Epidemiology of Lupus

Introduction

Epidemiological studies are essential to understand the overall burden of a disease, make policy decisions on healthcare utilization and cost and examine and mitigate disparities of disease burden. In a disease as heterogeneous as systemic lupus erythematosus (SLE), defining what counts as a case represents a basic challenge for epidemiological study. To meet this challenge, several SLE classification criteria have been created that are used as the basis for qualifying individual cases for inclusion into research studies, including epidemiological studies. The most widely used criteria are the 1982 ACR criteria, which received a revision in 1997 (Table 5-1). The ACR criteria include 19 clinical signs or laboratory findings grouped in 11 categories. An individual with signs or findings from 4 or more categories classifies as having SLE. In 2012, the SLE International Cooperating Clinics (SLICC) group developed a set of criteria (Table 5-2) to address the perceived shortcomings of the 1997 ACR criteria, including an overemphasis on cutaneous symptoms and insufficient emphasis on the presence of ANAs. The SLICC criteria have 17 categories (11 clinical and 6 immunologic). To be classified as having SLE, an individual must satisfy at least 4 criteria with at least 1 clinical and 1 immunologic criterion, or must have biopsy-proven lupus nephritis (LN) and be seropositive for ANAs or anti-dsDNA antibodies (see Pathogenesis and Pathophysiology for a discussion on ANAs and other SLE-associated antibodies). In 2019, EULAR and ACR collaboratively developed the newest set of SLE criteria (Figure 5-1), distinguished from earlier classification systems in that the presence of ANAs is counted as an essential “entry criterion”. Other clinical and immunologic criteria, called “additive” criteria, each have a point weight of 2-10 points. A patient is classified as having SLE if the entry criterion is satisfied and if the point total of additive criteria is 10 or above. Because they have only recently been introduced, the 2019 EULAR/ACR criteria have not yet been used in epidemiological studies.

Epidemiological data can be collected from a variety of sources, including surveillance registries, population surveys, death registries, clinic and hospital records and insurance claims databases. To reliably estimate the burden of SLE, high-quality data sources, robust surveillance systems and sound research methodologies, all of which may be lacking in low- and middle-income countries (LMICs), are essential. It thus remains difficult to estimate the global burden of SLE and to provide intra- or inter-regional comparisons. Nevertheless, increasing amount of data is becoming available on SLE prevalence (the overall number of cases in a defined timeframe), incidence (the number of new cases in a defined timeframe) and mortality. This module presents and reviews the burden of SLE in the United States (US) and globally based on currently available data.

Enlarge 

Enlarge 

Enlarge 

Prevalence of SLE in the United States

The collection of epidemiological data on SLE in the US has greatly benefited from the establishment, in 2002, of five population-based surveillance registries by the Centers for Disease Control (CDC). To provide an ethnically-diverse surveillance population, the registry sites included the Fulton and DeKalb counties (including Atlanta) in Georgia, Washtenaw and Wayne counties (including Detroit) in Michigan, New York county (Manhattan) in New York and San Francisco county (the city of San Francisco) in California. The fifth registry was focused on Native Americans and Alaskans, with data derived from the Indian Health Service (IHS) facilities in Alaska, Arizona (Phenix) and Oklahoma (Oklahoma City).

The first data from these registries were published in 2014; analyses of the 2002-2004 data reported an age-adjusted prevalence rate of 73 cases and 72.8 cases per 100,000 individuals for Georgia and Michigan, respectively. An analysis of the 2007-2009 data from the IHS registry, also published in 2014, reported an age-adjusted prevalence of 178 per 100,000 person-years. In 2017, two analyses of the 2007-2009 data from the remaining registries were published, reporting an age-adjusted prevalence of 84.8 per 100,000 in San Francisco and 62.2 per 100,000 person-years in Manhattan. A meta-analysis of data from the four state-specific registries estimated a pooled prevalence of 72.8 cases per 100,000 person years (95% confidence interval [CI] 65.3-81.0), translating to an estimated 204,295 individuals (95% CI 160,902–261,725) living with SLE in the US in 2018. All of these analyses used the 1997 ACR criteria to classify patients as having SLE. Data from all 5 registries is summarized in Table 3-1. An analysis of Medicare 20% random sample data found that the prevalence of SLE may be increasing, from an age- and sex-adjusted prevalence rate (per 100,000 Medicare population) of 301.1 in 2009 to 366.6 in 2016. This increase in prevalence may represent a global trend and may be partially due to increases in life expectancy.

Enlarge 

Incidence of SLE in the United States

The five CDC registries also provide valuable data on the incidence of SLE among diverse US populations (see Table 3-1 for a summary). Published reports indicate an age-adjusted annual incidence rate per 100,000 person-years of 5.6 in the Georgia registry (years 2002-2004), 5.5 in the Michigan registry (years 2002-2004), 7.4 in the IHS registry (years 2007-2009), 4.6 in the San Francisco registry (years 2007-2009), and 4.6 in the Manhattan registry (years 2007-2009). In the analysis of Medicare 20% random sample data, the incidence (per 100,000 Medicare population) was 46.9 in 2009 and 49.0 in 2016. Thus, incidence rates appear to stable, unlike the increasing prevalence rates.

Mortality in the United States

With the advances in diagnosis and treatment during the second half of the 20^th century, the 10-year survival in patients with SLE has improved from ~60% in the 1950s to >90% in the 1980s. A comprehensive analysis of multi-decade trends in SLE mortality in the US was published in 2017. Using data from the CDC’s National Vital Statistics System mortality database (representing >99% of deaths of US residents), the authors identified 50,249 death attributed to SLE (ie, deaths with SLE listed on the death certificate) and 100,851,288 non-SLE related deaths (ie, deaths without SLE listed on the death certificate) in the 46 years from 1968 to 2013. The overall age standardized mortality rate (ASMR) per 100,000 persons for SLE was 0.45 in 1968 and decreased to 0.34 in 2013, a 24.4% reduction. The non-SLE ASMR decreased from 1303.1 to 731.6 per 100,000 persons in the same period. While the non-SLE ASMR pattern was one of continuous decrease, the SLE ASMR initially (1968-1975) decreased, then increased or remained constant (1976-1999), and finally started decreasing again (2000-2013) (Figure 3-1). Several explanatory factors have been proposed for this, including changes in incidence and changes in the availability of diagnostic tests, classification criteria and therapeutic options (and associated adverse events).

Enlarge 

Prevalence of SLE Worldwide

The availability and quality of epidemiological data for regions outside North America is comparatively limited, though gradually improving. A comprehensive review published in 2021 collated a large number of epidemiological studies on SLE worldwide. The estimated prevalence ranges for countries for which data is available are shown in Figure 3-2. European prevalence estimates range from 29 to 210 per 100,000 individuals in Malta and Spain, respectively, reflecting both regional variation and differences in methodology and data quality. Similarly divergent prevalence rates per 100,000 persons have been reported in Asia (from 36 to 103 in the Anhui province of China and the United Arab Emirates, respectively), South America (from 34.9 to 91.9 in the Tucumán state of Argentina and in Colombia, respectively) and Africa (from 601.3 to 7,713.5 per 100,000 hospitalized patients in Cameroon and Senegal, respectively; note that these numbers are much higher because they derive from hospital-based populations, hospital records being among the few available epidemiological data in Africa).

A comprehensive systematic analysis of 112 publications with epidemiological data, covering 20% of countries worldwide, estimated a global prevalence of SLE of 43.7 per 100,000 individuals, translating to ~3.4 million people globally.A meta-analysis of 23 studies from 18 LMICs (including Argentina, Brazil, China, Cuba, Colombia, Democratic Republic Congo, Ecuador, Egypt, India, Kenya, Malaysia, Mexico, Nigeria, Pakistan, Turkey, Ukraine, Venezuela and Zimbabwe) found a pooled SLE prevalence of 103 per 100,000 persons.

Enlarge 

Incidence of SLE Worldwide

Estimates of worldwide SLE incidence are associated with the same methodological hurdles as that of prevalence. The estimated incidence ranges for countries with available data are shown in Figure 3-3. Incidence estimates tend to vary even more intra- and inter-regionally than prevalence estimates. In Europe, incidence ranges from 1.5 (Malta) to 7.4 (Crete) per 100,000 person-years and diverse estimates have also been reported in Asia (2.8 to 8.1 in Korea and Taiwan, respectively) and South America (4.2 to 6.3 in the Tucumán state and the city of Buenos Aires, Argentina); data from Africa are lacking.

In the 2023 comprehensive analysis of worldwide SLE epidemiology, the global incidence of SLE was estimated at 5.14 cases per 100,000 person-years, translating to ~400,000 new cases annually. The meta-analysis of epidemiologic studies from LMICs reported a pooled annual incidence rate of 5 cases per 100,000.

Enlarge 

Worldwide Mortality

The diversity of mortality measures used in epidemiologic studies globally makes comparisons difficult. In Europe, the standardized mortality ratio (SMR; the ratio of observed/expected number of deaths) among patients with SLE ranges from 1.6 in a Hungarian study to 3.1 in two UK cohorts, but is generally around 2. In Asia, the SMR ranges from 2.1 in a mainland Chinese cohort to as high as 11.1 in a Taiwanese population-based study. No SMR data are available from South America or Africa, but a study of Brazilian administrative data reported 4.76 SLE-related deaths per 100,000 inhabitants.

Epidemiologic Disparities in SLE

Lupus is associated with marked sex- and race/ethnicity-based disparities of disease burden. For patients with adult onset SLE, most studies report a female: male ratio of 9:1, although ratios as high as 15:1 have been reported. Data from the CDC registries in the US support this, showing much higher prevalence per 100,000 persons in women than in men (Georgia registry: 127.6 vs 14.7; Michigan registry: 128.7 vs 12.8; San Francisco registry: 155.6 vs 19.3; Manhattan registry: 107.4 vs 12.5; IHS registry: 215 vs 36). In the 2023 comprehensive analysis of global SLE epidemiology, the estimated prevalence of SLE was 78.73 and 9.26 per 100,000 persons in women and men, respectively. Incidence rates per 100,000 person-years are also higher for women than for men in the US (Georgia registry: 9.2 vs 1.8; Michigan registry: 9.3 vs 1.5; San Francisco registry: 9.6 vs 0.7; Manhattan registry: 7.9 vs 1.0; IHS registry [unadjusted rate]: 8.4 vs 1.7) and worldwide (8.82 vs 1.53). The higher incidence and prevalence of SLE in female patients make it one of the top 20 causes of death in women between 5 and 64 years of age in the US. Several explanations have been proposed to explain the sex-based differences in SLE epidemiology, including hormonal (estrogen-based immunomodulation), chromosomal (X-linked gene dosage effects) and epigenetic (sex-based genomic hypomethylation) factors, but the cause is not well understood.

Although less pronounced, racial/ethnic disparities are also present among patients with SLE. Within the US, Black, Hispanic and Asian populations typically show higher prevalence and incidence than White populations. For example, in the Georgia registry, the age-adjusted prevalence rate per 100,000 persons was 118.5 and 32.7 for Black and White patients, respectively and was especially high for Black women (196.2). Incidence per 100,000 person-years was also higher in Black patients (8.7) compared to White patients (2.7). Similar disparities have been observed in the other CDC registries with significant Black populations. Data from the Manhattan registry also show a higher prevalence per 100,000 persons in Asian (56.2) and Hispanic (82.8) patients compared to White patients (34.7). As per the IHS registry data presented above, SLE prevalence and incidence is highest in Native Americans and Alaskans. The cause of these racial disparities in prevalence and incidence is unknown. The putative existence of a “lupus prevalence gradient”, i.e., a gradient from low lupus prevalence in Africa to high prevalence in populations of African descent overseas, was suggested as a way to test the relative contributions of genetic and environmental factors to SLE development. However, more recent data suggest that the prevalence gradient may simply be a result of the limited resources for epidemiological study in Africa. In addition to the differences in prevalence and incidence, there are significant racial disparities in mortality as well. Data from the Georgia and California lupus registries show that cumulative mortality is significantly higher for Black patients than for White patients. In the Georgia registry, the mean age at death was ~52 years for Black patients and ~65 years for White patients. In the California registry, Black patients died an average of 6.8 years earlier than White patients while Hispanic patients died an average of 9.5 years earlier than non-Hispanic patients. Socioeconomic status is an important predictor of mortality outcomes in SLE; for example, the relative risk of mortality from LN-associated end-stage renal disease for Black patients compared to White patients is 1.2 is in bottom household income quintile but 1.0 in the top quintile.

Epidemiology of Other Types of Lupus

Most epidemiological studies to date have focused on the most common type of lupus, SLE. There are few studies of cutaneous lupus erythematosus (CLE) epidemiology. In one population-based study in Minnesota, the reported prevalence and incidence of CLE per 100,000 persons/person-years were 73.24 and 4.30, respectively. The prevalence of drug-induced lupus has been estimated at 10% of all SLE cases. Neonatal lupus (NL) is rare, with an estimated risk of 1:17,000 live births in the general population; even in women who are seropositive for anti-Ro/SSA antibodies, the risk in only 2% if they are first-time mothers or had previously had only children without NL. However, the risk is substantially higher in mothers who have previously had a child with cardiac NL (18%) or cutaneous NL (13%). Childhood-onset lupus is also rare, with a reported prevalence and incidence per 100,000 children of 3.3-24 and 0.3-0.9, respectively.

Remaining Challenges

Despite great progress in understanding the global epidemiology of SLE in recent decades, many important questions remain to be answered. Better data is needed from LMICs in order to answer questions about how the age structure, diagnostic delay and healthcare access in these countries impact intra-regional differences in SLE epidemiology. Even in developed countries, better data is needed on the epidemiology of SLE-associated conditions and adverse events and on how social determinants of health affect the epidemiology of SLE.