Ultrasound verified inflammation and structural damage in patients with hereditary haemochromatosis-related arthropathy

Patients with HH and hand OA (HOA) were recruited from clinical routine practice. HH was defined by a homozygous C282Y mutation or a compound mutation of the C282Y and H63D genes plus serological signs of iron overload at diagnosis: increased transferrin saturation (> 55% in men, > 45% in women) and either provisional (serum ferritin > 300 ng/ml for men and postmenopausal women, > 200 ng/ml for premenopausal women) or overt iron overload (serum ferritin > 1000 ng/ml or hepatic iron overload on biopsy) independent of the presence of clinical symptoms related to iron overload [9].

HH-A was defined as HH plus pain in the hands (VAS > 10 mm and/or ≥ 1 tender joint) plus at least one radiographic change compatible with HH-A on hand x-rays [9, 12]. Two patients with HH who declined to undergo x-rays were classified on a clinical basis only; both patients were considered to have HH-WA.

Patients with HOA were classified using the criteria of the American College of Rheumatology [13]. All patients with HOA had a negative medical and family history for HH and they were tested for the absence of serological signs of iron overload.

All patients underwent a structured history/chart review retrieving demographic data, medical history and current medication. Patients with HH were further questioned for the duration and frequency of phlebotomies, previous joint replacement and HH-related co-morbidities including hepatopathy, hypogonadism, cardiomyopathy, skin changes and diabetes mellitus.

Clinical assessments included the duration of morning stiffness (minutes), number of tender joints (TJ) (68-joint count), swollen joints (SJ) (66-joint count), number of bony swollen joints (assessed at the MCPs, proximal (PIP) and distal interphalangeal joints (DIP), wrists, knees, and ankles), patient’s (PGA) and evaluator’s global assessment of disease activity (EGA), hand function and pain related to hands, hips, knees and ankles. Function, global and pain assessments were all recorded on a 100-mm VAS, with 0 = best, 100 = worst [9, 12].

Blood investigations included erythrocyte sedimentation rate (ESR, measured by the Westergren method) and C-reactive protein (CRP) (measured by nephelometry), blood cell count, liver enzymes, renal functional tests and iron status (serum ferritin, transferrin and transferrin-saturation). The results of the HFE gene mutation (C282Y and H63D) tests were also retrieved in order to classify the patients.

Grey scale (GS) and power Doppler (PD) sonography were performed in 36 joints (wrists, MCPs, PIPs, DIPs, hips, knees and ankles) by one of two rheumatologists (CDe and CDu) who were unaware of the clinical findings (but not blinded to diagnosis of HH versus HOA). We used a MyLab Twice ultrasound device (Esaote, Genova, Italy) with two multi-frequence linear transducers (6–18 MHz, small and medium joints; 4–13 MHz, large joints). For GS imaging, parameters were adjusted to maximize the contrast between examined structures. PD settings were standardized accordingly: frequency 9.1 (small joints and entheses) or 6.3 (large joints) MHz, pulse repetition frequency 750 Hz and medium persistence. The PD gain was optimized by increasing the gain until noise appeared and then reducing it just enough to suppress the noise [14].

All ultrasonography comprised longitudinal and transverse scans in accordance with current guidelines and publications [15]. GSS was subjectively graded from 0 to 3 in which 0 represented no GSS, 1 = minimal, 2 = moderate and 3 = extensive GSS as defined in recent publications [14, 16‐18]. PD signals in large and small joints were also semi-quantitatively assessed on a scale of 0–3 with 0 = no PD-signal, 1 = up to three single or two confluent vessels, 2 = less than half of the synovia and 3 = half or more of the synovia covered by PD signals [14, 16‐18]. GSS and PD were independently graded on each view of each joint (e.g. palmar and dorsal) and the highest value for each joint was counted for the sum score. Tenosynovitis was scored as reported previously [14, 19]: GS tenosynovitis (GS-teno) was identified as hypoechoic or anechoic thickened tissue with or without fluid within the tendon sheath and was semi-quantitatively graded from 0 to 3 at the wrists and ankles, whereas at the level of MCPs, PIPs and DIPs it was recorded as 0 = absent or 1 = present. Tenosynovitis affecting the tendon at the level of phalangeal bones was assigned to the nearer joint (e.g. flexor tenosynovitis at the PIP bone was assigned either to the MCP or the PIP depending on which joint was closer). Tenosynovitis covering > 50% of the area of a phalangeal bone was scored for both related joints. At wrists, the extensor and flexor tendons were independently scored using the following grading: 0 = no tenosynovitis, 1 = tenosynovitis at one, 2 = at two and 3 = at three or more tendon compartments. At the ankles, only the extensor tendons were assessed and graded from 0 to 3, with 0 = no tendon, 1 = one tendon, 2 = two tendons and 3 = all three extensor tendons were involved. PD signals related to tenosynovitis (PD-teno) were graded from 0 to 3 (0 = no PD signal, 1 = up to three single or two confluent vessels, 2 = less than half of the tendon/tenosynovia and 3 = half or more of the area covered by PD signals). At the wrists and ankles the tendon with the maximal PD score was counted.

Erosions (assessed at MCPs, PIPs and DIPs only) or osteophytes (assessed at MCPs, PIPs, DIPs, hips and knees) were defined by a step-down or step-up contour defect, respectively that is visible in 2 perpendicular planes [20]. At the MCPs, PIPs and DIPs, erosions and osteophytes were independently assessed on dorsal and palmar views; lesions visible in the lateral compartments (e.g. erosions at the lateral aspects of the 2^nd or 5^th MCP joints) were counted as dorsal lesions. Osteophytes were further investigated by anterior scans of the hips and at the medial and lateral femorotibial spaces at the knees [15]. Grading of erosions was conducted from 0 to 3 as described previously [14] and was based on the maximum diameter of the cortical break for bone erosions (adapted from [21]) with grade 0 = no erosion, grade 1 = erosion ≤ 1 mm, grade 2 = erosion > 1 mm and ≤ 2 mm, grade 3 = erosion > 2 mm and/or large destruction of the joint. In the case of multiple erosions the largest lesion was counted. For osteophytes the maximum distance between the “original” and new cortical lining (= maximal height) was measured: grade 0 = no osteophyte, grade 1 = osteophyte ≤ 1 mm, grade 2 = osteophyte > 1 mm and ≤ 2 mm, grade 3 = osteophyte > 2 mm and/or large and diffuse osteophytes (adapted from [21]). Erosions and osteophytes were independently graded on each view of each joint and the highest value was counted for the sum score.

Cartilage was assessed at MCPs 2–5 on longitudinal scans with full flexion of the fingers, and at the knees on transverse scans with full flexion of the knees. Scoring was adapted from Filippucci et al. with a grading from 0 to 4, with 0 = no evidence of cartilage abnormalities, 1 = loss of sharpness of the superficial margin of the hyaline cartilage, 2 = partial thickness defect of the cartilage layer, 3 = full thickness defect of the cartilage layer with a normal subchondral bone profile and 4 = complete loss of the cartilage layer and subchondral bone involvement. CPPD deposits were defined as previously described and graded as 0 = absent or 1 = present [22]. Sum scores were calculated for GSS (range 0–108), PD (0 − 108), osteophytes (0–96), erosions (0–84), GS-teno (0–40), PD-teno (0–96) and cartilage (0–40).

Patients with HH underwent standard radiography of the hand and wrist, knee and ankle joints within 2 weeks of the study visit. All radiographs were assessed by one of two experienced radiologists (VT or SZ) and scored from 0 (no damage) to 6 (extensive damage) per joint as previously described [9]. The total score ranged from 0 to 72.

Statistical analysis was performed using SPSS (version 23.0). Descriptive statistics were used to summarize the data. The distribution of data was tested for normality using the Kolmogorov-Smirnov test. For continuous non-parametric data, we show the median and range whereas for parametric data, the mean and standard deviation are depicted. Comparisons between independent groups were conducted using the Mann-Whitney U test. Paired categorical data were analysed using the chi square test or Fisher’s exact test as appropriate. P values were not corrected for multiple testing.

Inter-rater reliability of the ultrasound results was determined by serial blinded assessments of 10% of patients’ scans by two investigators (CDe and CDu) and using the intra-class correlation coefficient (ICC).