Faculty Bibliography

The use of musculoskeletal ultrasound (MSKUS) in rheumatology practice and research has increased steadily over the last decade. An ever-growing body of literature shows parity and even superiority of MSKUS when compared to physical examination, plain radiography, and more expensive and static imaging modalities such as MRI. While many use the modality for procedure guidance, investigators continue to demonstrate its ability to impact diagnoses in a variety of rheumatic diseases. Initial efforts focused on establishing MSKUS as a helpful tool for rheumatoid arthritis (RA), especially in the detection of synovitis and joint erosions, but numerous studies are validating the use of MSKUS as a helpful diagnostic tool for the spondyloarthropathies, crystal diseases, osteoarthritis, and other rheumatic diseases. Advances in ultrasound technology are translating into more sensitive and accurate studies. Within the research community, current efforts aim at maximizing the direct clinical impact of MSKUS by developing global or patient level assessments and simplified joint scoring systems, with improvements in intra- and inter-reader reproducibility

Fewer United States rheumatologists perform or utilize musculoskeletal ultrasound (MSUS) than those in Europe, though this disparity is narrowing. To document perceptions and use of MSUS in the U.S. rheumatology community, we sent an anonymous electronic survey to American College of Rheumatology (ACR) physicians and tailored versions to fellows and program directors. A separately-conducted survey was sent to a smaller group of rheumatologists already utilizing MSUS. Acknowledging survey bias, we found that 20% of rheumatologists and fellows who responded are utilizing MSUS, and those using it primarily do so for diagnosis and injection guidance. Many rheumatologists across the country think that ultrasound should become a standard tool in rheumatology training, practice, and research. Despite an inherent survey bias likely overstating interest in MSUS, this study is valuable as the first to document this trend among U.S. rheumatologists

Background: Determination of monosodium urate (MSU) deposition in joints by musculoskeletal ultrasound (MSK-US) could have implications for uric acid (UA) management in patients with gout and possibly asymptomatic hyperuricemia (AH). Recently, criteria for sonographic diagnosis of MSU crystal deposition have been developed, but reproducibility of readings using these criteria has not been well established. Methods: We consecutively recruited male patients ages 55-85 during primary care visits to an urban VA hospital. We assessed all patients for gout by ACR criteria, and obtained serum UA levels. Patients were divided into 3 groups: gout, AH (no gout, UA >= 6.9 mg/dL), and controls (no gout, UA <= 6.8 mg/dL). 50 patients (14 with gout, 17 with AH, and 19 controls) returned for subsequent evaluation which included MSK-US of knees and 1st metatarsalphalangeal (MTP) joints to evaluate for the double contour sign (knees) and tophi (MTPs). All images were read blindly by two observers trained in rheumatology and MSK-US. Kappa statistics were used to estimate the amount of agreement between ultrasound measures scored by the two raters. We also calculated the total percent of observations in agreement. Results: Evidence of MSU crystal deposition was found in the same 10 patients by both observers (6 gout, 3 AH, 1 control), and in 3 additional patients by one of the observers (1 gout, 2 AH). These findings were further analyzed by site. MSU crystal deposition was identified in a total of 14 common joints by both observers, and in 4 additional joints by the first observer and 6 additional joints by the second observer. Percentage agreement and kappa statistics for our three primary ultrasound measures were as follows; total joints (n=200, 95% agreement, kappa 0.709), femoral articular cartilage (n=100, 95% agreement, kappa 0.679) and 1st MTPs (n=100, 95% agreement, kappa 0.734). Additional analyses by left and right side are shown in the table below. Ratings on only 10 out of 200 joints were in disagreement. (Table Presented) Conclusions: Both percentage agreement and agreement beyond chance between the raters (as estimated by kappa statistics) were very high for the three ultrasound measures. These findings support the use of MSK-US as a reliable modality for detecting MSU deposition. Since MSU deposition is an indication for urate lowering, this type of imaging could be performed noninvasively at the bedside or in the clinic to help direct therapy in gout patients, with possible implications for treatment in AH patients as well should these findings be reproducible in larger cohorts

Purpose: There is insufficient understanding regarding how generalized OA involving the hand and knee differs from isolated knee OA, which may result from other factors such as obesity or trauma. The purpose of these studies is to determine whether the presence of hand OA involving interphalangeal (IP) and first carpometacarpal (CMC) joints, alone or in combination, predicts progression of patients with symptomatic knee OA. Methods: Hand radiographs were obtained on 94 patients at NYUHJD who met ACR criteria for symptomatic knee OA, and who were enrolled in a two-year NIH-sponsored prospective study. The patients completed standardized fixed-flexion knee radiographs at baseline and 24 months, with progression the signal (more painful) knee OA determined by change in joint space width (JSW) and KL score. For these analyses, the patients were separated into two groups by results on their signal knee: 17 progressors, defined by at least 30% decreased JSW over 24 months, and 77 non-progressors. For each set of hand x-rays, 2 radiologists evaluated 18 IP joints and 2 CMC joints for joint space narrowing and/or osteophytes, and whether or not there was erosive change at the IP joints; we averaged the scores from the two readers. Results: Kappa scores between the two scoring radiologists for the IP and CMC joints, and for the presence of erosive IP disease, were 0.79, 0.87 and 0.96, respectively. The overall mean IP score was 5.6 and 1st CMC score was 0.9, while medians were 5 and 1.0, respectively. The 17 progressors had a higher average IP (but not CMC) score than the non-progressors, 7.2+/- 5.4 vs. 5.0+/-4.6, p=0.13. Since the IP scores were not normally distributed, we further analyzed data by dichotomizing the study populations into two groups using the median IP total (5) as the cutoff point. When so analyzed, the presence of hand OA increased the odds ratio of knee OA progression to 2.8 (p=0.096). Of interest, the severity of knee OA correlated with hand OA scores: the average total hand OA scores (out of 20 joints) increased with baseline KL score, with mean scores of 3.8+/-5.5, 6.1+/-6.1 and 7.2+/-5.6 for KL 1 to 3 (p=0.06). There is also an increasing trend of total hand OA joint scores by KL score (p=0.042) when dichotomized around the median (5 joints), and with IP scores alone (p=0.026). The 8 patients with radiographic evidence of erosive IP disease, as compared with the 31 non-erosive IP OA patients (>5 IP joints) and the 54 without IP OA, demonstrated faster knee OA progression over 2 years by average KL increases (1.00, 0.35, 0.30) and decreases in joint space width (0.65, 0.56, 0.36), although perhaps given small numbers, this was not statistically significant (p=0.839). Conclusions: In cross-sectional analysis, the quantitative burden of hand OA correlates with the radiographic severity of knee OA (KL). Moreover, radiographic hand OA at the IP joints, but not at the 1st CMC joint, predicts more rapid progression of knee OA. Erosive IP disease may be an even stronger predictor than non-erosive IP disease of accelerated progression of knee OA

Purpose: We have previously shown that carriage of an IL1RN haplotype (CTA) was associated with substantially lower odds of radiographic severity (KL score, joint space width [JSW]) (Ann Rheum Dis. 2010). In this 24 month prospective study we assessed whether IL1-RN haplotypes predicted disease progression in patients with symptomatic knee OA. Methods: Ninety-seven (N=97) patients from NYUHJD who met ACR criteria for symptomatic knee OA were genotyped for single nucleotide polymorphisms (SNPs) in the IL-1b and IL-1RN genes. Standardized fixed-flexion radiographs were taken on all patients at baseline and 24 months. Radiographic progression of signal (more painful) knee OA was determined by change in JSW over 24 months. To account for variations in baseline JSW, we defined progression as greater than 30% joint space narrowing (JSN) of the diseased compartment over 24 months, rather than in change in absolute JSW in millimeters. Results: Decreases in JSW ranged from zero to 3.7 mm over the 24 months; 19 of 97 patients exhibited < 30% JSN. (Figure presented) Patients with the IL-1RN (rs419598/rs315952/9005) TTG haplotype exhibited increased radiographic knee OA severity at baseline compared to those without TTG (p>0.08). These TTG patients exhibited increased risk for radiographic progression at 24 months that approached significance based on <=30% JSN [OR = 2.85; 95%CI=0.68-11.67; p>0.15]. In contrast, OA patients with IL-1RN CTA haplotype showed decreased risk for JSN over 24 months in the signal knee [OR= 0.33; 95%CI=0.170-1.014; p>0.05]. Differences in reported VAS pain between the CTA and TTG group were significant at 24 months (p> 0.01), indicating that while these patients were not distinguishable by radiograph or symptoms at onset, IL1RN haplotype predicted symptomatic differences at two years. Finally, the TTG haplotype group of patients expressed relatively increased IL-1b gene expression [15.683 +/- 9.407 (p>0.0001)] as assessed by TaqMan QPCR in peripheral blood leukocytes. The TTG patients also exhibited decreased sIL-1Ra [283.64 +/- 36.4 pg/ml (p>0.001) in plasma samples compared to IL-1RN CTA haplotype protective groups [IL-1b (fold change), 5.444 +/- 10.083; sIL-1Ra, 370.35 +/- 43.3pg/ml] of patients respectively. Conclusion: IL-1RN gene family polymorphisms, which appear to affect host production of IL-1Ra, merit evaluation as biomarkers that predict the risk of progression in patients with symptomatic knee OA

Objective: To evaluate the relationship of quantitative assessment of Bone Marrow Lesions (BML) with knee OA severity by radiographic findings. Methods: 58 OA patients (mean age 62+/-10, mean BMI 27+/-3, 59% female) underwent standardized nonfluoroscopic fixed-flexion knee radiographs. Two radiologists read the X-rays for KL grade, joint space width (JSW), and, using the OARSI atlas, joint space narrowing (JSN) and osteophytes; interreader agreement was assessed using Kappas and concordance correlation coefficients. Linear and logistic regression analysis was performed to assess associations while controlling the effects of age, sex and BMI. 3T-MRI included sagittal T2-weighted fat saturated spin-echo images (TR/TE=4000ms/75ms, FOV=15cm, matrix=256x128, slice thickness=3.0mm, receiver bandwidth 130Hz/pixel) and in/out of phase of FLASH images. Compartment-wise (medial tibial, lateral tibial, medial femur, lateral femur) BML volumes were quantified with T2-weighted fat saturated images and in/out of phase images respectively. BML volumes were dichotomized for statistical analysis. Results: KL score was a significant predictor of total BML volume (OR = 8.41, p = 0.0235). Medial tibial JSW, OARSI medial JSN, and medial tibial plateau osteophytes approached significance as predictors of BML volume at the medial tibia (OR = 0.71, p = 0.0551; OR = 2.16, p = 0.0597; and OR = 2.68, p = 0.0875, respectively). OARSI lateral JSN was a significant predictor of BML volume at the lateral tibia (OR = 3.62, p = 0.0169). Lateral tibial plateau osteophytes were predictors of total BML volume (OR = 4.58, p = 0.0299) and of BML volume at the lateral tibia (OR = 2.31, p = 0.0685). Lateral femoral condyle osteophytes approached significance as a predictor for BML at the lateral femur (OR = 2.25, p = 0.0651). Furthermore, quantitative BML volume strongly correlated with total quantitative synovial volume measured on MRI (beta= 0.22, p = 0.0003). Conclusions: Our data indicate that BML volume on MRI is a characteristic feature of progressive stages of OA, which not only correlates with JSN and osteophytes, but does so in a compartment-specific way. The data suggest that the altered mechanical forces that promote compartmental disease in OA lead to BML, JSN and osteophyte formation. Whether BML further contribute to cartilage loss, and are therefore targets of therapeutic intervention, remains to be determined

Purpose: Osteoarthritis is the most common form of arthritis, and the hereditable component of OA has been estimated to be 50-65%. The IL1 gene cluster region has been repeatedly associated with susceptibility to OA in various joints. This finding is consistent with the notion that inflammatory mediators, particularly IL-1, may be important in the pathogenesis and progression of OA. In these studies we examined the association of SNPs in inflammatory genes with the severity of knee OA and with the presence of generalized OA (GOA, here defined as knee plus hand OA). Method: One hundred forty-four OA patients (81 knee OA and 63 GOA) from two separate centers (NYUHJD, N=94 and Duke, N=50) met inclusion criteria: Caucasians of either sex, free of chronic disease other than osteoarthritis, and with a radiographic diagnosis of knee OA. Synovial fluid was available for the 50 Duke patients. Patients were genotyped for single nucleotide polymorphisms (SNPs). These subjects were divided into two groups: knee OA and GOA (knee and hand OA).We performed statistical analysis using Chi square and logistic regression, adjusting for age, gender and BMI to search for associations of radiographic knee OA severity and GOA with individual SNPs and with haplotypes. Results: After adjustment for age, gender, and BMI, individual TNF-a, IL1, IL1 receptor antagonist (IL1RN) SNPs were strongly associated with the presence of GOA (Table 1). In three of the four SNPs, alleles associated with increased incidence of GOA have been previously associated with increased cellular production of inflammatory mediators. These alleles are: IL1B (-3737) rs4848306 TT/CT, IL1B (-511) rs16944 CC/CT, and TNF-a (-308) rs1800629 AA/AG. In a separate analysis of knee OA severity based upon JSW, carriage of either one or two copies of a haplotype consisting of IL-1RN rs9005 (A), IL-1RN rs419598 (C), IL-1RN rs315952(T) was associated with lower odds of radiographic severity (OR 0.16; 95% CI 0.06-0.40), greater joint space width (JSW; p=0.0038) compared to those without these haplotypes. Patients with the protective haplotype had lower synovial fluid levels of IL-6 (p=0.018) and IL-10 (p=0.038). (Table Presented). Conclusion: Among the cytokines and receptor antagonist studied for genetic association, IL-1RN SNPs (rs9005, rs419598 & rs315952) predicted low risk for knee OA radiographic severity and haplotypes in IL-1B, TNFa ad IL-1RN (rs 315952) strongly associated with increased prevalence of GOA. These genetic markers provide insight into the possible role of inflammatory mediators in knee OA and could prove to be useful biomarkers in DMOAD drug development

Purpose: Wrist pain in rheumatoid arthritis (RA) is often complicated by carpal tunnel syndrome (CTS) secondary to active synovitis. In such cases it can be difficult to determine whether to inject the joint or the carpal tunnel - or treat systemically. Recent literature suggests that musculoskeletal ultrasound can diagnose CTS by identifying a median nerve cross sectional area of >10mm² at the inlet, or by using a wrist to forearm ratio of >1.4 (from inlet to 12 cm proximally) to correct for patients with confounding diagnoses that may cause larger nerves at baseline (Hobson-Webb 2008). In this pilot study, we aim to assess how well ultrasound can identify median nerve entrapment noninvasively specifically in RA patients with active wrist pain. Methods: We enrolled 19 consecutive patients with RA and active wrist pain in at least one wrist (total of 28 wrists) and assessed them for CTS by history and physical. Using gray scale ultrasound imaging at 12 MHz, we identified the median nerve at the level of the inlet (volar wrist between the scaphoid and pisiform) and 12 cm proximally to calculate the wrist-to-forearm ratio. More than a month after capture, we blindly evaluated median nerve cross-sectional area with a tracing function on the machine and calculated the ratio from wrist to forearm. We averaged the scores from two readings (one each by MMR and JS) for each patient. Results: In our pilot population of painful RA wrists, irrespective of CTS signs or symptoms, the overall average median nerve size at the inlet was 10.9 (range 5.5 to 26, SD of 3.9), while the overall average wrist-to-forearm ratio was 1.9 (range 1.32 to 2.80, SD of 0.47) Separating those with (n=12) and without (n=16) CTS signs and symptoms, the average median nerve areas at the inlet were 13 mm² and 9.42 mm² respectively (p=0.0097). The corresponding wrist to forearm ratios for patients with and without CTS signs and symptoms averaged 2.19 and 1.68 (p=.0032). Of the patients with CTS signs and symptoms, 92 % (11of 12) had an inlet nerve size >10 mm² and 100% (all 16) had a wrist-to-forearm ratio greater than the published 1.4 cutoff for CTS. Only 25% (4 of 16) of those without CTS signs and symptoms measured an inlet of >10 mm², but 75 % (12 of 16) of these asymptomatic patients had a wrist-to-forearm ratio >1.4. (Figure presented) Conclusion: This pilot data suggests that ultrasound evaluation of the median nerve by assessing cross sectional area at the wrist inlet is a sensitive and specific test in RA patients with active wrist pain, as it is in the general population. Conversely, the wrist-to-forearm ratio, introduced to compensate for size variability in healthy median nerves, may reveal false positive results for carpal tunnel syndrome in patients with RA and wrist pain. Measurements of the median nerve at the inlet can help distinguish RA synovitis of the joint from median nerve compression-and thus spare patients from painful nerve conduction studies