To describe long-term physical functioning and its association with somatic comorbidity and comorbid depression in patients with established rheumatoid arthritis (RA).
To describe long-term physical functioning and its association with somatic comorbidity and comorbid depression in patients with established rheumatoid arthritis (RA).
Longitudinal data over a period of 11 years were collected from 882 patients with RA at study inclusion. Patient-reported outcomes were collected in 1997, 1998, 1999, 2002, and 2008. Physical functioning was measured with the Health Assessment Questionnaire and the physical component summary score of the Short Form 36 health survey. Somatic comorbidity was measured by a questionnaire including 12 chronic diseases. Comorbid depression was measured with the Center for Epidemiologic Studies Depression Scale. We distinguished 4 groups of patients based on comorbidity at baseline.
Seventy-two percent of the patients at baseline were women. The mean ± SD age was 59.3 ± 14.8 years and the median disease duration was 5.0 years (interquartile range 2.0–14.0 years). For the total group of patients with RA, physical functioning improved over time. Patients with somatic comorbidity, comorbid depression, or both demonstrated worse physical functioning than patients without comorbidity at all data collection points. Both groups with comorbid depression had the lowest scores. Only patients with both somatic comorbidity and comorbid depression showed significantly less improvement in physical functioning over time.
Both somatic comorbidity and comorbid depression were negatively associated with physical functioning during an 11-year followup period. Furthermore, their combination seems to be especially detrimental to physical functioning over time. These results emphasize the need to take somatic comorbidity and comorbid depression into account in the screening and treatment of patients with RA.
Comorbidity, defined as any additional, coexistent condition in a patient with a particular index disease (), is highly prevalent in patients with rheumatoid arthritis (RA). The prevalence of comorbidity in patients with RA is higher than in the Dutch general population ([2-5]). Estimates of the percentages of at least one comorbid condition vary substantially, ranging from 27–81% due to differences in the definition of comorbidity and selection of the study population ([4, 6]). The average RA patient has approximately 1.6 comorbid conditions ([7, 8]).
Comorbidities in patients with RA are of both a somatic and psychological nature. Common somatic comorbidities are cardiovascular disease, hypertension, chronic pulmonary disease, gastrointestinal disease, osteoporosis, and infection ([3, 5, 9, 10]). With regard to psychological comorbidity, depression is common and occurs in 13–20% of patients with RA, which is 2 or 3 times more common than in the general population ([11-14]).
There is increasing evidence that comorbidity plays an important role in determining RA-related outcomes. Evidence points to poorer outcomes after comorbidity in patients with RA than in the general population (). Outcomes that are associated with comorbid conditions are disability, quality of life, health care costs, and mortality ([7, 15]). Both somatic comorbidity and comorbid depression are associated with negative outcomes, but different comorbid conditions have a different impact on health outcomes. For example, cardiac and pulmonary diseases are associated with greater rates of hospitalization and mortality, whereas depression is more strongly related to increased levels of disability ().
Physical functioning, a major patient-reported outcome, is rated as the most important outcome in patients with RA. Rupp and colleagues () have shown that a somatic comorbid condition negatively influences physical functioning. Furthermore, they have also shown that depression is of major importance in predicting physical functioning (). Radner and colleagues found that there was a negative influence of somatic comorbidity on all domains of physical functioning, independent of the level of disease activity ([17, 18]). One study that investigated the influence of comorbidity on change in physical functioning demonstrated that comorbidity at baseline contributed negatively to change over time in physical functioning (). They also found that nontreatment factors, such as age and comorbidity, had a greater effect on the progression in physical functioning than did treatment factors.
Most of the aforementioned studies investigated the influence of comorbidity over a 1-year period. Only one study used a 5-year followup period. However, little is known about the impact of comorbidity in the longer term, which is important because it provides clinicians with long-term information about the possible course of physical functioning for individual patients. If the presence of comorbidity was found to influence physical functioning on a longer-term basis, it would be of even greater importance for clinicians to adjust their treatment to the comorbidity accordingly. Furthermore, the aforementioned studies that investigated the effects of comorbidity on physical functioning did not compare the influence of the co-occurrence of somatic comorbidity and comorbid depression. Comparing somatic comorbidity and comorbid depression is important because of their differential impact on patient-reported outcomes (). Therefore, the aim of this study was to describe long-term physical functioning in patients with RA and its association with somatic comorbidity and comorbid depression. Research questions consisted of: 1) How does long-term physical functioning develop in an RA population? 2) How are somatic comorbidity and comorbid depression associated with physical functioning? And 3) how are somatic comorbidity and comorbid depression associated with change in physical functioning?
In 1997, a longitudinal study was started on comorbidity and health outcomes in patients with RA. At commencement, 1,251 patients were randomly selected from an outpatient clinic for rheumatology and rehabilitation in Amsterdam or from an affiliated outpatient clinic. For inclusion in the study, patients had to fulfill the following eligibility criteria: having a diagnosis of RA according to the American College of Rheumatology criteria for RA (), being age ≥16 years, having an adequate knowledge of the Dutch language, and having had at least one visit to a rheumatologist in the previous 2 years.
Data were collected in 1997, 1998, 1999, 2002, and 2008 by means of self-administered questionnaires. The questionnaires comprised questions about sociodemographic characteristics (age, sex, marital status, educational level, and employment status), clinical characteristics (including comorbidity), health status (including physical functioning), and the utilization of health care services. Information on disease duration was retrieved from the patients' medical records. In addition, we established whether participants had died during the period 1996–2010 from the mortality register of Statistics Netherlands.
Physical functioning was measured with the validated Dutch version of the Health Assessment Questionnaire (HAQ) and the physical functioning scales of the Dutch version of the RAND-36 (). We used the physical functioning scales of the RAND-36 in addition to the HAQ to provide a more generic overview of physical functioning. With respect to the HAQ, the category score was raised when aids or devices were indicated by the patient. The RAND-36 is almost the same as the 36-item Short Form (SF-36) health survey (). The physical component summary score of the SF-36 was calculated according to the manual for SF-36 health summary scales () using Dutch population means, SDs, and factor score coefficients ().
Somatic comorbidity as well as comorbid depression were assessed at baseline. Somatic comorbidity was measured with a self-report list adapted from the Health Interview Survey of Statistics Netherlands (). The Health Interview Survey covers 12 groups of chronic conditions, specifically lung diseases, cardiovascular diseases, diabetes mellitus, gastrointestinal diseases, cancer, kidney diseases, chronic infections, gallbladder and liver diseases, chronic back symptoms, skin diseases, thyroid gland diseases, and neurologic diseases. These chronic diseases are relatively most prevalent in The Netherlands. Respondents were asked to indicate whether they had had any of these conditions in the previous 12 months. Respondents indicating the presence of one or more conditions were classified as having somatic comorbidity.
Comorbid depression was assessed with the Center for Epidemiologic Studies Depression Scale (CES-D) (). The CES-D is a short, self-administered scale designed to measure depressive symptoms in the general population. The CES-D consists of 20 items and has a range from 0–60, with higher scores indicating more depressive symptoms. Scores ≥16 suggest the presence of depression.
The control variables included age, sex, socioeconomic status (SES), marital status, and disease duration; these characteristics are prognostic factors regarding health outcomes ([28-31]). SES was indicated by educational level. We divided SES into 3 categories: low SES, indicating patients with no education or education at the primary school level; medium SES, indicating patients with education at the secondary school level; and high SES, indicating patients with a college or university level education. Marital status was dichotomized into married and single.
To determine changes in physical functioning over time, we performed a longitudinal analysis analyzing how baseline comorbidity predicts long-term physical functioning. Four groups based on the absence or presence of comorbidity at baseline were distinguished: 1) patients without comorbidity, 2) patients with somatic comorbidity only, 3) patients with comorbid depression only, and 4) patients with both somatic comorbidity and comorbid depression.
Analyses were carried out with the use of a linear, mixed-effect, random intercept model with serial correlation of the residuals (). With this model, we controlled for intersubject correlation, taking into account that this correlation decreases with increasing time and for differences in duration between measurement moments. The outcome variable was physical functioning. The predictors were the comorbidity groups and their interaction with time. Time was entered as a continuous variable. We used 2 models: one model without and one model with the interaction between time and comorbidity groups. Separate analyses were performed for physical functioning measured with the HAQ and the SF-36. All models contained age, sex, SES, marital status, and disease duration in order to control for possible confounding by these factors. All analyses were carried out using R, package lme4 (). Results were considered statistically significant when P values were less than 0.05.
A flow chart summarizing the followup process of the 1,251 patients selected in 1997 is shown in Figure 1. Of the eligible patients, 882 (76%) returned the questionnaire in 1997. Of these patients, 755 (87% of the eligible patients) returned the questionnaire in 1998, 683 (81% of the eligible patients) returned the questionnaire in 1999, 529 (71% of the eligible patients) returned the questionnaire in 2002, and finally, 370 (62% of the eligible patients) returned the questionnaire in 2008.
Patient characteristics are shown in Table 1. Patients with comorbidity were older, were more often women, and had a lower SES and a longer disease duration when compared with patients without comorbidity.
|Total group (n = 882)||Comorbidity||Participants with complete followup (n = 367)||Participants with incomplete followup (n = 515)|
|None (n = 281)||Somatic (n = 328)a||Depression (n = 82)b||Somatic and depression (n = 163)c|
|Sex, no. (%)|
|Male||248 (28.1)||96 (34)||102 (31.1)||15 (18.3)||32 (19.6)||101 (27.5)||147 (28.5)|
|Female||634 (71.9)||185 (65.6)||226 (68.9)||67 (81.7)||131 (80.4)||266 (72.5)||368 (71.5)|
|Age, mean ± SD years||59.3 ± 14.8||56.7 ± 14.9||60.2 ± 13.9||59.2 ± 13.6||60.6 ± 15.2||53 ± 11.9||63 ± 15.3|
|Socioeconomic status, no. (%)|
|Low||220 (24.9)||50 (17.8)||76 (23.2)||21 (25.6)||60 (36.8)||57 (15.5)||163 (31.7)|
|Medium||526 (59.6)||179 (63.7)||206 (62.8)||49 (59.8)||85 (52.1)||235 (64.0)||291 (56.5)|
|High||123 (13.9)||51 (18.1)||43 (13.1)||11 (13.4)||17 (10.4)||73 (19.9)||50 (9.7)|
|Missing||13 (1.5)||1 (0.4)||3 (0.9)||1 (1.2)||1 (0.6)||2 (0.5)||11 (2.1)|
|Marital status, no. (%)|
|Single||315 (35.7)||84 (29.9)||111 (33.8)||38 (46.3)||70 (42.9)||92 (25.1)||223 (43.3)|
|Cohabitating||559 (63.4)||195 (69.4)||217 (66.2)||42 (51.2)||93 (57.1)||275 (74.9)||284 (55.1)|
|Missing||8 (0.9)||2 (0.7)||0 (0)||2 (2.4)||0 (0)||0 (0)||8 (1.6)|
|Disease duration, median (IQR) years||5 (2–14)||5 (2–13)||5 (2–14)||4 (1–14)||7 (2–15)||6 (2–13)||4 (1–15)|
|Disease activity (DAS28), mean ± SD||3.6 ± 1.3||3.4 ± 1.3||3.6 ± 1.3||3.8 ± 1.27||4.1 ± 1.39||3.5 ± 1.4||3.7 ± 1.35|
Table 2 provides the mean ± SD scores on the HAQ and the SF-36 at baseline and followup for all respondents. The HAQ scores showed a small improvement in physical functioning between baseline and 11-year followup. The SF-36 scores also improved between baseline and 11-year followup (Table 2).
|1997 (n = 882)||1998 (n = 755)||1999 (n = 683)||2002 (n = 529)||2008 (n = 370)|
|Total group||1.14 ± 0.80||1.18 ± 0.81||1.22 ± 0.80||1.11 ± 0.75||1.11 ± 0.75|
|None||0.92 ± 0.73||0.94 ± 0.74||1.02 ± 0.76||0.95 ± 0.71||1.01 ± 0.73|
|Somatic||1.07 ± 0.77||1.15 ± 0.79||1.16 ± 0.76||1.02 ± 0.68||1.13 ± 0.73|
|Depression||1.49 ± 0.84||1.39 ± 0.84||1.47 ± 0.82||1.39 ± 0.74||1.32 ± 0.69|
|Somatic and depression||1.47 ± 0.79||1.55 ± 0.81||1.59 ± 0.79||1.51 ± 0.81||1.70 ± 0.81|
|Total group||36.1 ± 10.9||37.2 ± 11.1||36.4 ± 10.8||36.6 ± 11.0||39.0 ± 11.3|
|None||38.7 ± 10.4||40.4 ± 10.4||39.2 ± 10.8||38.4 ± 10.3||41.5 ± 10.4|
|Somatic||35.5 ± 11.1||36.4 ± 11.3||35.3 ± 10.8||36.7 ± 10.8||38.0 ± 10.9|
|Depression||33.5 ± 10.3||34.9 ± 10.2||34.0 ± 10.0||34.7 ± 11.0||38.3 ± 11.6|
|Somatic and depression||32.3 ± 10.1||33.1 ± 9.0||32.1 ± 9.7||32.0 ± 9.9||32.8 ± 10.3|
Table 2 and Figure 2 provide the HAQ and SF-36 scores at baseline and followup for the comorbidity subgroups. Table 3 provides the results of the longitudinal analysis. Scores on the HAQ showed some difference in physical functioning between patients with somatic comorbidity and patients without comorbidity; however, this difference was not statistically significant. On the other hand, outcomes of the SF-36 did show a significant difference between patients with somatic comorbidity and patients without comorbidity. Patients with comorbid depression demonstrated lower physical functioning than patients with somatic comorbidity and patients without comorbidity. In particular, the outcomes on the HAQ showed that there is a strong negative association of comorbid depression with physical functioning. Patients with both somatic comorbidity and comorbid depression at baseline had the worst physical functioning.
|Difference with reference group||P||95% CI||Difference with reference group||P||95% CI|
|Group without comorbidity (reference group)||0||–||–||0||–||–|
|Somatic comorbidity||0.09||0.10||−0.02, 0.21||−2.80||< 0.01||−4.44, −1.18|
|Comorbid depression||0.45||< 0.01||0.27, 0.62||−4.41||< 0.01||−6.69, −1.60|
|Somatic comorbidity and comorbid depression||0.44||< 0.01||0.31, 0.58||−4.48||< 0.01||−6.85, −2.82|
Table 2 and Figure 2 outline changes in the HAQ and SF-36 scores at baseline and followup for the comorbidity subgroups. Table 4 provides the results of the longitudinal analysis. Patients without comorbidity improved in physical functioning between baseline and 11-year followup. There was no difference in the change in physical functioning between patients with somatic comorbidity, patients with comorbid depression, and patients without comorbidity, but there was a significant difference in the change in physical functioning between patients with both somatic comorbidity and comorbid depression and patients without comorbidity. This indicates that the difference in physical functioning between both groups increased between baseline and 11-year followup.
|Annual change||P||95% CI||Annual change||P||95% CI|
|Group without comorbidity (reference group)||−0.02||< 0.01||−0.03, −0.01||0.52||< 0.01||0.353, 0.679|
|Difference with reference group|
|Somatic comorbidity||0.004||0.46||−0.007, 0.016||−0.107||0.30||−0.031, 0.093|
|Comorbid depression||−0.004||0.68||−0.026, 0.016||−0.045||0.81||−0.397, 0.324|
|Somatic comorbidity and comorbid depression||0.018||0.02||0.002, 0.033||−0.369||0.01||−0.650, −0.083|
This study was conducted to determine the long-term change in physical functioning in patients with RA and the association with both somatic comorbidity and comorbid depression. Physical functioning improved slightly over 11 years for the total group. Somatic comorbidity and comorbid depression at baseline were negatively associated with physical functioning. This association remained over the entire 11-year period. RA patients with comorbid depression had the lowest level of physical functioning. Moreover, the combination of somatic comorbidity and comorbid depression was negatively associated with change in physical functioning over time.
Our results showed an improvement in physical functioning between baseline and 11-year followup for the total group on both the HAQ and the SF-36. Two other studies that investigated changes in physical functioning in patients with RA found a decline in physical functioning ([19, 34]). These studies showed an annual increase in HAQ score of 0.01. In both studies, a substantial part of the progression in HAQ score was attributable to age, which was more visible in patients ages >65–70 years. In our study, we controlled for age. Removing age from the model (data not shown) resulted in a smaller increase in physical functioning. Another explanation for the improvement in physical functioning could have been the introduction of new medication, which was introduced around 2001. Also, the process of response shift could explain the improvement in physical functioning. Response shift refers to the observation that many individuals tend to change the subjective evaluation of their health status as a result of changes in their objective health status (). Finally, and although we took measures to reduce this possibility (see below), selection bias could have occurred.
From our results, we can conclude that baseline comorbidity was negatively associated with long-term physical functioning. Patients with somatic comorbidity, comorbid depression, or both comorbidities at baseline displayed worse physical functioning than patients without comorbidity. These findings are in line with studies that have investigated the impact of somatic comorbidity on physical functioning in the short term ([17, 18]). Somatic comorbidity and comorbid depression are differentially associated with physical functioning. Comorbid depression resulted in lower physical functioning outcomes than somatic comorbidity. These results are in accordance with Baumeister et al, who emphasized the different impacts of somatic comorbidity and comorbid psychological disorders on quality of life, including physical functioning in patients with chronic, somatic diseases (). Several studies support the association between depressive symptoms and poor physical functioning in patients with RA ([13, 36]). Depression is linked to both the disease process and physical functioning by biologic, behavioral, cognitive, and social pathways (). Authors of the studies on depression in RA state that clinicians should pay more attention to the RA patient's psychological well-being (). Our results further emphasize this, given the long-term impact of depression. Clinicians rarely examine psychosocial factors in clinical practice, and depression in patients with chronic somatic illnesses is often underdiagnosed ([38, 39]). Screening and treatment of comorbid depression would most likely contribute to improving clinical outcomes, particularly regarding physical functioning ([36, 40]).
Only the combination of somatic comorbidity and comorbid depression was negatively associated with change in physical functioning over time, compared with the group of patients without comorbidity. Patients without comorbidity only slightly improve in physical functioning between baseline and 11-year followup. Patients with both somatic comorbidity and comorbid depression, however, did not improve in physical functioning. Michaud et al investigated which variables predicted change in physical functioning over time, as measured with the HAQ (). They found that comorbidity was an important predictor for change in physical functioning. In that study, however, although both somatic comorbidity and comorbid depression were measured, the effects of having both types of comorbidity were not compared. Moreover, the mean followup period in their study was shorter (3.7 years).
The major strengths of our study are its longitudinal design, its long-term followup period, the large patient sample, and the relatively high response rate during the 11-year period. To our knowledge, this is the first study on the long-term association between comorbidity and physical functioning, as well as the differential association of somatic comorbidity and comorbid depression with physical functioning.
Loss to followup might introduce bias into longitudinal studies. In this study, we used a statistical model that took into account the selection bias that may have occurred. We optimized the estimation of physical functioning by adding all baseline demographic variables that are regarded as prognostic factors for health outcomes. The model assumes that nonresponders have the same physical functioning during followup as comparable responders, i.e., responders with equal values for all variables in the model. Therefore, the model fills in the data of nonresponders with data of comparable responders (responders that have the same physical functioning at baseline and are comparable in other variables used in this model). Although this reduces the risk of selection bias, there is still a risk of an incorrect estimation, e.g., if nonresponders have the same baseline scores as comparable responders, but the course of their physical functioning develops differently. Therefore, although this method is appropriate in addressing selection bias (), it still remains a risk. In our study, information about physical functioning and comorbidity was collected with questionnaires. We did not have any information about use of medication. Because of this, we could not investigate the influence of new medications that were introduced between 2002 and 2008. However, we believe that all patients would have been treated according to the current insights at the time. Previous research has found that nontreatment factors have a greater effect on physical functioning than treatment factors ().
In this study, we only took into account the comorbid conditions that were present at baseline. It is possible that the improvement in long-term physical functioning was caused by a change in the number or severity of the comorbid conditions. Although it is possible that some comorbidities that were present at baseline disappeared or became less severe at followup, it seems more likely that during followup patients developed more or more severe comorbidities (). Therefore, a change in the number or severity of comorbidities does not seem to be a likely explanation for the improvement in long-term physical functioning.
Finally, comorbidity was measured as a dichotomous variable. Levels of severity were not measured. Stronger associations might have been found if the levels of severity had been assessed.
Our results stress the importance of paying attention to comorbid conditions in clinical practice. They show that patients without comorbidity report stable, or even a slight improvement in, physical functioning over time. Patients with comorbidity, and in particular patients with both somatic comorbidity and comorbid depression, however, remain behind in physical functioning over 11 years of followup. For the HAQ, the difference in change in physical functioning score for the group of patients with both somatic comorbidity and comorbid depression compared with the group without comorbidity was 0.018 annually. When multiplying this difference by 11 years, the difference in change between both groups after 11 years is clinically significant (). This means that if a patient with RA has both somatic comorbidity and comorbid depression at a particular time point, there is an increased risk that he or she shows clinically significantly worse outcomes in physical functioning after 11 years. As such, it is important to ascertain the presence of such comorbidities at an early stage so treatment can be adjusted. Integrating a routine screening process for both somatic comorbidity and comorbid depression in patients with RA is recommended. In addition, intervention strategies should be developed for the adequate treatment and management of somatic and psychological comorbidities to reduce their negative impact on physical functioning and to improve clinical outcomes. This requires integration of care of different professions. Alongside pharmacologic treatment aimed at improving disease activity, multidisciplinary rehabilitation can further improve physical functioning by booking improvements in functional ability and psychological and social health via nonpharmacologic treatment modalities ().
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Ms van den Hoek had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Roorda, Boshuizen, Rupp, van den Bos.
Acquisition of data. Van Hees, Rupp.
Analysis and interpretation of data. Van den Hoek, Roorda, Boshuizen, Tijhuis, Dekker, van den Bos.