Article Text

Extended report
Association of atherosclerosis with presence and progression of osteoarthritis: the Rotterdam Study
  1. Theun Antonius Hoeven1,
  2. Maryam Kavousi2,3,
  3. Stefan Clockaerts4,5,
  4. Hanneke J M Kerkhof3,6,
  5. Joyce B van Meurs3,6,
  6. Oscar Franco2,3,
  7. Albert Hofman1,2,3,
  8. Patrick Bindels1,4,
  9. Jacqueline Witteman2,3,
  10. Sita Bierma-Zeinstra1,4,5
  1. 1General Practice, Erasmus University Medical Centre, Rotterdam, The Netherlands
  2. 2Department of Epidemiology, Erasmus University Medical Centre, Rotterdam, The Netherlands
  3. 3Netherlands Genomics Initiative, Rotterdam, The Netherlands
  4. 4Department of Orthopaedic Surgery and Traumatology, University Hospital, Antwerp, Belgium
  5. 5Department of Orthopaedics, Erasmus University Medical Centre, Rotterdam, The Netherlands
  6. 6Department of Internal Medicine, Erasmus MC, Rotterdam, The Netherlands
  1. Correspondence to Theun Antonius Hoeven, Erasmus University Medical Centre, General Practice, Rotterdam 3015 CA, The Netherlands; t.hoeven{at}


Objective This study examined whether vascular alterations are associated with the presence and progression of osteoarthritis of the knee, the hip and the different hand joints in a large prospective cohort study.

Methods In this population-based study involving participants aged 55 years and older (Rotterdam Study I), men (n=2372) and women (n=3278) were analysed separately. x-Rays of the knee, hip and hand were scored using the Kellgren and Lawrence score for osteoarthritis at baseline, after 6.6 years and 10 years. Measures of atherosclerosis (carotid intima media thickness (IMT) and carotid plaque) and data on covariates (age, body mass index, hypertension, cholesterol ratio, diabetes mellitus and smoking) were collected at baseline. Multivariate logistic regression models with generalised estimated equations were used to calculate OR and corresponding 95% CI. Secondary multiple comparison adjustment resulted in a significance level of p<0.0021.

Results In women, IMT showed an independent association with the prevalence of knee osteoarthritis (adjusted OR (aOR) 1.7, 1.1 to 2.7), and carotid plaque with distal interphalangeal (DIP) osteoarthritis (aOR 1.4, 1.2 to 1.7) and with metacarpophalangeal osteoarthritis (aOR 1.5, 1.1 to 2.2). An independent association for IMT with progression of metacarpophalangeal osteoarthritis was found in women (aOR 2.9, 1.18 to 6.93). Additional adjustment for multiple testing yielded a significant association between carotid plaque and DIP osteoarthritis in women (p<0.001).

Conclusions This study showed independent associations of atherosclerosis with osteoarthritis of the knee and hand joints in women. The evidence was most solid for a relation with DIP osteoarthritis. More research is needed to confirm the associations and examine the differential association with various joints.

Statistics from

Osteoarthritis is a highly prevalent joint disorder that causes a huge burden of pain and disability.1 It is characterised by loss of cartilage structure, subchondral bone sclerosis, synovial inflammation and osteophyte formation, with involvement of the whole joint (ie, joint failure).2 Different risk factors have been suggested for osteoarthritis, among these age, female gender and obesity are the key ones. The involvement of other potential risk factors such as diabetes,3 menopause4 ,5 and cholesterol6 in the disease process suggests that osteoarthritis could be part of or linked to the metabolic syndrome.7 In addition, a higher risk of cardiovascular death has been reported for patients with knee or hip osteoarthritis.8

Atherosclerosis is an important feature of cardiometabolic disorders. Although some studies have indicated that atherosclerosis is associated with osteoarthritis, they are few in number, often lack sufficient power and are cross-sectional only. As a result, it is unclear whether atherosclerosis and osteoarthritis are associated, either as concurrent diseases due to a common aetiology or are causally related. Furthermore, whether there are differences in association with vascular diseases between joints affected by osteoarthritis remains undetermined.

Therefore, using data from a large longitudinal population-based study, we investigated the role that atherosclerosis might play in the prevalence, but also in the incidence and progression of osteoarthritis, and examined whether this association might vary by types of joints, gender and other key factors.


Study population

The Rotterdam Study is a prospective population-based cohort study that was set up in 1989 to investigate the occurrence and determinants of diseases in middle-aged and older populations.9 All 10 275 inhabitants aged 55 years and older who had been living for at least 1 year in the Ommoord district of the city of Rotterdam were invited to participate. The response rate was 78%, meaning that 7983 individuals responded. All participants gave written informed consent and the study was approved by the ethics committee of the Erasmus University Medical Centre, Rotterdam. Baseline measurements were obtained from 1990 to 1993 and consisted of a home interview and visits to the research centre for physical examinations. Follow-up data were collected during follow-up visits from 1997 to 2000 and from 2002 to 2005. In 2011, we included participants for whom radiographs of knees, hips or hands at baseline and follow-up were available and had been scored.

Measures of atherosclerosis

A 7.5 MHz linear array transducer with a duplex scanner (ATL UltraMark IV) was used for ultrasonography of both carotid arteries. On a longitudinal two-dimensional ultrasound image of the carotid artery, the near and far walls of the carotid artery were displayed as two bright white lines separated by a hypoechogenic space. The distance from the leading edge of the first bright line of the far wall (lumen–intima interface) to the leading edge of the second bright line (media–adventitia interface) indicated the intima media thickness (IMT). Data on IMT were available for 3369 women and 2275 men. Reproducibility assessed by the intraclass correlation coefficient was 0.74.10

The presence of plaques in the carotid artery was assessed by examining the ultrasonographic images of the common, internal and bifurcation sites of the carotid artery for the presence of atherosclerotic lesions. Plaques were defined as a focal widening relative to adjacent segments, with protrusion into the lumen composed of either only calcified deposits or a combination of calcified and non-calcified material.11 ,12 Data on the presence of plaque were available for 2665 women and 1808 men. Readers of the ultrasound images were unaware of the case status of the subject.

Assessment of osteoarthritis

Standard anteroposterior radiographs of both hands were scored by three trained assessors. Each joint was graded for overall radiographic osteoarthritis using a Kellgren–Lawrence (K/L) grade scaled 0–4 (see supplementary table S1, available online only). Radiographic osteoarthritis was defined as a K/L grade of 2 or greater. Distal interphalangeal (DIP), proximal interphalangeal (PIP), metacarpophalangeal and carpometacarpal/trapezoscaphoid (CMC/TS) groups were defined as positive if the K/L score in at least one joint in the group was 2 or greater. The interobserver reliability of a K/L score of 2 or more as a dichotomous variable expressed by the intraclass correlation coefficient was as follows: DIP, 0.69; PIP, 0.74; metacarpophalangeal, 0.70; CMC/TS (base of thumb), 0.84. After consultation with an expert, we excluded hand x-rays with signs of rheumatoid arthritis or fractures. x-Rays of the hand were available at baseline for 3210 women and 2428 men, and after 10 years of follow-up for 1385 women and 1057 men. Progression of hand osteoarthritis was defined as 1 grade increase in the K/L scale. This was done for each of the hand joint groups separately (H.H.M. Kerkhof, 2011, unpublished data).

Knee and hip x-rays were taken and scored with a K/L grade scaled 0–4 (see supplementary table S1, available online only). Radiographs of the pelvis were obtained when both feet were rotated 10° inward and the x-ray beam was centred on the umbilicus, and knee radiographs were taken with the knee extended and the patella in a central position. The intraclass correlation coefficient was 0.71 for the knee and 0.74 for the hip.13 ,14 Knee and hip osteoarthritis progression was defined as an increase in the K/L score between baseline and follow-up of 1 or more.15 x-Rays of the knee and hip were available at baseline for 3278 women and 2372 men, after 6.6 years for 2449 women and 1232 men, and after 10 years for 1149 women and 941 men. If x-rays were available after 6.6 years, but not after 10 years of follow-up, data were forwarded. Assessment of osteoarthritis was performed blinded for clinical and demographic data.


Trained interviewers gathered information on medical history and risk factors for chronic diseases. Participants were invited to visit the research centre for clinical examinations and laboratory assessments. The following information was collected: age; body mass index (BMI; weight (kg)/height (m2)); total cholesterol/high-density lipoprotein (HDL) ratio (serum total cholesterol determined by an enzymatic procedure; HDL measured after precipitation of non-HDL cholesterol); current smoking (self-reported); diabetes mellitus (use of glucose-lowering medication or non-fasting random or postload glucose levels exceeding 11.0 mmol/l); and arterial hypertension (systolic blood pressure of 140 mm Hg or higher, diastolic blood pressure of 90 mm Hg or higher, or use of antihypertensive drugs for hypertension).

Statistical analysis

We analysed the association between atherosclerosis and the presence of osteoarthritis using multivariate logistic regression models in which IMT and plaque were included separately for the calculation of OR and 95% CI. Thickness of the common carotid arteries (IMT) was analysed continuously. The presence of plaque was treated as a binary variable in the analysis. Different models were built to provide insight into how the link between atherosclerosis and osteoarthritis might be driven (non-adjusted, age adjusted, age and BMI adjusted and fully adjusted). The inclusion of potential confounding variables such as age, BMI, total cholesterol/HDL ratio, diabetes, hypertension and smoking was based on existing literature about risk factors for cardiovascular disease and osteoarthritis and following initial univariate evaluation. In addition, adjustments for follow-up time and baseline K/L score were made for the multivariate analyses on progression of osteoarthritis. We used generalised estimating equations, a mixed logistic regression model, to adjust for correlations between the right and left extremity in each individual and to increase discriminative ability. Because the definition of progression implies that participants are conditioned by the baseline presence of osteoarthritis the estimates of an effect of exposure (atherosclerosis) on outcome (osteoarthritis) might be biased. We therefore again included all potential confounders in the model to minimise conditioning by baseline.16 Subjects with missing values were excluded from the analysis. Stratification by sex was based on previous studies.17 ,18 Our a priori p value threshold was set at p≤0.05. However, due to multiple comparisons for the different joint groups, the two measures of atherosclerosis, and both genders we further examined the associations in a second analysis for the more rigorous p value of p<0.0021 (α/24). We performed all analyses with SPSS version 17.0.


General characteristics

Of the total 7983 participants, 3278 women and 2372 men were included in this study. Because they did not have x-rays taken at baseline, 2333 participants were excluded. Table 1 shows the baseline characteristics of the study population, overall and stratified by sex. The mean age of the population at baseline was 68.2 years; 58% were women. In this study, women were slightly older (68.6 years vs 67.5 years), had a higher BMI (26.8 kg/m2 vs 25.7 kg/m2), and more of them reported a history of hypertension (58% vs 53%) compared with men. However, there were more men with carotid plaques (66% vs 55%). In men, carotid IMT was higher (1.1 mm vs 1.0 mm), they smoked more (29% vs 19%) and had a higher total cholesterol/HDL ratio (5.5 vs 5.1). The commonest osteoarthritis sites in women were DIP joints (39%), CMC/TS joints (33%) and knee joints (20%). In men, osteoarthritis in the DIP joints was commonest (25%), followed by CMC/TS joints (19%). The prevalence of osteoarthritis was significantly higher in women for all joint groups.

Table 1

Baseline characteristics of the study population

Association of atherosclerosis with osteoarthritis

Table 2 shows the association between IMT, carotid plaques and the presence of knee and hip osteoarthritis in the different models by gender. After full adjustment IMT showed an association with the prevalence of knee osteoarthritis (fully adjusted OR (aOR) 1.7, 1.1 to 2.7) at a p value less than 0.05 in women. Associations between IMT or carotid plaque with hand osteoarthritis by gender are presented in table 3. After full adjustment, carotid plaque showed an association with DIP osteoarthritis (aOR 1.4, 1.2 to 1.7) and with metacarpophalangeal osteoarthritis (aOR 1.5, 1.1 to 2.2) in women. No independent associations existed between measures of atherosclerosis and hip osteoarthritis, PIP and CMC/TS osteoarthritis. Comparison of the crude and adjusted models in table 2 and table 3 shows that the associations between IMT and plaque with the presence of knee, hip, or hand osteoarthritis are largely attenuated by age adjustment.

Table 2

Analysis of vascular variables in relation to the presence of knee osteoarthritis and hip osteoarthritis

Table 3

Analysis of vascular variables in relation to the presence of hand osteoarthritis

The relationship between vascular variables and overall progression of osteoarthritis (table 4) showed a fully adjusted association between IMT and overall progression of metacarpophalangeal osteoarthritis (aOR 2.9, 1.18 to 6.93) at a p value of less than 0.05 in women. In men, no fully adjusted significant associations were observed between IMT or plaque and the presence (table 2 and table 3) or overall progression of osteoarthritis (data not shown).

Table 4

Analysis of vascular variables in relation to the overall progression of osteoarthritis in women

In a second step, we tested our results conservatively by using rigorous correction for multiple testing at p<0.0021 (α/24). This showed that the association between carotid plaque and DIP osteoarthritis prevalence remained (aOR 1.4, 1.2 to 1.7).


In the present study, independent correlations between measures of atherosclerosis with knee and hand osteoarthritis prevalence were found in women after adjustment for cardiovascular risk factors. The same applied for the progression of hand osteoarthritis. In the following sections, we will discuss the most important findings.

Gender differences

This study showed an independent association between atherosclerosis and osteoarthritis of the DIP, metacarpophalangeal and knee joints in women, whereas this was not the case for men. Jonsson et al18 previously demonstrated that hand osteoarthritis and atherosclerosis were associated in older women and hypothesised that circulatory disturbances in the synovial membrane and subchondral bone might contribute to the cartilage destruction and to the pathophysiological process of osteoarthritis; we confirmed their results in this study. Our results might indicate that due to hormonal changes during the menopause, women are more at risk of developing vascular disease and osteoarthritis. However, a systematic review found no clear association between female hormonal aspects and osteoarthritis.19 Furthermore, effect modification by years since the menopause was not present in our study. Recently, Gast et al20 showed that women with vasomotor menopausal symptoms are at risk of coronary heart disease. We studied a population of women aged 55 years and older and detailed data on the perimenopausal period were not available. It is probable that the relation between the menopause, atherosclerosis and osteoarthritis is complex and can not be shown in our study population. Younger perimenopausal women with early signs of osteoarthritis manifest a greater metabolic component, and research focusing on this group might elucidate why vascular disease and osteoarthritis are associated in women, but not in men.

Knee osteoarthritis, DIP osteoarthritis and metacarpophalangeal osteoarthritis

We showed associations between atherosclerosis and osteoarthritis in several joints in women, but not with carpometacarpal osteoarthritis or osteoarthritis of the hip. In the same way, Dahaghin et al21 showed that BMI was associated with osteoarthritis of DIP, PIP and metacarpophalangeal joints, but not with osteoarthritis of the thumb base (carpometacarpal osteoarthritis). Carpometacarpal osteoarthritis might be regarded a subset of hand osteoarthritis and aetiologically different from DIP, PIP and metacarpophalangeal osteoarthritis.22 Various studies also support a different pathogenesis between hip osteoarthritis and knee osteoarthritis.23,,25 Systemic factors such as disordered glucose metabolism, lipid metabolism and atherosclerosis, all related to the metabolic syndrome, seem to be more related to knee osteoarthritis than hip osteoarthritis. For example, BMI is highly associated with knee osteoarthritis, but not or to a lesser degree with hip osteoarthritis.23 ,24

Cardiometabolic disorders and osteoarthritis

The metabolic syndrome is a multiplex cardiovascular risk factor. Besides atherosclerosis, obesity, diabetes mellitus and dyslipidaemia are listed as part of the metabolic syndrome. The main characteristic is the presence of systemic inflammation caused by visceral adipose tissue that secretes proinflammatory cytokines and adipokines.26 These inflammatory processes are involved in the aetiopathogenesis of atherosclerosis, but also in rheumatic diseases such as psoriasis arthritis. There is increasing evidence that chronic systemic inflammation in patients with the metabolic syndrome might also initiate or stimulate the osteoarthritis disease process.7 We studied vascular variables that not only reflect generalised atherosclerosis, but are typically increased in individuals with the metabolic syndrome, or also cardiometabolic disorders in general.27 ,28 Based on our results, osteoarthritis of the DIP joints may be linked to cardiometabolic disorders in women. Additional investigations to test this hypothesis are warranted.

Progression of osteoarthritis

We noticed a tendency for an association between IMT and progression of metacarpophalangeal osteoarthritis and knee osteoarthritis in women. However, indications for contradictory estimates were also found for the association of vascular variables with the overall progression of hip osteoarthritis in women. We can not give a clear explanation for these results; different relations between the overall progression of knee and hip osteoarthritis have been reported previously.15 In addition, because the reasons for total hip replacements were not always clear, the overall progression of hip osteoarthritis remained a complex variable to analyse, although we also adjusted for hip fractures. Furthermore, the very limited power for hip progression can give spurious results.

Overall progression of osteoarthritis was defined as the combination of the incidence and the progression of existing osteoarthritis at baseline.15 When we analysed incidence or progression separately, as previously described in hands29 or in knees and hips,25 we found similar results (data not shown). However, associations between vascular variables and the incidence of knee osteoarthritis in women were stronger than with the true progression of knee osteoarthritis. These results do not fully agree with the hypothesis proposed by Conaghan et al30 that by causing subchondral bone ischaemia, atheromatous vascular disease contributes more to the progression of osteoarthritis than to its initiation. The mechanism leading to the initiation and progression of osteoarthritis still remains largely unknown.31

Strengths and limitations

The major strength of this study is its size. We used a large population cohort study, the Rotterdam Study, in which all relevant clinical and radiographic data were collected from a population at risk of osteoarthritis and cardiovascular disease. Detailed cardiovascular data, reflecting generalised atherosclerosis, were available,10 ,11 and in the adjusted models we included potential confounders that could influence the association between atherosclerosis and osteoarthritis. Furthermore, we adjusted for the baseline presence of osteoarthritis in our analyses of progression.16 Due to the longitudinal nature of the Rotterdam Study, we had the opportunity to analyse both the prevalence and overall progression of osteoarthritis.

Although our joint-specific analyses were hypothesis based, we tested our results conservatively by using rigorous adjustment for multiple testing in a second step. Some other hand joints as well as the knee joints showed effect estimates of similar magnitude in the same direction in both crude and adjusted analyses, but they did not reach statistical significance after these rigorous adjustments, possibly because of the lower prevalence of osteoarthritis in these joints. Therefore, we believe that our findings are supportive of an association between atherosclerosis and hand osteoarthritis and knee osteoarthritis in women, as found in previous studies.17 ,18

The selection of our participants, based on the availability of radiographs, might have introduced a healthy workers effect, because these participants survived the follow-up period and were fit enough to visit the research centre. We selected a younger and healthier sample, and the incidence and progression of osteoarthritis in these participants will be lower than in a population-based sample, so we can assume that if this healthy workers effect would have influenced the associations we found, it would underestimate the true effect. In addition, data on the overall progression of osteoarthritis were available for a small number of participants, which has limited our power to detect these associations. Unfortunately, we did not have information on activity levels. As a result, residual confounding can not be totally ruled out.

Future research

This study is the first large-scale prospective population-based cohort study that addresses the association between measures of atherosclerosis and osteoarthritis. Our study can therefore be viewed as one of the first steps to unravel the link between these two conditions. Many of the covariates are characteristics of a cardiovascular and metabolic profile.6 ,32 ,33 The associations between measures of atherosclerosis and osteoarthritis were highly driven by age, and additional links seem to exist in certain joint types. We do not know which risk factors are specific for osteoarthritis. We do know, however, that chronic inflammation plays an important role in both diseases. The accumulation of glycation end-products in cartilage and arteries, or alternatively from oxidative stress in cartilage and arteries indicates a shared aetiology in vascular disease and osteoarthritis.34 However, additional replication in other studies is recommended and more detailed research is necessary to identify the exact underlying mechanisms for the association between cardiovascular disease and osteoarthritis. A better understanding of how inflammatory activity in osteoarthritis and atherosclerosis in women interacts with other recognised risk factors such as increasing age, the menopause, diabetes and genetics might lead to improved treatment strategies in women. In addition, reverse causation in which, for example, painful knee or hip osteoarthritis leads to a sedentary lifestyle and a cardiometabolic syndrome might exist.


This study showed independent associations of atherosclerosis with osteoarthritis of the knee and hand joints in women. The evidence was most solid for a relation with DIP osteoarthritis. More research is needed to confirm the associations and examine the differential association with various joints.


The authors are very grateful to the participants and staff from the Rotterdam Study. The authors would like to thank M Reijman, A Bergink, S Dahaghin, A van Vaalen, M Kool, JS Dekkers, V Kahlmann, ES Stille and DAC Sluiter for the scoring of radiographs and E Oei for assistance.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:

    • Web Only Data - This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
  • Lay summary

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.


  • Funding This study was funded by the Dutch Arthritis Foundation.

  • Ethics approval The study was approved by the ethics committee of the Erasmus University Medical Centre, Rotterdam.

  • Patient consent Obtained.

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.