ARTHRITIS & RHEUMATISM Vol. 50, No. 2, February 2004, pp 674–680 © 2004, American College of Rheumatology LETTERS DOI 10.1002/art.20030 Dynamic gadolinium-enhanced magnetic resonance imaging of the wrist in patients with rheumatoid arthritis: comment on the article by Cimmino et al To the Editor: We read with interest the article by Cimmino et al (1) on the ability of dynamic gadolinium-enhanced magnetic resonance imaging (MRI) of the wrist to discriminate between active and inactive rheumatoid arthritis (RA). They described this technique as “an excellent candidate for the ideal method for the followup of patients with RA.” We take issue with this conclusion, which was based on results of a study of a small group of subjects who were not well matched for age (the average age in the control group was 39 years, compared with 62 years in RA patients) or disease duration. Importantly, no significant difference was found between patients with active disease and those with “intermediately active disease,” which was defined as any level of activity between active disease and complete clinical remission. Significance was reached only when the entire group of RA patients with any disease activity was compared with those with full disease remission combined with normal controls. These results suggest that the method used is unlikely to be sensitive enough for use in clinical trials, in which perhaps a 30% reduction in joint inflammation over time might be expected. The appeal of dynamic enhanced MRI as a means to measure synovitis lies in its apparent objectivity, providing, as it does, a numeric result upon which clinical and research decisions can confidently be based. Such confidence might be misplaced. Our own study of dynamic enhanced MRI of the wrist in 42 patients with early RA (2) revealed a number of points pertinent to the current study. We found that placement of the cursor at the region of highest signal intensity was critical to the results obtained. Although aiming for the brightest region on the slice showing most enhancement might seem straightforward, deviation by as little as 1–2 mm could greatly affect the rate of early enhancement (REE). This is because the region of interest (ROI) covers many highintensity pixels, all of which may change in brightness differently over time. Thus, the REE is an average of multiple pieces of information, obtained both spatially and temporally during the examination. In our study, blinded repositioning of the ROI cursor was performed in 10 patients and revealed 10% variability in the initial rate of enhancement (2). Figure 1 depicts cursor placement at 3 different regions of the wrist where there was bright enhancement of synovial membrane in a patient with early active RA. The rate of increase in signal intensity over a 42-second period varied by 18% according to cursor placement, from 26.2 to 28.3 to 32.1 pixels/second for points shown in Figures 1A, B, and C, respectively. Which of these points should be used for a followup study? The one with highest maximum signal intensity (Figure 1B), the one with the highest rate of increase in signal intensity (Figure 1A), or a different point altogether that happens to be very bright at followup? Cimmino et al have not provided data on reliability or reproducibility of their measurements, and we are not convinced of the suitability of this method for longitudinal followup of RA activity. Figure 1. A, Dynamic enhanced T1-weighted fat-suppressed spoiled gradient-recalled acquisition in the steady state coronal magnetic resonance image of the wrist with cursor placement at the radiocarpal joint, showing initial rate of enhancement as 32.1 pixels/second B, Cursor placement more distally within the midcarpal joint shows a rate of 28.3 pixels/second C, Placement within the midcarpal joint on a more dorsal slice produces a rate of 26.2 pixels/second. Cimmino and colleagues also reported a high degree of correlation between dynamic MRI and clinical measures of synovitis. Interestingly, these results were achieved in patients with widely disparate disease duration. Our data, from a more homogeneous group, revealed similar correlations at 1 year from disease onset (2), but in early disease only the pain score correlated with the initial rate of enhancement on dynamic MRI. This may be because MRI is capable of detecting the increased synovial vascularity of early synovitis before it becomes clinically apparent. For this reason, we would suggest that disease duration is of crucial importance in any comparison between MRI and clinical synovitis. We believe that MRI synovitis is more reliably measured by a global score quantifying synovial membrane thickness as well as postcontrast enhancement (3). This score is 674 LETTERS 675 derived visually from the whole joint scanned and is likely to reflect not only synovial angiogenesis (which best correlates with dynamic MRI results) but also hypertrophy and cellular infiltration, which correlate with synovial volume (4). Our global MRI synovitis score was a better predictor of erosions after 1 year than was synovitis scored using dynamic MRI (2 (1df) ⫽ 9.2, P ⫽ 0.003 versus 2 (1df) ⫽ 4.7, P ⫽ 0.03) (2) and closely resembles the score recommended by the Outcome Measures in Rheumatoid Arthritis Clinical Trials (OMERACT) MRI working party (5). Along with that group, we favor incorporating MRI measures of synovitis into clinical trials of new therapeutics for RA to more accurately determine the role of MRI in clinical and research practice. Fiona M. McQueen, MBChB, MD, FRACP Jeff Crabbe, MBChB, FRANZCR Neal Stewart, MBChB, FRANZCR University of Auckland Auckland, New Zealand 1. Cimmino MA, Innocenti S, Livrone F, Magnaguagno F, Silvestri E, Garlaschi G. Dynamic gadolinium-enhanced magnetic resonance imaging of the wrist in patients with rheumatoid arthritis can discriminate active from inactive disease. Arthritis Rheum 2003;48:1207–13. 2. Huang J, Stewart N, Crabbe J, Robinson E, McLean L, Yeoman S, et al. A 1-year follow-up study of dynamic magnetic resonance imaging in early rheumatoid arthritis reveals synovitis to be increased in shared epitope-positive patients and predictive of erosions at 1 year. Rheumatology (Oxford) 2000;39:407–16. 3. McQueen FM, Stewart N, Crabbe J, Robinson E, Yeoman S, Tan P, et al. Magnetic resonance imaging of the wrist in early rheumatoid arthritis reveals a high prevalence of erosions at four months after symptom onset. Ann Rheum Dis 1998;57:350–6. 4. Østergaard M, Hansen M, Stoltenberg M, Gideon P, Klarlund M, Jensen KE, et al. Magnetic resonance imaging–determined synovial membrane volume as a marker of disease activity and a predictor of progressive joint destruction in the wrists of patients with rheumatoid arthritis. Arthritis Rheum 1999;42:918–29. 5. Conaghan P, Edmonds J, Emery P, Genant H, Gibbon W, Klarlund M, et al. Magnetic resonance imaging in rheumatoid arthritis: summary of OMERACT activities, current status, and plans. J Rheumatol 2001;28:1158–62. DOI 10.1002/art.20137 Reply To the Editor: We appreciate the interest shown by McQueen et al in our recent article on dynamic MRI of the rheumatoid wrist. Their comments deserve some further consideration. McQueen et al state that they do not share our opinion that this technique could represent “an excellent candidate for the ideal method” of followup of RA patients. This is in contrast with their recent statement (1) that dynamic MRI “may have an important role in monitoring responses to new anti-rheumatic agents, including biologics. . . .” In their letter, McQueen et al observe that in our study we could not find any significant difference between RA patients with 2 different degrees of disease activity. Furthermore, they claim that this difference became significant only when the entire group of patients with any disease activity was compared with another group comprising patients with disease in remission and normal controls. As a matter of fact, our study showed that patients with high or intermediate disease activity had a significantly higher relative enhancement (RE) than did either patients in remission or normal controls. The fact that dynamic MRI could not differentiate patients with active RA according to the definition often used to enroll patients in clinical trials (ⱖ6 swollen joints and ⱖ9 tender joints plus at least 2 of the following features: morning stiffness lasting at least 45 minutes, erythrocyte sedimentation rate [ESR] ⱖ28 mm/hour, and C-reactive protein [CRP] value more than 3-fold the upper limit of normal) from those who had a milder disease but were not in remission as defined by Pinals et al (2) has already been discussed. We believe that the above-mentioned arbitrary definition of disease activity is too rough to reflect real differences in joint inflammation. In fact, dynamic MRI correlates well with several clinical and laboratory data of continuous distribution, such as number of swollen joints, number of tender joints, the Ritchie index, score on the Health Assessment Questionnaire (3), CRP value, and ␣2 globulins, as well as with the disease activity score (DAS) (4). The impression that the DAS is a more sensitive indicator of disease activity is confirmed by the fact that it is able to fully differentiate the 3 groups of patients (P ⬍ 0.00001 by analysis of variance). Therefore, our data support the view that dynamic MRI is sensitive enough to follow up patients receiving disease-controlling treatments or biologics. As an example, a patient recently treated with infliximab had a 62.3% decrease in RE (from 145 to 54.7) and a 68.5% decrease in the REE (from 2 to 0.63) after the first month of therapy, corresponding to 2 infusions. In the same period, the DAS diminished from 4.1 to 2.1 (49% decrease). Interestingly, among the components of the DAS, only the ESR, which is not correlated with either RE or the REE, did not decrease as much as the other components. We admit that this example is only anecdotal. Furthermore, the efficacy of anti–tumor necrosis factor ␣ therapy, which overwhelms that of any traditional diseasecontrolling antirheumatic drug, should be easily demonstrated clinically and with imaging systems. It is not clear to date whether dynamic MRI could evaluate minor changes of disease activity due to either spontaneous fluctuations or the effects of traditional disease-controlling therapy. Another point raised by McQueen et al concerns the choice of the area of the synovial membrane where the enhancement curve should be calculated. Our study is not concerned with the comparison of the different possibilities listed in their letter, i.e., the area with the maximum initial signal intensity or the one with the highest increase in signal intensity. This point needs further research and standardization. What we can state on the basis of our results is that identifying the slice with the highest visual enhancement after gadolinium infusion and positioning the ROI in the area with the highest rate of increase in signal intensity resulted in good discrimination of patients with disease in remission from those with active disease and yielded excellent correlations with clinical and laboratory data. We agree that ROI repositioning is crucial for comparison of results, especially in followup studies. Reproducibility of results depends on precise repositioning of the hand in the gantry in order to yield comparable slices for followup studies. Extremity-dedicated MRI is partic- 676 LETTERS Marco A. Cimmino, MD Enzo Silvestri, MD Giacomo Garlaschi, MD Università di Genoa Genoa, Italy Figure 1. Positioning of the patient’s wrist in the dedicated coil. The lower limit of the anatomic snuff-box should correspond to the internal border of the coil. ularly suited for this scope, because the hand can be precisely placed based on anatomic fiducials (5) (Figure 1). In this respect, we are developing a positioning device that should further improve the precision of the study by reducing the possibility of movements of the wrist. In this way, higher degrees of reproducibility should be ensured, unless major anatomic changes due to rheumatoid synovitis occur between 2 examinations. Finally, McQueen et al highlight the importance of disease duration in the evaluation of dynamic MRI in comparison with clinical data. Their finding that dynamic MRI correlates with the same clinical data of our study after only 1 year of followup, but not in the very early phase of RA (2), is difficult to explain and raises some concerns. If their explanation that MRI can detect synovitis in a preclinical phase is true, then their patients, being without arthritis, cannot fulfill the American College of Rheumatology (ACR; formerly, the American Rheumatism Association) criteria for RA (6). On the contrary, our study was concerned with patients with RA and synovitis of the wrist, who were identified according to the ACR criteria and attended a rheumatology clinic. The common goal of research in the area should be to develop a technique that is useful in prospective cohorts of patients, regardless of disease duration. In fact, in a series of 73 RA patients from our clinic, no correlation between disease duration and dynamic MRI was found (Cimmino MA: unpublished observations). Dynamic MRI is an appealing method to quantify synovial inflammation, because it is objective, easy to perform, rapid (15 minutes) and requires no time-consuming postexamination calculations. Studies on the correlation of dynamic MRI of the wrist with synovial membrane histology and with biologic markers of neoangiogensis and synovial cell density are needed to demonstrate whether this technique is a comprehensive indicator of the different aspects of synovial damage. Comparisons with more complex and time-consuming global scores of joint involvement by RA should also be performed to try to identify the best MRI approach to quantification of synovitis in RA. 1. Huang J, Stewart N, Crabbe J, Robinson E, McLean L, Yeoman S, et al. A 1-year follow-up study of dynamic magnetic resonance imaging in early rheumatoid arthritis reveals synovitis to be increased in shared epitope-positive patients and predictive of erosions at 1 year. Rheumatology (Oxford) 2000;39:407–16. 2. Pinals RS, Masi AT, Larsen A. Preliminary criteria for clinical remission in rheumatoid arthritis. Arthritis Rheum 1981;24:1308–15. 3. Fries JF, Spitz PW, Kraines RG, Holman HR. Measurement of patient outcome in arthritis. Arthritis Rheum 1980;23:137–45. 4. Van der Heijde DM, van’t Hof M, van Riel PL, van de Putte LB. Development of a disease activity score based on judgment in clinical practice by rheumatologists. J Rheumatol 1993;20:579–81. 5. Garlaschi G, Silvestri E, Satragno L, Cimmino MA. The rheumatoid hand: diagnostic imaging. Springer-Verlag Milan: SpringerVerlag Italia; 2002. 6. Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF, Cooper NS, et al. The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum 1988;31:315–24. DOI 10.1002/art.20003 Diagnostic value of Doppler ultrasonography in rheumatoid arthritis: comment on the article by Weidekamm et al To the Editor: We read with interest the recent article by Weidekamm et al (1). The authors should be congratulated on a solid contribution to the elucidation of the use of ultrasonography (US) in rheumatoid arthritis (RA). The article, however, raises some important issues that require comment. The title “Diagnostic Value of High-Resolution B-Mode and Doppler Sonography. . .” suggests that investigation of the diagnostic value of sonography in RA was the main purpose of the study. However, because all patients enrolled in the study had RA, the value of US for diagnosing RA could not be investigated. If by diagnostic value the authors mean the ability to correctly detect various joint pathologies (e.g., synovitis or erosions), this would require comparison with some kind of gold standard reference to determine the accuracy, sensitivity, and specificity of the applied method. The authors introduce a scoring system to quantify the severity of the joint involvement, using a 4-point scale (0 ⫽ no changes, 1 ⫽ slight changes, 2 ⫽ moderate changes, and 3 ⫽ strong changes of the soft tissue and erosions of the joint). The scoring system pools information on inflammation and destruction, and the rating of the different types of pathologies as well as the limits for the different grades are not further described, leaving some questions unanswered (e.g., does a small erosion without inflammatory changes result in grade 3?). Clear definitions and criteria for different grades could help testing and use by other groups. Furthermore, it may be controversial to pool reversible disease activity parameters such as synovitis with signs of cumulative joint damage such as erosions. Such a system has an inherent inability to be sensitive LETTERS 677 for detecting change, because therapy most often reduces signs of inflammation, while some erosive progression still occurs. We would suggest evaluating inflammatory and destructive parameters separately, as was previously done by several groups of investigators (2–6). The authors also introduce a 4-point scoring system for power Doppler signal, with no, small, moderate, and strong vascularization scored as 0, 1, 2, and 3, respectively. Very relevantly, they illustrate the different grades in a figure. However, 2 of the examples are sagittal cuts, while the other 2 are transverse cuts, making them difficult to use as reference images for scoring. Moreover, and very importantly, the majority of the signal in the illustration of grade 2 (Figure 1C) originates from the dorsal metacarpal artery, leaving only very little signal originating from the synovium. Probably, the remaining signal is not different from the illustration of grade 1 (Figure 1B), but, as mentioned, the cuts are difficult to compare. Similarly, the image illustrating grade 3 (Figure 1D) includes a large portion of physiologic signal from the dorsal metacarpal artery. Because these aspects are not mentioned in the figure legend, the figure may be misinterpreted. In conclusion, the fact that use of a certain method, in this case US, shows more abnormalities than does clinical examination does not directly translate into a higher diagnostic value, because without a gold standard reference it is unknown whether a finding is true or false positive. Furthermore, when new scoring systems are introduced, the benefits of thoroughly described methods, unambiguous, explicit definitions and illustrations of pathologies and grades, and biologically maximally meaningful scores cannot be overestimated. Once this is done, studies are needed to clarify the validity of the system (including construct, face, content, criterion, and discriminant validities) (7), as is currently being done for magnetic resonance imaging in RA (8), before the use of the system can be generally recommended. We acknowledge the quality of the study by Weidekamm et al and hope that discussion will further advance the process of clarifying the role and the optimal methodologies of US in RA. Marcin Szkudlarek, MD The Copenhagen University Hospital at Hvidovre Mikkel Østergaard, MD, PhD, DMSc The Copenhagen University Hospitals at Herlev and Hvidovre Copenhagen, Denmark 1. Weidekamm C, Köller M, Weber M, Kainberger F. Diagnostic value of high-resolution B-mode and Doppler sonography for imaging of hand and finger joints in rheumatoid arthritis. Arthritis Rheum 2003;48:325–33. 2. Backhaus M, Kamradt T, Sandrock D, Loreck D, Fritz J, Wolf KJ, et al. Arthritis of the finger joints: a comprehensive approach comparing conventional radiography, scintigraphy, ultrasound, and contrast-enhanced magnetic resonance imaging. Arthritis Rheum 1999;42:1232–45. 3. Wakefield RJ, Gibbon WW, Conaghan PG, O’Connor P, McGonagle D, Pease C, et al. The value of sonography in the detection of bone erosions in patients with rheumatoid arthritis: a comparison with conventional radiography. Arthritis Rheum 2000;43:2762–70. 4. Szkudlarek M, Court-Payen M, Strandberg C, Klarlund M, Klausen T, Østergaard M. Power Doppler ultrasonography for assessment of synovitis in the metacarpophalangeal joints of patients with rheumatoid arthritis: a comparison with dynamic magnetic resonance imaging. Arthritis Rheum 2001;44:2018–23. 5. Backhaus M, Burmester GR, Sandrock D, Loreck D, Hess D, Scholz A, et al. Prospective two year follow up study comparing novel and conventional imaging procedures in patients with arthritic finger joints. Ann Rheum Dis 2002;61:895–904. 6. Szkudlarek M, Court-Payen M, Jacobsen S, Klarlund M, Thomsen HS, Østergaard M. Interobserver agreement in ultrasonography of the finger and toe joints in rheumatoid arthritis. Arthritis Rheum 2003;48:955–62. 7. Felson DT, Anderson JJ, Boers M, Bombardier C, Chernoff M, Fried B, et al. The American College of Rheumatology preliminary core set of disease activity measures for rheumatoid arthritis clinical trials. Arthritis Rheum 1993;36:729–40. 8. McQueen F, Lassere M, Edmonds J, Conaghan P, Peterfy C, Bird P, et al. OMERACT Rheumatoid Arthritis Magnetic Resonance Imaging Studies: summary of OMERACT 6 MR Imaging Module. J Rheumatol 2003;30:1387–92. DOI 10.1002/art.20138 Reply To the Editor: We thank Drs. Szkudlarek and Østergaard for their comments on our recent article, and appreciate their valuable contribution. The purpose of our study was to point out the useful application of B-mode and power Doppler sonography in patients with RA. The predominant clinical issue was not to diagnose RA, but to determine the grade of severity of soft tissue lesions and hypervascularization. We noted that the severity of joint lesions, including inflammation and soft tissue abnormalities, should be ranked according to a standard score. The proposed new scoring system helps to determine individual therapy and the efficacy of therapy during followup. One of the limitations of our study is the lack of histologic data as a standard of reference, but our aim was to avoid invasive interventions in our patients. Therefore, we used the results of the final clinical outcome as a reference for our sonographic findings. We demonstrated a significant correlation between clinical examinations and sonography, and we found that sonography did not reveal any abnormalities over and above the clinical results. In our opinion, B-mode sonography should be used to score destruction of the joints, whereas power Doppler sonography should be used to score inflammation of the joints. However, we agree with Drs. Szkudlarek and Østergaard that some parameters of the high-resolution B-mode sonography score should include signs of inflammatory changes such as effusion or synovial thickening of the joint (Newman JS, Laing TJ, McCarthy CJ, Adler RS. Power Doppler sonography of synovitis: assessment of therapeutic response—preliminary observations. Radiology 1996;198:582–4). For a precise evaluation of the complete status of the joint, we recommend the additional use of power Doppler sonography with conventional B-mode sonography. We agree with the authors that inflammatory and destructive parameters should be evaluated separately, as proposed in a previous report (Backhaus M, Kamradt T, Sandrock D, Loreck D, Fritz J, Wolf KJ, et al. Arthritis of the finger joints: a comprehensive approach comparing conventional radiography, scintigraphy, ultrasound, and contrast-enhanced magnetic resonance imaging. Arthritis Rheum 1999;42:1232–45). 678 LETTERS The new 4-point scoring system is useful in daily clinical routine for the grading of lesions and hypervascularization. In our study, we combined each parameter in a 4-point scale, according to the criteria of the Larsen score (e.g., slight changes include periarticular tissue swelling and slight joint space narrowing without erosions). This summary of the individual parameters enabled us to make a better comparison of high-resolution B-mode sonography with conventional radiography and clinical examinations, and also simplified statistical analyses. The new power Doppler scoring system concentrates on the inflammatory changes of the joint, expressed as hypervascularization. Therefore, we recommend using power Doppler sonography for the evaluation of inflammatory changes in the followup period after therapy. Because sonography is an investigator-dependent examination mode, and motion artifacts can be misinterpreted as hypervascularization of the rheumatoid joints, inexperienced investigators should request a second opinion, and in case of controversy, scoring should be based on a consensus reading. We hope that results of further studies with power Doppler and B-mode sonography will support the usefulness of the new recommended scoring system and help standardize the investigation of inflammatory and destructive changes in RA. Claudia Weidekamm, MD University of Vienna Vienna, Austria DOI 10.1002/art.20017 Correlations of hip joint space among asymptomatic subjects: comment on the article by Lanyon et al To the Editor: We were interested in the findings of Lanyon et al regarding radiographic hip joint space width (JSW) among asymptomatic subjects without hip osteoarthritis (OA) in the UK (1), because their findings were in many ways similar to ours using a different demographic population (2). However, although our findings were consistent with those reported by Lanyon et al, that women have significantly smaller JSW than men, we found that these sex differences were no longer significant when normalized for height. Lanyon et al similarly observed that JSW and height were significantly related in both sexes, and we speculate that in the population described by them, the sex differences in JSW may be lost when similarly normalized. This issue is not trivial, because these authors and other investigators have used hip JSW as an epidemiologic index of OA prevalence, and at least some of the female preponderance that is often reported for JSW narrowing may well be an artifact of stature rather than truly related to sex differences. This is, in fact, implied by the authors’ findings after they reduced their threshold definition for JSW narrowing in women. In addition, whereas in the UK population there appeared to be an age-related decline of JSW among women but not men, no such decline was observed in the Turkish population that we studied. This discrepancy might be attributable to several reasons. First, geographic variation in the prevalence of radiographic hip OA has long been recognized. Because hip OA is less common in Turkey than in the UK (3,4), the progressive narrowing of JSW observed in the UK may be reflecting early OA (as Lanyon et al stated in the Discussion). Second, the higher mean age and narrower age range in the UK study may have resulted in a larger group of subjects with early OA, which could bias results toward a narrower measured JSW. Finally, patients who have knee deformity may slightly rotate and abduct their hips while lying supine, and if this occurs during positioning for the intravenous urography (IVP) it can result in an artifactual decrease in the measured radiographic JSW of clinically normal hips (5). In a population with a relatively high background prevalence of knee OA, the effect may be significant if there is subtle malpositioning by radiology technologists who are focusing on the abdomen to achieve the proper IVP views. Berna Goker, MD Gazi University School of Medicine Ankara, Turkey Joel A. Block, MD Rush Medical College Chicago, IL 1. Lanyon P, Muir K, Doherty S, Doherty M. Age and sex differences in hip joint space among asymptomatic subjects without structural change: implications for epidemiologic studies. Arthritis Rheum 2003;48:1041–6. 2. Goker B, Sancak A, Arac M, Shott S, Block JA. The radiographic joint space width in clinically normal hips: effects of age, gender and physical parameters. Osteoarthritis Cartilage 2003;11:328–34. 3. Goker B. Radiographic osteoarthritis of the hip joint in Turkey. Rheumatol Int 2001;21:94–6. 4. Lawrence JS, Sebo M. The geography of arthritis. In: Nuki G, editor. The etiopathogenesis of osteoarthrosis. Baltimore: University Park Press; 1980. p. 155–83. 5. Goker B, Sancak A, Haznedaroglu S, Arac M. The effects of minor flexion, abduction or adduction on the radiographic joint space width of the hip [abstract]. Arthritis Rheum 2002;46 Suppl 9:S149. DOI 10.1002/art.20139 Reply To the Editor: We thank Drs. Goker and Block for their interest in our study and for their speculations as to why the results of our study differ from theirs (Goker B, Sancak A, Arac M, Shott S, Block JA. The radiographic joint space width in clinically normal hips: effects of age, gender and physical parameters. Osteoarthritis Cartilage 2003;11:328–34). There are, of course, important differences in design between the 2 studies. The study by Goker at al included 118 Turkish subjects, of whom only 67 were over age 40 years, and just 25 were over age 60 years. Although none had hip pain, the definition of normal hips permitted inclusion of subjects with Kellgren/Lawrence grades of 0, 1, or 2. Therefore, some of their subjects had definite or possible osteophyte formation and possible joint space narrowing (unfortunately, the exact number in each grade was unspecified). In contrast, our starting study population included 689 UK subjects without hip pain, all of whom were over age 45 years. We then focused on 533 subjects whose radiographic definition of normal excluded subjects with any LETTERS 679 evidence of OA in either hip—that is, no subject had any evidence of osteophyte formation or narrowing in either hip, and all were Kellgren/Lawrence grade 0. These important differences in sample size and definition of normal are likely to be the major reason for differences between the 2 studies. It is likely that any sex differences in pelvic morphometry are multifactorial and are influenced by genetic, environmental, and hormonal factors. These same factors are likely to influence general body morphology, including height. In our UK study population, the influence of sex on joint space remained significant (P ⫽ 0.022), even with the inclusion of height as an independent variable. We have undertaken further regression analysis, and even when height is taken into account there is still a significant decline in JSW in women (P ⫽ 0.001), but not in men (P ⫽ 0.5). Thus, although height and sex are very closely associated, there does appear to be a difference that relates to sex, and not solely to height, with respect to diminution of JSW with age. Goker and Block interestingly speculate as to whether knee OA or deformity might have altered our results. The prevalence of knee pain in our subjects without hip pain and with normal hip radiographs was 26%. Not only is this prevalence identical to that in other unselected UK population studies (Peat G, McCarney R, Croft P. Knee pain and osteoarthritis in older adults: a review of community burden and current use of primary health care. Ann Rheum Dis 2001;60: 91–7), but we also observed no sex differences in knee pain prevalence. This negates the suggestion that a high prevalence of knee OA might have influenced these results. Goker and Block also suggest that there may be basic morphologic/racial differences between UK and Turkish individuals. This hypothesis would be interesting to test but could be examined only if a similar, rigorous definition of normal were to be used. We concur with Goker and Block that differences in JSW between sexes, which may in part be explained by differences in morphology, merit further study. Peter Lanyon, MD Kenneth Muir, PhD Sally Doherty Michael Doherty, MD Queen’s Medical Center, and Nottingham City Hospital Nottingham, UK DOI 10.1002/art.20041 Delayed onset of renal disease in mice expressing a human C-reactive protein transgene: comment on the article by Szalai et al To the Editor: In a recent study, Szalai et al showed that (NZB ⫻ NZW)F1 hybrid mice expressing a human C-reactive protein (CRP) transgene survived ⬃8 weeks longer and developed proteinuria ⬃5 weeks later than control (NZB ⫻ NZW)F1 mice without the transgene (1). In the mice with the human CRP gene, serum levels of human CRP were stable at ⬃1 g/ml, no CRP was detected in the urine, and human CRP accumulated in cells of renal tubules with increasing age of the mice. The level of messenger RNA for human CRP was also increased in the same cells. Of note is that human CRP was not present in renal glomeruli (Figure 7 of the article by Szalai et al and cover picture of the June 2003 issue of Arthritis & Rheumatism); the cover picture was mislabeled in that the red staining in cells of renal tubules was for CRP as in Figure 7 and not for mouse IgG). The authors discuss several possible mechanisms for the observed findings and favor a hypothesis that somehow the accumulation of human CRP in cells of renal tubules altered the deposition of immune complexes in glomeruli and/or the ensuing inflammation in glomeruli. There is no support for this hypothesis. Szalai et al point out, and illustrate with electron micrographs, that at 35 weeks the control (NZB ⫻ NZW)F1 mice had extensive subepithelial immune deposits and some subendothelial and mesangial immune deposits. In contrast, the mice with the human CRP transgene had sparse subepithelial and marked mesangial immune deposits. The authors do not discuss what mechanisms may have caused the differences observed by electron microscopy. A likely explanation is the effect of circulating CRP on deposition of histones or nucleosomes in glomeruli. Subepithelial immune deposits are formed locally, by first depositing the antigen mainly by charge–charge interaction, followed by binding of other molecules and subsequently antibodies. In (NZB ⫻ NZW)F1 mice as well as in kidney biopsy specimens from patients with systemic lupus erythematous, histones are present at the glomerular basement membrane (2,3). Furthermore, once histones are present at the glomerular basement membrane, these molecules can then bind DNA, nucleosomes, or other anionic antigens, followed by antibodies (4,5). Therefore, the human CRP present in the circulation of the transgenic mice would bind histones, an interaction noted by the authors. The histone–CRP complex would not penetrate the glomerular basement membrane owing to size, and thereby would decrease the formation of subepithelial immune deposits and delay proteinuria and loss of renal function. It would have been useful if Szalai and colleagues had stained the glomeruli of the control and transgenic mice for the presence of histones. A concern is that the authors should have examined the ultrastructural lesions of the transgenic mice beyond 35 weeks, at a time when these mice had proteinuria comparable with that in the control mice at 35 weeks. This comparison would have ruled out the possibility that the renal lesions were at different stages of development and were delayed in the mice with the CRP transgene. Mart Mannik, MD University of Washington Seattle, WA 1. Szalai AJ, Weaver CT, McCrory MA, van Ginkel FW, Reiman RM, Kearney JF, et al. Delayed lupus onset in (NZB ⫻ NZW)F1 mice expressing a human C-reactive protein transgene. Arthritis Rheum 2003;48:1602–11. 2. Schmiedeke T, Stoeckl F, Muller S, Sugisaki Y, Batsford S, Woitas R, et al. Glomerular immune deposits in murine lupus models may contain histones. Clin Exp Immunol 1992;90:453–8. 3. Stökl F, Muller S, Batsford S, Schmiedeke T, Waldherr R, Andrassy K, et al. A role for histones and ubiquitin in lupus nephritis? Clin Nephrol 1994;41:10–7. 4. Schmiedeke TM, Stökl FW, Weber R, Sugisaki Y, Batsford SR, Vogt A. Histones have high affinity for the glomerular basement 680 LETTERS membrane: relevance for immune complex formation in lupus nephritis. J Exp Med 1989;169:1879–94. 5. Gauthier VJ, Tyler LN, Mannik M. Blood clearance kinetics and liver uptake of mononucleosomes in mice. J Immunol 1996;156: 1151–6. Finally, Dr. Mannik points out that the cover picture on the June issue of Arthritis & Rheumatism is mislabeled. Although we were already aware of this mistake we take no credit for it—this unfortunate error was introduced by the editors. Alexander J. Szalai, PhD The University of Alabama at Birmingham DOI 10.1002/art.20141 Reply To The Editor: My colleagues and I are pleased by the attention Dr. Mannik paid to our recent article and appreciate his thoughtful letter. As Dr. Mannik reiterates, in our report we entertained several possible mechanisms for the observed findings and we favored a hypothesis that somehow (admittedly, we don’t yet know exactly how) the accumulation of CRP in cells of renal tubules altered the deposition of immune complexes in glomeruli and/or the ensuing inflammation there. We also pointed out that in human CRP–transgenic mice, subepithelial deposition of immune complexes was sparser than in nontransgenic animals, whereas mesangial immune deposition was more intense. Despite the entirely plausible nature of this model, Dr. Mannik expresses two concerns: first, that there is no support for our hypothesis that CRP altered the pattern of immune complex deposition, and second, that we did not discuss what mechanisms may have caused the differences revealed by electron microscopy. He then offers his own equally well-informed but no better supported hypothesis, i.e., that circulating CRP alters glomerular deposition of histones or nucleosomes. Our more mechanistic studies of these mice are ongoing, and we are considering many possible scenarios. We will certainly heed Dr. Mannik’s advice and stain the glomeruli of control and transgenic mice for the presence of histones. With regard to examining lesion morphology beyond 35 weeks of age, we did so and did see the same pattern of CRP and immune deposits as was seen at 35 weeks. For logistical reasons we chose to show only the images from the 35-weekold mice. DOI 10.1002/art.20031 Self-arthrocentesis in a man with joint pain To the Editor: With the help of his spouse, a patient underwent periodic home arthrocentesis. A 75-year-old man was seen for evaluation of joint pain. He had bilateral knee crepitus, large joint effusions, and radiographs characteristic of degenerative joint disease with chondrocalcinosis. His wife requested that an arthrocentesis be done to relieve his joint pain, and volunteered that she herself periodically “pinched” his knee joints to drain fluid. She explained that she would repeatedly pinch the skin overlying the medial compartment of the knee with her fingernails until the skin was broken, and would further abrade the underlying tissue until a path had been cleared into the joint and fluid liberated. She would apply pressure to facilitate drainage of synovial fluid. There was no apparent history of past episodes of joint infection. The couple was strongly discouraged from continuing this practice. Conventionally performed sterile diagnostic arthrocentesis yielded a bland joint fluid with occasional calcium pyrophosphate dihydrate crystals. Under some circumstances, it may be prudent to inquire about self-arthrocentesis when physical findings suggest self-induced trauma to the joint. The extent of this practice is unknown. Philip L. Cohen, MD University of Pennsylvania and Philadelphia Veterans Administration Medical Center Philadelphia, PA Applications Invited for Editor of Arthritis & Rheumatism, 2005–2010 and Editor of Arthritis Care & Research, 2005–2009 During the summer and fall of 2004, the American College of Rheumatology Committee on Journal Publications will review applications for the position of Editor, Arthritis & Rheumatism, 2005–2010 term and the position of Editor, Arthritis Care & Research, 2005–2009 term. The official term of the next Arthritis & Rheumatism editorship is July 1, 2005–June 30, 2010; however, some of the duties of the new Editor will begin during a transition period starting April 1, 2005. The official term of the next Arthritis Care & Research editorship is July 1, 2005–June 30, 2009; however, some of the duties of the new Editor will begin during a transition period starting April 1, 2005. Applications will be available beginning February 4, 2004. The deadline for completed applications is June 1, 2004, and the final selection will be announced by November 2004. It is requested, but not required, that those who plan to apply for either position submit a nonbinding letter of intent by April 15, 2004. For additional information or to request an application or submit a letter of intent, contact Jane Diamond, Managing Editor, at the ACR office.