In the vast majority of patients, the anatomical and mechanical axes of the tibia in the coronal plane are widely accepted to be equivalent. This philosophy guides the design and placement of orthopaedic implants within the tibia and in both the knee and ankle joints. However, the presence of coronal tibial bowing may result in a difference between these two axes and hence cause suboptimal placement of implanted prostheses. Although the prevalence of tibial bowing in adults has been reported in Asian populations, to date no exploration of this phenomenon in a Western population has been conducted. The aim of this study was to quantify the prevalence of coronal tibial bowing in a Western population.

This was an observational retrospective cohort study using anteroposterior long leg radiographs collected prior to total knee arthroplasty in our high volume arthroplasty unit. Radiographs were reviewed using a Picture Archiving and Communication System. Using a technique previously described in the literature for assessment of tibial bowing, two lines were drawn, each one third of the length of the tibia. The first line was drawn between the tibial spines and the centre of the proximal third of the tibial medullary canal. The second was drawn from the midpoint of the talar dome to the centre of the distal third of the tibial medullary canal. The angle subtended by these two lines was used to determine the presence of bowing. Bowing was deemed significant if more than two degrees. The position of the apex of the bow determined whether it was medial or lateral. Measurements were conducted by a single observer and 10% of measurements were repeated by the same observer and also by two separate observers to allow calculation of intraclass correlation coefficients (ICCs).

A total of 975 radiographs consecutively performed in the calendar years 2015–16 were reviewed, 485 of the left leg and 490 of the right. In total 399 (40.9%) tibiae were deemed to have bowing more than two degrees. 232 (23.8%) tibiae were bowed medially and 167 (17.1%) were bowed laterally. The mean bowing angle was 3.51° (s.d. 1.24°) medially and 3.52° (s.d. 1.33°) laterally. Twenty-three patients in each group (9.9% medial/13.7% lateral) were bowed more than five degrees. The distribution of bowing angles followed a normal distribution, with the maximal angle observed 10.45° medially and 9.74° laterally. An intraobserver ICC of 0.97 and a mean interobserver ICC of 0.77 were calculated, indicating excellent reliability.

This is the first study reporting the prevalence of tibial bowing in a Western population. In a significant proportion of our sample, there was divergence between the anatomical and mechanical axes of the tibia. This finding has implications for both the design and implantation of orthopaedic prostheses, particularly in total knee arthroplasty. Further research is necessary to investigate whether prosthetic implantation based on the mechanical axis in bowed tibias results in suboptimal implant placement and adverse clinical outcomes.