Visual Inspection for Lower Limb Malalignment Diagnosis Is Unreliable

Introduction

Varus malalignment is present in 53% to 76% of individuals with unicompartmental knee osteoarthritis (OA). ¹ Malalignment increases the risk of developing and progressing OA due to loading imbalance of the knee joint.^2,3 Realignment surgery is considered a good treatment option for unicompartmental knee OA with varus malalignment.^4,5 Next to the management of unicompartmental OA, lower limb (mal)alignment is an important variable in multiple knee pathologies. Treatment outcome increases in osteochondritis dissecans, meniscal lesions, ligamentous injuries, and (osteo)chondral lesions management when combined with a favorable limb alignment. ⁶

During physical examination, several methods are available for the assessment of lower limb alignment.^1,7,8 These methods make use of goniometers, calipers for intercondylar and intermalleolar distances, inclinometers, and physical tests.^1,7 All these alternatives for whole leg radiographs (WLRs) were tested and compared with the measured leg axis on WLRs for correlations, ranging from low to good.^1,7,8 In case of a suspected malalignment, physicians could indicate obtaining additional WLRs. ⁹ These physical examination methods are advantageous in terms of lower radiation exposure, decrease in health care costs, and in cases of no available WLR system.^1,7

The reliability of only a visual inspection (visual observation of the patient in stance on both legs in the examination room) for the detection of lower limb malalignment is to the best of our knowledge unknown. This potentially affects decision making during clinical workup when physicians rely too much on only a visual assessment and omit further investigations in the examination room and/or additional radiographs such as a WLR.

Our goal was to investigate whether visual inspection of the leg alignment is a reliable method. The hypothesis of this study is that only a visual inspection of the lower limb is not sufficient for the detection of lower limb malalignment.

Materials and Methods

This study was approved by the local ethics board of the medical center (METC number 20-568). Patients were recruited at the outpatient clinic of the University Medical Center (UMC) Utrecht, a tertiary orthopedic referral center for knee pathology. Patients were eligible when a WLR was obtained on the same day as their appointment at the outpatient clinic. Exclusion criteria were age below 18 years; inability to read, communicate, and/or speak the Dutch language; and incapable of providing informed consent. When willing to participate, informed consent was signed. Body mass index (BMI), age, gender, and possible surgical indication were subtracted from the electronic health records.

Digital Photograph of Lower Limbs

Included patients were all placed in the examination room following the WLR protocol used at our radiology department. ¹⁰ The conditions for obtaining the digital photograph were kept similar to the WLR, making the images from the 2 modalities consistent for assessments. The patient’s feet were placed 10° externally rotated and 10 cm apart on a template. The knees were in full extension and both arms alongside the body. The researcher also placed the hips and upper body of the patient straightforward. In addition, the researcher instructed the patients to distribute their weight equally over both legs. A Sony Cybershot digital camera was used to obtain visual digital photographs of the lower limbs in a standing position. Digital photographs were taken from knee height, with 2 meters from the patient (Fig. 1).

OPEN IN VIEWER

Figure 1.

Included patient with on the left her WLR and on the right her digital photograph. The HKA measured on the WLR was 1.3° valgus for the left leg and 3.2° varus for the right leg. All observers overlooked the varus deformity present in the right leg and classified it as neutrally aligned. WLR = whole leg radiograph; HKA = hip knee angle.

Results

This study included 50 WLRs of patients, of which 31 (62%) were males and 19 (38%) females. The median age was 37 (18-58) years, the mean BMI was 25.4 (SD = 4.2), and the mean HKA was 178.9 (SD = 2.7). Reported surgical indications included an osteotomy, total/partial knee arthroplasty, and (osteo)chondral lesion treatment, for example, autologous chondrocyte implantation and microfracture. Characteristics of each alignment group are listed in Table 1.

OPEN IN VIEWER

Table 1.

Distribution of Patients’ Leg Alignments Categorized into Severe Valgus (HKA > 185°), Valgus (HKA 183°-185°), Neutral (HKA 182°-178°), Varus (HKA 177°-175°), and Severe Varus (HKA < 175°).

	n	BMI (M, SD)	Age (M, SD)	Gender (n)
Severe valgus	0	—	—	—
Valgus	6	24.7 ± 3.2	29.8 ± 13.9	3 males (50%)
Neutral	67	25.2 ± 4.5	36.8 ± 11.9	36 males (53.7%)
Varus	16	26.3 ± 4.4	40.4 ± 10.9	13 males (81.3%)
Severe varus	11	25.1 ± 1.8	40.9 ± 10.9	10 males (90.9%)

BMI = body mass index.

Visual Assessments

Spearman’s rho between the visual leg axis assessment of all observers and measured HKA on WLRs was moderate with 0.478 (P < 0.001). ICCs between the first and second assessments of the observers are moderate to good and listed in Table 2. Table 3 summarizes the inter-observer correlations between the 4 observers included in this study, which were all moderate. All visual assessments of patients’ leg axes are presented in Figure 2. The histogram plots the visual assessments per patient lower limb alignment, based on the measured HKA on the WLR.

OPEN IN VIEWER

Table 2.

Intra-Observer ICCs between the 2 Visual Assessments on Different Moments of the Observers, with 95% Confidence Intervals.

	ICC	95% Confidence Interval
Researcher	0.860	0.799-0.904
Orthopedic resident	0.718	0.607-0.801
Orthopedic surgeon—5	0.733	0.628-0.812
Orthopedic surgeon—10	0.786	0.679-0.857

ICC = intraclass correlation coefficient.

OPEN IN VIEWER

Table 3.

Inter-Observer Intraclass Correlation Coefficients between the Visual Assessments of the 4 Observers. The Researcher, Orthopedic Resident, Orthopedic Surgeon with 5 Years’ Experience (Orthopedic Surgeon—5), and Orthopedic Surgeon with 10 Years’ Experience (Orthopedic Surgeon—10) Are Included in This Table.

	Researcher	Orthopedic Resident	Orthopedic Surgeon—5	Orthopedic Surgeon—10
Researcher	—	0.598	0.716	0.700
Orthopedic resident	0.598	—	0.530	0.524
Orthopedic surgeon—5	0.716	0.530	—	0.653
Orthopedic surgeon—10	0.700	0.524	0.653	—

OPEN IN VIEWER

Figure 2.

Visual assessments in numbers (n) and percentages (%) of patients’ leg axes per category, based on measured hip knee angle (HKA) on whole leg radiographs (WLRs): severe valgus, valgus, neutral, varus, and severe varus patients. The y-axis illustrates the alignment categories based on the measured HKA on the WLR. The x-axis summarizes the 400 observations of 4 observers of 100 lower limbs, while the bars with different gray scales denote the visual assessment.

Incorrect Visual Assessments

The percentage of incorrect visual leg axis assessments ranged between 45.8% and 75.0% (Fig. 3). The errors were lowest in patients with a moderate valgus alignment and highest when the patient presented a severe varus deformity (Fig. 3). Figure 3 also visualizes the 4 observers with different experience levels separately. Chi-square tests for homogeneity between the 4 observers resulted in no significant differences (χ² = 1.651, P = 0.199). Figure 1 illustrates an example case, in which all the observers visually assessed the leg axis incorrectly.

OPEN IN VIEWER

Figure 3.

Percentage of erroneous visual assessments by all observers on the left. On the right is the percentage of erroneous visual assessments by the different observers. Researcher, orthopedic resident, orthopedic surgeon (5 years’ experience), and orthopedic surgeon (10 years’ experience).

Discussion

This study researched the reliability of a visual inspection in terms of detecting lower limb malalignments. The visual inspections of the observers were moderately correlated to the measured HKA on WLR. The percentage of incorrect assessments ranged between 45.8% and 75.0%, making a visual inspection as a standalone diagnostical method unreliable. Patients with a neutral leg alignment were assessed to be pathological in 50.7% of the times. Given this, patient indication for undergoing additional examinations such as WLRs based on only a visual inspection is not sufficient.

Physicians should perform physical tests and possibly subsequently request a WLR for reliable leg geometry measurements. ¹¹ Additional tests during physical examination include using goniometers, calipers for intercondylar and intermalleolar distances, and inclinometers.^1,7 In addition to physical tests and WLRs, plain knee radiographs are an excellent option for the measurement of femoro-tibial angles which correlates to the HKA. ¹²

A favorable leg axis, in which the mechanical axis passes through the healthy knee compartment, is a powerful tool in the treatment of knee pathologies.^4,5,13-15 Missed lower limb deformities could therefore have a serious clinical impact. The results of this study show substantial amounts of neutrally visually assessed lower limbs, while they were severe varus (>5°), moderate varus (2°-5°), and moderate valgus (2°-5°). This study reported a rate of 18.2% in cases of severe varus deformities, where the lower limb was visually assessed to be neutral (Fig. 2). For varus deformities, the rate of overlooked deformities was 23.4% (Fig. 2). In cases of valgus deformities, the rate of neutrally assessed lower limbs was 45.8% (Fig. 2).

The difference between the severe varus and neutral categories was at least 3° (18.2% of the cases) and between the valgus categories at least 5° (4.5% of the cases). Due to the used visual rating method, the mismatch between visual rating and measured HKA could not be provided in absolute degrees. But these results indicate that a mismatch in visual assessment and actual HKA of more than 3° is common with 18.2% in severe varus legs, which is, for instance, higher than the possible clinical accuracy of <1° in osteotomies.^16-18 To put this in another perspective, for the detection of varus/valgus malalignment, an accuracy of 3° from natural is needed. ¹⁹ This means that diagnoses based only on visual assessments are clinically not sufficient.

The ICC between the resident and surgeon with 10 years of experience was lowest at 0.524. This is a notable result as residents are working closely together with the surgeons as part of their training. Of note, the differences between all raters in this study were not statistically significant. The researcher and orthopedic surgeons are focused on malalignments and knee pathologies, in both their research and treatment of patients. This means that in real clinical settings, the difference in visual malalignment assessments between physicians could be higher.

The results of this study show that gender influences the visual assessment of patients’ leg axis. Women do present an increased risk for incorrect visual ratings. A possible explanation could be the difference in the anatomical axis of the femur between men and women.^20,21 The femoro-tibial angle was observed to be more valgus in women and of influence on the correlation with the HKA. ²⁰ Interestingly, BMI was also considered a possible inducer of errors, due to increased soft tissue enveloping the bones of these patients. But in this study, we found no significant effects of BMI, comparably with reports of Gibson et al.,^22,23 where they included participants with a mean BMI of 31.3. A possible explanation is that high BMI ratios do not necessarily mean high-fat concentrations in the leg and therefore around the landmarks for HKA assessments. ²³

Our study had some limitations. First, visual assessments were performed on two-dimensional (2D) photographs instead of physical inspection in the examination room. However, the author rated the patients in real life as on photographs with no significant differences between the 2 methods. Second, this study included no patients with severe valgus deformities. Bellemans et al. ¹⁹ reported the likelihood of severe valgus deformities to be very low (<1%). The third limitation of this study was that only one observer measured the HKAs on WLRs. Therefore, there was no control for the inter-observer reliability regarding the HKA measurements. Earlier reports, however, showed high reliabilities of HKA measurements on WLRs.^11,24-27 The HKA measurements performed in this study resulted in high intra-observer reliability, with a low mean absolute error of 0.15° between the 2 measurement moments. This is comparable with most literature.^11,24-27

Conclusion

Visual assessment of the lower limb alignment does not provide clinically relevant information. Lower limb malalignment diagnoses cannot be performed using only a visual inspection. Physical examination tests and radiographical assessments are advised.