To assess the impact of ultrasonography (US) on local corticosteroid (CS) injections of painful ankle, hindfoot, and midfoot in chronic inflammatory diseases.
To assess the impact of ultrasonography (US) on local corticosteroid (CS) injections of painful ankle, hindfoot, and midfoot in chronic inflammatory diseases.
Consecutive patients with chronic rheumatic diseases admitted to the hospital for local CS injections of painful ankle, hindfoot, or midfoot were enrolled in this study. Clinical and radiographic evaluation was performed by the physician in charge of CS injections, and US examination was performed by an independent examiner blinded to the clinical and radiologic findings. According to a randomized weekly-periods design, the physician planned CS injections either aware (G1 group) or unaware (G2 group) of US results. In the latter case, he was nonetheless informed of US results after he had performed the injections. Impact of this information on the treatment planning was assessed in all cases. Prognostic impact of US was also evaluated by comparing the change in global assessment of efficacy of CS injections, in activity of the disease, and in the Western Ontario and McMaster Universities (WOMAC) subscales after 1 and 3 months, between G1 and G2 groups.
The knowledge of US findings led the physician to change his decision of local CS injections in 56 (82%) of 68 patients studied. Among 1,131 assessed sites, by clinical, radiographic, and US evaluation, injection was cancelled in 37 (15%) of 242 proposed sites, whereas it was decided in 74 (8%) additional sites. After 1 month, there was an improvement in G1 as compared with G2 groups. The mean ± SD change in WOMAC physical function subscale was 15.6 ± 17.5 in G1 versus 8 ± 13 in G2 (P = 0.0305). After 3 months, only the global assessment of efficacy of CS injections was statistically greater in G1 than in G2 group (P = 0.0170).
US frequently led the physician to change his diagnosis of inflammatory lesions in painful foot, and consequently his planning of CS injections. Moreover, US could improve the response to local treatment.
Foot structures are commonly affected during the course of chronic inflammatory diseases such as rheumatoid arthri tis (RA) or spondylarthropathies (SpA) (1, 2). Their involvement induces local pain and deformity, leading to standing and walking impairment. Functionally, the foot can be divided into 4 anatomic units: ankle, hindfoot, midfoot, and forefoot, which comprise various articular or tenosynovial structures. All of these structures can be involved in the painful foot, which may induce considerable dysfunction (1, 3). It is likely that chronic synovitis, which induces joint destruction, represents a major primary cause of deformity and pain (4). Thus, treatment of chronic synovitis might prevent the subsequent destruction and deformities.
In the management of chronic inflammatory diseases, corticosteroid (CS) injection therapy is still of great importance in the local treatment of arthritic joints (5–8). Also, the use of imaging techniques helps the physicians to evaluate the progression of disease and to determine the best methods of management (3–5, 9–15). In clinical practice, the detection of involved structures is usually made by the clinical examination associated with radiographic evaluation (4, 10, 16, 17).
There is accumulating evidence that the detection of involved sites, i.e., joint or tendon, and their degree of inflammation is underestimated by clinical and radiographic evaluation (1, 17). Therefore the response to CS injections may vary considerably among patients because the physician can perform the injection in an unaffected site or can fail to inject all affected sites (18–21). Thus, radiographic guidance for injection of certain joints is strongly recommended to ensure accuracy of injection (18). Moreover, the presence of inflammatory signs (synovitis or effusion) seems to correlate with the degree of improvement (22). An improved outcome after the use of magnetic resonance imaging to show synovitis in the subtalar joint, and after injection into both tibiotalar and subtalar joints using radiographic guidance, was demonstrated in children (23).
Over the last years, ultrasonography (US) has already proved to be an excellent, noninvasive, and economic instrument to detect joint and tendon involvement in rheumatic diseases (12–15, 24–26). It has also been suggested that US guiding could improve the efficacy of local CS injections in inflammatory rheumatism (27–31). Nevertheless, despite such results, and considering that appropriate therapy might strongly depend on accurate diagnosis, studies addressing the impact of US on the anatomic diagnosis by physicians and on therapeutic decision are scarce (32).
The objective of this study was to evaluate the impact of US on therapeutic decision and on the prognosis of local CS injection performed as a treatment of painful ankle, hindfoot, and midfoot in chronic inflammatory diseases.
Consecutive patients with SpA fulfilling the criteria of Amor et al (33) or the European Spondylarthropathy Study Group criteria (34), or with RA fulfilling the American College of Rheumatology (formerly the American Rheumatism Association) criteria (35), hospitalized in our department of rheumatology for local CS injection of painful ankle, hindfoot, and midfoot were eligible for the study. Inclusion criteria required that the symptomatic clinical involvement of the foot lasted >2 weeks, and that the referring physician was not the clinical or the US investigator responsible for conducting the study. Exclusion criteria were contraindication to local CS injection and previous surgery of the foot.
At baseline, each patient enrolled in the study had clinical, radiologic, and US evaluations. We also recorded age, sex, diagnosis (RA or SpA), and disease duration. An experienced rheumatologist (XA), in charge of local CS injections, performed clinical evaluation and radiographic interpretation, whereas an independent examiner (MADA) performed and interpreted US examination blinded to diagnosis and to the clinical and radiographic findings.
The physician evaluated all joints, tendons, and bursa of the ankle, hindfoot, and midfoot of included patients. The decision to perform CS injections was based on the presence of inflammatory signs, i.e., synovitis, tenosynovitis, and bursitis. The studied structures were tibiotalar, subtalar, talonavicular, calcaneocuboidal, cuneonavicular, tarsometatarsal joints, the tibialis anterior and posterior tendons, the peroneus longus and brevis tendons, and calcaneal and retrocalcaneal bursa. The presence of synovitis was determined by the presence of spontaneous pain at the site of the joint, and/or local swelling of the joint, and/or tenderness during pressure and mobilization of the joint.
Tenosynovitis was determined by the presence of spontaneous pain at the site of the tendon and/or local swelling along the tendon sheath, and tenderness during pressure and/or during stretching and contraction of the tendon against resistance. The involvement of calcaneal and retrocalcaneal bursa (bursitis) was determined by the presence of spontaneous pain at the site of the bursa, and swelling or tenderness during pressure of the bursa.
The Western Ontario and McMaster Universities (WOMAC) instrument (36), a multidimensional, self-administered health status instrument utilized for this study to evaluate ankle, hindfoot, and midfoot, and the patient global assessment of disease activity of these structures were also obtained for each patient at baseline, and at 1 and 3 months after injections.
Patient assessment of pain, stiffness, and physical function (WOMAC), and patient global assessment of disease activity in the ankle, hindfoot, and midfoot were performed using 5-point Likert scales (none, slight, moderate, severe, extreme).
Each patient underwent a bilateral radiographic series at baseline including anteroposterior view of the ankle, lateral view of the ankle and foot, and dorsoplanar and medial oblique views of the feet. The anteroposterior radiograph of the ankle combined with the lateral view allowed the physician to explore the tibiotalar and subtalar joints and to seek calcaneal bursitis, with or without adjacent calcaneum erosion. The dorsoplanar and medial oblique views allowed the physician to explore the hindfoot and midfoot. The radiographic involvement of joints was defined as a reduction of joint space width >25%, as compared with the opposite side. (37).
US was performed using ESAOTE AU5 Epi (ESAOTE, Genoa, Italy) with a 10–13 MHz linear probe at the same anatomic sites listed above. Each joint and tendon was scanned in both longitudinal and transverse planes. In B mode, the following abnormal findings were considered features of tenosynovitis: the presence of an anechoic area within the tendon sheath, associated with a well-defined area of decreased echogenicity, i.e., synovial thickening (38). Joint effusion was identified as an anechoic area in the joint space. The presence of a well-defined area of decreased echogenicity within capsule and joint space was considered to indicate synovial thickening (39). In addition, an enlarged bursa was considered evidence of bursitis (38, 39). A comprehensive analysis of normal US features in B mode was not specifically conducted for the purpose of this study because normal findings of the structures examined have already been extensively described by others (38, 40). In particular, we considered a joint normal when the capsule stretched linearly between the 2 bones, delimiting the joint, and when the only anechoic area present in the joint was the cartilage line.
After performing clinical examination and radiographic interpretation, the physician recorded the anatomic structures involved by synovitis and planned CS injections without knowledge of US findings. Then, according to the study design, the physician was informed of US results either before or after performing the planned CS injections. In all cases he recorded his decision to change or keep his diagnosis (presence or absence of synovitis) and thus the anticipated treatment plan for each site examined in a given patient. All CS injections were performed under radiographic guidance to control the correct position of the needle.
We conducted a prospective, weekly-periods, randomized, controlled study (Figure 1). The weeks were randomized into intervention or control weeks. During the intervention weeks, US results were given to the physician before he performed the CS injections. The patients enrolled during the intervention weeks belonged to the intervention group (G1). In this group, a change in the diagnosis of involved structures and consequently a change in the plan of CS injections was defined as modification a priori. Patients included during control weeks belonged to the control group (G2). In this latter group, the clinician performed CS injections without knowing US results. Afterwards, he was informed of US results and was asked whether US results would have changed his diagnosis and his planning of CS injections. These changes were defined as modifications a posteriori.
The impact of US on the clinician's diagnosis was assessed by calculating the percentage of involved structures for which the clinician modified his final diagnosis after knowing US results. The impact of US on clinician's planning of CS injections was assessed by calculating the proportion of change of the sites of CS injections after knowing US results, in the entire population. Changes a priori in G1 group, and a posteriori in G2 group, were also calculated.
The prognostic impact of US on clinical outcome of CS injections was assessed by comparing clinical response at 1 and 3 months between G1 and G2 groups, for the following parameters: patient's global assessment of treatment efficacy (response to question was measured on a 5-point Likert scale: very good, good, moderate, slight, none); patient's global assessment of disease activity in ankle, hindfoot, and midfoot; and mean change in pain, stiffness, and physical function (WOMAC). The comparison after 1 and 3 months was made for all feet assessed, and thereafter only for the feet for which the planning of CS injections was changed for at least one site (G1+ and G2+ groups).
The differences were tested using the chi-square, Fisher's exact, Student's t, and Wilcoxon's rank sum (for paired and unpaired data) tests as appropriate. A P value < 0.05 was considered statistically significant.
Sixty-eight patients (48 RA and 20 SpA) were enrolled in this study. Eighty-seven feet (both feet in 19 patients) were involved. We assessed clinically, radiographically, and by US 1,131 sites, 559 in the G1 group and 572 in the G2 group. Patient characteristics are shown in Table 1. There were no significant differences in the number of patients, proportions of SpA and RA, age, sex distribution, and number of feet involved between G1 and G2 groups. However, the groups statistically differed in the duration of disease, which was longer in the G2 group (P = 0.0154). They statistically differed also on the patient global assessment of disease activity of hindfoot on a 5-point Likert scale (P = 0.0114), but not on WOMAC score.
|G1 (n = 34)||G2 (n = 34)||P†|
|Women, no. (%)||19 (55.9)||20 (58.8)||NS|
|Age, mean ± SD years||47.2 ± 17.7||51.1 ± 10.9||NS|
|Rheumatoid arthritis, no. (%)||22 (64.7)||26 (76.5)||NS|
|Spondylarthropathies, no. (%)||12 (35.3)||8 (23.5)||NS|
|Disease duration, mean ± SD years||7.2 ± 6.5||10.9 ± 7.4||0.015|
|One foot involved, n (%)||25 (73.5)||24 (70.6)||NS|
|Both feet involved, n (%)||9 (26.5)||10 (29.4)||NS|
|Patient global assessment of disease activity on AHMF, n (%)||0.011|
|None||0 (0)||0 (0)|
|Slight||0 (0)||7 (15.9)|
|Moderate||24 (58.5)||14 (31.8)|
|Severe||11 (26.8)||16 (36.4)|
|Extreme||6 (14.6)||7 (15.9)|
|WOMAC score on AHMF (range 0–100)|
|Patient assessment of pain, mean ± SD||47.0 ± 19.0||47.9 ± 23.4||NS|
|Patient assessment of stiffness, mean ± SD||51.7 ± 27.7||53.3 ± 23.2||NS|
|Patient assessment of physical function, mean ± SD||40.8 ± 17.6||44.8 ± 19.1||NS|
The clinician's diagnosis based on his clinical examination and radiologic interpretation disagreed with US findings in 333 (29.5%) of 1,131 sites (lack of information about possible discrepancy in 4 sites). The clinician changed his diagnosis in 108 (32.4%) of these 333 disagreements; among these 108 sites, 72 were reconsidered as involved sites and 36 were reconsidered as uninvolved sites.
Disagreements between physician's diagnosis and US findings in the G1 and G2 groups were very similar. In G1, such disagreements were observed in 156 sites (27.9%). The physician changed his diagnosis in 46 (29.4%) of these disagreements: 28 out of 46 sites were reconsidered as involved sites and 18 were reconsidered as normal sites. In G2, such disagreements were observed in 177 sites (31.2%). The physician changed his diagnosis in 62 (35.0%) of these disagreements: 44 out of 62 sites were reconsidered as involved sites and 18 were reconsidered as normal sites.
The clinician reported at least one change of the anticipated treatment plan of CS injections in 56 (82.5%) of 68 patients. The knowledge of US findings lead to a modification a priori (G1) and a posteriori (G2) of the treatment plan in 26 (76.4%) and 29 (85.2%) patients, respectively.
When analyzing the total population, the clinician initially planned to inject 242 sites and not to inject 889 sites (125 and 434 in G1, 117 and 455 in G2, respectively). Knowing US results led more to abandon injections initially planned than to add injections. Abandonment of planned injections was decided in 37 sites (15.3% of the sites initially planned), whereas the addition of unplanned injection was decided in 74 sites (8.3% of the sites finally injected, P = 0.0012) (Table 2). Comparable rates of change were observed both in G1 and G2 (15.2% and 15.4% of the abandoned sites, and 6.7% and 9.9% of the final sites added, respectively). The sites for which the clinician abandoned the planned injections most frequently were the tibiotalar (18.6%) and tarsometatarsal (16.0%) joints, and the retrocalcaneal bursa (33.3%). The sites for which the US results lead the clinician to most frequently add the initially unplanned CS injection were the cuneonavicular (58.3% of the final sites), calcaneocuboidal (35.8%), and subtalar (34.5%) joints, and the tibialis posterior tendon (33.3%). Details of the treatment plan modifications introduced after knowledge of US results are listed in Table 2, whereas Figure 2 shows 2 examples of US involved joints.
|Decided beforehand n (n/examined)||Abandoned afterwards n (n/decided)||Unchanged afterwards n (n/decided)||Added afterwards n (n/decided + added)||Performed afterwards n (n/examined)||Not planned and unchanged n (n/examined)|
|Talocrural||59 (67.8)||11 (18.6)||48 (81.3)||6 (9.2)||54 (62.0)||22 (25.2)|
|Subtalar||36 (41.3)||3 (8.3)||33 (91.6)||19 (34.5)||52 (59.7)||32 (36.7)|
|Talonavicular||37 (42.5)||5 (13.5)||32 (86.4)||8 (17.7)||40 (45.9)||42 (48.2)|
|Cuneonavicular||10 (11.4)||0 (0)||10 (100)||14 (58.3)||16 (18.3)||71 (81.6)|
|Calcaneocuboid||25 (28.7)||4 (16.0)||21 (84.0)||14 (35.8)||35 (40.2)||48 (55.1)|
|Tarsometatarsa||25 (28.7)||4 (16.0)||21 (84.0)||4 (13.7)||25 (28.7)||58 (66.6)|
|Tibialis posterior||10 (11.4)||1 (10)||9 (90.0)||5 (33.3)||14 (16.0)||72 (82.7)|
|Fibular||22 (25.2)||3 (13.6)||19 (86.3)||9 (29.0)||28 (32.1)||56 (64.3)|
|Tibialis anterior||0 (0)||0 (0)||0 (0)||1 (100)||1 (1.1)||86 (98.8)|
|Retrocalcaneal||9 (10.3)||3 (33.3)||6 (66.6)||2 (18.1)||8 (9.1)||76 (87.3)|
|Calcaneal||5 (5.7)||3 (60)||2 (40.0)||0 (0)||2 (2.3)||82 (94.2)|
|Plantaris fascia||4 (5.5)||0 (0)||4 (100)||0 (0)||4 (4.5)||83 (95.4)|
|Total||242 (21.3)||37 (15.3)||205 (84.7)||74 (23.4)||279 (24.6)||815 (72.0)|
Forty-three and 44 feet were assessed in G1 (modification a priori) and G2 (modification a posteriori), respectively (the information was missing for 3 feet). After 1 month, although the efficacy of treatment was reported as at least moderate for 38 (92.7%) feet by G1 patients and 32 (80%) feet by G2 patients, the 2 groups did not differ statistically in the patient global assessment of efficacy of CS injections. Similarly, the global assessment of disease activity in ankle, hindfoot, and midfoot was reported as none or slight for 13 (31.7%) feet by G1 patients (0 foot had none at entry) and 10 (25%) feet by G2 patients (7 feet had none at entry); there was no statistically significant difference between the 2 groups.
However, there was a greater improvement in G1 than in G2 in WOMAC subscales of physical function and stiffness (Table 3). Nonetheless, the 2 groups statistically differed only in mean change of physical function (P = 0.0305). Although the mean improvement in WOMAC pain subscale was greater in G2 than in G1, the 2 groups did not statistically differ (Table 3).
|G1 group (n = 34)||G2 group (n = 34)||P†|
|1 Month patient global assessment of efficacy of treatment, no. (%)||0.3880|
|Very good||5 (12.2)||6 (15.0)|
|Good||23 (56.0)||16 (40.0)|
|Moderate||10 (24.3)||10 (25.0)|
|Slight||3 (7.3)||5 (12.5)|
|None||0 (0)||3 (7.5)|
|1 Month patient global assessment of disease activity in AHMF, no. (%)||0.3304|
|None||3 (7.5)||0 (0)|
|Slight||10 (25.0)||10 (25.0)|
|Moderate||16 (40.0)||14 (35.0)|
|Severe||6 (15.0)||11 (27.5)|
|Extreme||5 (12.5)||5 (12.5)|
|1 Month WOMAC score, mean ± SD (range 0–100)|
|Change in patient assessment of pain||10.2 ± 22.1||19.2 ± 20.9||0.0654|
|Change in patient assessment of stiffness||17.3 ± 25.5||12.3 ± 26.1||0.3876|
|Change in patient assessment of physical function||15.6 ± 17.5||8.0 ± 13.0||0.0305|
|3 Month patient global assessment of efficacy of treatment, no. (%)||0.0170|
|Very good||14 (37.8)||4 (10.2)|
|Good||9 (24.3)||16 (41.0)|
|Moderate||7 (18.9)||11 (28.2)|
|Slight||3 (8.1)||7 (17.9)|
|None||4 (10.8)||1 (2.5)|
|3 Month patient global assessment of disease activity of AHMF, no. (%)||0.1704|
|None||5 (13.5)||5 (12.8)|
|Slight||12 (32.4)||6 (15.3)|
|Moderate||7 (18.9)||17 (43.5)|
|Severe||10 (27.0)||9 (23.0)|
|Extreme||3 (8.1)||2 (5.1)|
|3 Month WOMAC score, mean ± SD (range 0–100)|
|Change in patient assessment of pain||20.3 ± 23.3||13.5 ± 20.6||0.1817|
|Change in patient assessment of stiffness||22.1 ± 26.9||22.4 ± 27.0||0.9561|
|Change in patient assessment of physical function||17.0 ± 21.6||11.5 ± 17.7||0.2379|
After 3 months, the clinical outcome of treatment, i.e., patient global assessment of efficacy of CS injections, was statistically stronger among patients in G1 than in G2 (P = 0.017) (Table 3). The efficacy of CS injections was reported as at least good for 23 (62.1%) feet by G1 patients and 20 (51.2%) by G2 patients. The improvement in WOMAC pain and physical function subscales was greater in G1 than in G2, but there was no statistically significant difference between the 2 groups. Finally, the mean change in WOMAC stiffness subscale was similar in G1 and G2.
When comparing only patients in G1 group for whom the physician's treatment planning was really modified by the knowledge of US results (i.e., G1+ group) with patients in G2 group for whom the knowledge of US results would have changed the planning of CS injections (i.e., G2+ group), a statistically significant difference was observed in the improvement of physical function at 1 month: the mean ± SD change in physical function was 16.3 ± 18.3 and 8.8 ± 12.0, respectively (P = 0.0215) (Table 4). There were no statistically significant clinical differences at 1 and 3 months when comparing patients in G1 without modification a priori (i.e., G1-) with patients in G2 without modification a posteriori (i.e., G2-) (data not shown).
|G1+ (n = 26)||G2+ (n = 29)||P†|
|1 Month patient global assessment of efficacy of treatment, no. (%)||0.3081|
|Very good||3 (10.1)||3 (10.3)|
|Good||18 (60.0)||13 (44.8)|
|Moderate||7 (23.3)||5 (17.2)|
|Slight||2 (6.6)||5 (17.2)|
|None||0 (0)||3 (10.3)|
|1 Month patient global assessment of disease activity in AHMF, no. (%)||0.3129|
|None||1 (3.4)||0 (0)|
|Slight||7 (24.1)||6 (20.6)|
|Moderate||14 (48.2)||11 (37.9)|
|Severe||3 (10.3)||9 (31.0)|
|Extreme||4 (13.7)||3 (10.3)|
|1 Month WOMAC score, mean ± SD (range 0–100)|
|Change in patient assessment of pain||17.9 ± 18.2||11.4 ± 17.0||0.2586|
|Change in patient assessment of stiffness||12.1 ± 20.1||13.4 ± 25.0||0.7429|
|Change in patient assessment of physical function||16.3 ± 18.3||8.8 ± 12.0||0.0215|
|3 Month patient global assessment of efficacy of treatment, no. (%)||0.1881|
|Very good||11 (39.2)||4 (13.3)|
|Good||6 (21.4)||10 (33.3)|
|Moderate||6 (21.4)||10 (33.3)|
|Slight||3 (10.7)||5 (16.6)|
|None||2 (7.14)||1 (3.3)|
|3 Month patient global assessment of disease activity in AHMF, no. (%)||0.2136|
|None||4 (14.2)||3 (10.0)|
|Slight||9 (32.1)||5 (16.6)|
|Moderate||5 (17.8)||13 (43.3)|
|Severe||7 (25.0)||8 (26.6)|
|Extreme||3 (10.7)||1 (3.3)|
|3 Month WOMAC score, mean ± SD (range 0–100)|
|Change in patient assessment of pain||21.8 ± 22.4||17.1 ± 20.5||0.5237|
|Change in patient assessment of stiffness||22.7 ± 26.1||24.1 ± 27.8||0.6914|
|Change in patient assessment of physical function||12.0 (18.2)||19.0 (22.3)||0.3325|
In this study, we have clearly shown that the US findings lead the physician to change his diagnosis of involved structures of the ankle, hindfoot, and midfoot and, consequently, his planning of CS injections. We have also shown that such modifications were associated with a trend towards improved short-term symptomatic treatment effect.
The difficulties in determining clinically and radiographically which structures are precisely involved in painful ankle, hindfoot, and midfoot of rheumatic diseases are well documented. Hence, an accurate, safe, and quickly performed diagnostic examination has long been sought to improve diagnostic accuracy before therapeutic procedures (1, 3, 4). Recently, the importance of US imaging technique in the diagnosis and management of rheumatic diseases has been assessed (13, 24–32).
The planning of CS injections was changed for 56 of 68 patients, and changes could concern any of the structures (joints, tendons, and bursa) studied. In this study, 58% of the injections performed in the cuneonavicular, 36% in the calcaneocuboidal, and 35% in the subtalar joints were not initially planned. Indeed, these joints are difficult to examine. Moreover, as reported in other studies, location of subtalar joint pain and swelling was frequently confused with tibialis posterior and/or fibulars tendons involvement (2–4). These results confirmed the importance of an accurate diagnostic imaging technique to support a clinical judgment.
To our knowledge, this is the first study evaluating the impact of US on the management of chronic inflammatory rheumatism, especially the modification of treatment plans. Although Fineberg's hierarchical step-by-step method of evaluation of diagnostic tests has been widely adopted (41–45), the evaluation of the diagnostic and therapeutic impact of new diagnostic imaging procedure is not systematic (42). In Fineberg's method, the evaluation of diagnostic impact on one hand and of therapeutic on the other are separated through consecutive steps (42, 44, 45). Actually, the impact of diagnostic imaging on the planning and delivery of therapy is closely related to its diagnostic impact (44). In this study, both the diagnostic impact and therapeutic impact were evaluated. The change of diagnosis of involved structures, expressed by the percentage of diagnoses modified after knowledge of US results, recorded a priori and a posteriori, was directly related to the change of CS injections. However, we carried out particular attention on evaluation of therapeutic impact. Moreover, methods that can be used to measure diagnostic and therapeutic impact may be similar. In these evaluations, clinical findings, diagnoses, and management plans should be recorded before and after diagnostic imaging (44).
Several study designs can be used to make these evaluations (42–49). Randomized controlled trial is the best method to prove clinical impact of diagnostic imaging (43, 45). However, randomized studies may be ethically contentious and, logistically and financially, difficult to conduct (44). The controlled observational design, in which each patient is his own control through the record of clinical findings and therapeutic plans before and after diagnostic imaging, is an alternative design. One limitation of this design, the contrary of randomized controlled trial design, is that the participating physician determines the diagnosis and management in the absence of imaging (44). There is nonetheless minimal interference to the normal care afforded to patients (41).
In our study, although the patient was his own control, we used a randomized weekly-period design. Hence, during intervention weeks, the therapeutic management of patients was decided by the clinician informed of US results. Conversely, during control weeks, the physician decided the therapeutic management of patients without US results. However, this experienced rheumatologist used well-experimented diagnostic imaging techniques in clinical practice such as radiographs.
The experimental design used in this study allowed us to demonstrate that knowing US results changes the planning of CS injections. During control weekly periods, the physician was also informed of US results, but only after he had performed the treatment plan. Thus, a feedback effect on the results observed cannot be excluded. The awareness of US results may have blunted assessment of the real US impact.
Finally, as proposed by Lijmer and Bossuyt (50), this study has also assessed prognostic impact of US in chronic inflammatory rheumatism. Actually, although our results showed a better improvement in G1 patients, only the mean change in function disability at 1 month and the patient global assessment of efficacy of treatment at 3 months were statistically greater in G1 than in G2. A lack of study power cannot be excluded for the other comparisons. Indeed, the assessment of the prognostic impact was a secondary objective in this study. Due to the relatively low number of patients, these results should be considered preliminary. Thus, other studies will be necessary. Further steps of Fineberg's method (41–45) should be considered. However, our results show clearly that the use of US in the management of CS injections in chronic inflammatory diseases may have an important impact on the diagnosis and therapeutic deliveries. Furthermore, US results could have a prognostic impact on the efficacy of local treatment of painful ankle, hindfoot, and midfoot.