Article Text
PDF
Abstract
Objective To assess treatment patterns and the association between long-term glucocorticoid (GC) and hydroxychloroquine (HCQ) use and damage accrual in patients with systemic lupus erythematosus (SLE).
Methods A retrospective study including patients with SLE using the computerised database of a large health maintenance organisation. Patients were matched with subjects from the general population. Multivariable logistic regression models were used to assess the association between GC cumulative daily doses, HCQ and comorbidities: Osteoporosis, cardiovascular disease (CVD), hypertension and diabetes mellitus. Models were adjusted for age, sex, socioeconomic status, smoking, disease duration and HCQ use.
Results A total of 1073 patients with SLE were included, 87.79% were women. The age at first diagnosis was 37.23±14.36 and the SLE disease duration was 12.89±6.23 years. Initiation of HCQ within 12 months of SLE diagnosis increased from 51.02% in 2000 to 83.67% in 2010 and 93.02% in 2018. The annual usage of GC gradually decreased from 45.34% in 2000 to 30.76% in 2020. CVD and osteoporosis were more prevalent in SLE than in the general population. Multivariable logistic regression models revealed increased odds for comorbidities in patients receiving a mean daily dose of prednisone of more than 5 mg/day compared with those receiving 5 mg/day or less.
Conclusions CVD and osteoporosis were more prevalent in SLE than in the general population. The dose and frequency of GC treatment in patients with SLE have decreased over the years. Prednisone usage in doses exceeding 5 mg/day is associated with significantly increased odds of osteoporosis and CVD.
- Systemic Lupus Erythematosus
- Glucocorticoids
- Cardiovascular Disease
- Osteoporosis
- Epidemiology
Data availability statement
Data are available upon reasonable request. Data may be obtained from a third party and are not publicly available. The data sets generated during and/or analysed during the current study are not publicly available because the data that support the findings of this study originate from Maccabi Healthcare Services and restrictions apply to the availability of these data. Due to restrictions, these data can be accessed only by request to the authors and/or Maccabi Healthcare Services.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Prevalence and clinical characteristics of systemic lupus erythematosus (SLE) are affected by race, ethnicity and socioeconomic status which are not amendable to change. However, the treatment of SLE is a controllable variable. Treatment recommendations have recently been updated emphasising the need to reduce glucocorticoid exposure, although patient compliance remains a significant challenge.
WHAT THIS STUDY ADDS
This study evaluated the prevalence and characteristics of patients with SLE in Israel’s genetically unique population which has not been studied in depth. When analysing medication prescription over time, it is evident that treatment patterns have shifted in recent years along with the adoption of a more judicious use of steroids and an increased awareness of the importance of hydroxychloroquine (HCQ). However, the uptake of HCQ has not significantly improved since 2006. As previously reported, the use of prednisone doses over 5 mg/day was associated with osteoporosis and cardiovascular disease in this population as well.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
Our findings reinforce the need for better uptake of HCQ and better strategies to improve patient compliance emphasising the importance of a treat-to-target approach in a real-life setting.
Introduction
Systemic lupus erythematosus (SLE) is a systemic autoimmune disease characterised by autoantibody production and heterogeneous manifestations.1 Genetic predisposition, environmental triggers and hormonal factors interplay in disease development and activity.2 Consequently, the prevalence of SLE varies by geographical region, race and ethnicity.3–5 Prevalence estimates vary widely in the USA 48–366 per 100 000, in Europe it is 29–210 per 100 000, in Asia 26.5–103 per 100 000 and in Australia 13–89 per 100 000 individuals according to the population and region studied and methodology used.4
SLE is associated with increased morbidity and mortality; in Europe mortality is double while in female patients from the USA all-cause mortality was 2.6 higher for SLE than the general population.4 5 Cardiovascular diseases (CVDs) are a major cause of mortality.4
Glucocorticoid (GC) usage in SLE is associated with both disease control and morbidity or damage. In the current 2023 European League Against Rheumatism (EULAR) treatment recommendations for SLE,6 it is stated that GCs should be dosed to 5 mg/day or less compared with the 7.5 mg/day cut-off dose that was previously considered acceptable.7
Israel is a Middle Eastern country of ∼10 million people with a unique and diverse population consisting of a large Jewish population and also Israeli Muslim and Christian Arabs. However, the prevalence and clinical characteristics of SLE in Israel are unknown. The current study aimed to estimate SLE prevalence in Israel, assess past and current treatment patterns and the association between long-term GC use, hydroxychloroquine (HCQ) use and association with related comorbidities.
Methods
Study design and data sources
This observational retrospective study was conducted using the computerised database of Maccabi Healthcare Services (MHS), a health maintenance organisation (HMO) in Israel with 2.5 million members accounting for ∼25% of the Israeli population. According to the Israeli National Health Insurance Law of 1995, each citizen chooses one of four HMOs operating nationwide which are obligated to accept any citizen who wishes to join them irrespective of their health status. Therefore, all sectors of the Israeli population are represented in the MHS.
The study was conducted in accordance with the protocol, applicable regulations and the Declaration of Helsinki.
Study population
Prevalent patients with SLE on 31 December 2020 were identified according to the following inclusion criteria: A diagnosis of SLE according to International Classification of Diseases (ICD)-9 code 710.0 (≥2 by a rheumatologist/immunologist and/or ≥3 by a primary care physician (PCP)/paediatrician and/or ≥2 from a tertiary medical centre); first diagnosed by 31 December 2018; age≥10 years on date of first diagnosis; MHS member for ≥12 months before first diagnosis and ≥24 months after diagnosis; ≥1 purchase of HCQ since first diagnosis up to 31 December 2020; ≥1 positive antinuclear antibody (ANA) test by multiplex or immunofluorescence assay (IFA). Patients diagnosed with dermatomyositis and/or systemic sclerosis were excluded.
Patients with SLE were matched by age, sex and socioeconomic status (SES) at a ratio of 1:5 to individuals from the general MHS population.
Sociodemographic and clinical characterisation
Patients with SLE and the matched general MHS subjects were characterised according to age; sex; and SES on a 1–10 scale based on data from the Israeli Central Bureau of Statistics categorised into low (1–4), medium (5, 6) and high (7–10) SES; smoking status, categorised into ever smoked and never smoked; body mass index categorised using standard cutoff-points and Charlson’s comorbidity index (CCI).
Patients with SLE were also characterised according to age at disease onset; disease duration on 31 December 2020; SLE manifestations (online supplemental table S1). Serology tests included: ANA by multiplex and IFA; anti-double stranded DNA (dsDNA) antibodies by ELISA and if positive by the Crithidia luciliae IFA as well; anti-Smith, anti-Sjögren-Syndrome A (SSA) (Ro) and anti-Sjögren-Syndrome B (SSB) (La) antibodies were performed by ELISA; Complement 3 (C3) and complement 4 (C4); antiphospholipid antibodies (aPL) were recorded positive if ≥2 anti-cardiolipin (CL) immunoglobulin (Ig) G or IgM antibodies or anti beta2-glycoprotein I (b2GPI) IgG or IgM antibodies or lupus anticoagulant (LAC) were positive, 12 weeks apart. For anti-CL and anti-b2GPI, the cut-off for positivity was ≥40 units or the 99th percentile of the upper limit of normal. LAC positivity was defined by the MHS lab according to partial thromboplastin time and/or Russell viper venom time ratios above the upper normal limit. Additional recorded laboratory tests included urine protein-creatinine ratio (UPCR), 24-hour urine protein excretion and haematology laboratory tests. Detailed definitions of all laboratory tests are included in online supplemental table S2.
Supplemental material
Comorbidities and healthcare resource utilisation
All comorbid conditions were defined on 31 December 2020. The following conditions were defined according to corresponding MHS registries: Myocardial infarction (MI), ischaemic heart disease (IHD), congestive heart failure (CHF), atrial fibrillation (AF), peripheral vascular disease (PVD), any CVD (a composite of IHD, CHF, PVD, AF, cerebrovascular disease, cerebrovascular atherosclerosis, valve disease, cardiomyopathies or arrhythmias), osteoporosis, chronic kidney disease (CKD), hypertension (HTN) and diabetes mellitus (DM). History of a thromboembolic event was defined by any of the following conditions: Pulmonary embolism (PE), arterial embolism or upper or lower limb deep vein thrombosis (DVT) and/or by stroke (according to MHS cerebrovascular disease registries). An antiphospholipid syndrome (APS) event was any one of PE, arterial embolism, DVT, stroke, MI, abortions or miscarriages, in patients with positive aPL recorded twice, 12 weeks apart.
Healthcare resource utilisation was assessed during 2020 by visits to the PCP/paediatrician; visits to specialists (haematologists, dermatologists, cardiologists, neurologists, nephrologists); ≥1 hospitalisation (all-cause); days of hospitalisation stay.
Treatment patterns
Past treatment was defined as at least one purchase of any of the following drugs from the first SLE diagnosis until 31 December 2020: HCQ; oral GC (prednisone is the oral GC used in >90% of cases and hence this was the referent GC); synthetic immunosuppressants (azathioprine, methotrexate, mycophenolate mofetil, cyclosporine, cyclophosphamide, tacrolimus); biological therapies (rituximab and/or belimumab). Anifrolumab was unavailable during the study period and other biological therapies such as tumour necrosis factor inhibitors that are not usually indicated for SLE were not included.
Initiation of HCQ use in the first 12 months after the first SLE diagnosis was examined annually, during 2000–2018. Annual HCQ use during 2000–2020 was categorised into <300 tabs/year and ≥300 tabs/year.
Annual GC use among prevalent patients with SLE was examined per each calendar year during 2000–2020 and categorised into: ≤5 mg/day, >5–≤7.5 mg/day and >7.5 mg/day of prednisone. Annual GC use was also examined in the first 5 years after the diagnosis using the same categories. To estimate the actual daily GC exposure the cumulative daily dose of prednisone from the first SLE diagnosis was calculated according to the amount purchased divided by disease duration (in days) and categorised into: No GC use or ≤5 mg/day, >5–≤7.5 mg/day and >7.5 mg/day.
Annual treatment patterns among patients with SLE during 2000–2020 were defined by the following categories: HCQ only (≥1 purchase, with/without GC); synthetic immunosuppressants (≥1 purchase, with/without GC and/or HCQ); biological therapy (≥1 purchase, with/without GC and/or HCQ and/or immunosuppressants); GC only (without HCQ, immunosuppressants or biological therapy); none (no purchases of HCQ, immunosuppressants, biological therapy or GC).
Statistical analyses
Descriptive analyses were conducted to characterise patients with SLE, assess treatment patterns and matched subjects from the general population according to sociodemographic and clinical variables and evaluate the proportion of patients with each of the defined comorbid conditions. Descriptive statistics are presented as number (n), per cent (%), mean±SD or median, IQR, as appropriate. Multivariate logistic regression models were used to analyse the association between GC cumulative daily doses since the first diagnosis of SLE and comorbidities: Osteoporosis, CVD, HTN and DM. The models were adjusted for: Age (31 December 2020), sex, SES, smoking status, disease duration and HCQ cumulative annual doses since the first diagnosis of SLE (<300 tabs and ≥300 tabs per year). The CVD model was also adjusted for HTN and DM. Age, sex, smoking, HTN and DM were chosen to be included in the models as these are known cardiovascular risk factors. Dyslipidaemia, a complex variable to define that is based on cut-offs that rely on other risk factors and variables (cardiovascular risk estimates, age, sex, smoking and blood pressure) was not assessed in this study.8 SLE disease duration was included to assess the impact of the inflammatory disease and years of exposure to HCQ and prednisone. The fitness of the models was assessed by the area under the curve and Hosmer-Lemeshow goodness of fit tests.
Statistical significance was defined as a p value≤0.05 and standardised mean difference (SMD)>0.1 or <–0.1. Analyses were performed by IBM-SPSS V.25.0 (IBM, Amrock, New York, USA) and R Foundation for Statistical Computing (Vienna, Austria).
Patient and public involvement
Patients were not involved in the research design or process. However, the findings of this study are relevant to the Israeli SLE population as a whole and the authors plan to disseminate these findings in patient conferences and talks in order to achieve better outcomes for all patients with SLE.
Results
Study population and SLE prevalence
We identified 1073 eligible patients with SLE (online supplemental figure S1), SLE prevalence estimation in 2020 was 52.2 per 100 000 or 1:1916 among MHS members aged≥12 years. The proportion of women was 87.79% (n=942) and the mean age was 50.67±14.89 years. The age at disease onset was 37.23±14.36 years and the disease duration on 31 December 2020 was 12.89±6.23 years. The most prevalent SLE manifestations were cutaneous (85.18%), followed by: Haematological (74.28%), musculoskeletal (39.51%), renal (27.59%); neurological (11.09%) and serositis (10.81%). ANA by IFA was available for 773 patients (72.04%) of whom 87.97% were positive. Anti-dsDNA antibodies detected by an ELISA (available for 89.66%) were positive at least once for 53.01%; an IFA (available for 49.30%) was positive in 79.02% of cases (all with positive anti-dsDNA ELISA). C3 and C4 levels (available for 98.97%) were below the normal range of 50.75% and 32.11%, respectively. Positive aPL was identified in 40.37% of eligible patients. In patients who ever had a UPCR test done, 32.43% had a UPCR of over 0.5 mg/mg; when a 24-hour urine collection for protein was performed, 28.64% excreted more than 500 mg protein/day (table 1).
Characterisation of patients with systemic lupus erythematosus (N=1073)
Treatment patterns
Initiation HCQ in the 12 months after the first SLE diagnosis increased from 51.02% among those diagnosed in 2000 to 83.67% in 2010 and reached 93.02% in 2018 (figure 1a). Annual purchase of HCQ (figure 1b) remained stable since 2006 ranging between 66.42% and 69.49% while the proportion of those purchasing≥300 tabs (ie, ≥200 mg/day) increased annually.
(a) Initiation of hydroxychloroquine treatment in the first 12 months after the diagnosis of systemic lupus erythematosus by diagnosis year 2000–2018. (b) Annual hydroxychloroquine purchase among prevalent patients with systemic lupus erythematosus 2000–2020.
Analysis of the annual treatment patterns during 2000–2020 (online supplemental figure S2) showed that since 2000, 20.18–28.61% of all patients with SLE did not receive any medication. HCQ with no additional immunosuppressant/biologic (with/without GC) was used by 51.07% of all patients in 2020; immunosuppressant use decreased slightly during follow-up reaching 18.08% in 2020; use of biologics was <1% until 2012 when belimumab was first introduced to the Israeli market and increased since then reaching 5.31% in 2020. The proportion of patients treated only with GC (without HCQ or other medication) decreased from 10.59% in 2000 to 5.67% in 2010 and to 3.82% in 2020.
Annual GC use gradually decreased during follow-up (figure 2a) from 45.34% in 2000 to 36.40% in 2010 and 30.76% in 2020. The proportion of patients treated with prednisone at a mean daily dose of >7.5 mg/day also decreased from 11.47% in 2010 to 7.18% of all patients in 2020. GC use was also assessed in the first 5 years following the diagnosis (figure 2b). In the first year, 47.74% of the patients were treated with prednisone, half of them (23.14%) were treated with a mean dose of >7.5 mg/day. After the first year, the use of GC decreased and reached 29.40% in the fifth year with more than half of the treated patients (16.32%) receiving an annual dose of ≤5 mg/day. GC daily dose in the first 12 months after the initial diagnosis (online supplemental figure S3) showed a slight increase in the proportion of patients treated with >7.5 mg/day from 24.15% during 2000–2004 to 26.32% among those first diagnosed during 2015–2018.
(a) Annual glucocorticoids use among prevalent patients with systemic lupus erythematosus 2000–2020. (b) Glucocorticoids use in the first 5 years from diagnosis among patients diagnosed during 2000–2018.
Comorbidities in SLE and the general population
The 1073 patients with SLE were matched with 5365 subjects from the general MHS population (table 2). Compared with the general population, patients with SLE had higher CCI scores (2.00 IQR 1.00–3.00, 0.00 IQR 0.00–1.00, respectively, p<0.001, SMD=−0.9). CVD was present in 17.99% of the patients with SLE versus 7.59% in the general population (p<0.001, SMD=−0.32). A history of MI, IHD, CHF, PVD and a history of CVA and/or TIA was more prevalent in patients with SLE compared with the general population (2.61% vs 1.27%, SMD=−0.1; 6.34% vs 2.89%, SMD=−0.16; 1.58% vs 0.47%, SMD=−0.11; 1.58% vs 0.73%, SMD=−0.08; 2.98% vs 1.23%, SMD=−0.12; respectively, p<0.01 for all comparisons). A diagnosis of any thromboembolic event was recorded for 21.81% of the patients with SLE versus 5.72% of the matched general population (p<0.001, SMD=−0.48). The prevalence of PE, DVT and an arterial embolic event was also higher in patients with SLE than in the matched general population (5.87% vs 0.75%, SMD=−0.29; 16.22% vs 4.16%, SMD=−0.41; 1.03% vs 0.09%, SMD=−0.13; respectively, p<0.001 for all comparisons). Compared with patients with SLE without aPL, patients with SLE with positive aPL had higher rates of all thromboembolic events and of venous thromboembolism while the rates of arterial thromboembolism were similar in patients with SLE with and without aPL (online supplemental table S3). Additional comorbidities were more prevalent in patients with SLE compared with the general population including osteoporosis (29.82% vs 11.16%, p<0.001, SMD=−0.48), CKD (36.72% vs 11.76%, p<0.001, SMD=−0.61) and HTN (29.64% vs 21.90%, p<0.001, SMD=−0.18), while diabetes rates were similar in both groups (7.64% vs 8.72%, p=0.2, SMD=0.04). During 2020, visits to the PCP and to specialists were more frequent in patients with SLE (15.65±9.80 vs 8.95±7.68, p<0.001, SMD=−0.76 and 2.94±9.80 vs 8.95±7.68, p<0.001, SMD=−0.18, respectively). Patients with SLE were more likely to be hospitalised at least once (14.82% vs 7.88%, p<0.001, SMD=0.22).
Characterisation of patients with systemic lupus erythematosus and matched subjects from the general population
Comorbidities and their association with GC use
The multivariable logistic regression models (table 3) showed that compared with patients with SLE who were not treated with GC or were treated with a cumulative daily dose of prednisone of ≤5 mg, those treated with a cumulative daily dose of >5–≤7.5 mg/day and those treated with >7.5 mg/day had increased odds for osteoporosis (OR 3.13, 95% CI 1.92 to 5.13 and OR 5.37, 95% CI 3.05 to 9.52, respectively). For CVD, the OR for patients treated with a prednisone dose of >5–7.5 mg/day was 1.74, 95% CI 1.03 to 2.88, p=0.035 and OR 1.68, 95% CI 0.91 to 3.01, p=0.088 for doses >7.5 mg/day. As expected, all known cardiovascular risk factors (age, sex, smoking, HTN and DM had a significant effect on CVD (table 3). The odds for HTN were increased for those treated with >5–≤7.5 mg/day (OR 2.39, 95% CI 1.47 to 3.92) and the odds for DM were increased for those treated with >7.5 mg/day (OR 3.09, 95% CI 1.41 to 6.33). A cumulative annual HCQ dose of 300 tablets or more was associated with an increase in the odds for osteoporosis (OR 1.63, 95% CI 1.18 to 2.27, p=0.003), however, there was no significant association with CVD, HTN or DM. An additional analysis (online supplemental table S4) found a correlation between purchasing ≥300 HCQ tablets per year and higher daily doses of prednisone (p<0.001).
Multivariable logistic regression models for the association between cumulative daily dose of GC and various comorbidities (n=1073)
Discussion
Race, ethnicity and SES are well-known factors that impact disease severity and prognosis in SLE, however, are not amendable to change while controllable treatment patterns may have a significant effect on disease outcomes.1 5
This is the first study to assess the prevalence of SLE in Israel using a stringent definition of SLE cases. We report an estimated prevalence of 52.2 per 100 000 (1:1916) in Israel similar to the most conservative estimates4 and less than the more recent estimation of SLE prevalence in the USA of 72.8 per 100 000.5
The rate of renal involvement was relatively low compared with rates commonly reported in the literature.9 Several possible explanations may include the retrospective design. Additionally, the relatively high SES of patients in MHS may explain the low rate. In MHS, 16.3% are of low SES as opposed to 34% of low SES in the Israeli population.10 Accordingly, Sagy et al reported that low compared with medium and high SES, regardless of ethnicity was associated with higher rates of lupus nephritis.11 Another possible explanation—may be due to ethnic and racial determinants. Jewish people of North African descent show more genetic similarity to other Jewish communities than to other non-Jewish geographically near populations.12A retrospective study of two medical centres in Israel did find a difference in the rate of lupus nephritis in Ashkenazi (29.9%), non-Ashkenazi (32.4%) Jews, and Arabs (42.9%), (p=0.11).13
Treatment patterns relating to HCQ usage mirror other global SLE patient populations and recommendations over time. Initiation of HCQ treatment in the first year after diagnosis increased significantly in 2009 reaching approximately 90%, in line with the publications showing its efficacy in controlling disease, preventing flares and damage, metabolic advantages and safety.14–17 However, the annual proportion of patients purchasing HCQ remained stable (nearly 70%) since 2007 similar to rates of HCQ usage in other studies.18 19
In general, treatment patterns of GC usage were encouraging. In 2020, the rate of GC use decreased to under a third (30.8%) of all patients and the mean daily dose also decreased with most receiving less than 5 mg/day of prednisone following the 2023 EULAR recommendations.6 The rate of GC use in this Israeli SLE patient population was lower than that of other real-world studies. A study describing a contemporary USA claims database between 2010–2019 found that 49.7% of new and 51.3% of existing patients with SLE were treated with prednisone at a median daily dose of 14 mg/day and 12.3 mg/day, respectively.20 Another recent Swedish national cohort study found that at 1 year and 15 years after diagnosis, 64.5% and 48% of patients with SLE respectively were treated with GC.21 In an older report in 2015 The Systemic Lupus International Collaborating Clinics cohort reported a higher use of GC, 69.98% of new patients with SLE and 56.9% of patients 5 years from diagnosis.22 In a Spanish multicentric study of new patients with SLE, 65% were prescribed prednisone at a mean dose of 11 mg/day.23 A French claims database study found that 48.2% of patients with SLE were treated with GC and 11.7% of patients were treated only with GC.24 Our findings showing lower use of GC may be due to the good accessibility of the healthcare system in Israel and the unique ethnic and genetic characteristics of the Israeli population with possibly lower prevalence and severity of lupus nephritis.
Patients with SLE had a three times higher prevalence of osteoporosis and CKD compared with the general population. IHD was twice as prevalent and thromboembolism was four times more prevalent in SLE compared with the general population. Our results are in line with published literature from other countries.25–27 Accordingly, patients with SLE used more healthcare resources, had more hospitalisations and ambulatory clinic visits similar to studies from Europe, the USA and Asia.28–30
In our study, GC use was associated with higher odds of osteoporosis in a dose-dependent manner with HTN and CVD in doses above 5 mg/day and with DM in doses over 7.5 mg/day. Our findings are in line with other large studies assessing the dose-dependent harm of GC. The Hopkins lupus cohort found an increased risk for organ damage with increasing cumulative monthly doses of prednisone with very low risk for doses lower than 180 mg/month.17 The risks for osteoporotic fractures and cardiovascular damage were increased 2.16 and 1.54 respectively with prednisone doses over 7.5 mg/day.31 Similarly, the Swedish national SLE cohort showed higher risks than the general population for osteoporosis and cardiovascular risk factors. The increased osteoporosis risk was strongly dose-dependent with an HR of 3.08 and 5.11 for prednisone doses of 5–7.5 mg/day and >7.5 mg/day respectively, p<0.001. The HR for HTN was also dose-dependent but smaller risks were reported (1.23, 1.6, respectively). Regarding DM, similar to our results, risks were not significant for the lowest GC dosage (over 0 and up to 5 mg/day), but were high for doses above 5 mg/day.21 Lastly, a French claims database also found that prednisone doses over 5 mg/day were associated with an increased risk of osteoporosis and CVD.24
We found increased odds for osteoporosis in patients adherent to HCQ therapy. The effect of HCQ on osteoporosis risk, bone mineral density and fractures is mostly described as positive, although the strength of evidence is low and some studies found an inverse association similar to our study.14 32–34 The use of HCQ may be confounded by higher disease activity and hence higher GC doses used to suppress it (as shown by our additional analysis), a better adherence to bone mineral density testing and hence diagnosis, less sun exposure and lower vitamin D levels.
Our study has several limitations. The MHS represents about a quarter of the Israeli population and SES levels may be higher than national levels. The database is based on ICD-9 codes, so some clinical data may be missing or miscoded, and SLE cases were not included based on formal classification criteria. The cumulative HCQ and prednisone doses are based on MHS pharmacy purchase records but few purchases may have occurred in other pharmacies leading to possible underestimation of doses; while purchase of medications may not always reflect medication intake or adherence leading to overestimation of doses. Deceased patients with SLE were excluded and we could not assess the association of treatment patterns with mortality. In addition, for the multivariable analysis we cannot rule out the possibility of residual confounding. Previous studies have found various factors associated with the development of each of the outcomes tested. In the setting of our study, we were able to adjust for major confounders (age, sex, SES, smoking, SLE disease duration in addition to HCQ cumulative annual dose and prednisone cumulative daily dose; the model for CVD was also adjusted for HTN and diabetes) which may help reduce the risk of residual confounding. Our study also has several important strengths. The definition of an SLE case was strict but inclusive and probably representative of most SLE cases in MHS. We included a wide range of clinical and laboratory information that may serve to reinforce the specificity of our patient population and the accurate characterisation of patients with SLE in Israel. The study spanned almost two decades and captures the dynamics in treatment patterns.
Conclusions
This is the first study to characterise patients with SLE and treatment patterns in Israel. GC use in our population of patients with SLE is lower than that reported in other real-world studies and may reflect different ethnic and genetic populations and/or different management strategies. We, like others, show that the use of prednisone in doses over 5 mg/day is associated with an increase in the risk of osteoporosis and CVD. Our findings reinforce the need for better uptake of HCQ, judicious use of GC, better strategies to mitigate comorbidities and better steroid-sparing agents.
Data availability statement
Data are available upon reasonable request. Data may be obtained from a third party and are not publicly available. The data sets generated during and/or analysed during the current study are not publicly available because the data that support the findings of this study originate from Maccabi Healthcare Services and restrictions apply to the availability of these data. Due to restrictions, these data can be accessed only by request to the authors and/or Maccabi Healthcare Services.
Ethics statements
Patient consent for publication
Ethics approval
This study involves human participants and was approved by Maccabi Healthcare Services (MHS) Institutional Review Board (IRB) 0047-21-MHS. The MHS IRB approved the study protocol (0047-21-MHS) and waived the requirement to obtain informed consent for secondary analysis of existing data.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
X @talieshua
VR and DP contributed equally.
Contributors TE took part in the conception and design of the study, data analysis and interpretation, drafted the first and later drafts and revised the work. RY was part of the conception and design of the study, data interpretation and revised the manuscript. IL was part of the conception and revised the manuscript. ME was part of the conception and design, data interpretation and revised the manuscript. OE assisted in data interpretation and manuscript revisions. GC assisted in acquisition of data, data interpretation and manuscript revisions. VR took part in the conception and design of the study, acquisition of data, data analysis and interpretation, first draft and revisions of the manuscript. DP took part in conception and design of the study, data analysis and interpretation and manuscript revisions and supervised the study. All authors discussed the results and interpretation of them, contributed to the final manuscript and approved its final version. TE and VR are guarantors.
Funding The study was sponsored by AstraZeneca. The company was not involved in the gathering or analysis of data.
Competing interests RY, IL and ME are employees of AstraZeneca. TE and DP received consultant fees from AstraZeneca.
Patient and public involvement Patients and/or the public were involved in the design, or conduct, or reporting, or dissemination plans of this research. Refer to the Methods section for further details.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.