Systemic lupus erythematosus in Asia: is it more common and more severe?

Authors


: Associate Professor Julian Thumboo, Department of Rheumatology and Immunology, Singapore General Hospital, Outram Road, Singapore 169608. Email: julian.thumboo@sgh.com.sg

Abstract

Aim:  We summarized published data on prevalence and comparative disease manifestations of systemic lupus erythematosus (SLE) to determine if SLE is more common or more severe in Asians compared with Caucasians.

Methods:  A PubMed search was used to identify studies reporting prevalence or clinical manifestations (limited to comparative studies in multi-ethnic Asian populations or between Asians and Caucasians) of SLE among Asians.

Results:  We identified 18 studies of prevalence and six studies of comparative disease manifestations and/or prevalence (from 1725 retrieved abstracts). Prevalence (per 100,000) of SLE in community-based studies ranged from 3.2 to 70.4 in Asian studies and 46.7–161.0 in Asian migrant studies compared with 20.6–78.5 in North America and 24.6–91.0 in Europe. Studies in Asian migrants showed double the prevalence of SLE among Indians (2/2 studies) and Chinese (1/2 studies) as compared with Caucasians. Two studies found differences in clinical manifestations of SLE among Chinese, Malays and Indians living in the same country. Four comparative studies variously found more renal, central nervous system or other major organ involvement among Asians and Asian migrants than Caucasians with SLE over the disease course.

Conclusions:  SLE is generally not more prevalent in Asian than Caucasian populations, although it is more prevalent among Asian migrants. SLE has more severe clinical manifestations among Asians compared with Caucasians.

INTRODUCTION

Asia is the world's largest continent and home to almost half the world's 6272 million people,1 with China and India alone accounting for 21% and 17%, respectively, of the world's population. Marked regional differences in the prevalence and severity of SLE have been reported, with studies suggesting that SLE may be more prevalent and more severe in non-Caucasian populations, including Asians, African-Americans and Hispanics.2–4 Such epidemiological studies of the prevalence and clinical manifestations of SLE are important for several reasons. First, by estimating disease burden, they facilitate planning and provision of health care resources. Second, comparative studies of the prevalence or clinical manifestations of SLE may provide insights into pathogenesis and natural history (including the role of genetic and environmental factors) and may identify factors influencing disease severity, potentially contributing to clinical management. The aims of this review were thus (i) to summarize published data on SLE prevalence among Asians, and (ii) to evaluate if SLE is more common and more severe among Asians compared with Caucasians.

METHODS

We identified relevant studies from the computerized bibliographic database of the National Library of Medicine (Medline, 1966 to 15 July 2006) using PubMed (see Table 1 for details of the search strategy). We reviewed the abstracts and retrieved those articles (and relevant references from these articles) in English and other languages (where available) that reported data on prevalence or comparative clinical manifestations of SLE among Asians. As an excellent review of clinical manifestations of SLE among Asians by Navarra and King forms part of this themed issue (p. XX), this review focuses on comparative studies that directly compared clinical manifestations between Asian and Caucasian populations or among multiple ethnic groups within Asia. Studies were excluded if they were performed in subjects < 18 years of age, did not report prevalence or comparative clinical manifestations, or did not clearly define the target population or methods used for case identification. We reported prevalence as the number of cases per 100,000 population per year and calculated the 95% confidence interval (95% CI) of prevalence in studies where the denominator for computing prevalence was reported,5–20 setting a lower bound of 0% since prevalence cannot be negative. Studies were categorized as community-based versus hospital-based, and by ethnic groups of interest within each category.

Table 1.  Prevalence of systemic lupus erythematosus among Asians/Asian migrants (community-based studies)
Country/ascertainment strategyDenominator for computing prevalencePrevalence (95% CI‡) (per 100,000)
  • ANA, antinuclear antibodies; ACR, American College of Rheumatology.

  • PubMed search strategy: Keyword searches were performed using (lupus AND Asia) OR (lupus AND (Chin

  • *

    * OR India OR Japan OR Hong Kong OR Malaysia OR Vietnam OR Thailand OR Indonesia OR Brunei OR Taiwan OR Korea)). For each article, the following information was abstracted, where available: author(s), journal, year of publication, country/geographical area studied, study strategy (community- or hospital-based), number of study subjects, ethnic groups studied, incidence and/or prevalence estimates.

  • A lower bound of zero was set for the 95% confidence interval since prevalence values cannot be negative.

Asian studies – Chinese
Taiwan, China; screening questionnaire (29% response)7383626.0 (0, 77.0) (one case < 4 ACR criteria)
Taiwan, China; screening questionnaire (50% response)8899833.0 (0, 70.5)
Textile factory, Shanghai, China932,66870.4 (41.6, 99.2)
Beijing (north), Shantou (south), China; Village administration registers (80, 92% response)10a) Beijing: 4192
b) Shantou: 5057
Total = 9249
43.2 (0.8, 85.6) (4 cases)
Asian studies – Japanese
Japan, except Okinawa; screening questionnaire1111775.0 (0, 45.4)
Japan, Okinawa; screening questionnaire12a) 1972: 566
b) 1991: Not stated
a) 3.7 (0, 53.8)
b) 37.7
Japan, Fukuoka city and Hospitals; Six sources: National hospital registers, local hospital registers, private physicians, government registers, nation-wide survey, death certificates.131,002,21510.8 (8.8, 12.8)
Japan; Hospitals, General Practices38Not stated20.9
Asian studies – others
Saudi Arabia; Screening Questionnaire1410,37219.3 (0, 46.0)
North India; a) Screening ANA Assay
b) Screening Questionnaire15
a) 52,061
b) 39,826
Total = 91,887
3.2 (0, 6.9)
Asian migrant studies
San Francisco, USA; HMO inpatient and outpatient registers16121,444Chinese: 50.8 (38.1, 63.5) (same as Caucasians)
Birmingham (UK); Six sources: notification by attending and primary care physicians, the lupus patient support group, and hospital inpatient and laboratory data.17872,877Overall: 27.7 (24.2, 31.2)
Caucasian: 20.7 (17.7, 23.7)
Indian: 46.7 (42.2, 51.2)
Afro-Caribbean: 111.8 (104.8, 111.8)
Leicester (UK); Five sources: hospital, general practice, Lupus Society, laboratory18191,469Caucasians: 20.2 (13.8, 26.6)
Indians: 50.4 (40.3, 60.5)
Hawaii, US; Medical cases from medical facilities and a patient support group19Not statedChinese: 161.0
Fillipino: 104.0
Hawaiian: 53.0
Japanese: 81.0
Caucasian: 71.0

RESULTS

PubMed search

The majority of 1725 identified abstracts did not meet inclusion criteria as they were case reports/series, laboratory or genetic studies and were therefore excluded. This review was thus based on 24 articles (23 English and one Chinese language publication7) fulfilling criteria, with 18 studies reporting prevalence and six comparative studies reporting prevalence and/or clinical manifestations in adults aged 18 years or older.

SLE prevalence: community-based and hospital-based studies

Prevalence (per 100,000) of SLE in community-based studies ranged from 3.2–70.4 in studies in Asia (n = 10), and 46.7–161.0 in Asian migrant studies (n = 4) (Table 1). The majority of community-based studies in Asia were among subjects of Chinese or Japanese ethnicity, in which the prevalence ranged from 26.0–70.4 and 3.7–20.9, respectively. The 95% CI of point estimates of prevalence in these community-based studies generally overlapped. Of the four Asian migrant studies identified, two studies from the UK with multiple sources of ascertainment showed approximately double the prevalence of SLE in Indians as compared with Caucasians,15,16 which was statistically significant (i.e. non-overlapping 95% CI) in one study.16 Two studies comparing Caucasians with Chinese yielded differing results, with Chinese having the same prevalence (50.8, 95% CI 38.1–63.5) as Caucasians in one study among members of a Health Maintenance Organization in San Francisco,14 and more than twice the prevalence (161.0 vs. 71.0) in another study in Hawaii.17 The prevalence of SLE in the latter study was also higher among Filipinos (104.0) compared with Caucasians (71.0) but was similar among Japanese (81.0) and Caucasians.17

In five hospital-based studies, the prevalence of SLE (per 100 000) ranged from 12.0–1300.0 (Table 2). Two studies from Hong Kong reported differing prevalences of 58.8 (95% CI 54.9–62.7)18 and 250.0 (95% CI 207.2–292.8).19 This difference may be related to the use of different denominators for computing prevalence: the number of people living in the catchment area was used in the former, the number of medical referrals over a 2-year period in the latter.

Table 2.  Prevalence of systemic lupus erythematosus among Asians (hospital-based studies)
Country/ascertainment strategyDenominator for computing prevalencePrevalence (95% CI) (per 100,000)
  • ANA, antinuclear antibodies; ACR, American College of Rheumatology.

  • PubMed search strategy: Keyword searches were performed using (lupus and Asia) OR (lupus and (Chin

  • *

    * OR India OR Japan OR Hong Kong OR Malaysia OR Vietnam OR Thailand OR Indonesia OR Brunei OR Taiwan OR Korea)). For each article, the following information was abstracted, where available: author(s), journal, year of publication, country/geographical area studied, study strategy (community- or hospital-based), number of study subjects, ethnic groups studied, incidence and/or prevalence estimates.

  • A lower bound of zero was set for the 95% confidence interval since prevalence values cannot be negative.

Asian studies – Chinese
Hong Kong, China; retrospective chart review201.49 million (total adult population in the catchment areas of 2 hospitals under study)58.8 (54.9, 62.7)
Hong Kong, China; retrospective chart review21Medical referrals to two hospitals:
a) 44,897
b) 7426
Total = 52,323
250.0 (207.2, 292.8)
Asian Studies – others
South India; retrospective chart review39Not stated1300.0
Malaysia; retrospective chart review22Extrapolated from prevalence of amyotrophic lateral sclerosis which has a remarkably uniform rate throughout most parts of the world.Overall: 43.0
Chinese: 46.0
Malays: 26.0
Indians: 12.0
Pakistan; retrospective chart review40Not statedNumber of cases: 196 (49 cases < 3 ARC criteria)

By way of comparison, in a recent review,21 worldwide prevalence of SLE ranged from 20.6–78.5 in the USA and Canada, 24.6–91.0 in Europe (including UK), 13.4–89.3 in Australia and 64.2 in Martinique.

SLE severity: clinical manifestations

We identified six comparative studies on clinical manifestations of SLE, with two studies comparing different ethnic groups living in the same Asian country, one study comparing Asians in Singapore with Caucasians in America, three studies comparing Asian migrants and Caucasians, and no studies comparing different ethnic groups in Asia living in their respective home countries. These studies found ethnic differences in severity of disease manifestations among ethnic groups in Asia and more severe disease manifestations among Asians than Caucasians. However, the results of these studies need to be interpreted in the light of the many factors influencing clinical manifestations of SLE, and the fact that these manifestations can develop at or after diagnosis. Some of these studies controlled for some but not all factors that could potentially influence disease manifestations.

Two comparative studies were performed among Chinese, Malay and Indian SLE patients living in the same Asian country.20,22 In a study in Malaysia among 539 SLE patients,20 Malays and Indians were less likely to present with a malar rash than Chinese (57%vs. 21%vs. 64%, P < 0.001, without adjustment for duration of symptoms or gender). In a study in Singapore among 472 SLE patients,22 after adjusting for age, gender and duration of symptoms prior to diagnosis, as compared with Chinese, Malays were more likely to present with renal (odds ratio [OR] 2.26, 95% CI 1.21–4.21) or central nervous system (CNS) (OR 3.97, 95% CI 1.01–9.34) involvement while Indians were less likely to present with malar rash (OR 0.30, 95% CI 0.13–0.68) but more likely to present with oral ulcers (OR 2.90, 95% CI 1.45–7.34) at diagnosis.

In the one study comparing Asian and Caucasian SLE patients both living in their respective home countries (Caucasians in Rochester, Minnesota [n = 46] and Chinese in Singapore [n = 175]),23 after controlling for subject- and SLE-related factors, Caucasians were more likely than Chinese to have serositis (OR 3.11, 95% CI 1.01–9.71) or a haematologic disorder (OR 6.95, 95% CI 2.20–21.97), but far less likely to have a malar rash (OR 0.19, 95% CI 0.07–0.54) or positive antinuclear antibodies (OR 0.11, 95% CI 0.03–0.52) at diagnosis. After diagnosis, Caucasians were less likely to develop proteinuria (relative risk [RR] 0.47, 95% CI 0.19–1.15) or other major organ involvement (RR 0.22, 95% CI 0.05–1.04), compared with Chinese.23

Three comparative studies among Asian migrants and Caucasians were identified. In a follow-up study among 80 SLE patients (9 Indians) performed in the UK,24 at 5 years after diagnosis, neuropsychiatric damage measured using the Systemic Lupus International Collaborative Clinics/American College of Rheumatology (SLICC/ACR) Damage Index in Indians was significantly higher than in Caucasians (0.22 vs. 0.12, P = 0.027). In the same study, at 10 years after diagnosis, a higher prevalence of renal damage was found among Indians compared with Caucasians (57.1 vs. 20.8%, P = 0.006). In a retrospective community-based study in Leicester (UK), in 50 SLE patients (19 Indians, 31 Caucasians), proteinuria and neuropsychiatric disease were significantly more common in Indians compared with Caucasians (58%vs. 35%, P < 0.02; 38%vs. 13%, P < 0.01, respectively).14 Indians were also more likely than Caucasians to be on immunosuppressive therapy (46%vs. 20%, P < 0.05) and had a higher mortality rate (27%vs. 8%, P = 0.04). However, socioeconomic status and other variables were not controlled for in this study. An Australian study compared SLE disease activity between Caucasians (n = 121) and South-East Asian/Chinese (n = 59) patients admitted to a tertiary healthcare facility due to active disease.25 Compared with Caucasians, South-East Asian/Chinese were more likely to be diagnosed with WHO class III/IV glomerulonephritis both at presentation (23%vs. 54%, P = 0.04) and during subsequent admissions (27%vs. 64%, P = 0.02), readmitted for flare (41%vs. 63%, P = 0.03), and to have higher rates of severe organ damage as measured by renal (OR 4.9, 95% CI 3.1–8.1) and central nervous system (CNS) (OR 2.4, 95% CI 1.2–4.8) involvement and higher rates of usage of cyclophosphamide (OR 7.5, 95% CI 4.2–13.8) after adjusting for age and gender, but not duration of symptoms, prior to presentation.

DISCUSSION

In this review of published data on prevalence and severity of disease manifestations among Asians with SLE, we found that the prevalence of SLE was broadly similar among Asians and Caucasians in studies performed in their respective home countries. In contrast, the prevalence of SLE appeared to be higher among Asian migrants when compared with Caucasians in the same geographical location. Differences in severity of disease manifestations were observed among Chinese, Malays and Indian SLE patients sharing a common environment, as well as among Asians compared with Caucasians, with more renal and CNS involvement in Chinese and Indian than Caucasian patients.

In interpreting the prevalence data summarized in this study, the influence of methodological issues or actual differences (which in turn may be related to differences in genetic, environmental and other factors, which may also affect disease severity and are discussed below) need to be taken into account. Possible methodologic issues include the definition of SLE (the American College of Rheumatology 1982 criteria were used by studies reported here which were performed after 1982)26 and method/s of case ascertainment.15,17 Given the rarity of SLE, case ascertainment has generally involved use of a screening questionnaire (the Liang questionnaire often being used27) followed by clinical evaluation of subjects identified through the questionnaire. Case ascertainment using a screening ANA assay in the general population has been pioneered by Malaviya et al.13 Both these methods allow diagnosis of previously undetected cases of SLE, but are influenced by response rates and selection bias. Other approaches to case ascertainment have identified cases through hospital/community-based physicians, patient support groups or from laboratory results. These approaches would only allow detection of known or suspected cases of SLE and are also influenced by response rates and selection bias, as illustrated by the generally higher prevalence of SLE observed in hospital-based studies compared with community-based studies. Future studies could employ capture-recapture methods28 to determine the completeness of ascertainment. Of note, the factors listed above do not explain the higher prevalence of SLE found in the four studies in Asian migrants, because similar methods of ascertainment were used for ethnic groups assessed in each study.

Differences in SLE prevalence and/or severity among different ethnic groups in Asia and among Asians and Caucasians may be related to the complex interplay among subject-related factors such as gender,29 age at onset,30 socioeconomic status,31 genotype,32 environmental,3 behavioural factors,33 access to and compliance with medical care31 and comorbidities34 and SLE-related factors such as flares,35 autoantibody profiles36 and therapy.37 Thus some of these factors need to be adjusted for comparative studies to provide meaningful results. One aim of such studies is to identify patients at risk of severe disease, who would benefit from close monitoring and therapy.

There are several possible reasons for the higher prevalence of SLE observed among Asian migrants. First, Asian migrants could be genetically predisposed. For example, most Indian subjects developed SLE before migrating to the UK.14 Second, selection bias could have been introduced, for example, if severe SLE among migrants was more readily detected than mild SLE among Caucasians. Third, decreased survival in the home country could explain the observed lower prevalence for Asians compared with Asian migrants.

In conclusion, SLE is generally not more prevalent in Asian than Caucasian populations, although it is more prevalent among Asian migrants. However, SLE is clearly more severe among Asians compared with Caucasians, highlighting the need for a better understanding of factors influencing severity of SLE among Asians, which may lead to insights into the pathogenesis and treatment of this condition.

Ancillary