Age-dependent efficacy of transconjunctival lower eyelid fat removal: receiver operating characteristic curve-derived thresholds and multimodal recommendations
Article information
Abstract
Background
Transconjunctival lower eyelid fat removal is widely performed in Asian aesthetic practice, yet the domain-specific outcomes and the effects of age or adjunctive procedures remain unclear.
Methods
Fifty patients (median age, 50 years) completed a 1-month postoperative assessment. Satisfaction with bulging, pigmentation, and laxity was evaluated using the 5-point Periorbital Satisfaction Likert Scale (PSLS). Preoperative and postoperative photographs were independently assessed by two blinded physicians, and their mean score constituted the physician rating. Nonparametric tests, receiver operating characteristic (ROC) curve analysis (area under the curve, AUC), and ordinal logistic regression were performed.
Results
Median patient scores were 5 for bulging and 4 for pigmentation and laxity. Physician scores were significantly lower for pigmentation (P<0.001) and laxity (P=0.002) but not for bulging (P=0.080). Patient–physician correlations were moderate to strong (ρ=0.55–0.70), and inter-rater reliability was good (intraclass correlation coefficient=0.78). Patients younger than 50 years rated bulging (P=0.026) and laxity (P=0.003) more favorably than older patients. ROC analysis identified age cutoffs of 54 years for bulging (AUC=0.61) and pigmentation (AUC=0.51) and 47 years for laxity (AUC=0.75). Age negatively predicted bulging (odds ratio [OR], 0.95) and laxity (OR, 0.94), whereas periosteal hyaluronic acid filler positively predicted bulging (OR, 6.90) and pigmentation (OR, 5.05).
Conclusions
One month after surgery, satisfaction is highest for bulging but declines beyond the late forties, particularly for laxity. Deep periosteal hyaluronic acid augmentation enhances outcomes in bulging and pigmentation. The PSLS demonstrated good convergent validity and may support age-specific, multimodal strategies for periorbital rejuvenation.
INTRODUCTION
Dark circles and puffiness of the lower eyelid—colloquially known as “kuma” in Japan—arise from at least three interacting mechanisms: (1) bulging caused by anterior herniation of orbital fat, (2) pigmentation due to vascular congestion or dermal melanosis, and (3) laxity resulting from age-related attenuation of the skin and supporting structures. Among Asian patients, orbital fat protrusion is often more pronounced, leading to substantial aesthetic and psychosocial consequences [1]. Transconjunctival lower eyelid fat removal is widely used as a minimally invasive technique to reduce bulging; because it leaves no external scar and involves minimal downtime, it ranks among the most frequently performed procedures in our clinics, following double-eyelid surgery and thread lifting [2].
Recent studies suggest that transconjunctival fat removal may affect not only bulging but also pigmentation and laxity, potentially through alterations in local microcirculation [3]. Nevertheless, few investigations have separately quantified these three components using domain-specific patient-reported outcomes, and the influence of aging or adjunctive treatments, such as periosteal hyaluronic acid (HA) filler, Sunekos, type I/III collagen injections, or thread lifting, on satisfaction remains poorly defined.
The primary aim of this study was to assess 1-month postoperative patient satisfaction across three domains—bulging, pigmentation, and laxity—following transconjunctival fat removal and to characterize the overall distribution of scores. The secondary aims were threefold: (1) to examine the effect of age on satisfaction using dichotomous and receiver operating characteristic (ROC) curve analyses to identify optimal age cutoffs; (2) to determine whether adjunctive procedures independently influence satisfaction in each domain through ordinal logistic regression; and (3) to compare patient self-ratings with blinded physician assessments to evaluate the convergent validity of the newly developed Periorbital Satisfaction Likert Scale (PSLS). This short-term, exploratory pilot study is intended to provide foundational data for larger, long-term investigations and to guide the development of individualized algorithms for periorbital rejuvenation.
METHODS
This prospective observational study included consecutive patients who underwent transconjunctival lower eyelid fat removal at Tokyo Chuo Beauty Clinic and Kobe Clinic between December 2024 and February 2025. Among 91 initially registered cases, 50 patients (6 men and 44 women) completed both the questionnaire and periorbital photography at the 1-month follow-up and were therefore included in the final analysis. The study protocol was approved by the Tokyo Chuo Beauty Clinic Institutional Review Board (UMEDA-ERB-2025-May-006), and written informed consent was obtained from all participants.
Outcome measures
Postoperative satisfaction was assessed using the newly developed PSLS. Patients rated three domains (bulging, pigmentation, and laxity) on a 5-point scale ranging from 1 (very dissatisfied) to 5 (very satisfied); anchor definitions are provided in Supplementary Table 1. Additionally, anonymized close-up photographs taken preoperatively and at 1 month postoperatively were independently evaluated by two physicians using the same anchors. The mean of the two ratings was used as the physician score. Inter-rater reliability was calculated using the intraclass correlation coefficient (ICC(2,1)).
Statistical analysis
All analyses were performed using EZR (version 1.68). Because Shapiro-Wilk tests indicated non-normal distributions across all domains, data were summarized as medians with interquartile ranges (IQR), and nonparametric tests were applied. Paired patient–physician scores were compared using the Wilcoxon signed-rank test, and rank correlations were calculated using Spearman’s ρ. Patients were dichotomized into younger (<50 years) and older (≥50 years) groups based on the sample median; between-group differences were examined using the Mann-Whitney U test. Age was also treated as a continuous variable, and ROC analysis was performed to identify age cutoff values that best predicted high satisfaction (score ≥4) for each domain, with optimal thresholds defined by the Youden index (J).
To identify factors associated with higher satisfaction, proportional-odds ordinal logistic regression was conducted using the five-point domain score as the dependent variable. Independent variables included age (continuous) and the presence or absence (dummy-coded) of adjunctive procedures—periosteal HA filler, Sunekos, type I/III collagen (Baby Collagen) injection, and thread lifting. Variance inflation factors (VIFs) were computed, and multicollinearity was considered negligible for VIF <2. Because this was an exploratory pilot study, no adjustment for multiple testing was applied. Sample-size estimation using G*Power 3.1 (effect size=0.25, α=0.05, power=0.80) for a three-level repeated-measures Friedman test indicated that 28 cases would be sufficient; nevertheless, all 50 complete cases were included in the analysis. Missing data were not imputed.
RESULTS
A total of 91 patients were enrolled, of whom 50 (6 men and 44 women) completed the 1-month follow-up (Table 1). The median age was 50 years. The median satisfaction scores on the 5-point PSLS were 5 (4–5) for bulging, 4 (3–5) for pigmentation, and 4 (3–5) for laxity. The Friedman test revealed no significant differences among the three domains (p=0.623) (Table 2). All score distributions were non-normal according to the Shapiro-Wilk test (P<0.01).
Patient demographics and adjunctive procedures (n=50)
Satisfaction scores at 1 month
Patient self-ratings were compared with blinded physician ratings (Table 3; anchor definitions in Supplementary Table 1). The median physician scores were 4 (4–5) for bulging, 3 (3–4) for pigmentation, and 3 (2–3) for laxity. The Wilcoxon signed-rank test indicated significant differences for pigmentation (P<0.001) and laxity (P=0.002), whereas the difference for bulging did not reach significance (P=0.080). Spearman rank correlations between patient and physician scores were ρ=0.70 for bulging, ρ=0.60 for pigmentation, and ρ=0.55 for laxity (all P<0.001). Inter-rater reliability between the two physicians was good (ICC(2,1)=0.78; 95% CI, 0.66–0.86).
One-month satisfaction scores: patients versus blinded expert evaluators (Periorbital Satisfaction Likert Scale)
When patients were dichotomized at the median age of 50 years, the younger group (<50 years) demonstrated higher satisfaction scores for bulging (P=0.026) and laxity (P=0.003), whereas pigmentation showed no significant difference between age groups (P=0.078) (Table 4). ROC analysis identified age thresholds of 54 years for both bulging and pigmentation (area under the curve [AUC]=0.606 [95% CI, 0.411–0.801] and 0.512 [95% CI, 0.332–0.692], respectively) and 47 years for laxity (AUC=0.754 [95% CI, 0.613–0.894]). The corresponding sensitivity and specificity values are presented in Table 5 and Fig. 1.
Age-stratified comparison of satisfaction
ROC-derived age thresholds for high satisfaction
Age cutoff analysis for high postoperative satisfaction (score ≥4) 1 month after transconjunctival lower eyelid fat removal. (A) Bulging: optimal threshold 54 years (sensitivity, 82.4%; specificity, 46.2%; area under the curve [AUC], 0.606). (B) Pigmentation: threshold 54 years (sensitivity, 80.0%; specificity, 33.3%; AUC, 0.512). (C) Laxity: threshold 47 years (sensitivity, 83.3%, specificity, 58.1%; AUC, 0.754). Diagonal dotted line=reference (AUC 0.5).
Results of the ordinal logistic regression are shown in Table 6 and Fig. 2. For bulging, both age (odds ratio [OR], 0.95; 95% CI, 0.89–0.99; P=0.024) and periosteal HA filler (OR, 6.90; 95% CI, 1.56–30.60; P=0.013) were significant predictors. For pigmentation, HA filler remained a significant predictor (OR, 5.05; 95% CI, 1.18–23.60; P=0.033). For laxity, age was the only significant predictor (OR, 0.94; 95% CI, 0.89–0.98; P=0.009). VIFs were <2 across all models, indicating no relevant multicollinearity.
Ordinal logistic regression for high satisfaction
Forest plots from ordinal logistic regression predicting postoperative satisfaction (5-point scale) at 1 month. Odds ratios (ORs) are displayed on a logarithmic scale; horizontal bars represent 95% confidence intervals (CIs); and the vertical dashed line marks OR=1. (A) Bulging: age per year (OR, 0.95; 95% CI, 0.89–0.99; P=0.024); periosteal hyaluronic acid (HA) filler (OR, 6.90; 95% CI, 1.56–30.60; P=0.013); Sunekos (OR, 0.82; 95% CI, 0.18–3.75; P=0.79); type I/III collagen (OR, 1.10; 95% CI, 0.21–5.73; P=0.92); thread lift (OR, 1.48; 95% CI, 0.35–6.21; P=0.59). (B) Pigmentation: age (OR, 0.98; 95% CI, 0.92–1.03; P=0.410); periosteal HA filler (OR, 5.05; 95% CI, 1.18–23.60; P=0.033); Sunekos, collagen injection, and thread lift were non-significant. (C) Laxity: age (OR, 0.94; 95% CI, 0.89–0.98; P=0.009); all adjunctive procedures were non-significant.
DISCUSSION
This study evaluated 1-month postoperative satisfaction after transconjunctival lower eyelid fat removal across three domains—bulging, pigmentation, and laxity—and examined the influence of age and adjunctive procedures. Although median patient ratings were ≥4 in all domains, physician ratings obtained under blinded conditions were significantly lower than patient ratings for pigmentation and laxity (P<0.001 and P=0.002, respectively), while no significant difference was observed for bulging (P=0.080). The moderate-to-strong rank correlations between patient and physician scores (ρ=0.55–0.70) support the convergent validity of the PSLS. Because patients tended to rate pigmentation and laxity more favorably than physicians, preoperative counseling should carefully manage expectations for these domains.
A nationwide Japanese survey by Funakoshi et al. [1] reported that lower-eyelid “kuma” surgery ranks as the third most common cosmetic procedure after suture blepharoplasty and HA filler injection. Tamura et al. [2] analyzed 127,505 cases and found that the largest patient group was aged 40–49 years. The median age of 50 years and the ROC-derived cutoffs of 47–54 years in the present study are consistent with these national datasets, confirming that the highest demand for lower eyelid rejuvenation occurs among middle-aged patients.
With regard to the surgical approach, Rancati et al. [3] emphasized that the transconjunctival technique avoids external scarring and preserves lateral canthal support. For Asian patients, a volume-preserving strategy that avoids excessive fat removal is recommended. Both Guo et al. [4] and Sarhaddi et al. [5] have advocated combining fat preservation or repositioning with augmentation techniques. The high satisfaction with bulging reduction observed in the present study likely reflects adherence to a similar “preservation plus deep augmentation” approach.
For pigmentation, shadowing at the lid–cheek junction (tear trough) is a major contributing factor, as noted by Vrcek et al. [6]. Consistent with this, periosteal HA filler was an independent positive predictor of satisfaction in both pigmentation and bulging (OR, 5.05 and 6.90, respectively). This finding aligns with the multifactorial intervention algorithm proposed by Bhattacharjee et al. [7] and supports deep HA augmentation as an effective adjunct for short-term aesthetic enhancement in Asian patients.
Age independently decreased satisfaction with bulging (OR, 0.95) and particularly with laxity (OR, 0.94), with a marked decline at 47 years (AUC, 0.75). Progressive loss of dermal elasticity and septal integrity may limit the tightening effect achievable through fat removal alone. Thus, patients in their fifties and beyond are likely to benefit from a multimodal strategy combining conservative fat removal with deep HA augmentation or other skin-tightening modalities. Notably, the AUC for pigmentation was 0.51, indicating that the age threshold for this domain has limited discriminative value and should be interpreted cautiously.
This single-center pilot study is limited by its small sample size and short 1-month follow-up period; therefore, long-term outcomes and inter-institutional reproducibility remain to be confirmed. Because consent for scholarly image publication was not obtained, patient photographs were not published, and objective third-party evaluation was restricted to blinded physician scoring. Although the PSLS demonstrated good convergent validity, its construct validity and minimal clinically important difference have yet to be established. Adjunctive procedures were analyzed as binary variables, precluding evaluation of dose–response or technique effects. No correction for multiple testing was applied, and baseline severity of fat protrusion, pigmentation, and skin laxity was not stratified, leaving possible residual confounding by age. Dedicated, objective tools for lower eyelid evaluation are currently lacking, and unified re-photography was not feasible retrospectively. Additionally, the proportional-odds assumption was not formally tested; future analyses will incorporate diagnostic checks. Prospective studies employing VECTRA 3D imaging or skin analysis devices are warranted.
In conclusion, transconjunctival lower eyelid fat removal provides high short-term satisfaction for bulge reduction; however, discrepancies between patient and physician ratings are more likely in pigmentation and laxity. Satisfaction declines markedly after 47–54 years of age, whereas periosteal HA filler independently enhances satisfaction in bulging and pigmentation. Age-tailored, multimodal treatment algorithms are therefore essential to optimize outcomes and maximize patient satisfaction in lower eyelid rejuvenation.
Notes
No potential conflict of interest relevant to this article was reported.
Ethical approval
The study was approved by the Tokyo Chuo Beauty Clinic Institutional Review Board (No. UMEDA-ERB-2025-May-006) and performed in accordance with the principles of the Declaration of Helsinki. Written informed consent was obtained from all participants prior to the study.
Supplemental material
Supplementary materials can be found via https://doi.org/10.14730/aaps.2025.01445
Supplementary Table 1.
Domain-specific anchors for the Periorbital Satisfaction Likert Scale