Efficacy and Safety of Liposomal Bupivacaine in Direct Anterior Approach Total Hip Arthroplasty: A Retrospective Cohort Study

Introduction

Total hip arthroplasty (THA) is an important surgical method for the treatment of hip joint diseases and has been widely recognized for improving hip joint mobility and enhancing patients’ quality of life [1,2]. The direct anterior approach (DAA), as a commonly used surgical approach for THA, is highly favored by clinicians due to its clear surgical field, minimal physical trauma, and faster recovery speed [3]. However, postoperative pain remains a significant issue, particularly in the early period after surgery [4]. Acute postoperative pain not only hinders early rehabilitation and increases the risk of chronic postsurgical pain but also increases the incidence of perioperative complications, such as delirium, deep vein thrombosis, and prolonged hospital stays, especially in older adults [5,6]. Moreover, postoperative pain often disrupts sleep, with poor sleep further amplifying pain perception, weakening immunity, and delaying tissue repair. This bidirectional relationship increases healthcare costs and reduces patient satisfaction, thereby emphasizing the need for optimized analgesic strategies.

Local infiltration analgesia (LIA) has become an important component of multimodal perioperative analgesia [7]. By directly injecting local anesthetics into the tissue of the surgical area, LIA can effectively inhibit the transmission of pain signals, thereby reducing postoperative pain [8]. Traditional agents such as ropivacaine have a short analgesic duration (6–8 hours), failing to cover the 24- to 72-hour postoperative pain peak [9]. This often requires supplementary opioids, which carry risks of nausea, vomiting, constipation, and respiratory depression, which is particularly dangerous for older adults with reduced physiological reserve [5,10].

Liposomal bupivacaine (LB) is a novel long-acting local anesthetic. Using liposomal encapsulation, it sustains bupivacaine release for over 72 hours while maintaining stable plasma concentrations [11]. Previous studies have reported inconsistent findings: some show LB improves pain control and reduces opioid use in posterolateral approach THA [12], while others fail to confirm such advantages [13,14]. Moreover, DAA differs from the posterolateral approach in terms of anatomy and soft tissue handling, and specific data regarding its use remain limited [15,16]. Therefore, this retrospective cohort study aims to evaluate the early efficacy of LB vs ropivacaine for LIA using a multi-point injection technique in DAA-THA, with a focus on pain control, sleep quality, and complication rates, in order to provide a clinical data reference for the current inconsistencies and unresolved questions in the existing literature.

Material and Methods

STUDY DESIGN:

This was a retrospective cohort study approved by the Medical Ethics Committee of Shaoxing Municipal Hospital of Traditional Chinese Medicine (approval No. 2025/024). We retrieved the records of 141 patients who underwent DAA-THA between January 1, 2025, and September 30, 2025, from the hospital’s electronic medical record (EMR) system (Lianzhong Inpatient Physician Workstation V2.57.0603). This integrated EMR platform includes routinely recorded structured data fields, including demographic characteristics, surgical parameters, visual analog scale (VAS) scores, laboratory test results, medication records, and adverse event documentation; and unstructured clinical notes, including operative reports, progress notes, and analgesia management records. These patients were divided into 2 groups based on whether LIA with LB was administered during the operation. Sleep quality scores were post hoc constructed composite outcomes and derived from modified Pittsburgh Sleep Quality Index (PSQI) items, which were retrospectively abstracted from clinical notes and nursing assessment records.

ETHICS STATEMENT:

This study was approved by the Medical Ethics Committee of Shaoxing Municipal Hospital of Traditional Chinese Medicine (approval No. 2025/024; approved on January 15, 2025). Written informed consent was obtained from all patients prior to surgery, in accordance with the requirements of the ethics committee. For patients with limited decision-making capacity, consent was obtained from their legal guardians. To protect patient confidentiality, all records were de-identified during data extraction (replacing personal identifiers with unique study IDs), and access to the de-identified dataset was restricted to only the members of the research team.

EXPOSURE ASSIGNMENT:

The choice of local infiltration agent (LB or ropivacaine) was determined by routine clinical practice during the study period and was not randomized. Selection was primarily driven by institutional protocol and medication availability, rather than individualized prognostic assessment. Residual confounding related to non-random treatment selection cannot be fully excluded.

DATA EXTRACTION AND VERIFICATION PROCEDURES:

Two independent researchers (Ye Chen and Lanyong Chen) extracted data from the EMR system using a standardized data collection form. The extracted variables included baseline characteristics, surgical details, perioperative management, and outcome measures. For the specification of sleep quality data, sleep-related information was prospectively recorded by bedside nursing staff during clinical care following a standardized nursing assessment protocol. Assessments were conducted once daily during the morning nursing round (approximately 7: 00–9: 00 AM) on the day of surgery (postoperative day 0 [POD 0]), and on postoperative days 1 and 2 (POD 1, POD 2). Nurses documented patient-reported sleep parameters, such as difficulty falling asleep and nighttime awakenings, along with their observations in the nursing records, as part of the standard inpatient symptom monitoring protocol. All nursing staff and research personnel responsible for sleep documentation and outcome assessment adhered strictly to the institutional pain and symptom management protocol and had completed the corresponding standardized training and certification. This training program explicitly defined the assessment method and documentation standards for the VAS and standardized the procedures for collecting sleep-related observation items, thereby ensuring consistency, comparability, and reliability of data collection across the hospital. Discrepancies between the 2 researchers were resolved through discussion with a third senior researcher (Chunxiao Zhang) who was not involved in the initial extraction. A 10% random sample of records was audited by the study supervisor (Songfeng Hu) to verify data accuracy, with a consistency rate of over 95%.

GENERAL INFORMATION:

The inclusion criteria were as follows: (1) diagnosis of femoral neck fracture or femoral head necrosis; (2) no preoperative infection, with normal inflammatory indicators; (3) full knowledge of the treatment and rehabilitation plan and signing of the informed consent form; and (4) complete preoperative and postoperative data.

The exclusion criteria were as follows: (1) allergy to the anesthetic drugs used in this study; (2) severe liver and kidney dysfunction, with severe liver dysfunction defined as alanine aminotransferase (ALT) or aspartate aminotransferase (AST) ≥3 times the upper limit of normal or total bilirubin ≥2 times the upper limit of normal, and severe renal dysfunction defined as serum creatinine (Scr) ≥133 μmol/L (male) or ≥124 μmol/L (female), or estimated glomerular filtration rate <30 mL/min/1.73 m²; (3) mental and psychological diseases or cognitive impairment (including cognitive disorders such as Alzheimer disease or vascular dementia confirmed by clinical evaluation or neuropsychological testing), inability to cooperate with functional exercises and follow-up assessments; (4) history of hemorrhagic diseases; and (5) need to undergo surgery at other sites during the same period.

Alzheimer disease and other cognitive disorders were excluded primarily to ensure the reliability of the core outcome measures of VAS and sleep quality scores, as cognitive impairment would lead to inconsistent or unreliable assessments that could compromise the validity of between-group comparisons.

Surgical Methods

PERIOPERATIVE MANAGEMENT:

Perioperative management consisted of 5 aspects, as follows. (1) Health education: psychological intervention and guidance on cardiopulmonary function exercises were provided to patients, and the precautions during the perioperative period were explained in detail. (2) Nutritional support: during the perioperative period, the patient’s albumin level was improved through a high-protein diet, and albumin was infused if necessary to correct hypoalbuminemia. (3) Blood management: 1 g of tranexamic acid was intravenously infused 10 minutes before the start of the operation. After the operation, ice compress and pressure bandaging were used to reduce bleeding. For patients with a hemoglobin level ≤70 g/L, or 70–90 g/L accompanied by symptoms such as dizziness and palpitations, blood transfusion therapy was required. (4) Analgesic and anesthetic management: preoperative pain education was provided to patients to improve compliance, along with intravenous lornoxicam (8 mg) every 12 hours for twice-daily preemptive analgesia. Intraoperatively, general anesthesia combined with fascia iliaca block was used for regional analgesia, and controlled hypotension was implemented to alleviate surgical trauma-related pain. After the operation, ice compress and elevation of the affected limb were used to reduce joint swelling and inflammatory response. Patients were encouraged to get out of bed and ambulate early to relieve their psychological pressure, and non-steroidal anti-inflammatory drugs (NSAIDs; intravenous lornoxicam 8 mg, twice daily) were continued to be administered for pain relief in accordance with the preoperative plan. Rescue analgesia with oral tramadol (50–100 mg) was made available upon patient request for a pain intensity of greater than 4 on the VAS. (5) Anticoagulant therapy: preoperative routine color Doppler ultrasound of the lower extremity veins was performed to rule out lower extremity thrombosis. Low-molecular-weight heparin was administered 8 hours after the operation for anticoagulation, and it was changed to oral rivaroxaban after discharge. (6) Functional exercise: on the first day after the operation, under the guidance of a rehabilitation therapist, the patient attempted to sit by the bed and then transitioned to standing. If there was no discomfort, the patient got out of bed and moved with the support of a walker [17–21].

EFFICACY EVALUATION INDICATORS:

Primary outcomes were resting and activity-related pain intensity measured using a 0–10 VAS score (0=no pain, 10=worst imaginable pain) [22] recorded on POD 1, POD 2, POD 3, and POD 7. The resting VAS score was assessed at rest, and the activity-related VAS score was assessed during mobilization and physiotherapy, as documented in routine nursing records. To contextualize the clinical relevance of between-group differences in pain scores, we referenced a threshold from prior THA research, in which a 2-point difference on the VAS was reported as the minimal clinically important difference [23,24].

Secondary outcomes were as follows. (1) An exploratory composite sleep quality score was derived retrospectively from nursing documentation, inspired by selected components of the PSQI and adapted for acute inpatient postoperative documentation [25]. This modified score has not been formally validated and should be interpreted as an exploratory secondary outcome. Two components of the original PSQI were excluded from this modified score, namely “daytime dysfunction” and “use of hypnotic drugs”. Justification for these exclusions is as follows. The daytime activities of hospitalized patients are significantly affected by medical interventions, such as rehabilitation exercises and routine examinations, which cannot objectively reflect daytime dysfunction associated with sleep quality. The postoperative use of hypnotic drugs is a physician-ordered intervention rather than a spontaneous sleep state, potentially confounding the true assessment of sleep quality. The 5 retained components included subjective sleep quality, sleep latency, sleep duration, sleep efficiency, and sleep disturbances, all of which are directly relevant to the acute postoperative sleep state of hospitalized patients. Each component was scored on a 0–3 scale according to standard PSQI criteria, resulting in a total score ranging from 0 (no disturbance) to 15 (extreme disturbance). (2) The outcome of rescue opioid use within 72 hours postoperatively was defined as administration of tramadol or other opioids for pain control. (3) White blood cell (WBC) count, C-reactive protein (CRP) level, liver and kidney function measurements were further secondary outcomes. (4) Adverse events were systematically monitored throughout the perioperative period (from the initiation of surgery until hospital discharge). All adverse events were documented in detail, specifying the type, time of onset, severity (graded as mild, moderate, or severe), and the clinical management undertaken. Any severe adverse events, such as anaphylaxis or major cardiovascular events, were reported to the institutional ethics committee immediately upon occurrence.

BIAS CONTROL AND OUTCOME ASCERTAINMENT:

Given the retrospective observational design, blinding of patients or care providers could not be implemented or verified. To reduce information bias, postoperative VAS scores and sleep-related items were extracted from routine nursing assessment records using a prespecified data abstraction form. Two independent reviewers abstracted the data and cross-checked key outcomes and exposure variables; discrepancies were resolved by consensus with a senior adjudicator. Data were de-identified prior to statistical analysis.

STATISTICAL ANALYSIS:

All analyses were performed using SPSS 26.0 and R (version 4.3.1). Continuous variables were inspected for distributional characteristics using histograms and Q-Q plots. VAS and PSQI scores were approximately normally distributed and therefore treated as continuous variables. Categorical variables are presented as frequencies and percentages.

BASELINE COMPARISONS:

Baseline characteristics were compared using independent-samples t tests for continuous variables, or chi-square or Fisher exact tests for categorical variables. Given the study’s retrospective cohort design, these baseline comparisons are purely descriptive and not intended for inferential purposes. The purpose is to characterize the study population rather than test for group equivalence.

LONGITUDINAL OUTCOMES (VAS AND SLEEP QUALITY SCORES):

Postoperative VAS (POD 1–3, POD 7) and sleep quality scores (POD 0–2) were analyzed using a linear mixed-effects model to account for repeated measurements within individuals and potential confounders. Model selection was justified based on the hierarchical structure of the data (repeated measurements within patients), which required accounting for intra-individual correlation. All models included the following covariates: age, sex, body mass index (BMI), American Society of Anesthesiologists (ASA) classification, disease type, operation time, incision length, and surgical side. These covariates were selected based on their potential to confound the relationship between the intervention and the outcomes. The model included a random effect: patient ID (random intercept); and fixed effects, including treatment group (LB vs ropivacaine), time (categorical variable), group×time interaction, and adjustment covariates, including age, sex, BMI, ASA classification, disease type, operation time, incision length, and surgical side.

The unstructured covariance matrix was used when the model converged; otherwise, an AR(1) structure was applied. Estimated marginal means were used to generate the figures. When a significant group × time interaction was detected, post hoc comparisons at each time point were performed using Bonferroni correction.

RESCUE OPIOID USE (SENSITIVITY ANALYSIS):

All patients were analyzed according to their original group assignment (analogous to intention-to-treat principles). Rescue opioid use (yes vs no) within 72 hours was analyzed as a secondary binary outcome using logistic regression, adjusting for the same covariates, and as an exploratory subgroup analysis among patients who did not require opioids.

The justification for this model was that it was scientifically appropriate given the binary nature of the outcome (rescue opioid use: yes vs no). Adopting the identical covariate set as the linear mixed-effects model ensured consistency in confounding adjustment across primary and secondary outcomes, enabling isolation of the independent effect of LB vs ropivacaine on postoperative opioid requirements, independent of baseline patient and surgical characteristics. No patients were excluded on the basis of receiving postoperative opioids.

OTHER LABORATORY AND SAFETY INDICATORS:

Between-group comparisons of WBC counts, CRP levels, liver enzymes (ALT/AST) levels, and Scr levels were analyzed using a linear mixed-effects model with random intercepts. Categorical adverse events were compared using chi-square and Fisher exact tests.

STATISTICAL THRESHOLDS:

Two-sided P<0.05 was considered statistically significant. For multiple post hoc comparisons of VAS and PSQI scores across time points, Bonferroni-adjusted significance thresholds were applied.

HANDLING OF MISSING DATA:

For longitudinal outcomes analyzed using linear mixed-effects models, all available repeated measurements were included under maximum likelihood estimation, without requiring complete observations at every time point. Patients with missing key exposure information or without any postoperative outcome record were excluded. The overall completeness of the extracted dataset was high.

Results

BASELINE COMPARISONS:

The 2 groups were comparable with respect to sex distribution, age, preoperative ASA classification, disease type, operation time, and incision length (all |standardized mean difference [SMD]| <0.10; all P>0.05). However, meaningful between-group imbalances were observed in BMI and surgical side. To control for these potential confounders, both variables were adjusted for as covariates in the subsequent primary outcome analyses (Table 1).

PRIMARY OUTCOME: VAS SCORES:

A significant group×time interaction was observed for resting VAS (F=3.142, P=0.025) and active VAS (F=5.725, P<0.001) scores. After adjusting for covariates (age, sex, BMI, ASA class, disease type, operation time, incision length, surgical side), post hoc comparisons showed that the LB group had statistically significant lower pain scores than the ropivacaine group on POD 1–3 for both resting and activity-related pain (all adjusted P<0.05 scores). The adjusted mean differences ranged from −0.68 to −0.01 points (Table 2), all below the 2-point minimal clinically important difference for VAS scores in THA, suggesting limited clinical relevance despite statistical significance. No between-group difference was observed at POD 7 (adjusted P>0.05). Trends are displayed in Figure 2. Unadjusted observed scores are provided in Table 3.

SECONDARY OUTCOME: SLEEP QUALITY SCORES:

A significant group×time interaction was also found for sleep quality scores (F=14.17, P<0.001). After the same covariate adjustment (including age, sex, BMI, ASA class, and surgical factors), the LB group reported significantly better scores than the ropivacaine group on POD 1 (MD=−1.55, adjusted P<0.001) and POD 2 (MD=−1.82, adjusted P<0.001), but not on POD 0. Detailed results are provided in Table 2 and illustrated in Figure 3. Observed, unadjusted sleep quality scores are available in Table 3.

RESCUE OPIOID USE:

The LB group showed a numerically lower rate of rescue opioid use within 72 hours (4.92% vs 7.35%), although this difference was not statistically significant (adjusted OR = 0.629, 95% CI: 0.093–4.235, P =0.633).

WBC COUNT AND CRP LEVEL:

There were no statistically significant differences in WBC count and CRP levels between the 2 groups before and after surgery (P>0.05; Table 4).

LIVER AND KIDNEY FUNCTION INDICATORS AND ADVERSE EVENTS:

Before and after surgery, there were no statistically significant differences in ALT, AST, and Scr levels between the 2 groups (P>0.05). The incidence of nausea and vomiting was significantly lower in the LB group (4.92%) than in the ropivacaine group (17.65%, P=0.024). There were no significant differences in the incidence of blood glucose fluctuation, lower extremity deep vein thrombosis, or constipation between the 2 groups (P>0.05; Table 5).

Discussion

KEY FINDINGS:

This retrospective analysis compared outcomes associated with LB vs ropivacaine for LIA in patients undergoing DAA-THA. The key observations were that LB-based LIA was associated with modestly lower pain scores in the early postoperative period (POD 1–3), improved short-term sleep quality scores (POD 1–2), and a lower incidence of nausea and vomiting. No significant between-group differences were found in rescue opioid use within 72 hours, other complications, or inflammatory and liver and kidney function indicators.

As a novel long-acting local anesthetic, LB has been extensively studied in joint arthroplasty in recent years [26]. Vij et al [27] found that LB can prolong the duration of postoperative analgesia following THA, and the findings of the present study are consistent with their results. The LB group’s significantly lower VAS scores on POD 1–3 are consistent with its known characteristic of sustained drug release over 72 hours, although a direct causal link cannot be confirmed [28,29]. The relatively shorter duration of action of ropivacaine may limit its ability to address peak postoperative pain, which provides a plausible context for the observed between-group differences in early pain outcomes [9,30]. Postoperative pain is driven by surgical trauma-induced inflammatory responses and direct nerve injury [31,32]. The absence of inter-group differences in systemic WBC counts and CRP levels suggests that the analgesic effect of LB may be more likely mediated by its local anesthetic action on nerve conduction than by a modulating effect on these systemic inflammatory markers.

We observed no statistically significant reduction in rescue opioid use with LB. Given the retrospective design of this study, this negative finding does not itself prove a lack of opioid-sparing potential for LB. The low overall rate of rescue opioid use suggests that the comprehensive multimodal analgesic regimen used in our protocol may have created a “ceiling effect”, which could have rendered between-group differences in opioid requirements difficult to detect, thereby masking any opioid-sparing benefit that LB might have demonstrated in a setting with less intensive background analgesia. This interpretation constitutes a post hoc analysis rather than a pre-specified hypothesis. Consequently, the opioid-sparing effect of LB might become more apparent in clinical environments with a less aggressive background analgesic regimen.

The multimodal analgesic regimen used in this study represents the standard of care for older adult patients undergoing DAA-THA at our institution. Consisting of preoperative administration of NSAIDs, intraoperative fascia iliaca block, and postoperative NSAID therapy with optional opioid rescue analgesia, this protocol aims to optimize pain control and early mobility while minimizing associated risks. We conducted close monitoring of adverse events associated with NSAIDs, regional blocks, and opioids. All monitoring and intervention measures strictly adhered to institutional clinical practice standards to ensure the safety and rationality of the analgesic strategy. While these concurrent interventions are clinically justified, they may act as potential confounding factors when evaluating the overall analgesic efficacy of LIA. For instance, preoperative preemptive administration of lornoxicam exerts anti-inflammatory and analgesic effects, potentially reducing baseline pain sensitivity prior to surgical trauma. Intraoperatively, fascia iliaca block provides regional analgesia by targeting the femoral nerve, lateral femoral cutaneous nerve, and obturator nerve [4]. Although these interventions are clinically necessary for optimizing patient care, they may independently mitigate pain and reduce opioid requirements, thereby attenuating the magnitude of between-group differences in analgesic outcomes that might otherwise be attributed to LB-based LIA vs ropivacaine.

The lower sleep quality scores observed in the LB group are compatible with the recognized interplay between pain and sleep [33,34]. This pattern is of particular interest in older adult patients, in whom sleep disruption is a known risk factor for delirium [18]. This is supported by the study by Shen et al [35], which confirmed that improving postoperative sleep quality in patients can alleviate early pain and reduce the use of analgesic medications, which aligns with the positive feedback loop observed in the LB group. However, the sleep quality score used in this study was a post hoc constructed composite metric not validated for hospitalized patients. Therefore, the observed between-group differences should be considered a preliminary finding, which may indicate that better early pain control potentially exerts a positive influence on patients’ acute postoperative sleep experience.

LB’s lower incidence of nausea and vomiting is consistent with its pharmacokinetic profile, characterized by slowed absorption from the liposomal formulation [36]. This structural property is thought to contribute to more stable plasma concentrations, which may reduce direct stimulation of the central nervous system [29,37]. Rice et al [28] showed LB reduces the maximum plasma concentration by 30% and prolonged the time to peak concentration by 2.5-fold vs conventional bupivacaine. These properties could plausibly account for a reduction in drug-related gastrointestinal irritation. However, the association observed in this retrospective study and the aforementioned pharmacokinetic explanation are both post hoc inferences, which cannot establish a clinical causal relationship between LB and the reduction of nausea and vomiting. This favorable safety profile was further supported by the absence of significant between-group differences in liver and kidney function parameters or other complications, indicating that LB was well-tolerated across the patient cohort.

COMPARATIVE LITERATURE:

Our findings of meaningful early pain reduction and improved sleep quality with LB differ from several randomized trials reporting null or clinically irrelevant benefits. Bowen et al [13] conducted a prospective randomized non-inferiority trial including 118 THA patients and found that LB added to periarticular injections resulted in only a minor reduction in opioid use, with no significant difference in postoperative pain scores or hospital length of stay. Perets et al [14] reported similar findings in a randomized trial of 107 THA patients: LB infiltration showed no superiority over conventional bupivacaine in pain scores, opioid consumption, hospital stay, or functional recovery up to 72 hours postoperatively. The divergent findings can be attributed to several key distinctions. First, our study used a multi-point, anatomically targeted injection protocol, which may promote more optimal distribution of the sustained-release formulation to critical pain-generating structures. In contrast, the comparator study used a more standardized periarticular injection. Second, our cohort included cases with an acute pain component, whereas the contrasting trial primarily enrolled patients with chronic osteoarthritic pain; the neuroinflammatory milieu of acute pain may be more responsive to prolonged local anesthetic blockade.

LIMITATIONS:

Due to its retrospective design, this study has limitations in the interpretation of the results. While statistical models adjusted for available confounders, this study is still subject to retrospective interpretive bias. For example, the extraction and interpretation of sleep-related data from nursing notes can be influenced by subjective factors of the researchers, and the assessment of pain intensity can be affected by unrecorded patient-related factors, such as pain tolerance and psychological state. Additionally, residual confounding cannot be completely eliminated. Unmeasured variables, such as preoperative opioid tolerance, patient compliance with rehabilitation exercises, and psychosocial factors, may have influenced the study outcomes but were not captured in the EMR data. These biases and confounding factors highlight the need for cautious interpretation of the study results.

This study has several additional limitations. First, its retrospective single-center design and limited sample size carry inherent risks of selection bias. The exclusion of 7 patients–due to Alzheimer disease, postoperative pulmonary infection, or incomplete data–may further affect the representativeness of the sample, as these subgroups have distinct pain perception and rehabilitation potential, thereby limiting the generalizability of the findings. Second, sleep quality was assessed using derived “sleep quality scores” rather than a validated instrument for daily sleep evaluation in hospitalized patients; this score requires further validation against established tools in future studies. Third, although the standardized, intensive multimodal analgesic protocol aligns with clinical ethics and patient care, it likely attenuated observable between-group differences in rescue opioid consumption. This ceiling effect may have obscured any opioid-sparing benefits of LB that might be evident in settings with less comprehensive background analgesia. Fourth, the short follow-up period precludes assessment of long-term functional recovery. Fifth, the cost-effectiveness of LB was not evaluated, which is an important consideration for its broader clinical adoption. Sixth, all surgeries in the present study were performed by the same team of orthopedic surgeons with specialized experience in DAA-THA, which may limit the applicability of the study findings to medical institutions adopting different surgical techniques or with varying provider expertise. Finally, the study cohort was predominantly composed of older adult patients, mostly with femoral neck fracture or femoral head necrosis; thus, the study findings may not be generalizable to younger populations, patients with other underlying hip pathologies, or those undergoing DAA-THA in multicenter settings characterized by diverse patient demographics and clinical practices.

Two of these limitations are of particular importance for interpreting the findings. First, as a retrospective single-center study, residual confounding and treatment-selection bias cannot be fully eliminated despite multivariable adjustment. Second, the sleep composite score was constructed from routine nursing documentation and adapted from PSQI components; therefore, the sleep-related findings should be regarded as exploratory and hypothesis-generating.

Conclusions

MANUSCRIPT PREPARATION:

No third-party manuscript preparation services (whether paid or unpaid) were used in the development of this manuscript. All authors independently contributed to manuscript drafting, revision, and finalization, and all content reflects the authors’ original work.