Co-morbidities increase the risk of disability pension among MS patients: a population-based nationwide cohort study
© Tinghög et al.; licensee BioMed Central Ltd. 2014
Received: 2 April 2014
Accepted: 28 May 2014
Published: 3 June 2014
Multiple sclerosis (MS) is a chronic and often disabling disease. In 2005, 62% of the MS patients in Sweden aged 16–65 years were on disability pension. The objective of this study is to investigate whether the presence of common co-morbidities increase MS patients’ risk for disability pension.
This population-based cohort study included 4 519 MS patients and 4 972 174 non-MS patients who in 2005 were aged 17–64 years, lived in Sweden, and were not on disability pension. Patients with MS were identified in the nationwide in- and outpatient registers, while four different registers were used to construct three sets of measures of musculoskeletal, mental, and cardiovascular disorders. Time-dependent proportional hazard models with a five-year follow up were performed, adjusting for socio-demographic factors.
All studied disorders were elevated among MS patients, regardless of type of measure used. MS patients with mental disorders had a higher risk for disability pension than MS patients with no such co-morbidities. Moreover, mental disorders had a synergistic influence on MS patients’ risk for disability pension. These findings were also confirmed when conducting sensitivity analyses. Musculoskeletal disorders appeared to increase MS patients’ risk for disability pension. The results with regard to musculoskeletal disorders’ synergistic influence on disability pension were however inconclusive. Cardiovascular co-morbidity had no significant influence on MS-patients’ risk for disability pension.
Co-morbidities, especially mental disorders, significantly contribute to MS patients’ risk of disability pension, a finding of relevance for MS management and treatment.
KeywordsMultiple sclerosis Co-morbidity Disability pension Sick leave Synergistic effects Insurance medicine
Multiple sclerosis (MS) is an often progressive neurological disorder that may lead to substantial disability [1–3]. Some MS patients quickly experience permanent work incapacity while others maintain a high level of work capacity for several years [4, 5]. Co-morbidity has been suggested as a key factor for understanding heterogeneity of the MS progression .
Research on how MS-patients are affected by co-morbidities has so far focused on other outcomes than disability pension (DP), such as ambulatory disability,  health-related quality of life,  and physical functioning . It has been reported that MS patients with vascular disorders are more likely to suffer from ambulatory disability,  that MS patients with musculoskeletal disorders have a more rapid decline of motor functions,  and that mental disorders among MS patients are linked to decreased physical functioning  and increased perceived disability . No population-based study with a comparison group of non-MS patients has, to our knowledge, been conducted to determine if MS in combination with other disorders has a synergistic influence on a disability outcome. We have chosen DP as an outcome as it also involves the social consequences of the reduced function, in terms of permanent work incapacity.
This study aimed at analyzing; 1) the presence of musculoskeletal, cardiovascular, and mental disorders in MS patients and in the general population of working ages; 2) if musculoskeletal, cardiovascular, and mental co-morbidity increase the risk of DP among MS patients; and 3) if these three types of disorders act synergistically on MS patients’ risk for DP.
This study shows that co-morbidities, especially mental disorders, significantly contribute to MS patients’ risk of disability pension.
Baseline descriptives (2005) in percentages and incidence rates (IRs) for DP per 100 000 person-years (2006–2010) among MS patients and the general population, respectively
All (n = 10 750)
At risk for DP (n = 4 519)
All (n = 5 553 120)
At risk for DP (n = 4 972 174)
Person-years at risk
Person-years at risk
11 324 670
12 271 042
Age (mean years)
4 033 844
5 284 065
5 719 389
4 874 419
3 683 995
Living with partner
11 733 753
11 861 959
Compulsory School (≤9 years)
4 258 677
High School (10–12 years)
11 339 322
University (≥13 years)
7 997 713
Country of birth
20 482 499
Other Nordic countries
2 052 786
Type of living area
8 854 806
8 505 141
6 235 765
3 951 156
East Middle Sweden
2 053 102
Småland and Islands
3 421 054
4 720 280
2 056 790
North Middle Sweden
1 302 967
3 951 156
Linkage and data sources
By using the Personal Identity Number (a unique ten-digit number assigned to all Swedish residents), data from the following five nationwide registers were linked for each of the included individuals: 1) Statistics Sweden’s Longitudinal Integration Database for Health Insurance and Labor Market Studies (LISA) regarding data on socio-demographics and migration; 2) Social Insurance Agency’s database Micro Data for Analysis of the Social Insurance (MiDAS) regarding data on disability pension and diagnosis-specific sick-leave; 3) National Board of Health and Welfare’s databases National Patient Register (PAR), 4) Swedish Prescribed Drug Register (PDR), and 5) the Causes of Death Register from which data about diagnosis-specific in- and specialized outpatient care, prescribed drugs, and year of death, respectively, were obtained. All five registers are longitudinal, but differ with regard to when they were instigated. Important to mention in relation to this study is that nationwide specialized outpatient data only is available from 2001 and onwards, that reliable data on sick-leave diagnoses is available from 2004, and that the PDR register started 1 July 2005.
The study was approved by the Regional Ethical Review Board in Stockholm, Sweden.
In Sweden, all adult residents with a disease or injury that has led to permanent work incapacity are entitled to disability pension. Disability pension covers up to 64% of the lost income. The customary age for old-age pension is 65 years, but may be taken earlier. Also, all people above the age of 16 with income from work or unemployment benefits can be entitled to sickness benefits if a disease or injury has led to work incapacity.
The MS patients were identified using the nationwide PAR; that is, those who had at least one hospitalization or outpatient specialist visit due to MS as a main or secondary diagnosis during 2000–2005, classified according to the International Statistical Classification of Diseases and Related Health problems ICD-10 ; G35.
Separate time-dependent dummy variables were constructed for musculoskeletal, cardiovascular, and mental disorders. Year-specific data for these disorders were retrieved from the PDR, MiDAS, and PAR, respectively. The first year the disorder was observed and the years following were coded as 1, while the preceding years were coded as 0. Individuals without the respective disorder were consistently coded as 0. To circumvent some of the potential drawback inherited with using register data to identify individuals with these three classes of disorders , three different types of measures were constructed.
In the first, and most inclusive, measure, i.e. model 1, individuals were classified as having musculoskeletal disorder at baseline if they had been hospitalized or received specialized outpatient care between 2000 and 2005 with a musculoskeletal disorder (ICD-10: M00-M99), or had been sickness absent due to musculoskeletal diagnoses (ICD-10: M00-M99) in 2004 or 2005. Also, from the PDR we used prescriptions for dispensed drugs licensed for musculoskeletal disorders (Anatomical Therapeutical Chemical Classification (ATC)-codes: M01-M09) in 2005. Similarly, individuals were classified as suffering from cardiovascular and mental disorders in the same manner, using the following ATC and diagnostic codes: cardiovascular disorders ATC: C01-C10; ICD-10: I00-I99 and mental disorders ATC: N05-N06; ICD-10: F00-F99.
As we were concerned about overestimation and that differential misclassification may bias the estimates obtained when applying the above described measures – in particular relevant for PDR data since no information on indication is included in this register that may make it a less specific proxy for diagnosis – two additional and more conservative measures were constructed. In the first of these, i.e. model 2, the classes of disorders were defined on the sole basis of the sick leave and the in- and outpatient diagnoses, i.e., PAR and MiDAS. The second type of alternative measures, i.e. model 3, were based on all four registers as described above (i.e. in and outpatient PAR, MiDAS and PDR), with the exception that drugs belonging to ATC groups hypnotics and sedatives (N05C), centrally acting sympathomimetics (N06BA), anesthetics (M01), and muscle relaxants (M03) were excluded, as drugs of these kinds may be prescribed to treat MS symptoms.
Cohabiting status was also constructed as a time-dependent variable, but in this case individuals were only classified in reference to the preceding year.
Those living in Sweden all of 2005 were identified through LISA and the following fixed covariates, i.e. at baseline, were retrieved from LISA: age-groups (17–24, 25–34, 35–44, 45–54, 55–64); educational level [compulsory school (≤9 years), high school (10–12 years), university (≥13 years)]; country of birth (Sweden, other Nordic countries, EU 25 or other countries); type of living area [based on the H-region classification scheme  into the following 3 categories: larger cities (H1-H2), medium-sized municipalities (H3-H4), or smaller municipalities (H5-H6)]; and geographic region [in 8 categories; Stockholm County, South Sweden, East Middle Sweden, North Middle Sweden, Middle Norrland, Småland and Islands, West Sweden, or Upper Norrland in accordance with Eurostat’s Nomenclature of Territorial Units for Statistics, (NUTS) classification (level 2)].
The cohort was followed from 2006 through 2010 or the year the individual turned 65, emigrated, died, or received DP, whichever came first. Hazard ratios (HRs) with 95% confidence intervals (CIs) were estimated by time-dependent proportional hazards model.
First, descriptive analyses were performed to explore the distribution of the baseline covariates among MS patients and in the general population, respectively. The absolute risks for DP by baseline characteristics were calculated for MS patients and the general population and presented as incidence rates (IRs) per 100 000 person-years.
Second, 5-year prevalence estimates, based on the three specified types of measures, of musculoskeletal, cardiovascular, and mental disorders were computed for MS patients and the general population. The MS patients’ and the general population’s 5-year prevalence estimates were compared in adjusted logistic regression analyses. Separate analyses were conducted for those at risk for DP and all individuals, i.e. including also those on early old-age pension or DP. Incidence rates for DP (IRs) per 100 000 person-years were provided for the MS patients and the general population, respectively.
Third, models based on the three specified types of measures of disorders were tested to establish whether the studied co-morbid conditions influenced the MS patients’ risk for DP. To illustrate effect-modifications also HRs with 95% CIs were calculated for the general population. Effect-modifications were evaluated using Wald ×2 tests.
Fourth, Rothman’s synergy index (SI) and attributable proportion due to interaction (AP) were calculated . These statistics were obtained with 95% CIs, following Andersson et al’s recommendations . A SI above 1 indicates a synergistic effect and a SI lower than 1 indicates an antagonistic synergistic effect. Models based on the three pre-specified types of proxy measures were tested separately.
Table 1 shows that MS patients have a different socio-demographic profile than the general population. The MS patients were more often women, cohabiting, university educated, and born in Sweden. It was also noticeable that the socio-demographic differences between MS patients and the general population became more pronounced when comparing only those at risk for DP. Furthermore, older age, lower educational level, living in a small municipality or in the northern part of Sweden seemed to be predictors of DP in the MS population. In the general population, similar trends were observed, though the absolute DP risks were overall much lower.
Five year prevalence estimates for different measures (2000–2005) for musculoskeletal, cardiovascular and mental disorders among MS patients and the general population, with incidence rates (IRs) for DP per 100 000 person-years and adjusted odds ratios (ORs) with 95% confidence intervals (CIs)
MS patients vs. General population (All)d
MS patients vs. General population (at risk for DP)d
All (n = 10 791)
At risk for DP (n = 4 519)
All (n = 5 618 191)
At risk for DP (n = 4 972 174)
Model 1 b
Model 2 a
Model 3 c
Musculoskeletal disorders (exl. ATC: M01, M03)
Mental disorders (exl. ATC:N06AB, N05c)
The influence of different measures for musculoskeletal, cardiovascular, and mental disorders on DP among MS patients and the general population during follow-up 2006–2010 estimated as hazard ratios (HRs) with 95% confidence intervals (CI)
MS patients HRs (95% CI)
General population HRs (95% CI)
Effect modifications Wald X 2(p-values)
Model 1 b
Model 2 c
Model 3 d
Musculoskeletal, cardiovascular, and mental disorders’ synergistic influence on DP in a five-year follow up, presented as hazard ratios (HRs), attributable proportion due to interaction (AP), and synergy index (SI) a
HRs (95% CI)
AP % (95% CI)
SI (95% CI)
No musculoskeletal disorder and no MS
Musculoskeletal disorder (only)
Musculoskeletal disorder and MS
No cardiovascular disorder and no MS
Cardiovascular disorder (only)
Cardiovascular disorders and MS
No mental disorder and no MS
Mental disorder (only)
Mental disorder and MS
Age- (16–44 and 45–64 years) and gender-stratified analyses were conducted to evaluate the fit of the models (data not shown). The estimates from these analyses (based on model 1) concerning the influence of co-morbidity were comparable across genders and age-groups. However, worth mentioning is that cardiovascular disorders were associated with a higher HR for DP among the younger MS patients, i.e. HR 1.38 (1.13-1.69).
This prospective and population-based register study is, as far as we know, the first dealing with how co-morbidity influences MS patients’ risk for DP. As expected, MS patients with musculoskeletal and mental co-morbidity had a higher risk for DP, but contrary to our expectation, cardiovascular disorders did not increase MS patients’ risk for DP compared to MS patients without such co-morbidity. Our results also showed that musculoskeletal, cardiovascular, and mental disorders were more common among MS patients of working ages but were, in a relative sense, stronger predictors for DP in the general population than in the MS population. Furthermore, mental disorders had a synergistic influence on MS patients’ risk for DP. The results regarding musculoskeletal disorders synergistic influence on DP were inconclusive.
The finding that musculoskeletal and mental disorders increased MS patients’ risk for DP is in accordance with previous research where different disability measures have been used [10–12]. It was, however, unexpected that cardiovascular disorders did not predict DP among MS patients. This may be interpreted as that this specific co-morbid condition is negligible in the context of MS and work incapacity, as MS in itself is a severe and disabling disorder. It may also be a result of that a cardiovascular disorder often are attained after the age of 50, when many MS patients already have experienced a reduced work capacity and been granted disability pension.
In contrast to our results, a large US cohort study found that MS patients with vascular co-morbidity at diagnosis had more than a 1.5 folded increased risk of ambulatory disability . However, important differences exist; we used another outcome measure and incorporated co-morbid conditions occurring during follow-up. Moreover, the methods for defining the co-morbid disorders differed. We used four nationwide registers to identify occurrences of co-morbidity, while Marrie et al.  relied on self-reported data. Marrie et al used the term vascular disorders, including e.g. diabetes, while we employed ICD-10 chapters and pre-established groups of ATC-codes when defining the co-morbid disorders.
That mental disorders are highly overrepresented among MS patients has often been reported [18, 19]. Several studies have also shown that the severity of MS cannot be linked to having depression or anxiety in a straight forward manner, instead they are common in all forms and stages of MS, [18, 20–23] yet other studies have reported somewhat contradictory findings [11, 24]. However, the majority of prior studies support the notion that the higher risk for DP among MS patients with mental disorders cannot be explained only as a consequence of especially high rates of mental disorders among severe cases of MS. Still, when interpreting the influence of mental disorders, some caution is warranted as a common pathogenic agent that influences inflammatory markers may be involved in both MS and depression . and MS may sometimes cause mental disorders through purely psychological mechanisms.
In contrast to mental disorders, co-morbidity of musculoskeletal and cardiovascular disorders among MS patients has seldom been studied. Previous attempts to compare the presence of these disorders in an MS population to that in a representative population without MS have reported contradictory findings [26–28]. In the present study, all results support the notion that musculoskeletal and cardiovascular disorders are more common among individuals with MS than they are in the general population.
The strengths of the present study is its population-based and prospective cohort design, the large cohort covering a whole country, no loss to follow up, i.e. avoiding selection bias, and the use of several data sources to estimate the prevalence of co-morbidities, i.e. information about in-patient and outpatient specialized care, on specific prescribed drugs, as well as on sick-leave diagnoses – rather than self-reports. We know of no other study using such a wide spectrum of data on co-morbidity.
A potential weakness with this study concerns the potential influence of differential misclassification. First, some drugs used to operationalize musculoskeletal and mental disorders can also be prescribed for MS symptoms, e.g. hypnotics and sedatives, centrally acting sympathomimetics, anesthetics, and muscle relaxants. Second, MS patients consume more specialized health care and may thereby be more likely to become diagnosed with an additional disorder in this study. Third, it is possible that MS patients are more likely to at some point before receiving their MS diagnosis have been misdiagnosed with a musculoskeletal or mental disorder. We thus recognize that all used registers have their flaws that may both underestimate and overestimate the true differences between MS patients and the general population with regard to prevalence rates of co-morbidities. To deal with these limitations additional analyses, based on different case ascription methods were conducted. On most occasions, but not all, these analyses corroborated one another.
To conclude; this study suggests that attention should be given to co-morbidity in order to better understand the DP trajectory among MS patients. This study was based on fairly broad categories of disorders and it is possible that different and/or more specific case ascriptions would nuance our findings. Additional population-based register studies focusing on how specific diagnoses or drugs influence MS patients’ work incapacity would thus be valuable.
- Kobelt G, Berg J, Lindgren P, Fredrikson S, Jönsson B: Costs and quality of life of patients with multiple sclerosis in Europe. J Neurol Neurosur Ps. 2006, 77: 918-926. 10.1136/jnnp.2006.090365.View ArticleGoogle Scholar
- Tinghög P, Hillert J, Kjeldgård L, Wiberg M, Glaser A, Alexanderson K: High prevalence of sickness absence and disability pension among multiple sclerosis patients: a nationwide population-based study. Mult Scle J. 2013, 19: 1923-1930. 10.1177/1352458513488234.View ArticleGoogle Scholar
- WHO: World Report on Disability. 2011, Geneva: WHOGoogle Scholar
- Lassmann H, Bruck W, Lucchinetti CF: The immunopathology of multiple sclerosis: an overview. Brain Pathol. 2007, 17: 210-218. 10.1111/j.1750-3639.2007.00064.x.View ArticlePubMedGoogle Scholar
- Noseworthy JH: Progress in determining the causes and treatment of multiple sclerosis. Nature. 1999, 399: A40-A47.View ArticlePubMedGoogle Scholar
- Marrie RA, Horwitz RI: Emerging effects of comorbidities on multiple sclerosis. Lancet Neurol. 2010, 9: 820-828. 10.1016/S1474-4422(10)70135-6.View ArticlePubMedGoogle Scholar
- Marrie R, Rudick R, Horwitz R, Cutter G, Tyry T, Campagnolo D, Vollmer T: Vascular comorbidity is associated with more rapid disability progression in multiple sclerosis. Neurology. 2010, 74: 1041-1047. 10.1212/WNL.0b013e3181d6b125.View ArticlePubMedPubMed CentralGoogle Scholar
- Turpin K, Carroll L, Cassidy J, Hader W: Deterioration in the health-related quality of life of persons with multiple sclerosis: the possible warning signs. Mult Scler. 2007, 13: 1038-1045. 10.1177/1352458507078393.View ArticlePubMedGoogle Scholar
- Dallmeijer AJ, Beckerman H, de Groot V, van de Port IG, Lankhorst GJ, Dekker J: Long-term effect of comorbidity on the course of physical functioning in patients after stroke and with multiple sclerosis. J Rehabil Med. 2009, 41: 322-326. 10.2340/16501977-0335.View ArticlePubMedGoogle Scholar
- Marrie R, Horwitz R, Cutter G, Tyry T, Campagnolo D, Vollmer T: Comorbidity delays diagnosis and increases disability at diagnosis in MS. Neurology. 2009, 72: 117-124. 10.1212/01.wnl.0000333252.78173.5f.View ArticlePubMedPubMed CentralGoogle Scholar
- Chwastiak L, Ehde DM, Gibbons LE, Sullivan M, Bowen JD, Kraft GH: Depressive symptoms and severity of illness in multiple sclerosis: epidemiologic study of a large community sample. Am J Psychiat. 2002, 159: 1862-1868. 10.1176/appi.ajp.159.11.1862.View ArticlePubMedGoogle Scholar
- Smith S, Young C: The role of affect on the perception of disability in multiple sclerosis. Clin Rehabil. 2000, 14: 50-54. 10.1191/026921500676724210.View ArticlePubMedGoogle Scholar
- World Health Organization: International statistical classification of diseases and related health problems. 2004, Geneva, 10th revision, 2Google Scholar
- Alexanderson K: Measuring health. Indicators for working women. Women’s health at work. Edited by: AMK K, Bildt Thorbjörnsson C. 1998, Stockholm: National Institute for Working LifeGoogle Scholar
- Statistics Sweden: Rikets indelningar: årsbok över regionala indelningar med koder, postadresser, telefonnummer m m. 2003 [Country classifications: yearbook of regional classifications with codes, postal addresses, phone numbers etc. 2003]. 2003, StockholmGoogle Scholar
- Rothman KJ, Greenland S, Lash TL: Modern epidemiology. 2008, Philadelphia: Lippincott Williams & WilkinsGoogle Scholar
- Andersson T, Alfredsson L, Källberg H, Zdravkovic S, Ahlbom A: Calculating measures of biological interaction. Eur J Epidemiol. 2005, 20: 575-579. 10.1007/s10654-005-7835-x.View ArticlePubMedGoogle Scholar
- Beiske A, Svensson E, Sandanger I, Czujko B, Pedersen E, Aarseth J, Myhr K: Depression and anxiety amongst multiple sclerosis patients. Eur J Neurol. 2008, 15: 239-245. 10.1111/j.1468-1331.2007.02041.x.View ArticlePubMedGoogle Scholar
- Patten SB, Svenson LW, Metz LM: Psychotic disorders in MS: population-based evidence of an association. Neurology. 2005, 65: 1123-1125. 10.1212/01.wnl.0000178998.95293.29.View ArticlePubMedGoogle Scholar
- Gottberg K, Einarsson U, Fredrikson S, von Koch L, Holmqvist LW: A population-based study of depressive symptoms in multiple sclerosis in Stockholm county: association with functioning and sense of coherence. J Neurol Neurosur Ps. 2007, 78: 60-65. 10.1136/jnnp.2006.090654.View ArticleGoogle Scholar
- Janssens AC, Buljevac D, Van Doorn P, van der Meché FG, Polman C, Passchier J, Hintzen R: Prediction of anxiety and distress following diagnosis of multiple sclerosis: a two-year longitudinal study. Mult Scler. 2006, 12: 794-801. 10.1177/1352458506070935.View ArticlePubMedGoogle Scholar
- Möller A, Wiedemann G, Rohde U, Backmund H, Sonntag A: Correlates of cognitive impairment and depressive mood disorder in multiple sclerosis. Acta Psych Scand. 1994, 89: 117-121.View ArticleGoogle Scholar
- Arnett P, Randolph J: Longitudinal course of depression symptoms in multiple sclerosis. J Neurol Neurosur Ps. 2006, 77: 606-610. 10.1136/jnnp.2004.047712.View ArticleGoogle Scholar
- Figved N, Klevan G, Myhr KM, Glad S, Nyland H, Larsen JP, Harboe E, Omdal R, Aarsland D: Neuropsychiatric symptoms in patients with multiple sclerosis. Acta Psych Scand. 2005, 112: 463-468.View ArticleGoogle Scholar
- Gold SM, Irwin MR: Depression and immunity: inflammation and depressive symptoms in multiple sclerosis. Immunol Allergy Clin. 2009, 29: 309-320. 10.1016/j.iac.2009.02.008.View ArticleGoogle Scholar
- Jadidi E, Mohammadi M, Moradi T: High risk of cardiovascular diseases after diagnosis of multiple sclerosis. Mult Scler. 2013, 19: 1336-1340. 10.1177/1352458513475833.View ArticlePubMedGoogle Scholar
- Kang JH, Chen YH, Lin HC: Comorbidities amongst patients with multiple sclerosis: a population-based controlled study. Eur J Neurol. 2010, 17: 1215-1219. 10.1111/j.1468-1331.2010.02971.x.View ArticlePubMedGoogle Scholar
- Marrie R: The influence of comorbid diseases and health behaviors on clinical characteristics, disability at diagnosis, and disability progression in multiple sclerosis. 2007, Cleavland, Ohio: Case Western Reserve UniversityGoogle Scholar
- The pre-publication history for this paper can be accessed here:http://www.biomedcentral.com/1471-2377/14/117/prepub
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.