Visual inspection for lower limb malalignment diagnosis is unreliable

Key points

1. 01Visual ratings matched the X-ray measured alignment only moderately (Spearman's rho 0.478).
2. 02Observers, including experienced surgeons, got the visual call wrong 45.8% to 75.0% of the time.
3. 03Nearly 1 in 5 severely bow-legged (severe varus) legs were judged as normal alignment.
4. 04Women were about 3.7 times more likely to be misjudged by eye than men.
5. 05Whole leg X-ray measurement was extremely consistent (intra-rater ICC 0.993).

How it was conducted

Design

Diagnostic reliability study (Level II); 2D photographs rated by eye versus whole leg radiograph

Participants

50 patients (100 legs) at a tertiary center, with same-day whole leg radiograph

Raters

4 observers (researcher, orthopedic resident, surgeon with 5 years, surgeon with 10 years), each rated photos twice 1 week apart, blinded to the X-ray

Reference standard

Hip knee angle (HKA) measured on whole leg radiograph, measured twice

Outcome

Agreement of visual category (severe valgus, moderate valgus, neutral, moderate varus, severe varus) with measured HKA

What they found

• Spearman's rho between visual assessment and measured HKA was moderate at 0.478 (P < 0.001).
• Incorrect visual assessments ranged from 45.8% to 75.0% per observer, with the most errors in severe varus and the fewest in moderate valgus.
• 18.2% of severe varus legs were visually assessed as neutral.
• Female gender carried an odds ratio of 3.7 for incorrect visual assessment (P = 0.001).
• Higher HKA increased the odds ratio for an erroneous assessment to 1.4 (P = 0.003); BMI and age were not significant.
• Intra-observer ICC for visual rating ranged from 0.718 to 0.860; inter-observer ICC was moderate (0.524 to 0.716).
• Whole leg radiograph HKA intra-rater ICC was 0.993 (95% CI 0.989 to 0.995) with a mean absolute error of 0.15 degrees (SD 0.29 degrees).

Limitations

• Single center with only 50 patients, and no severe valgus cases were present, so that category could not be tested.
• Visual ratings were made from 2D photographs rather than in person, though the authors report no significant difference.
• Only one rater measured the whole leg radiographs, relying on prior studies showing high radiographic reliability.
• Sample skewed toward neutral alignment (67 of 100 legs), limiting tested range.

Why it matters

For patients

If your leg alignment matters for treatment, expect an X-ray rather than relying on how your legs look.

For clinicians

Do not use eyeballing alone to grade lower limb alignment; confirm with whole leg radiograph and physical examination, especially before osteotomy planning.

For readers

Visual estimation of knee alignment is unreliable, and an objective measurement is required for decisions needing sub-degree accuracy.