Abstract

1. Introduction

When evaluating periarticular wounds, assessing the integrity of the joint capsule is a critical step.¹ Violation of the joint capsule has been associated with a high incidence of septic arthritis and potentially significant sequelae to the joint.² The biggest challenge orthopaedic surgeons face when evaluating a periarticular wound is reliably diagnosing a traumatic arthrotomy without subjecting the patient to the morbidity of a surgical procedure.

Currently, the gold standard for diagnosis of a traumatic elbow arthrotomy is the saline load test (SLT). Recent studies, however, have shown that computed tomography (CT) can more accurately and efficiently diagnose knee arthrotomies with 100% sensitivity.³ Furthermore, we know that CT scans are able to detect as low as 0.1 cc of air within the knee joint.⁴ These results are promising, yet there are no studies in the literature using CT scans to evaluate joints other than the knee. Understanding the sensitivity of CT scans to diagnose arthrotomies in different joints is crucial for a more thorough and reliable use of this technology. The purpose of this study was to validate the use of CT imaging as a diagnostic tool for elbow arthrotomies by evaluating the absence or presence of air in the joint. We hypothesize that CT will be more sensitive in diagnosing elbow arthrotomies compared with SLT while decreasing the morbidity and technical challenges associated with making the diagnosis.

2. Materials and Methods

Before initiation of the study, approval by our institutional review board was obtained. Twenty thawed, fresh frozen cadaveric arm specimens with amputation through the shoulder joint were obtained. Each elbow was visualized for signs of trauma, prior surgical scars, and skin or capsule compromise. One elbow was found to have a violated elbow joint capsule on visual inspection. This was believed to have occurred during specimen preparation, and the specimen was excluded from the study.

2.1. Arthrotomies

Nineteen fresh frozen cadaver elbows were permitted to thaw. Before an arthrotomy, all 19 elbows underwent CT imaging described below and used as controls. Arthrotomies were subsequently performed in all 19 specimens, using the posterocentral arthroscopic portal site of the elbow, just proximal to the olecranon with the elbow flexed at 90 degrees as described by Feathers et al.⁵ A 1-cm skin incision was performed just proximal to the tip of the olecranon. A 4.5 millimeter trocar was used to perform the arthrotomy (Fig. ​(Fig.1A).1A). Arthrotomy was confirmed by inserting an arthroscopy camera through the trocar and visualizing the articular surface (Fig. ​(Fig.11B).

Open in a separate window

Figure 1.

A, Representation of the arthrotomy site with the arthroscopic trocar intra-articular. LE, lateral epicondyle; R, radial head; O, olecranon. B, Representative arthroscopic image confirming arthrotomy with visualization of radial head marked with the asterisk.

2.2. Computed Tomography imaging

As described by Konda et al,³ a CT scan using 2 mm cuts in the plane of the joint with sagittal and coronal reformats was performed for each specimen before any arthrotomy. These images were used as controls (Fig. ​(Fig.2A).2A). After performing the arthrotomy as described above, all elbows underwent a second CT scan with the same protocol (Fig. ​(Fig.2B).2B). These images were used as arthrotomy elbows. CT images were randomized and independently reviewed by 2 blinded, fellowship-trained orthopaedic trauma surgeons. Bimodal scoring was performed for each specimen with regard to the presence of an arthrotomy indicated by the presence of air in the joint. One reviewer again scored images 6 months after initial review.

Open in a separate window

Figure 2.

A, Representative coronal and sagittal images of a control elbow CT scan. B, Representative coronal and sagittal images of an elbow CT scan after an arthrotomy was performed showing air present within the joint.

2.3. Saline Load

After arthrotomy creation, and after the second CT acquisition, an SLT was performed on each specimen. An 18 gauge needle and 30 mL syringe were used to inject normal saline into the elbow joint. With the elbow at 90 degrees of flexion, the soft spot posterior lateral to the radial head demarcated by the radial head, lateral epicondyle, and tip of the olecranon was used for the injection (Fig. ​(Fig.3).3). Saline was injected at a rate of 2 mL per second. As saline was being injected, the arthrotomy site was being observed for evidence of saline leakage. Once saline was observed leaving the arthrotomy wound, the test was positive and the volume injected was recorded.

Open in a separate window

Figure 3.

Asterisk showing the soft spot where saline was injected. LE, lateral epicondyle; R, radial head; O, olecranon.

3. Results

Nineteen elbows were deemed intact thus used for evaluation. CT scans were found to have 100% sensitivity and 86% specificity for diagnosing elbow arthrotomies, that is, presence of air in the joint. Interobserver reliability calculated with Cohen kappa coefficient was near perfect at r = 0.89. Intraobserver reliability calculated with Cohen kappa coefficient was perfect at r = 1.

Regarding the SLT, injection of 20 mL confirmed the arthrotomy in 15 of 19 elbows (sensitivity 79%). A total of 25 mL of saline was required to be injected for a sensitivity of greater than 95% to be achieved with a range from 9 mL to 25 mL. Comparison of sensitivities between CT imaging and the SLT are represented in Table ​Table11.

4. Discussion

The importance of accurate diagnosis of arthrotomies in the setting of traumatic wounds has been previously described, with studies showing association of arthrotomies with a high incidence of septic arthritis. Hendry et al, studying soldiers on World War II, reported a 27% infection rate of patients with a penetrating knee joint injury who underwent nonoperative management of the arthrotomy.² The current challenge for orthopaedic surgeons is to accurately assess when the joint has been violated. The elbow joint is the second most common site of traumatic arthrotomies after the knee. Currently, the SLT is the standard method for diagnosing an arthrotomy. Our study was able to show the value of CT scans in diagnosing traumatic arthrotomies of the elbow, with a sensitivity of 100% and specificity of 86% in a cadaveric model.

Previous studies have estimated the volume of the elbow joint to be between 15 mL and 20 mL,^6,7 with studies specifically looking at the SLT using 20 mL volumes.^5,8 Some authors have suggested loading the elbow until the joint is too distended not allowing for further injection or until the patient can no longer tolerate.^9,10 We believe that this is unreliable due to multiple variables including patient pain and provider experience, altering the sensitivity of the test. We opted to use the previously described 20 mL volume in this study, and we continued to load the joint until all the elbows were positive by SLT. Previously described sensitivity of the SLT in the elbow joint when 20 mL was injected was 72%. For a 95% sensitivity of the test, 40 mL of saline had to be injected.⁵ In our study, we found similar results with a sensitivity of 79% when 20 mL was used. We did however achieve 95% sensitivity with only 25 mL injected, similar to a recent published cadaver study.¹¹ This difference may be in part due to cadaveric variations or due to inherent variability on how the test is performed. The SLT depends on accurate intra-articular injection, physician's experience, elbow position, and injection location in relation to the arthrotomy, all of which can influence volume needed for a positive test.¹² Moreover, in the clinical setting, the SLT is invasive and can cause patient discomfort and potential for introduction of infection. Overall, the results regarding the SLT seem comparable with previous studies.

Our study demonstrated that CT imaging is a reliable method of diagnosing elbow arthrotomies with 100% sensitivity. Previous studies have shown similar results on the knee.³ Based on this study and the study by Konda et al, the high sensitivity of CT scans is consistent when evaluating different joints (knee or elbow) and with different study methods; we used a cadaveric method while Konda et al performed a retrospective review. We also showed that CT imaging has a greater sensitivity compared with the gold standard SLT when 20 mL are injected, with sensitivity of 100% versus 79%, respectively. Specificity was 86% indicating that a number of CT scans were read as positive for an arthrotomy before arthrotomy creation. It is unclear whether this was a product of the CT scan itself or the use of cadavers, and it would warrant further studies in an observational clinical trial. In this situation, the 100% sensitivity may justify use despite concern for false-positive tests. False-negative tests, although not seen in this study, is a possibility. Bunyasaranand et al¹³ showed a case report of a positive arthrotomy not seen on CT. We likely have not had a sample size big enough to account to these rare cases of missed arthrotomy on CT.

There are various benefits in adopting CT scan to diagnose arthrotomies in comparison with the SLT. CT scans are quick to perform and are reliably interpreted by radiologists, orthopaedic surgeons, or other trained providers. In our study, the interobserver reliability score was r = 0.89 and intraobserver score was r = 1. CT scans are also noninvasive and less painful to the patient. Moreover, CT scans rely only on the interpretation of standardized imaging, while SLT assumes the injection is intra-articular, relying primarily on the experience of the provider to accurately perform the test. Jackson et al¹⁴ reported accurate intra-articular placement of a needle into the knee joint 71%–93% of the time. The elbow is a smaller and technically more challenging joint to perform intra-articular injections; thus, one may assume equal or lower rates of accurate intra-articular needle placement. Finally, the use of CT allows for a diagnosis to be made without the presence of a provider who is trained at performing an SLT. This is particularly relevant at institutions that do not have trained in-house providers at all times of the day. A diagnosis can be obtained rapidly without waiting for a consultant to arrive and perform an SLT.

There are downsides to this technology. An elbow CT scan costs more to the patient and to the overall health system than an SLT. This study did not include a cost analysis of CT imaging versus SLT, but it is important to acknowledge that a cost difference is likely to exist. That being said, if involvement of a consulting service can be avoided based on negative test results, overall cost may decrease. Radiation exposure should also be considered specially when dealing with children or pregnant patients. Iordache et al¹⁵ looking at radiation exposure to elbow CT scans found it to be 0.21 mSv when the arm was above the head and significantly higher 13.1 mSv when the arm was next to the torso. Imaging protocols and positioning should be patient-specific with the goal of limiting radiation exposure to the extent possible. Elbows were not all in the same degree of flexion, and we did not investigate whether elbow position or axis of reformatting of the CT can affect sensitivity of the test. Both factors could potentially influence detection.

This study is limited because it is a cadaveric study. Although we are using human specimens, there are soft tissue differences between a live specimen and the cadavers used here that could have affected how an arthrotomy is seen or detected on CT. In addition, we used one standard arthrotomy site with one standard size. We were unable to account for how arthrotomies at different location or sizes will affect the sensitivity of CT scans. Moreover, recent studies showed evidence that size of the arthrotomy may dictate operative versus nonoperative management.¹⁶ We believe that by doing a small arthrotomy (roughly 4 mm) would allow for the lowest threshold for detection of intra-articular air. Another limitation of the study is that the methods did not account for vacuum phenomenon. This phenomenon has been described where intra-articular air is seen in CT images in the absence of an arthrotomy, often in association with a high-energy injury. This study did not account possible false positives that can come from this phenomenon. Finally, a recent study showed very poor sensitivity of CT scans in the elbow, illustrating the need for further studies in the matter.¹⁷

Based on our findings, we conclude that CT scans are 100% sensitive and 86% specific in the detection of traumatic arthrotomies of the elbow in this cadaveric model. CT scans seem to have better sensitivity compared with SLT, but this is dependent on volume used. We also showed a high correlation between observers regarding the ability to identify a positive test. This study further illustrates the benefit and accuracy of CT scans to detect arthrotomies on joints other than the knee. Large clinical studies are needed to validate these findings, but we believe this technique will improve clinical practice of patients with periarticular elbow wounds by promoting the use of a faster, less invasive, and more sensitive test. We believe our results support the routine use of CT in the detection of arthrotomies for the elbow. If concern exists for a false positive, an SLT can also be performed in conjunction with physical examination and clinical judgment.

Footnotes

REFERENCES