Introduction: In total hip arthroplasty (THA) an optimal fit and fill of the stem is essential for stable fixation. Thus femur morphology must be studied during pre-op planning (implant choice, sizing, positioning) or when a new stem is to be designed. Using plain AP x-ray analysis and the definition of a simple two-level parameter (canal flare index, CFI), Noble et al. identified an age related transition of the endosteal canal in AP view from a ‘champagne flute’ to a ‘stove pipe’. This reference data is 2D only, limited to the endosteal geometry and the elderly age range was defined as 60–90yrs so that the number of octogenerians > 80yrs was too low to analyze morphological features of this rapidly growing and critical THA patient population.

In this study the endosteal and periosteal femur morphology of subjects > 80yrs was studied using 3D CT analysis. It was the goal to

describe age related changes of the femur morphology in 3D,

Methods: High-resolution CT-scans (slice thickness 1mm) were made of 170 consenting volunteers (m/ f=101/69). The old group consisted of 119 subjects ≥80yrs (m/f=65/54, mean age: 84.1yrs [80–105]) and the young group of 51 subjects < 80yrs (m/f=36/15, mean age 67.8yrs [39–79]). After thresholding the bone boundaries in Mimics V12 (Materialise, B), the endos-teal and periosteal coordinates were analyzed for width, wall thickness, surface areas and various CFI’s relating dimensions at 20mm above LT and at a distal level (e.g. 60mm below LT, isthmus): Surface CFI (3D-CFI), frontal and lateral CFI based on the AP and ML projections (2D-CFI) and flaring in each of the four directions (1DCFI).

Results: The surface CFI was sign. lower in subjects ≥80yrs (5.08 ±1.23) than in subjects < 80yrs (6.61 ±1.72, p< 0.0001). This difference was sign. larger in females than in males (−32% vs. −17%), an observation valid with reference to any distal level. Equivalent age differences were found in both the frontal and lateral 2D-CFI as well as the medial, lateral and anterior 1D-CFI with changes in the anterior direction (−26.3%) being most dominant. In addition wall thickness was sign. reduced in the very elderly. E.g. at 20mm above LT, the medial wall measured 10.40mm at < 80yrs and 7.61 at ≥80yrs, a reduction of −27% (p< 0.001). In females (−35%) this difference was sign. larger (males: −23%, p< 0.001) even when corrected for height.

Discussion: The age driven transition of proximal femur morphology continues in the octogenarian population. This transition is not limited to two discrete levels in the AP plane as previously reported but it is a continuous 3D phenomenon with high directional asymmetry. In addition, this transition also affects the wall thickness and the periosteal shape. Furthermore a strong gender effect was identified with aging females showing increasingly and asymmetrically less flaring and thinner walls. An age and gender specific THA stem design seems necessary to fit the morphed femur. The asymmetric transition prohibits the effective use of current implant systems with proportionally scaled dimensions but favors a matrix sizing scheme with frontal and lateral dimensions changing independently.

Due to demographic changes patients > 80yrs (octogenarians) are a rapidly growing group in total hip arthroplasty (THA). Stem design, choice, sizing and surgical insertion are more important in these patients as complications such as fractures are critical. Age and gender driven differences regarding canal shape (flare index, CFI), cortical wall thickness (WT) and bone mineral density (BMD) have been studied before only in isolation. Using CT, this study aims to investigate these parameters in combination and in 3D with a focus on the very elderly, identifying the regions critical for THA.

High-resolution CT-scans (1mm slices) of 168 femora (M/F=100/68) were analyzed in 3D (Mimics V12). Flaring indices were based on the dimensions measured 20mm proximal to the lesser trochanter (LT) and 60mm distal to LT: intramedullary surface area (3D-CFI), frontal/lateral planes (2D-CFI) and flaring of the 4 sides medial (med), lateral (lat), anterior (ant), posterior (post) (1D-CFI). WT was calculated subtracting periosteal and endosteal dimensions and BMD was measured in Hounsfield Units (HU). An octogenarian group (80+: n=117, mean age 84yrs [80–105]) was compared to a typical THA age group (80−: n=51, mean age 68yrs [39–79]).

Age and gender had significant effects on several parameters but at different levels, e.g. 2D frontal CFI was more influenced by the small age difference (80+ vs 80−=12%, p< 0.01) than gender (F vs M=2%). However, regarding lateral canal width, gender (F vs M=7%, p< 0.01), was more influential than age (80+ vs 80−=3%). The age-related changes on the shape occurred in 3D (3D-CFI 80+ vs 80−=23%, p< 0.01), but were asymmetrical between the 4 sides (e.g. 1D-CFI 80+ vs 80−: med=11%, p< 0.01) vs ant=27%, p< 0.01). Age and gender did not only effect shape, but also cortical WT, e.g. proximally octogenarian females had 35% less WT than the typical THA age group while males only had 14% lower WT (p< 0.01). Age, gender and shape asymmetry was also reflected in BMD distribution. on the medial side, the BMD gender difference in the octogenarians was small (=1%, p=0.61) but high on the anterior side (12%, p< 0.01). The most critical configurations for the octogenarians were found proximally on the posterior side with the lowest WT, lowest BMD and largest gender difference.

The complex transition of the proximal femur affects shape, WT and BMD, continues in the very elderly and differs between genders. It produces femoral canals and bone stock different from the typical THA patient group. Conventional stems may not fit properly. Surgical implant choice, sizing and templating should consider this asymmetric age plus gender effect on shape, WT and BMD to avoid complications such as periprosthetic fracture, excessive migration or luxation in this vulnerable age group. A major risk zone is the posterior wall where age transition and gender differences are high and WT and BMD low.