Lupus
http://lup.sagepub.com/
Mortality and years of potential life loss in systemic lupus erythematosus: a population-based cohort study K Lerang, I-M Gilboe, D Steinar Thelle and J T Gran Lupus 2014 23: 1546 originally published online 10 September 2014 DOI: 10.1177/0961203314551083 The online version of this article can be found at: http://lup.sagepub.com/content/23/14/1546
Published by: http://www.sagepublications.com
Additional services and information for Lupus can be found at: Email Alerts: http://lup.sagepub.com/cgi/alerts Subscriptions: http://lup.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav
>> Version of Record - Nov 11, 2014 OnlineFirst Version of Record - Sep 10, 2014 What is This?
Downloaded from lup.sagepub.com at TEMPLE UNIV on November 17, 2014
Lupus (2014) 23, 1546–1552 http://lup.sagepub.com
LUPUS AROUND THE WORLD
Mortality and years of potential life loss in systemic lupus erythematosus: a population-based cohort study K Lerang1, I-M Gilboe1, D Steinar Thelle2,3 and JT Gran1,4 1
Department of Rheumatology, Oslo University Hospital, Rikshospitalet, Oslo, Norway; 2Department of biostatistics, Institute of Basic Medical Sciences, University of Oslo, Oslo, Norway; 3Department of Public Health and Community Medicine, Gothenburg University, Gothenburg, Sweden; and 4Institute of Clinical Medicine, University of Oslo, Oslo, Norway
Multiple sources were used to identify 325 systemic lupus erythematosus (SLE) patients within the city of Oslo during 1999–2009 who met 4 of the American College of Rheumatology (ACR) criteria. The survival, standard mortality rate (SMR), years of potential life loss before 60 years of age (YPLL60) and causes of death of these patients were examined and compared to a matched control population. Only inception cases (127) were studied in the calculation of survival. The analysis includes underlying, immediate and contributing causes of death. The five- and 10-year survival was 95% and 90%, respectively, which was significantly reduced when compared to the general population. A total of 50 SLE patients died during the study period. Overall SMR was 3.0 (95% confidence interval (CI) 2.2–3.8) with the highest SMR found for female patients aged 16–39 years old. SLE patients had a 10 times higher rate of YPLL60 compared to the control group. YPLL emphasizes active disease and reduces the importance of cancer as a cause of death in SLE. This study demonstrates that YPLL gives additional and useful information for the prognosis of SLE, supplementing traditional methods of measuring mortality. Lupus (2014) 23, 1546–1552. Key words: Systemic lupus erythematosus; epidemiology; mortality; outcome; survival rate
Introduction The changes in systemic lupus erythematosus (SLE) treatment during the 1950s and 1960s brought about tremendous improvements in survival.1 Globally, and at the same time, there has been a large decline of 40%–50% in the mortality rates of women 15–54 years old and a generally lower decline (15%–35%) for men 15–54 years old and an essential shift in causes of death.2 Considering allocation of resources in the management of SLE, it is therefore important to separate improvement in SLE from improvement that has affected the general population. In most studies on SLE, causes of death are determined through review of hospital charts, autopsy reports and contact with the physician who cared for the patient during the terminal illness.3 Although such methods give Correspondence to: Karoline Lerang, Department of Rheumatology, Oslo University Hospital, Rikshospitalet, Pb 4950 Nydalen, 0424 Oslo, Norway. Email: [email protected] Received 11 March 2014; accepted 14 August 2014
better quality of causes of death, they do not permit comparison of causes of death with the general population. This study aimed to measure the impact of SLE on years of potential life lost (YPLL) before 60 years of age, to examine causes of death within a population-based SLE cohort and to compare these with the general population.
Patients and methods Patients and controls The study included all (325) identifiable adult (16 years) SLE patients fulfilling four or more of the updated 1997 American College of Rheumatology (ACR) criteria,4 alive and living within the city of Oslo from January 1, 1999 to January 1, 2009. The cases were found by the use of multiple sources: inpatient and outpatient hospital discharge diagnosis registers, a local cohort from 1995,5 The Norwegian Systemic connective tissue disease and Vasculitis Registry (NOSVAR) and private
! The Author(s), 2014. Reprints and permissions: http://www.sagepub.co.uk/journalsPermissions.nav
Downloaded from lup.sagepub.com at TEMPLE UNIV on November 17, 2014
10.1177/0961203314551083
Mortality and years of potential life loss in SLE K Lerang et al.
1547
rheumatologists. Additionally, Norway’s Cause of Death Registry was reviewed, which did not add any new cases. Details of inclusion are described elsewhere.6 Of the 325 SLE patients, 129 adults were diagnosed with SLE during the study period (incident cases). Three SLE patients migrated out of Norway during the study period and were lost to follow-up. All incident cases were inception-cases, diagnosed at one of the hospitals in Oslo, except two patients who migrated to Oslo more than one year after the time of diagnosis. Some of the patients had been followed since 1986 as part of an earlier, non-population-based study in Oslo.5 In a sub-analysis including follow-up time since 1986, an additional 71 patients became inception cases. The patients were the same, except for the inclusion of one who died before 1999. The patients were each randomly assigned five population control individuals resident in Oslo matched for year of birth, sex, and parents’ ethnicity. Ethnicity was defined as the parents’ country of birth. If one parent was born in Norway, the ethnicity was defined as Norwegian. To allow comparison with a control group, the date and causes of death from January 1, 1999 through January 1, 2010 were retrieved from the data files at the Cause of Death Registry. Follow-up time Follow-up time was defined as the time interval from the first year after January 1, 1999 SLE patients were registered with an International Classification of Diseases, 10th revision (ICD-10) code of SLE, until time of death, emigration or end of the study period January 1, 2010. Juvenile SLE patients were included in the study after they turned 16 years. The controls’ observational period started at the same time as the corresponding SLE patients’. Data collection The information collected from medical records included year of diagnosis (as received by a physician), ACR criteria defined according to the guidelines,4 antiphospholipid syndrome according to the Sapporo criteria,7 reduced renal function (estimated glomerular filtration rate of less than 60 ml/min per 1.73 m2 or creatinine >100 m mol/l for men and >90 m mol/l for women) and specific medical treatment. Demographic data were collected from the National Population Register and hospital records. The first time the patient was registered with an ICD-10 code within the study period was found in
the inpatient and outpatient hospital discharge diagnosis registers. Immunological tests were assessed as described elsewhere.6 Population figures in Oslo according to year and age were retrieved from Statistics Norway.8 Time and causes of death for SLE patients and their controls were retrieved from the Cause of Death Registry at Statistics Norway, except for the causes of death in one SLE patient who had an unknown cause in the registry. Additionally, we collected data regarding active disease as a cause of death from the medical records, evaluated and reached consensus on by two researchers (IMG, KL). Other causes were classified as follows: Death from cardiovascular disease (CVD) was defined as all ICD-10 I00–I99 except pulmonary embolism and cerebral bleeding, but including sudden death R96. Death of infections was A00–B99, J10–J18, N39, M86, kidney disease all N00–N29 except infections, malignancy C00– C97, injury V0–Y3, respiratory diseases J40–J99, digestive system K00–K93, central nervous system (CNS) included psychiatry and dementia F00–F99, G03–G99. Study area All inhabitants in Oslo are registered in a governmental database (the National Population Register) with a unique 11-digit number. All deaths are reported by physicians who are required to complete a death certificate and, if an autopsy is performed, additionally by pathologists, and registered in the Cause of Death Registry. In Norway, most specialists are hospital based, and at the time of the study, there were two rheumatologists in private practice in Oslo. The health care system in Norway is almost entirely publicly funded with an all-encompassing insurance for all inhabitants. The study was approved by the Regional Committee for Medical Research Ethics and the National Data Inspectorate. Statistical analysis For survival analysis we computed the KaplanMeier survival probabilities and curves for cases and controls and for women and men. The log rank test was used to test for differences. To estimate the risk of death for SLE patients in the time period, we included deaths both in prevalent and incident cases in the calculation of standardized mortality ratio (SMR). The expected number of deaths was taken from the matched control population. In the SMR analysis the number of Lupus
Downloaded from lup.sagepub.com at TEMPLE UNIV on November 17, 2014
Mortality and years of potential life loss in SLE K Lerang et al.
1548
deaths was divided by number of years under observation in every sub-group studied. The 95% confidence interval (CI) of SMR was calculated using the exact method.9 Annual age-specific mortality rates were calculated from the number of deaths in SLE patients in that specific age group in the time period divided by the person-years for all inhabitants in Oslo in the same period. To process the variables we used SPSS 18 and Microsoft Excel 2010. We set the level of significance at p < 0.05. The YPLL was the numerical difference between a predetermined end-point age (defined as 60 years old) and the age at death for deaths that occurred prior to that end-point age. The age of 60 was chosen by approximation to the nearest 10 from median age of death. The YPLL for each death in SLE patients was summed up to represent the total YPLL for all SLE patients resident in Oslo 1999– 2010. The YPLL60 rate is found by dividing YPLL by years of observation for the population under age 60.
Figure 1 Kaplan-Meier estimated survival function for incident cases 1999–2010 captured within one year after diagnosis (n ¼ 127).
12
Results
SLE females 10
There were 127 inception cases in the study period with 812 years of follow-up. Five cases died within five years and an additional two within 10 years. Survival calculated with Kaplan-Meier (Figure 1) gave a five- and 10-year survival of 95.0% and 90.0%, respectively, for SLE and 98.8% and 96.0% for controls (p ¼ 0.02 and p ¼ 0.006 for difference between groups). The five- and 10-year survival rates were similar (94.5% and 90.0%) when we included follow-up time from 1986 in a subgroup of patients followed from an earlier nonpopulation-based study within Oslo (2493 years of follow-up and 29 dead among 200 inception cases). Five- and 10-year survival rates in male patients (78%; all died within five years) and 10-year survival in female patients (95%) were reduced compared to controls (p < 0.001 and p ¼ 0.03). Mortality Fifty (10/34 men and 40/292 women) of the SLE patients (15%) and 90 (12/170 men and 78/1460 women) of the controls (6%) who lived in Oslo in the study period died between 1999 and 2010, and their age distribution is shown in Figure 2. Of these, 94% were of European descent and 56% died in the
Number of deaths
Survival
Control females SLE males
8
Control males 6 4 2 0 16-29 30-39 40-49 50-59 60-69 70-79
80
Age at death, years
Figure 2 Distribution of age at death for male and female systemic lupus erythematosus (SLE) patients who died during 1999–2010 and their matched controls.
hospital. There were 2665 years of follow-up in SLE patients in the time period, and the mean follow-up time was eight years. The median disease duration at the time of death was 13.5 for female patients and 7.0 years for men. SMR for SLE compared to controls was 3.0 (95% CI 2.2–3.8). The values stratified for sex and age are shown in Table 1. The highest SMR in female patients was for 16–39 years of age and their case fatality rate was seven per 1000 annually. The crude annual
Lupus Downloaded from lup.sagepub.com at TEMPLE UNIV on November 17, 2014
Mortality and years of potential life loss in SLE K Lerang et al.
1549
Table 1 Standard mortality ratio (SMR), stratified for age and sex N deaths Age group: female 16–39 40–59 60–79 80 Total Age group: male 16–39 40–59 60–79 80 Total
SMR
CI
7 8 19 6 40
35.8 2.8 3.6 1.8 2.7
(15.7–70.8) (1.3–5.2) (2.3–5.6) (0.7–3.7) (2.0v3.7)
0 5 4 1 10
27.2 4.7 1.4 4.6
(10.0–60.4) (1.5–11.4) (0.07–6.7) (2.3–8.1)
N: number; CI: 95% confidence interval.
death rate for SLE was 1.0 (95% CI 0.8–1.3) per 100,000 Oslo inhabitants. Of those who died, the median age of death was 63 (women 64 and men 59) years for SLE patients and 77 years for the control group (p < 0.001). SLE patients who had or had not experienced lupus nephritis (LN) had a median age at death of 59 and 69 years, respectively (p ¼ 0.05 for difference). Characteristics according to age at death The characteristics of the patients who died very young (49 years old) and the total SLE group, are shown in Table 2. Seven SLE patients died before the age of 40. One had another comorbid condition unrelated to SLE and died after four years. The other six patients lived for 10 years or longer, all had had LN and three had a kidney transplant. In a subanalysis the seven patients who died at a young age were compared with 48 SLE patients who received diagnosis at a similar age, but survived (median follow-up time of 21 years). The patients who died had more LN (86% vs 39%), kidney transplant (43% vs 8%) and total American Rheumatism Association (ACR) criteria (7 vs 5). Causes of death Those who died among the SLE patients and the control individuals were registered with a mean of 2.4 (SD 1.3 and 1.2, respectively) causes of death, when SLE as a cause was removed from the analysis. Two SLE patients and no controls were registered with unknown cause of death. Of the 50 SLE patients who died, six patients (12%) had active
Table 2 Characteristics of all SLE patients and patients dying before 40 years of age compared to those who died after 49 years of age
Female Age at diagnosis, number
http://lup.sagepub.com/
Mortality and years of potential life loss in systemic lupus erythematosus: a population-based cohort study K Lerang, I-M Gilboe, D Steinar Thelle and J T Gran Lupus 2014 23: 1546 originally published online 10 September 2014 DOI: 10.1177/0961203314551083 The online version of this article can be found at: http://lup.sagepub.com/content/23/14/1546
Published by: http://www.sagepublications.com
Additional services and information for Lupus can be found at: Email Alerts: http://lup.sagepub.com/cgi/alerts Subscriptions: http://lup.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav
>> Version of Record - Nov 11, 2014 OnlineFirst Version of Record - Sep 10, 2014 What is This?
Downloaded from lup.sagepub.com at TEMPLE UNIV on November 17, 2014
Lupus (2014) 23, 1546–1552 http://lup.sagepub.com
LUPUS AROUND THE WORLD
Mortality and years of potential life loss in systemic lupus erythematosus: a population-based cohort study K Lerang1, I-M Gilboe1, D Steinar Thelle2,3 and JT Gran1,4 1
Department of Rheumatology, Oslo University Hospital, Rikshospitalet, Oslo, Norway; 2Department of biostatistics, Institute of Basic Medical Sciences, University of Oslo, Oslo, Norway; 3Department of Public Health and Community Medicine, Gothenburg University, Gothenburg, Sweden; and 4Institute of Clinical Medicine, University of Oslo, Oslo, Norway
Multiple sources were used to identify 325 systemic lupus erythematosus (SLE) patients within the city of Oslo during 1999–2009 who met 4 of the American College of Rheumatology (ACR) criteria. The survival, standard mortality rate (SMR), years of potential life loss before 60 years of age (YPLL60) and causes of death of these patients were examined and compared to a matched control population. Only inception cases (127) were studied in the calculation of survival. The analysis includes underlying, immediate and contributing causes of death. The five- and 10-year survival was 95% and 90%, respectively, which was significantly reduced when compared to the general population. A total of 50 SLE patients died during the study period. Overall SMR was 3.0 (95% confidence interval (CI) 2.2–3.8) with the highest SMR found for female patients aged 16–39 years old. SLE patients had a 10 times higher rate of YPLL60 compared to the control group. YPLL emphasizes active disease and reduces the importance of cancer as a cause of death in SLE. This study demonstrates that YPLL gives additional and useful information for the prognosis of SLE, supplementing traditional methods of measuring mortality. Lupus (2014) 23, 1546–1552. Key words: Systemic lupus erythematosus; epidemiology; mortality; outcome; survival rate
Introduction The changes in systemic lupus erythematosus (SLE) treatment during the 1950s and 1960s brought about tremendous improvements in survival.1 Globally, and at the same time, there has been a large decline of 40%–50% in the mortality rates of women 15–54 years old and a generally lower decline (15%–35%) for men 15–54 years old and an essential shift in causes of death.2 Considering allocation of resources in the management of SLE, it is therefore important to separate improvement in SLE from improvement that has affected the general population. In most studies on SLE, causes of death are determined through review of hospital charts, autopsy reports and contact with the physician who cared for the patient during the terminal illness.3 Although such methods give Correspondence to: Karoline Lerang, Department of Rheumatology, Oslo University Hospital, Rikshospitalet, Pb 4950 Nydalen, 0424 Oslo, Norway. Email: [email protected] Received 11 March 2014; accepted 14 August 2014
better quality of causes of death, they do not permit comparison of causes of death with the general population. This study aimed to measure the impact of SLE on years of potential life lost (YPLL) before 60 years of age, to examine causes of death within a population-based SLE cohort and to compare these with the general population.
Patients and methods Patients and controls The study included all (325) identifiable adult (16 years) SLE patients fulfilling four or more of the updated 1997 American College of Rheumatology (ACR) criteria,4 alive and living within the city of Oslo from January 1, 1999 to January 1, 2009. The cases were found by the use of multiple sources: inpatient and outpatient hospital discharge diagnosis registers, a local cohort from 1995,5 The Norwegian Systemic connective tissue disease and Vasculitis Registry (NOSVAR) and private
! The Author(s), 2014. Reprints and permissions: http://www.sagepub.co.uk/journalsPermissions.nav
Downloaded from lup.sagepub.com at TEMPLE UNIV on November 17, 2014
10.1177/0961203314551083
Mortality and years of potential life loss in SLE K Lerang et al.
1547
rheumatologists. Additionally, Norway’s Cause of Death Registry was reviewed, which did not add any new cases. Details of inclusion are described elsewhere.6 Of the 325 SLE patients, 129 adults were diagnosed with SLE during the study period (incident cases). Three SLE patients migrated out of Norway during the study period and were lost to follow-up. All incident cases were inception-cases, diagnosed at one of the hospitals in Oslo, except two patients who migrated to Oslo more than one year after the time of diagnosis. Some of the patients had been followed since 1986 as part of an earlier, non-population-based study in Oslo.5 In a sub-analysis including follow-up time since 1986, an additional 71 patients became inception cases. The patients were the same, except for the inclusion of one who died before 1999. The patients were each randomly assigned five population control individuals resident in Oslo matched for year of birth, sex, and parents’ ethnicity. Ethnicity was defined as the parents’ country of birth. If one parent was born in Norway, the ethnicity was defined as Norwegian. To allow comparison with a control group, the date and causes of death from January 1, 1999 through January 1, 2010 were retrieved from the data files at the Cause of Death Registry. Follow-up time Follow-up time was defined as the time interval from the first year after January 1, 1999 SLE patients were registered with an International Classification of Diseases, 10th revision (ICD-10) code of SLE, until time of death, emigration or end of the study period January 1, 2010. Juvenile SLE patients were included in the study after they turned 16 years. The controls’ observational period started at the same time as the corresponding SLE patients’. Data collection The information collected from medical records included year of diagnosis (as received by a physician), ACR criteria defined according to the guidelines,4 antiphospholipid syndrome according to the Sapporo criteria,7 reduced renal function (estimated glomerular filtration rate of less than 60 ml/min per 1.73 m2 or creatinine >100 m mol/l for men and >90 m mol/l for women) and specific medical treatment. Demographic data were collected from the National Population Register and hospital records. The first time the patient was registered with an ICD-10 code within the study period was found in
the inpatient and outpatient hospital discharge diagnosis registers. Immunological tests were assessed as described elsewhere.6 Population figures in Oslo according to year and age were retrieved from Statistics Norway.8 Time and causes of death for SLE patients and their controls were retrieved from the Cause of Death Registry at Statistics Norway, except for the causes of death in one SLE patient who had an unknown cause in the registry. Additionally, we collected data regarding active disease as a cause of death from the medical records, evaluated and reached consensus on by two researchers (IMG, KL). Other causes were classified as follows: Death from cardiovascular disease (CVD) was defined as all ICD-10 I00–I99 except pulmonary embolism and cerebral bleeding, but including sudden death R96. Death of infections was A00–B99, J10–J18, N39, M86, kidney disease all N00–N29 except infections, malignancy C00– C97, injury V0–Y3, respiratory diseases J40–J99, digestive system K00–K93, central nervous system (CNS) included psychiatry and dementia F00–F99, G03–G99. Study area All inhabitants in Oslo are registered in a governmental database (the National Population Register) with a unique 11-digit number. All deaths are reported by physicians who are required to complete a death certificate and, if an autopsy is performed, additionally by pathologists, and registered in the Cause of Death Registry. In Norway, most specialists are hospital based, and at the time of the study, there were two rheumatologists in private practice in Oslo. The health care system in Norway is almost entirely publicly funded with an all-encompassing insurance for all inhabitants. The study was approved by the Regional Committee for Medical Research Ethics and the National Data Inspectorate. Statistical analysis For survival analysis we computed the KaplanMeier survival probabilities and curves for cases and controls and for women and men. The log rank test was used to test for differences. To estimate the risk of death for SLE patients in the time period, we included deaths both in prevalent and incident cases in the calculation of standardized mortality ratio (SMR). The expected number of deaths was taken from the matched control population. In the SMR analysis the number of Lupus
Downloaded from lup.sagepub.com at TEMPLE UNIV on November 17, 2014
Mortality and years of potential life loss in SLE K Lerang et al.
1548
deaths was divided by number of years under observation in every sub-group studied. The 95% confidence interval (CI) of SMR was calculated using the exact method.9 Annual age-specific mortality rates were calculated from the number of deaths in SLE patients in that specific age group in the time period divided by the person-years for all inhabitants in Oslo in the same period. To process the variables we used SPSS 18 and Microsoft Excel 2010. We set the level of significance at p < 0.05. The YPLL was the numerical difference between a predetermined end-point age (defined as 60 years old) and the age at death for deaths that occurred prior to that end-point age. The age of 60 was chosen by approximation to the nearest 10 from median age of death. The YPLL for each death in SLE patients was summed up to represent the total YPLL for all SLE patients resident in Oslo 1999– 2010. The YPLL60 rate is found by dividing YPLL by years of observation for the population under age 60.
Figure 1 Kaplan-Meier estimated survival function for incident cases 1999–2010 captured within one year after diagnosis (n ¼ 127).
12
Results
SLE females 10
There were 127 inception cases in the study period with 812 years of follow-up. Five cases died within five years and an additional two within 10 years. Survival calculated with Kaplan-Meier (Figure 1) gave a five- and 10-year survival of 95.0% and 90.0%, respectively, for SLE and 98.8% and 96.0% for controls (p ¼ 0.02 and p ¼ 0.006 for difference between groups). The five- and 10-year survival rates were similar (94.5% and 90.0%) when we included follow-up time from 1986 in a subgroup of patients followed from an earlier nonpopulation-based study within Oslo (2493 years of follow-up and 29 dead among 200 inception cases). Five- and 10-year survival rates in male patients (78%; all died within five years) and 10-year survival in female patients (95%) were reduced compared to controls (p < 0.001 and p ¼ 0.03). Mortality Fifty (10/34 men and 40/292 women) of the SLE patients (15%) and 90 (12/170 men and 78/1460 women) of the controls (6%) who lived in Oslo in the study period died between 1999 and 2010, and their age distribution is shown in Figure 2. Of these, 94% were of European descent and 56% died in the
Number of deaths
Survival
Control females SLE males
8
Control males 6 4 2 0 16-29 30-39 40-49 50-59 60-69 70-79
80
Age at death, years
Figure 2 Distribution of age at death for male and female systemic lupus erythematosus (SLE) patients who died during 1999–2010 and their matched controls.
hospital. There were 2665 years of follow-up in SLE patients in the time period, and the mean follow-up time was eight years. The median disease duration at the time of death was 13.5 for female patients and 7.0 years for men. SMR for SLE compared to controls was 3.0 (95% CI 2.2–3.8). The values stratified for sex and age are shown in Table 1. The highest SMR in female patients was for 16–39 years of age and their case fatality rate was seven per 1000 annually. The crude annual
Lupus Downloaded from lup.sagepub.com at TEMPLE UNIV on November 17, 2014
Mortality and years of potential life loss in SLE K Lerang et al.
1549
Table 1 Standard mortality ratio (SMR), stratified for age and sex N deaths Age group: female 16–39 40–59 60–79 80 Total Age group: male 16–39 40–59 60–79 80 Total
SMR
CI
7 8 19 6 40
35.8 2.8 3.6 1.8 2.7
(15.7–70.8) (1.3–5.2) (2.3–5.6) (0.7–3.7) (2.0v3.7)
0 5 4 1 10
27.2 4.7 1.4 4.6
(10.0–60.4) (1.5–11.4) (0.07–6.7) (2.3–8.1)
N: number; CI: 95% confidence interval.
death rate for SLE was 1.0 (95% CI 0.8–1.3) per 100,000 Oslo inhabitants. Of those who died, the median age of death was 63 (women 64 and men 59) years for SLE patients and 77 years for the control group (p < 0.001). SLE patients who had or had not experienced lupus nephritis (LN) had a median age at death of 59 and 69 years, respectively (p ¼ 0.05 for difference). Characteristics according to age at death The characteristics of the patients who died very young (49 years old) and the total SLE group, are shown in Table 2. Seven SLE patients died before the age of 40. One had another comorbid condition unrelated to SLE and died after four years. The other six patients lived for 10 years or longer, all had had LN and three had a kidney transplant. In a subanalysis the seven patients who died at a young age were compared with 48 SLE patients who received diagnosis at a similar age, but survived (median follow-up time of 21 years). The patients who died had more LN (86% vs 39%), kidney transplant (43% vs 8%) and total American Rheumatism Association (ACR) criteria (7 vs 5). Causes of death Those who died among the SLE patients and the control individuals were registered with a mean of 2.4 (SD 1.3 and 1.2, respectively) causes of death, when SLE as a cause was removed from the analysis. Two SLE patients and no controls were registered with unknown cause of death. Of the 50 SLE patients who died, six patients (12%) had active
Table 2 Characteristics of all SLE patients and patients dying before 40 years of age compared to those who died after 49 years of age
Female Age at diagnosis, number
