Performance of Fibromyalgia Rapid Screening Tool (FiRST) to detect fibromyalgia syndrome in rheumatic diseases

Transcription

1 RHEUMATOLOGY Rheumatology 2016;55: doi: /rheumatology/kew244 Advance Access publication 15 June 2016 Concise report Performance of Fibromyalgia Rapid Screening Tool (FiRST) to detect fibromyalgia syndrome in rheumatic diseases Angelique Fan 1, Anne Tournadre 1, Bruno Pereira 2, Zuzana Tatar 1, Marion Couderc 1, Sandrine Malochet-Guinamand 1, Sylvain Mathieu 1, Martin Soubrier 1 and Jean-Jacques Dubost 1 Abstract Objective. To evaluate the performance of the Fibromyalgia Rapid Screening Tool (FiRST) selfquestionnaire for the detection of FM associated with inflammatory rheumatic diseases. Methods. This cross-sectional, French single-centre study was carried out between September 2014 and April 2015 in all patients who consulted for RA, SpA or CTD. Diagnosis of FM was based on ACR 90 criteria and rheumatologist opinion. Results. The self-questionnaire was completed by 605 patients (279 RA, 271 SpA, 57 CTD). It detected 143 concomitant FMs (24.4%). When assessed against ACR 90 criteria, FiRST had a sensitivity of 74.5%, a specificity of 80.4%, a positive predictive value of 26.6% and a negative predictive value (NPV) of 97.1%. Specificity was lower in the CTD group (RA: 84.4%, SpA: 80.2%, CTD: 59.6%) (P = 0.001). When assessed against the rheumatologist s opinion, FiRST had a sensitivity of 75.8%, a specificity of 85.1%, a positive predictive value of 48.3% and an NPV of 95%. Sensitivity was lower in the SpA group than in the CTD group (66% vs 94.4%) (P = 0.004). Performance varied according to self-questionnaire items. CLINICAL SCIENCE Conclusion. Although it performs less well in inflammatory rheumatic disease, FiRST s opinion is close to that of the rheumatologist. It can be used by the rheumatologist in clinical practice for patients facing an apparent treatment failure and to rule out a potential FM diagnosis which could interfere with the treatment response. Key words: fibromyalgia, inflammatory rheumatic disease, FiRST Rheumatology key messages. The diagnosis of concomitant fibromyalgia is complex, especially in spondyloarthritis.. The Fibromyalgia Rapid Screening Tool showed good performance for detecting fibromyalgia associated with rheumatic disease.. The Fibromyalgia Rapid Screening Tool can aid rheumatologists in everyday practice. Introduction A European study estimated the prevalence of FM syndrome to be 4.7% in the general population and it is more 1 Rheumatology Department and 2 Biostatistics Unit, DRCI, CHU Gabriel-Montpied, Clermont-Ferrand, France Submitted 5 October 2015; revised version accepted 13 May 2016 Correspondence to: Jean-Jacques Dubost, Rheumatology Department, CHU Gabriel-Montpied, 58 rue Montalembert, Clermont-Ferrand Cedex 1, France. jjdubost@chu-clermontferrand.fr frequent in patients with inflammatory rheumatic diseases [1]. The prevalence has been estimated to be between 6.6 and 17% in patients with RA [2 4], 4 15% in patients with AS [5, 6], 5 22% in patients with SLE [4, 7 9] and 12% in patients with SS [8]. Overlapping symptoms between inflammatory rheumatic diseases and FM make the differential diagnosis difficult and lead to overestimation of disease activity, causing non-justified therapeutic intensification especially in patients with non-radiographic axial spondyloarthritis [2 5, 10 12]. Self-questionnaires for easy detection of FM have been developed and validated.! The Author Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For Permissions, please journals.permissions@oup.com

2 Performance of Fibromyalgia Rapid Screening Tool in rheumatic diseases Currently, the most rapid and best performing tool is the Fibromyalgia Rapid Screening Tool (FiRST), created by the French Society of Rheumatology and validated by Perrot et al. [13]. It is used to detect FM in <3 min with a sensitivity of 90.5% and a specificity of 85.7%. Validation was made in a population of 162 patients with FM (meeting ACR 90 criteria) or rheumatic disease (RA, AS, OA) [13]. The Spanish-translated version of FiRST was validated in a population with chronic pain of more varied aetiology and showed a sensitivity of 89% and a specificity of 55% [14]. All the tools have been validated in order to discriminate between FM and other painful conditions. The purpose of the present work was to evaluate the performance of the FiRST questionnaire for detecting concomitant FM in patients with inflammatory rheumatic disease. Methods The study was approved by the local ethics committee, Comité de Protection des Personnes Sud Est VI (IRB: ), and the subject s written consent was obtained from all patients who agreed to participate in the study according to the Declaration of Helsinki Patients and diagnoses All patients who consulted in the department of rheumatology of Clermont-Ferrand University Hospital for RA, SpA or CTD and who had been experiencing joint, muscular or tendon pain for at least 3 months were invited to participate in the study. Classification criteria used were the following: ACR/EULAR 2010 criteria for RA; New York criteria for AS; ASAS criteria for non-radiographic SpA and peripheral SpA; CASPAR criteria for PsA; SLICC/ACR criteria for SLE; American-European consensus group criteria for SS; ACR/EULAR 2013 criteria for scleroderma; and Troyanov et al. criteria for myositis. ACR 90 criteria were used as the gold standard to calculate the performance of FiRST [15]. Study description This cross-sectional, single-centre study was conducted between September 2014 and April The rheumatologist determined the FM status of the patient (rheumatologist judgement and ACR 90 criteria) and was blinded as to the result of the FiRST questionnaire. Patients completed the questionnaire anonymously in an isolated location. FiRST The FiRST self-questionnaire consists of six closed questions covering the different dimensions of FM: widespread pain (item 1), fatigue (item 2), pain characteristics (item 3), non-painful abnormal sensations (item 4), functional somatic symptoms (item 5), sleep and cognitive problems (item 6). Questionnaires with missing data or ambiguous responses, for which the score could not be determined, were not taken into consideration for the calculation of performance measures. The cut-off score was determined using questionnaires for which all items had been completed (n = 557). The cut-off score of 55/6 achieved the highest sensitivity and specificity. Statistical analysis Statistical analysis was carried out by the biostatistician affiliated with the study. Data verification was performed on a random subset of data. Statistical analysis was performed using Stata 13 software (StataCorp LP, College Station, TX, USA). The tests were two-sided, with the type I error set at a = Comparisons between independent groups were performed using Student s t-test or the Mann Whitney test when conditions for the t-test were not met. Comparisons concerning categorical data were performed using the chi-squared test or Fisher s exact test. In addition, the following psychometric properties of FiRST were calculated: internal consistency: determined through Kuder Richardson s alpha coefficient (minimum accepted value: 0.60); item difficulty (proportion of patients with an item or proportion of right answers); item test correlations (corrected item test point-biserial correlation coefficients, also known as a discrimination index); discriminant validity: the kappa coefficient was used for each item to determine accuracy of the FiRST scale with regard to ACR 90 criteria and the physician s opinion. Following these analyses, diagnostic values [sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV)] were presented with 95% CIs. The comparisons between pathologies according to these values were performed by the chi-squared test. A receiver operating characteristic (ROC) curve was constructed to evaluate the discriminant power of FiRST. Results Description of patients The total population consisted of 605 patients with an inflammatory rheumatic disease and chronic pain, including 400 women (66.1%) and 205 men (33.9%) of mean age 55.7 (15.1) years. The sample included 279 RA [208 women (75%), 183 RF positive and 189 anti-ccp positive]; 271 SpA [141 women (52%), 52 PsA, 126 AS, 64 non-radiographic axial SpA, 29 peripheral SpA] and 57 CTD [53 women (93%), 21 SLE, 22 SS, 11 scleroderma, 4 myositis]. Exploitable questionnaires were returned by 586 patients. Performance of FiRST A score 55/6 was obtained for 143/586 patients [24.4% (21 28%)]. It detected three times more FM than the ACR 90 criteria (n = 52, 8.6%) and more than the rheumatologist (n = 93, 15.4%). The proportion of male FM is higher with FiRST (n = 36, 25.2%) than with ACR90 criteria (n = 3, 5.8%, P< 0.001) or the rheumatologist (n = 12, 13%, P= 0.05). It detected FM more often in CTD (n = 25, 43.9%) than in RA (n = 50, 18%, P < 0.001) or SpA (n = 58, 21.4%, P < 0.001). The frequency of ACR 90- defined FM was lower in RA than in SpA (P = 0.05) (RA: 5.4%; SpA: 11.1%; CTD: 12.3%). The rheumatologist

3 Angelique Fan et al. TABLE 1 Diagnostic performance of FiRST according to ACR 90 criteria and the rheumatologist s opinion Comparison criteria Sensitivity (%) Specificity (%) PPV (%) NPV (%) Concordance (%) i FiRST vs ACR 90 Total Chronic 74.5 (60.4, 85.7) 80.4 (76.8, 83.7) 26.6 (19.5, 34.6) 97.1 (95.0, 98.4) pain population RA 71.4 (41.9, 91.6) 84.4 (79.4, 88.6) 20.0 (10.0, 33.7) 98.2 ( ) SpA 73.3 (54.1, 87.7) 80.2 (74.5, 85.1) 32.4 (21.5, 44.8) 95.9 ( ) CTD 85.7 (42.1, 99.6) 59.6*(44.3, 73.6) 24.0 (9.36, 45.1) 96.6 ( ) FiRST vs Rheumatologist Total chronic 75.8 (65.7, 84.2) 85.1 (81.6, 88.1) 48.3 (39.8, 56.8) 95.0 (92.6, 96.9) pain population RA 82.6 (61.2, 95.0) 87.5 (82.7, 91.3) 38.0 (24.7, 52.8) 98.2 (95.4, 99.5) SpA 66.0*(51.2, 78.8) 83.5 (77.8, 88.2) 48.5 (36.2, 61.0) 91.2 (86.3, 94.8) CTD 94.4 (72.7, 99.9) 77.8 (60.8, 89.9) 68.0 (46.5, 85.1) 96.6 (82.2, 99.9) *P < FiRST: Fibromyalgia Rapid Screening Tool; NPV: negative predictive value; PPV: positive predictive value. detected also less FM in RA than in SpA (P < 0.003) but evenly in CTD (RA: 8.6%; SpA: 18.5%; CTD: 33.3%; P = 0.001). FiRST did not find a significantly higher frequency of FM in non-radiographic (nr) axial forms of SpA than in other forms of SpA (nr-axial SpA: 36.4%; AS: 22.3%; PsA: 22.9%; P = 0.1) unlike both the ACR 90 criteria (nr-axial SpA: 23.9%; AS: 6.4%; PsA: 9.6%; P = 0.001) and the rheumatologist (nr-axial SpA: 37.3%; AS: 7.2%; PsA: 13.5%; P < 0.001). Convergence with ACR 90 criteria FiRST detected ACR 90-defined FM patients with a sensitivity of 74.5% and a specificity of 80.4%. The NPV was 97% and the PPV was 26.6%. These performance measures translate into 80% of patients being correctly classified (k = 0.3). The ability of FiRST to discriminate between FM and non-fm was reasonable (ROC area under the curve (AUC) = 0.86). Sensitivity did not vary by rheumatic disease type. However, specificity was lower in the CTD group (P = 0.001). A score 54/6 had a sensitivity of 95.7%, a specificity of 61.2%, a PPV of 18.7% and an NPV of 99.3% (Table 1). Convergence with the rheumatologist s opinion The performance of FiRST against the rheumatologist s opinion was comparable to that achieved against ACR 90 criteria. The sensitivity was 75.8%, specificity was 85.1% and concordance was 83.6% (k = 0.49). The PPV was more satisfactory than the one achieved with ACR 90 criteria (48.3%), while NPV was similar (95%). Specificity did not vary by rheumatic disease type. However, sensitivity was lower in the SpA group than in the CTD group (P = 0.004) (Table 1). Performance of each FiRST item in comparison with ACR 90 criteria All six FiRST items displayed acceptable internal validity [Kuder Richardson coefficient: 0.69; item difficulty: 0.50 ( ); item-test correlation: 0.42 ( )]. The performance for each item is presented in Table 2. Item 2, fatigue, and item 6, sleep and concentration, obtained the highest sensitivities and were positive in at least 60% of questionnaires. However, they were also the least specific. Widespread pain (item 1) and non-painful abnormal sensations (item 4) were considered the most specific symptoms for concomitant FM secondary to rheumatic disease. Discussion FiRST detects FM associated with inflammatory rheumatic disease and meeting ACR 90 criteria with acceptable sensitivity (74.5%) and specificity (80.4%). The NPV is excellent (97%), whereas the PPV is poor (26%). The ROC AUC indicates a good ability to discriminate between FM and non-fm. The tool correctly diagnoses 83.6% of patients who were diagnosed with FM by a rheumatologist (k = 0.49) with a sensitivity of 75.8% and a specificity of 85.1%. This shows that the FiRST diagnosis is similar to that of the rheumatologist. Nevertheless, these performance measures are lower than those observed during the tool s validation by Perrot et al. in 2010 [13]; however, their study was conducted in a population of patients with a diagnosis of FM at inclusion. The sensitivity is also lower than that obtained during validation of the Spanish version of FiRST, although our specificity is more performant. However, performance is similar when a FiRST cut-off threshold of 54/6 is used instead of 55/6. In contrast to the Spanish version, our high NPV and low PPV suggest that FiRST has better performance for excluding the diagnosis of concomitant FM rather than confirming it. Its routine use by the rheumatologist would quickly exclude this condition and improve patient care. Overall, the lower performance was expected because signs of rheumatic disease interfere substantially with those of FM. Above all, FiRST does not detect several patients with SpA and concomitant FM identified by the physician. Discrimination between these two diseases is the most difficult because they have several signs in common. FiRST identifies a different population, notably a greater proportion of male FM (female (F)/male (M) ratio: 3/1) than

4 Performance of Fibromyalgia Rapid Screening Tool in rheumatic diseases TABLE 2 Performance of each FiRST item in comparison with ACR 90 criteria Items Sensitivity (%) Specificity (%) ROC AUC Rate of concordant cases i Widespread pain (item 1) 72.5 (58.3, 84.1) 65.8 (61.6, 69.8) Fatigue (item 2) 94.1 (83.8, 98.8) 41.8 (37.6, 46.2) Pain characteristics (item 3) 88.2 (76.1, 95.6) 41.4 (37.2, 45.7) Non-painful abnormal sensations (item 4) 71.4 (56.7, 83.4) 62.0 (57.8, 66.1) Associated somatic disorders (item 5) 84.3 (71.4, 93.0) 55.8 (51.5, 60.0) Effect on sleep and concentration (item 6) 98.1 (89.7, 100) 41.3 (37.1, 45.6) AUC ROC: receiver operating characteristic area under the curve; FiRST: Fibromyalgia Rapid Screening Tool. the rheumatologist (F/M ratio: 7/1) and the ACR 90 criteria (F/M ratio: 16/1). This sex ratio is comparable to that found with the modified ACR 2010 criteria [16]. FiRST identifies a sizeable amount of FM in patients with CTD but not meeting ACR 90 criteria. However, its findings strongly resemble the opinion of the rheumatologist; the agreement is very satisfactory (83.3%, k = 0.66). FiRST items did not all have the same performance. The low specificity and the very high sensitivity of item 2, fatigue, and item 6, effect on sleep and concentration, are expected, because even though they are usual signs of FM, fatigue and sleep impairment due to inflammatory pain are also very common in chronic inflammatory rheumatic diseases [4]. Inversely, widespread pain and non-painful abnormal sensations are symptoms more characteristic of FM. The ACR 90 criteria only consider the pain dimension of this syndrome. However these criteria are the only ones to specify that the presence of a second clinical entity that can explain the chronic pain does not rule out the diagnosis of FM in contrast to the ACR 2010 and modified 2010 criteria [17, 18]. The physician is able to distinguish FM from RA but not from SpA, in which pain is widespread, often axial and identical to FM. FiRST did not have poorer performance than other available tools. Sensitivity was similar to but specificity was higher than the Fibrodetect screening tool [19] and the FM Diagnostic Screen [20]. One strength of this study was the unknown FM status, because tender points were evaluated on the day of inclusion, such that the study population was representative of the target population. The main limitations were the low proportion of FM patients meeting ACR 90 criteria and the limited size of and heterogeneity of the CTD group. Conclusion The FiRST questionnaire is a simple and rapid tool for screening FM. Although it performs less well in inflammatory rheumatic disease, FiRST s opinion is close to that of the rheumatologist. It can be used by the rheumatologist in clinical practice in patients facing an apparent treatment failure and to rule out a potential FM diagnosis which could interfere with the treatment response. Further studies should be considered to highlight its usefulness in patient care and its performances in predicting TNF-a inhibitor treatment failure. Funding: This work received an internal funding from the CHU of Clermont-Ferrand. Disclosure statement: The authors have declared no conflicts of interest. References 1 Branco JC, Bannwarth B, Failde I et al. Prevalence of fibromyalgia: a survey in five European countries. Semin Arthritis Rheum 2010;39: Ranzolin A, Brenol JC, Bredemeier M et al. Association of concomitant fibromyalgia with worse disease activity score in 28 joints, health assessment questionnaire, and short form 36 scores in patients with rheumatoid arthritis. Arthritis Rheum 2009;61: Pollard LC, Kingsley GH, Choy EH, Scott DL. Fibromyalgic rheumatoid arthritis and disease assessment. Rheumatology 2010;49: Haliloglu S, Carlioglu A, Akdeniz D, Karaaslan Y, Kosar A. Fibromyalgia in patients with other rheumatic diseases: prevalence and relationship with disease activity. Rheumatol Int 2014;34: Almodóvar R, Carmona L, Zarco P et al. Fibromyalgia in patients with ankylosing spondylitis: prevalence and utility of the measures of activity, function and radiological damage. Clin Exp Rheumatol 2010;28:S Azevedo VF, Paiva Edos S, Felippe LR, Moreira RA. Occurrence of fibromyalgia in patients with ankylosing spondylitis. Rev Bras Reumatol 2010;50: Middleton GD, McFarlin JE, Lipsky PE. The prevalence and clinical impact of fibromyalgia in systemic lupus erythematosus. Arthritis Rheum 1994;37: Giles I, Isenberg D. Fatigue in primary Sjögren s syndrome: is there a link with the fibromyalgia syndrome? Ann Rheum Dis 2000;59: Friedman AW, Tewi MB, Ahn C et al. Systemic lupus erythematosus in three ethnic groups: XV. Prevalence and correlates of fibromyalgia. Lupus 2003;12: Di Franco M, Iannuccelli C, Bazzichi L et al. Misdiagnosis in fibromyalgia: a multicentre study. Clin Exp Rheumatol 2011;29:S Marchesoni A, Atzeni F, Spadaro A et al. Identification of the clinical features distinguishing psoriatic arthritis and fibromyalgia. J Rheumatol 2012;39:

5 Angelique Fan et al. 12 Joharatnam N, McWilliams DF, Wilson D et al. A crosssectional study of pain sensitivity, disease-activity assessment, mental health, and fibromyalgia status in rheumatoid arthritis. Arthritis Res Ther 2015;17: Perrot S, Bouhassira D, Fermanian J; Cercle d Etude de la Douleur en Rhumatologie. Development and validation of the Fibromyalgia Rapid Screening Tool (FiRST). Pain 2010;150: Torres X, Collado A, Gómez E et al. The Spanish version of the Fibromyalgia Rapid Screening Tool: translation, validity and reliability. Rheumatology 2013;52: Wolfe F, Smythe HA, Yunus MB et al. The American College of Rheumatology 1990 Criteria for the Classification of Fibromyalgia. Report of the Multicenter Criteria Committee. Arthritis Rheum 1990;33: Jones GT, Atzeni F, Beasley M et al. The prevalence of fibromyalgia in the general population: a comparison of the American College of Rheumatology 1990, 2010, and Clinical vignette modified 2010 classification criteria. Arthritis Rheumatol 2015;67: Wolfe F, Clauw DJ, Fitzcharles MA et al. The American College of Rheumatology preliminary diagnostic criteria for fibromyalgia and measurement of symptom severity. Arthritis Care Res 2010;62: Wolfe F, Clauw DJ, Fitzcharles MA et al. Fibromyalgia criteria and severity scales for clinical and epidemiological studies: a modification of the ACR Preliminary Diagnostic Criteria for Fibromyalgia. J Rheumatol 2011;38: Baron R, Perrot S, Guillemin I et al. Improving the primary care physicians decision making for fibromyalgia in clinical practice: development and validation of the Fibromyalgia Detection (FibroDetect Õ ) screening tool. Health Qual Life Outcomes 2014;12: Arnold LM, Stanford SB, Welge JA, Crofford LJ. Development and testing of the fibromyalgia diagnostic screen for primary care. J Womens Health 2012;21: Rheumatology 2016;55:1750 doi: /rheumatology/kew250 Advance Access publication 30 June 2016 Vein inflammation and ultrasound in Behc et s syndrome Venous thrombosis (VT) in Behc et s syndrome (BS) results in part from inflammation. Ambrose et al. [1] found increased wall thickness of veins in BS by MRI, suggesting venous inflammation in the absence of VT. Recently Alibaz-Oner et al. [2]found that one-third of BS patients had venous insufficiency despite no history of VT. We report a 22-year-old patient of Moroccan origin who experienced recurrent lower limb deep VT and oral ulcers. HLA B-51 was present and diagnosis of BS was considered. He was treated with rivaroxaban and colchicine. He consulted in July 2015 for pain in the right thigh along the great saphenous vein. Duplex US (DUS) found hypoechogenic thickening of the wall of the right great saphenous vein, without VT (Fig. 1). An intimal border was observable and was not evocative of VT reperfusion, a finding similar to that already obtained by MRI. Steroids were prescribed but not taken. One month later he experienced arthritis of the right ankle, suggesting BS in an active inflammatory state. He was given 50 mg prednisone. DUS showed a dramatic decrease in venous wall thickness. DUS may be useful for detecting venous inflammation in the event of limb pain in BS, even in the absence of superficial VT. Acknowledgements We thank Ray Cooke for copy-editing. Funding: No specific funding was received from any bodies in the public, commercial or not-for-profit sectors to carry out the work described in this manuscript. Disclosure statement: The authors have declared no conflicts of interest. Carine Boulon 1, Sophie Skopinski 1 and Jo el Constans 1 1 Service de médecine vasculaire, H^opital St André, Bordeaux, France Correspondence to: Joel Constans, Médecine vasculaire, H^opital St André, 1 rue Jean Burguet, Bordeaux 33075, France. joelconstans@gmail.com References 1 Ambrose N, Pierce IT, Gatehouse PD, Haskard DO, Firmin DN. Magnetic resonance imaging of vein wall thickness in patients with Behc et s syndrome. Clin Exper Rheumatol 2014;32:S Alibaz-Oner F, Karatay E, Akpinar IN, Ergun T, Direskeneli H. Evaluation of asymptomatic venous disease by venous Doppler ultrasonography in patients with Behc et s disease without overt thrombosis. Clin Rheumatol 2014;33: FIG. 1Hypoechogenic thickening of the wall of the great saphenous vein! The Author Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For Permissions, please journals.permissions@oup.com