fetching data ...

POS0238 (2021)
SICK LEAVE AND ITS PREDICTORS IN EARLY AXIAL SPONDYLOARTHRITIS: THE ROLE OF CLINICAL AND SOCIOECONOMIC FACTORS. FIVE-YEAR DATA FROM THE DESIR COHORT
E. Nikiphorou1,2, P. Carvalho3,4,5, A. Boonen6,7, B. Fautrel8, P. Richette9, P. M. Machado10, D. Van der Heijde11, R. B. M. Landewé12,13, S. Ramiro11,12
1King’s College London, Centre for Rheumatic Diseases, London, United Kingdom
2King’s College Hospital NHS Trust, Rheumatology, London, United Kingdom
3Centro Hospitalar Universitário do Algarve, Rheumatology, Faro, Portugal
4Centro Académico Médico de Lisboa, Rheumatology, Lisboa, Portugal
5Algarve Biomedical Center, Rheumatology, Faro, Portugal
6Maastricht University Medical Center and Care and Public Health Research Institute (CAPHRI), Rheumatology, Maastricht, Netherlands
7Academic Hospital Maastricht, Internal Medicine, Div. of Rheumatology, Maastricht, Netherlands
8Sorbonne University – Assistance Publique Hopitaux de Paris, Pitie Salpetriere Hospital, Pierre Louis Institute for Epidemiology and Public Health, INSERM UMRS 1136, PEPITES Teams, Rheumatology, Paris, France
9Université de Paris, Hopital Lariboisière, Rheumatology, Paris, France
10University College London, Centre for Rheumatology, London, United Kingdom
11Leiden University Medical Center (LUMC), Rheumatology, Leiden, Netherlands
12Zuyderland MC, Rheumatology, Heerlen, Netherlands
13Amsterdam University Medical Center, Rheumatology, Amsterdam, Netherlands

Background: Sick leave (SL) represents an often poorly studied adverse work outcome especially in early axSpA, with speculation around the potential role of clinical and socioeconomic (SE) factors.


Objectives: To investigate the occurrence of SL and the impact of clinical and SE factors on SL in early axSpA.


Methods: Patients with a clinical diagnosis of axSpA from the DESIR cohort up to 5 years of follow-up (6-month visits in the first 2 years, followed by annual visits) were studied. Time to SL and potential baseline and time-varying predictors were explored, with a focus on SE variables: age, gender, smoking status since last visit, ethnicity (Caucasian vs other), job type based on ‘collar’ (blue vs white), educational status (low vs high -university), marital status (married vs not) and parental status (number of children); and clinical factors including disease activity (ASDAS/BASDAI), function (BASFI), mobility (BASMI), at each time point. The incidence of SL was calculated as the number of SL events over the total number of person-days under observation. Univariable analyses, followed by collinearity and interaction tests, guided subsequent multivariable Cox survival model building.


Results: In total, 704 axSpA patients with work-related data were included in this study: mean (SD) age 33.8 (8.6); 46% male. At baseline, 80% of patients were employed; of these, 5.7% reported being on SL, with people shifting in and out of different work states over time. The distribution of first and recurrent SL episodes over time is shown in the figure 1 . The incidence of SL amongst those at risk (n=620, 88%) and across the five years of DESIR was 0.05 (95% CI 0.03, 0.06) per 1000 days calculated in a total of 913,559 observed person-days. In survival analyses, 7% (n=43) of those at risk developed SL at some point. Mean (SD) time to SL was 806 (595) days (min 175, max 2021 days). In people who developed SL, 25% did so at 364 days; 50% and 75% at 545 and 1172 days, respectively. Significant differences were seen between baseline socio-demographic, clinical variables and treatment in patients who developed SL at any point, compared to those who did not. In multivariable models ( Table 1 ) older age, higher disease activity, smoking and use of TNFi, the latter likely a proxy to worse disease, were all significantly associated with more SL. Male gender and higher education were associated with less SL. There were no relevant interactions between SE factors and clinical variables.

Univariable and multivariable model analyses with Sick Leave as outcome.

Type of analysis Univariable analysis Multivariable model
HR (95% CI) HR (95% CI) (N = 614)
Explanatory variables
Age 1.04 (1.01, 1.08) 1.05 (1.01, 1.09)
Male gender 0.37 (0.19, 0.74) 0.41 (0.20, 0.86)
High education 0.33 (0.17, 0.61) 0.48 (0.24, 0.95)
Marital status 2.44 (1.12, 5.27) NS
ASDAS (CRP) 1.83 (1.34, 2.50) 1.49 (1.04, 2.13)
BASFI, 0-10 1.24 (1.09, 1.40) *
BASMI, 0-10 1.76 (1.31, 2.38) *
Comorbidity count 1.77 (1.22, 2.57) NS
HLA-B27 positive 0.51 (0.28, 0.93) NS
Smoking (current vs not) 2.40 (1.31, 4.37) 2.55 (1.32, 4.91)
NSAID score last week, 0-400 1.01 (1.00, 1.01) NS
Oral Corticosteroid use (vs no) 3.90 (1.80, 8.46) NS
TNF use 2.86 (1.55, 5.28) 2.41 (1.27, 4.58)

*Variables tested in models separate from ASDAS. NS=Not significant in multivariable model.

Distribution of first and recurrent sick leave episodes over time in the study population at risk.


Conclusion: In this early axSpA cohort of young, working-age individuals, older age and worse disease activity were associated with more SL, whereas male gender and higher education were associated with less SL. The findings suggest a role of SE factors such as gender and level of education in adverse work outcomes, alongside active disease.


Disclosure of Interests: None declared

Citation: Ann Rheum Dis, volume 80, supplement 1, year 2021, page 340
Session: Clinical aspects of spondyloarthritis (Poster Tours)