Do ankle, hindfoot, and heel ultrasound findings predict the symptomatology and quality of life in rheumatoid arthritis patients?

Consecutive RA patients presenting to the day-hospital department were enrolled between April and December 2018 in this observational, cross-sectional study. The inclusion criteria were: diagnosis of RA according to the 2010 American College of Rheumatology (ACR)/EULAR classification criteria⁽¹⁾, stable treatment in the past 3 months, and age over 18 years. The exclusion criteria included: unstable treatment in the past 3 months before enrollment [need for nonsteroidal anti-inflammatory or antalgic drugs, increase or decrease in oral steroid therapy, synthetic or biological disease-modifying anti-rheumatic drugs (DMARD) initiation, switch or combination, or topical administration of steroid drugs], history of foot and ankle surgery, trauma, tendinosis, other causes of architectural changes within the feet and ankles (e.g. diabetes, osteoarthritis), other inflammatory rheumatoid diseases (e.g. crystal arthropathies, spondylarthritis, etc.), and history of cancer with bone metastasis, including suspect pulmonary nodules where further evaluation was indicated.

The group of patients was compared to healthy control (HC) subjects matched for age, sex, and body mass index (BMI). The HC status was defined as asymptomatic subjects with no history of surgery, trauma, inflammatory and non-inflammatory diseases (including osteoarthritis) on the foot and ankle levels, and no architectural changes of the feet and ankles.

All the patients and subjects provided written informed consent before study enrollment. Also, the approval of the University Ethics Committee, number 51, of 22 January 2018, was obtained.

The RA patients were examined by a rheumatologist with 4 years’ experience in rheumatology (IP) following a pre-established protocol. Demographic data and the presence of pain at the level of each ankle, quantified by the Visual Analog Scale (VAS), were recorded. The status of symptomatic ankle was established by the presence of pain at the level of the ankle (patient’s personal assessment) at the time of enrollment in the study. The swollen and tender joints were counted. The postural deformities of the arches of the feet (flat feet, high arched feet), postural changes of the hindfeet (hindfoot valgus or varus), pain caused by palpation of the anterior, medial, lateral and posterior ankles and hindfeet were assessed. The level of motion, and pain caused by passive or active movements of the ankle (dorsal flexion, plantar flexion, inversion, eversion) and hindfoot joints (inversion, eversion), were also recorded.

The disease activity in RA patients was estimated by calculating the Clinical Disease Activity Index (CDAI)⁽¹⁸⁾ and DAS 28 with CRP (DAS28-CRP)⁽¹⁹⁾.

The functionality and the effect of RA on daily life was assessed by applying the Routine Assessment of Patient Index Data 3 (RAPID3) questionnaire⁽²⁰⁾, measuring the physical function, the patient’s assessment of pain (global, at any site), and overall health. The quality of life was evaluated for each patient using the validated Romanian language version of the RA Quality of Life (RAQoL) questionnaire^(21,22) with a score range of 0 to 30 (the higher the score, the lower the QoL).

The plantar pressure measurement was performed by a rheumatologist (CDB) with more than 15 years’ experience, using the P-WALK baropodometric platform (BTS S.p.A., Milan, Italy), calibrated according to the manufacturer’s recommendations. The plantar pressure distribution and the contact surface of the footprints were measured using static analysis, with the barefoot patient standing still in the middle of the platform for 10 seconds, with the feet slightly outlying, forming an angle of about 30º. The Pressure Color Scale for each measurement was preset at 100 kPa. The baropodometric evaluation provided an arch index that allowed the classification of each foot in high arch, normal foot or flat foot, a colored map of the foot contact area, and the maximum and average pressure for the 1^st toe, 2^nd to 5^th toes, 1^st to ^5th metatarsal regions, midfoot (MF), medial heel (MH), and lateral heel (LH). For the purpose of the study, we only used the pressure results obtained at the MF, MH and LH levels.

The US evaluation was performed using GE Logiq S7 ultrasound machine (GE Healthcare, Chicago, USA) with ML6-15 linear transducer (at a frequency adapted to the examined region). In the Power Doppler (PD) mode, the frequency was 9 MHz, the gain was set at a minimal level to avoid noise artifacts, and the Pulse Repetition Frequency (PRF) was set at 0.8 kHz.

The US examination was performed by two examiners with 5 years’ (OS) and 20 years’ (DF) of experience in musculoskeletal US, respectively, in a blinded mode for the patient’s complaints, or for other results. US images were stored for each structure following a pre-established protocol. In the case of disagreements between the two examiners, a third examiner (MCM), with more than 15 years’ experience in musculoskeletal US, reviewed the images.

Following the Outcome Measures in Rheumatology (OMERACT)/EULAR definitions and recommendations^(23-28), the tibiotalar (TT), anterior and posterior approaches, talonavicular (TN) and subtalar (ST), medial, lateral and posterior approaches were examined. The synovial hypertrophy (SH) in gray scale, PD signal, osteophytes and erosions were also identified and graded accordingly.

The tibialis tendons, long flexor and extensor tendons and peroneal tendons of the ankles were evaluated following the same guidelines⁽²³⁾. The presence of tenosynovitis, PD signal and tendinosis were also recorded and graded⁽²⁹⁾.

In the heel region, the Achilles tendon (enthesophytes, retrocalcaneal bursitis and posterior calcaneal erosions) and plantar fascia (plantar fasciitis, perifasciitis, inferior calcaneal erosions) were evaluated according to the previous recommendations^(27,30).

We included the TT joint and tendons in the ankle region, and the TN and ST joints in the hindfoot region. The Achilles tendon, plantar fascia and their adjacent structures were assigned to the heel region.

The sample size was calculated to be 35 for each group at a level of significance of 0.05 and a power of 0.80. For the estimation of sample size, the rearfoot peak plantar pressure was taken into account⁽³¹⁾ with a difference between RA and HC groups of 18 kPa and a standard deviation of 30 kPa, the effect size of 0.6, and also the frequencies of ankle synovitis⁽³²⁾ of 18.3% in the RA group and 2.5% in the HC group, the effect size of 0.6.

The assessment for the normal distribution of continuous variables was performed using the Shapiro-Wilk test. Descriptive statistics were obtained, and the results were presented as numbers (percentages) for the frequencies of the categorical variables, mean ± standard deviation (SD) for normally distributed continuous variables, and median (interquartile range – IQR) for non-normally distributed continuous variables. Chi-square test or Fisher’s exact test (as appropriate) were used to assess the association between categorical variables. The comparison of medians between the two groups was performed using Man-Whitney U test. The inter-observer agreement was assessed using the Cohen’s kappa for dichotomous variables and weighted kappa for ordinal variables. The interpretation of the k coefficient values was as follows: 0–0.20 poor, 0.20– 0.40 fair, 0.40–0.60 moderate, 0.60–0.80 good, and 0.80–1 very good.

Using the univariable and multivariable logistic regression analysis, the unadjusted and adjusted (for the other variables included) odds ratios for the prediction of hindfoot valgus and for the symptomatic ankle were calculated. The predictive variables of RAQoL were established using an ordinal logistic regression model. The variables to be included in the univariable logistic regression analysis were those having clinical relevance. The variables to be incorporated into the multivariable logistic regression models were selected from those that had statistical significance after the univariable logistic regression and using the forward method. P-values under 0.05 were considered statistically significant. The analysis was performed using Microsoft Excel in Office 365 (Microsoft, Redmond, Washington, SUA), and SPSS Statistics v.23 (IBM, Armonk, New York, USA).

Fifty-five RA patients were examined clinically, but only 35 RA patients were enrolled (Fig. 1). After matching for sex (85.7% females), age and BMI, a total of 35 HC individuals were recruited (HC vs RA: 58.91 ± 11.59 vs 59.2 ± 11.25 years old, p = 0.743, and 25.72 ± 3.40 vs 26.49 ± 5.64 kg/m², p = 0.244). The RA patients were split into 2 subgroups depending on the presence/absence of symptoms in the ankle region (Tab. 1).

Fig. 1.

The inter-observer agreement for all the US findings in the joint, tendon and heel structures was very good (kappa: 0.82–0.88), except for the subtalar joint, where the level of agreement was good (kappa: 0.71–0.75).

The US findings for the ankle/hindfoot joint, tendon and heel structures are summarized in Tab. 2, Tab. 3, Tab. 4, and Tab. 5.

The presence of flat foot (37 ankles) was significantly associated with calcaneus valgus (34 ankles) (p <0.001), 31 ankles (83.8%) having both modifications, and with tibialis posterior (TP) tendinosis (p = 0.029).

Considering only the ankles with calcaneus valgus, ST joint synovitis was detected in 20 (58.8%) ankles: 8 (40%) from the medial scan, 3 (15%) from the lateral scan, 6 (30%) from both medial and lateral scans, 1 (5%) from medial and posterior scans, 1 (5%) from lateral and posterior scans, and 1 (5%) from all three scans. Lateral ST synovitis was detected significantly less frequently in patients with hindfoot valgus (32.4% vs 55.6%, p = 0.041) (Fig. 2).

Fig. 2.

Including the relevant parameters in a logistic regression model for the prediction of hindfoot valgus, we found that age (OR = 1.1, p = 0.006), disease duration (OR = 1.02, p = 0.001), and DAS28-CRP (OR = 1.46, p = 0.035) were independent predictors of hindfoot valgus.

The rest of the clinical examination was poorly associated with the US findings. The exceptions included pain in the anterior ankle and TT joint synovitis (p = 0.013), pain in the Achillean region and enthesophytes (p = 0.018), and pain in the inferior heel and plantar fasciitis (p = 0.008) or perifasciitis (p <0.001).

In 47 hindfeet of the RA patients, ST synovitis was identified from the medial approach in 34 (72.3%) hindfeet, from the lateral approach in 31 (66%), and from the posterior approach in 5 (10.6%). In 15 (31.91%) hindfeet, ST synovitis was visualized both from the medial and lateral approaches, and in 3 (6.38%) hindfeet from all the three approaches.

Out of the 35 RA patients, 14 (40%) were symptomatic, with median (IQR – interquartile range) VAS of 5 (3–7). The distribution of US findings in the asymptomatic and symptomatic ankle groups is detailed in Tab. 2 Tab. 3, Tab. 4 and Tab. 5.

The results of the unadjusted univariable and adjusted multivariable logistic regression for the prediction of the symptomatic ankle are listed in Tab. 6.

The maximum and mean plantar pressures and contact surfaces in the HC subjects and RA patients (with or without ST synovitis) are shown in supplementary Tab. 7. Significant differences were found in the medial and lateral plantar pressures in the RA patients (with or without ST joint synovitis) (Tab. 8), but in no relation to the US changes in other joints, tendons or heels.

The maximum and mean midfoot pressures were higher in the symptomatic ankle [72 (32.5–96.5) vs 49 (16–129) kPa, p = 0.043 and 25 (18.5–39.5) vs 18 (9–59), p = 0.040, respectively], and in hindfoot valgus patients [94.5 (29–162) vs 25 (15–95.5), p = 0.046]. In flat feet, medians of MF maximum and mean pressures [72 (32.5–96.5) vs 49 (16–129) kPa, p = 0.043, and 40 (16–60.5) vs 10 (7–25) kPa, p = 0.002, respectively], MF contact surface [24.15 (9.9–31.55) vs 6.4 (2.7–22.4) cm², p = 0.022], MH maximum and mean plantar pressures [140 (65.5–216) vs 50 (34–144) kPa, p = 0.008, and 76 (31–85.5) vs 24 (16–75) kPa, p = 0.005, respectively], LH maximum and mean pressures [136 (64–210.5) vs 52 (33–136) kPa, p = 0.0010, and 71 (30–82.5) vs 25 (15–73) kPa, p = 0.010, respectively] and total maximum pressures [162.4 (106.7–267.1) vs 52.6 (35.4– 197.3), p = 0.006] were higher comparing with no-flat foot RA patients.

The RAQoL score of RA patients with symptomatic ankle was significantly higher compared to the asymptomatic ankle group (p = 0.037). Including all the relevant variables (clinical, US and baropodometric findings) in the multivariable ordinal logistic regression model for the prediction of the RAQoL score, the DAS28-CRP (p <0.001), RAPID3 score (p <0.001), disease stages (p = 0.016), hindfoot valgus (p = 0.010), ankle joint synovitis [TT synovitis (p <0.001), ST synovitis (p <0.001)], ankle tendon pathology [TP PD positive tenosynovitis (p = 0.029), TP tendinosis (p = 0.005), peroneal tendinosis (p <0.001)], Achilles tendon enthesophytes (p <0.001), posterior calcaneal erosions (p = 0.011), plantar fasciitis (p = 0.04), and perifasciitis (p = 0.007) were found to be independent predictors.