Skip to main content
NCBI home page
Medicine logoLink to Medicine
. 2023 Dec 15;102(50):e36202. doi: 10.1097/MD.0000000000036202

Serum glutathione peroxidase 4 as a novel biomarker for nontraumatic osteonecrosis of the femoral head: A retrospective case-control study

Qiang Zhao a,b, Jianhong Dong c, Shiying Wang b, Biaofang Wei b,*
PMCID: PMC10727552  PMID: 38115372

Abstract

There are no serum biomarkers available in nontraumatic osteonecrosis of the femoral head in clinical practice. This study aimed to evaluate the clinical value of serum glutathione peroxidase 4 in nontraumatic osteonecrosis of the femoral head. This retrospective study analyzed serum glutathione peroxidase 4 levels and clinical data of 80 patients with nontraumatic osteonecrosis of the femoral head and 80 healthy controls between August 2021 and May 2022. Serum glutathione peroxidase 4 levels were analyzed using an enzyme-linked immunosorbent assay. The Association Research Circulation Osseous classification system determined disease progression. Clinical severity was assessed by Harris hip score and visual analogue scale. Correlations between serum glutathione peroxidase 4 and disease progression as well as clinical severity were evaluated statistically. The diagnostic accuracy of serum glutathione peroxidase 4 in nontraumatic osteonecrosis of the femoral head was determined using receiver operating characteristic analysis. The baseline characteristics of participants between 2 groups were comparable. Patients with nontraumatic osteonecrosis of the femoral head displayed a decreased glutathione peroxidase 4 level compared with healthy controls (11.87 ± 2.76 μU/mL vs 16.54 ± 4.89 μU/mL, P < .01). The levels of glutathione peroxidase 4 were inversely correlated with Association Research Circulation Osseous stage (P < .01) and visual analogue scale scores (P < .01), and positively correlated with Harris score (P < .01). Receiver operating characteristic analyses showed that area under curves of glutathione peroxidase 4 was 0.808 (95% CI 0.721–0.858) and 0.847 (95% CI 0.743–0.951) with regard to diagnosis and collapse prediction in nontraumatic osteonecrosis of the femoral head, respectively. Serum glutathione peroxidase 4 could serve as a novel biomarker for diagnosing nontraumatic osteonecrosis of the femoral head and predicting collapse of the femoral head.

Keywords: biomarker, diagnosis, femur head necrosis, glutathione peroxidase 4, propensity score matching

1. Introduction

Nontraumatic osteonecrosis of the femoral head (NONFH) is a devastating disease characterized by hip pain and dysfunction worldwide.[1] It is estimated that there are 8.12 million NONFH cases aged 15 years and over in China.[2] It mainly affects adults aged 30 to 50 years and usually progresses to femoral head collapse if left untreated.[3] Femoral head collapse is a key turning point in disease progression. In the pre-collapse stage, hip preservation treatment is a possible way to delay or even block the collapse. After the collapse, most patients have to undergo total hip arthroplasty (THA). Given the limited operational life span of hip prostheses, young patients may require revision surgery after THA. Both THA and revision surgery place a heavy economic burden on patients and decrease their quality of life. Therefore, early diagnosis and timely treatment are necessary to prevent or delay the collapse of the femoral head. NONFH is commonly asymptomatic at the initial occurrence, making it challenging to screen with the most conclusive method like magnetic resonance imaging (MRI). However, there are no blood biomarkers available for diagnosing NONFH and predicting collapse of the femoral head.

Cell death is a basic physiological process during the normal development and maintenance of tissue homeostasis in multicellular organisms. Several new forms of cell death have been discovered recently, suggesting that a cell can die in different mechanisms. Ferroptosis was identified as a novel form of programmed cell death characterized by iron-dependent lipid peroxidation and metabolic constraints.[4] Ferroptosis-inducing factors can directly or indirectly affect glutathione peroxidase through different pathways, resulting in a decrease in antioxidant capacity and accumulation of lipid reactive oxygen species (ROS) in cells, ultimately leading to oxidative cell death.[5] As a key regulator of ferroptosis, glutathione peroxidase 4 (GPx4) is one of the enzymes that play an important role in maintaining cellular homeostasis by protecting cells from oxidative stress.[6] When the activity of GPx4 was inhibited, excessive lipid peroxides and ROS were generated, leading to cell damage and eventually ferroptosis. Recently, an experimental study observed a decreased expression of GPx4 protein in the femoral head of rat model of osteonecrosis.[7] This highlights the crucial role of GPx4 in the pathogenesis of NONFH.

Blood GPx4 was considered as a potential biomarker in various diseases related to oxidative stress. For example, serum GPx4 showed good performance in diagnosing gestational diabetes mellitus[8] and diabetic retinopathy.[9] GPx4 was also associated with disease progression in diabetic kidney disease.[10] No study has investigated the relationship between blood GPx4 and patients with NONFH. In the present study, we hypothesized that serum GPx4 was associated with disease severity of NONFH and showed good performance in diagnosing NONFH. Our findings may provide a reference for diagnosis and individualized treatment of NONFH in clinical practice.

2. Methods

2.1. Study design and participants

This retrospective case-control study was performed at Linyi People’s Hospital, China, between August 2021 and May 2022. Clinical data of 141 NONFH patients and 202 healthy subjects, including gender, age, body mass index (BMI) and past history were retrospectively reviewed and analyzed. After the exclusions and propensity score matching (PSM) method, 80 NONFH cases and 80 healthy subjects were well-matched and selected for further analysis. Blood samples of 80 NONFH patients and 80 healthy controls were retrieved from the bioBank in our hospital. Additional tests of GPx4 levels were performed in those blood serum samples. The flow diagram of this study is shown in Figure 1. This study was reviewed and approved by the Ethics Committee with a reference number of YX200342.

Figure 1.

Figure 1.

Open in a new tab

Flow diagram of the present study.

2.2. Inclusion and exclusion criteria

Patients were included if they met the following criteria: (1) diagnosed as NONFH (ICD code: FB81) according to a Guideline for clinical diagnosis and treatment of NONFH,[11] (2) age at diagnosis over 18 years. The exclusion criteria were as follows: (1) history of hip, or femur surgery or trauma, (2) history of malignancy, cardiovascular diseases, and metabolic bone disease, (3) treated with medication affecting bone metabolism, such as zoledronic acid, denosumab, and traditional Chinese Medicine, (4) involvement in other clinical trials within 3 months, (5) pregnant women.

2.3. Case-control matching method

PSM is a reliable statistical method to control selection bias and confounding bias in the non-randomized study.[12,13] We implemented PSM using the R package “MatchIt” with the following covariates, including gender, age, and BMI. The present case-control study was matched at a ratio of 1:1. Ultimately, 80 healthy subjects and 80 NONFH cases were well matched and selected for further analysis.

2.4. Measurement of serum GPx4 levels

Measurement of GPx4 concentrations in serum samples were performed in the Centre Laboratory at Linyi People’s Hospital. We used an enzyme-linked immunosorbent assay as described by the manufacturer (Cusabio, China). The intra-assay coefficient of variation was below 8%. The lowest detection of serum GPx4 concentrations was 7.8 μU/mL.

2.5. Assessment of disease stage

We assessed the disease stage of osteonecrosis using Association Research Circulation Osseous (ARCO) staging classification in plain radiograph and/or MRI.[14] The ARCO stages consist of I to IV stages. In plain radiograph, no evidence can be observed in stage I, while trabecular bone changes can be recognized in stage II, but without changes of subchondral fracture. When subchondral fracture or flattening of the femoral head can be found in plain radiograph, it is stage III. Evidence of hip osteoarthritis can be seen in stage IV in plain radiographs. MRI is usually undertaken for the detection of stage I in clinical practice. The typical pattern of signal change is a “double-line” sign which is presented as double low-intensity bands.

2.6. Definition of clinical severity

The 10-point visual analog scale (VAS) was used to assess pain intensity in the hip joint, where 0 represents no pain and 10 score indicates extremely pain. The Harris hip score is a popular tool for the measurement of hip functional capacity. It comprises a total score of 100 points, a higher score indicates better functional capacity.[15]

2.7. Assessment of size and location of osteonecrotic lesions

The Japanese Investigation Committee classification establishes 3 types for describing the location and stage of osteonecrotic lesions in the femoral head.[16] Type A indicates that such lesions occupy up to medial one-third of the surface; type B, up to medial two-thirds; or type C, more than medial two-thirds. Moreover, type C is subdivided into 2 types: type C1, the osteonecrotic lesions do not exceed the acetabular edge; and type C2, the osteonecrotic lesions exceed the acetabular edge. In this study, the measurement of the area of osteonecrotic lesions was using the modified Kerboul method.[17]

2.8. Post hoc statistical power calculation

Use the Power and Sample Size Calculators (http://powerandsamplesize.com) to calculate statistical power (1 − β) to obtain data on different average GPx4 levels, standard errors, and the number of enrolled patients in each group.[18] When >0.8, the statistical power is considered strong.

2.9. Statistical analysis

Statistical analysis was performed using SPSS Statistics software version 20.0. Continuous data were expressed as means ± standard deviations and were compared using unpaired Student t test. Categorical data were expressed as frequencies and percentages and were analyzed with the 2-tailed chi-square test. Pearson or Spearman test was used to evaluate correlations between GPx4 and clinical severity as well as radiographic progression, where appropriate. One-way ANOVA analysis was performed to test GPx4 levels among different ARCO stages and different etiologies. Receiver operating characteristic (ROC) curves analysis for serum GPx4 was employed to determine the diagnostic and prognostic values for NONFH and disease progression, respectively. P-value < .05 was accepted as statistically significant.

3. Results

3.1. Characteristics of participants

The demographic characteristics of all participants are listed in Table 1. A total of 80 NONFH cases and 80 controls were analyzed in this study. After PSM, a balance of baseline variables between the above 2 groups was achieved. The NONFH group consists of 14 women and 66 men, while the control group consists of 20 women and 60 men. The mean age of NONFH patients was (49.58 ± 10.83) years, while in controls it was (48.38 ± 14.44) years. The mean BMI was (24.84 ± 3.63) kg/m2 in NONFH patients and (23.93 ± 2.34) kg/m2 in controls, respectively. Gender distribution was approximately the same in both groups (P = .246). No significant differences were detected regarding age and BMI between NONFH patients and controls. Of 80 NONFH patients, 31 had a history of steroid therapy, while 17 had a history of alcohol abuse. Among NONFH patients, 22 (27.50%) presented with ARCO II, 21 (26.25%) suffered from ARCO III, and 37 (46.25%) were ARCO IV stage. After calculation, the statistical power was 0.82, suggesting that the sample size of 160 was sufficient to obtain the conclusion.

Table 1.

Demographic data of NONFH patients and controls.

NONFH patients (n = 80) Controls (n = 80) P value
Gender (male/female) 64/16 60/20 .246
Age (yr) 49.58 ± 10.83 48.38 ± 14.44 .832
BMI (kg/m2) 24.84 ± 3.63 23.93 ± 2.34 .557
ARCO stage (II/III/IV) 22/21/37 / /
VAS 5.63 ± 1.31 / /
Harris score 60.25 ± 16.34 / /
Etiology (steroid/alcohol/idiopathic) 31/17/32 / /
Side (unilateral/ bilateral) 37/43 / /
Open in a new tab

ARCO = Association Research Circulation Osseous, BMI = body mass index, NONFH = nontraumatic osteonecrosis of the femoral head, VAS = visual analogue scale.

3.2. Comparisons of serum GPx4 levels

As shown in Figure 2A, the mean serum GPx4 levels were significantly lower in NONFH patients (11.87 ± 2.76 μU/mL) than those in controls (16.54 ± 4.89 μU/mL) (P < .01). Serum GPx4 levels were all markedly lower in ARCO stage IV and III than stage II in the subgroups of steroid-induced, alcohol-induced or idiopathic ONFH. (P < .05) (Fig. 2B). This indicates the potential role of GPx4 in the collapse of the femoral head. As to the size of osteonecrotic lesions, serum GPx4 levels were lower in modified Kerboul angle > 200° group (11.81 ± 2.88 μU/mL) than ≤ 200° group (15.87 ± 1.92 μU/mL) (P = .042) (Fig. 2C). As to the location of osteonecrotic lesions, serum GPx4 levels were lower in Type C2 group (16.67 ± 2.96 μU/mL) than Type C1 (14.81 ± 2.80 μU/mL), Type B (12.21 ± 2.76 μU/mL) and Type A group (10.59 ± 2.30 μU/mL) (P < .05) (Fig. 2D).

Figure 2.

Figure 2.

Open in a new tab

(A) Comparison of serum GPx4 levels between NONFH cases and healthy controls (**P < .01). (B) Comparison of serum GPx4 levels between modified Kerboul angle > 200° group and ≤200° group (*P < .05). (C) Comparison of serum GPx4 levels between ARCO III + IV stages and ARCO II stage in the subgroups of steroid-induced, alcohol-induced, or idiopathic ONFH (*P < .05). (D) Comparisons of serum GPx4 levels among different JIC types (**P < .05). ARCO = Association Research Circulation Osseous, GPx4 = glutathione peroxidase 4, JIC = Japanese Investigation Committee, NONFH = nontraumatic osteonecrosis of the femoral head.

3.3. Correlations of serum GPx4 and ARCO stage as well as other indices

We investigated the correlation of serum GPx4 with ARCO stages and other indices including Harris score and VAS to illustrate whether serum GPx4 is related to disease progression and clinical severity. As shown in Figure 3, serum GPx4 levels were positively correlated with Harris score (R = 0.538, P < .01) and negatively correlated with ARCO stages (r = −0.593, P < .01) as well as VAS scores (r = −0.331, P < .01).

Figure 3.

Figure 3.

Open in a new tab

(A) Correlation of serum GPx4 levels with ARCO stages. (B) Correlation of serum GPx4 levels with Harris score. (C) Correlation of serum GPx4 levels with VAS scores. ARCO = Association Research Circulation Osseous, GPx4 = glutathione peroxidase 4, VAS = visual analogue scale.

3.4. Diagnostic performance of serum GPx4

We further performed ROC curve analysis to explore the diagnostic and prognostic values of serum GPx4 for NONFH. As shown in Figure 4, the results reveal ROC of serum GPx4 between NONFH patients and healthy controls with an area under the curve (AUC) of 0.808. At a cutoff of 15.635 μU/mL, the sensitivity was 55.50%, the specificity was 87.50%, and the accuracy was 71.25%. The AUC was 0.847 (95%CI 0.743–0.951) for ROC of serum GPx4 with regard to pre-collapse vs post-collapse. At a cutoff of 13.369 μU/mL, the sensitivity was 77.27%, the specificity was 86.21%, and the accuracy was 83.75% for predicting femoral head collapse. These findings indicate that serum GPx4 may serve as a potential biomarker for diagnosis and collapse prediction of NONFH.

Figure 4.

Figure 4.

Open in a new tab

(A) ROC curve analysis of serum GPx4 levels with regard to control versus NONFH. (B) ROC curve analysis of serum GPx4 levels with regard to pre-collapse versus post-collapse. GPx4 = glutathione peroxidase 4, NONFH = nontraumatic osteonecrosis of the femoral head, ROC = receiver operating characteristic.

4. Discussion

This study was the first to evaluate diagnostic and prognostic values of serum GPx4 in patients with NONFH. The results showed that serum GPx4 levels were significantly lower in patients with NONFH. Serum GPx4 levels were positively correlated with Harris hip score and negatively correlated with ARCO stages and VAS scores. Moreover, serum GPx4 showed a high diagnostic and prognostic accuracy in patients with NONFH.

GPx4 is a 22kDa selenoprotein belonging to GPx family which plays an important role in protecting cells from oxidative stress.[19] Oxidative stress is suggested to be one of the possible pathogeneses of osteonecrosis by contributing to the imbalance of bone homeostasis.[20] Bone is a dynamic tissue that undergoes constant remodeling, which is regulated by a balance between bone-forming osteoblasts and bone-resorbing osteoclasts.[21] This process is tightly regulated by various signaling pathways. ROS are highly reactive molecules that can cause damage to cellular components, including DNA, proteins, and lipids. higher levels of ROS can lead to oxidative stress. Recent research has shown the critical role of ferroptosis. Ferroptosis is an iron-dependent, non-apoptotic form of cell death characterized by the accumulation of lipid peroxides and ROS.[22] GPx4 has been recognized as a central repressor of ferroptosis.[23] The above literature implied the crucial role of GPx4 in the pathogenesis of NONFH.

In the present study, we found that NONFH patients displayed a decreased GPx4 in serum compared with healthy controls. Circulating levels of GPx4 have been studied in various diseases related to oxidative stress. In women with gestational diabetes mellitus, serum GPx4 levels were decreased compared with normal subjects.[8] The decreased concentrations of serum GPx4 were also observed in patients with diabetic retinopathy compared with healthy subjects.[9] The acute coronary syndrome patients had lower plasma GPx4 levels than controls.[24] The reduction of GPX4 plays a central regulatory role in ferroptosis. This confirmed that the pathogenesis of NONFH is related to ferroptosis. Our findings are consistent with the above studies, suggesting the reliability of our results.

Our results also revealed that serum GPx4 levels were markedly lower in ARCO stage IV and stage III than stage II, implicating the reduction of GPx4 may be involved in the disease progression of NONFH. In addition, serum GPx4 levels were negatively correlated with ARCO stages. It is well known that the size and location of osteonecrosis are important factors to predict collapse of the femoral head. Our results showed that the serum GPx4 was lower in both bigger size and the location where are frequent to collapse. The predictive accuracy may be improved by a nomogram including size, location of osteonecrosis and GPx4 levels in future study.[25,26]

Finally, ROC curve analysis was performed to test the efficacy of serum GPx4 in the diagnosis and progression prediction of NONFH. The AUC of ROC was 0.808 for serum GPx4 in the diagnosis of NONFH and 0.847 in the prediction of collapse of the femoral head. Our results regarding AUC are similar to a previously published study. Wang et al[27] reported an AUC of 0.806 for serum nicotinamide phosphoribosyltransferase in diagnosing NONFH. The ideal AUC would be close to 1, with a minimum of 0.7 to indicate clinical utility as a biomarker.[28] Taken together, monitoring GPx4 levels can provide valuable information for detection and collapse prediction of NONFH.

There are still several limitations in this study. First, this was a retrospective study including a limited number of subjects, thus prospective studies with a larger sample size should be further conducted. Second, the use of serum GPx4 as a diagnostic and prognostic biomarker must be validated in diverse ethnic populations, because the clinical specimens analyzed in our study were solely from patients of Chinese origin. Nonetheless, this is the first study to explore the diagnostic and prognostic values in NONFH.

5. Conclusion

Our results indicated that serum GPx4 may serve as a potential biomarker in the diagnosis of NONFH and prediction of collapse of the femoral head.

Author contributions

Conceptualization: Qiang Zhao.

Data curation: Qiang Zhao, Shiying Wang.

Formal analysis: Jianhong Dong, Shiying Wang.

Investigation: Jianhong Dong.

Methodology: Qiang Zhao.

Project administration: Biaofang Wei.

Software: Jianhong Dong.

Supervision: Biaofang Wei.

Validation: Jianhong Dong.

Visualization: Qiang Zhao, Shiying Wang.

Writing - original draft: Qiang Zhao.

Writing – review & editing: Biaofang Wei.

Abbreviations:

ARCO
Association Research Circulation Osseous
AUC
area under curve
BMI
body mass index
GPx4
glutathione peroxidase 4
MRI
magnetic resonance imaging
NONFH
nontraumatic osteonecrosis of the femoral head
PSM
propensity score matching
ROC
receiver operating characteristic
ROS
reactive oxygen species
THA
total hip arthroplasty
VAS
visual analogue scale

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

The authors have no funding and conflicts of interest to disclose.

How to cite this article: Zhao Q, Dong J, Wang S, Wei B. Serum glutathione peroxidase 4 as a novel biomarker for nontraumatic osteonecrosis of the femoral head: A retrospective case-control study. Medicine 2023;102:50(e36202).

Contributor Information

Qiang Zhao, Email: zhuoer889@163.com.

Jianhong Dong, Email: jianhongdong@126.com.

Shiying Wang, Email: newshiying@126.com.

References

  • [1].Mont MA, Salem HS, Piuzzi NS, et al. Nontraumatic osteonecrosis of the femoral head: where do we stand today?: a 5-year update. J Bone Joint Surg Am. 2020;102:1084–99. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [2].Zhao DW, Yu M, Hu K, et al. Prevalence of nontraumatic osteonecrosis of the femoral head and its associated risk factors in the Chinese population: results from a nationally representative survey. Chin Med J (Engl). 2015;128:2843–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [3].Cui Q, Jo WL, Koo KH, et al. ARCO consensus on the pathogenesis of non-traumatic osteonecrosis of the femoral head. J Korean Med Sci. 2021;36:e65. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [4].Dixon SJ, Lemberg KM, Lamprecht MR, et al. Ferroptosis: an iron-dependent form of nonapoptotic cell death. Cell. 2012;149:1060–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [5].Li J, Cao F, Yin HL, et al. Ferroptosis: past, present and future. Cell Death Dis. 2020;11:88. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [6].Xu S, Wu B, Zhong B, et al. Naringenin alleviates myocardial ischemia/reperfusion injury by regulating the nuclear factor-erythroid factor 2-related factor 2 (Nrf2)/System xc-/glutathione peroxidase 4 (GPX4) axis to inhibit ferroptosis. Bioengineered. 2021;12:10924–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [7].Fang L, Zhang G, Wu Y, et al. SIRT6 prevents glucocorticoid-induced osteonecrosis of the femoral head in rats. Oxid Med Cell Longev. 2022;2022:6360133. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [8].Zhang B, Zhang T, Hu S, et al. Association of serum lipid peroxidation and glutathione peroxidase 4 levels with clinical outcomes and metabolic abnormalities among patients with gestational diabetes mellitus: a case-control study in the Chinese population. Front Biosci (Landmark Ed). 2022;27:68. [DOI] [PubMed] [Google Scholar]
  • [9].Mu L, Wang D, Dong Z, et al. Abnormal levels of serum ferroptosis-related biomarkers in diabetic retinopathy. J Ophthalmol. 2022;2022:3353740. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [10].Wang YH, Chang DY, Zhao MH, et al. Glutathione peroxidase 4 is a predictor of diabetic kidney disease progression in type 2 diabetes mellitus. Oxid Med Cell Longev. 2022;2022:2948248. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [11].Zhao D, Zhang F, Wang B, et al. Guidelines for clinical diagnosis and treatment of osteonecrosis of the femoral head in adults (2019 version). J Orthop Translat. 2020;21:100–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [12].Kane LT, Fang T, Galetta MS, et al. Propensity score matching: a statistical method. Clin Spine Surg. 2020;33:120–2. [DOI] [PubMed] [Google Scholar]
  • [13].Jupiter DC. Propensity score matching: retrospective randomization? J Foot Ankle Surg. 2017;56:417–20. [DOI] [PubMed] [Google Scholar]
  • [14].Yoon BH, Mont MA, Koo KH, et al. The 2019 revised version of association research circulation osseous staging system of osteonecrosis of the femoral head. J Arthroplasty. 2020;35:933–40. [DOI] [PubMed] [Google Scholar]
  • [15].Lau BC, Scribani M, Lassiter T, et al. Correlation of single assessment numerical evaluation score for sport and activities of daily living to modified harris hip score and hip outcome score in patients undergoing arthroscopic hip surgery. Am J Sports Med. 2019;47:2646–50. [DOI] [PubMed] [Google Scholar]
  • [16].Sugano N, Atsumi T, Ohzono K, et al. 2001 revised criteria for diagnosis, classification, and staging of idiopathic osteonecrosis of the femoral head. J Orthop Sci. 2002;7:601–5. [DOI] [PubMed] [Google Scholar]
  • [17].Ha YC, Jung WH, Kim JR, et al. Prediction of collapse in femoral head osteonecrosis: a modified Kerboul method with use of magnetic resonance images. J Bone Joint Surg Am. 2006;88(Suppl 3):35–40. [DOI] [PubMed] [Google Scholar]
  • [18].Ye T, Yi Y. Sample size calculations in clinical research, third edition, by Shein-Chung Chow, Jun Shao, Hansheng Wang, and Yuliya Lokhnygina. Stat Theory Relat Fields. 2017;1:265–6. [Google Scholar]
  • [19].Weaver K, Skouta R. The selenoprotein glutathione peroxidase 4: from molecular mechanisms to novel therapeutic opportunities. Biomedicines. 2022;10:891. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [20].Chen K, Liu Y, He J, et al. Steroid-induced osteonecrosis of the femoral head reveals enhanced reactive oxygen species and hyperactive osteoclasts. Int J Biol Sci. 2020;16:1888–900. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [21].Kim JM, Lin C, Stavre Z, et al. Osteoblast-osteoclast communication and bone homeostasis. Cells. 2020;9:2073. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [22].Rochette L, Dogon G, Rigal E, et al. Lipid peroxidation and iron metabolism: two corner stones in the homeostasis control of ferroptosis. Int J Mol Sci. 2022;24:449. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [23].Seibt TM, Proneth B, Conrad M. Role of GPX4 in ferroptosis and its pharmacological implication. Free Radic Biol Med. 2019;133:144–52. [DOI] [PubMed] [Google Scholar]
  • [24].Li MN, Bao BW, Si-Yu D, et al. Correlation between plasma glutathione peroxidase 4 and N-acetylneuraminic acid levels with clinical risk stratification and prognosis of patients with acute coronary syndrome. Saudi Med J. 2022;43:1103–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [25].Li W, Dong S, Lin Y, et al. A tool for predicting overall survival in patients with Ewing sarcoma: a multicenter retrospective study. BMC Cancer. 2022;22:914. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [26].Li W, Jin G, Wu H, et al. Interpretable clinical visualization model for prediction of prognosis in osteosarcoma: a large cohort data study. Front Oncol. 2022;12:945362. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [27].Wang S, Zhan H, Xu L, et al. Serum nicotinamide phosphoribosyltransferase as a novel biomarker for non-traumatic osteonecrosis of the femoral head. J Orthop Surg Res. 2022;17:514. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [28].Mandrekar JN. Receiver operating characteristic curve in diagnostic test assessment. J Thorac Oncol. 2010;5:1315–6. [DOI] [PubMed] [Google Scholar]

Articles from Medicine are provided here courtesy of Wolters Kluwer Health

RESOURCES