MRI Features of Lyme Arthritis in Children

Transcription

1 Pediatric Imaging Ecklund et al. MRI of Lyme Arthritis in Children Kirsten Ecklund 1 Sigella Vargas 1 David Zurakowski 2 Robert P. Sundel 3 Ecklund K, Vargas S, Zurakowski D, Sundel RP Received June 30, 2004; accepted after revision September 13, Department of Radiology, Children s Hospital, 300 Longwood Ave., Boston, MA Address correspondence to K. Ecklund (kirsten.ecklund@childrens.harvard.edu). 2 Departments of Orthopaedics and Biostatistics, Children s Hospital, Harvard Medical School, Boston, MA Rheumatology Program, Division of Immunology and the Department of Pediatrics, Children s Hospital, Harvard Medical School, Boston, MA AJR 2005;184: X/05/ American Roentgen Ray Society MRI Features of Lyme Arthritis in Children OBJECTIVE. Oligoarthritis is the most common manifestation of late Lyme disease in children. Considerable overlap can occur in the clinical presentation of Lyme arthritis and acute septic arthritis. Early differentiation is critical, given the disparate therapeutic implications; Lyme arthritis is treated with outpatient oral antibiotics, while septic arthritis requires hospitalization, IV antibiotics, and, often, surgical drainage. We wanted to identify MRI features that may distinguish Lyme arthritis from septic arthritis in children. MATERIALS AND METHODS. Knee MR images in 11 children with Lyme arthritis and 7 with septic arthritis, with a mean age 10.6 years old and 11.7 years old, respectively, were reviewed by a radiologist blinded to the final diagnosis. Joint effusion size, synovial thickness, adenopathy, subcutaneous, marrow, and muscle edema on MRI; and clinical parameters including age, sex, fever, WBC, erythrocyte sedimentation rate, C-reactive protein, and joint fluid WBC in the two patient groups were compared using univariate and multivariate analyses. RESULTS. Subcutaneous edema was seen in all septic arthritis patients but in only one of 11 patients with Lyme arthritis (p < 0.01). Myositis and adenopathy were present in all Lyme arthritis patients and two of seven patients with septic arthritis (both p < 0.01). No significant difference was present in synovial thickness, marrow edema, or joint fluid size. There were no statistically significant differences in the clinical parameters assessed. CONCLUSION. Our results identified three MRI features, specifically, myositis, adenopathy, and lack of subcutaneous edema, that strongly suggest the diagnosis of Lyme arthritis rather than septic arthritis in children with acute inflammation of the knee. Awareness of these characteristic MRI features may avoid unnecessary invasive procedures and cost. yme disease, caused by infection L with the spirochete Borrelia burgdorferi, is the most common vector-borne illness in the United States. The annual number of cases reported to the Centers for Disease Control and Prevention has increased steadily since surveillance was initiated in In the year 2002 alone, 23,763 new cases of Lyme disease were reported in the U.S. [1]. The disease has striking seasonal and geographic predilections. Nearly 80% of cases are reported between May and August. Most cases occur in New England, the Mid-Atlantic states, and, to a lesser extent, the Midwest. The highest reported incidence occurs in children 5 9 years old [2]. Lyme disease is divided into early and late phases, with early infection characterized by constitutional symptoms and the hallmark rash, erythema migrans. Fewer than half of patients present with early-phase symptoms, and even fewer recall a tick bite. In contrast to adults, who typically develop neurologic symptoms in the late phase of Lyme disease, children more often present with oligoarticular arthritis, especially involving the knee [3]. Timely diagnosis of Lyme arthritis is difficult. No rapid laboratory test reliably identifies serologic markers of borellial infection, and there is considerable overlap in the clinical, laboratory, and radiographic presentation of children with Lyme arthritis and those with other acute arthritides. In fact, Lyme arthritis is diagnosed on initial presentation in fewer than 20% of children [4]. Yet early differentiation from septic arthritis is particularly important because of the disparate therapeutic implications of each diagnosis. Treatment for septic arthritis involves hospitalization, joint aspiration, a lengthy course of IV antibiotics, and, often, surgical drainage. In contrast, initial therapy for Lyme arthritis is a 4-week course of oral doxycycline or amoxicillin [3] AJR:184, June 2005

2 MRI of Lyme Arthritis in Children Fig year-old boy with bacterial arthritis of left knee. Sagittal intermediateweighted fat-suppressed MR image shows large joint effusion with maximal distention of suprapatellar bursa. Even when Lyme is suspected, the diagnosis can be delayed pending the results of serologic analysis. Formal diagnosis of Lyme disease is dependent on visualization of antibodies to the causative spirochete by serum enzyme immunoassay (EIA), followed by Western immunoblot confirmation, which can take up to a week [5]. The purpose of this study was to determine whether unique MRI characteristics of Lyme arthritis in children may differentiate the disease from septic arthritis. Awareness of such findings could avoid delay in diagnosis and unnecessary invasive interventions. TABLE 1 Materials and Methods A review of radiology and rheumatology department databases from a large New England children s hospital between 1997 and 2004 revealed 11 children who underwent MRI of the knee and ultimately were diagnosed with Lyme arthritis, and seven children similarly imaged who were found to have bacterial septic arthritis. All patients were imaged within 48 hr of their presentation with acute inflammatory arthritis of the knee. The Lyme group consisted of seven boys and four girls with an age range of years; mean, 10.6 ± 4.5 (SD) years. In the septic arthritis group, there were three Comparison of Clinical Findings Between Lyme and Septic Arthritis Patients Demographic/Clinical Variable Lyme Arthritis (n = 11) Septic Arthritis (n =7) p Age (yr) 10.6 ± ± Sex (M/F) 7/4 3/ Presence of fever 4 (36%) 6 (86%) 0.07 Serum WBC count ( 10 9 cells/l) 10.4 ± ± ESR (mm/hr) 56.4 ± ± CRP (range) (mg/l) 4.9 (0 11) 0.8 (0 2) 0.53 Joint fluid WBC (range) (cells/µl) 30,000 (600 62,200) Fig year-old boy with septic arthritis of right knee. Maximum synovial thickness measures 4.11 mm along anterior aspect of suprapatellar bursa on this sagittal T1-weighted postcontrast MR image with fat suppression. 16,150 (1, ,000) Note. ESR = erythrocyte sedimentation rate, CRP = C-reactive protein. Age, WBC, and ESR are presented as mean ± SD and were compared using the two-sample Student s t test. CRP and joint fluid WBC are given as medians with ranges and compared by the Mann-Whitney U test. Sex and fever were compared using Fisher s exact test boys and four girls with an age range of years; mean, 11.7 ± 6.6 years. The requirement for the diagnosis of Lyme arthritis was a positive serologic EIA with Western immunoblot confirmation. Septic arthritis was confirmed by joint fluid culture, either by percutaneous aspirate or operative drainage. Joint fluid was aspirated percutaneously within 24 hr before MRI in seven of 11 children with Lyme arthritis and four of seven with septic arthritis. One patient from each group underwent arthroscopic drainage of the knee before MRI. MRI was performed on a 1.5-T MR scanner using a standard extremity coil with the following sequences: axial and sagittal fast spin-echo inversion recovery or T2-weighted spin-echo with fat suppression, sagittal intermediate-weighted spin-echo, and coronal T1-weighted spin-echo. All but two patients also had sagittal T1-weighted images with fat suppression obtained after the IV administration of gadopentetate dimeglumine (Magnevist, Berlex). Images were viewed on the radiology department s Fuji Synapse PAC system (Fujifilm Medical Systems). An experienced pediatric radiologist blinded to the final diagnosis reviewed the MRI examinations. MR images were reviewed for size of joint effusion; presence of marrow, muscle, and subcutaneous edema; synovial thickness; synovial enhancement; and presence of popliteal lymph nodes. The amount of joint fluid was considered small if the suprapatellar AJR:184, June

3 Ecklund et al. bursa was mildly distended, moderate if extension into the posterior recesses of the joint was present, and large if maximal distention of the suprapatellar bursa was present (Fig. 1). Subcutaneous edema was considered present only if it was generalized rather than focal, which could be related to recent joint aspiration. Maximum synovial thickness was measured along the anterior aspect of the suprapatellar bursa on the sagittal postcontrast images using the standard PAC system measurement tool (Fig. 2). The patients clinical records were reviewed for the following data at the time of presentation and MRI: age, sex, presence of fever, serum WBC, serum erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), joint fluid WBC, and Lyme titer results. Institutional review board approval was obtained for review of the MR images and clinical and laboratory records of the 18 patients. Statistical Analysis Univariate and multivariate analyses were performed to identify differences in clinical and MRI variables between patients with Lyme arthritis and those with septic arthritis. All continuous variables were tested for normality to determine the most appropriate statistical tests. CRP and joint fluid WBC were evaluated as median values and ranges with groups compared using the nonparametric Mann- Whitney U test. All other continuous variables, including age, serum WBC, ESR, and synovial thickness, were presented in terms of the mean ± SD and compared using the Student s t test. Categoric data, including sex; fever; presence of marrow, muscle, and subcutaneous edema; and hemarthrosis were compared using Fisher s exact test for binomial proportions. Joint effusion size was assessed using the chi-square test with Yates correction. Multiple logistic regression with the backward stepwise procedure was applied to identify which variables were independently associated with differentiation of Lyme disease and septic arthritis using the likelihood ratio test to determine significance [6]. Data analysis was conducted using the Statistical Package for the Social Sciences (version 12.0, SPSS). Two-tailed values of p < 0.05 were considered statistically significant. TABLE 2 Comparison of MRI Findings Between Lyme and Septic Arthritis Patients MRI Finding Lyme Disease (n = 11) Septic Arthritis (n =7) p Synovial thickness (mm) 4.7 ± ± Size of joint effusion 0.94 Small 3 (27%) 2 (29%) Moderate 4 (36%) 3 (43%) Large 4 (36%) 2 (29%) Marrow edema 0 (0%) 2 (29%) 0.14 Muscle edema 11 (100%) 2 (29%) < 0.01 a Subcutaneous edema 1 (9%) 7 (100%) < 0.01 a Popliteal lymph nodes 11 (100%) 2 (29%) < 0.01 a Hemarthrosis 2 (18%) 0 (0%) 0.14 Note. Synovial thickness is presented as mean ± SD and was compared using the two-sample Student s t test. The presence of marrow, muscle edema, subcutaneous edema, and popliteal lymph nodes was compared using Fisher s exact test. Size of joint effusion was assessed using the chi-square test with Yates correction. Numbers in parentheses are percentages of the total patients in each group. a Statistically significant. Results The comparison of the clinical variables between the patients with Lyme arthritis and those with septic arthritis is shown in Table 1. No significant differences in age, sex, serum WBC, ESR, or CRP were present between the patients with Lyme disease and those with septic arthritis. Although fever was more common in children with septic arthritis, the difference was not statistically significant. Similarly, there was no significant difference in the median joint fluid WBC between the two groups, although all patients with a joint fluid WBC higher than 70,000 cells/µl had septic arthritis. The comparison of MRI findings between the patients with Lyme arthritis and those with septic arthritis is shown in Table 2. All of the children had joint effusions. All 16 patients who received IV contrast showed synovial enhancement. No difference was present in synovial thickness between the two groups of children. Two patients with Lyme arthritis did not have postcontrast imaging. A mild amount of marrow edema was present in two children with septic arthritis but none with Lyme arthritis. All 11 children with Lyme arthritis had edema within the adjacent muscles, especially within the popliteus (Fig. 3), while associated myositis was seen in only two of the seven children with septic arthritis. Similarly, popliteal adenopathy was seen in all patients with Lyme arthritis (Fig. 4), but in just two patients with septic arthritis. Subcutaneous edema was present in all patients with septic arthritis (Fig. 5). Only one child with Lyme arthritis showed subcutaneous edema at MRI, and that examination was obtained 24 hr after surgical drainage. There was no significant difference in the presence of subcutaneous edema between the children who underwent preimaging arthrocentesis and those who did not. Hemarthrosis was seen in two children with Lyme arthritis and none with septic arthritis. Popliteal cysts were present in three patients with Lyme arthritis and none with septic arthritis. Based on univariate analysis, three MRI variables were significantly different between patients with Lyme disease and those with septic arthritis. The findings of muscle edema, subcutaneous edema, and popliteal lymphadenopathy showed statistical significance (all p < 0.01). None of the other imaging variables were significantly different between the two study groups (all p > 0.05). Multiple logistic regression analysis confirmed that the best independent predictor of Lyme arthritis versus septic arthritis was the lack of subcutaneous edema on MRI with a likelihood ratio test of 10.89, p < This implies that independently of all of the other variables evaluated in this study, patients who were found to have subcutaneous edema on MR images were significantly more likely to have septic arthritis compared with Lyme disease. The final logistic regression model indicated that irrespective of the other variables, when subcutaneous edema was absent in the patient groups evaluated, the estimated probability that a patient had Lyme disease was greater than 99% and the probability of septic arthritis was less than 1%. Discussion Lyme disease in children is occurring with increasing frequency in endemic regions in the United States. The highest reported incidence occurs in children ages 5 9 years old [7]. Oligoarthritis is the most common manifestation of late phase Lyme disease in children and usually involves the knee, although the hip, ankle, elbow, and wrist are other common sites [8]. Initial reports claimed that more than half of children had a history of the characteristic erythema migrans rash. More 1906 AJR:184, June 2005

4 MRI of Lyme Arthritis in Children Fig year-old boy with Lyme arthritis of knee. Sagittal T2-weighted fat-suppressed MR image shows joint effusion, synovial thickening, and high-signal-intensity fluid within popliteus muscle (arrow). Fig year-old girl with septic arthritis of left knee. Sagittal intermediateweighted fat-suppressed MR image reveals joint effusion, synovial thickening, and extensive high-signal-intensity edema within subcutaneous tissues (arrow). Fig year-old boy with Lyme arthritis of left knee. Sagittal T2-weighted fat-suppressed MR image shows suprapatellar effusion, synovial hypertrophy, and popliteal lymph nodes (arrows). recently, however, it has been recognized that fewer than 20% of children with Lyme arthritis ever had erythema migrans. With the increasing recognition of this and other clinical and laboratory similarities between Lyme arthritis and acute bacterial arthritis, more groups are reporting diagnostic uncertainty when children present with an acute inflammatory monoarthritis [9]. Articular involvement due to B. burgdorferi may develop from 2 weeks to 2 years after infection or onset of systemic symptoms, making attribution to Lyme disease difficult. Patients usually present with fever, limp, joint pain, and elevated acute phase reactants, all of which are also seen in children with acute septic arthritis. Radiographs reveal joint effusion in both groups. Our finding that clinical and laboratory parameters, including serum WBC, ESR, CRP, or joint fluid WBC, cannot distinguish between Lyme and septic arthritis are consistent with earlier reports [9]. MRI may be requested in children with acute infectious arthritis to assess severity or to exclude associated osteomyelitis. Our results identified three MRI features, specifically myositis, lymphadenopathy, and lack of subcutaneous edema, that help differentiate Lyme arthritis from septic arthritis of the knee in children (Fig. 6). AJR:184, June

5 Ecklund et al. A C Fig year-old girl with Lyme arthritis of right knee. A, Sagittal T2-weighted fat-suppressed image shows joint effusion, synovial thickening, and high-signal-intensity masslike lesion posteriorly (arrow). B, Sagittal T1-weighted postcontrast fat-suppressed MR image at the same level as A confirms that posterior lesion is distended popliteal cyst with thickened, enhancing synovium peripherally (arrow). C, Sagittal T1-weighted postcontrast fat-suppressed MR image just medial to B shows associated lymphadenopathy (white arrows) and myositis (black arrow). All inflammatory arthritides result in synovial hyperemia. It is not surprising that joint effusion, synovial hypertrophy, and enhancement were seen in both groups of patients, with no significant difference in synovial thickness. In contrast, myositis and lymphadenopathy were seen in all of the children with Lyme arthritis but in only two with septic arthritis. Lyme myositis has been reported [10], so this finding may represent a fundamental difference in the tissue tropism of borellial and bacterial causes of arthritis. Conversely, edema within the subcutaneous tissues was seen in all patients with septic arthritis and in only one child with Lyme. Similar proportions of patients with Lyme arthritis (64%) and septic arthritis (57%) underwent percutaneous joint aspiration before MRI. The effect of this procedure on the MRI findings is likely negligible with the exception of insignificant foci of subcutaneous edema related to needle insertion. Arthroscopy, on the other hand, is more invasive and likely complicates the MRI appearance of the joint, increasing the soft-tis- B 1908 AJR:184, June 2005

6 MRI of Lyme Arthritis in Children sue swelling and synovial enhancement. The one patient with Lyme arthritis who exhibited subcutaneous edema was also the only patient in that group who had undergone preimaging arthroscopy, likely accounting for the finding. Overlying cellulitis is much more common in bacterial septic arthritis, which may account for the increased incidence of superficial edema in those patients. Marrow edema, a hallmark of osteomyelitis, was rare in both groups, as would be expected. In addition, the MR images of two patients with Lyme arthritis revealed hemarthrosis. This feature of Lyme arthritis has been described [11] and may be the result of subclinical trauma to the hyperemic synovium. Although not seen in our series, hemarthrosis has also been reported in septic arthritis and probably cannot be used as a distinguishing feature. One recent study of Lyme arthritis in children stated that MRI was not useful in the diagnostic evaluation [9]. However, only two of 10 patients in that study underwent MRI and no control patients with bacterial septic arthritis were used for comparison. Differences between our conclusions and those of that study may be related to the small number of MR images that they evaluated. The results of this study are limited by its retrospective nature and the small sample size. This is in part because patients thought to have septic arthritis are not routinely referred for MRI on acute presentation at our institution. MRI is most often requested when there is lack of clinical improvement after joint aspiration and antibiotic therapy. It is interesting to note that eight of the 18 total patients underwent MRI in the last year of the 6- year study period. This coincides with the time of our initial observation that associated soft-tissue findings might help differentiate Lyme from septic arthritis. It is likely that the increase in referrals for MRI relates to increased awareness of these findings on the part of referring physicians. Our study was confined to a comparison of the MRI findings of Lyme arthritis with those of acute septic arthritis in children. We did not evaluate other common inflammatory monoarthridites, such as juvenile rheumatoid arthritis (JRA). JRA is somewhat less likely to be confused clinically with acute infectious arthritis. Some children with Lyme arthritis, however, come to medical attention in the chronic phase of their illness, when the disease is more likely to be confused with JRA. It will be important for future studies to compare the MRI features of Lyme arthritis with those of JRA. Despite the small number of patients, we identified three MRI features in children with acute onset of arthritis of the knee myositis, lymphadenopathy, and lack of subcutaneous edema that are more suggestive of Lyme than septic arthritis. In geographic regions where Lyme disease is endemic, MRI may play an important role in the diagnosis and management of children with acute arthritis. This should not be interpreted as a recommendation for MRI in all children presenting with acute arthritis, but rather as a report that MRI can offer additional information in those cases where clinical and laboratory data are inconclusive for septic arthritis. If myositis, lymphadenopathy, and especially lack of subcutaneous edema are identified, Lyme disease should be suggested. However, caution should be used when MRI is performed after operative drainage procedures that may alter the findings, especially with respect to subcutaneous edema. Since our initial recognition of the soft-tissue features accompanying Lyme arthritis, we have found that the radiologist is often the first to suggest the diagnosis. Confirmation of our findings would allow educational efforts to be directed toward disseminating these MRI manifestations of Lyme arthritis among other caregivers responsible for the diagnosis and treatment of children with acute arthritis, including emergency department physicians, orthopedists, and rheumatologists. References 1. Centers for Disease Control and Prevention. Final 2002 reports of notifiable diseases. MMWR 2003;52: Orloski KA, Hayes EB, Campbell GL, Dennis DT. Surveillance for Lyme disease: United States, MMWR CDC Surveill Summ 2000;49: Steere AC. Lyme disease. N Engl J Med 2001;345: Culp RW, Eichenfield AH, Davidson RS, Drummond DS, Christofersen MR, Goldsmith DP. Lyme arthritis in children. An orthopaedic perspective. J Bone Joint Surg Am 1987;69: Bunikis J, Barbour AG. Laboratory testing for suspected Lyme disease. Med Clin North Am 2002;86: Hosmer DW, Lemeshow S. Interpretation of the fitted logistic regression model. In: Applied logistic regression. New York: Wiley, 2000: Marshall S, Hayes EB, Dennis DT. Lyme disease: United States, MMWR 2002;51: Gerber MA, Zemel LS, Shapiro ED. Lyme arthritis in children: clinical epidemiology and longterm outcomes. Pediatrics 1998;102: Willis AA, Widmann RF, Flynn JM, Green DW, Onel KB. Lyme arthritis presenting as acute septic arthritis in children. J Pediatr Orthop 2003;23: Ilowite NT. Muscle, reticuloendothelial, and late skin manifestations of Lyme disease. Am J Med 1995;98[suppl 4A]:63S 68S 11. Seldes R, Glasgow SG, Torg JS. Atraumatic spontaneous hemarthrosis associated with Lyme arthritis. Clin Orthop Relat Res 1993;297: AJR:184, June