Tendon involvement and its association with pain and hand function in patients with osteoarthritis of the hand

Abstract

Objective

To characterize the frequency and influence of tenosynovitis and tendon damage on pain and hand function using clinical examination and US in hand OA.

Methods

We included 86 patients with hand OA and 23 age- and sex-matched control subjects. Extensor and flexor tendons of both hands were assessed by clinical examination and US for tenosynovitis and tendon damage. Conventional radiographs were acquired. Hand function was evaluated by the function subtest of the M-SACRAH (modified Score for the Assessment and Quantification of Chronic Rheumatoid Affections of the Hands) questionnaire and the Moberg pick-up test. K-means cluster analyses was calculated to assess clusters based on radiographic features and sonographic tendon scores.

Results

US identified the involvement of one or more tendon in 60/86 (69.8%) hand OA patients compared with 2/23 (8.7%) subjects (P < 0.01) in the control group. In the hand OA group, US detected tendon damage more often in flexor tendons compared with extensor tendons (2.1% vs 0.9%, P = 0.03), while tenosynovitis was observed more often in extensor tendons compared with flexor tendons (8% vs 0.6%, P < 0001). The sensitivity and specificity of clinical examination to detect tendon involvement was 81.4% and 34.6%, respectively, on the patient level and 14.5% and 83.8% on the tendon level. The cluster analyses revealed one cluster with more radiographic features of hand OA and more tendon damage while more tenosynovitis was found in cluster 2. M-SACRAH function did not correlate with tendon involvement on US.

Conclusion

This study revealed a high frequency of tendon involvement in hand OA. Tendon involvement on US did not impact hand function or self-reported pain.

Introduction

Tenosynovitis is an inflammatory process affecting the synovial sheath of tendons which occurs in various rheumatic and musculoskeletal diseases (RMDs). It is associated with pain and persistent tenosynovitis, and can lead to adhesions and dysfunction. Both inflammatory and mechanical factors may cause tendon damage and ultimately result in the complete rupture of the tendon with loss of function. Early diagnosis and treatment are necessary to prevent further complications from both types of lesions [1–3].

Tenosynovitis is one of the most common early manifestations of RA [4–7]. Consequently, tendon damage is often found in RA and is associated with radiographic damage [8]. Tendon and entheseal involvement, particularly tenosynovitis and enthesitis, are typical findings in PsA [9–11]. However, data on tendon involvement in hand OA, a highly prevalent musculoskeletal condition, is limited. While hand OA is characterized primarily by progressive loss of articular cartilage and changes to subchondral bone, it is a disease that affects the entire joint, with various manifestations such as synovial hypertrophy, joint effusion, osteophytes, erosions, tendon damage and, in some cases, tendon rupture [12–14]. In addition to the lack of data on the frequency of tendon involvement in hand OA, the influence of such involvement on hand function or pain is not known.

A major limitation of research in this area is the lack of standardization of clinical assessment of tendons, which is generally considered to be more difficult as compared with examining joint tenderness and swelling. Therefore, musculoskeletal US (MSUS) has been successfully applied in various studies in inflammatory RMDs as a sensitive tool to detect both tenosynovitis and tendon damage [3, 7, 15, 16].

This study aimed to characterize tendon involvement in hand OA, assess its frequency and influence on pain and hand function, and its potential association with radiographic features.

Methods

This was a cross-sectional, monocentre observational study at the Division of Rheumatology of the Medical University of Vienna, a tertiary centre for RMDs. Inclusion criteria for the hand OA group included: diagnosis of hand OA according to the ACR classification criteria [17]. To assess and interpret the frequency of sonographically detected tendon involvement in hand OA we also enrolled a cohort of age- and sex-matched subjects with the absence of bony enlargement in any PIP, DIP or CMC 1 joint, and no self-reported pain of the hands as a control group. Exclusion criteria for both groups included: the documented history of inflammatory RMD; prior or current treatment with DMARDs or oral glucocorticoids; and recent documented trauma or surgery (<2 months) of the hands. All patients provided informed consent. This study complies with the Declaration of Helsinki and was approved by the ethics committee of the Medical University of Vienna (EK1206/2016).

Clinical evaluation

Clinical examination was performed on hand OA patients by an experienced biometrician blinded to sonographic and radiographic data. The following tendons of the hand and fingers were examined bilaterally: flexor digitorum superficialis and profundus tendons of the II–V finger, flexor pollicis longus tendon, as well as the extensor tendons of extensor compartments 1–6. Clinical symptoms of tenosynovitis and tendon damage were assessed in each wrist and finger in terms of volar or dorsal pain, crepitus and swelling involving the hand, wrist or forearm during active movement of the tendon against resistance according to the Birmingham consensus criteria [18], as well as in terms of tendon function in a binary fashion for each tendon.

Sonographic evaluation

Sonographic examination was performed in both groups, with patients sitting with their hands resting on the examination table. In order to ensure objectivity as much as possible, sonographers had no access to the clinical diagnosis prior to the examination and the examination room was darkened. However, the nature of the sonographic examination, prohibits complete blinding to clinical diagnosis. The examinations were performed on a GE Logiq E9 US unit equipped with multifrequency linear transducers (6–15 MHz and 8–18 Mhz) by expert sonographers with >10 years of experience in musculoskeletal sonography. Longitudinal and transverse scans were performed using both greyscale (GS) and power Doppler flow according to the EULAR standardized procedures [19]. The following settings were used for power Doppler: frequency 11.9 MHz, PRF 600 Hz and medium persistence, Doppler gain optimized to suppress noise, particularly at the bony cortex.

MCP and PIP joints were examined in the extended position; wrists in a neutral position. The following tendons of the hand and fingers were examined bilaterally at the level of the wrist and individual fingers: flexor digitorum superficialis and profundus tendons of the II–V finger, flexor pollicis longus, flexor carpi radialis and extensor tendons of extensor compartments 1–6.

Tenosynovitis and tendon damage was defined according to the definitions of the OMERACT Ultrasound Working Group [20, 21].

Both a binary scoring system (0–1 for absence and presence, respectively) [22], as well as a four-grade semiquantitative scoring system [21, 23], were used for the sonographic grading of each tendon for GS signs of tenosynovitis and tendon damage, respectively, and for tenosynovial/intratendinous power Doppler signals. Global sonographic scores were constructed on the patient level for the binary scoring system for GS signs of tenosynovitis (range 0–22) and for tendon damage (range 0–22). A combined score was calculated where either tenosynovitis or damage on GS and/or power Doppler results in a scoring of 0 or 1 (range 0–22).

Furthermore, the wrists, proximal (1–5) and distal (2–5) IP joints were scored for osteophytes (presence/absence for each joint) using the definition proposed by the OMERACT Ultrasound Working Group [24]. A global sonographic score for osteophytes on the patient level was calculated (range 0–20).

To test interrater reliability, two expert sonographers (G.S., P.M.) performed US examinations grading both GS signs of tenosynovitis, tendon damage, tenosynovial/intratendineous power Doppler signal and osteophytes in five randomly selected patients. The two independent examinations subsequently took place on the same day, and the sonographers were blinded both to the clinical findings and the findings of the other sonographer.

Radiologic evaluation

Plain posterior–anterior radiographs were used to evaluate radiographic signs of hand OA using the recently established interphalangeal OsteoArthritis Radiographic Simplified (iOARS) score on the PIP and DIP joints of included patients [25].

Functional assessment

Hand OA patients were asked to fill out the M-SACRAH questionnaire, a modified, shortened version of the SACRAH (Score for the Assessment and Quantification of Chronic Rheumatoid Affections of the Hands) [26, 27] questionnaire, which is used to quantify hand function, stiffness and pain in hand OA and RA patients [26] during the patient visit to assess the potential role of tenosynovitis/tendon damage in hand function. To quantify function, the M-SACRAH function subtest was used. In addition, the Moberg picking-up Test (MPUT) was carried out to assess hand function physically [28]. Pain levels were evaluated using a visual analogue scale (VAS 0–10) asking for pain in the joints of the hands according to the M-SACRAH [27].

Statistical analysis

Patient characteristics are described as numbers and percentages or mean and s.d. for categorical or continual variables, respectively. Differences in baseline characteristics were calculated by Student’s t-test or χ² test, respectively. Differences in the prevalence of tendon involvement between the left and the right hand and between extensor and flexor tendons were calculated with the χ² test. The agreement between sonographic and clinical diagnosis was calculated by Cohen’s kappa for individual tendons and on the patient’s level. The interrater reliability between the two sonographers was also calculated by Cohen’s kappa and interpreted according to Landis and Koch [29].

The global sonographic tendon scores were correlated with the M-SACRAH function score, the results of the MPUT, pain VAS, the number of joints with sonographically detected osteophytes, as well as the iOARS and the iOARS subscores using Spearman correlation coefficient. The number of joints with osteophytes was correlated with M-SACRAH function, MPUT and pain VAS scores.

The number of joints with osteophytes detected on US along each tendon was assessed and analysed by age- and sex-adjusted logistic binary regression using tendon involvement (tenosynovitis and tendon damage detected on US) as the dependent variable.

To assess potential subentities among the hand OA group according to radiographic and sonographic features, we performed a K-means clustering analyses including the number of tendons with tenosynovitis and tendon damage as well as the iOARS subscores. The number of clusters was set at 2.

We performed a sample size calculation to detect a significant correlation between the scores with r ≥ 0.3, with alpha set at 0.05 and beta at 0.8. This yielded a total of 85 patients as a necessary sample size. Statistical significance was defined as P < 0.05. All statistical analyses were carried out using SPSS 29.0 (SPSS, Chicago, IL, USA).

Results

Patient characteristics

Ninety patients with hand OA fulfilled the required inclusion and exclusion criteria. Four patients had to be excluded due to incomplete data; in total, we analysed data from 86 patients. Most patients (75/86, 87.2%) were female, with a mean age of 65.9 ± 9.5 years. We included 23 patients in the age- and sex-matched control group. Mean pain level on the VAS was 3.2 ± 2.4 in the hand OA group and 0 ± 0 (P < 0.01) in the control group (Table 1). Mean iOARS was 30.8 ± 16.9 in the hand OA group.

All individuals among the control group (23/23, 100%) had at least one comorbidity, compared with 47/64 (73.4%) hand OA patients (P = 0.03). Among hand OA patients, 16/64 (25%) took NSAIDs, compared with no patient in the control group (Table 1).

Interrater reliability

Interrater reliability was excellent for GS signs of tenosynovitis [kappa 0.94 (95% CI 0.78–0.96)], GS signs of tendon damage [kappa 0.91 (95% CI 0.67–0.97)] and tenosynovial/intratendinous power Doppler [kappa 0.90 (95% CI 0.71–0.94)].

Prevalence and distribution of tendon involvement detected by US

Any tendon involvement was found on US in 60/86 hand OA patients (69.8%). US detected GS tenosynovitis of one or more tendon in 56/86 patients (65.1%) with 1.0 ± 1.0 affected tendons per patient on average. Tendon damage (one or more tendon) was found in 14/86 patients (16.3%), with 0.4 ± 1.0 affected tendons per patient on average. Power Doppler (PD) positivity of one or more tendon was found in 20/86 patients (23.3%). In the control group, we found significantly less tenosynovitis [2/23 (8.7%), patients, P < 0.01] and no tendon damage [0/23 (0%) patients, P = 0.04] (Table 1).

In the hand OA group, tenosynovitis was found in 89/2064 (4.3%) tendons and graded as 1 in 2.2% (45/2064), 2 in 1.5% (30/2064) and 3 in 0.2% (4/2064) of tendons. Tendon damage was detected in 31/2064 (1.5%) tendons and graded as 1 in 1% (21/2064) and 2 in 0.1% (3/2064) of tendons. No grade 3 lesion was detected. Power Doppler grade 1 and 2 were found in 14/2064 (0.7%) and 9/2064 (0.4%) tendons, respectively. No grade 3 power Doppler signal was detected. Only two tendons (2/552, 0.4%) with tenosynovitis were detected in the control group. Both were graded as GS 1, power Doppler 0.

In the hand OA group, tendon damage was found more than twice as often in flexor as compared with extensor tendons (2.1% vs 0.9%, P = 0.03), while tenosynovitis much more commonly affected the extensor tendons (8% vs 0.6%, P < 0.001). Similarly, extensor tendons were more often power Doppler positive (3.8% vs 1.2%, P < 0.05). Tendon damage (right hand 1.6% vs left hand 1.5%, P = 1) and tenosynovitis (left 5% vs right 3.6%, P = 0.13) were equally distributed between both sides. Among all tendons, tenosynovitis most frequently affected the sixth extensor compartment (55/172, 32%). Flexor digitorum II (8/172, 4.7%) was the tendon most frequently affected by tendon damage (Fig. 1).

Tendon involvement on US and radiographic features of hand OA

We found a significant, but low negative correlation between the iOARS and the tenosynovitis score (r = −0.26, P = 0.03), while no significant correlation was found with the tendon damage score (r = 0.19, P = 0.12). A low, but significant positive correlation of the tendon damage score with subchondral sclerosis was seen (Table 2). The cluster analysis revealed one cluster with lower iOARS subscores and tendon damage score, but a higher tenosynovitis score as compared with cluster 2 (Table 3).

The number of joints with osteophytes detected on sonography did not correlate with the number of detected tenosynovitis (r = −0.06, P = 0.60), while significant correlation was found with the number of detected tendon damage (r = 0.23, P = 0.03) on the patient level. The age- and sex-adjusted binary logistic regression revealed a significant association between the number of joints with osteophytes along a tendon and the occurrence of tenosynovitis (odds ratio 1.36, 95% CI 1.1–1.69; P < 0.01) of this tendon, while no significant association was found with tendon damage (odds ratio 1.18, 95% CI 0.92–1.51; P = 0.2) (Fig. 2).

Tendon involvement (on US) and hand function and pain

As expected, we found a significant correlation between M-SACRAH function and MPUT: r = 0.31, P < 0.01. However, neither the M-SACRAH function subset (r = 0.09, P = 0.42) nor the MPUT (r = 0.17, P = 0.15) correlated with the number of affected tendons detected by US (Table 4). The number of clinically affected tendons correlated significantly with M-SACRAH function (r = 0.36, P < 0.01), MPUT (r = 0.28, P = 0.02) and VAS pain (r = 0.24, P = 0.03) (Table 4). We found no association between pain and the number of affected tendons on US (r = −0.05, P = 0.65), those with tenosynovitis (r = −0.0, P = 0.84), damage (r = −0.09, P = 0.42) or positive power Doppler signal (r = 0.01, P = 0.98) (Table 4).

Agreement between US and clinical assessment

We found no agreement between sonographic assessment (any tendon involvement) and clinical examination on the level of individual tendons (kappa −0.009) and only slight agreement on the level of individual patients (any affected tendon yes/no) (kappa 0.17). The sensitivity was found to be 14.5%, the specificity 83.8% for clinical examination on tendon level with sonographic examination as the gold standard. On the patient level, defined as presence/absence (binary) of affected tendons per patients, the sensitivity and specificity were 81.4% and 34.6%, respectively.

Discussion

In this study, we found a significantly higher prevalence of tendon involvement in hand OA patients compared with age- and sex-matched healthy subjects, however no association with hand function or pain.

Data on the prevalence of tendon involvement in hand OA are scarce. One study performed MRI of the dominant hand of patients with hand OA and found 1 (median; interquartile range 0–3) flexor tenosynovitis per patient, while extensor tendons were not assessed [30]. One study reported a frequency of tenosynovitis of 30% in extensor and 16% in flexor tendons [31]. Our study revealed an even higher prevalence of tenosynovitis, and similar to that study, we found that tenosynovitis affects extensor tendons more frequently than flexor tendons.

The majority of included patients were female. This gender imbalance in the included patients is thus in line with studies demonstrating that symptomatic hand OA is more common in females [32, 33], underlining that our cohort represents a good overview of such patients.

In RA, tenosynovitis is found in up to 90% of patients [34–36]. It is regarded as part of the disease’s gestalt and associated with subsequent flares [37]. In the past, hand OA was thought to be induced by ageing and environmental, biochemical and biomechanical factors. Recently, a number of studies have emphasized the role of inflammation in the development and progression of hand OA [38–41]. Synovitis has been detected in a significant number of patients with hand OA and was associated with tenderness [30, 42, 43]. These findings raise the question of whether tenosynovitis in hand OA may be yet another manifestation of inflammation, similar to that seen in RA. Furthermore, interestingly, similar to what we found in hand OA, tenosynovitis was found often (or most often) at the extensor carpi ulnaris tendon (ECU) in studies assessing RA patients [1, 44–46]. In one study, this was compared with PsA patients, where a significant difference was found [45].

Another possible reason for the high prevalence of tenosynovitis and tendon damage might be mechanical irritation brought upon by osteophytes, which are pathognomonic for OA. In our study, the probability of tenosynovitis increased with the number of osteophytes found along the tendon in a binary logistic regression adjusted for age and sex. This suggests that osteophytes’ mechanical irritation might play a role in developing such tenosynovitis. In contrast, tendon damage was not found to be more prevalent on fingers with more osteophytes.

Interestingly, in contrast, a low, but significant negative correlation was found between the tenosynovitis score and the iOARS, while low but significant positive correlations were found between the tendon damage score and the subchondral sclerosis subscore and the osteophytes score. These results should be interpreted with caution due to the low correlation coefficients and significance levels barely meeting the threshold of 0.05. We did not adjust for multiple testing. There might be a difference between tendon damage and tenosynovitis where tendon damage is associated with the severity of hand OA while tenosynovitis is found in patients with a higher inflammatory burden but not necessarily more severe hand OA. This hypothesis was further supported by the results of the cluster analysis and should be assessed and addressed in further studies. To our knowledge, this has not been reported in other studies.

We found no association between tendon involvement with hand function. The frequency of tendon damage may have been too low to detect a significant correlation. A weak but significant correlation existed between the number of osteophytes on US and MPUT. This suggests that functional impairment may be better explained by structural damage (osteophytes) rather than by tendon involvement. In one study, the osteophyte sum score was associated with grip strength in patients with hand OA [30].

Similar to hand function, we found no association between self-reported pain and sonographically detected tendon involvement, which was in contrast to clinically detected tendon involvement. Furthermore, the number of osteophytes did not correlate with self-reported pain. One study assessed hand OA and found a significant association between osteophytes and joint tenderness by palpation. However, like in our study, the osteophyte sum score did not correlate with pain assessed by a self-reported questionnaire [30]. Self-reported pain might not be specific enough to assess pain associated with tendon involvement or the number of osteophytes.

Interestingly, we found only very slight agreement on the level of individual patients between clinical and US detection of tendon involvement and no agreement on the level of individual tendons. We further found low sensitivity but higher specificity on the tendon level, but higher sensitivity and a lower specificity on the patient level for clinical examination using sonography as a gold standard. Considering that the clinical assessment of tendons is not standardized and more difficult compared with joint examination, our findings suggest that patients with tendon involvement might be identified by clinical examination, but accurate assignment to the respective tendon may not always be possible. Furthermore, clinical examination might not differentiate between tendon involvement or other causes for pain. Pain, bony deformation and loss of function, which are known characteristics of hand OA, likely all contribute to the difficulties facing clinical assessment, and prominent joint involvement may ‘overshadow’ concomitant tendon pathology. Imaging modalities such as US may facilitate this designation and provide an opportunity to treat tendon involvement more specifically.

Limitations of our study include the lack of a standardized clinical examination and that of a gold standard to detect tenosynovitis and/or tendon damage. Our study did not assess clinical tenderness of individual joints separately, which may have allowed a more detailed evaluation of pain. Finally, the M-SACRAH questionnaire is validated to assess overall hand function in hand OA and RA, but not tendon involvement specifically.

In conclusion, we found a high prevalence of tenosynovitis in patients with hand OA. It was associated with the number of osteophytes along the tendon. We found high sensitivity but low specificity of clinical examination for tendon involvement on the level of patients. Physicians treating patients with hand OA should keep the high prevalence of tendon involvement in mind. In case of clinical suspicion, a sonographic examination should be performed. If tenosynovitis or tendon damage is detected, treatment may be tailored accordingly. Whether tendon involvement develops only secondary to biomechanical factors (osteophytes) or whether it can be regarded as an independent inflammatory feature of hand OA should be evaluated in future studies.

Data availability

Data available on reasonable request. P.M. takes responsibility for the integrity of the work as a whole, from inception to finished article.

Contribution statement

I.G. made substantial contributions to the conception or design of the work, or the acquisition, analysis or interpretation of data, drafting the work or revising it critically for important intellectual content. A.V. made made substantial contributions to the acquisition, analysis or interpretation of data. G.S. made substantial contributions to the conception or design of the work, or the acquisition, analysis or interpretation of data. M.Z. made substantial contributions to the acquisition, analysis or interpretation of data. M.D. made substantial contributions to the acquisition, analysis or interpretation of data. L.L. made substantial contributions to the acquisition, analysis or interpretation of data. V.R. made substantial contributions to the conception or design of the work. T.S. revised the work critically for important intellectual content. J.S. drafted the work or revised it critically for important intellectual content. D.A. made substantial contributions to the conception or design of the work, drafting the work or revising it critically for important intellectual content. P.M. made substantial contributions to the conception or design of the work, or the acquisition, analysis or interpretation of data, drafting the work or revising it critically for important intellectual content, and final approval of the version published, agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Funding

This study was supported by the Medical Scientific Fund of the Mayor of the City of Vienna (Grant No. 15213).

Disclosure statement: T.S. reports personal fees: AOP Health, AbbVie, Pfizer, Roche and Takeda. J.S.: grants: Abbvie, AstraZeneca, Lilly, Novartis and Roche; royalties or licenses: UpToDate, Elsevier-Rheumatology Textbook; consulting fees: AbbVie, Amgen, AstraZeneca, Astro, Bristol-Myers Squibb, Celltrion, Gilead-Galapagos, Janssen, Lilly, Pfizer, R-Pharma, Samsung and Sanofi; speaker’s bureaus: AbbVie, Amgen, AstraZeneca, Astro, Bristol-Myers Squibb, Chugai, Janssen, Lilly, Merck Sharp & Dohme, Novartis-Sandoz, Pfizer, R-Pharma, Roche, Samsung and UCB; advisory board: AstraZeneca; leadership or role in board/society: Editor, Annals of the Rheumatic Diseases, Editor, Rheumatology 8E (Textbook). D.A.: grants: Galapagos, Lilly; speaker’s bureau/consultancy: Abbvie, Gilead, Janssen, Lilly, Merck, Novartis. Peter Mandl: grants: AbbVie, Novartis, UCB; speaker’s bureau: Abbvie, Novartis, Janssen and UCB. The other authors have no potential conflict of interest to declare.

Acknowledgements

This study was conducted in conjunction with the Ludwig Boltzmann Institute for Arthritis and Rehabilitation, Vienna, Austria

References