome and prognostic factors in critically ill patients with systemic lupus erythematosus: A retrospective study. Crit Care 2005; 9: R177–183.

12.

Namendys-Silva

Baltazar-Torres

Rivero-Sigarroa

. Prognostic factors in patients with systemic lupus erythematosus admitted to the intensive care unit. Lupus 2009; 18: 1252–1258.

13.

Godeau

Boudjadja

Dhainaut

. Outcome of patients with systemic rheumatic disease admitted to medical intensive care units. Ann Rheum Dis 1992; 51: 627‐–631.

14.

Godeau

Mortier

Roy

. Short and longterm outcomes for patients with systemic rheumatic diseases admitted to intensive care units: A prognostic study of 181 patients. J Rheumatol 1997; 24: 1317–1323.

15.

Thong

Tai

Goh

Johan

. An audit of patients with rheumatic disease requiring medical intensive care. Ann Acad Med Singapore 2001; 30: 254–259.

16.

Moreels

Melot

Leeman

. Prognosis of patients with systemic rheumatic diseases admitted to the intensive care unit. Intensive Care Med 2005; 31: 591–593.

17.

Camargo

Tobon

Fonseca

. Autoimmune rheumatic diseases in the intensive care unit: Experience from a tertiary referral hospital and review of the literature. Lupus 2005; 14: 315–320.

18.

Cavallasca

Del Rosario Maliandi

Sarquis

. Outcome of patients with systemic rheumatic diseases admitted to a medical intensive care unit. J Clin Rheumatol 2010; 16: 400–402.

19.

Ranzani

Battaini

Moraes

. Outcomes and organ dysfunctions of critically ill patients with systemic lupus erythematosus and other systemic rheumatic diseases. Braz J Med Biol Res 2011; 44: 1184–1193.

20.

Lee J, Yim JJ, Yang SC, et al. Outcome of patients with connective tissue disease requiring intensive care for respiratory failure. Rheumatol Int 2012; 32: 3353–3358. DOI: 10.1007/s00296-011-2158-6.

21.

National Health Insurance Research Database. http://nhird.nhri.org.tw/en/index.htm. Accessed 28 January 2013.

22.

Chiu

Huang

Chan

. Increased risk of ischemic stroke in patients with systemic lupus erythematosus: A nationwide population-based study. Intern Med 2012; 51: 17–21.

23.

Chen

Wang

Chou

. Predictors of longterm mortality in patients with and without systemic lupus erythematosus on maintenance dialysis: A comparative study. J Rheumatol 2011; 38: 2390–2394.

24.

Lin

Liang

Liu

. Tuberculosis as a risk factor for systemic lupus erythematosus: Results of a nationwide study in Taiwan. Rheumatol Int 2012; 32: 1669–1673.

25.

Liang

Sun

Yeh

. Malignancies associated with systemic lupus erythematosus in Taiwan: A nationwide population-based cohort study. Rheumatol Int 2012; 32: 773–778.

26.

Chen

Chang

Wang

. Malignancy in systemic lupus erythematosus: A nationwide cohort study in Taiwan. Am J Med 2010; 123: 1150.e1–1150.e6.

27.

Chiu

Lai

. Nationwide population-based epidemiologic study of systemic lupus erythematosus in Taiwan. Lupus 2010; 19: 1250–1255.

28.

Wang

Lin

. Relationship between caseload volume and outcome for systemic lupus erythematosus treatment: The experience of Taiwan. J Rheumatol 2008; 35: 1795–1800.

29.

Huang

Yao

See

. Prevalence of pediatric systemic lupus erythematosus and juvenile chronic arthritis in a Chinese population: A nation-wide prospective population-based study in Taiwan. Clin Exp Rheumatol 2004; 22: 776–780.

30.

Shen

Yang

. Epidemiologic trend of severe sepsis in Taiwan from 1997 through 2006. Chest 2010; 138: 298–304.

31.

Shen

. Incidence, resource use, and outcome of acute pancreatitis with/without intensive care: A nationwide population-based study in Taiwan. Pancreas 2011; 40: 10–15.

32.

Shen

Yang

. Increased risks of acute organ dysfunction and mortality in intensive care unit patients with schizophrenia: A nationwide population-based study. Psychosom Med 2011; 73: 620–626.

33.

Charlson

Pompei

Ales

MacKenzie

. A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. J Chronic Dis 1987; 40: 373–383.

34.

Deyo

Cherkin

Ciol

. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol 1992; 45: 613–619.

35.

Hanley

Negassa

Edwardes

Forrester

. Statistical analysis of correlated data using generalized estimating equations: An orientation. Am J Epidemiol 2003; 157: 364–375.

36.

Ward

. Hospital experience and mortality in patients with systemic lupus erythematosus. Arthritis Rheum 1999; 42: 891–898.

37.

Urowitz

Gladman

Tom

Ibanez

Farewell

. Changing patterns in mortality and disease outcomes for patients with systemic lupus erythematosus. J Rheumatol 2008; 35: 2152–2158.

38.

Lerang

Gilboe

Gran

. Differences between rheumatologists and other internists regarding diagnosis and treatment of systemic lupus erythematosus. Rheumatology (Oxford) 2012; 51: 663–669.

39.

Ruiz-Irastorza

Khamashta

. Hydroxychloroquine: The cornerstone of lupus therapy. Lupus 2008; 17: 271–273.

40.

Lee

Silverman

Bargman

. The role of antimalarial agents in the treatment of SLE and lupus nephritis. Nat Rev Nephrol 2011; 7: 718–729.

41.

Vincent

. Critical care nephrology: A multidisciplinary approach. Contrib Nephrol 2007; 156: 24–31.

42.

Mok

Wong

Lau

. Damage accrual in southern Chinese patients with systemic lupus erythematosus. J Rheumatol 2003; 30: 1513–1519.

43.

Gladman

Ginzler

Goldsmith

. The development and initial validation of the Systemic Lupus International Collaborating Clinics/American College of Rheumatology Damage Index for systemic lupus erythematosus. Arthritis Rheum 1996; 39: 363–369.

44.

Jonsen

Clarke

Joseph

. Association of the Charlson comorbidity index with mortality in systemic lupus erythematosus. Arthritis Care Res (Hoboken) 2011; 63: 1233–1237.