Systemic lupus erythematosus incidence and prevalence in a large population-based study in northeastern Italy

Abstract

Objectives

We aimed at estimating the incidence and prevalence of SLE in northeastern Italy over the period 2012–20.

Methods

A retrospective population-based study was conducted in Veneto Region (4.9 million people) using the population registry, an administrative health database where all residents are recorded. Between 2012 and 2020, SLE prevalence was defined by a healthcare co-payment exemption for SLE (national registry code 028) or any hospital diagnosis of SLE (International Classification of Disease , Ninth Revision, Clinical Modification 710.0), whichever came first. Incident SLE was defined from 2013 to 2020 to exclude prevalent cases. Standardized incidence and prevalence rates were reported by age and sex.

Results

During the study period, we identified 4283 SLE patients (85% female), with 1092 incident cases. Across the study period, SLE standardized point prevalence increased from 63.5 (95% CI 61.2, 65.8) to 70.6 (95% CI 68.3, 73.0) per 100 000 residents, corresponding to an annual increment of 1.14% (P < 0.0001). The highest prevalence was observed in females aged 60–69 years. SLE incidence corresponded to 2.8 per 100 000 person-years (95% CI 2.6, 2.9), with an annual decline of 7.3% (P < 0.0001). Incidence was 5-fold higher in females (female-to-male incidence rate ratio: 5.00, 95% CI 4.25, 5.87; P < 0.0001), with a peak among women aged 30–39 years. At diagnosis, women were significantly younger (45 years, IQR 33–58) than men (52 years, IQR 38–64).

Conclusions

Over the last decade, SLE prevalence has increased, while incidence has stably declined. In view of the introduction of new high-cost drugs, a clear definition of the epidemiology of SLE is crucial for all healthcare stakeholders.

Introduction

SLE is a complex autoimmune disease with a wide spectrum of clinical manifestations, from mild to life-threatening conditions. This disease leads to the inflammation of one or multiple organs including the skin, kidneys, joints and nervous system, contributing to the heterogeneity of clinical presentations. The reported incidence and prevalence of SLE vary widely across different countries and are strongly influenced by ethnicity, as well as environmental and genetic factors [1]. The diversity of reported epidemiological estimates of SLE is also partly related to population demographics as well as study design.

Four studies carried out in Italy estimated the incidence of SLE to be 2.0–2.6 per 100 000 person-years [2–4] and the prevalence 39.2–81 per 100 000 people [2–5]. According to these data, the incidence and prevalence of SLE in Italy are among the lowest identified in European countries. Nevertheless, the very small study population limits the generalizability of these results. Incidence rates in Europe vary from 1.2 to 31 per 100 000 person-years, and the prevalence ranges from 28 to 210 per 100 000 [1], with the lowest incidence in Central Europe (Germany and Estonia) [6, 7] and the highest prevalence in Southern Europe (Greece and Spain) [8, 9]. In North America, epidemiological estimates vary largely, with reported incidence rates ranging from 3.7 to 49 and prevalence varying from 72.8 to 366 per 100 000 [1, 10]. In a population-based study in Olmsted County, Minnesota, USA, based on more recent data, an overall annual SLE incidence of 4.77 cases (95% CI 4.09, 5.46) per 100 000 and a prevalence of 97.4 in the last decade were found [11]. Prevalence was shown to be 34–204 in South America and 18–100 per 100 000 in Asia and Australasia [1]. A different scenario has been depicted by studies carried out in Africa, where prevalence has been reported to be much higher, but the methodologies used were not reliable enough, with substantial heterogeneity between studies [12]. Interestingly, recent studies from UK and North America found that SLE prevalence has increased over time [11, 13–15]. Studies reporting variations in incidence and prevalence of SLE in Italy and other European countries over the past decade are currently lacking.

Our aim was to estimate variations in trends of SLE incidence and prevalence between 2012 and 2020 in a large population of 4.9 million inhabitants in Italy. Some new high-cost drugs have recently been approved by the Food and Drug Administration and the European Medicines Agency, and have become available for the treatment of the disease [16–19]. Thus, a clear definition of the epidemiology of SLE could represent a useful tool for allocating health resources.

Methods

This is a retrospective population-based study conducted in the Veneto Region, northeastern Italy, with a population of about 4.9 million residents. In Veneto Region, healthcare services are provided by nine local social and healthcare units, two university hospitals, two hospitals for scientific research and private accredited providers, based on a hub-and-spoke hospital model. Hospital care in Italy is free of charge for all residents and covered by general taxation. The Veneto Region is also one of the most populous regions of the country, accounting for 8% of the entire Italian population (it is the fourth most populous Italian region) and contributing to 9.2% of the Italian gross domestic product; the Veneto Region demographics is mainly composed of Caucasians (95%), but other ethnicities are also represented, including residents from Africa (2.1%), Asia (2.0%) and other minorities (data derived by the Veneto population registry, 2020).

All residents in Veneto Region between January 2012 and December 2020 were identified by means of the population registry, which has virtually full coverage of all residents of this region. Residents were then linked with the hospital discharge records, healthcare co-payments exemptions database and regional mortality records. The healthcare co-payment exemption database includes information on all individuals with a diagnosis performed by a medical specialist with specific conditions for which the national health service provides specific inpatient and outpatient free of charge services. Hospital discharge records contain data on all inpatient episodes, including (i) socio-demographic characteristics (sex, age, city of residence, educational level); (ii) clinical information (primary and secondary diagnoses, any surgical or medical procedure performed on the patient, and the mode of discharge); and (iii) data on the hospital (hospitalization ward, admission and discharge date).

All diagnoses are coded according to the International Classification of Diseases Revision, Ninth Revision, Clinical Modification (ICD-9-CM) coding system, currently used in Italy.

All residents in the Veneto Region between 1 January 2012 and 31 December 2020 were included in the analyses. This time window was chosen according to the availability of data required for SLE case identification in the registry, as the healthcare co-payments exemptions database was available since January 2012. Any resident with a healthcare co-payment exemption for SLE (code 028) or at least one hospitalization with a primary or secondary diagnosis of SLE (ICD-9-CM 710.0) was defined as affected by SLE. To estimate the point prevalence of SLE, only cases alive and residing in the region on 1 January of each year, between 2012 and 2020, were included. Details on the selection of the cohort are reported in Supplementary Fig. S1, available at Rheumatology online.

The incidence date was defined as the earliest date between healthcare co-payment exemptions and hospital discharge records. Since the electronic health record database for healthcare co-payment exemption was established in 2012, to minimize the risk of prevalent cases with SLE being identified as incident cases, analyses on incident SLE started after 1 January 2013.

The study was conducted using anonymized records of data routinely collected by healthcare services. All regional health records undergo a standardized anonymization process that assigns a unique anonymous code to each individual, allowing record linkage between electronic health records without any possibility of back retrieving the subject’s identity. All data in the Local Health Authority registries are recorded with the patient’s consent and can be used as aggregate data for scientific studies without further authorization (Garante per la protezione dei dati personali, Resolution n.85 of 1 March 2012). This study complies with the Declaration of Helsinki and the Italian Decree n.196/2003 on personal data protection.

Patient and public involvement

There were no funds or time allocated for patient and public involvement, so no patients were involved in the study.

Statistical analysis

Univariate and bivariate analyses were performed to summarize data with respect to the patient demographic characteristics. Continuous variables were reported with descriptive statistics [mean, s.d., median and interquartile range (IQR)]. For categorical variables, frequencies and percentages were calculated. The difference between groups was examined by Student’s t-test or Mann–Whitney test for continuous variables, Pearson’s χ² or Fisher’s exact test for categorical variables, as appropriate. A P-value <0.05 was considered statistically significant.

Point prevalence for each year was calculated as the ratio between the number of prevalent cases and the resident population in the Region on 1 January of each year.

The yearly incidence rates were estimated by dividing the number of new patients diagnosed with SLE each year by the number of residents in the Veneto Region in that year.

Incidence and prevalence per 100 000 people, along with the 95% CI, were presented for each year and stratified by age and sex. Direct standardization by age and sex was performed using the resident population of Veneto Region in 2017 as a reference. To examine the presence of a trend across the study period we fitted a Poisson regression model with year as explanatory variable, and point prevalence or incidence rates as outcome. The average annual percentage change was also estimated. All statistical analyses were conducted using Statistical Analysis Software (SAS) software version 9.4 (SAS Institute Inc., Cary, NC, USA).

Results

Prevalence

During the study period, 4283 prevalent cases with SLE were identified, of which 3636 were females (85%). The prevalence of SLE in the Veneto Region increased from 2012 to 2020 and was markedly higher among females (Table 1). The overall standardized point prevalence ranged from 63.5 (95% CI 61.2, 65.8) in 2012 to 70.6 SLE patients (95% CI 68.3, 73.0) per 100 000 persons in 2020, with an average annual increment of 1.1% (P < 0.0001).

SLE prevalence by age groups and sex for the year 2020 is reported in Table 2 and Fig. 1A, and shows a peak at age 60–69 years. Prevalence ranged from 5.9 (95% CI 3.6, 8.3) per 100 000 in the 0–19 years age group to 192.9 (95% CI 177.5, 208.4) per 100 000 in the 60–69 years age group among women, and from 2.0 (95% CI 0.7, 3.3) per 100 000 in the 0–19 years age group to 35.9 (95% CI 28.1, 43.6) per 100 000 in the 70–79 years age group among men. The prevalence was significantly higher among females in all age groups.

Incidence

During the study period, we identified 1092 incident SLE cases, corresponding to a standardized incidence of 2.8 per 100 000 person-years (95% CI 2.6, 2.9). Table 3 shows the yearly SLE incidence trend. Overall, there was an average annual decline of 7.3% (P < 0.0001), with the highest incidence rate observed in 2013 (3.4 per 100 000 person-years) and the lowest between 2018 and 2020 (2.2 per 100 000 person-years).

The median age at diagnosis was 45 years (IQR 34–59), lower among women than men (45, IQR 33–58 years vs 52, IQR 38–64 years; P = 0.002). Most incident cases were diagnosed at 30–39 years among women, and 50–59 years among men (Fig. 1B). Mean and median age at diagnosis were relatively stable across the study period (P = 0.3702). Incidence rates in females were 5-fold higher than in males (4.5 per 100 000 person-years and 0.9 per 100 000 person-years, respectively; P < 0.001). The female-to-male incidence ratio reached a peak in the 30–39 years age group (female-to-male ratio 10.5). The incidence was always higher in females compared with males regardless of age, as reported in Table 4.

Discussion

This is the largest study on the epidemiology of SLE carried out in Italy and covers the longest observation period among all region-wide population-based studies of SLE in our country [2–5].

In the present population-based study, we found that in 2020 SLE standardized point prevalence was 70.6 per 100 000 residents (95% CI 68.3, 73.0), whereas SLE standardized incidence rate between 2013 and 2020 was 2.8 per 100 000 person-years (95% CI 2.6, 2.9). We provide updated information regarding the epidemiology of SLE, which is crucial for clinicians, healthcare providers and companies, for planning interventions and allocating health resources, also in view of the availability of new targeted therapies and of upcoming high-cost medications for SLE.

Incidence and prevalence estimated in our population are comparable to those observed in Sweden, Denmark and France, but are lower than those observed in other countries such as the UK, Greece and Hungary, and higher than those reported in Germany, Estonia and Malta [6, 7, 20–24]. Environmental and genetic factors are known to play a relevant role in SLE onset [25], but discordant estimates across Europe may partly derive from different methodological approaches.

Different methodological strategies have been applied to describe the epidemiology of SLE, mainly depending on the population size and the availability of regional health registers. In small areas, community-based, individual-case validated studies based on medical charts and administrative data are more likely to be chosen [9, 11]. Conversely, larger study populations make this approach less feasible. In these cases, either representative population samples [8, 13], or, if present, healthcare registers with high coverage of the target population represent more appropriate options [6, 11, 20, 21]. Both approaches have some limitations but are generally considered reliable [1]. For example, two population-based studies using representative samples of 12 million and 2.3 million people, respectively, were carried out in the UK and Germany [6, 13], reporting an average incidence of 4.91 per 100 000 person-years (95% CI 4.73, 5.09) vs a maximum incidence of 3.6 (95% CI 2.9, 4.3) cases per 100 000 person-years, respectively. In 2010, using the French health insurance system (covering about 58.2 million people, ∼86% of the population) an annual crude incidence rate of 3.32 per 100 000 person-years was observed [21]. Differences in incidence estimates across studies with similar designs reinforce the relevance that specific genetic, environmental and socio-economic conditions play in determining SLE susceptibility. Nevertheless, these studies depict the epidemiology of SLE in their geographically and demographically determined settings only up to 2002, 2012 and 2010, respectively, so more recent trends and estimates of SLE in Europe are lacking.

The validity of studies carried out through population-based health registers has also been tested by selecting a representative population sample, and verifying the accuracy of the diagnosis in the register by reviewing individual medical charts [13]. A previous study in the UK performed this validation and found no statistically significant differences in prevalence estimates based solely on the population-based health register, compared with the prevalence obtained by verifying SLE cases by means of medical charts to confirm register-based diagnoses [13]. This finding supports the solidity of our results. In our study, chart review was not possible due to the high number of subjects included in the study, the anonymization process in the co-payment exemption database and hospital discharge records, and the lack of a linkage between the population registry and electronic medical charts.

According to our data, the incidence of SLE has been constantly decreasing in recent years, in contrast to the stable trend observed in the early 2000s in studies conducted in Europe and North America [9, 13–15, 22, 24, 26]. This finding might reflect a progressive increase in the recognition of milder SLE cases in the early 2010s, with a consequent progressive reduction of the pool of undiagnosed cases and a decline in incidence rates in the following years. Also, the role of preventive strategies largely implemented in the past two decades has yet to be established. These comprise strong recommendations for UV protection and smoking cessation, vitamin D supplementation and the use of HCQ to prevent disease onset in patients with UCTD. We did not find a decline in the age of incident cases through the study period, which would support a reduction in the delay of diagnosis, leading to a progressive reduction in the pool of undiagnosed cases.

A recent study carried out in a population of ∼144 248 individuals in the USA showed a higher incidence of SLE in the last decade (2009–18) compared with the period 1976–88 (6.44 vs 3.32 per 100 000 person-years) [11]. The authors attributed this rise to the change in the ethnic diversity of the study population. However, an additional explanation might be the improvement and the wider availability of diagnostic tools, which significantly changed over the past 20 years compared with the early 1980s. Furthermore, EULAR/ACR classification criteria for SLE were retrospectively applied, and this may represent a limitation of the study [11]. Despite being of extreme relevance, all these potential biases contribute in making long-term trends in SLE very difficult to assess.

We observed an increase in the prevalence of SLE: the same trend has been reported by other studies [11, 13–15, 22]. The analysis of the Clinical Practice Research Datalink, considered to be representative of the UK population, suggests that the age- and sex-adjusted prevalence of SLE in 1999 was 64.9 cases per 100 000 individuals and rose to 97.0 per 100 000 by 2012 [13]. Between 1993 and 2006, in Sweden, a study reported an increase from 55 to 65 SLE cases per 100 000 inhabitants [22]. According to administrative databases, SLE prevalence increased from 301.1 per 100 000 individuals in 2009 to 366.6 per 100 000 individuals in 2016 in the USA [14], while according to another recent population-based study it increased from 30.6 in 1985 to 97.4 per 100 000 persons in 2015. Similarly, SLE prevalence has increased from 48 to 90 per 100 000 individuals between 2000 and 2015 in Alberta, Canada [15].

The increase in prevalence observed worldwide may reflect the improved survival in patients affected with SLE, owing to earlier diagnosis and advances in management and therapy [27–29], with significant improvements in clinical outcomes [30–32].

As expected, SLE predominantly affected females, with an overall 5-fold higher incidence rate compared with males. Only among individuals aged 30–39 years SLE incidence was 10-fold higher in women than in men, which is the female-to-male ratio usually reported in SLE, while for all the other age groups the ratio was lower. However, other recent epidemiological studies found similar data: the female-to-male ratio was 2:1 in Germany [6], 3.9:1 in Estonia [7], 6:1 in France [21], 4–6.7:1 in Sweden [22] and 5.7 in Denmark [24]. Future studies are required to examine the reasons for these sex-based differences in incidence ratios.

Our study has many strengths: it examined the most recent epidemiological trends of SLE over almost a decade, and therefore our data are indicative of the current epidemiology of the disease, reflecting the substantial progress in early recognition of SLE which has been reported in the last decade, owing to a better understanding of disease heterogeneity and broader use of reliable diagnostic tools [26]. Our analyses span over almost a decade, allowing the evaluation of trends of both incident and prevalent cases. This is one of the largest studies conducted in Europe to date [13, 20–22]. In Europe, larger studies have been carried out only in the UK, Hungary, France and Sweden [13, 20–22]. Through the population registry and linkage with healthcare administrative databases, we were able to examine the health records of a population of 4.9 million inhabitants, corresponding to 8% of the overall Italian population. We used the Veneto population registry, which includes all the residents in the Region and linked data with the healthcare co-payment exemptions database, and with hospital discharge records to further increase the sensitivity of the incidence and prevalence algorithms.

Our study has some limitations: data on ethnicity were not available in the registry, thus we cannot assess the influence of this variable. Another limitation was the unfeasibility of verifying the diagnostic accuracy of those identified as SLE cases, but based on a previous validation study [13], the use of the healthcare co-payment database alone should have allowed us to provide very accurate epidemiological estimates; furthermore, the sensitivity of our incidence and prevalence algorithms was further improved by the use of hospital discharge records. Considering all sources used by our algorithm are based on medical diagnoses, we believe the risk of misclassification bias should be marginal [13]. Lastly, we cannot exclude that some SLE patients did not receive a co-payment exemption after their diagnosis, but these should represent a small minority, as patients are well informed of their right to be exempted from the cost of medical services that SLE patients require, so the incentive to claim for their exemption is high.

In conclusion, we observed that SLE prevalence has increased in both genders, probably due to improvements in diagnostic accuracy and disease management, with a possible improvement in survival. Data on mortality trends in SLE patients in Italy are warranted to support this hypothesis. In contrast, SLE incidence has stably declined in northeastern Italy over the past 8 years; overall, the incidence was 5-fold higher in females than in males and the highest incidence rate among females was found at age 30–39 years.

Supplementary data

Supplementary data are available at Rheumatology online.

Data availability

There are no additional unpublished data from this study to share.

Contribution statement

M.Z. contributed to the conception and design of the work, and interpretation of data, and she drafted the paper; L.S., C.B.A. and U.F. contributed to the acquisition and interpretation of data, and performed statistical analyses; S.B., A.G. and L.I. helped in revising the final work; A.D. revised the final work for important intellectual content; M.S. contributed to the conception and design of the work, and interpretation of data, and revised the manuscript for important intellectual content. All the authors approved the final version of the manuscript and gave their agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Funding

No specific funding was received from any bodies in the public, commercial or not-for-profit sectors to carry out the work described in this article.

Disclosure statement: The authors have declared no conflicts of interest.

References

Author notes