Arthritis Care & Research

Vol. 66, No. 4, April 2014, pp 608 – 616

DOI 10.1002/acr.22173
© 2014, American College of Rheumatology
ORIGINAL ARTICLE

Overall and Cause-Specific Mortality in

Patients With Systemic Lupus Erythematosus:
A Meta-Analysis of Observational Studies
MARKO YURKOVICH,1 KATERYNA VOSTRETSOVA,1 WENJIA CHEN,1 AND
J. ANTONIO AVIÑA-ZUBIETA2

Objective. To determine the magnitude of risk from all-cause and cause-specific mortality in patients with systemic
lupus erythematosus (SLE) compared to the general population through a meta-analysis of observational studies.
Methods. We searched the Medline and Embase databases from their inception to October 2011. Observational studies
that met the following criteria were assessed: 1) a prespecified SLE definition; 2) overall and/or cause-specific deaths,
including cardiovascular disease (CVD), infections, malignancy, and renal disease; and 3) reported standardized mor-
tality ratios (SMRs) and 95% confidence intervals (95% CIs). We calculated weighted–pooled summary estimates of SMRs
(meta-SMRs) for all-cause and cause-specific mortality using the random-effects model and tested for heterogeneity using
the I2 statistic by using Stata/IC statistical software.
Results. We identified 12 studies comprising 27,123 patients with SLE (4,993 observed deaths) that met the inclusion
criteria. Overall, there was a 3-fold increased risk of death in patients with SLE (meta-SMR 2.98, 95% CI 2.32–3.83) when
compared with the general population. The risks of death due to CVD (meta-SMR 2.72, 95% CI 1.83– 4.04), infection
(meta-SMR 4.98, 95% CI 3.92– 6.32), and renal disease (SMR 7.90, 95% CI 5.50 –11.00) were significantly increased.
Mortality due to malignancy was the only cause-specific entity not increased in SLE (meta-SMR 1.19, 95% CI 0.89 –1.59).
Conclusion. The published data indicated a 3-fold increase in all-cause mortality in patients with SLE compared to the
general population. Additionally, all cause-specific mortality rates were increased except for malignancy, with renal
disease having the highest mortality risk.

INTRODUCTION findings may be attributed to earlier diagnosis and treat-

Systemic lupus erythematosus (SLE) is a chronic autoim-
mune disease that has the potential to affect all organ
systems and can be associated with severe morbidity and
mortality. Recent studies have reported that the survival of
patients with SLE has improved significantly (1– 4). For
example, the present 5-year survival rate has increased to
⬎95% compared with 64 – 87% in the 1980s (2). These
Dr. Aviña-Zubieta holds salary awards from the Canadian
Arthritis Network and the Arthritis Society of Canada and is
currently the British Columbia Lupus Society Scholar.
1
Marko Yurkovich, MD, Kateryna Vostretsova, MD,
Wenjia Chen, MS, MPH: University of British Colum-
bia, Vancouver, British Columbia, Canada; 2J. Antonio
Aviña-Zubieta, MD, PhD: University of British Columbia,
Vancouver, and Arthritis Research Centre of Canada, Rich-
mond, British Columbia, Canada.
Address correspondence to J. Antonio Aviña-Zubieta,
MD, PhD, Arthritis Research Centre of Canada, First Floor,
5591 No. 3 Road, Richmond, British Columbia, V6X 2C7
Canada. E-mail: azubieta@arthritisresearch.ca.
Submitted for publication May 11, 2013; accepted in re-
vised form September 10, 2013.
ment, as well as more aggressive management of disease-
related comorbidities (2,3). Despite these improvements,
patients with SLE still have mortality rates 2–5 times
higher than the general population (5), and the 15-year
survival rate has been reported to range between 76% and
85% (4 –7). In addition, mortality rates have been shown to
vary between countries and to be higher in men (7).
Although SLE can potentially affect any organ system,
some of the more serious complications include cardiovas-
cular and end-stage renal disease (8 –10). Additionally,
people who have SLE may be more prone to certain infec-
tions and malignancies compared with the general popu-
lation (9,10). Although the literature regarding mortality in
SLE has been growing, few studies to date have reported
standardized mortality ratios (SMRs) for all-cause mortal-
ity and cause-specific mortality. The use of SMRs seems to
give the best estimate when assessing mortality, given that
the comparator group is standardized from the general
population. No study to date has summarized the magni-
tude of the problem by conducting a systematic review and
meta-analysis of the observational studies available on
SLE-related mortality.

608
Mortality in SLE
Significance & Innovations
● Patients with systemic lupus erythematosus (SLE)
had a 3-fold increased risk of premature mortality
from any cause compared with the general popu-
lation. This risk was similar between sexes.
● Over time, cardiovascular mortality decreased in
the general population; however, a similar level of
improvement was not seen in SLE patients. Pa-
tients with SLE still had a 3-fold increased risk of
premature cardiovascular mortality.
● Mortality secondary to end-stage renal disease and
infections were the causes with the highest mor-
tality in patients with SLE, being 8 and 5 times
greater than the general population, respectively.
However, only 1 study reported renal disease us-
ing standardized mortality ratios.
● Studies assessing mortality in SLE from cohorts
assembled during the last 2 decades are very lim-
ited; therefore, research evaluating the contempo-
rary mortality in patients with SLE is urgently
needed.
The purpose of this study was to determine the magni-
tude of the risk from all-cause mortality and cause-specific
mortality in patients with SLE. In particular, cardiovascu-
lar disease (CVD), infection, malignancy, and renal disease
were examined in patients with SLE compared to the
general population by conducting a meta-analysis of ob-
servational studies.
MATERIALS AND METHODS
Search strategies. The Medline and Embase databases
were searched from their inception (1946 and 1980, re-
spectively) to October 2011 by an experienced medical
librarian to identify primary studies and reviews in the
field of mortality in SLE. The following search terms were
used alone or in combination: systemic lupus erythemato-
sus, mortality, standardized mortality ratio, and all-cause
mortality. SMRs and specific causes searched included
CVD (including ischemic heart disease [IHD] and cerebro-
vascular accidents [CVAs]), infections, cancer (malig-
nancy), and renal disease (renal failure). The authors also
hand searched the reference lists of textbooks and key
articles retrieved for additional references.
We selected peer-reviewed articles (case– control and
cohort studies) that met the following inclusion criteria: 1)
predefined SLE definition; 2) overall and/or cause-specific
deaths, including CVD, infections, malignancy, and renal
disease; and 3) reported SMRs and 95% confidence inter-
vals (95% CIs) or the data for calculating them. If data from
a single study were reported in ⬎1 article, only the results
from the most recent study were included in the meta-
analysis. We only included studies of adults ages ⬎18
years.
609
Data extraction. Two researchers (MY and KV) inde-
pendently assessed the studies for eligibility and extracted
data on the year of publication, type of study, source of the
SLE sample, sample size, female:male ratio in the sample,
mean age at diagnosis, mean duration of SLE, enrollment
period, SLE definition, extent of loss to followup, and
number of observed and expected deaths. Any differences
among the 2 authors were discussed until consensus was
reached involving a third researcher (JAA-Z).
Quality scores of the included studies. We assessed
study quality based on a 12-point scale that included ele-
ments of previously published scales for observational
studies (11,12). This scale was also used in our previously
published meta-analyses on the risk of cardiovascular
events and mortality in rheumatoid arthritis (13,14). Each
study was scored according to 5 characteristics related to
its patient sample and methods. Each item was scored as 0,
1, or 2; specifically, we determined the source of the study
sample (community based ⫽ 2, clinic based ⫽ 1, and
undefined ⫽ 0); cohort type, if applicable (inception ⫽ 2
and noninception ⫽ 1); definition of SLE (use of the Amer-
ican College of Rheumatology classification criteria for
SLE [15] ⫽ 2, other validated criteria ⫽ 1, and other pre-
defined but nonvalidated criteria ⫽ 0); ascertainment of
death outcome (validated criteria ⫽ 2, nonvalidated but
clearly defined criteria [e.g., death certificates] ⫽ 1, and
not mentioned ⫽ 0); and SLE exposure (⬎10 years ⫽ 2, ⬎5
years but ⬍10 years ⫽ 1, and ⬍5 years or not defined ⫽ 0).
For stratification purposes, studies that scored ⱖ7 were
considered to be of higher quality and the rest were con-
sidered to be of lower quality. Quality scoring was per-
formed independently by 2 reviewers (MY and KV). Dis-
agreement was resolved by consensus involving a third
reviewer (JAA-Z).
Statistical analysis. We calculated weighted–pooled
summary estimates of SMRs (meta-SMRs) for all-cause
mortality, as well as for cause-specific mortality, which
included CVD (IHD and CVA), infection, malignancy, and
renal disease. The meta-SMR represents a summary esti-
mate of the increased risk of death in patients with SLE
compared with the general population, weighted by the
sample size of each study. Separate meta-SMRs were cal-
culated for men and women when available. The calcula-
tions were performed on the log of the SMRs from the
individual studies, and the resulting pooled values were
subsequently transformed back to the SMR scale. We used
the random-effects model and tested for heterogeneity us-
ing the bootstrap version of the Q and I2 statistics using
Stata/IC statistical software, version 12.1.
Because heterogeneity is expected in meta-analyses of
observational studies, a subgroup analysis was carried out
to assess heterogeneity. The studies were stratified based
on study population (community-based versus clinic-
based samples), cohort type (inception versus nonincep-
tion), quality score (⬍7 versus ⱖ7), and the midpoint of
the enrollment period (before versus after 2000). Statistical
inferences about the difference in the meta-SMRs between
subgroups of the studies were performed using a univari-
610 Yurkovich et al

Figure 1. Funnel plot of 14 cohorts from the 12 studies evaluating systemic lupus
erythematosus mortality. SMRs ⫽ standardized mortality ratios; CI ⫽ confidence
interval.

ate meta–regression analysis (16). A multivariate meta–
regression analysis to evaluate the adjusted effect of the
above study characteristics was not performed because of
the small number of studies and because, in the majority of
studies, at least 1 study characteristic could not be esti-
mated, leaving too few observations for the multivariate
regression.
Robustness of the result was evaluated using a jackknife
sensitivity analysis (i.e., the analysis was repeated multi-
ple times, each time with removal of a single study from
the baseline group of studies) (17).
Assessment of publication bias/small-study effect. To
detect the presence of publication bias (i.e., the bias result-
ing from the greater likelihood of published studies report-
ing positive results compared with negative results) or the
small-study effect (i.e., a tendency for treatment effect
estimates in small studies to differ from those in larger
studies) (18), we constructed a funnel plot, in which a
measure of the study size was plotted as a function of the
measure of interest (19). The funnel plot distribution of the
data points (Figure 1) was considered symmetric; there-
fore, we considered no publication bias. Egger’s test was
also used to test the funnel plot asymmetry, and this was
not significant (P ⫽ 0.4).
RESULTS
All-cause mortality. We screened 566 abstracts pub-
lished over the last 65 years (Figure 2). Upon completion,
we identified a total of 12 studies (2– 4,8 –10,20 –25) with
SMRs evaluating the risk of all-cause and/or cause-specific
mortality, comprising a total of 27,210 patients with SLE
and a total of 4,989 observed deaths (Table 1). The meta-
demographics for age and sex for all studies were 33.4

Figure 2. Flow chart of the study selection from the literature search. SMR ⫽
standardized mortality ratio.
 Table 1. Characteristics of the 12 studies included in the meta-analysis of all-cause mortality in patients with SLE*

Mean age Mean

Mortality in SLE

Number at duration Quality

Author, year Enrollment of Sample Inception diagnosis, of SLE, Women, SLE Death outcome score
(ref.) Country N period deaths type cohort years years % definition ascertainment (of 12)

Alamanos et al, Greece 178 1982–2001 12 Community Yes 38.8 8.25 88 ACR NA 7
2003 (20)
Bernatsky et al, Multinational† 9,547 1970–2001 1,255 Clinical No NA NA 90 ACR DC 7
2006 (21)
Bernatsky et al, Canada 2,688 1958–2001 360 Clinical No 34.6 NA 90 ACR DC 6
2006 (22)
Bjornadal et al, Sweden 4,737 1964–1994 2,314 Community No NA NA 78 ICD-7, ICD-8, DC 5
2004 (8) ICD-9
Chun and Bae, South Korea 434 1992–2002 10 Clinical No 36.1 5.6 100 ACR NA 6
2005 (9)
Hersh et al, US 957 2002–2003 72 Community No 52, 3‡ 16.5, 19.5‡ 91 ACR DC 10
2010 (2)
Jacobsen et al, Denmark 513 1976–1995 122 Clinical Yes 34.3 6.2 88 ACR DC, MR, autopsy 10
1999 (10)
Mok et al, 2011 Hong Kong 5,243 1999–2008 514 Community No NA NA 91 ICD-9 DC 5
(23)
Mok et al, 2008 Hong Kong 442 2000–2006 30 Community No NA 8.5 91 ACR DC 6
(24)
Uramoto et al, US 61 1950–1992 18 Community No NA NA NA ACR NA 6
1999 (4)
Urowitz et al, Canada 1,241 1970–2005 211 Clinical No NA NA 83–92§ ACR DC, GP, family 6
2008 (3)
To et al, 2009 Hong Kong 1,082 2007–NA 75 Clinical No 30.5 10.3 90 ACR NA 7
(25)¶ (total) (total)
Cluster 1# 347 NA 28.5 9.9 92
Cluster 2** 409 NA 35.1 8.2 88
Cluster 3†† 326 NA 26.9 12.2 89

* SLE ⫽ systemic lupus erythematosus; ACR ⫽ American College of Rheumatology; NA ⫽ not available (or not reported); DC ⫽ death certificate; ICD-7 ⫽ International Classification of Diseases, Seventh
Revision; MR ⫽ medical record; GP ⫽ general practitioner.
† Multinational cohort from Canada, the US, the UK, Sweden, and South Korea (23 centers total).
‡ The data for mean age and mean duration of SLE in this study are shown for the adult-onset and childhood-onset cohorts, respectively.
§ This study included 4 separate cohorts stratified by enrollment period, with the percentage of women in each cohort ranging from 83–92%.
¶ This study divided patients into 3 distinct groups using K-means cluster analysis. Clinical features, prevalence of proliferative lupus nephritis, autoantibody profile, and treatment data were compared
and standardized mortality ratios were calculated for each cluster of patients.
# Cluster 1: characterized by patients with predominantly mucocutaneous manifestations and arthritis, but having the lowest prevalence of serositis, hematologic manifestations, and proliferative lupus
nephritis.
** Cluster 2: characterized by patients with mainly renal and hematologic manifestations, but having the lowest prevalence of mucocutaneous manifestations.
†† Cluster 3: characterized by patients with the most heterogeneous features, having the highest prevalence of mucocutaneous manifestations, serositis and hematologic manifestations, renal
involvement, and proliferative lupus nephritis among the 3 clusters.
611
612 Yurkovich et al

Figure 3. Meta-analysis of 12 studies on all-cause mortality in patients with

systemic lupus erythematosus. meta-SMR ⫽ weighted–pooled summary estimates
of standardized response means.

years (95% CI 30.4 –36.4 years) and 89.6% women (95% CI
86.3–92.1%), respectively.
All 12 studies assessed all-cause mortality. Overall,
there was a 3-fold increase in mortality risk in patients
with SLE (meta-SMR 2.98, 95% CI 2.32–3.83) compared
with the general population (Figure 3). Six studies pro-
vided SMR estimates by sex separately (3,8,10,21,23,24).
Overall, there was no significant difference in the risk of
mortality in women compared with men (meta-SMR 4.06,
95% CI 3.11–5.30 and meta-SMR 3.41, 95% CI 2.56 – 4.53,
respectively). The meta-SMR remained significant when
studies were excluded one at a time in the jackknife sen-
sitivity analyses, with the point estimates ranging from
2.84 –3.20 and the corresponding 95% CIs remaining ⬎1 in
all analyses. The results from the jackknife sensitivity
analyses are shown in Table 2.
There was significant heterogeneity among the studies
included in our sample (P ⫽ 0.001). However, in an at-

Table 2. Sensitivity analysis using the jackknife approach, where each study is
excluded one at a time to test robustness of the overall SMR*

All-cause mortality, Study excluded,

Author, year (ref.) SMR (95% CI) meta-SMR (95% CI)

All studies 2.98 (2.32–3.83) Not applicable

Alamanos et al, 2003 (20) 1.5 (1.3–1.8) 3.16 (2.45–4.07)
Bernatsky et al, 2006 (21) 2.4 (2.3–2.5) 3.04 (2.35–3.93)
Bernatsky et al, 2006 (22) 3.3 (3.0–3.6) 2.95 (2.24–3.88)
Bjornadal et al, 2004 (8) 3.63 (3.49–3.78) 2.91 (2.12–4.00)
Chun and Bae, 2005 (9) 3.02 (1.45–5.55) 2.98 (2.30–3.86)
Hersh et al, 2010 (2) 2.5 (2–3.2) 3.02 (2.33–3.92)
Jacobsen et al, 1999 (10) 4.6 (3.8–5.5) 2.88 (2.21–3.74)
Mok et al, 2011 (23) 5.25 (4.79–5.70) 2.84 (2.18–3.70)
Mok et al, 2008 (24) 3.67 (2.52–5.31) 2.94 (2.27–3.81)
Uramoto et al, 1999 (4) 2.7 (1.65–4.18) 3.00 (2.32–3.89)
Urowitz et al, 2008 (3) 4.53 (3.96–5.19) 2.88 (2.21–3.75)
To et al, 2009 (25), cluster 1 0.95 (0.5–1.7) 3.19 (2.47–4.12)
To et al, 2009 (25), cluster 2 7.23 (6.7–7.7) 2.79 (2.27–3.44)
To et al, 2009 (25), cluster 3 1.27 (1.1–1.5) 3.20 (2.50–4.11)

* SMR ⫽ standardized mortality ratio; 95% CI ⫽ 95% confidence interval; meta-SMR ⫽ weighted–pooled
summary estimates of standardized response means.
Mortality in SLE
Figure 4. Weighted–pooled summary estimates of standardized response means
(meta-SMR) on cause-specific mortality in patients with systemic lupus erythem-
atosus. CVD ⫽ cardiovascular disease; CVA ⫽ cerebrovascular accident; IHD ⫽
ischemic heart disease.
tempt to explain the source of this heterogeneity, we per-
formed a univariate meta–regression analysis using the
cohort type, quality assessment, sample type, and mid-
point of enrollment period (after 2000 versus before). None
were statistically significant at P ⬍ 0.05.
CVD. When assessing cause-specific mortality, 2 studies
with a total of 14,284 SLE patients reported the risk of
mortality from CVD (including IHD and CVA) using SMRs
(8,21). Overall, there was a nearly 3-fold increase in the
risk of death from CVD (meta-SMR 2.72, 95% CI 1.83–
4.04) compared with the general population (Figure 4).
Two studies examined the mortality risk specifically of
IHD in SLE patients and found a significantly increased
risk of death (meta-SMR 2.28, 95% CI 1.29 – 4.02). Addi-
tionally, 3 studies that examined mortality risk due to CVA
were identified (8,21,22). A significantly increased risk of
death from CVA in patients with SLE was found (meta-
SMR 1.68, 95% CI 1.10 –2.55).
Malignancy. Two studies reporting CVD mortality also
reported mortality from malignancy (8,21). Overall, the
pooled estimates showed no increased risk of death from
malignancy in SLE patients (meta-SMR 1.16, 95% CI 0.57–
2.35) (Figure 4). Another study assessed the risk of mor-
tality due to malignancy by sex, but did not report an
overall SMR for malignancy (26). In this study of 116
patients with SLE, Nived et al (26) reported an ⬃2-fold
increase in the risk of death in men in their cohort (SMR
2.24, 95% CI 0.6 –5.7), while the risk in women was com-
parable to that of the general population (SMR 1.02, 95%
CI 0.4 –2.1). Because this study did not include an overall
613
SMR for malignancy for the entire cohort, we calculated
the weighted mean of the reported sex-specific SMRs to
estimate the overall SMR. When this was done, the new
pooled estimate for malignancy from all 3 studies (8,21,26)
was still not significant (meta-SMR 1.19, 95% CI 0.89 –
1.59).
Infection. In assessing the risk of mortality related to
infections, we identified 2 studies (8,21) with a total of
14,284 SLE patients. There was an ⬃5-fold increase in the
risk of death from infection in patients with SLE (meta-
SMR 4.98, 95% CI 3.92– 6.32) when compared with the
general population (Figure 4).
Renal disease. Only 1 study evaluated the risk of mor-
tality from renal disease (21). Overall, there was a nearly
8-fold increase in the risk of death in patients with SLE
(SMR 7.9, 95% CI 5.5–11) when compared with the gen-
eral population.
DISCUSSION
This is the first systematic review and meta-analysis of
published observational studies assessing all-cause and
cause-specific risk of mortality in patients with SLE. We
found that all-cause and cause-specific mortality were in-
creased in patients with SLE when compared to the gen-
eral population; overall, there was a 3-fold increase in the
risk of death in SLE patients (meta-SMR 2.98, 95% CI
2.32–3.83). Almost every study used in our meta-analysis
showed that cause-specific mortality was increased in pa-
614
tients with SLE when compared to the general population.
We found a 3-fold increased risk of cardiovascular death
(meta-SMR 2.72, 95% CI 1.83– 4.04), a 5-fold increased
risk of mortality due to infections (meta-SMR 4.98, 95% CI
3.92– 6.32), and an 8-fold increased risk of renal mortality
(SMR 7.9, 95% CI 5.5–11). Malignancy was the only spe-
cific cause of mortality that was not increased in our meta-
analysis (meta-SMR 1.19, 95% CI 0.9 –1.6).
A progressive decline in the risk of mortality (and there-
fore an increase in survival) among SLE patients over the
last 40 years has been well documented in a number of
long-term longitudinal studies (3,8,21). One possible rea-
son for this decrease in mortality is an increase in and
more aggressive use of immunosuppressive agents and
antimalarials and a more judicious use of corticosteroids
to decrease disease activity and limit damage (3,5). It can
also be presumed that mortality has improved with in-
creased recognition of milder cases of the disease. An
example of this is shown in the multinational study by
Bernatsky et al, in which cohorts derived from population-
based centers had the lowest SMRs (21). Furthermore, the
emergence of better antihypertensive and lipid-lowering
agents to treat comorbid conditions in SLE patients also
may have helped to improve mortality outcomes (5,8).
Finally, it was suggested that there has been more effective
recognition and treatment of infectious complications of
SLE in recent years (5,27). Despite these advancements in
improving survival of SLE patients, the risk of mortality
remains significantly higher than the general population.
We found no statistically significant difference between
the SMRs of men and women. Previous studies have sug-
gested higher mortality among men with SLE (28,29); how-
ever, Bernatsky et al highlighted that these studies did not
calculate mortality rates according to SMRs (21). SMRs
would allow a more precise comparison with age- and
sex-adjusted mortality rates among the general population.
In their cohort of SLE patients, Jacobsen et al found that
men had higher rates of death due to SLE end-stage disease
and causes unrelated to SLE or infection when compared
with women (10). Despite the prevailing belief that men
have a more severe form of the disease than women, this
study found no difference in mortality between the sexes.
Regardless, only 6 studies calculated sex-specific SMRs;
therefore, further research to compare mortality rates be-
tween men and women with SLE according to SMRs is
warranted.
Our data supported a nearly 3-fold increase in the risk of
death due to CVD among SLE patients. It has been recog-
nized that a chronic inflammatory state, as seen in patients
with SLE, plays an important role in atherosclerosis
(30,31). Despite the fact that many studies have demon-
strated a decline in the number of deaths due to SLE
disease activity over time (3,21), there has not been a
corresponding decline in the number of deaths due to CVD
in SLE patients (8,21,22). As observed by Urowitz et al (3),
despite more judicious use of immunosuppressive medi-
cations and improved management of CVD risk factors
(including hypertension and hyperlipidemia), the preva-
lence of coronary artery disease in SLE patients increased
significantly over time, reaching nearly 30% after 3 de-
cades in their study cohort. Therefore, traditional risk
Yurkovich et al
factors alone do not fully account for the accelerated rate
of atherosclerosis and CVD in SLE patients. It has been
demonstrated that disease-specific characteristics, such as
SLE duration and severity, contribute to the significantly
increased prevalence of atherosclerosis and consequently
the increased risk of death due to CVD, especially in the
subgroup of patients with highly active disease (32,33).
We did not find a statistically significant increase in the
risk of death due to malignancy in patients with SLE when
compared to the general population. However, in a land-
mark study, Bernatsky et al demonstrated a considerably
significant increase in the risk of death due to specific
malignancies, with SMRs of 2.1 for all hematologic can-
cers, 2.8 for non-Hodgkin’s lymphoma (NHL), and 2.3 for
lung cancer (21). Similarly, in an inception cohort of SLE
patients from Southern Sweden, Nived et al found a sig-
nificantly increased risk of death due to NHL and lung
cancer, with SMRs of 11.6 and 5.5, respectively (26).
Therefore, although the overall risk of death due to malig-
nancy does not appear significantly increased in SLE pa-
tients, the risk of death due to specific malignancies, par-
ticularly NHL, is significantly increased among SLE
patients when compared with the general population. This
is consistent with previously reported associations show-
ing an increased risk of developing NHL in SLE patients
(34 –36). A more recent study by Bernatsky et al has not
only reaffirmed this increased risk of NHL, in addition to
leukemia, vulvar cancer, lung cancer, and possibly thyroid
cancer, but also interestingly showed a decreased risk for
breast, endometrial, and possibly ovarian cancers (37). The
increased risk for certain cancers, most notably hemato-
logic malignancies, with a paradoxical decreased risk for
others, most notably gynecologic malignancies, may help
in the understanding of the pathogenesis of cancer in SLE
and thus deserves further research.
Our calculated meta-SMR demonstrated a nearly 5-fold
increase in the risk of death as a result of infection in SLE
patients when compared with the general population. It
has been well documented that SLE patients are at an
increased risk of developing and potentially dying from
infections that develop as a result of immunosuppression,
which is a significant side effect of nearly all medications
used to treat SLE (5,21).
We found only 1 study that calculated an SMR for renal
disease. Bernatsky et al found that there was a nearly
8-fold increase in the risk of mortality due to renal disease,
most commonly lupus nephritis, among patients with SLE
compared with the general population (21).
Our study has some limitations. We included cohorts
with different clinical settings, diagnostic criteria, ages at
enrollment, durations of SLE exposure, and study designs.
As recommended, we used the random-effects model to
include an estimate of variability (38); however, results
from the univariate meta–regression analysis using cohort
type, quality assessment, sample type, and enrollment pe-
riod were unable to explain the source of observed het-
erogeneity. Therefore, the heterogeneity in the study co-
horts included in our meta-analysis may be explained by
other factors that were not available in the primary studies
(e.g., disease duration and organ involvement). Further-
more, in this meta-analysis, the SMR evaluating the asso-
Mortality in SLE
ciation between SLE and both all-cause mortality and
cause-specific mortality was adjusted for age and sex only.
Although other confounding factors may influence the risk
of mortality in patients with SLE, there is no method for
adjusting the results of meta-analyses using SMRs. For
example, it has been established that there is a close cor-
relation between race and mortality in SLE patients;
however, this association appears to be complicated and
confounded by socioeconomic, sociocultural, and envi-
ronmental factors (5). Bernatsky et al (21) calculated race-
specific SMRs for white (SMR 1.4, 95% CI 1.2–1.7) and
African American (SMR 2.6, 95% CI 2.0 –2.4) patients from
their US cohort. These results indicated that African
American patients with SLE had a significantly increased
risk of mortality when compared with white patients in
the US. However, because this study was the only one
used in our meta-analysis that calculated race-specific
SMRs, we were unable to calculate specific pooled meta-
SMRs for race in our study. We would like to recommend
that race/ethnicity should be reported in future studies
assessing mortality in SLE.
The enrollment period of individual studies included in
our meta-analysis varied considerably, from 1950 –2008.
The historical cohorts did not provide an accurate contem-
porary estimate of SMRs in patients with SLE because they
did not capture the improvement in mortality in SLE pa-
tients that has occurred over the past 40 –50 years. There-
fore, our pooled meta-SMR, which included these out-
dated studies, may have contributed to an overly
pessimistic assessment of SLE mortality. Nevertheless,
over time, mortality has decreased in the general popula-
tion, particularly for CVD. However, as previously dis-
cussed, a similar level of improvement in mortality due to
CVD has not been seen in SLE patients (8,21,22). This may
at least partially explain why the SMRs of earlier cohorts
remain largely similar when compared with contemporary
cohorts. The studies on SMRs for SLE cohorts enrolled
during the last decade are limited and we believe that
further research on this is needed. Last, the few studies
available on malignancy may have limited the conclusion
on mortality risk due to malignancy. Moreover, competing
risks from death by other causes in historical cohorts,
where higher early mortality from infections occurred, are
important to consider prior to drawing definitive conclu-
sions. Further research on this topic with more recent
cohorts is needed.
In summary, our meta-analysis of published observa-
tional studies indicated that people with SLE have an
overall 3-fold increased risk of death when compared with
the general population. An examination of mortality due
to specific causes indicated a 3-fold increase in the risk of
death due to CVD, a 5-fold increase in the risk of death due
to infection, and an 8-fold increase in the risk of death due
to renal disease in patients with SLE when compared to
age- and sex-matched controls from the general popula-
tion. We did not find a significantly increased risk of death
due to malignancy among SLE patients. We found that the
risk of death between the sexes was approximately equal,
with no significant difference between men and women.
There currently is a paucity of recent comprehensive lon-
gitudinal studies that evaluate all-cause mortality and, in
615
particular, cause-specific mortality in patients with SLE.
Further research in this area is warranted.
AUTHOR CONTRIBUTIONS
All authors were involved in drafting the article or revising it
critically for important intellectual content, and all authors ap-
proved the final version to be published. Dr. Aviña-Zubieta had
full access to all of the data in the study and takes responsibility
for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Yurkovich, Vostretsova, Aviña-
Zubieta.
Acquisition of data. Yurkovich, Vostretsova, Aviña-Zubieta.
Analysis and interpretation of data. Yurkovich, Chen, Aviña-
Zubieta.
REFERENCES
1. Alamanos Y, Voulgari PV, Siozos C, Katsimpri P, Tsintzos S,
Dimou G, et al. Epidemiology of systemic lupus erythemato-
sus in Northwest Greece 1982-2001. J Rheumatol 2003;30:
731–5.
2. Hersh AO, Trupin L, Yazdany J, Panopalis P, Julian L, Katz P,
et al. Childhood-onset disease as a predictor of mortality in an
adult cohort of patients with systemic lupus erythematosus.
Arthritis Care Res (Hoboken) 2010;62:1152–9.
3. Urowitz MB, Gladman DD, Tom BD, Ibanez D, Farewell VT.
Changing patterns in mortality and disease outcomes for pa-
tients with systemic lupus erythematosus. J Rheumatol 2008;
35:2152– 8.
4. Uramoto KM, Michet CJ Jr, Thumboo J, Sunku J, O’Fallon
WM, Gabriel SE. Trends in the incidence and mortality of
systemic lupus erythematosus, 1950-1992. Arthritis Rheum
1999;42:46 –50.
5. Borchers AT, Keen CL, Shoenfeld Y, Gershwin ME. Surviving
the butterfly and the wolf: mortality trends in systemic lupus
erythematosus. Autoimmun Rev 2004;3:423–53.
6. Hopkinson N. Epidemiology of systemic lupus erythemato-
sus. Ann Rheum Dis 1992;51:1292– 4.
7. Ward MM. Cardiovascular and cerebrovascular morbidity and
mortality among women with end-stage renal disease attrib-
utable to lupus nephritis. Am J Kidney Dis 2000;36:516 –25.
8. Bjornadal L, Yin L, Granath F, Klareskog L, Ekbom A. Cardio-
vascular disease a hazard despite improved prognosis in pa-
tients with systemic lupus erythematosus: results from a
Swedish population based study 1964-95. J Rheumatol 2004;
31:713–9.
9. Chun BC, Bae SC. Mortality and cancer incidence in Korean
patients with systemic lupus erythematosus: results from the
Hanyang lupus cohort in Seoul, Korea. Lupus 2005;14:635– 8.
10. Jacobsen S, Petersen J, Ullman S, Junker P, Voss A, Rasmus-
sen JM, et al. Mortality and causes of death of 513 Danish
patients with systemic lupus erythematosus. Scand J Rheu-
matol 1999;28:75– 80.
11. Takkouche B, Etminan M, Montes-Martinez A. Personal use
of hair dyes and risk of cancer: a meta-analysis. JAMA 2005;
293:2516 –25.
12. Bhutta AT, Cleves MA, Casey PH, Cradock MM, Anand KJ.
Cognitive and behavioral outcomes of school-aged children
who were born preterm: a meta-analysis. JAMA 2002;288:
728 –37.
13. Avina-Zubieta JA, Thomas J, Sadatsafavi M, Lehman AJ, La-
caille D. Risk of incident cardiovascular events in patients
with rheumatoid arthritis: a meta-analysis of observational
studies. Ann Rheum Dis 2012;71:1524 –9.
14. Avina-Zubieta JA, Choi HK, Sadatsafavi M, Etminan M, Es-
daile JM, Lacaille D. Risk of cardiovascular mortality in pa-
tients with rheumatoid arthritis: a meta-analysis of observa-
tional studies. Arthritis Rheum 2008;59:1690 –7.
15. Tan EM, Cohen AS, Fries JF, Masi AT, McShane DJ, Rothfield
NF, et al. The 1982 revised criteria for the classification of
616
systemic lupus erythematosus. Arthritis Rheum 1982;25:
1271–7.
16. Thompson SG, Higgins JP. How should meta-regression analy-
ses be undertaken and interpreted? Stat Med 2002;21:1559 –
73.
17. Miller R. The jackknife: a review. Biometrika 1974;61:1–15.
18. Sterne JA, Egger M. Funnel plots for detecting bias in meta-
analysis: guidelines on choice of axis. J Clin Epidemiol 2001;
54:1046 –55.
19. Begg C, Berlin J. Publication bias: a problem in interpreting
medical data. J R Stat Soc 1988;54:629 –34.
20. Alamanos Y, Voulgari PV, Papassava M, Tsamandouraki K,
Drosos AA. Survival and mortality rates of systemic lupus
erythematosus patients in Northwest Greece: study of a 21-
year incidence cohort. Rheumatology (Oxford) 2003;42:
1122–3.
21. Bernatsky S, Boivin JF, Joseph L, Manzi S, Ginzler E, Gladman
DD, et al. Mortality in systemic lupus erythematosus. Arthri-
tis Rheum 2006;54:2550 –7.
22. Bernatsky S, Clarke A, Gladman DD, Urowitz M, Fortin PR,
Barr SG, et al. Mortality related to cerebrovascular disease in
systemic lupus erythematosus. Lupus 2006;15:835–9.
23. Mok CC, Kwok CL, Ho LY, Chan PT, Yip SF. Life expectancy,
standardized mortality ratios, and causes of death in six rheu-
matic diseases in Hong Kong, China. Arthritis Rheum 2011;
63:1182–9.
24. Mok CC, To CH, Ho LY, Yu KL. Incidence and mortality of
systemic lupus erythematosus in a Southern Chinese popula-
tion, 2000-2006. J Rheumatol 2008;35:1978 – 82.
25. To CH, Mok CC, Tang SS, Ying SK, Wong RW, Lau CS.
Prognostically distinct clinical patterns of systemic lupus ery-
thematosus identified by cluster analysis. Lupus 2009;18:
1267–75.
26. Nived O, Bengtsson A, Jonsen A, Sturfelt G, Olsson H. Malig-
nancies during follow-up in an epidemiologically defined
systemic lupus erythematosus inception cohort in Southern
Sweden. Lupus 2001;10:500 – 4.
27. Bernatsky S, Ramsey-Goldman R, Clarke AE. Malignancy risk
in systemic lupus: recent research and ongoing challenges.
Curr Rheumatol Rev 2008;4:59 – 62.
Yurkovich et al
28. Ward MM, Pyun E, Studenski S. Long-term survival in sys-
temic lupus erythematosus: patient characteristics associated
with poorer outcomes. Arthritis Rheum 1995;38:274 – 83.
29. Manger K, Manger B, Repp R, Geisselbrecht M, Geiger A,
Pfahlberg A, et al. Definition of risk factors for death, end
stage renal disease, and thromboembolic events in a mono-
centric cohort of 338 patients with systemic lupus erythema-
tosus. Ann Rheum Dis 2002;61:1065–70.
30. Goodson NJ, Solomon DH. The cardiovascular manifestations
of rheumatic diseases. Curr Opin Rheumatol 2006;18:135– 40.
31. Nikpour M, Urowitz MB, Gladman DD. Premature atheroscle-
rosis in systemic lupus erythematosus. Rheum Dis Clin North
Am 2005;31:329 –54.
32. Karp I, Abrahamowicz M, Fortin PR, Pilote L, Neville C,
Pineau CA, et al. Recent corticosteroid use and recent disease
activity: independent determinants of coronary heart disease
risk factors in systemic lupus erythematosus? Arthritis
Rheum 2008;59:169 –75.
33. Roman MJ, Shanker BA, Davis A, Lockshin MD, Sammaritano
L, Simantov R, et al. Prevalence and correlates of accelerated
atherosclerosis in systemic lupus erythematosus. N Engl
J Med 2003;349:2399 – 406.
34. Bernatsky S, Ramsey-Goldman R, Rajan R, Boivin JF, Joseph
L, Lachance S, et al. Non-Hodgkin’s lymphoma in systemic
lupus erythematosus. Ann Rheum Dis 2005;64:1507–9.
35. Mellemkjaer L, Andersen V, Linet MS, Gridley G, Hoover R,
Olsen JH. Non-Hodgkin’s lymphoma and other cancers among
a cohort of patients with systemic lupus erythematosus. Ar-
thritis Rheum 1997;40:761– 8.
36. Agudelo CA, Schumacher HR, Glick JH, Molina J. Non-Hodg-
kin’s lymphoma in systemic lupus erythematosus: report of 4
cases with ultrastructural studies in 2. J Rheumatol 1981;8:
69 –78.
37. Bernatsky S, Ramsey-Goldman R, Labrecque J, Joseph L,
Boivin JF, Petri M, et al. Cancer risk in systemic lupus: an
updated international multi-centre cohort study. J Autoim-
mun 2013;42:130 –5.
38. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring
inconsistency in meta-analyses. BMJ 2003;327:557– 60.