Abstract
Objective
The aim of this study was to investigate the clinical significance of serum homocysteine (Hcy) as a biomarker for early diagnosis of diabetic nephropathy (DN) in type 2 diabetes mellitus (T2DM) patients.
Methods
Fifty-five T2DM patients with DN and 51 T2DM patients without DN were prospectively recruited from January 2016 to May 2020 in our hospital. The serum Hcy was tested by electrochemiluminescence assay in DN and T2DM groups and compared. The diagnostic efficacy of serum Hcy as a biomarker for early diagnosis of DN was evaluated by calculating the diagnostic sensitivity, specificity and area under the ROC curve (AUC).
Results
The serum levels of Hcy were 15.49 ± 5.40 and 9.23 ± 3.15 μmol/L for DN and T2DM patients, respectively, with statistical difference (t = 7.21, P < 0.001). In the DN group, the serum Hcy levels for patients with hyperfiltration, intermittent proteinuria, microalbuminuria, macroalbuminuria and uremic were 10.99 ± 2.57, 13.90 ± 2.86, 15.38 ± 4.77, 18.98 ± 4.36 and 23.31 ± 5.22 μmol/L, respectively, which indicated that serum Hcy levels in DN were higher than those of T2DM patients and correlated with patient’s renal damage. Using the serum Hcy level as the reference, the diagnostic sensitivity, specificity and AUC were 84.31 (71.41–92.98)%, 74.55 (61.00–85.33)% and 0.85 (0.78–0.92)%, respectively, with the cutoff value of 12.08 between DN and T2DM. The serum Hcy also had relatively good differential diagnostic efficacy between different DN stages with high sensitivity, specificity and AUC.
Conclusion
Serum Hcy was obviously elevated in DN compared to T2MD and correlated with the renal damage severity, which can be applied as a potential serological marker for early diagnosis of DN.
1 Introduction
Diabetes mellitus (DM) is one of the most frequent metabolic diseases and the prevalence of DM is rising [1,2]. Diabetic nephropathy (DN), mostly characterized by proteinuria, is one of the most common microvascular complications of type 2 diabetes mellitus (T2DM) [3,4]. For diagnosis of DN, serum creatinine and urinary albumin were generally measured [5–7]. Hypertension, edema, proteinuria, systemic microvascular disease and even renal failure were often identified in the uremia stage of DN patients [8]. Therefore, early identification of DN and proper intervention had important clinical significance in the aspects of delaying the development of renal failure [9–11].
Hcy is a kind of sulfur-containing amino acid, which is an intermediate product in the process of methionine metabolism [12]. It metabolizes through methylation and trans-sulfurization, maintaining the two major abilities of human body: methylation and antioxidation. Folic acid and vitamin B12 play an important role in the Hcy metabolic cycle [13–15]. Hcy was first isolated from bladder stones in 1931 [16]. Subsequent studies found that the increased serum Hcy level was an independent risk factor for cardiovascular and cerebrovascular diseases [17,18]. In recent years, it has been reported that serum Hcy level was correlated with degree of renal function damage [19]. With the development of renal damage, serum Hcy was also increased correspondingly [20].
Therefore, we performed a prospective study in order to investigate the clinical significance of serum Hcy as a biomarker for early diagnosis of DN in T2DM patients. In this study, we selected T2MD patients with DN with different clinical stages and T2MD cases without DN as the study objects to explore whether the serum Hcy level is different in different stages of DN and T2DM.
2 Materials and methods
2.1 Patients
Fifty-five T2DM patients with DN and 51 T2DM patients without DN were prospectively recruited from January 2016 to May 2020 in the Sixth Affiliated Hospital of Wenzhou Medical University. The inclusion criteria were as follows: (1) the patients were older than 18 years; (2) all the patients were diagnosed with T2MD according to the WHO diagnostic criteria for DM in 1999; (3) the patients also met the diagnostic criteria for DN designated by the American Diabetes Association in 1977; (4) the patients did not take any drugs affecting renal function within 4 weeks before enrollment; and (5) no folic acid and VB12 intake within 3 months. The exclusion criteria were as follows: (1) patients with type 1 diabetes mellitus; (2) patients with malignant tumor; (3) patients with other primary kidney diseases such as IgA nephropathy and glomerulonephritis; (4) cases of acute infection; (5) patients with acute cerebral tissue or myocardial infarction; and (6) pregnant or lactating women.
In the DN group, there were 55 cases with 28 males and 27 cases, aged 41–78 years (58.6 ± 12.8). Of the 55 patients, 14 cases had hyperfiltration, 16 cases had intermittent proteinuria, 8 cases had microalbuminuria, 10 cases had massive proteinuria and 7 cases had uremia. There were 25 males and 26 females with a mean age of (54.3 ± 14.1) years (range, 33–79 years) in the T2MD group (Table 1). The gender distribution and mean age were not statistically different between DN and T2MD groups.
Basic data of the DN and T2DM groups
| Group | DN (n = 55) | T2DM (n = 51) |
|---|---|---|
| Gender ( n ) | ||
| Male | 28 | 25 |
| Female | 27 | 26 |
| Age (years) | ||
| Range | 41–78 | 33–79 |
| Mean ± SD | 58.6 ± 12.8 | 54.3 ± 14.1 |
| Hyperfiltration (n) | 14 | 0 |
| Intermittent proteinuria (n) | 16 | 0 |
| Microalbuminuria (n) | 8 | 0 |
| Massive proteinuria (n) | 10 | 0 |
| Uremia (n) | 7 | 0 |
-
Ethical approval: The research related to human use has been complied with all the relevant national regulations, institutional policies and in accordance with the tenets of the Helsinki Declaration and has been approved by ethical committee of the Sixth Affiliated Hospital of Wenzhou Medical University.
-
Informed consent: Informed consent has been obtained from all individuals included in this study.
2.2 Serum Hcy examination
After fasting for 12 h, 5 mL of the peripheral venous blood was collected from the included cases. The peripheral venous blood was centrifuged for 10 min under 3,000 rpm and then the upper serum was collected for Hcy detection through electrochemiluminescence assay.
2.3 Statistical analysis
All data were analyzed through the statistical software STATA11.0 (http://www.stata.com). Serum Hcy was expressed by
3 Results
3.1 Serum Hcy in DN was higher than that of T2DM patients and correlated with patient’s renal damage
The serum levels of Hcy were 15.49 ± 5.40 and 9.23 ± 3.15 μmol/L for DN and T2DM patients, respectively, with statistical difference (t = 7.21, P < 0.001) (Figure 1a). In the DN group, the serum Hcy levels for patients with hyperfiltration, intermittent proteinuria, microalbuminuria, macroalbuminuria and uremic were 10.99 ± 2.57, 13.90 ± 2.86, 15.38 ± 4.77, 18.98 ± 4.36 and 23.31 ± 5.22 μmol/L, respectively, which indicated that serum Hcy in DN was higher than that of T2DM patients and correlated with patient’s renal damage (Figure 1b).
Scatter plot of serum Hcy in DN and T2DM patients and correlated with patient’s renal damage. (a) Serum Hcy level between DN and T2DM patients. (b) Serum level of Hcy of DN patients with different renal damage.
3.2 Serum Hcy had high efficacy for early diagnosis of DN in T2DM patients
Using serum Hcy as the reference, the diagnostic sensitivity, specificity and AUC were 84.31 (71.41–92.98)%, 74.55 (61.00–85.33)% and 0.85 (0.78–0.92)%, respectively, with the cutoff value of 12.08 between DN and T2DM groups (Figure 2a). The serum Hcy also had relatively good differential diagnostic efficacy between different DN stages with high sensitivity, specificity and AUC (Table 2 and Figure 2b–k).
ROC curve of serum Hcy as a biomarker for early diagnosis of DN in T2DM patients. (a) DN vs T2DM; (b) hyperfiltration vs intermittent proteinuria; (c) hyperfiltration vs microalbuminuria; (d) hyperfiltration vs macroalbuminuria; (e) hyperfiltration vs uremic; (f) intermittent proteinuria vs microalbuminuria; (g) intermittent proteinuria vs macroalbuminuria; (h) intermittent proteinuria vs uremic; (i) microalbuminuria vs macroalbuminuria; (j) microalbuminuria vs uremic; and (k) macroalbuminuria vs uremic.
Diagnostic efficacy of serum Hcy as a biomarker for early identification of DN in T2DM patients
| Renal damage | Sensitivity (%) | Specificity (%) | AUC | Cutoff |
|---|---|---|---|---|
| DN vs T2DM | 84.31 (71.41–92.98) | 74.55 (61.00–85.33) | 0.85 (0.78–0.92) | 12.08 |
| Hyperfiltration vs intermittent proteinuria | 81.25 (54.35–95.95) | 71.43 (41.90–91.61) | 0.79 (0.63–0.95) | 12.50 |
| Hyperfiltration vs microalbuminuria | 75.00 (34.91–96.81) | 78.57 (49.20–95.34) | 0.79 (0.57–1.00) | 12.93 |
| Hyperfiltration vs microalbuminuria | 90.00 (55.50–99.75) | 85.71 (57.19–98.22) | 0.94 (0.84–1.00) | 13.56 |
| Hyperfiltration vs uremic | 100 (59.04–100) | 85.71 (57.19–98.22) | 0.98 (0.92–1.00) | 13.81 |
| Intermittent proteinuria vs microalbuminuria | 62.50 (24.49–91.48) | 62.50 (35.43–84.80) | 0.62 (0.35–0.90) | 14.68 |
| Intermittent proteinuria vs macroalbuminuria | 80.00 (44.39–97.48) | 81.25 (54.35–95.95) | 0.84 (0.66–1.00) | 16.66 |
| Intermittent proteinuria vs uremic | 85.71 (42.13–99.64) | 93.75 (69.77–99.84) | 0.95 (0.84–1.00) | 18.05 |
| Microalbuminuria vs macroalbuminuria | 80.00 (44.39–79.48) | 62.50 (24.49–91.48) | 0.73 (0.49–0.96) | 16.93 |
| Microalbuminuria vs uremic | 85.71 (42.13–99.64) | 87.50 (47.35–99.68) | 0.89 (0.71–1.00) | 19.43 |
| Macroalbuminuria vs uremic | 71.43 (29.04–96.33) | 80.00 (44.39–97.48) | 0.76 (0.51–1.00) | 22.59 |
4 Discussion
In the present work, we found that the serum level of Hcy was obviously elevated in DN cases compared to T2DM patients and correlated with patient’s renal damage severity. Serum Hcy can also be used as a serological marker for early diagnosis of DN with relatively high sensitivity and specificity. Wang et al. evaluated the association between plasma Hcy and progression of early nephropathy in T2DM patients and found that plasma Hcy was significantly associated with eGFR decline after controlling for other progression parameters. The results demonstrated that plasma Hcy was an independent risk factor as well as an early predictor for DN progression in T2DM patients [21]. Deebukkhum et al. [22] also found that Hcy was correlated with nephropathy development and could be used as a potential biomarker for early diagnosis of DN. All of the aforementioned relevant studies and our present work demonstrated that serum Hcy may be a potential biomarker for early diagnosis of DN. Furthermore, compared with the previously relevant study, our work also presents the diagnostic sensitivity, specificity and AUC for early diagnosis of DN, which was more useful for clinical practice. DN is a common chronic microvascular complication in T2DM patients [23]. The main cause of DN is the kidney damage caused by persistent hyperglycemia, which can injure the whole kidney, including glomerulus, renal interstitium and renal vessels. In addition, if the T2DM cases also smoke, have hypertension, hyperlipidemia and other risk factors, the patients would be more likely to develop DN [24–26]. In clinical practice, there are several parameters, including serum creatinine, urea nitrogen, serum cystatin C and serum Hcy, which can be used as reference for clinical diagnosis of DN.
Although serum creatinine and urea nitrogen levels can reflect the degree of renal damage in patients with DN, in clinical application, it is found that serum creatinine level was easy to be affected by body mass index, drug influence, metabolic difference, protein intake and other factors in patients with DN [27,28]. Usually, the level of serum creatinine will increase when the renal function damage is more than 50% [29]. Urea nitrogen is also easily affected by heart function and other factors, and the sensitivity of early diagnosis of DN is relatively low [30].
Hcy is the intermediate product of cysteine and methionine [31,32]. Relevant studies have shown that serum Hcy had positive correlation with proteinuria in DN patients, which can reflect the degree of renal function damage in DN patients [33]. At the same time, compared with serum creatinine, urea nitrogen and other indicators, serum Hcy is not easily affected by the patient’s age, gender, inflammation and other factors. Xu et al. [19] evaluated serum Hcy and cystatin C as biomarkers for progression of DN. They found that serum levels of Hcy and cys-C in DN patients were elevated compared to those of simple T2DM cases, making them potential biomarkers for early diagnosis of DN in T2DM patients. In our present work, we further divided the DN cases into five subgroups: hyperfiltration, intermittent proteinuria, microalbuminuria, massive proteinuria and uremia and evaluated the differential diagnosis ability of serum Hcy for the four subgroup cases. We found that serum Hcy was obviously elevated in DN compared to T2MD and correlated with the renal damage severity, which can be applied as a potential serological marker for early diagnosis of DN. However, many factors may affect Hcy levels in DM such as 5-methylenetetrahydrofolate reductase, cystathionine-β-synthase, insulin, fasting, non-fasting states and so on [34]. Serum Hcy is also easily affected by folate [35] and VB12 levels in vivo. Therefore, if the patients take folic acid and VB12 during Hcy detection [36], Hcy cannot reflect the renal function damage of DN patients.
In conclusion, Hcy was increased in serum of DN cases and correlated with the severity of renal damage in T2DM cases. Combined detection of serum Hcy and other parameters such as serum creatinine and urinary albumin may improve the early DN diagnosis efficacy and patient prognosis. However, the work also had several limitations. First, the sample size is relatively small with 55 DN and 51 T2DM cases. Second, all of the cases were from a single medical center, and the patients’ selection bias seems to be ineluctable. Third, due to short-term follow-up, the correlation between serum Hcy and DN patient prognosis was not investigated.
-
Conflict of interest: The authors state no conflict of interest.
-
Data availability statement: The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.
References
1 Li H, Lu W, Wang A, Jiang H, Lyu J. Changing epidemiology of chronic kidney disease as a result of type 2 diabetes mellitus from 1990 to 2017: estimates from global burden of disease 2017. J Diabetes Investig. 2021 Mar;12(3):346–56.10.1111/jdi.13355Search in Google Scholar PubMed PubMed Central
2 Glovaci D, Fan W, Wong ND. Epidemiology of diabetes mellitus and cardiovascular disease. Curr Cardiol Rep. 2019;21:21.10.1007/s11886-019-1107-ySearch in Google Scholar PubMed
3 Zhang XX, Kong J, Yun K. Prevalence of diabetic nephropathy among patients with type 2 diabetes mellitus in China: a meta-analysis of observational studies. J Diabetes Res. 2020 Feb;2020:2315607.10.1155/2020/2315607Search in Google Scholar PubMed PubMed Central
4 Reutens AT, Atkins RC. Epidemiology of diabetic nephropathy. Contrib Nephrol. 2011;170:1–7.10.1159/000324934Search in Google Scholar PubMed
5 Warram JH, Gearin G, Laffel L, Krolewski AS. Effect of duration of type I diabetes on the prevalence of stages of diabetic nephropathy defined by urinary albumin/creatinine ratio. J Am Soc Nephrol. 1996 Jun;7(6):930–7.10.1681/ASN.V76930Search in Google Scholar PubMed
6 Gross JL, Zelmanovitz T, Oliveira J, de Azevedo MJ. Screening for diabetic nephropathy: is measurement of urinary albumin-to-creatinine ratio worthwhile? Diabetes Care. 1999 Sep;22(9):1599–600.10.2337/diacare.22.9.1599Search in Google Scholar PubMed
7 Pathania M, Rathaur VK, Yadav N, Jayara A, Chaturvedi A. Quantitative micro-albuminuria assessment from ‘random voided urinary albumin: creatinine ratio’ versus ’24 hours urinary albumin concentration’ for screening of diabetic nephropathy. J Clin Diagn Res. 2013 Dec;7(12):2828–31.10.7860/JCDR/2013/6589.3768Search in Google Scholar PubMed PubMed Central
8 Selby NM, Taal MW. An updated overview of diabetic nephropathy: diagnosis, prognosis, treatment goals and latest guidelines. Diabetes Obes Metab. 2020 Apr;22(S1 Suppl 1):3–15.10.1111/dom.14007Search in Google Scholar PubMed
9 Zürbig P, Jerums G, Hovind P, Macisaac RJ, Mischak H, Nielsen SE, et al. Urinary proteomics for early diagnosis in diabetic nephropathy. Diabetes. 2012 Dec;61(12):3304–13.10.2337/db12-0348Search in Google Scholar PubMed PubMed Central
10 Huang M, Liang Q, Li P, Xia J, Wang Y, Hu P, et al. Biomarkers for early diagnosis of type 2 diabetic nephropathy: a study based on an integrated biomarker system. Mol Biosyst. 2013 Aug;9(8):2134–41.10.1039/c3mb25543cSearch in Google Scholar PubMed
11 Pasko N, Toti F, Strakosha A, Thengjilli E, Shehu A, Dedej T, et al. Prevalence of microalbuminuria and risk factor analysis in type 2 diabetes patients in Albania: the need for accurate and early diagnosis of diabetic nephropathy. Hippokratia. 2013 Oct;17(4):337–41.Search in Google Scholar
12 Olszewski AJ, McCully KS. Homocysteine metabolism and the oxidative modification of proteins and lipids. Free Radic Biol Med. 1993 Jun;14(6):683–93.10.1016/0891-5849(93)90151-JSearch in Google Scholar
13 Özkale Y, Erol İ, Kılıçarslan B, Özkale M, Saygı S, Sarıtürk Ç, et al. Serum vitamin B12, folic acid, and homocysteine levels in children with febrile seizure. Turk J Pediatr. 2015 Jul-Aug;57(4):345–52.Search in Google Scholar
14 Narang M, Singh M, Dange S. Serum homocysteine, vitamin B12 and folic acid levels in patients with metabolic syndrome. J Assoc Phys India. 2016 Jul;64(7):22–6.Search in Google Scholar
15 Esnafoğlu E, Yaman E. Vitamin B12, folic acid, homocysteine and vitamin D levels in children and adolescents with obsessive compulsive disorder. Psychiatry Res. 2017 Aug;254:232–7.10.1016/j.psychres.2017.04.032Search in Google Scholar PubMed
16 Finkelstein JD. Homocysteine: a history in progress. Nutr Rev. 2000 Jul;58(7):193–204.10.1111/j.1753-4887.2000.tb01862.xSearch in Google Scholar PubMed
17 Foscolou A, Rallidis LS, Tsirebolos G, Critselis E, Katsimardos A, Drosatos A, et al. The association between homocysteine levels, Mediterranean diet and cardiovascular disease: a case-control study. Int J Food Sci Nutr. 2019 Aug;70(5):603–11.10.1080/09637486.2018.1547688Search in Google Scholar PubMed
18 Vadini F, Santilli F, Casalini G, Dell’Isola M, Iuliani O, D’Ardes D, et al. Homocysteine and education but not lipoprotein (a) predict estimated 10-year risk of cardiovascular disease in blood donors: a community based cross-sectional study. BMC Cardiovasc Disord. 2019 Jul;19(1):177.10.1186/s12872-019-1157-5Search in Google Scholar PubMed PubMed Central
19 Xu W, Tang S, Xiang M, Peng J. Serum homocysteine, cystatin C as biomarkers for progression of diabetic nephropathy. Pteridines. 2019;30(1):183–8.10.1515/pteridines-2019-0024Search in Google Scholar
20 Ma N, Xu N, Yin D, Liu W, Wu M, Cheng X. Relationship between plasma total homocysteine and the severity of renal function in Chinese patients with type 2 diabetes mellitus aged ≥75 years. Medicine (Baltimore). 2020 Jul;99(27):e20737.10.1097/MD.0000000000020737Search in Google Scholar PubMed PubMed Central
21 Wang H, Cui K, Xu K, Xu S. Association between plasma homocysteine and progression of early nephropathy in type 2 diabetic patients. Int J Clin Exp Med. 2015 Jul;8(7):11174–80.Search in Google Scholar
22 Deebukkhum S, Pingmuangkaew P, Tangvarasittichai O, Tangvarasittichai S. Estimated creatinine clearance, homocysteine and high sensitivity-C-reactive protein levels determination for early prediction of nephropathy and atherosclerosis risk in type 2 diabetic patients. Indian J Clin Biochem. 2012 Jul;27(3):239–45.10.1007/s12291-012-0192-0Search in Google Scholar PubMed PubMed Central
23 Lu Y, Liu D, Feng Q, Liu Z. Diabetic nephropathy: perspective on extracellular vesicles. Front Immunol. 2020 Jun;11:943.10.3389/fimmu.2020.00943Search in Google Scholar PubMed PubMed Central
24 Al-Rubeaan K, Youssef AM, Subhani SN, Ahmad NA, Al-Sharqawi AH, Al-Mutlaq HM, et al. Diabetic nephropathy and its risk factors in a society with a type 2 diabetes epidemic: a Saudi national diabetes registry-based study. PLoS One. 2014 Feb;9(2):e88956.10.1371/journal.pone.0088956Search in Google Scholar PubMed PubMed Central
25 Ahmed MA, Kishore G, Khader HA, Kasturirangan MN. Risk factors and management of diabetic nephropathy. Saudi J Kidney Dis Transpl. 2013 Nov;24(6):1242–7.10.4103/1319-2442.121310Search in Google Scholar
26 Unnikrishnan RI, Rema M, Pradeepa R, Deepa M, Shanthirani CS, Deepa R, et al. Prevalence and risk factors of diabetic nephropathy in an urban South Indian population: the Chennai urban rural epidemiology study (CURES 45). Diabetes Care. 2007 Aug;30(8):2019–24.10.2337/dc06-2554Search in Google Scholar
27 Kronborg J, Jenssen T, Njølstad I, Toft I, Eriksen BO. Metabolic risk factors associated with serum creatinine in a non-diabetic population. Eur J Epidemiol. 2007;22(10):707–13.10.1007/s10654-007-9164-8Search in Google Scholar
28 Butani L, Polinsky MS, Kaiser BA, Baluarte HJ. Dietary protein intake significantly affects the serum creatinine concentration. Kidney Int. 2002 May;61(5):1907.10.1046/j.1523-1755.2002.00342.xSearch in Google Scholar
29 Lee EY, Lee YM, Choi KH, Lee HC, Lee BW, Kim BS. Comparison of two creatinine-based equations for predicting decline in renal function in type 2 diabetic patients with nephropathy in a Korean population. Int J Endocrinol. 2013;2013:848963.10.1155/2013/848963Search in Google Scholar
30 Matsue Y, van der Meer P, Damman K, Metra M, O’Connor CM, Ponikowski P, et al. Blood urea nitrogen-to-creatinine ratio in the general population and in patients with acute heart failure. Heart. 2017 Mar;103(6):407–13.10.1136/heartjnl-2016-310112Search in Google Scholar
31 Finkelstein JD. Pathways and regulation of homocysteine metabolism in mammals. Semin Thromb Hemost. 2000;26(3):219–25.10.1055/s-2000-8466Search in Google Scholar
32 Aguilar B, Rojas JC, Collados MT. Metabolism of homocysteine and its relationship with cardiovascular disease. J Thromb Thrombolysis. 2004 Oct;18(2):75–87.10.1007/s11239-004-0204-xSearch in Google Scholar
33 Bostom AG, Kronenberg F, Jacques PF, Kuen E, Ritz E, König P, et al. Proteinuria and plasma total homocysteine levels in chronic renal disease patients with a normal range serum creatinine: critical impact of true glomerular filtration rate. Atherosclerosis. 2001 Nov;159(1):219–23.10.1016/S0021-9150(01)00502-0Search in Google Scholar
34 Muzurović E, Kraljević I, Solak M, Dragnić S, Mikhailidis DP. Homocysteine and diabetes: role in macrovascular and microvascular complications. J Diabetes Complications. 2021 Mar;35(3):107834.10.1016/j.jdiacomp.2020.107834Search in Google Scholar PubMed
35 Girgis S, Suh JR, Jolivet J, Stover PJ. 5-Formyltetrahydrofolate regulates homocysteine remethylation in human neuroblastoma. J Biol Chem. 1997 Feb;272(8):4729–34.10.1074/jbc.272.8.4729Search in Google Scholar PubMed
36 Sombolos K, Papaioannou A, Christidou F, Natse T, Bamichas G, Gionanlis L, et al. The effect of two different doses comprising the simultaneous administration of intravenous B-complex vitamins and oral folic acid on serum homocysteine levels in hemodialysis patients. Int Urol Nephrol. 2006;38(3–4):725–30.10.1007/s11255-005-5791-8Search in Google Scholar PubMed
© 2021 Bin Ye et al., published by De Gruyter
This work is licensed under the Creative Commons Attribution 4.0 International License.
