Pelvic tilt, impingement simulation, and the limits of static planning in robotic total hip arthroplasty

­

Despite the technical success of total hip arthroplasty (THA), instability remains one of the leading indications for revision.^1,2 Robotic assistance has refined implant positioning accuracy,^3-7 yet conventional CT-based planning carries an inherent blind spot in assuming a static pelvis. In doing so, it does not account for functional pelvic tilt variation or spinal stiffness, both of which have been reported to directly affect cup orientation, stability, and functional outcome in THA.^8-15 The study by Kouyoumdjian et al¹⁶ confronts this gap directly, asking whether integrating pelvic tilt variation, during seated and standing positions, into a 3D virtual collision detection model reduces early complications after robotic-assisted THA.

This single-centre retrospective cohort study included 656 consecutive patients undergoing robotic-assisted THA (MAKO; Stryker, USA) using either conventional CT-based static planning (V3; n = 406) or functional kinematic-based planning incorporating spinopelvic kinematics (V4; n = 250). The V4 group had sacral slope measured in seated and standing positions, which was fed into a virtual range of motion (vROM) model simulating impingement across the hip’s functional positions preoperatively. Subsequently, implant positioning was changed to minimize the risk of impingement and dislocation, which also guided implant selection, including the use of dual-mobility when deemed appropriate. Patients were assigned by date, into V3 before July 2022 and V4 after. To reduce baseline imbalances, propensity-score matching yielded 243 balanced pairs for multivariable analysis. Outcomes at one year included dislocations, peri-prosthetic fractures, reoperation, implant revision, and patient-reported outcome measures (PROMs).

Complications occurred in 2.8% of V4 patients versus 8.4% in V3 (p = 0.007). There were no dislocations in the V4 cohort compared with 1.7% in V3 (p = 0.048), and reoperations rates were significantly lower (2.0% vs 6.9%, p = 0.013). Reoperation-free survival at two years was 97.2% with V4 versus 94.1% with V3. In the multivariate analysis, V3 planning was the only independent predictor of reoperation (odds ratio 3.51, 95% CI 1.32 to 9.33), with age, sex, BMI, ASA grade, smoking, and surgical approach (anterior vs posterior) all failing to reach significance. PROMs were higher in V4 across all scores (Harris Hip Score,¹⁷ Oxford Hip Score,¹⁸ Forgotten Joint Score¹⁹), but differences did not reach the minimally important difference.

Although these findings are promising, they deserve cautious interpretation. This is a retrospective single-centre study with a time-based software change (all V3 cases preceding V4). As a result, any differences in outcomes may partly reflect surgical learning or evolving implant preferences rather than functional planning alone. Moreover, it is worth acknowledging that the V3 group had no spinopelvic imaging, whereas V4 introduced both spinopelvic assessment and vROM modelling. This is important as a V3 surgeon with access to spinopelvic images could still have adjusted cup targets using hip spine classifications.^20,21 Therefore, the difference in outcomes between the two cohorts cannot be attributed vROM alone, and may reflect a combined effect of vROM and spinopelvic assessment. Also, although not explicitly stated by Kouyoumdjian et al,¹⁶ a relaxed seated positioning was most likely used to measure the change in pelvic tilt. While there is no consensus on whether relaxed- or a flexed-seated measurement is preferable, there is some evidence suggesting that a flexed seated positioning could more accurately reflect dynamic spinopelvic kinematic during everyday activities, such as rising from a chair.²² Finally, dual-mobility implants were used more frequently in the V4 group (21.6% vs 12.3%). Since dual-mobility is known to reduce dislocation risk independently,^23-27 it is difficult to fully separate their effect from that of the planning software. Nevertheless, the decision to use dual-mobility was chosen based on the virtual ROM model, supporting the value of preoperative functional planning.

The study adds to the important conceptual evolution in THA planning towards a more patient-specific approach.^9,10,12 The classical Lewinnek safe zone (long the reference standard for acetabular orientation) has shown well-documented limitations and dislocations have been reported within the safe zone.^28,29 This led to the development of the functional safe zone, which includes postural pelvic tilt and combined anteversion to reduce instability.³⁰ Tezuka et al³¹ introduced the combined sagittal index (combination of sagittal acetabular and femoral positioning) as a tool to identify patients at impingement risk despite meeting conventional positioning criteria, while O'Connor et al¹¹ proposed functional combined anteversion, which accounts for both cup and stem position, as a method to determine the ideal implant orientation for impingement-free range of motion in patients with spinal stiffness. Kouyoumdjian et al¹⁶ build on these concepts by using spinopelvic assessment and vROM in a real clinical setting and reported how it may reduce early complications and improve outcomes. These results are in line with recent data presented at the American Academy of Orthopedic Surgeons 2026 annual meeting by Burgio et al,³² who reported a stepwise reduction in dislocation incidence with advancing technology across large series: 1.30% with conventional THA (n = 15.542), 1.00% with computer-navigated THA (n = 5.190), 0.61% with robotic THA without vROM (n = 1,793), and 0.24% with robotic THA with vROM (n = 6,647).

It has long been known that the pelvis changes position with movement and posture. This study adds to the evidence that static planning, which does not account for change in pelvic movement, is associated with higher rates of reoperation and dislocation in THA. Further advancements in technology, in the form of functional kinematic-based planning by vROM simulation, has given us the tools to begin addressing this issue. The fact that this was the only independent factor protecting against dislocation, stronger than any patient or surgical variable, is an important clinical finding. Our understanding of this complex relationship is still in its infancy. For example, what is less often appreciated is that the spinopelvic kinematics used for preoperative planning may not reflect the postoperative state.⁸ Anterior pelvic tilt caused by flexion contracture often resolves following surgery, whereas posterior pelvic tilt generally does not. This means that spinopelvic assessment used for preoperative planning may no longer reflect postoperative spinopelvic kinematics. Patient-specific algorithms that can predict standing and sitting pelvic tilt one year after THA have been published,³³ and the next step is to run functional planning against predicted postoperative kinematics rather than measured preoperative ones. AI and machine learning represent the most practical route to scaling this, and integrating them into functional planning is the next step towards enhancing THA outcomes.