Life expectancy, standardized mortality ratios, and causes of death in six rheumatic diseases in Hong Kong, China

Authors


Abstract

Objective

To examine the life expectancy, standardized mortality ratios (SMRs), and causes of death in 6 groups of patients from Hong Kong with different rheumatic diseases.

Methods

Patients with a diagnosis of systemic lupus erythematosus (SLE), rheumatoid arthritis (RA), ankylosing spondylitis (AS), psoriatic arthritis (PsA), systemic vasculitis (SV), or systemic sclerosis (SSc) registered in 37 public hospitals between 1999 and 2008 were identified in the hospital registry. SMRs were calculated by comparing the mortality rate in patients with each disease with that in the general population. Life expectancy was calculated by abridged life-table analysis, and the causes of death were compared.

Results

In 2008, data on 8,367 RA, 5,243 SLE, 2,154 AS, 1,636 SV, 778 PsA, and 449 SSc patients were available in our registry. The age- and sex-adjusted SMRs were highest for SLE (5.25 [95% confidence interval 4.79–5.70]), SSc (3.94 [95% confidence interval 3.20–4.68]), and SV (2.64 [95% confidence interval 2.36–2.93]). In female patients, the loss in life expectancy was greatest for SSc (34.1 years), SV (19.3 years), and SLE (19.7 years). In male patients, the loss in life expectancy was highest for SV (28.3 years), SLE (27 years), and SSc (16 years). There were 2,486 deaths during the study period (1999–2008), and the principal causes were infections (28%), cardiovascular complications (18%), cancer (16%), and disease activity (7%). Infection was the leading cause of death in SLE, RA, AS, and PsA, whereas deaths from disease-related activity and cardiovascular complications were most frequent in SSc. Cancer was the most common cause of death in SV.

Conclusion

Our findings indicate that patients with SLE, RA, AS, PsA, SV, and SSc have increased mortality rates and reduced life expectancy. SLE has the highest adjusted SMR, and female SSc patients have the greatest loss in life expectancy. Infection is the leading cause of death, followed by cardiovascular complications and malignancies.

Rheumatic diseases are chronic multisystemic medical illnesses that are associated with significant mortality and morbidity. Organ dysfunction as a result of disease involvement is a major cause of reduced survival. Examples are renal failure in patients with systemic lupus erythematosus (SLE) and systemic vasculitis (SV), pulmonary hypertension and interstitial lung fibrosis causing cardiorespiratory failure in patients with systemic sclerosis (SSc), and extraarticular complications, such as heart, lung, and renal diseases, in patients with inflammatory arthritides, including rheumatoid arthritis (RA), ankylosing spondylitis (AS), and psoriatic arthritis (PsA). In addition to disease-related complications, adverse events related to treatment may aggravate the risk of mortality. Immunosuppressive therapies for various rheumatic diseases increase the incidence of common bacterial and viral, as well as opportunistic, infections and other long-term adverse effects, such as liver fibrosis, malignancies, premature menopause, and osteoporotic fractures.

Common to all of these rheumatic disorders is chronic inflammation in the joints and the visceral organs. Chronic inflammation is associated with the persistent elevation of levels of proinflammatory markers, such as tumor necrosis factor α, interleukin-6 (IL-6), and C-reactive protein (CRP), which are associated with an increased risk of atherosclerosis and arterial thrombosis (1–3). The standardized incidence ratios of coronary heart disease and cerebrovascular disease (stroke) are increased in patients with RA, SLE, AS, and PsA (4–7), and these cardiovascular complications are major causes of death, second to infection.

The cumulative survival rate of an inception cohort of patients who were diagnosed as having SLE between 1991 and 2003 in our geographic area was 92% 5 years after disease onset and 83% 10 years after disease onset (8). The unadjusted standardized mortality ratio (SMR) for our local SLE patients in the year 2006 was estimated to be 2.2 (95% confidence interval [95% CI] 0.7–6.7) (9). However, data on the life expectancy of SLE patients and on mortality rates in other non-SLE rheumatic diseases in our locality are unavailable. This prompted the current study, which used the hospital registry system and population census data to analyze the life expectancy and SMRs for 6 diseases that are commonly encountered in rheumatologic practice. The causes of death in patients with these diseases were also evaluated and compared.

PATIENTS AND METHODS

Data source.

Hong Kong is a small city situated in the southern part of China. In the year 2009, it had an estimated population of 7 million, 95% of whom were ethnic Chinese. All citizens of Hong Kong are entitled to receive heavily government-subsidized public medical services by paying a nominal fee for medical consultations in hospitals and outpatient clinics. These medical services are provided by 37 hospitals run by the Hospital Authority, and they account for >90% of the medical care provided in Hong Kong.

The numbers of patients with International Classification of Diseases, Ninth Revision (ICD-9), diagnostic codes for SLE, RA, AS, PsA, SV, and SSc who were registered and treated in 37 Hospital Authority hospitals between 1999 and 2008 were retrieved using the Clinical Data Analysis and Reporting System (CDARS). This central system captures data on both patients treated in outpatient clinics and inpatient hospital admissions. Patients with these disease codes who died during the specified period were identified using the same system. The causes of death, which were recorded in the death registry according to ICD-9 codes, were evaluated and compared across the 6 rheumatic diseases.

Determination of SMRs.

The SMR was used to compare the mortality risk of patients with any of the 6 rheumatic diseases to that of the general population of Hong Kong. SMR and its asymptotic 95% CI were calculated as follows:

equation image
equation image

where O is the observed number of deaths in the study population and E is the expected number of deaths (10).

For each of the 6 rheumatic diseases, the expected number of deaths was calculated as the total number of person-years contributed by the study population (i.e., patients with that disease) multiplied by the mortality rate of the general population. The age- and sex-adjusted SMRs were calculated similarly, except that the expected number of deaths was stratified by age and sex.

The size of the general population and the mortality rate in the general population from 1999 to 2008 were obtained from the Census and Statistics Department of the Hong Kong government, which maintains population census data from the years 1996, 2001, and 2006, with annual adjustments according to birth, death, and immigration/emigration status. The mean values of crude death rates over 10 years were used in the SMR analysis described above.

Life expectancy analysis.

Life expectancy for patients with each of the 6 rheumatic diseases was estimated by standard single-decrement life-table analysis as described by Preston et al (11), using CDARS prevalence and death data and Hong Kong population census data for 1999–2008. The following assumptions were made in the construction of period life tables for each sex and disease. First, the number of patients and the number of deaths for each age group in a given year were assumed to be independent of those in the subsequent year, and second, deaths were assumed to be symmetrically distributed in each age group. Under the first assumption, the age-specific death rates mx could be estimated by the maximum likelihood estimator for binomial distribution parameter, given by

equation image

where dmath image is the number of deaths for age group x in the kth year and nmath image is the number of patients for age group x in the kth year (12).

Although the second assumption might not be valid in patients younger than 5 years old, this did not affect the results of the life expectancy analysis, since no deaths were observed in age groups 0–1 year and 1–4 years for patients with the 6 diseases that were studied.

Analysis of the causes of death.

The principal causes of death, as coded by the attending physician in the death registry of the CDARS, were obtained for patients with the 6 diseases who died between 1999 and 2008. These were grouped under the following categories: infections, malignancies, cardiovascular and circulatory complications, cerebrovascular complications, gastrointestinal complications, hepatic complications, respiratory causes (excluding infections), renal failure, trauma and poisoning (including suicide), and related to rheumatic disease activity but with unspecified exact causes.

Statistical analysis.

Unless stated otherwise, values are expressed as the mean ± SD. Comparison of the frequencies of deaths due to various causes across different rheumatic diseases was performed by the chi-square test. P values less than 0.05 (2-tailed) were considered significant. All statistical analyses were performed using SPSS software, version 11.5 for Windows XP.

RESULTS

Demographic characteristics of the patients.

Table 1 shows the age, sex, and cumulative number of patients with the diagnostic codes for each of the 6 rheumatic diseases captured by the CDARS at the beginning of 2008 (including all new and old patients and living patients who were lost to followup). There were 8,367 RA patients, 5,243 SLE patients, 2,154 AS patients, 1,636 SV patients, 778 PsA patients, and 449 SSc patients. Their mean ages in 2008 were 44.3 years for the SLE patients, 56.4 years for the RA patients, 46.4 years for the AS patients, 51.7 years for the PsA patients, 51.8 years for the SSc patients, and 51.4 years for the SV patients. Patients with SLE, SSc, and RA were predominantly female, whereas patients with AS were predominantly male. There were slightly more women than men with SV, and the sex distribution was roughly equal for PsA.

Table 1. Demographic characteristics of the patients with each rheumatic disease*
 SLE (n = 5,243)RA (n = 8,367)AS (n = 2,154)PsA (n = 778)SSc (n = 449)SV (n = 1,636)
  • *

    Data are for all living patients with records in the hospital registry at the beginning of 2008. SLE = systemic lupus erythematosus; RA = rheumatoid arthritis; AS = ankylosing spondylitis; PsA = psoriatic arthritis; SSc = systemic sclerosis; SV = systemic vasculitis.

Women, no. (%)4,782 (91)6,657 (80)369 (17)354 (46)382 (85)1,021 (62)
Men, no. (%)461 (9)1,710 (20)1,785 (83)424 (54)67 (15)615 (38)
Ratio of female to male patients10.4:13.9:11:4.81:1.25.7:11.7:1
Age, mean ± SD years44.3 ± 15.156.4 ± 17.046.4 ± 14.951.7 ± 13.251.8 ± 17.251.4 ± 19.4

SMRs.

The SMRs for the 6 rheumatic diseases within the study period are shown in Table 2. The age- and sex-adjusted SMR was highest for SLE (5.25 [95% CI 4.79–5.70]), followed by SSc (3.94 [95% CI 3.20–4.68]), SV (2.64 [95% CI 2.36–2.93]), AS (1.87 [95% CI 1.61–2.13]), RA (1.68 [95% CI 1.59–1.77]), and PsA (1.59 [95% CI 1.16–2.03]). For women, the age-adjusted SMRs, in descending order, were as follows: SLE (5.59 [95% CI 5.07–6.12]), SSc (4.32 [95% CI 3.45–5.20]), SV (2.66 [95% CI 2.27–3.05]), PsA (1.96 [95% CI 1.14–2.77]), RA (1.72 [95% CI 1.61–1.82]), and AS (1.38 [95% CI 0.74–2.02]). The corresponding values for men were SLE (3.94 [95% CI 3.08–4.80]), SV (2.62 [95% CI 2.20–3.05]), SSc (2.59 [95% CI 1.32–3.87]), AS (1.94 [95% CI 1.65–2.22]), RA (1.59 [95% CI 1.42–1.76]), and PsA (1.40 [95% CI 0.89–1.90]). The mortality rates for patients with these rheumatic diseases were 59–425% higher than those of the age- and sex-matched general population. Of note, the adjusted SMR for SLE patients was much higher than the unadjusted value (5.25 versus 2.63), indicating a much greater impact of mortality on the younger population with this disease.

Table 2. Standardized mortality ratios (SMRs) for the 6 rheumatic diseases within the study period*
 SLERAASPsASScSV
  • *

    Values are the SMR (95% confidence interval). See Table 1 for other definitions.

Women2.88 (2.61–3.15)5.05 (4.73–5.37)2.07 (1.11–3.03)2.76 (1.61–3.91)8.55 (6.82–10.28)6.67 (5.70–7.65)
Men4.00 (3.13–4.87)4.97 (4.43–5.50)2.86 (2.44–3.28)2.27 (1.44–3.10)6.53 (3.33–9.73)6.82 (5.70–7.93)
All2.63 (2.41–2.86)4.58 (4.33–4.83)3.07 (2.64–3.50)2.50 (1.81–3.19)7.29 (5.93–8.65)6.47 (5.77–7.18)
Age-adjusted SMR (women)5.59 (5.07–6.12)1.72 (1.61–1.82)1.38 (0.74–2.02)1.96 (1.14–2.77)4.32 (3.45–5.20)2.66 (2.27–3.05)
Age-adjusted SMR (men)3.94 (3.08–4.80)1.59 (1.42–1.76)1.94 (1.65–2.22)1.40 (0.89–1.90)2.59 (1.32–3.87)2.62 (2.20–3.05)
Age- and sex-adjusted SMR5.25 (4.79–5.70)1.68 (1.59–1.77)1.87 (1.61–2.13)1.59 (1.16–2.03)3.94 (3.20–4.68)2.64 (2.36–2.93)

Results of life expectancy analysis.

Figures 1 and 2 show the life expectancy curves for the 6 rheumatic diseases, stratified according to sex, in comparison with the life expectancy for the general population within the same time period. The life expectancy curve for male patients with PsA could not be plotted because of the absence of deaths in the older age groups. For female patients, the loss in life expectancy at the time of birth compared with the general population was greatest for patients with SSc (34.1 years), followed by SV (19.3 years), SLE (19.7 years), RA (6.9 years), PsA (6.5 years), and AS (1.2 years). For male patients, the loss in life expectancy was greatest for patients with SV (28.3 years), followed by SLE (27.0 years), SSc (16.0 years), AS (7.0 years), and RA (5.2 years). The loss in life years appeared to be greatest for younger patients (particularly patients with an age at onset of disease of younger than 40 years).

Figure 1.

Life expectancy curves for female patients (A) and male patients (B) with rheumatoid arthritis (circles), systemic lupus erythematosus (squares), or systemic sclerosis (triangles), as compared to the life expectancy of the general population.

Figure 2.

Life expectancy curves for female patients (A) and male patients (B) with ankylosing spondylitis (circles), psoriatic arthritis (squares), or systemic vasculitis (triangles), as compared to the life expectancy of the general population (diamonds) from 1999 to 2008. The life expectancy curve for male patients with PsA could not be plotted because of the absence of deaths in the older age groups.

Mortality rates and causes of death.

Between January 1999 and December 2008, there were 2,486 deaths (1,705 [69%] women). The distribution of deaths for each of the underlying rheumatic diseases was 52% RA, 21% SLE, 13% SV, 8% AS, 4% SSc, and 2% PsA. The mean ± SD age at the time of death was 52.0 ± 17 years for SLE patients, 75.1 ± 11.0 years for RA patients, 68.6 ± 17.0 years for SV patients, 69.5 ± 11.4 years for AS patients, 67.2 ± 13.6 years for PsA patients, and 64.5 ± 15.2 years for SSc patients.

Table 3 shows the causes of death according to the ICD-9 codes recorded in the death registry. The principal causes of death were, in descending order of frequency, infections (28%), cardiovascular/cerebrovascular complications (18%), cancer (16%), disease activity (7%), renal failure (6%), pulmonary causes (6%), gastrointestinal and hepatic complications (4%), and accidents, injury, or poisoning (including suicide) (0.7%). The cause of death was recorded as “other” in 4% of the cases, and the cause of death was missing for 9%.

Table 3. Causes of death in the patients with rheumatic diseases (January 1999 to December 2008)*
Cause of deathSLE (n = 514)RA (n = 1,289)AS (n = 197)PsA (n = 51)SSc (n = 110)SV (n = 325)Total (n = 2,486)
  • *

    Values are the number (%) of patients. See Table 1 for definitions.

Infection114 (22)442 (34)57 (29)17 (33)19 (17)58 (18)707 (28)
 Septicemia25 (5)61 (5)4 (2)1 (2)2 (2)9 (3)
 Pneumonia72 (14)324 (25)43 (22)12 (24)12 (11)44 (14)
 Soft tissue3 (0.6)3 (0.2)0 (0)0 (0)0 (0)0 (0)
 Tuberculosis9 (1.8)10 (0.8)2 (1)1 (2)1 (1)1 (0.3)
 Other5 (1)44 (3)8 (4)3 (6)4 (4)4 (1)
Cancer61 (12)175 (14)34 (17)10 (20)11 (10)108 (33)399 (16)
Cardiovascular53 (10)160 (12)23 (12)10 (20)18 (16)34 (10)298 (12)
 Coronary heart disease29 (6)89 (7)9 (4.6)5 (10)8 (7)15 (5)
 Heart failure7 (1.4)45 (3.5)11 (5.6)5 (10)4 (4)14 (4)
 Arrhythmia4 (0.8)10 (0.8)1 (0.5)0 (0)0 (0)3 (1)
 Pulmonary embolism2 (0.4)3 (0.2)1 (0.5)0 (0)0 (0)0 (0)
 Pulmonary hypertension5 (1)0 (0)0 (0)0 (0)3 (3)0 (0)
 Other6 (1.2)13 (1)1 (0.5)0 (0)3 (3)2 (0.6)
Cerebrovascular33 (6.4)85 (6.6)9 (4.6)2 (4)3 (3)16 (5)148 (6)
Gastrointestinal24 (4.7)44 (3.4)7 (3.6)1 (2)2 (2)8 (2.5)86 (3)
Hepatic7 (1.4)16 (1.2)2 (1)2 (4)2 (2)4 (1)33 (1)
Respiratory system9 (1.8)84 (6.5)28 (14)1 (2)10 (9)19 (6)151 (6)
 Obstructive airway disease3 (0.6)37 (3)19 (10)1 (2)0 (0)8 (2.5)
 Bronchiectasis0 (0)15 (1.2)2 (1)0 (0)0 (0)4 (1)
 Pulmonary fibrosis4 (0.8)21 (1.6)1 (0.5)0 (0)8 (7)3 (1)
 Other2 (0.4)11 (0.9)6 (3)0 (0)2 (2)4 (1)
Renal failure44 (9)75 (6)9 (4.6)1 (2)5 (4.5)16 (5)150 (6)
Poisoning/injury7 (1.4)7 (0.5)1 (0.5)2 (4)0 (0)0 (0)17 (0.7)
Disease activity110 (21)28 (2.2)3 (2)0 (0)26 (24)5 (1.5)172 (7)
Other13 (2.5)49 (4)8 (4)3 (6)5 (4.5)23 (7)101 (4)
Unknown (missing data)39 (8)124 (10)16 (8)2 (4)9 (8)34 (10)224 (9)

Infection was the leading cause of death in SLE, RA, AS, and PsA, whereas disease-related and cardiovascular complications were the most frequent causes of death in patients with SSc. Cancer was the most frequent cause of death in patients with SV, while renal failure was the most frequent cause of death in SLE and RA.

DISCUSSION

This is the first study to examine the mortality statistics for different rheumatic disorders in a large number of patients from the Hong Kong central hospital registry. We demonstrated that the age- and sex-adjusted mortality rates in patients with RA, SLE, AS, PsA, SSc, and SV were significantly higher than those in the general population, by 0.6 to 4.3 times. The age- and sex-adjusted SMR was highest for SLE (5.3), SSc (3.9), and SV (2.6). For women, the loss in life expectancy at birth was greatest for patients with SSc (34 years), SV (19 years), and SLE (20 years). For men, the loss in life expectancy was greatest for patients with SV (28 years), SLE (27 years), and SSc (16 years). The most common cause of death in all of the diseases studied was infection, followed by cardiovascular complications, malignancies, and uncontrolled disease activity. The loss in life expectancy in patients with rheumatic diseases, especially during the early years of life, is an important issue because it causes a tremendous loss of economic productivity in the community.

Although there have been a number of previous studies of cumulative survival rates in patients with these rheumatic diseases, not many studies focusing on age-adjusted mortality rates compared to the general population have been published. A PubMed search revealed 6 studies of SMRs in SLE patients, 12 in RA patients, 3 in PsA patients, 3 in SSc patients, and 3 in SV patients (9, 13–38). There was only one study on SMR in AS patients, published in 1993 (39). The reported unadjusted SMRs for RA and SLE ranged from 1.15 to 2.64 (17–28) and from 2.17 to 3.81 (9, 13–17), respectively. SSc had the highest SMR, with results ranging from 1.46 to 4.29 (33–35). The SMR of antineutrophil cytoplasmic antibody–associated vasculitis ranged from 1.60 to 4.80 (36–38). The SMR of PsA patients was reported to be 1.62 in a cohort from the 1980s (32), but the SMR determined in the past decade was reported to be 0.56–0.82, i.e., not higher than that of the general population (30, 31).

Table 4 summarizes the SMRs for different rheumatic diseases found in different studies. The SMRs for the patients with SLE and SSc in the present study were similar to those found in patients from most other geographic areas. The SMR for patients with PsA in the present study was 2.50, and the mortality rate was significantly higher than that of the general population. The SMRs for patients with RA, AS, and SV in the present study were higher than those reported in the literature. After adjustment for age and sex, the SMRs for these 3 diseases ranged from 1.68 to 2.64. The reasons for the discrepancies in the SMRs between patients with RA, AS, and SV in the present study and those in previous studies are open to speculation. Undercoding of RA and AS in the present study was more likely to occur, since most patients are seen and treated in the outpatient and day-ward setting. In the present study, SV included all types of vasculitis as judged by the attending physician, including paraneoplastic vasculitis. This might have caused a higher mortality rate compared to those reported for antineutrophil cytoplasmic antibody–associated vasculitis.

Table 4. Standardized mortality ratios for various rheumatic diseases determined in the past decade*
Author, year (ref.)No. of patientsCountrySMR (95% CI)Study period
  • *

    SMR = standardized mortality ratio; 95% CI = 95% confidence interval; MPA = microscopic polyangiitis; WG = Wegener's granulomatosis; CSS = Churg-Strauss syndrome (see Table 1 for other definitions).

SLE    
 Present study5,686Hong Kong, China2.63 (2.41–2.86)1999–2008
 Hersh et al, 2010 (13)957US2.5 (2.0–3.2)2002–2007
 Urowitz et al, 2008 (14)389Canada3.81 (1.98–7.32)1997–2005
 Mok et al, 2008 (9)442Hong Kong, China2.17 (0.7–6.7)2000–2006
 Bernatsky et al, 2006 (15)9,547Multinational2.4 (2.3–2.5)1970–2001
 Chun and Bae, 2005 (16)434 (women)Korea3.02 (1.45–5.55)1992–2001
 Bjornadal et al, 2004 (17)4,737Sweden3.63 (3.49–3.78)1964–1995
RA    
 Present study9,469Hong Kong, China4.58 (4.33–4.83)1999–2008
 Nakajima et al, 2010 (18)7,926Japan1.46 to 1.902000–2007
 Bergstrom et al, 2009 (19)161Sweden1.15 (0.82–1.60)1995–2002
 Young et al, 2007 (20)1,429UK1.27 (1.04–1.46)1986–1997
 Gonzalez, 2007 (21)822US1.35 (1.23–1.49)1965–2005
 Sihvonen et al, 2004 (22)1,042Finland2.64 (2.63–2.68)1988–1999
 Minaur et al, 2004 (23)100UK2.13 (1.26–3.60)1957–1999
 Thomas et al, 2003 (24)41,344UK 1981–2000
  Male  2.07 (2.01–2.13) 
  Female  1.97 (1.93–2.01) 
 Gabriel et al, 2003 (25)609US1.27 (1.13–1.41)1955–1994
 Bjornadal et al, 2002 (26)61,899Sweden2.03 (2.00–2.05)1964–1995
 Gordon, 2001 (27)289UK1.30 (1.08–1.56)1986–1996
 Martinez et al, 2001 (28)182Spain1.851989–1998
 Chehata et al, 2001 (29)309UK1.65 (1.34–1.98)1981–1998
PsA    
 Present study822Hong Kong, China2.50 (1.81–3.19)1999–2008
 Buckley et al, 2010 (30)453UK0.82 (0.58–1.13)1985–2007
 Ali et al, 2007 (31)680Canada0.56 (0.14–2.25)1996–2004
 Wong et al, 1997 (32)428Canada1.62 (1.21–2.12)1978–1994
SSc    
 Present study547Hong Kong, China7.29 (5.93–8.65)1999–2008
 Hissaria et al, 2010 (33)786Australia1.46 (1.28–1.69)1993–2007
 Ioannidis et al, 2005 (34)1,645Multinational1.5 to 7.21982–1995
 Simeon et al, 2003 (35)79Spain4.29 (2.22–7.50)1976–1996
AS    
 Present study2,322Hong Kong, China3.07 (2.64–3.50)1999–2008
 Lehtinen, 1993 (39)398Finland1.51961–1993
SV    
 Present study1,907Hong Kong, China6.47 (5.77–7.18)1999–2008
 Eriksson et al, 2009 (36)63 (MPA/WG)Sweden1.6 (0.6–3.2)1997–2005
 Mohammad et al, 2009 (37)140 (MPA/WG/CSS)Sweden2.77 (2.02–3.71)1997–2006
 Lane et al, 2005 (38)99 (MPA/WG/CSS)UK4.8 (3.0–6.6)1988–2000

The causes of death in patients with the 6 rheumatic diseases investigated in this study were consistent with those found in most previous studies (13–39). Infections remained the most common immediate cause of death. This was the result of immunomodulatory therapies and the intrinsic immune aberrations of the diseases themselves, causing an increased susceptibility to invasion by microorganisms. Chronic inflammation leading to persistent elevation of cytokine levels, which is central to rheumatic disease activity, accelerates atherosclerosis in addition to traditional Framingham risk factors. This explains the contribution of cardiovascular and cerebrovascular complications as an important cause of death in these diseases. Uncontrolled or refractory disease manifestations, particularly for some diseases such as SSc, SLE, and SV, lead to organ damage and account for the decreased survival. On the other hand, the long-term use of immunosuppressive agents and defective immune surveillance increases the risk of malignancies, which is another important cause of death in rheumatic diseases.

Finally, it should be noted that while the factors that contribute to mortality mentioned above are disease-related, other factors, such as genetic background, environment, cultural beliefs, smoking, alcohol abuse, education level, socioeconomic status, medical and specialist referral system, delay in seeking medical advice, availability of immunization programs, and compliance with treatment are also involved. This may also explain the discrepancies in the mortality rates of the same group of diseases observed in different geographic areas.

The present study has some limitations. First, this is a study derived from data obtained from a hospital registry. Undercoding of the diagnoses among patients is bound to occur, whereas the number of deaths recorded in the registry is accurate and mandatory. This could have led to an underestimation of the total number of patients with the 6 diseases studied and subsequent overestimation of the SMRs. Second, as with all registry studies, the accuracy of the clinical diagnosis as judged by the attending physician cannot be verified. Third, we incorporated some assumptions into the calculation of the mortality ratios and life expectancy, such as uniform distribution of the number of deaths within the age ranges. Finally, since the hospital registry data management system only became available in the late 1990s, we did not have data on the trend of these mortality statistics in the last decade compared to two decades ago. Nevertheless, this is the first study to describe the mortality statistics for various chronic rheumatic diseases in our geographic area. This serves as an important reference for future survival analysis and epidemiology surveys in our community and in the field of rheumatology.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Mok 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. Mok, Kwok, Ho, Chan, Yip.

Acquisition of data. Mok, Kwok, Ho, Chan, Yip.

Analysis and interpretation of data. Mok, Kwok, Ho, Chan, Yip.

Acknowledgements

The authors would like to thank all of the rheumatologists and physicians working in public hospitals in Hong Kong who contributed to the management of patients with rheumatic diseases within the study period. We would also like to thank the Hospital Authority headquarters staff members who are responsible for maintaining the CDARS data system. Without this system and their technical support, this study would not have been successful.

Ancillary