Cortical discontinuity
A negative finding of cortical discontinuity on PsUS is useful in ruling out femoral proximal fractures (Sn, 0.96), whereas the presence of cortical discontinuity is suggestive of femoral proximal fractures (Sp, 0.92). The diagnostic accuracy of cortical discontinuity for femoral proximal fractures was comparable to that seen in a report using conventional US for fractures in various parts of the body (Sn, 0.94; Sp, 0.92) [3]. Furthermore, the diagnostic accuracy of cortical discontinuity for femoral proximal fractures was comparable to that of radiography (Sn, 0.97-0.98; Sp, 1.00) [16, 17]. The diagnostic accuracy of cortical discontinuity was equal to or more than that of the patellar-pubic percussion test (Sn, 0.79-0.96), which was reported as one of the most representative physical examinations to identify fractures [18, 19]. However, physicians should be attentive to false-negative findings when they cannot detect cortical discontinuity (e.g., localized fracture on the greater trochanter) and false-positive findings when they can only detect cortical discontinuity (e.g., misinterpretation of an acetabular head gap to cortical discontinuity) (Additional file 1 and Additional file 2).
Joint fluid retention
A negative finding of joint fluid retention on PsUS is useful to rule out acute hip arthritis (Sn, 1.00); however, it cannot effectively rule out femoral proximal fracture. Positive findings of joint fluid retention on PsUS lead physicians to consider femoral neck fracture (positive predictive value, 0.73) rather than acute hip arthritis (positive predictive value, 0.27). The diagnostic accuracy of joint fluid retention for femoral neck fracture (Sp, 0.77; 95% CI, 0.56-0.91) is comparable to that of the fat pad sign that suggests joint swelling on radiographic finding (Sp, 0.86; cutoff point, 1.5 mm) [20]. In addition, physicians can increase the diagnostic accuracy by repeated US examinations with no radiation (i.e., to decrease false-negative finding of no joint fluid retention in the hip joint) [14, 21]. To diagnose acute septic or aseptic hip arthritis, joint fluid needle aspiration is recommended [1], which may develop complications, such as vascular or neurological injuries. In this study, we performed joint fluid needle aspiration in 2 of 6 patients (33%) who had acute hip arthritis. The final diagnosis for all the 6 patients was acute aseptic hip arthritis (100%).
Combination of cortical discontinuity and joint fluid retention
PsUS finding of neither cortical discontinuity nor joint fluid retention is useful to rule out femoral proximal fractures and acute hip arthritis. The Sn of either cortical discontinuity or joint fluid retention was 0.97 for femoral proximal fracture or acute hip arthritis. Our study included 6 pubic and ischial fractures (12%); nevertheless, cortical discontinuity or joint fluid retention could not rule out other fractures, except femoral proximal fractures. A previous study to diagnose acute hip pain after post-traumatic injuries with normal radiographic finding showed 8 pubic and ischial fractures (27%), of which 4 (50%) were identified using conventional US [22]. Therefore, physicians should consider pubic and ischial fractures when they are unable to detect cortical discontinuity using US for patients with acute hip pain. In addition to pubic and ischial fractures, physicians should also consider other diseases, such as vascular abnormalities of the hip, soft-tissue abnormalities, and neurogenic causes [23], which were not seen in our study.
Diagnostic accuracy of PsUS
We adopted a new unified approach of PsUS probe handling to evaluate femoral proximal fractures and acute hip arthritis (Figure 2). Recent studies indicated that the accuracy among beginners was comparable to that among well-skilled US examiners for musculoskeletal examinations after a short training of 3 hours for use of US [24] or stepwise from off-the-job to on-the-job educational training for 4 months [25]. Nevertheless, US demonstrates lower intra- and inter-rater reliability than radiography, CT, or MRI because US imaging depends on the examiner’s skill [26-28]. In this study, antalgic positions caused by femoral proximal fractures and acute hip arthritis were sometimes more difficult to examine using PsUS, especially in cases of excessive flexion and internal rotation of the hip joints. In this situation, the diagnostic sensitivity of PsUS might decrease. Furthermore, there could be heuristic bias in that the examiner might estimate the higher possibility of fracture if the patient complained of severe pain when the examiner performed PsUS. Therefore, to reach adequate intra-rater reliability, the PsUS examiner (TA) in our study was well-trained for the unified approach before the study started. Further evaluation of the intra- and inter-rater reliability of the approach should be performed in the future.
The diagnostic accuracy of US depends on the type of US. Higher-quality US imaging may have a higher diagnostic accuracy. Recently, portable and high-image-quality PsUS devices have been rapidly developing around the world. Therefore, we believe that the diagnostic accuracy of PsUS for femoral proximal fractures and acute hip arthritis will further improve.
Clinical perspective
PsUS decreases unnecessary referral to orthopedic specialists by optimizing medical examination for acute hip pain. Physicians are recommended to take radiographs when fracture or dislocation is suspected [1]. However, if a physician cannot use these standard diagnostic devices in rural clinics or nursing homes, PsUS findings of the absence of cortical discontinuity or joint fluid retention can help rule out femoral proximal fractures or acute hip arthritis (Sn, 0.97), leading to a decrease in unnecessary outpatient visits, referrals, and imaging. Physicians must consider referral to orthopedic specialists when they confirm cortical discontinuity and joint fluid retention due to hip joint fractures on US.
Limitations
Our study has the following limitations. First, we could not validate the reliability of PsUS for the patient’s body weight because we could not use a stretcher scale with beds. The quality of the US image depends on the patient’s body shape, especially the thickness of the subcutaneous tissue. Second, we conducted this study at a single medical facility in Japan. Therefore, further studies with larger samples are warranted to generalize our results. Third, we used a Vscan dual probe as the PsUS. We did not evaluate other PsUS devices. Therefore, the diagnostic accuracy of other PsUS devices is unknown. Fourth, the US image with the linear probe may have the same accuracy as that of a convex or sector probe.