Five-factor modified frailty index is a reliable indicator for predicting postoperative mortality in geriatric intertrochanteric fracture patients aged 80 years and older: a case-control study

Abstract

Background

To investigate the reliability and threshold of five-factor modified frailty index (mFi-5) for predicting postoperative mortality in patients aged 80 years and older with intertrochanteric fractures (ITF).

Methods

This retrospective study included advanced-age ITF patients who underwent surgery between January 2016 and January 2020 at a single center, and the eligible patients were sequentially divided into six groups according to mFi-5 scores. The mortalities in in-hospital, 3-months and 1-year after surgery were defined as observed outcomes. Baseline characteristics, including age, gender, time to surgery, surgery time, length of hospital, anesthesia, anemia and hypoproteinemia at admission, were collected for logistic regression analysis. Then, Receiver-Operating Characteristic (ROC) was used to assess reliability, and decision curve analysis (DCA) was used to evaluate the clinical benefit of the mFi-5.

Results

263 patients were included in the final analysis. The in-hospital, 3-month and 1-year all-cause mortalities of all patients were 4.2% (11/263), 14.1% (37/263) and 24.7% (65/263). Overall, the mortalities increased with mFi-5 score. There were no significant between-group differences of baseline characteristics between the groups. Multivariate regression analysis indicated only gender was an independent factor for in-hospital morality, gender, anemia and high mFi-score for 3-month mortality, and age, gender, anemia and high mFi-score for 1-year mortality. The ROC analysis shown that the areas under the curve of mFi-5 for predicting in-hospital, 3-month and 1-year morality were 0.862, 0.773 and 0.756, respectively P < 0.01). DCA demonstrated that using an mFi-5 threshold of ≥ 3 provided a higher net benefit.

Conclusions

In conclusion, mFi-5 is a reliable prognostic predictor in the advanced-age ITF patients following surgery. Also, high risk of mortality is associated with a mFi-5 score no less than “three” especially in the much older male patient with anemia.

Background

The incidence of hip fractures is increasing rapidly due to the rapid development of aging and osteoporosis, which is predicted to affect 1.3 million in people aged 65 years and older in China by year [1]. Surgery is the gold standard for treating these hip fractures. Despite the improvements of perioperative plan and deepening concept of accelerated surgery in hip fractures, the prognosis remains suboptimal, as evidenced by a high 1-year postoperative mortality rate (up to 36%) [2, 3]. Intertrochanteric fracture (ITF) is one of the most common types of hip fractures. Intertrochanteric fracture (ITF) is one of the most common types of hip fractures. Although hip fractures are often discussed collectively, significant characteristic differences exist between ITF and femoral neck fracture (FNF) [4–8]. The comparative mortality rates between these fracture types remain a topic of discussion. While some studies have reported higher 1-year mortality following ITF compared to FNF [5, 6], other evidence suggests that mortality may be higher for FNF [7, 8]. This apparent contradiction may be partly explained by the primary treatment modality employed for each fracture type. Critically, the choice of procedure itself is a significant determinant of outcomes. Thus, the debate on mortality may be more about the complex interplay between fracture pattern, subsequent treatment selection, and the patient’s baseline physiology. Given this multifactorial complexity, there is an urgent need for reliable predictive tools, such as frailty indices, to improve prognosis for all hip fracture patients, with a particular interest in high-risk patients.

Early recognition of high-risk ITF patients is not only a prerequisite for timely effective intervention during perioperative period, but also helpful for both orthopedists and patients in finding common ground on the treatment plan and prognosis. Frailty has shown to be strongly associated with adverse outcomes in geriatric patients underwent orthopedic surgery, which is also commonly found in the hip fractures [9–11]. Frailty refers to a geriatric state that reflects the body’s decline in physiological function, which may reduce the ability to respond to stress [12]. Higher frailty levels imply higher risk of postoperative complication and worse prognosis. There are various risk stratification tools for identifying high-risk populations including American Society of Anesthesiologists (ASA) classification [13], FRAIL scale [14] and Charlson comorbidity index (CCI) [15] and Elixhauser Comorbidity Measure (ECM) [16]. However, these measures have some shortcomings in clinical application and study. ASA and FRAIL scale were subjective and unverifiable assessments, of which the results may vary widely from different evaluators. As for CCI, although they have shown to be effective in assessing preoperative condition and predict postoperative complications in the elderly with orthopedic surgery [17, 18], these measurement tools account for over ten comorbid variables that were not routine data in medical and nursing notes. This limitation renders complex tools like CCI and ECM impractical for routine clinical use. Thus, a simple and reliable measurement tool for risk stratification in ITF should be explored and validated.

The 5-factor modified frailty index (mFi-5) is an objective frailty indicator which is adjusted on the basis of traditional mFi-11 within the National Surgical Quality Improvement Program (NSQIP) database [19]. In contrast to mFi-11, mFi-5 include five comorbidities which are commonly seen in the elderly and easily available in electronic medical records. Also, multiple studies have demonstrated mFi-5 have equal sensitivity and accuracy with mFi-11 in predicting postoperative complication after orthopedics surgery [19–21]. And a study involved 58,603 patients with hip fractures based on NSQIP database, have demonstrated that mFi-5 is an effective predictor of morbidity and mortality in elderly patients following hip fracture surgery [20]. However, this study included over 20% patients aged no more 75 years and investigated only short-term prognosis, and the conclusion may be not applicable to the long-term prognosis of advanced-age (aged 80 years and older) patients. On the other hand, as the average life expectancy in urban China has been over 80 years [22, 23], it is likely more meaningful to focus on this advanced-age population. Furthermore, it has been demonstrated that postoperative mortality of advanced-age ITF patients was much higher than those relatively young patients [6], which indicates this advanced-age population may have different characteristics compared with young ITF patients. Few studies have been validated the predictive power and value of mFi-5 in this advanced-age population. Therefore, whether mFi-5 is a reliable prognosis predictor in advanced-age ITF patients remain unclear.

Although lower mFi-5 score has been confirmed to attribute to healthier status and better prognosis [20], advanced-age ITF patients with lower mFi-5 score may still suffer suboptimal outcome because of their worse physical function and anti-stress ability. We speculate that mFi-5 is reliable in predicting postoperative mortality in advanced-age ITF patients. Besides, the cut-off values of mFi-5 are likely be different in predicting short- and long-term mortality. Therefore, we conduct a retrospective case-control study to investigate the reliability of mFi-5 for predicting mortality in advanced-age ITF patients, and the predictive power of mFi-5 for mortality at different time points.

Patients and methods

Study design and participants

This study was a retrospective case-control study using the database from orthopedic center in Guangzhou First People’s Hospital, which was approved by the Ethics Committee of Guangzhou First People’s Hospital (B-2022-065-01). Inclusion criteria were radiographic diagnosis of femoral intertrochanteric fractures and patients who underwent treatment from January 2016 to January 2020. Exclusion criteria were (1) aged less than 80 years, (2) non-surgical treatment, (3) Multiple trauma or high-energy injury, (4) pathological or atypical fractures, (5) hematologic disorders, (6) acute or chronic infection, (7) long-term of anti-platelet therapy (8) end-stage malignant tumor, (9) chronic renal insufficiency, (10) missing complete blood count data, 11) insufficient follow-up information (less than one year).

Grouping and variables

All patients in our study underwent internal fixation using the same standardized AO techniques to minimize variability. Furthermore, all surgeries were performed by the same senior orthopedic surgeon following consistent protocols. The mFi-5 comprises five comorbidities: diabetic status, chronic obstructive pulmonary disease or current pneumonia, congestive heart failure, hypertension requiring medication and non-independent functional status [20]. A patient with no comorbid conditions at admission would have a mFi-5 of zero, and each comorbidity was assigned one point. The included patients in the study were divided into six groups based on the mFi-5 score (from zero to five, the Group 1 means mFi-5 = 0, the Group 2 means mFi-5 = 1, and so on). Demographic characteristics and laboratory data were obtained from the electronic medical records. Considering the potential impact of other factors on prognosis, demographic characteristics including age, gender, time to surgery, surgery time, anesthesia, length of stay and anemia (hemoglobin < 70 g/L) were collected for analysis.

Outcomes and follow-up

All-cause mortalities at discharge, within three months and one year after surgery were determined as study outcomes. The date of in-hospital death was obtained from electronic medical records. To those living patients at discharge, telephone follow-up was performed to evaluate the survival at 3 months and one year after surgery.

Statistical analysis

All statistical analyses were performed with SPSS Version 25.0. Continuous variables were assessed for normal distribution with Shapiro-Wilk test. Normally distributed variables were expressed as means (standard deviation, SD) while non-normally distributed variables were expressed as median (interquartile range, IQR). One-way analysis of variance was applied for normally distributed continuous variables and the Kruskal-Wallis H test for non-normally distributed continuous variables. Chi-squared or Fisher’s exact tests were used to assess differences for categorical variables. A two-tailed p-value below 0.05 was considered statistically significant. The significant variables in univariate analysis (p<0.1) and potential significant predictors (including age, gender, time to surgery, and anemia) were entered in multivariate analysis with a by two-way stepwise regression method for the best model. Moreover, Receiver operating characteristic (ROC) curve was drawn to calculate predictive power of the discriminative ability of the continuous mFi-5 score according to area under the curve (AUC). Clinical utility of the best threshold of mFi-5 was evaluated by reporting threshold-specific metrics including sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), accuracy, and Youden’s index at three timepoints. Decision curve analysis (DCA) was performed to further assess the clinical value of the mFi-5 score in decision-making. This method quantifies the net benefit of using the prediction model across a range of threshold probabilities, comparing it against strategies of intervening for all patients or no patients. DCA was conducted separately for each timepoint to evaluate the model’s utility in predicting in-hospital, 3-month, and 1-year mortality. These plots were generated using R studio (Version 4.1.2).

Results

Between January 2016 and January 2020, a total of 504 patients diagnosed with ITF underwent treatment in our department. In this study, 241 patients were excluded according to the exclusion criteria. Finally, 263 eligible patients were included in the analysis, and there were eight patients in Group 1, 70 in Group 2, 70 in Group 3, 69 in Group 4, 30 in Group 5 and five in Group 6. The flow chart was presented in Fig. 1.

Fig. 1.

Patients flow chart

Baseline characteristics were shown in Table 1. The median [IQR] age of all patients is 85.0 [83,0, 89.0] years, and the proportion of female patients was 67.7% (178/263). There were no significant between-group differences in the baseline characteristics including age, gender, time to surgery, anesthesia, surgery time, LOS, and anemia (P > 0.05).

Table 1.

Baseline characteristics

Variables	Total (n = 263)	mFi-5 = 0 (n = 8)	mFi-5 = 1 (n = 70)	mFi-5 = 2 (n = 69)	mFi-5 = 3 (n = 81)	mFi-5 = 4 (n = 30)	mFi-5 = 5 (n = 5)	P
Age, year	85.0 [83.0, 89.0]	82.0 [80.8,83.5]	86.0 [83.0,89.0]	85.0 [83.0,87.0]	85.0 [83.0,90.0]	87.5 [83.3,91.8]	85.0 [81.0,92.0]	0.052
Female (%)	178 (67.7)	5 (62.5)	47 (67.1)	50 (72.5)	53 (65.4)	19 (63.3)	4 (80.0)	0.906
TTS < 48 h	18 (6.8)	0 (0.0)	5 (7.1)	4 (5.8)	5 (6.2)	4 (13.3)	0 (0.0)	0.767
Surgery time, min	75.0 [60.0, 95.0]	94.0 [85.3,101.3]	70.0 [60.0,90.0]	72.0 [60.0,95.0]	80.0 [62.5,95.0]	82.0 [65.0,105.0]	80.0 [53.0,80.0]	0.184
LOS, day	15.0 [11.0, 18.3]	17.5 [9.8,20.0]	14.0 [10.0,17.0]	16.0 [13.0,20.0]	15.0 [12.0,19.0]	13.5 [10.3,17.0]	16.0 [13.0,23.0]	0.329
Spinal anesthesia (%)	197 (74.9)	5 (62.5)	52 (74.3)	55 (79.7)	57 (70.4)	24 (80.0)	4 (80.0)	0.694
Hgb < 70 g/L (%)	7 (2.7)	0 (0.0)	3 (4.3)	1 (1.5)	2 (2.5)	1 (3.3)	0 (0.0)	0.830

mFi-5 Five Factor Modified Fragility Index, TTS Time from Admission to Surgery, LOS Length of Stay, Hgb Hemoglobin

The in-hospital, 3-month and 1-year all-cause mortalities of all the 263 patients were 4.2% (11/263), 14.1% (37/263) and 24.7% (65/263), respectively. The mortalities at different time points in the six groups were shown in the Table 2 and Fig. 2. No patients in the Group 1, Group 2, Group 3 and Group 6 died during hospitalization. The in-hospital mortality is 4.9% in the Group 4 and 23.3% in the Group 5. The 3-month and 1-year mortality were increased with the mFi-5 score. Fitting curves between mortality rates at the three time points and mFi-5 score were shown in Fig. 2.

Table 2.

Comparison of the in-hospital, 3-month and 1-year mortality in the different groups

Mortality	Total (n = 263)	mFi-5 = 0 (n = 8)	mFi-5 = 1 (n = 70)	mFi-5 = 2 (n = 69)	mFi-5 = 3 (n = 81)	mFi-5 = 4 (n = 30)	mFi-5 = 5 (n = 5)
In-hospital (%)	11 (4.2%)	0	0	0	4 (4.9%)	7 (23.3%)	0
3-month (%)	37 (14.1%)	0	4 (5.7%)	1 (1.5%)	17 (20.1%)	13 (43.3%)	2 (40.0%)
1-year (%)	65 (24.7%)	1 (12.5%)	8 (11.4%)	3 (4.4%)	32 (39.5%)	18 (60.0%)	3 (60.0%)

Fig. 2.

The relationship between the mFi-5 score and in-hospital, 3-month and 1-year mortality. (a) The histograms shown that mortality rate at the three time points increased with mFi-5 score. No death was observed during hospitalization in mFi-5 score 0, 1, 2 and 5. The in-hospital mortality rates in the score of 3 and 4 were 4.9% and 23.3%. No patients died within 3 months in mFi-5 score 0. In order from low score to high score, the 3-month mortality rates were 0, 5.7%, 1.5%, 20.1%, 43.3% and 40.0%, and the 1-year mortality rates were 12.5%, 11.4%, 4.4%, 39.5%, 60.0% and 60.0%. (b) The fitting curve shown that mortalities at the three time points increases with mFi-5 score at different growth rates. The increase of in-hospital morality was linear while those of 3-month and 1-year mortality were exponential

The ROC curve analysis of mFi-5 for predicting mortality was shown in Fig. 3. For all patients, the AUCs of mFi-5 in predicting in-hospital, 3-month and 1-year mortality after surgery were 0.862, 0.773 and 0.756 (P < 0.01), “with an optimal cut-off value of 2.5. For convenience of clinical application, the threshold-specific metrics of mFi-5 ≥ 3 was done to explore their predictive sensitivity and specificity. The sensitivity, specificity, accuracy, PPV, NPV, and Youden’s index of mFi-5 ≥ 3 for predicting mortality at the three timepoints are presented in Table 3.

Fig. 3.

ROC analysis for the predictive threshold for mFi-5 on mortality at the three time points. (a-c) Results from ROC curve shown mFi-5 was reliable in predicting in-hospital (AUC = 0.862, P < 0.01), 3-month (AUC = 0.773, P < 0.01) and 1-year (0.756, P < 0.01) mortality

Table 3.

Performance metrics of mFi-5 ≥ 3 for predicting mortality at three timepoints

Timepoint	Sensitivity	Specificity	PPV	NPV	Accuracy	Youden’s index
In-hospital	1.000	0.583	0.095	1.000	0.601	0.583
3-Months	0.865	0.628	0.276	0.966	0.662	0.493
1-Year	0.815	0.682	0.457	0.918	0.715	0.497

mFi-5 5-factor modified Frailty Index, PPV Positive Predictive Value, NPV Negative Predictive Value, Youden’s index Calculated as (Sensitivity + Specificity − 1)

The DCA for mFi-5 ≥ 3 is shown in Fig. 4. The net benefit curves intersected with the ‘treat-all’ reference line at threshold probabilities of approximately 5% and 10% for in-hospital mortality, 15% and 28% for 3-month mortality, and 25% and 45% for 1-year mortality. The y-intercepts of the net benefit curves were 0.04 for in-hospital mortality, 0.11 for 3-month mortality, and 0.18 for 1-year mortality.

Fig. 4.

Decision Curve Analysis for the mFi-5 Score in Predicting in-hospital Mortality, 3-month and 1-year mortalities after surgery. Decision curve analysis showing the net benefit of using the mFI-5 score (dichotomized at ≥ 3) for predicting mortality compared to strategies of treating all patients or no patients. The net benefit curves intersected with the ‘treat-all’ reference line at threshold probabilities of approximately 5% and 10% for in-hospital mortality, 15% and 28% for 3-month mortality, and 25% and 45% for 1-year mortality

To control for confounding variables, age, gender, time to surgery within 48 h, anemia and mFi-5 greater than or equal to three, were included for multiple logistic regression (Table 4). The multivariate regression model for in-hospital morality shown that only gender was an independent factor (P = 0.046, OR = 0.28). The result for 3-month morality revealed that gender (P = 0.044, OR = 0.45), anemia (P = 0.043, OR = 7.48) and high mFi-score (P = 0.001, OR = 11.67) were risk factors. Also, the multivariate regression result found that age (P = 0.001, OR = 1.15), gender (P = 0.002, OR = 0.34), anemia (P = 0.044, OR = 6.58) and high mFi-score (P = 0.001, OR = 11.05) were significantly associated with 1-year mortality.

Table 4.

Multivariate logistic regression of mFi-5 and confounding variables for in-hospital, 3-month and 1-year mortality

Variables	In-hospital mortality			3-month mortality			1-year mortality
	OR	95%CI	P	OR	95%CI	P	OR	95%CI	P
Age	1.10	0.97–1.25	0.137	NA	NA	NA	1.15	1.06–1.24	0.001
Female	0.28	0.08–0.98	0.046	0.45	0.21–0.98	0.044	0.34	0.17–0.68	0.002
TTS > 48 h	NA	NA	NA	NA	NA	NA	2.97	0.86–10.18	0.084
Hgb < 70 g/L	NA	NA	NA	7.48	1.07–52.38	0.043	6.58	1.05–41.23	0.044
mFi-5 ≥ 3	NA	NA	NA	11.67	4.23–32.17	0.001	11.05	5.11–23.86	0.001

OR Odds Ratio, CI Confidence Interval, TTS Time to Surgery, Hgb Hemoglobin, mFi-5 Five Factor Modified Fragility Index

NA Not applicable

Discussion

As highlighted in the literature, ITF was strongly associated with decreased functional outcome and increased morbidity and mortality, especially to those patients with advanced ages [6, 24]. In this context, an effective and simple prognostic assessment for early recognition in this population is markedly important. mFi-5 is a composite indicator which has been demonstrated its predictive value for mortality follow different surgeries [19, 20, 25]. However, the impact of mFi-5 on in advanced-age ITF patients has received little attention, and a threshold of mFi-5 for identifying high mortality risk is unknown in this population. Thus, we conduct a retrospective study to investigate the reliability of mFi-5 for predicting mortality in advanced-age ITF patients, and the criterion of mFi-5 value for identifying poor prognosis in this population. Our study indicates that mFi-5 is a reliable prognostic predictor in advanced-age ITF patients following surgery, which particularly apply to long-term prognosis. Besides, this advanced-age population with mFi-5 greater than or equal to three points suffers high risk of 3-month and 1-year mortality. mFi-5 is simple, easily available and high sensitivity measurement to evaluate prognosis in ITF with advanced-age, which can be used by orthopedists to attach more importance in early.

The most important finding in our study was that mFi-5 was associated with postoperative mortality in advanced-age ITF patients. The risk of mortality increased with the mFi-5 values in this advanced-age population (Fig. 2). Our results were consistent with some previous conclusions that mFi-5 independently predicted 1-year survival after hip fracture surgery [20, 25]. It is not difficult to understand that postoperative prognosis gets worse with higher frailty level or more comorbidities. However, the unexpected finding was no death occurred in the patients with mFi-5 scoring five points. The main reason might be only five patients in this group. Actually, there are few patients with mFi-5 equal to five points and fewer of them underwent surgical treatment in clinical practice. Patients with frailty and multiple comorbidities were more likely have complications and reduced mobility, which made enhanced recovery after surgery and post-discharge rehabilitation difficult to implement [26–28]. These cumulative adverse effects would probably lead to suboptimal long-term prognosis.

There were various research focusing specifically on whether mFi-5 was associated with postoperative outcomes in hip fracture [20, 25, 29, 30], but little published literature has addressed the impact of mFi-5 on geriatric patients with fracture. There were also few studies focused on the detailed predictive value of mFi-5. We considered the elderly with fracture have different characteristics and threshold of mFi-5. The comprehensive assessment of mFi-5 ≥ 3 revealed its substantial utility in predicting mortality across different timepoints (Fig. 3 and Table 3). Although our model shown that the Youden’s index at 3 months and 1 year is only marginally close to 0.5, with only the in-hospital value reaching 0.583, the principal utility may be derived from the consistently high sensitivity and negative predictive value across all follow-up periods. This profile is a screening tool for identifying a low-risk cohort that may be managed with standardized pathways. This high sensitivity, however, inherently results in a higher false-positive rate, meaning the model captures most true high-risk patients but also classifies some lower-risk individuals into the high-risk group (mFI-5 ≥ 3). In clinical practice, this trade-off can be justified, as it ensures that scarce resources for intensive perioperative optimization (e.g., comprehensive MDT and ERAS protocols) are prioritized for all potential candidates, thereby minimizing the chance of missing a vulnerable patient who would benefit greatly from such intervention. The decision curve analysis (DCA) complemented these findings by quantifying the net clinical benefit of the risk stratification strategy derived from the mFI-5. The DCA evaluated the strategy of classifying patients as high-risk based on an mFI-5 score ≥ 3. This strategy demonstrated a higher net benefit across a range of threshold probabilities. This implies that implementing this simple classification rule to guide decision-making—specifically, to identify high-risk patients who may benefit from intensified perioperative management—can improve patient outcomes and optimize resource use. The increasing net benefit with longer follow-up intervals aligns with the improved PPV and specificity over time, reinforcing the utility of this risk stratification approach for long-term prognosis.

Frailty has been defined as a complex multiorgan condition in clinical rather than in physiological [31], but there was no gold standard for frailty measure in the elderly with hip fractures. Various study had investigated the comparison of different frailty measurement tools including CCI, mFi-11 and mFi-5 following hip surgery [30, 32, 33]. Our study did not compare mFi-5 with these tools directly. We considered that although the CCI and ECM have been shown to outperform the mFi-5 in discriminative ability, it seemed to be plausible because a tool with more variables could have greater sensitivity and specificity for predicting outcomes. Moreover, most of variables in CCI and ECM could not be obtained routinely, which caused their application limited in clinical usage and research. As for mFi-5, it comprised five common diseases that can be easily extracted from the patient history. This advantage would make mFi-5 a more streamlined and practical tool for routine preoperative screening. However, due to a small sample size, we did not make more refined grouping in high mFi-5 score (greater than or equal to three) to further analyze the differences of mortality risk between different subgroups. Further large sample size studies are required to explore this issue.

There are several weaknesses in our study that are worth noting. The first one is the retrospective design, limited sample size and inadequate follow-up. Second, our study only included ITF patients over 80 years to increase the probability of outcome events so application value of mFi-5 cannot be verified in the other patients. However, there may be limited implication to investigate the mortality of those relatively young and healthy patients. Third, as mentioned above, patients were divided according to the five variables given equal weight in mFi-5, which might not accurately reflect each comorbidity’s severity and impact on patients’ overall health. Finally, this stratification could not well consider other potential comorbidities. Although we have tried to reduce bias and potential errors by excluding some comorbidities such as hematologic disorders, advanced malignant tumor and so on (Fig. 1), the clinical usefulness of mFi-5 may only be suitable to those ITF patients without special comorbidities.

In conclusion, mFi-5 is a reliable prognostic predictor in advanced-age ITF patients following surgery. Also, high risk of mortality is associated with a mFi-5 score no less than “three” especially in the much older male ITF patient with anemia.

Abbreviations

mFi-5

Five-Factor Modified Frailty Index

ITF

Intertrochanteric Fractures

ROC

Receiver-Operating Characteristic

FNS

Femoral Neck Fractures

ASA

American Society of Anesthesiologists

CCI

Charlson Comorbidity Index

ECM

Elixhauser Comorbidity Measure

NSQIP

National Surgical Quality Improvement Program

AUC

Area under The Curve

PPV

Positive Predictive Value

NPV

Negative Predictive Value

DCA

Decision Curve Analysis

Authors’ contributions

The study was designed by Yun-fa Yang. Material preparation and study performance were conducted by all authors. The data collection and analysis were performed by Jian-wen Huang and Xiao-sheng Gao. The follow-up was done by Jian-wen Huang and Xiao-sheng Gao. Surgery evaluation was done postoperatively by Yun-fa Yang and Zhong-he Xu. The first draft of the manuscript was written by Jian-wen Huang and revised by Yun-fa Yang. All authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Funding

Partially supported by the fund of Guangzhou Municipal Health Commission (No. 2023 C-TS32), Guangzhou Planned Project of Science and Technology (No. 2024A03J1179) and General Guidance Project of Guangzhou Municipal Health Commission (No. 20251A011006).

Data availability

The datasets generated and analyzed during the current study are not publicly available because other studies coming from the same dataset are currently ongoing, but are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The study was approved by the Ethics Committee of Guangzhou First People’s Hospital (B-2022-065-01). All patients provided written informed consent after the nature and possible consequences of the study had been explained. All methods were carried out in accordance with the Declaration of Helsinki.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.