Article Text
PDF
Abstract
Objective Few epidemiological studies are reported in the published literature on the incidence or prevalence of antiphospholipid syndrome (APS), and available results are heterogeneous. This study aimed to estimate the incidence and prevalence of APS in the USA, overall and by APS subtype.
Design A retrospective analysis of APS disease incidence and a cross-sectional analysis of disease prevalence.
Setting Merative MarketScan Commercial Claims and Encounters Database, and the Medicare Supplemental and Coordination of Benefits Database.
Participants All individuals with claims for at least two antiphospholipid antibody tests undertaken at least 12 weeks apart and a diagnosis claim for APS as a primary or secondary diagnosis on or after the second antibody test, during the period 1 January 2016 to 31 December 2019.
Main outcome measures Annual incidence and prevalence of APS and APS subtypes.
Results In total, 1708 cases of APS were identified during the study period (2016–2019), of which 83% were women. The overall annual standardised incidence rate of APS per 100 000 person-years increased slightly over the study period, from 2.31 in 2016 to 2.71 in 2019. In 2019, the estimated annual prevalence of APS per 100 000 persons was 10.42 per 100 000 persons (95% CI 9.96–10.90). Based on this and US census data, we have estimated that 34 000 persons in the USA were affected by APS in 2019.
Conclusions These data add to the estimates of prevalence and incidence of APS in the literature, all of which have different strengths and limitations of the different data sources and case ascertainment methods.
- EPIDEMIOLOGY
- Chronic Disease
- Retrospective Studies
- Cross-Sectional Studies
Data availability statement
No data are available. Data from non-interventional studies are outside of UCB's data sharing policy and are unavailable for sharing.
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
STRENGTHS AND LIMITATIONS OF THIS STUDY
The Merative MarketScan Commercial Claims and Encounters and Medicare Supplemental and Coordination of Benefits used in this study are large databases.
The identification of the APS claims was realised by using a unique APS-specific International Classification of Disease code (D68.61), which was made available during the study period, whereas previous studies used a non-specific code.
Instead of the full Sydney criteria, a case definition proxy was applied to identify APS cases.
Given our estimate of incidence is calculated using claims from only a subset of the total US population, which is not a representative sample, it cannot be extrapolated to the full US population or populations in other countries.
Introduction
Antiphospholipid syndrome (APS) is an autoimmune disease caused by antiphospholipid antibodies (aPLs) that results in a wide spectrum of clinical manifestations, involving the haematological, obstetrical, neurological, cardiovascular, renal and orthopaedic systems, among others.1 Criteria aPLs include lupus anticoagulant (LAC), anticardiolipin (aCL) and anti-β2-glycoprotein I (anti-β2GPI) immunoglobulin (Ig) G and IgM antibodies.
According to the Sydney 2006 International Classification Criteria Consensus, clinical diagnosis of APS requires the presence of clinical symptoms associated with APS (vascular thrombosis or pregnancy morbidity) and two positive test results for aPL, 12 weeks apart.2 Furthermore, older epidemiology studies were undertaken before LAC, aCL and anti-β2GPI antibodies were recognised as aPLs in the 2006 update of the classification criteria.2 As a result, few studies have been carried out to assess the epidemiology of APS, while the few data that have been published are inconsistent.3
Since completion of this study, a manuscript supported by the American College of Rheumatology Board of Directors and the European Alliance of Associations for Rheumatology has been published, in which a new classification criterion of APS has been proposed.4
In the USA, Duarte-Garcia et al (2019) estimated the overall annual incidence and prevalence rates of APS for patients aged ≥18 years at 2.1 per 100 000 person-years (PY) and 50 per 100 000 persons, respectively.5 Studies worldwide have estimated incidence and prevalence ranging from 7.5 per 100 000 PY in Korea6 to 40 per 100 000 PY in Spain,7 and from 16.8 per 100 000 people in the Piedmont and Aosta Valley regions8 to 61.9 per 100 000 people in Korea,6 respectively. Previous studies have also shown that the peak incidence of APS differs for men and women. In the UK, Rodziewicz et al (2019) used data from the UK Clinical Practice Research Datalink to estimate the peak APS incidence of 7.5 per 100 000 PY for women occurring between the ages of 35 years and 39 years, while the peak APS incidence for men was 2.2 per 100 000 PY and occurred later in life, between the ages of 55 years and 59 years.9 Prevalence of APS in the UK was also higher in women compared with men: 50 and 9.8 per 100 000 persons, respectively.9
The Merative MarketScan Commercial Claims and Encounters (CCAE) and the Medicare Supplemental and Coordination of Benefits (MDCR) databases represent insurance claims from US employees and their dependents, from all US census regions, covering approximately 40 million persons in the USA annually (figure 1).10 Using this large, real-world database, we aimed to retrospectively and cross-sectionally estimate the incidence and prevalence of APS in the US population during 2016 and 2019. A proxy definition for APS cases was developed using clinical variables available in the Marketscan database, with the aim to be as close as possible to the Sydney 2006 classification criteria. This study was already complete prior to the 2023 publication of the new APS classification by Barbhaiya et al.4
MarketScan research databases.10 CCAE, Commercial Claims and Encounters; MDCR, Medicare Supplemental and Coordination of Benefits.
Methods
Study design and data source
This was a retrospective, cross-sectional analysis of APS incidence and prevalence, using the CCAE and MDCR databases during the study period of 1 January 2015 to 31 December 2019 (figure 2). The database is Health Insurance Portability and Accountability Act compliant, and all patient data were deidentified before delivery to the study team. No ethics committee approval was required.
Study design. aPL, antiphospholipid antibody; APS, antiphospholipid syndrome; ICD, International Classification of Diseases.
Patient and public involvement
Patients and their families were not involved in the design, implementation or in setting the research question or the outcome measures.
Case definition and identification
APS cases were defined as claims for at least two aPL tests undertaken at least 12 weeks apart, and a diagnosis claim for APS as a primary or secondary diagnosis on or after the second antibody test, identified from International Classification of Diseases, 10th revision (ICD-10) codes. In October 2015, APS was given a unique ICD code (D68.61) with the introduction of the Clinical Modification (CM). Prior to ICD-10-CM, APS was coded together with multiple other hypercoagulation defects into 289.81 (primary hypercoagulable state). Thus, cases of APS can only be ascertained from October 2015 onwards.
The Sydney International Classification Criteria Consensus case definition requires laboratory criteria (at least 12 weeks between the initial and repeated positive aPL test) and at least one of the clinical criteria (vascular thrombosis or pregnancy morbidity). An aPL test is confirmed positive if one or more of the following aPLs are detected on two or more occasions at least 12 weeks apart: (i) LAC present in plasma detected according to the International Society on Thrombosis and Haemostasis guidelines;2 11 12 (ii) aCL of IgG and/or IgM isotype in serum or plasma at medium or high titre (>40 IgG or IgM phospholipid units, or >99th percentile) measured by a standardised ELISA or (iii) anti-β2GPI of IgG and/or IgM isotype in serum or plasma (>99th percentile) measured by a standardised ELISA.2
As aPL test results are not available from the CCAE and MDCR databases, a proxy was used for a positive result. If a second, repeated aPL test was carried out 12 weeks later than the first test, it was assumed at least one of the tests was positive. aPL test codes were 86148, 86146 and 86147, corresponding to aPL, anti-β2GPI, and aCL, respectively. At least one clinical criterion (vascular thrombosis or pregnancy morbidity) was added to the case definition as a sensitivity analysis. The index date was defined as the date of the first APS diagnosis which could be before, on or after the second aPL test.
Study populations and APS subtypes
The study population comprised all patients identified as having APS according to the criteria described above. Patients were also required to have at least 12 months of continued medical and pharmacy benefits memberships prior to the index date (not required for children aged <1 year) from 1 January 2016 to 31 December 2019 and, for incident patients, no diagnosis record of APS (ICD-9 or ICD-10) at any time prior to the index date. APS was classified as primary APS in the absence of systemic autoimmune disease (per the Sydney 2006 International Classification Criteria Consensus2); however, primary and secondary APS were not analysed separately. APS subtypes were classified into five mutually exclusive categories as thrombotic APS only (with ≥1 clinical criteria as described in online supplemental table 1), obstetric APS only (women aged ≥14 years only with ≥1 clinical criteria as described in online supplemental table 1), mixed APS (both thrombotic APS and obstetric APS events; women aged ≥14 years only), other APS manifestations not included in the Sydney criteria (online supplemental tables 1 and 2) and APS type unknown (APS cases that were not linked to any clinical criteria relating to the listed subtypes). In addition, catastrophic APS (CAPS) was analysed as a subtype of thrombotic and mixed APS, defined as thrombosis in three or more organs developing in less than a week. This definition was as close as possible to the 10th and 14th International Congress on Antiphospholipid Antibodies definitions,13–15 without the required information on aPL status and biopsy, which was not available in this database.
Supplemental material
Study outcomes and statistical analyses
Distribution of APS subtypes by age and sex during 2016–2019, and by specialty involved in the first APS diagnosis for the period 2018–2019, was assessed. Overall distribution of CAPS was calculated and presented as one result for the aggregated incident cohort. The annual incidence and prevalence of APS and APS subtypes were calculated by age and sex in each calendar year from 2016 to 2019. Incidence of APS was age- and sex-standardised indirectly to the US 2020 population by using age- and sex-specific population census data as weights and applying them to the age- and sex-specific incidence rates.16
Incidence, the number of new APS cases during a specific time period,17 was reported as a rate per 100 000 PY. Person-time was defined as the sum of each patient’s duration of follow-up from 1 January to 31 December of each year, to the end of the patient record, to the end of the first insurance claim if there was a gap >60 days or to the APS diagnosis—whichever came first. Incidence was calculated as the number of incident cases divided by the person-time of observation per year. The annual incidence rate was calculated as the number of incident cases in a year divided by the person-time corresponding to that year. Prevalence, the proportion of the population with APS in a given time period,17 was expressed as prevalence per 100 000 persons. The denominator was all patients in the CCAE and MDCR databases on 1 July of each year. The annual prevalence rate was calculated as prevalent cases in a year divided by the overall population corresponding to that year. The 95% CI of the prevalence and incidence rates was estimated by the Poisson distribution.18
Sensitivity analysis
To assess the robustness of the case definition in this study, a number of alternative case definitions were considered for the incidence and prevalence calculation. Patients could have one of the following: (a) at least one APS diagnosis (primary or secondary) with no requirement of an aPL test; (b) at least one APS diagnosis (primary or secondary) with at least one aPL test; (c) at least one APS diagnosis with two aPL tests undertaken 12–26 weeks apart and 2–13 weeks before an APS diagnosis code; (d) for a patient with a membership gap during follow-up, the follow-up included all time at risk including during and after the gap (ie, they were analysed as if there was no gap); (e) at least one diagnosis record of LAC syndrome (ICD-10 code 68.62) with two aPL tests at least 12 weeks apart and ≥1 diagnostic record of thrombotic or obstetric events according to the Sydney criteria or (f) at least one of the clinical criteria (vascular thrombosis or pregnancy morbidity per the definitions in online supplemental table 1) with at least two aPL tests undertaken at least 12 weeks apart, and a diagnosis claim for APS as a primary or secondary diagnosis on or after the second antibody test (online supplemental table 3). All statistical analyses were performed using SAS V.9.4.
Results
Baseline characteristics
A total of 1708 cases of APS that met our case definition during 2016–2019 were identified, with the majority being women (n=1411, 83%; table 1, online supplemental table 4). For women, the number of incident APS cases was highest among the 25–44 years age category (n=706, 50.0%), and 376 (26.6%) patients had obstetric APS. For men, the number of incident APS cases was highest among patients aged 45–64 years (n=201, 67.7%). The overall number of paediatric APS cases was small (n=23), with most cases identified within the 15–17 years age category (n=16) for both female and male patients. Thrombotic APS was the most frequent subtype identified (n=728, 42.6%), followed by APS of unknown subtype (n=404, 23.7%), obstetric APS (n=376, 22.0%), other APS (n=138, 8.1%) and mixed APS (n=62, 3.6%). 75 cases of CAPS were identified, representing 4.4% of all APS cases. The acute care hospital was the most frequently visited healthcare specialty site for diagnosis (n=227, 26.9%), followed by laboratory (n=161, 19.1%), rheumatology (n=89, 10.5%), internal medicine (n=87, 10.3%), obstetrics and gynaecology (n=72, 8.5%) and haematology (n=70, 8.3%). Other categories for APS diagnoses included oncology, family practice, pathology, neurology, multispecialty and medical doctors.
Overall incident APS cases by baseline gender and age distribution and APS subtype
Incidence of APS
The overall annual incidence rate of APS per 100 000 PY (95% CI) standardised to the US 2020 population census increased slightly over the study period, from 2.31 (2.11–2.53) in 2016 to 2.84 (2.59–3.11) and 2.71 (2.45–2.99) in 2018 and 2019, respectively (figure 3). Over the whole study period, the APS incidence rates were higher in women, potentially due to certain subtypes being applicable to women only. In 2019, the annual incidence rate (standardised to the US 2020 population census) was 4.45 (3.99–4.96) per 100 000 PY for women and 0.91 (0.71–1.17) per 100 000 PY for men. The incidence rate of APS in women was always highest in patients aged 30–39 years old, ranging from 10.09 per 100 000 PY in 2016 to 11.80 and 11.22 per 100 000 PY in 2018 and 2019 (online supplemental table 5). However, in men, the age categories with the highest incidence rates varied each year (online supplemental table 5).
Yearly incidence and prevalence rates of APS cases. Lines represent incidence rates per 100 000 PY and bars represent prevalence rates per 100 000 persons. APS, antiphospholipid syndrome; PY, person-years.
Prevalence of APS
The estimated annual prevalence of APS per 100 000 persons (95% CI) was 10.42 per 100 000 persons (9.96–10.90) in 2019 (figure 3). In 2019, APS prevalence among women was four times higher than in men: 16.59 (15.78–17.45) and 4.00 (3.60–4.44) per 100 000 persons, respectively. Based on this and US census data, we have estimated that slightly more than 34 000 persons in the USA were affected by APS in 2019. Among women, the prevalence rate of APS was highest in the 40–49 years age group, ranging from 14.71 in 2016 to 30.61 per 100 000 persons in 2019. Among men, APS prevalence was highest in patients aged 60–69 years old between 2016 (6.27 per 100 000 persons) and 2018 (10.85 per 100 000 persons) and in patients aged 80+ years in 2019 (13.57 per 100 000 persons) (online supplemental table 6).
Incidence rates of APS subtypes
Incidence rates of APS subtypes were standardised to the US 2020 population census. In 2019, the overall annual incidence (95% CI) of thrombotic APS was 1.28 (1.10–1.49) per 100 000 PY, which was the highest incidence over the study period. The 2019 annual incidence rate of thrombotic APS patients was higher in women (1.76 per 100 000 PY) than in men (0.79 per 100 000 PY) (online supplemental figure 1A). In 2019, among women-only, the incidence rate of obstetric APS was 1.24 (1.01–1.53) per 100 000 PY (online supplemental figure 1B), and the incidence rate of mixed APS was 0.17 (0.10–0.30) per 100 000 PY. The other APS cases represent 8.1% of the total APS population, with an overall annual incidence rate in 2019 of 0.23 (0.16–0.32) per 100 000 PY. The unknown APS cases represent almost a quarter (23.65%) of the total APS population, with an overall incidence rate in 2019 of 0.48 (0.38–0.60) per 100 000 PY (online supplemental table 7).
Prevalence rates of APS subtypes
The overall annual prevalence (95% CI) of thrombotic APS was 5.38 (5.05–5.73) per 100 000 persons in 2019, and the annual prevalence rate of thrombotic APS was higher in women compared with men over the whole study period. The annual prevalence of obstetric APS subtype among women-only was 3.70 (3.33–4.12) per 100 000 persons in 2019, and the highest prevalence rate of obstetric APS was in those aged 30–39 years old (12.98 (11.24–15.00) per 100 000 persons). The annual prevalence of mixed APS was 1.19 (0.99–1.44) per 100 000 persons in 2019, among women-only. The overall annual prevalence rates of other APS and unknown APS in 2019 were 0.90 (0.77–1.05) per 100 000 persons and 2.19 (1.99–2.42) per 100 000 persons, respectively (online supplemental table 7).
Sensitivity analysis
Outcomes from a sensitivity analysis for different predefined scenarios for the last year of the study period, 2019, are shown in online supplemental figure 2. The outcomes obtained from the rest of the study years were similar to those for 2019, and the subtypes of APS showed the same behaviours for the sensitivity analysis as for the overall APS cases.
Discussion
The incidence and prevalence of APS in the USA have been assessed using the large CCAE and MDCR databases, which include claims data on more than 60 million insured US employees and their dependents. Our data suggest there is a trend towards a yearly increase in the incidence and prevalence of APS over time, which may be a product of generally increased awareness of the disease.3 Additionally, adoption of the ICD-10 code in 2015 may have also contributed to the observed increase in prevalence.
There are a number of published studies that aimed to assess the incidence and prevalence of APS in different countries around the world,5–9 19 which revealed important clinical and epidemiological information on APS in their respective regions (online supplemental table 8). However, the incidence and prevalence estimates reported vary, likely due to the variation in APS case definitions, design across studies and the type of database analysed, which may result in variation in socioeconomic demographics among the populations included.3
In particular, our sensitivity analysis shows that differences in incidence rates are apparent when using different diagnostic scenarios, thus results may be sensitive to the case definition used. This is evident when evaluating studies in the literature, as one pertinent difference between them was the method used for the identification of APS cases. A study of the incidence and prevalence of APS cases in Korea, identified from the Korean Health Insurance and Review Agency, used the Korean Classification of Disease, seventh edition code D68.6 (other thrombophilia), which corresponds to ICD-10: D68.6 and includes aCL syndrome, APS and the presence of LAC, and V253 code of rare intractable disease to confirm APS cases.6 Therefore, the identification and retrieval of APS cases were not exhaustive, which might affect the accuracy of identification. Radin et al (2020) estimated the incidence and prevalence from the Piedmont and Aosta Valley Regional Registry, a part of the Italian National Registry of Rare Diseases, but data from the National Registry are not reported.8 The study was carried out between 2010 and 2019, but the authors do not report on how the cases were ascertained or coded, or whether the Sydney criteria were used, and the potential completeness of the registry has not been described.8 Duarte-Garcia et al (2019) included data from a population-based study run in the Mayo Clinic, Olmsted County, between 2000 and 2015.5 While this study provided a rich clinical and epidemiological knowledge of APS, the authors acknowledged that outcomes could only be generalised to populations with the same demographic profile.5 In addition, the study sample size of only 33 patients after applying the Sydney criteria does not allow extrapolation of outcomes to the country level.5 In the present study, the specific APS code, D68.61, was used to identify patients since the switch from ICD-9-CM to ICD-10-CM in October 2015, and the prior period was excluded to avoid any misclassification or underestimation of APS cases.
Our standardised estimate of APS incidence (2.31 per 100 000 PY in 2016 to a high of 2.71 per 100,000 PY in 2019) is higher than the incidence reported from the Piedmont and Aosta Valley regions in Italy8 and also higher than that reported in the previous US (2.1 per 100 000 PY) and UK (1.8 per 100 000 PY) studies.5 9 The study in South Korea reported a higher incidence rate (7.5 per 100 000 PY) than found in any of the other studies.6 Our estimation of prevalence rate (10.42 per 100 000 persons in 2019) was close to that estimated for the Piedmont and Aosta Valley region (16.8 per 100 000 persons).8 In other studies, prevalence rates ranged from 40 per 100 000 persons in Spain7 to 61.9 per 100 000 persons in South Korea.6 The UK study estimated that prevalence rates were 43 per 100 000 persons9 while the Olmsted County study in the USA had an estimated prevalence of 50 per 100 000 persons.5 The difference in the population inclusion and the design of the study might explain the heterogeneity of prevalence rate estimation. When assessing APS subtype, the overall annual standardised incidences of thrombotic APS and obstetric APS in 2019 were 1.28 per 100 000 PY, and 1.24 per 100 000 PY, respectively. These incidences are close to those reported by Duarte-Garcia et al (2019) but far from those reported by Andreoli et al (2013), who estimated the overall annual incidence of thrombotic APS by the indirect method as 1.8 and 65 per 100 000 PY, respectively, and 0.2 and 15 per 100 000 PY, respectively, for obstetric APS.5 20
It has been previously estimated that CAPS patients represent <1% of all patients with APS, owing to the life-threatening nature of CAPS that requires high clinical awareness.13 However, the proportion of CAPS cases in the present study was substantially higher (4.4%), and more in line with the proportion (3.0%) reported by Hwang et al (2020) in Korea.6 This discrepancy could be attributed to the classification criteria for CAPS, which require the knowledge of aPL test result and a biopsy, neither of which were available in the present study and in the study conducted in Korea. As a result, it is possible that the number of CAPS cases were overestimated in both studies.
This study has several limitations, inherent to retrospective epidemiological studies, and the use of insurance claims databases. First, the Merative MarketScan CCAE and MDCR are commercial insurance databases and under-represent smaller employers and persons aged under 65 years with no occupational health insurance. The CCAE also excludes those with state-funded insurance. Additionally, any medical history prior to membership in an insurance plan will also be excluded, and it is possible that some patients were diagnosed with APS prior to entering the MarketScan database. Similarly, the MDCR captures information only for the subset of Medicare patients who have supplemental insurance paid by their employers. As a result, our estimate of incidence is calculated using claims from only a subset of the total US population, and therefore may not be fully accurate or generalisable and extrapolated. Second, the data is collected for billing of insurance claims, and the validity of the claims is reliant on the accuracy of the data in the claims, therefore may not truly reflect medical diagnosis. Third, antibody titres were not available, and not all APS claims were linked to the clinical criteria as required per the Sydney classification.2 Thus, the case definition used in this study, although based on the Sydney classification criteria,2 is not complete. We have minimised this limitation by using a case definition that requires two aPL test claims at least 12 weeks apart, plus an APS diagnosis claim on or after the second laboratory diagnosis. However, there is no data to confirm that these tests yielded positive and persistent results. This missing information could result in cases being classified as APS which do not meet the Sydney criteria, thus overestimating the incidence and prevalence of APS in the USA. The lack of aPL data is a limitation inherent to the database, which may have contributed to very few populational studies in the literature on the prevalence of aPL and a lack of standardisation between aPL tests. Relatedly, to ensure that we captured every possible APS case, we did not limit APS claims to include only those with a clinical manifestation. Nevertheless, a sensitivity analysis (Scenario F) has been considered including the condition of clinical criteria (per the Sydney classification) linked to APS claims. These patients, classified as ‘Unknown’ in our study, may have therefore had a specific clinical manifestation or APS subtype that was not captured. For example, data may have been incomplete for obstetric APS if patients were misclassified as ‘Unknown’ but were pregnant at the time or even after the period of identification. Indeed, a peak in incidence and prevalence of APS, irrespective of subtype, is observed in women of childbearing age that is not observed in men (online supplemental tables 5 and 6). Additionally, some of the confidence intervals were wide-ranging, owing to variability in the number of events per category, and should be interpreted with caution. A further limitation of the case definition is that primary and secondary APS were not separated. Finally, the absence of a unique ICD code for APS prior to the introduction of ICD-10-CM in October 2015 limits the ability to estimate APS incidence and prevalence prior to 2016. Our estimation of prevalence from 2016 to 2019 is likely to underascertain prevalent cases diagnosed prior to October 2015. The impact of this underestimation of prevalence will be most pronounced in the years immediately after the change in ICD coding.
Very few epidemiological studies are reported in the published literature on the incidence or prevalence of APS, and estimates reported in these studies are heterogeneous. A careful interpretation should be considered when comparing these results to other countries that may have a different healthcare system with variations in APS management, including treatment administration, as well as possible socioeconomic differences.
This study identified a trend towards a yearly increase in the incidence and prevalence of APS in the USA over the study period. The results of this study add to existing estimates published in the literature, but further studies are needed to fully elucidate the global epidemiology of APS using the 2023 classification criteria of APS by Barbhaiya et al4
Data availability statement
No data are available. Data from non-interventional studies are outside of UCB's data sharing policy and are unavailable for sharing.
Ethics statements
Patient consent for publication
Ethics approval
The CCAE and MDCR databases are Health Insurance Portability and Accountability Act compliant, and all patient data were deidentified before delivery to the study team. No ethics committee approval was required.
Acknowledgments
The authors thank Margarita Lens, MSci, CMPP of UCB for publication and editorial support.
Footnotes
Contributors MK, PM, MS, PZ, AJ and AS provided substantial contributions to the conception or design of the work, analysis or interpretation of data. A. Scowcroft is responsible for the overall content as guarantor. All authors revised the work for important intellectual content and provided approval of the final version to be published.
Funding This research was funded by UCB. The study design, data collection, analysis and interpretation of the data were completed by UCB employees or contractors. Medical writing support was provided by Rachel Price of Ogilvy Health, London, UK, and funded by UCB, in accordance with Good Publications Practice (GPP3) guidelines (http://www.ismpp.org/gpp3). Editorial support was funded by UCB. The decision to submit the paper for publication was provided by all authors who are employees or contractors of UCB.
Competing interests MK, PZ and AJ are contractors employed by UCB. PM, MS and AS are employees and stockholders of UCB.
Patient and public involvement Patients and/or the public were not involved in the design, conduct, reporting or dissemination plans of this research.
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.