Objective To investigate the association between clinical and ultrasonographic (US) evidence of inflammation in psoriatic arthritis (PsA), as well as to compare clinical and US remission criteria.
Methods In this cross-sectional study 141 PsA outpatients were included. Minimal disease activity (MDA), 28-joint Disease Activity Score (DAS28), Disease Activity Index for PSoriatic Arthritis (DAPSA) and modified versions of Composite Psoriatic Disease Activity Index (CPDAI) and Psoriatic ArthritiS Disease Activity Score (PASDAS) were assessed. Remission criteria were explored. US evaluation was performed on 34 joints, in addition to joints being tender/swollen by 66/68 joint count, 30 tendons, 10 entheses and additionally entheses found to be tender by clinical examination of 19 other entheses. Power Doppler (PD) and grey scale global scores on joints, entheses and tendons were assessed. US remission was defined as no PD activity in joints, entheses and tendons.
Results DAPSA and DAS28, but not CPDAI and PASDAS, were associated with PD activity. MDA was fulfilled in 22.7% and the clinical remission criteria in 5.7%–9.9% of the patients. US remission was found in 49.6% of the patients. The prevalence of PD activity at joints, entheses and tendons was similar for patients fulfilling versus not fulfilling MDA/clinical remission criteria. MDA (OR 2.3, p=0.048), DAPSA ≤3.3 (OR 4.2, p=0.025) and Boolean's (OR=7.8, p=0.033) definitions of remission were found to predict US remission.
Conclusions We found major discrepancies between US and clinical findings. DAPSA and DAS28 reflected US findings better than CPDAI and PASDAS. MDA, DAPSA and Boolean's remission criteria predicted US remission.
- Psoriatic Arthritis
- Disease Activity
Statistics from Altmetric.com
Clinical joint examination in psoriatic arthritis (PsA) may be challenging.1 In clinical practice the use of composite scores validated only for rheumatoid arthritis (RA) are widely used in PsA.2 The need for PsA specific composite scores has been emphasised by the Group for Research and Assessment of Psoriasis and Psoriatic Arthritis.3 In recent years, several methods for scoring have been developed, and among them are the Disease Activity index for PSoriatic Arthritis (DAPSA),4 Composite Psoriatic Disease Activity Index (CPDAI)5 and the Psoriatic ArthritiS Disease Activity Score (PASDAS).6 CPDAI is a five-component score including involvement of joints, entheses, dactylitis, axial skeleton and skin, and PASDAS includes joints, entheses and dactylitis, whereas DAPSA is strictly joint focused.
During the recent years the use of ultrasound (US) evaluation in ordinary clinical practice has become more common, and US evaluation of arthritis including PsA has shown to be more sensitive in detecting arthritis than clinical examination.7 ,8
The associations between clinical and US findings have been well explored in RA, and are mostly described as weak to moderate.9 ,10 The associations between clinical and US findings of arthritis, tenosynovitis and enthesitis in PsA remain much less investigated to date. One recent study, including CPDAI and DAPSA, described disparity between US and clinical findings.11 Studies comparing both RA and PsA specific composite scores as well as PASDAS with US findings in PsA are lacking in the current literature.
To date there are no validated remission criteria for PsA, but criteria for minimal disease activity (MDA) in PsA have been established.12 Several remission criteria in PsA have been proposed.11 ,13 ,14 Similar to RA, there is consensus that a treat-to-target approach in PsA should result in better disease outcome.15
The present objectives were to investigate the association between clinical and US evidence of inflammation in consecutive patients with PsA from the outpatient clinic. Further, we compared clinical and ultrasonographic remission criteria. As part of the clinical evaluation we included DAPSA, modified versions of CPDAI and PASDAS as well as composite scores initially developed for RA.
The study size was specified at the beginning of the study. Assuming a correlation between clinical and ultrasonographic findings of 0.25, power 80% and significance 5%, we needed at least 124 patients. In total, there were 581 patients with PsA registered in the outpatient clinic of the Hospital of Southern Norway Trust, Norway, during the study period from January 2013 to May 2014, of whom 471 fulfilled the ClASsification for Psoriatic ARthritis criteria.16 Of these 471 patients 141 were included in the study in a random manner at consecutive clinic visits. All the included patients had peripheral inflammatory involvement clinically.16 A comparison of the characteristics of the included and non-included patients is listed in online supplementary table S1. The study was approved by the Norwegian Regional Committees for Medical and Health Research Ethics (Regional komité for Medisinsk og helsefaglig forskningsetikk Midt-Norge 2012/101) and written informed consent was obtained from each patient. The patients performed registration of demographics and patient-reported outcome measures in a computer system (GoTreatIT),17 including Modified Health Assessment Questionnaire (MHAQ, range 0–3),18 Bath Ankylosing Spondylitis Disease Activity Index (BASDAI, range 0–10),19 Dermatology Life Quality Index (DLQI, range 0–30)20 and Rheumatoid Arthritis Impact of Disease (RAID, range 0–10).21 C reactive protein (CRP, mg/L) and erythrocyte sedimentation rate (ESR, mm/h) were assessed. Two specially trained nurses, unaware of the US results, performed 66/68 joint count, tender dactylitis count, “Maastricht Ankylosing Spondylitis Enthesitis Score”22 (MASES, range 0–13, including first and seventh costosternal joints, anterior superior iliac spine, iliac crest, fifth lumbar spinous process, posterior superior iliac spine, Achilles) and evaluation of additionally 16 other entheses (lateral and medial epicondyle, triceps, great trochanter, quadriceps, proximal and distal patellar tendons, plantar fasciae) and Psoriasis Area Severity Index (PASI, range 0–72).23 The presence of enthesitis was defined as tenderness on firm palpation. Visual analogue scales for evaluator's global assessment of disease activity (EGA) and patient's global assessment of disease activity (PGA), pain and joint pain were recorded (range 0–100 mm). The presence of erosions on X-rays of hands and feet was assessed by a radiologist as part of general care.
Simple Disease Activity Index (SDAI),24 disease activity score for 28 joints (DAS28)25 and DAPSA score were calculated. Modified versions of PASDAS (including EGA, PGA, MHAQ, SJC66, TJC68, MASES, dactylitis count, CRP) and CPDAI (including 66/68 joint count, MHAQ, PASI, DLQI, MASES, dactylitis count, BASDAI, RAID), including CPDAI joint, entheses and dactylitis domains (CPDAI-JED), were calculated (see online supplementary table S2). In the PASDAS and CPDAI calculations we used MHAQ instead of HAQ26/SF36-PCS (Physical Component Summary score of the Medical Outcome Survey Short Form-36),27 RAID instead of AsQol (Ankylosing Spondylitis Quality of Life)28 and MASES instead of Leeds enthesitis index, as MHAQ, RAID and MASES but not HAQ, SF36-PCS, AsQol or Leeds enthesitis index29 were assessed in this study.
The patients were classified as achieving MDA12 when meeting five of the seven following criteria: TJ≤1, SJ≤1, PASI≤1, pain≤15, PGA≤20, MHAQ≤0.5 and MASES≤1. The following suggested criteria for clinical remission were assessed: (1) DAPSA≤3.3,11 (2) CPDAI<213 or CPDAI joint, entheses and dactylitis domains (CPDAI-JED)=0,11 (3) PASDAS < 2.4,13 (4) a Boolean's definition of remission modified for PsA,11 meeting all of the following criteria: TJ≤1, SJ≤1, MASES≤1, dactylitis count ≤1, EGA≤10, PGA≤10 and CRP≤1 mg/dL.
For the complete US protocol see the online supplementary files, including overview of US assessments in online supplementary table S3 and examples of characteristic US findings in online supplementary figure S1. US evaluation was performed by a rheumatologist experienced in US (APD) on the day of clinical examination. The duration of each US examination was 45 min. The sonographer was aware of the clinical results. Grey scale (GS) and power Doppler (PD) sonography were performed at 34 mandatory joints (bilateral metacarpo-phalangeal 1–5, radio-carpal, inter-carpal, radio-ulnar, knees, talo-crural, subtalar, talo-navicular, metatarso-phalangeal 1–5) and additional joints found to be swollen or tender by 66/68 joint count; 10 mandatory entheses (bilateral quadriceps, proximal and distal patellar tendons, Achilles, plantar fasciae) and additionally entheses found to be tender by clinical examination of 19 other entheses (first and seventh costosternal joints, anterior superior iliac spine, iliac crest, fifth lumbar spinous process, posterior superior iliac spine, lateral and medial epicondyle, triceps, great trochanter); 30 tendons were examined for tenosynovitis (bilateral finger flexor digiti 1–5, flexor (digitorum superficialis and profundus) and extensor (compartments 1–6) groups in the wrists, medial, anterior and lateral tendon sheets in the ankle, flexor digiti pedis 1–5). The US examinations were performed using two US devices (Siemens Acuson S2000 and GE logic E) with two multifrequency linear transducers (6–18 MHz and 4–10 MHz), as it is standard practice in the clinic to share US machines between physicians. A head-to-head comparison of the two US devices showed similar results including GS and PD synovitis, enthesitis and tenosynovitis (kappa with linear weighting 0.83–1.0). GS and PD signals in joints and tendons were semiquantitatively graded 0–3.30–34 GS and PD sum scores for joints (both range 0–228) and tendons (both range 0–90) were calculated. The Global OMERACT-European League Against Rheumatism Synovitis Score (GLOESS)35 was additionally used to score synovitis. The number of entheses with enthesitis defined as GS structural pathology (thickening, hypoechogenicity, fibrillar separation, calcifications, erosions or enthesophytes) and/or PD signal at the entheses were registered (scored absent/present, both range 0–29).34 ,36 GS and PD sum scores on joints, entheses and tendons (both range 0–347) were calculated. To date no US remission criteria in PsA has been established. We defined US remission as no PD activity in all examined joints, entheses and tendons, in accordance with the previously proposed definition of US remission in RA33 ,37 and PsA.11
Statistical analyses were performed using IBM SPSS Statistics (V.220.127.116.11). Descriptive statistics were used to calculate patients' demographic variables. The median and range were calculated for non-parametric data and the mean and SD for parametric data. Proportions were analysed using χ2 test or Fisher's exact test with mid-p correction as appropriate. Quantitative results were compared using the Mann–Whitney U-test and correlation analyses were performed by Spearman's rank correlation test (non-parametric distribution of the data). Univariable and multivariable logistic regressions were used to examine the possible association between clinical factors (Pain, ESR, CRP, SJC66, TJC68, MHAQ, MASES, EGA, PGA) and US remission. A significance level of 0.1 was set as discriminatory for included variables. Hosmer–Lemeshow statistics was performed for testing of goodness of fit.
Patients' demographic characteristics are summarised in table 1.
Significantly more patients with bDMARD treatment were in US remission, compared with patients without bDMARD treatment. No such difference was found for csDMARD and prednisolone treatment (see online supplementary table S4).
Patient-reported outcome measures, inflammatory markers and composite scores of disease activity are listed in online supplementary table S5. CPDAI categories are listed in online supplementary table S6.
Associations between clinical and US findings
Of the PsA specific composite scores, only DAPSA was associated to the PD global score. Similar weak correlations were found between the PD global score and DAS28ESR(4) and SDAI. The GLOESS showed the same correlations to clinical findings as the GS arthritis score. In online supplementary table S8 the findings for patients with discrepancies between US examination and 66/68 joint count are further explored.
Clinical remission criteria versus US pathology
The MDA criterion was fulfilled by 22.7% of the patients. Of the remission criteria DAPSA≤3.3 was fulfilled by 9.9%, CPDAI<2 by 8.6%, CPDAI-JED domain=0 by 7.8%, PASDAS<2.4 by 7.8% and Boolean's by 5.7% of the patients.
As shown in table 3 the PD global score was higher in patients with active versus inactive disease according to DAPSA and Boolean's definitions of remission and the GS sum score was higher in patients with active versus inactive disease according to CPDAI-JED definition of remission.
In online supplementary table S9, US findings of arthritis, enthesitis and tenosynovitis were explored for patients with active versus inactive disease according to clinical remission/MDA criteria. The GS entheses score was higher for patients with active versus inactive disease according to DAPSA and CPDAI-JED. The other US scores were similar for patients fulfilling versus not fulfilling the clinical remission criteria.
US remission criteria versus clinical remission criteria
US remission with PD=0 in all joints, entheses and tendons was found in 70 (49.6%) patients. GLOESS=0 was found in 35 (24.8%) patients. Only three (2.1%) patients had PD=0 and GS=0 for joints, entheses and tendons.
The prevalence of patients with PD score>0 at joints, entheses and tendons was similar for patients fulfilling clinical remission/MDA criteria compared with patients not fulfilling these criteria (see online supplementary table S10). The MDA, DAPSA and Boolean's PsA remission criteria were useful to identify patients in US remission according to PD signal=0 at joints, entheses and tendons (table 4). The specificity of these indices was good, but the sensitivity poor.
Association between clinical factors and ultrasound remission
The possible association between clinical factors and US remission with PD=0 on joints, entheses and tendons was explored with pairwise comparisons and univariable and multivariable logistic regression analyses.
MASES, MHAQ, CRP and ESR were similar in the US activity and US remission groups. TJC68, SJC66, pain, PGA and EGA tended to be lower in the US remission group compared with the US activity group (see online supplementary table S11). By means of multivariable logistic regression analysis, only SJC66 was associated to US remission (OR 0.51, 95% CI 0.34 to 0.79, p=0.002). Correction of the model for age, gender, disease duration, body mass index, current use of biologics and steroids did not change the primary outcome (OR 0.50, 95% CI 0.32 to 0.79, p=0.003).
We found poor correlations between clinical and US evidence of joint inflammation in PsA. However, the correlations between clinical composite scores and US findings improved to some extent when evaluating entheses and tendons in addition to joints. This underlines the heterogeneity of inflammation in PsA, often involving joints and periarticular structures. To date there is no consensus on the US definitions of active versus chronic enthesitis changes, and often both will be present.36 A high prevalence of subclinical enthesitis in patients with PsA has been described.38 ,39 It is important to be aware that some US findings may be permanent changes, for example, enthesophytes, erosions and calcifications at the entheses. Enthesitis is demonstrated to be seen in a majority of patients with spondyloarthropathy, but not in controls.40 However, the extent to which healthy controls may have inflammatory findings on US assessment is poorly investigated. Witt et al41 reported recently grade 1 GS findings in the joints of healthy controls.
In RA, subclinical arthritis is found to be of negative prognostic value,42 and whether this is also valid in PsA remains to be explored.
It is important to keep in mind that US is prone to operator influence and experience. The interobserver reliability of US findings in PsA is reported as fair to excellent,10 ,43 and can be improved by standardisation of US examinations.44
The PD global score was weakly correlated to both composite scores developed for RA, but frequently used also for PsA (DAS28ESR, SDAI) as well as to DAPSA, specifically developed for PsA. DAS28, SDAI and DAPSA are heavily peripheral joint focused, but did not perform better when compared with the PD joints score or the GLOESS. The PsA specific composite scores PASDAS and CPDAI-JED domains did not correlate significantly to US activity, although incorporating arthritis, enthesitis and dactylitis, in contrast to DAPSA, DAS28 and SDAI, which only incorporate arthritis. None of the composite scores were found to adequately reflect US findings. Similar results were described in a recent study by Husic et al,11 exploring DAPSA and CPDAI, but not DAS28, SDAI or PASDAS.
In contrast to MDA, which is validated for PsA, remission criteria in PsA have not yet been established. However, cut-off values for clinical remission have been proposed, but not validated for the PsA specific composite scores; DAPSA≤3.3,11 CPDAI-JED=011 and PASDAS<2.4.13 We found the MDA, DAPSA and Boolean's remission criteria somewhat useful for discriminating between patients with and without US activity, which is in line with Husic et al11 reporting DAPSA and Boolean's remission definitions useful for discriminating between patients with and without minimal US activity. In both studies the specificity of the remission criteria was good, but the sensitivity poor.
To date no US remission criteria in PsA have been established. We defined US remission as the absence of PD signals at joints, entheses and tendons in accordance with previous studies in RA33 ,37 and PsA.11 About half of our patients were in US remission according to this definition. We also explored the newly developed GLOESS, previously examined only in RA, which did not perform better than the PD joints score. Interestingly, the GLOESS in our population was very similar to the GS joints score, as the GLOESS is more weighted after GS than PD findings and most of the patients in our population had higher GS score compared with PD score in individual joints.
ESR and CRP are regularly used in the clinic to monitor disease activity in inflammatory arthritides. Both were weakly correlated to the PD score on joints, which is consistent with the previous findings for ESR11 ,45 ,46 and CRP45 in PsA.
US is an important, sensitive tool to detect inflammation in articular and periarticular structures. Our study underlines the value of US examination in addition to clinical examination in the evaluation of disease activity in PsA. It also underlines the need for further development of both clinical and ultrasonographic PsA specific composite scores. According to our findings, evaluator's global assessment and 66 swollen joint count are the clinical scores best correlated to US findings and ought to be included in further development of PsA specific composite scores.
Recently, specific US composite scores for treatment monitoring in PsA have been proposed.43 ,47 Interestingly, these scores were shown to have only weak to moderate correlations to clinical findings, which is in accordance with our study.
Our study has several strengths. The basis of the study was a comprehensive clinical and ultrasonographic examination of joints, but also periarticular structures, which gave a good basis for comparison between clinical and US findings. In addition, we compared US findings with recently developed PsA specific composite scores, as well as clinical composite scores developed for RA, but frequently used in the clinic also for PsA. Similar comparison between PsA specific composite scores and a thorough US examination has to our knowledge only been reported in one previous study in a smaller patient sample, excluding composite scores initially developed for RA.11 Further, evaluation of the discriminatory capacity of MDA and newly proposed PsA remission criteria to predict US remission is previously explored only in one smaller study, which did not include PASDAS.11 Finally, our study is based on ‘real-life’ data, as consecutive patients from the outpatient clinic were included.
Limitations of the study include the relatively low disease activity of the patients. This is to be expected in a clinic where the goal is to get patients to the lowest possible disease activity, but it may impact the results of assessments such as this, because the strength of association between US and clinical findings in RA is demonstrated to diminish as patients are in or near to remission.33 Another limitation of the study is that the ultrasonographer was aware of the clinical findings, which may have biased the results. Further, the modifications of CPDAI and PASDAS used in the study are not validated and represent a limitation of the generalisability of the results.
In conclusion, we report discrepancy between clinical and US evidence of inflammation in PsA. DAPSA and DAS28 reflected US findings better than CPDAI and PASDAS. The MDA, DAPSA and Boolean's definitions of remission predicted US remission. Further advances in US, with development and validation of US composite scores including joints and periarticular structures, may improve sensitivity and specificity for diagnosis and assessment in PsA.
The authors thank the patients for participating in this study and the local rheumatology staff for data collection.
Handling editor Gerd R Burmester
Contributors GH, DMS, APD, HBH, AK and BM were responsible for study design. BM, APD, DMS and GH were responsible for data acquisition. BM analysed the data and wrote the manuscript. All authors critically revised the manuscript and approved the final version.
Funding Unrestricted grant from Pfizer ID WS1948310 (GH). Clinical research fellowship from the Hospital of Southern Norway Trust (BM).
Competing interests APD: advisory board for Novartis and Pfizer, speaker fees from Abbvie; HBH: honorary fee from Abbvie, Pfizer, UCB, Roche; GH: founder and shareholder in DiaGraphIT. Unrestricted grant from Pfizer Norway.
Ethics approval Norwegian Regional Committees for Medical and Health Research Ethics Midt-Norge.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement All important data are within the text.
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.