Patients undergoing unilateral hip arthroscopy for FAIS hip preservation at a single sports medicine center were invited to participate in a prospective institutional review board approved hip registry (Table 1). Inclusion criteria were participants with a clinical and radiographic diagnosis of FAIS with either cam or pincer morphology. Patients with prior surgical treatment on the hip of interest, unrecovered contralateral surgery, or a secondary syndrome/diagnosis such as cerebral palsy, Charcot-Marie-Tooth, or Down syndrome were excluded. All participants were analyzed prior to ipsilateral surgery. Participants with bilateral symptomatic FAIS and/or prior contralateral surgery were included, but only one limb was assessed per participant. Informed consent was obtained for each adult participant while children provided assent with parental consent. The rights of participants were protected.

Two squat types were employed in this study: a TSquat (10) and an HSquat. Each squat type was performed for three official repetitions after a practice period, with the start of each repetition beginning after the research staff's audible prompt. Participants were instructed to stand comfortably with their feet shoulder-width apart and their arms out in front to ensure motion marker visibility. However, participants were given no instruction with respect to trunk position, foot placement (whether flat or heel raised), squat depth, or the velocity of squat descent and/or ascent to assess natural mechanics. The HSquat was performed first, and participants were instructed to squat down to their lowest comfortable squat depth, hold their lowest position for a count of three, and then return to a standing position. For the TSquat, participants were instructed to squat down to their lowest comfortable point then immediately return to their original standing position (11). After each squat type, participants were asked to state their pain level between 0 and 10, where 0 was defined as “no pain” while 10 was defined as “the worst pain imaginable” (12). To ensure that no participant was pressured to continue with unintended discomfort, participants were continually monitored for signs of visible discomfort or reported pain. A pain score of ≥7 prompted a pause in testing, and the research team confirmed that they could continue the protocol.

During the squat task, participants underwent 3D motion analysis and were instrumented with the Plug-in-Gait (PIG) marker set (13). Additional tracking markers were added on the medial knees, medial ankles, and the most prominent superior region of the left and right iliac crests to account for potential visual loss during the squat. Maximum squat depth was calculated by dividing the vertical displacement of the hip joint center by the highest vertical position of the hip joint center (static standing position) and was utilized to determine a representative trial of each squat type (HSquat, TSquat). Kinematic variables included peak motion of the trunk, pelvis, hip, knee, and ankle. External joint moments (Newton-meter/kilogram) and powers (Watt/kilogram) at the hip, knee and ankle were also compared between the two squat types as a method to identify potential biomechanical compensation. Kinematic data were collected using a 12-camera VICON motion capture system (Vicon Motion Systems Ltd., Oxford, United Kingdom) sampled at 120 Hz. Ground reaction force data were collected using two AMTI force plates (Advanced Mechanical Technology Inc., Watertown, MA, USA) sampled at 3,000 Hz. A Woltring filtering routine with a predicted mean square error of 10 mm² was applied to the trajectory data (14). Squat events for each trial were automatically identified using a custom MATLAB script run within Vicon Nexus (11). The squat cycle included the following: start of descent, end of descent, start of ascent, and end of ascent. The hold period in the HSquat was not assessed for peak kinematic/kinetic values.

Participants also completed a series of questionnaires at the time of testing. The Harris Hip Score (HHS; 0–100) is commonly used to assess pain and overall function of individuals with hip disorders, with higher values indicating a better score (15). Activity levels were assessed via the UCLA activity score, which is a ten-point scale that describes activities ranging from wholly inactive (score of 1) to regularly participating in impact sports (score of 10; (16). These patient-reported outcomes were included given that pain may impact squat depth or the duration that a participant can hold a flexed position, and activity level may improve muscular control even with symptoms of FAIS.

Mean and standard deviations were computed for all biomechanical variables. Statistical analyses were performed using SPSS software (Statistical Package for Social Sciences Version 24; IBM, Armonk, NY, USA). Shapiro–Wilk tests for normality revealed deviations from normal distribution; therefore, non-parametric tests were used. Wilcoxon's signed rank tests were performed comparing the biomechanical variables of the two squat types for each participant. Spearman's rank correlations were also conducted to identify an association between the participant's self-reported HHS/UCLA activity scores and the participant's squat depth for each squat type. To consider time-normalized data over the entire squat, mean waveforms and statistical parametric mapping (17) were additionally employed via MATLAB for pelvic tilt, hip flexion, hip internal rotation, trunk lean, hip flexion moment, and knee flexion moment (MathWorks, Version 2024b, Natick, MA, USA). The alpha level was set to 0.05 for all statistical tests.

Peak kinematic differences between squat types can be found in Table 2. Maximum trunk flexion, trunk lean, and pelvic tilt were not significantly different between squat types. Maximum hip flexion was significantly greater during the TSquat (mean difference: 4.2° ± 0.7°; p = 0.002) and participants were less abducted at the hip during the HSquat (mean difference: 1.5° ± 1.7°; p = 0.020). There was no significant difference in maximum internal hip rotation between the two squat types. Maximum knee flexion was significantly greater during the TSquat (mean difference: 10.2° ± 1.2°; p < 0.001). In addition, there was significantly greater ankle dorsiflexion during the TSquat (mean difference: 1.4° ± 0.1°; p = 0.002). Time-normalized statistical parametric mapping between both squat types revealed differences in descent, deep squat, and ascent phases of pelvic tilt, hip flexion, and hip internal rotation. No differences in trunk lean were seen at any phase (Figure 2).

Peak flexor moments at the hip, knee, and ankle in the sagittal plane can be found in Table 3. Participants generated a greater peak sagittal moment during the TSquat both at the hip (mean difference: 0.10 ± 0.05 Nm/kg; p < 0.001) and at the knee (mean difference: 0.09 ± 0.08 Nm/kg; p < 0.001). Time-normalized statistical parametric mapping revealed differences in descent, deep squat, and ascent phases for both hip and knee moments, as well (Figure 2). There was no significant difference in peak sagittal ankle moment between the two squat types.

There was significantly greater peak power generation at the knee during the TSquat (mean difference: 0.40 ± 0.16 W/kg; p = 0.001; Figure 3). No differences in peak power absorption or hip/ankle peak power generation were observed.

Certain limitations must be noted in the current study. Squat order was not randomized, and data was taken from a single representative trial, though all participants were offered practice trials and experienced biomechanists ensured consistent mechanics. While attempting to reduce the impact of fatigue or adaptation, selection bias may be introduced. Bilaterally symptomatic participants were also included, and this may impact double-leg squat movement patterns more than single-leg squats, which are another frequently utilized functional movement test. Future studies should assess other potential variations of the squat such as stance width, foot position, etc. Studies may also investigate a cohort of healthy participants for increased generalizability or consider patient-reported outcomes of pain/kinesiophobia during different squat types. Although statistical parametric mapping offers temporal change insight, future work may include hip joint force using a musculoskeletal model estimate for more precise estimates of impingement beyond the surrogate of external hip joint moment (33–35).

Participants with FAIS tend to squat to a lower depth when performing a traditional squat compared to a hold squat. Potentially as a result, maximum flexion at the hip, knee, and ankle are greater in a traditional squat, as well as peak moments and power generation at certain joints. It is likely that anticipation of the hold alters movement patterns at maximum squat depth. In instances where the kinematic effects of deformity and/or pain are not the primary outcome, hold time should be standardized across a cohort. Regardless of the source of the kinematic and kinetic discrepancies identified in this study, standardization offers greater accuracy in identifying biomechanical abnormalities as well as longitudinal care during rehabilitation. Clinicians and researchers must consider the effect of potential compensatory adjustments in double-leg squat studies of FAIS patients, particularly in comparison to healthy controls better equipped for a deep squat.