Ulnar variance – Reliability of three different methods of measurement

INTRODUCTION

Distal radius fractures (DRFs) are a common type of fracture and make up approximately 17% of all fractures. The long-term consequences of a DRF can be chronic pain, osteoarthritis, joint instability, and diminished strength and range of motion.^[1] These complications can impact a patients’ overall quality of life and ability to maintain employment. As a result, it is important to diagnose DRFs efficiently and precisely.

Radiographic imaging and measurements are typically used when diagnosing DRFs, as well as in the process of determining whether to pursue surgical or conservative treatment.^[2,3] One such measurement is ulnar variance (UV), which describes the length of the ulna relative to the length of the radius distally.^[4,5] In relation to a compacted DRF, UV can be used as an indirect estimate for shortening of the radius.^[6] It is considered good practice to offer surgical treatment of the DRF when UV is more than 2–3 mm, i.e., the ulna is 2–3 mm longer than the radius distally, indicating fracture compression.^[2,7] Since UV is an important radiographic measurement, it is equally important to have a reliable method of measuring UV. Various methods have been proposed for measuring UV, three of which are described in Parker et al.,^[4] namely the method of perpendiculars (MoP), the method of central reference point (CRP), and the lateral method (LM). Three orthopedic surgeons measured UV on 100 radiographs using above-mentioned three methods and found that each method had a clinically acceptable intra- and interobserver agreement.^[4] However, in daily clinical practice, the responsibility of assessing radiographs and measuring UV falls is not only on experienced surgeons and radiologists but also on a diverse range of clinicians with varying levels of training, education, and experience. If the measured value of UV is influenced by the clinician’s experience or the chosen method of measurement, it could potentially result in either over-treatment or under-treatment of fractures. The overall aim of this study is to explore observer variance when measuring ulnar variance. The specific objectives are to estimate the inter- and intra-rater reliability of measuring UV using, respectively, (i) the method of CRP, (ii) the LM, and (iii) the MoP when measured by a group of raters with varying levels of experience and professional background.

MATERIAL AND METHODS

Measuring UV

The three methods used in this study to measure UV were the CRP, the LM, and the MoP.

UV measured using the CRP method was measured in a posterior-anterior radiograph. On the ulnar aspect of the radius, a point midway between the volar and the dorsal aspect of the articular surface of the radius was defined as the CRP. The distance from this point to the most distal part of the articulating surface of the ulna was measured and defined as UV according to CRP [Figure 1]. The length of the ulna relative to the radius was measured according to LM in the lateral radiograph as the vertical distance between the articular surfaces of the distal radius and the distal ulna [Figure 2]. Finally, UV defined according to the MoP was measured as distance between the articular surface of the distal ulna and the volar, ulnar aspect of the articular surface of the radius [Figure 3].

Close

Figure 1:

The method of central reference point. Line A is the longitudinal axis of the radius. Line B is perpendicular to Line A and is drawn through the ulnar volar part of the distal radius that is often depicted as a sclerotic border. Line C is parallel with line B and is drawn through the ulnar-dorsal part of the distal radius. Line D is the central reference point. It is placed parallel with Line B and C at the halfway point between these two lines. Finally, Line E is drawn through the most distal part of the articulating surface of the ulna, perpendicular with line A. The distance between line D and E is measured and is defined as ulnar variance.

Export to PPT

Close

Figure 2:

Line A is the longitudinal axis of the radius. Line B is drawn through the most distal part of the articulating surface of the ulna. Line C is drawn through the most proximal, sclerotic part of the articulating surface of the distal radius. Lines B and C are perpendicular to the longitudinal axis of the radius. The distance between the two lines B and C is measured and defined as ulnar variance.

Export to PPT

Close

Figure 3:

The method of perpendiculars. Line A is the longitudinal axis of the radius. Line B is perpendicular to line A and is drawn through the ulnar volar part of the distal radius that is often depicted as a sclerotic border. Line C is parallel with line B and is drawn through the most distal part of the articulating surface of the ulna. The distance between line B and C is measured and defined as ulnar variance.

Export to PPT

RESULTS

The mean UV (±SD) for the CRP, the LM, and the MoP was −2.76 (±2.2) mm, −0.93 (±2.11) mm, and −0.33 (±2.02) mm, respectively. Negative values indicate ulna minus, i.e., an ulna that is shorter than the radius at the joint level.

All three methods showed excellent interrater reliability according to the calculated ICC. For the CRP, the ICC was 0.91 (95% CI, 0.83, 0.96), and for the LM, the ICC was 0.96 (95% CI, 0.92, 0.98). The MoP had the highest ICC value and the narrowest 95% CI at 0.97 (95% CI, 0.95, 0.99).

Using the BA LoA analyses to assess intrarater agreement, the broadest LoA was found for rater 1 using the CRP, with a mean measured difference, bias SD (95% CI), of 0.38 (1.50) (−0.30, 1.06) and LoA of ±2.94. Conversely, the narrowest BA LoA was found by rater 2 using the MoP, with a bias of 0.43 (0.68) (−0.90, 1.75) and LoA of ± 1.33. BA intrarater agreement with LoA and 95% CI for both raters and methods of measurement is summarized in Table 1. The intrarater agreement is presented visually in BA plots with LoA and 95% CI [Figure 4].

Close

Figure 4:

Bland–Altman plot illustrating the intraobserver agreement for rater 1 and rater 2 using the method of central reference point, the lateral method and the method of perpendiculars. The dashed lines depict the mean differences, and the surrounding green area depicts the 95% confidence intervals (CIs). The solid lines represent the upper and lower limits of agreement and the blue areas illustrate corresponding 95% CI.

Export to PPT

DISCUSSION

UV is an important measurement for several pathologies of the wrist, especially when assessing fractures of the distal radius. UV is included in the decision-making process when determining whether treatment should be surgical or conservative.^[2,6,13] Different methods have been defined for measuring UV.^[4,14,15] This study estimated and compared inter- and intrarater agreement of the method of CRP, the LM, and the MoP. The calculated ICCs for all three methods were >0.9, which indicate excellent interrater reliability with minimal differences in ICC between methods.^[11] Conversely, when looking at intra-rater agreement, the LoA was broad, indicating that the mean measured difference between two ratings made by the same rater may vary from 2.65 mm (rater 2 MoP) to 5.88 mm (rater 1 CRP) [Table 1]. Intra-rater agreement using the MoP on radiographs of 63 cadaver arms obtained in various degrees of forearm rotation has previously been estimated by BA LoA and reported to be, respectively, 3.12 and 6.0 mm for two observers, which is slightly broader than in the current study. A possible explanation for this difference in LoA could be that different degrees of forearm rotation were applied in the aforementioned study, where radiographs were obtained with a degree of forearm rotation ranging from −11.8° of supination to 11.2° of pronation.^[9] Whereas, in the current study, all radiographs were obtained in a strictly neutral rotation using the scaphopisocapitate relationship to define the true lateral position of the forearm. It has previously been reported that an increase in forearm pronation can cause the wrist to appear more ulnar positive.^[4,16] Since an UV of 2–3 mm is commonly used as a cut-off for when surgical treatment is recommended, it is important to define the most reliable method of measurement.^[2,17]

In the current study, the highest ICC reached 0.97, while the narrowest BA LoA was 2.65 mm. Both results were achieved using the MoP method, which may be attributed to its simplicity, as it involves only three lines. In addition, the MoP is the preferred method routinely used by most of the raters in their daily practice at this institution, making it the method they are most experienced with and most accustomed to using. The finding that the MoP is the method of measurement with the least variance is supported by Parker et al.,^[4] who reported that the MoP had the best inter- and intrarater agreement when compared to the CRP and the LM. The CRP method is the most intricate of the three approaches, as it requires an additional step to estimate the central point between the dorsal and volar aspects of the radial articular surface. This extra step entails drawing an additional line, which may increase observer variability in measurements [Figure 1].

It is worth mentioning that the mean UV found using the CRP is noticeably more negative than the UV found using the LM and the MoP. This is because the CRP measures UV from the CRP, which is a point midway between the volar and dorsal rim of the distal radius, as opposed to the LM and the MoP, where UV is measured from the volar rim of the radius. As a result, the CRP measures from a more distal point on the radius, while all three methods define the most distal articular surface of the ulna similarly. Consequently, when employing the CRP method, the UV is measured more negatively. This could potentially imply conservative treatment for patients who might have otherwise been candidates for fracture reduction or surgical intervention using the LM or the MoP methods.

If clinical practice guidelines do not explicitly specify the preferred method of measurement, it is important to be aware of potential discrepancies in the UV values derived from different measurement techniques. These variations may potentially impact treatment decisions, emphasizing the need for careful interpretation of UV results across different measurement methodologies.

The LM is less complex, but the landmarks can sometimes be harder to locate due to the overlap of the radius and ulna in the lateral projection.^[4,18] However, previous studies have suggested that the LM is the most clinically relevant of the three techniques. According to Bernstein et al., the UV measured using the LM is the most accurate and correlates highly with measurements made directly on the bones.^[4,18] It has also been reported that when using the MoP, the UV is underestimated by up to 1.5 mm in comparison to the actual variance.^[19] Nonetheless, a comprehensive systematic review evaluating the accuracy of radiographic measurement of DRFs, including UV, has indicated a lack of sufficient evidence to establish the accuracy of UV measurements obtained radiographically.^[20] This underscores the need for further research and standardized protocols to define the method of measurement and radiographic positioning.

The current study possesses inherent limitations that should be acknowledged. While the raters in this study represented a diverse range of experience and education, certain populations that frequently use and measure UV in their daily clinical practice were not included, such as hand surgeons or junior doctors in the emergency room. The latter often plays a key role in making initial diagnoses based on acute radiographs. In addition, only two out of the nine raters repeated the measurements, which may lead to an insufficient sample size for calculating intra-rater agreement. While the cadaver arms were useful for obtaining radiographs of the wrist in true neutral pronosupination, it is worth noting that the radiographs may not be fully representative of the wrist obtained from living patients. In the present study, our focus was primarily on estimating measures of reliability, recognizing that while the issue of measurement accuracy is equally significant, it falls beyond the scope of our investigation.

CONCLUSION

The MoPs displayed the highest interrater reliability (ICC 0.97). However, the method of CRP and the LM also had excellent interrater reliability with ICC’s of, respectively, 0.91 and 0.96. Intra-rater agreement based on BA LoA ranged from 2.65 mm for the MoP to 5.88 mm for the method of CRP. The results suggest that UV measurements can be obtained consistently using all three methods of measuring across raters with different professional backgrounds and levels of experience.