Abstract
Objectives
Multimorbidity is burdensome for people with RA. We investigated differences in multimorbidity and comorbidities by sex and age in the RA population.
Methods
This cross-sectional analysis used national administrative claims (OptumLabs® Data Warehouse) from people with RA and non-RA comparators (matched on age, sex, race, census region, index year and length of baseline insurance coverage) from 2010–2019. RA was determined using a validated algorithm. Multimorbidity was defined as ≥2 (MM2+) or ≥5 (MM5+) comorbidities from a validated set of 44 chronic conditions. We used logistic regression to assess associations between characteristics and multimorbidity.
Results
The sample included 154 391 RA patients and 154 391 non-RA comparators. For people aged 18–50 years, RA women (vs RA men) had 7.5 and 4.4 (vs 3.2 and 0.9 in non-RA women vs non-RA men) percentage point increases for MM2+ and MM5+, respectively. For people aged 51+ years, RA women (vs RA men) had 2.1 and 2.5 (vs 1.2 and 0.3 in non-RA women vs non-RA men) percentage point increases for MM2+ and MM5+, respectively. Interactions revealed that differences in multimorbidity between women and men were exacerbated by RA (vs non-RA) (P < 0.05), with more pronounced effects in people aged 18–50. Men had more cardiovascular-related conditions, whereas RA women had more psychological, neurological and general musculoskeletal conditions. Other comorbidities varied by sex and age.
Conclusion
Multimorbidity disproportionately impacts women with RA. Research, clinical and policy agendas for rheumatic diseases should acknowledge and support the variation in care needs by sex and gender across the lifespan.
Keywords: RA, multimorbidity, comorbidity, sex, disparities
Rheumatology key messages.
Women with RA have a higher prevalence of multimorbidity than men with RA, especially among those aged 18–50.
Women with RA had more psychological and musculoskeletal conditions and men with RA had more cardiovascular-related conditions.
Women with RA experience more multimorbidity than women without RA.
Introduction
For people with RA, a chronic inflammatory autoimmune disease that is systemic in nature [1], the presence of co-occurring comorbidities and multimorbidity (i.e. the co-occurrence of two or more chronic conditions) is commonplace [2, 3]. A population-based cohort study found that people with RA had a prevalence of 86% (vs 71% for non-RA) for ≥2 morbidities and 55% (vs 38% for non-RA) for ≥5 morbidities [2]. Despite the introduction of advanced therapeutics for RA, patients with RA and multimorbidity continue to experience a higher risk for mortality, reduced functional status and poorer quality of life [4]. Given the increased prevalence and burden of multimorbidity in the RA population, a greater understanding of the factors associated with multimorbidity and comorbidities in people with RA is needed.
Biological sex and gender are two risk factors that play an essential role in the treatment and outcomes for rheumatic diseases and associated pathologies. RA predominately affects women and hormonal factors [5] and encoding of genes in chromosomes (X [6] and Y [7]) are known to contribute to differences in immune system responses. For example, oestrogen impacts inflammation and RA disease activity [5, 8]. Chronic inflammation also contributes to the development or worsening of many chronic conditions. For instance, it accelerates atherosclerosis and increases the risk for cardiovascular diseases [9], which often occurs at younger ages in men than in women [10]. Sociocultural aspects of gender may also influence the course of RA. For example, men with RA more frequently develop extra-articular manifestations of disease (e.g. interstitial lung disease) which may be related to differences in accompanying behavioural risk factors (e.g. smoking) when compared with women with RA [10]. Moreover, sex and gender inequities have been demonstrated in the treatment of RA wherein, compared with men, women demonstrated decreased adherence (e.g. immunotherapy [11] and biologics [12]), less symptom remission [13] and poorer treatment responses [14].
Disparities based on sex and gender have also been observed in the general population for multimorbidity in which nine of fourteen studies demonstrated a higher prevalence of multimorbidity in women compared with men [15]. However, limited research has explicitly investigated these disparities in the RA population. This knowledge gap is problematic because multimorbidity and its associated complexity complicate treatment and impact outcomes for patients with RA [16]. Differences in multimorbidity in the RA population based on sex and gender may also vary with age because of the rise and fall of various hormones among other factors across the lifespan [8]. However, there is limited evidence on how sex and age together influence the development of multimorbidity in people with RA.
To address this gap in knowledge, we leveraged a national claims database to evaluate whether the relationship between sex and multimorbidity varied by RA status in two populations: people aged 18–50 years; people aged 51+ years. We hypothesized that multimorbidity would be more common in women compared with men in the RA (vs non-RA) population and that these differences would be more pronounced among people aged 18–50 years.
Methods
This study follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines. This study used de-identified administrative claims data and thus was exempt from Institutional Review Board review.
Study design and setting
We conducted a 1:1 matched cross-sectional study using administrative claims between 1 January 2010 and 30 June 2019 from the OptumLabs Data Warehouse (OLDW) [17], a database that includes administrative claims from over 149 million people with commercial or Medicare Advantage (MA) insurance plans over the past 20 years [18]. There were 53 000 000 unique individuals in the OLDW during the study period. The OLDW contains longitudinal health information from individuals throughout the US including geographically diverse regions with the greatest representation in Midwestern and Southern states [18]. Characteristics of individuals in the OLDW have been shown to mirror the commercial and MA population in the US across age, race/ethnicity and sex (Supplementary Table S1, available at Rheumatology online). Data include any provider, pharmacy, facility and laboratory claim that is submitted to insurance for reimbursement, as well as individual information from enrolment files including enrolment dates.
Sample
The sample included people with RA, as well as matched non-RA comparators from the OLDW during the study period. All individuals were required to be 18 years or older. A detailed description of the inclusion criteria is displayed in the eligibility flow diagram (Supplementary Fig. S1, available at Rheumatology online).
Rheumatoid arthritis identification
Patients with RA were identified based on the requirements of at least two RA diagnoses between 30 and 365 days apart and at least one disease-modifying antirheumatic drug (DMARD) prescription fill, during the study period. The date of the second RA diagnosis code was used as the index date. We used International Classification of Diseases, Ninth Revision (ICD-9) codes 714.0, 714.1 and 714.2 and Tenth Revision (ICD-10) codes M05.xxx, M06.0xx, M06.8x and M06.9 to identify an RA diagnosis, using professional and facility claims. This approach was based on a previously validated algorithm and demonstrated a positive predictive value (PPV) of 92% for identifying patients with RA [19, 20].
Non-RA comparator identification
Non-RA comparators included people who had (i) at least one evaluation and management (E/M) visit during the study period and (ii) at least 365 days of continuous enrolment prior to the E/M visit date. From the qualifying E/M visits, one E/M visit date was randomly assigned as the index date. Non-RA comparators were excluded if they had an ICD-9 or ICD-10 code for RA or a prescription fill for a DMARD any time during the study period.
Measures
Comorbidities and multimorbidity
Comorbidities were identified from a validated set of 44 chronic conditions developed by England et al. [21] These conditions have been shown to be clinically relevant and prevalent in the RA and general population [2]. Comorbidities were identified based on the occurrence of at least two ICD-9 or ICD-10 codes for a given diagnosis within 365 days prior to the index date. Multimorbidity was defined in two ways: (i) multimorbidity ≥2 [2, 21] (MM2+), which required having at least two or more chronic conditions from the validated set and (ii) multimorbidity ≥5 (MM5+), also termed substantial multimorbidity [3], which required having five or more chronic conditions from the validated set.
Covariates
Characteristics included sex (male, female), age (years), age group (18–50, 51+), race/ethnicity [Asian, non-Hispanic Black (Black), non-Hispanic White (White), Hispanic, Other/Unknown], geographical region (Midwest, Northeast, South, West), calendar year of index date (2010 through 2019) and length of baseline insurance coverage (years). The cut-point for age group was selected based on the average age of the onset of menopause, 51 years. Sex and year of birth were ascertained from updated enrolment files that are provided by the health plan. Race/ethnicity and geographical region were also retrieved from the OLDW. To facilitate interpretation, males and females are referred to as men and women, respectively. While gender was not directly available in claims data, the variable, sex, used in our study may represent biological processes as well as serve as a context-specific proxy for sociocultural aspects.
Analysis
We used greedy matching to match patients with RA 1:1 with non-RA comparators on age, sex, race, geographical region, calendar year of index date and length of baseline insurance coverage (within 30 days). Pearson χ2 tests for categorical variables and analysis of variance and t tests for continuous variables were conducted to analyse bivariate associations between cohort characteristics and RA. Multivariable logistic regression models were performed, with adjustment for pertinent covariates, to assess associations between cohort characteristics and multimorbidity. Odds ratios for MM2+ and MM5+ are reported for observable cohort characteristics. Because we were particularly interested in how the absolute risk of multimorbidity in patients with RA (vs non-RA comparators) may vary by sex in people aged 18–50 and 51+ years, we report predicted probabilities and average marginal effects (AMEs) of MM2+ and MM5+ for women, with and without RA, and men, with and without RA, based on adjusted multivariate logistic regression models for each age group. We tested the differences between AMEs for RA women (vs non-RA women), as well as RA men (vs non-RA men) for both age groups. The prevalence of comorbidities for patients with RA and non-RA comparators by sex are also reported. Lastly, to further portray how RA and sex and age together influence the development of multimorbidity, we plotted the adjusted predicted probability of MM2+ and MM5+ for patients with RA and non-RA comparators by sex across the lifespan. Benign prostatic hyperplasia and non-inflammatory gynaecologic conditions were excluded from multimorbidity-specific analyses so that direct comparisons could be made between sexes.
Results
Cohort characteristics and RA diagnosis
The final cohort included 308 782 (154 391 RA and 154 391 matched non-RA) individuals. Non-RA comparators had similar observable baseline characteristics as patients with RA (Table 1). Patients with RA and non-RA comparators had a mean age of 59.6 (s.d.: 13.6) years, and 76% were women. Overall, about 74% were White, 11% were Black, 2% were Asian, 11% were Hispanic, and 3% were other/unknown. Most individuals were from the South (50%) and Midwest (26%), and the median length of baseline insurance coverage was ∼1.6 years. The unadjusted prevalence for MM2+ was 69.0% (vs 54.0%) for RA women (vs non-RA women) and 69.0% (vs 54.0%) for RA men (vs non-RA men). For MM5+, the unadjusted prevalence was 29.3% (vs 18.5%) for RA women (vs non-RA women) and 31.1% (vs 18.2%) for RA men (vs non-RA men) (Table 2).
Table 1.
Cohort characteristics in RA and non-RA
| RA (N = 154 391) | Non-RA (N = 154 391) | |
|---|---|---|
| Sex | ||
| Female | 117 389 (76.0) | 117 389 (76.0) |
| Male | 37 002 (24.0) | 37 002 (24.0) |
| Age, mean (s.d.) | 59.6 (13.6) | 59.6 (13.6) |
| Age group | ||
| 18–39 | 13 247 (8.6) | 13 247 (8.6) |
| 40–64 | 80 585 (52.2) | 80 585 (52.2) |
| 65+ | 60 559 (39.2) | 60 559 (39.2) |
| Region | ||
| Midwest | 39 989 (25.9) | 39 989 (25.9) |
| Northeast | 17 393 (11.3) | 17 393 (11.3) |
| South | 77 288 (50.1) | 77 288 (50.1) |
| West | 19 721 (12.8) | 19 721 (12.8) |
| Insurance coverage (yrs), median (IQR) | 1.6 (1.3, 3.1) | 1.6 (1.3, 3.1) |
| Race/Ethnicity | ||
| Asian | 3664 (2.4) | 3664 (2.4) |
| Black | 16 361 (10.6) | 16 361 (10.6) |
| Hispanic | 16 320 (10.6) | 16 320 (10.6) |
| Other/Unknown | 3832 (2.5) | 3832 (2.5) |
| White | 114 214 (74.0) | 114 214 (74.0) |
| Index date yr, mean (s.d.) | 2014.52 (3.2) | 2014.52 (3.2) |
IQR: interquartile range.
Table 2.
Individual comorbidities in non-RA and RA by sex in ages 18–50 and 50+ years
| Ages 18–50 years |
Ages 51+ years |
|||||||
|---|---|---|---|---|---|---|---|---|
| Non-RA |
RA |
Non-RA |
RA |
|||||
| Comorbidity | Female (N = 29 910) | Male (N = 7220) | Female (N = 29 910) | Male (N = 7220) | Female (N = 87 479) | Male (N = 29 782) | Female (N = 87 479) | Male (N = 29 782) |
| Alcohol abuse | 99 (0.3) | 53 (0.7) | 93 (0.3) | 64 (0.9) | 369 (0.4) | 295 (1.0) | 279 (0.3) | 272 (0.9) |
| Anaemia | 848 (2.8) | 71 (1.0) | 2541 (8.5) | 307 (4.3) | 6889 (7.9) | 2070 (7.0) | 12 141 (13.9) | 3511 (11.8) |
| Anxiety | 2128 (7.1) | 306 (4.2) | 2994 (10.0) | 383 (5.3) | 6737 (7.7) | 1134 (3.8) | 7044 (8.1) | 1286 (4.3) |
| Asthma | 893 (3.0) | 152 (2.1) | 1779 (5.9) | 207 (2.9) | 3895 (4.5) | 681 (2.3) | 6336 (7.2) | 1007 (3.4) |
| Bipolar disorder | 341 (1.1) | 47 (0.7) | 590 (2.0) | 89 (1.2) | 937 (1.1) | 192 (0.6) | 1033 (1.2) | 224 (0.8) |
| Benign prostatic hyperplasia | N/A | 45 (0.6) | N/A | 54 (0.7) | N/A | 2613 (8.8) | 0 | 2836 (9.5) |
| Cancer | 564 (1.9) | 80 (1.1) | 538 (1.8) | 87 (1.2) | 6208 (7.1) | 2975 (10.0) | 5242 (6.0) | 2822 (9.5) |
| Cardiac arrhythmia | 253 (0.8)a | 70 (1.0)a | 585 (2.0)a | 135 (1.9)a | 5727 (6.5) | 2911 (9.8) | 6321 (7.2) | 3240 (10.9) |
| Cerebrovascular disease | 82 (0.3)a | 21 (0.3)a | 189 (0.6) | 23 (0.3) | 2934 (3.4) | 1173 (3.9) | 2887 (3.3)a | 1053 (3.5)a |
| Chronic back pain | 3258 (10.9) | 679 (9.4) | 7372 (24.6) | 1580 (21.9) | 15 297 (17.5) | 4219 (14.2) | 26 774 (30.6) | 7701 (25.9) |
| Congestive heart failure | 85 (0.3)a | 25 (0.3)a | 184 (0.6)a | 53 (0.7)a | 3581 (4.1) | 1511 (5.1) | 3977 (4.5) | 1741 (5.8) |
| COPD | 117 (0.4)a | 29 (0.4)a | 364 (1.2)a | 106 (1.5)a | 5448 (6.2) | 2000 (6.7) | 7712 (8.8) | 3259 (10.9) |
| Coronary artery disease | 123 (0.4) | 80 (1.1) | 275 (0.9) | 133 (1.8) | 6114 (7.0) | 4492 (15.1) | 7300 (8.3) | 5291 (17.8) |
| Dementia | S | S | S | S | 2086 (2.4) | 523 (1.8) | 1059 (1.2) | 228 (0.8) |
| Depression | 2628 (8.8) | 391 (5.4) | 3998 (13.4) | 478 (6.6) | 9069 (10.4) | 1594 (5.4) | 10 823 (12.4) | 1896 (6.4) |
| Diabetes | 1153 (3.9) | 498 (6.9) | 2057 (6.9) | 575 (8.0) | 17 903 (20.5) | 6993 (23.5) | 17 918 (20.5) | 6886 (23.1) |
| Diverticular disease | 51 (0.2) | 24 (0.3) | 89 (0.3) | 39 (0.5) | 1210 (1.4) | 333 (1.1) | 1565 (1.8) | 472 (1.6) |
| Drug abuse | 144 (0.5) | 58 (0.8) | 405 (1.4)a | 102 (1.4)a | 518 (0.6) | 212 (0.7) | 1004 (1.1)a | 333 (1.1)a |
| Fibromyalgia | 619 (2.1) | 91 (1.3) | 3572 (11.9) | 367 (5.1) | 1925 (2.2) | 269 (0.9) | 7087 (8.1) | 1087 (3.6) |
| GERD | 839 (2.8)a | 226 (3.1)a | 2179 (7.3) | 406 (5.6) | 8288 (9.5) | 2110 (7.1) | 12 438 (14.2) | 3034 (10.2) |
| Gout | 14 (0.0) | 67 (0.9) | 103 (0.3) | 294 (4.1) | 862 (1.0) | 798 (2.7) | 1790 (2.0) | 1925 (6.5) |
| Headaches | 1547 (5.2) | 107 (1.5) | 2776 (9.3) | 215 (3.0) | 3032 (3.5) | 393 (1.3) | 4014 (4.6) | 542 (1.8) |
| Hearing loss | 78 (0.3)a | 26 (0.4)a | 116 (0.4)a | 26 (0.4)a | 912 (1.0)a | 304 (1.0)a | 1026 (1.2) | 413 (1.4) |
| Hyperlipidemia | 2021 (6.8) | 928 (12.9) | 3008 (10.1) | 1089 (15.1) | 32 743 (37.4) | 12 367 (41.5) | 32 923 (37.6) | 12 800 (43.0) |
| Hypertension | 3040 (10.2) | 1104 (15.3) | 4797 (16.0) | 1529 (21.2) | 40 881 (46.7) | 14 681 (49.3) | 45 848 (52.4) | 16 105 (54.1) |
| Hyperthyroidism | 1901 (6.4) | 144 (2.0) | 3531 (11.8) | 312 (4.3) | 13 204 (15.1) | 1666 (5.6) | 16 515 (18.9) | 2221 (7.5) |
| Inflammatory skin disease | 362 (1.2) | 52 (0.7) | 1765 (5.9) | 688 (9.5) | 1466 (1.7)a | 458 (1.5)a | 4868 (5.6) | 1937 (6.5) |
| Interstitial lung disease | S | S | 1089 (3.6)a | 261 (3.6)a | 276 (0.3)a | 116 (0.4)a | 4060 (4.6) | 1605 (5.4) |
| Liver disease | 120 (0.4) | 44 (0.6) | 349 (1.2) | 114 (1.6) | 979 (1.1) | 432 (1.5) | 1390 (1.6)a | 519 (1.7)a |
| Neuropathy | 401 (1.3) | 67 (0.9) | 1650 (5.5) | 319 (4.4) | 3244 (3.7)a | 1037 (3.5)a | 6865 (7.8) | 2129 (7.1) |
| Non-inflammatory gynecologic | 2556 (8.5) | N/A | 2720 (9.1) | N/A | 4145 (4.7) | 0 (0.0) | 3900 (4.5) | 0 |
| Obesity | 1219 (4.1) | 234 (3.2) | 2162 (7.2) | 358 (5.0) | 4893 (5.6) | 1330 (4.5) | 6248 (7.1) | 1584 (5.3) |
| Osteoarthritis | 406 (1.4)a | 92 (1.3)a | 4047 (13.5) | 1069 (14.8) | 10 290 (11.8) | 2450 (8.2) | 31 962 (36.5) | 9528 (32.0) |
| Osteoporosis | S | S | 559 (1.9) | 97 (1.3) | 4051 (4.6) | 152 (0.5) | 12 482 (14.3) | 1222 (4.1) |
| Parkinson's disease | S | S | S | S | 456 (0.5) | 297 (1.0) | 356 (0.4) | 173 (0.6) |
| Pulmonary circulation disorder | S | S | 186 (0.6)a | 38 (0.5)a | 963 (1.1) | 274 (0.9) | 1550 (1.8) | 455 (1.5) |
| Peptic ulcer disease | S | S | 81 (0.3)a | 19 (0.3)a | 279 (0.3)a | 88 (0.3)a | 517 (0.6)a | 162 (0.5)a |
| Peripheral vascular disease | 58 (0.2) | 26 (0.4) | 163 (0.5)a | 47 (0.7)a | 3842 (4.4) | 1728 (5.8) | 4371 (5.0) | 1859 (6.2) |
| PTSD | 161 (0.5) | 17 (0.2) | 324 (1.1) | 21 (0.3) | 325 (0.4) | 57 (0.2) | 355 (0.4) | 64 (0.2) |
| Renal disease | 81 (0.3) | 51 (0.7) | 285 (1.0)a | 73 (1.0)a | 4583 (5.2) | 1901 (6.4) | 5646 (6.5) | 2187 (7.3) |
| Severe vision reduction | 293 (1.0) | 91 (1.3) | 947 (3.2) | 177 (2.5) | 11 159 (12.8) | 3288 (11.0) | 14 294 (16.3) | 4001 (13.4) |
| Sleep disorder | 672 (2.2)a | 157 (2.2)a | 1379 (4.6) | 270 (3.7) | 3114 (3.6) | 886 (3.0) | 4197 (4.8) | 1212 (4.1) |
| Urinary incontinence | S | S | S | S | 1402 (1.6) | 235 (0.8) | 1832 (2.1) | 244 (0.8) |
| Valvular heart disease | 97 (0.3)a | 28 (0.4)a | 228 (0.8)a | 43 (0.6)a | 2380 (2.7) | 1034 (3.5) | 2933 (3.4) | 1110 (3.7) |
Not significant at P < 0.05.
COPD: chronic obstructive pulmonary disease; GERD: gastroesophageal reflux disease; PTSD: post-traumatic stress disorder; S: supressed due to small cell size of data.
Cohort characteristics and multimorbidity
Adjusted associations for the relative risk between cohort characteristics and multimorbidity are presented in Fig. 1. Among patients with RA, women (vs men), along with people who were older (vs younger) were more likely to have MM2+ and MM5+. The same predictors were associated with an increased likelihood of MM2+ and MM5+ for non-RA comparators except for sex which was associated with MM2+ but not MM5+. Regarding adjustors, for both patients with RA and comparators, Black or Hispanic (vs White) were associated with increased likelihood of MM2+ and MM5+ and Asian (vs White) race was inversely associated with MM2+ and MM5+, whereas having a more recent index date year was positively associated with MM2+ and MM5+. West census region (compared with the Midwest) and a greater length of baseline insurance coverage were inversely associated with MM2+ and MM5+ in non-RA comparators. Reported differences were statistically significant at P < 0.001.
Figure 1.
Association between cohort characteristics and multimorbidity. Panel A: 2 morbidities. Association between cohort characteristics and multimorbidity. Panel B: 5 morbidities
RA and multimorbidity by sex and age interaction
Fig. 2 presents the adjusted absolute risk of MM2+ and MM5+ and 95% CIs in men and women with RA and their non-RA comparators aged 18–50 and 51+ years. For people aged 18–50 years, compared with RA men, RA women had 7.5 and 4.4 percentage point increases for MM2+ and MM5+, respectively (P < 0.001). In contrast, non-RA women, compared with non-RA men, had 3.2 and 0.9 percentage point increases for MM2+ and MM5+, respectively (P < 0.001). An interaction demonstrated that the difference between RA women (vs RA men) compared with the difference between non-RA women (vs non-RA men) was significant for MM2+ (P < 0.001) and MM5+ (P < 0.05).
Figure 2.
Adjusted absolute risk of multimorbidity in RA (vs non-RA) by sex and age group. Statistical significance of the absolute risk of MM2+ between males and females was tested with Wald tests based on adjusted logistic regression models for four groups: (i) Non-RA (ages 18–50); (ii) RA (ages 18–50); (iii) non-RA (Ages 51+); (iv) RA (ages 51+). Bars represent prevalence for MM2+ (Panes A) and MM5+ (Panel B). Capped lines are 95% Cis
For people aged 51+ years, differences in multimorbidity between RA and non-RA men and women were significant but less pronounced. RA women had an increase of 2.1 percentage points for MM2+ and 2.5 percentage points for MM5+ when compared with RA men (P < 0.05). Compared with non-RA men, non-RA women had a 1.2 percentage point and 0.3 percentage point increase for MM2+ and MM5+, respectively (P < 0.05).
Fig. 3 further demonstrates that the magnitude of the differences in MM2+ and MM5+ by sex in patients with RA compared with their non-RA comparators were more pronounced for younger ages and waned as age increased.
Figure 3.
Adjusted absolute risk of multimorbidity in non-RA and RA by sex and age
Comorbidities by RA, sex and age
Table 2 displays the overall prevalence for each comorbidity stratified by sex in patients with RA and their non-RA comparators aged 18–50 and 51+ years. For people aged 18–50 years, a higher prevalence of alcohol abuse, coronary artery disease, diabetes mellitus, diverticular disease, gout, hyperlipidaemia, hypertension, inflammatory skin disease, liver disease, osteoarthritis and peripheral vascular disease was observed for RA men. In contrast, RA women showed a higher prevalence of anaemia, anxiety, asthma, bipolar disorder, cancer, cerebrovascular disease, chronic back pain, depression, diverticular disease, fibromyalgia, gastro-esophageal reflux disease (GERD), headaches, hypothyroidism, neuropathy, obesity, osteoporosis, post-traumatic stress disorder (PTSD), severe vision reduction, sleep disorder and urinary incontinence (P < 0.05 for all). Men and women without RA had a similar prevalence of many comorbidities with the exception that women had more inflammatory skin disease and males had more diverticular disease, drug abuse, renal disease and severe vision reduction (P < 0.05 for all). There were also no significant differences by sex for cerebrovascular disease, GERD, osteoarthritis, osteoporosis and sleep disorder in non-RA comparators.
RA men and women aged 51+ years tended to follow a similar pattern as those aged 18–50 years but tended to have higher prevalences for many comorbidities. The main differences were that RA men had more cancer, cardiac arrhythmia, cerebrovascular disease, heart failure, chronic obstructive pulmonary disease (COPD), hearing loss, interstitial lung disease, Parkinson’s disease, renal disease and valvular heart disease, whereas RA women had more dementia, osteoarthritis and pulmonary circulation disorder (P < 0.01 for all). Men and women without RA also tended to show higher prevalences, as well as follow a similar pattern, as their respective younger age group. The exceptions were that non-RA men had more cancer, cerebrovascular, heart failure, COPD, drug abuse, Parkinson’s disease and valvular heart disease compared with their younger counterparts. On the other hand, non-RA women had a higher prevalence of dementia, diverticular disease, osteoarthritis, osteoporosis, pulmonary circulation disorder and sleep disorders (P < 0.01 for all).
Discussion
In this study, we found differences in multimorbidity based on sex in the RA population, wherein women with RA were more likely to have multimorbidity compared with men with RA. The differences in multimorbidity between women and men were larger in the RA population compared with those without RA and began to wane for both populations as the number of comorbidities compounded with increasing age. We also found variation in the types of comorbidities between women and men with RA aged 18–50 and 51+ years.
Our primary finding was that sex was an independent and significant risk factor for multimorbidity in the RA population, with the strongest effect observed in people aged 18–50 years. The increased prevalence of multimorbidity in women with RA is consistent with the multimorbidity differences observed in the general population [15]. Notably, sex may serve as a direct determent, as well as a context-specific proxy for underlying social, behavioural and environmental factors, of multimorbidity [22]. Concordantly, women with greater access barriers (e.g. elderly, low-SES) have shown an increased risk of multimorbidity [15]. Women, compared with men, have also demonstrated a greater propensity for healthcare seeking (e.g. primary care, outpatient visits) [23]. These differences may contribute to chronic conditions being diagnosed earlier in women and later and at more severe stages in men. Our study adds to the literature by confirming such disparities are even greater in the RA population.
Other notable findings were the differences in the prevalence of certain comorbidities between men and women aged 18–50 and 51+ years in the RA population. First, men with RA, compared with women with RA, exhibited a higher prevalence of cardiovascular-related conditions including coronary artery disease, hyperlipidaemia and hypertension for those aged 18–50 years with the addition of cardiac arrhythmia, heart failure and valvular heart disease for those aged 51+ years. While these elevated risks are also observed for men in the general population, for people with RA, chronic inflammation increases the progression of atherosclerosis, a primary catalyst for cardiovascular conditions [9]. This process usually occurs earlier in men than women [10, 24] and may contribute to the increased prevalence of cardiovascular-related conditions found in men in the youngest age group in our study. Next, for both age groups, women with RA tended to show an increased prevalence of psychological (e.g. depression, anxiety, bipolar, PTSD), neurological (e.g. headaches, sleep disorders, vision reduction, neuropathy) and general musculoskeletal (e.g. chronic back pain, fibromyalgia) conditions compared with men with RA. Aligned with these findings, in the RA population, women compared with men have shown greater functional impairment and anxiety symptoms, along with an increased incidence of depression and osteoporosis [25]. The latter two symptoms also contributed to a lower quality of life in women compared with men [25]. Interestingly, in our study, in people with RA aged 18–50, osteoarthritis was more common for men (vs women) and the prevalence of osteoporosis was low (<2%) for both men and women; however, in people with RA aged 51+, the prevalence for both conditions was disproportionately high for women with RA. Indeed, the onset of menopause is a known to escalate disease activity for rheumatic-associated conditions such as osteoporosis [8].
Other comorbidities that were elevated in the RA population and exhibited distinctions by sex and age were kidney and digestive diseases. For instance, men demonstrated a higher prevalence of liver disease in the youngest age group and kidney disease in the oldest age group, whereas women exhibited a higher prevalence of GERD in both age groups and urinary incontinence in the oldest age group. For people with RA, these conditions may be exacerbated by inflammation from RA as well as toxicity from medications (e.g. methotrexate or nonsteroidal anti-inflammatory drugs [26]). As mentioned above, delayed healthcare-seeking behaviours in men may also contribute to forgoing care for conditions perceived to be less severe (e.g. GERD, urinary incontinence) and, in turn, receiving a diagnosis at a later stage (e.g. end-stage kidney disease or liver disease). Consistent with this explanation, our study found that women with RA demonstrated a higher prevalence of cancer in the youngest age group and men with RA showed a higher prevalence of cancer in the oldest age group. Further research is needed to identify how other health behaviours (e.g. substance use), biological processes (e.g. hormonal activity), and social/environmental factors (e.g. income inequities, financial toxicity) contribute to disease disparities and burden in women and men with RA.
There are important implications from this study. First, multimorbidity may exacerbate existing treatment inequities (e.g. poorer adherence [11] and response [14]) for women with RA and be underrepresented in men with RA who may have a lower propensity to seek care. Routine screening in both men and women with RA may facilitate early identification of comorbidities that may impact treatment decisions. The variation in comorbidities by sex and age also may underscore a need for individualized care plans and care coordination that target disease and symptom management for women and men with RA across the lifespan. Finally, research and policy efforts should strive to reduce sex-based inequities in the treatment and outcomes for RA. This necessitates increased vigilance to screening for comorbidities and multimorbidity in the RA population earlier on to optimize prevention and timely management, along with ensuring that sex and/or gender are incorporated into the study design and analysis [27] (e.g. requiring that clinical trials report treatment effects by sex and/or gender).
This study has several limitations. The first is the potential for misclassification of RA and comorbidities given the use of code-based definitions with the latter combined requiring only a one-year period to identify diagnoses. However, the algorithms used demonstrated a PPV of 90% and 95%, respectively. Further, requiring more enrolment may induce selection bias [17] and underrepresent people with RA who may experience greater barriers to accessing care. Second, we were not able to observe prescriptions and health services not submitted to the insurance plan. Third, patients with RA may have more encounters with the health care system and, in turn, may be more likely to receive a comorbidity diagnosis. Fourth, we could not directly ascertain gender. When possible, studies should report effects by sex and gender identity. Fifth, we cannot disentangle the role of specific health behaviours (e.g. smoking) or access to enabling services (e.g. employment). Future research should evaluate the mediating and moderating role of specific risk factors with respect to the development of multimorbidity. Lastly, these results may not be generalizable beyond people with commercial or MA health insurance. Future research is needed to identify whether similar disparities in multimorbidity are observed in vulnerable populations that may be disproportionately impacted by RA and multimorbidity inequities (e.g. patients with Medicaid insurance). Despite these limitations, to our knowledge, this is the first study that has evaluated how the relationship between RA and multimorbidity, along with key comorbidities, vary by sex across the lifespan using a large and representative sample.
Conclusion
Our findings underscore that multimorbidity disproportionately impacts women with RA, but differences waned as the number of comorbidities increased with age. There was also variation in comorbidities across sex and age groups in the RA population. These findings highlight a need for individualized treatment and supporting research and policies that account for the variation in care needs in people with RA. It is essential that research, clinical and policy agendas for rheumatic diseases acknowledge and support the differences in care needs by sex and gender across the lifespan.
Supplementary Material
Contributor Information
Maria A Stevens, Mayo Clinic Robert D. and Patricia E. Kern Center for the Science of Health Care Delivery, Rochester, MN, USA; OptumLabs, Eden Prairie, MN, USA; Department of Health Policy and Management, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA.
Hayley J Dykhoff, Mayo Clinic Robert D. and Patricia E. Kern Center for the Science of Health Care Delivery, Rochester, MN, USA.
Vanessa L Kronzer, Division of Rheumatology, Department of Internal Medicine, Mayo Clinic, Rochester, MN, USA.
Elena Myasoedova, Division of Rheumatology, Department of Internal Medicine, Mayo Clinic, Rochester, MN, USA; Department of Quantitative Health Sciences, Mayo Clinic, Rochester, MN, USA.
John M Davis, III, Division of Rheumatology, Department of Internal Medicine, Mayo Clinic, Rochester, MN, USA.
Alí Duarte-García, Mayo Clinic Robert D. and Patricia E. Kern Center for the Science of Health Care Delivery, Rochester, MN, USA; Division of Rheumatology, Department of Internal Medicine, Mayo Clinic, Rochester, MN, USA.
Cynthia S Crowson, Division of Rheumatology, Department of Internal Medicine, Mayo Clinic, Rochester, MN, USA; Department of Quantitative Health Sciences, Mayo Clinic, Rochester, MN, USA.
Supplementary material
Supplementary material is available at Rheumatology online.
Data availability
Maintaining OptumLabs data as de-identified is a key part of OptumLabs’ mission and structure. Users are therefore not permitted to disclose individual level data from OLDW and are only permitted to use summary level data for purposes of publication or presentation, as it pertains to an approved project or study. To maintain the de-identified nature of the OLDW, all summary tables or summary charts shared with recipients outside of the Project Team and OptumLabs’ employees must adhere to CMS cell size suppression policy.
Funding
This work was supported by grants from the National Institutes of Health, NIAMS (R01 AR46849) and NIA (R01 AG068192). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Disclosure statement: The authors declare no conflicts of interest. In the past three years, A.D.-G. has received unrelated grant funding from the Centers for Disease Control and Prevention, the Rheumatology Research Foundation Career Development Award, the Robert D. and Patricia E. Kern Center for the Science of Health Care Delivery, and Mayo Clinic. E.M.’s research effort is funded by the NIA (R01 AG068192, K24 AG078179-01).
References
- 1. Wasserman AM. Diagnosis and management of rheumatoid arthritis. Am Fam Physician 2011;84:1245–52. [PubMed] [Google Scholar]
- 2. Crowson CS, Gunderson TM, Dykhoff HJ et al. Comprehensive assessment of multimorbidity burden in a population-based cohort of patients with rheumatoid arthritis. RMD Open 2022;8:e002022. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Gunderson TM, Myasoedova E, Davis JM 3rd, Crowson CS. Multimorbidity burden in rheumatoid arthritis: a population-based cohort study. J Rheumatol 2021;48:1648–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Canning J, Siebert S, Jani BD et al. Examining the relationship between rheumatoid arthritis, multimorbidity, and adverse health-related outcomes: a systematic review. Arthritis Care Res (Hoboken) 2022;74:1500–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Hughes GC, Choubey D. Modulation of autoimmune rheumatic diseases by oestrogen and progesterone. Nat Rev Rheumatol 2014;10:740–51. [DOI] [PubMed] [Google Scholar]
- 6. Libert C, Dejager L, Pinheiro I. The X chromosome in immune functions: when a chromosome makes the difference. Nat Rev Immunol 2010;10:594–604. [DOI] [PubMed] [Google Scholar]
- 7. Bellott DW, Hughes JF, Skaletsky H et al. Mammalian Y chromosomes retain widely expressed dosage-sensitive regulators. Nature 2014;508:494–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Talsania M, Scofield RH. Menopause and rheumatic disease. Rheum Dis Clin North Am 2017;43:287–302. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Kaplan MJ. Cardiovascular complications of rheumatoid arthritis: assessment, prevention, and treatment. Rheum Dis Clin North Am 2010;36:405–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Targońska-Stępniak B, Biskup M, Biskup W, Majdan M. Gender differences in cardiovascular risk profile in rheumatoid arthritis patients with low disease activity. Biomed Res Int 2019;2019:3265847. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Klein SL, Morgan R. The impact of sex and gender on immunotherapy outcomes. Biol Sex Differ 2020;11:24–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Souto A, Maneiro JR, Gómez-Reino JJ. Rate of discontinuation and drug survival of biologic therapies in rheumatoid arthritis: a systematic review and meta-analysis of drug registries and health care databases. Rheumatology (Oxford) 2016;55:523–34. [DOI] [PubMed] [Google Scholar]
- 13. Forslind K, Hafström I, Ahlmén M, Svensson B; BARFOT Study Group. Sex: a major predictor of remission in early rheumatoid arthritis? Ann Rheum Dis 2007;66:46–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Jawaheer D, Olsen J, Hetland ML. Sex differences in response to anti-tumor necrosis factor therapy in early and established rheumatoid arthritis — results from the DANBIO registry. J Rheumatol 2012;39:46–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Violan C, Foguet-Boreu Q, Flores-Mateo G et al. Prevalence, determinants and patterns of multimorbidity in primary care: a systematic review of observational studies. PLoS ONE 2014;9:e102149. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Radner H, Yoshida K, Frits M et al. The impact of multimorbidity status on treatment response in rheumatoid arthritis patients initiating disease-modifying anti-rheumatic drugs. Rheumatology (Oxford) 2015;54:2076–84. [DOI] [PubMed] [Google Scholar]
- 17. OptumLabs. OptumLabs and OptumLabs Data Warehouse (OLDW) descriptions and citation. Eden Prairie, MN: n.p., 2019. PDF. Reproduced with permission from OptumLabs.
- 18. Wallace PJ, Shah ND, Dennen T, Bleicher PA, Crown WH. Optum Labs: building a novel node in the learning health care system. Health Aff (Millwood) 2014;33:1187–94. [DOI] [PubMed] [Google Scholar]
- 19. Chung CP, Rohan P, Krishnaswami S, McPheeters ML. A systematic review of validated methods for identifying patients with rheumatoid arthritis using administrative or claims data. Vaccine 2013;31(Suppl 10):K41–61. [DOI] [PubMed] [Google Scholar]
- 20. MacLean CH, Traina SB, Liu H et al. Positive predictive value of an administrative data-based algorithm for the identification of patients with rheumatoid arthritis (RA). Arthritis Rheum 2001;44:324. [Google Scholar]
- 21. England BR, Roul P, Yang Y et al. Burden and trajectory of multimorbidity in rheumatoid arthritis: a matched cohort study from 2006 to 2015. Ann Rheum Dis 2021;80:286–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. The Academy of Medical Sciences. Multimorbidity: a priority for global health research. 2018. https://acmedsci.ac.uk/file-download/82222577 (1 July 2023, date last accessed).
- 23. Glynn LG, Valderas JM, Healy P et al. The prevalence of multimorbidity in primary care and its effect on health care utilization and cost. Fam Pract 2011;28:516–23. [DOI] [PubMed] [Google Scholar]
- 24. Fairweather D. Sex differences in inflammation during atherosclerosis. Clin Med Insights Cardiol 2015;8:49–59. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Aurrecoechea E, llorcadiaz J, Diezlizuain ML, Mcgwin G, Calvo-alen J. Impact of gender in the quality of life of patients with rheumatoid arthritis. J Arthritis 2015;4:160. [Google Scholar]
- 26. Svanström H, Lund M, Melbye M, Pasternak B. Concomitant use of low-dose methotrexate and NSAIDs and the risk of serious adverse events among patients with rheumatoid arthritis. Pharmacoepidemiol Drug Saf 2018;27:885–93. [DOI] [PubMed] [Google Scholar]
- 27. Getting serious about sex and gender. Lancet Rheumatol 2021;3:E313. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Maintaining OptumLabs data as de-identified is a key part of OptumLabs’ mission and structure. Users are therefore not permitted to disclose individual level data from OLDW and are only permitted to use summary level data for purposes of publication or presentation, as it pertains to an approved project or study. To maintain the de-identified nature of the OLDW, all summary tables or summary charts shared with recipients outside of the Project Team and OptumLabs’ employees must adhere to CMS cell size suppression policy.