Autoimmune gastritis (AIG) is an immune-mediated condition characterized by chronic atrophic gastritis.^1,2 It is distinguished by the presence of parietal cell antibodies (PCA) and/or intrinsic factor antibodies (IFA), which contribute to the atrophy of the gastric fundic mucosa.³ AIG is associated with severe complications, such as pernicious anemia (PA) and gastric neoplasms. Reduction in parietal cell numbers can lead to hypochlorhydria or achlorhydria. Hypochlorhydria prompts G cells to release gastrin, causing hypergastrinemia. Elevated stomach pH resulting from this condition fosters the proliferation of bacteria and diminishes the absorption of essential elements from food.⁴ Gastrin stimulates enterochromaffin-like (ECL) cells in the gastric body, promoting their proliferation. Excessive ECL cell proliferation may result in gastric neuroendocrine tumors (G-NET).^3,5 Reduced gastric acid impairs iron absorption, leading to iron deficiency anemia (IDA). Intrinsic factor, a glycoprotein secreted by parietal cells, aids in vitamin B₁₂ absorption in the distal ileum. Intrinsic factor deficiency hinders vitamin B₁₂ absorption, impairing tetrahydrofolate regeneration and causing PA.⁶ Numerous studies highlight the association between AIG and various autoimmune diseases, notably autoimmune thyroid diseases (AITD), type 1 diabetes mellitus, Addison's disease, and vitiligo.^3,7-11

Recent studies have identified distinct clinical characteristics of AIG among different subgroups.^10,12-14 Therefore, this study aims to investigate potential differences or similarities in demographics and laboratory findings associated with AIG among groups defined by age, sex, and concurrent AITD.

1. Patients

We conducted a retrospective study on patients diagnosed with AIG at the First Affiliated Hospital of Xiamen University from January 2018 to October 2023. Data, including age, sex, presence of PCA and/or IFA, hemoglobin levels, mean corpuscular volume , serum gastrin levels, serum pepsinogen (PG) levels, serum vitamin-B₁₂ levels, histological characteristics of gastric mucosa, and comorbid autoimmune disorders, were extracted from electronic medical records. The approving institution is the Ethics Committee of the First Affiliated Hospital of Xiamen University (approval number: SL-2025KY021-01). For the retrospective nature of this study, written informed consent was not required.

2. Diagnostic criteria

1) Inclusion criteria

A total of 183 patients were included if they exhibited histological atrophy of the gastric body with PCA and/or IFA, or endoscopic atrophy of the gastric body with PCA and/or IFA, accompanied by hypergastrinemia and a decreased serum PG-I to PG-II ratio (PG-I/PG-II).

2) Exclusion criteria

Patients were excluded if they met any of the following criteria: (1) recent use of proton pump inhibitors, H2 receptor blockers, or potassium ion competitive acid blockers within the past 2 months; (2) history of gastrectomy; (3) unavailable serological data; or (4) renal and/or liver function impairment.

3. Gastroscopic and histological procedures

Endoscopic confirmation of atrophy was defined by discoloration of the mucosa with visible vascular markings on the lesser and greater curvature of the corpus, along with the absence of folds, and normal or inflamed antrum. The biopsy site was chosen based on the findings of the endoscopy. A biopsy would be performed when there was atrophy or any obvious lesions in the gastric corpus and antrum. Atrophy, inflammation, and metaplasia were evaluated using the visual analog scale in the updated Sydney System.¹⁵

4. Laboratory tests

Anemia was defined as hemoglobin levels <11.0 g/dL for women or 12.0 g/dL for men. PA was diagnosed based on anemia, mean corpuscular volume >100 fL, and low vitamin B₁₂ levels (<233 pg/mL). IDA was defined by microcytic (mean corpuscular volume <80 fL) anemia with serum ferritin <30 ng/mL. IFA and PCA were tested using an indirect immunofluorescence technique with a kit (INOVA Diagnostics, Inc., San Diego, CA, USA). Gastrin levels were measured using the Gastrin Radioimmunoassay Kit (Siemens Healthcare Diagnostics, Gwynedd, UK). PG and gastrin 17 levels were measured using enzyme-linked immunosorbent assay. Decreased PG-I was defined as <90 ng/mL. PG-I/PG-II <3.0 was considered a decrease, while gastrin 17 >43.68 pg/mL was considered elevated. Hypergastrinemia was defined as gastrin levels >106.5 pg/mL and/or gastrin 17 >43.68 pg/mL. Helicobacter pylori (HP) infection status was defined as current infection and previous infection. Those with at least one positive urea breath test (UBT) or histological evidence (hematoxylin-eosin stain and an anti-HP immunohistochemical stain, VENTANA anti-HP SP48 Rabbit monoclonal, Ventana-Roche) were classified as “current HP infection patients.” Patients who had not undergone eradication therapy for HP and tested positive for serum HP antibodies were also classified as having a “current HP infection.” Patients with a confirmed history of successful eradication, indicated by a negative result in either a urea breath test, were classified as “previous HP infection patients.” In the UBT, the cutoff value for a positive diagnosis was set at 4‰ for delta over baseline .

5. Statistical analysis

Continuous variables were presented as mean±standard deviation or median with interquartile range. Categorical variables were expressed as numbers and percentages (n, %). The t-test and Mann-Whitney U test were used for normally and non-normally distributed variables, respectively. Categorical variables were compared using the chi-square test or Kruskal-Wallis H test as appropriate. Statistical analyses were conducted using SPSS Statistics 27.0 (IBM Corp., Armonk, NY, USA), with significance set at p<0.05 (two-sided).

1. Patient characteristics

A total of 183 patients diagnosed with AIG were included for analysis (Table 1). Patients ranged in age from 21 to 79 years with a mean age of 54.6±12.2 years and a female-to-male ratio of 2.4:1. The primary complaints reported by patients encompassed gastrointestinal, hematological, and neurological symptoms. Of these, 73.2% (134 patients) presented with gastrointestinal symptoms, including bloating, poor appetite, belching, and epigastric pain, with bloating being the most common symptom, affecting 23.9% (32 patients). Hematological symptoms were reported by 9.8% (18 patients), mainly presenting anemia-related signs such as pallor, fatigue, and dizziness. Additionally, 4.9% (9 patients) experienced neurological symptoms, characterized by limb numbness. Notably, 12.0% (22 patients) displayed no significant clinical symptoms and were diagnosed with AIG during routine health check-ups.

Table 1 . Clinicopathological Characteristics of 183 Autoimmune Gastritis Patients.

Characteristic	Value
Age, yr	54.6±12.2
Sex
Female	129 (70.5)
Male	54 (29.5)
Gastric histopathology
Corpus (n=172)
Atrophy
Normal	15 (8.7)
Mild	84 (48.8)
Moderate	59 (34.3)
Marked	14 (8.1)
Metaplasia	140 (81.4)
Intestinal metaplasia	79 (56.4)
Pseudo-pyloric metaplasia	37 (26.4)
Intestinal metaplasia+pseudo-pyloric metaplasia	24 (17.1)
ECL cell hyperplasia	104 (60.5)
Polyps	36 (20.9)
G-NET	9 (5.2)
Cancer	2 (1.1)
Antrum (n=111)
Atrophy
Normal	67 (60.4)
Mild	34 (30.6)
Moderate	8 (7.2)
Marked	2 (1.8)
Intestinal metaplasia	37 (33.3)
Gastric antibody
PCA positive	177 (96.7)
IFA positive	49 (26.8)
Hypergastrinemia (n=175)	155 (88.6)
Serum pepsinogen (n=89)
Decreased PG-I	78 (87.6)
Decreased PG-I/PG-II	53 (59.6)
Anemia (n=144)	60 (41.7)
IDA (n=85)	15 (17.6)
PA (n=81)	14 (17.3)
Comorbidity
Autoimmune disease	73 (39.9)
HP status (n=159)
Current infection	42 (26.4)
UBT	25
HP antibody	17
Histological evidence	11
Previous infection	13 (8.2)
Absence of HP infection	104 (65.4)

Data are presented as mean±SD or number (%)..

ECL cell, enterochromaffin-like cell; G-NET, gastric neuroendocrine tumors; PCA, parietal cell antibody; IFA, intrinsic factor antibodies; PG, pepsinogen; IDA, iron deficiency anemia; PA, pernicious anemia; HP, Helicobacter pylori; UBT, urea breath test..

Gastric histopathology results were available for 172 patients for the gastric corpus and 111 patients for the gastric antrum. In the corpus, 91.3% of patients exhibited atrophy, and metaplasia was observed in 81.4%. ECL cell hyperplasia was present in 60.5%, while G-NETs were found in 5.2%. Additionally, 20.9% of patients were diagnosed with gastric polyps in the corpus. Two cases of gastric cancer were identified. In the gastric antrum, atrophy was observed in 39.6% of patients, and 34.6% tested positive for HP. The positive rates of PCA and IFA were 96.7% and 26.8%, respectively. Anemia was present in 41.7% of AIG patients. Furthermore, 39.9% of AIG patients had at least one other autoimmune disease, with AITD being the most common (64.2%). Other comorbidities included type 1 diabetes mellitus (1.4%), ankylosing spondylitis (1.4%), rheumatoid arthritis (1.4%), systemic lupus erythematosus (1.4%), Sjogren's syndrome (1.4%), and autoimmune polyendocrine syndromes (5.6%).

2. Age groups

Patients were categorized into two age groups: younger than 60 years (n=115) and 60 years or older (n=68) (Table 2). The mean ages were 47.4±8.9 years and 66.9±5.0 years, respectively. Among older patients, 36.8% tested positive for IFA compared to 20.9% of younger patients (p=0.019). PA was found in 26.7% of older patients and 11.8% of younger patients (p=0.087), while IDA affected 26.4% of younger patients and 3.1% of older patients (p=0.006). Patients with IDA and PA had mean ages of 45.3±11.9 and 60.6±14.9 years, respectively (p=0.005). The prevalence of comorbid autoimmune disease was higher in the young group (46.1%) than in the older group (29.4%) (p=0.026). However, there were no significant differences between the groups in terms of sex ratio, degree of gastric atrophy, gastrin levels, or PG levels.

Table 2 . Age Differences Analysis of Autoimmune Gastritis Patients.

Variable	Young patients (<60 yr, n=115)	Old patients (≥60 yr, n=68)	p-value
Age, yr	47.4±8.9	66.9±5.0	-
Sex			0.516
Female	83 (72.2)	46 (67.6)
Male	32 (27.8)	22 (32.4)
Histopathology of corpus†
Atrophy			0.332
Normal	9 (8.4)	6 (9.2)
Mild	48 (44.9)	36 (55.4)
Moderate	43 (40.2)	16 (24.6)
Marked	7 (6.5)	7 (10.8)
Intestinal metaplasia	65 (60.7)	38 (58.5)	0.767
Pseudo-pyloric metaplasia	35 (32.7)	26 (40.0)	0.333
ECL cell hyperplasia	66 (61.7)	38 (58.5)	0.675
Polyps	13 (20.0)	23 (21.3)	0.839
G-NET	8 (7.0)	1 (1.5)	0.157
Gastric antibody
PCA positive	111 (96.5)	66 (97.1)	1.000
IFA positive	24 (20.9)	25 (36.8)	0.019*
Hypergastrinemia‡	99 (90.0)	56 (86.2)	0.440
Serum pepsinogen§
Decreased PG-I	54 (87.1)	25 (89.3)	1.000
Decreased PG-I/PG-II	36 (59.0)	17 (60.7)	0.880
PG-I levels, ng/mL	30.5 (6.3–67.0)	13.9 (7.4–58.3)	0.708
PG-II levels, ng/mL	10.4 (6.6–15.6)	9.7 (6.3–12.7)	0.387
PG-I/PG-II	2.7 (0.8–5.6)	1.9 (1.1–6.1)	0.989
Anemia
PAⅡ	6 (11.8)	8 (26.7)	0.087
IDA¶	14 (26.4)	1 (3.1)	0.006*
Comorbidity
Autoimmune disease	53 (46.1)	20 (29.4)	0.026*
Current HP infection	28 (27.5)	14 (24.6)	0.692

Data are presented as mean±SD, number (%), or median (interquartile range)..

ECL cell, enterochromaffin-like cell; G-NET, gastric neuroendocrine tumors; PCA, parietal cell antibodies; IFA, intrinsic factor antibody; PG, pepsinogen; PA, pernicious anemia; IDA, iron deficiency anemia; HP, Helicobacter pylori..

*p<0.05; ^†n=172, missing in 8 patients of the young group and 3 patients of the old group; ^‡n=175, missing in 5 patients of the young group and 3 patients of the old group; ^§n=89, missing in 54 patients of the young group and 40 patients of the old group; Ⅱn=81, missing in 64 patients of the young group and 38 patients of the old group; ^¶n=85, missing in 62 patients of the young group and 36 patients of the old group..

3. Sex groups

Table 3 presents the findings stratified by sex. ECL cell hyperplasia and pseudo-pyloric metaplasia were more prevalent among women (69.4%, 42.1%, respectively) compared to men (39.2%, 19.6%, respectively) (p<0.001 and p=0.005). The incidence of comorbid autoimmune disorders was 45.0% for women and 27.8% for men (p=0.030). All 15 patients with IDA were female (p=0.004). Additionally, women exhibited a higher occurrence of gastric polyps compared to men (24.6% vs 11.8%, p=0.058).

Table 3 . Sex Differences Analysis of Autoimmune Gastritis Patients.

Variable	Female (n=129)	Male (n=54)	p-value
Age, yr	54.7±11.3	54.3±14.2	0.861
Histopathology of corpus†
Atrophy			0.811
Normal	10 (7.8)	5 (9.3)
Mild	60 (46.5)	24 (44.4)
Moderate	43 (33.3)	16 (29.6)
Marked	8 (6.2)	6 (11.1)
Intestinal metaplasia	73 (60.3)	30 (58.8)	0.854
Pseudo-pyloric metaplasia	51 (42.1)	10 (19.6)	0.005
ECL cell hyperplasia	84 (69.4)	20 (39.2)	<0.001*
Polyps	30 (24.6)	6 (11.8)	0.058
G-NET	8 (6.2)	1 (1.9)	0.285
Cancer	1 (0.8)	1 (1.9)	0.295
Gastric antibody
PCA positive	125 (96.9)	52 (96.3)	1.000
IFA positive	33 (25.6)	16 (29.6)	0.573
Hypergastrinemia‡	113 (90.4)	42 (84.0)	0.123
Serum pepsinogen§
Decreased PG-I	60 (93.8)	19 (73.1)	0.012*
Decreased PG-I/PG-II	37 (58.7)	16 (61.5)	0.806
PG-I levels, ng/mL	25.9 (6.5–61.3)	23.5 (9.0–99.9)	0.510
PG-II levels, ng/mL	10.2 (6.2–13.7)	10.3 (6.6–14.9)	0.885
PG-I/PG-II	2.3 (0.8–5.7)	2.5 (1.5–6.5)	0.465
Anemia
PAⅡ	7 (12.3)	7 (29.2)	0.105
IDA¶	15 (25.0)	0	0.004
Comorbidity
Autoimmune disease	58 (45.0)	15 (27.8)	0.030*
Current HP infection	28 (25.2)	14 (29.2)	0.605

Data are presented as mean±SD, number (%), or median (interquartile range)..

ECL cell, enterochromaffin-like cell; G-NET, gastric neuroendocrine tumors; PCA, parietal cell antibodies; IFA, intrinsic factor antibody; PG, pepsinogen; PA, pernicious anemia; IDA, iron deficiency anemia; HP, Helicobacter pylori..

*p<0.05; ^†n=172, missing in 8 females and 3 males; ^‡n=175, missing in 4 females and 4 males; ^§n=89, missing in 66 females and 28 males; ^Ⅱn=81, missing in 72 females and 30 males; ^¶n=85, missing in 69 females and 29 males..

4. Comparison of AIG patients with and without AITD

Table 4 describes the findings among patients with and without comorbid AITD. Of the 68 AIG patients with AITD, the mean age was 50.9±12.0 years old, which was 8 years younger than those without AITD (p=0.003). Patients with AITD also showed a higher proportion of females compared to those without (p=0.023). The patients with AITD had a milder atrophy of the corpus in comparison to those without (p=0.009). No other significant differences in clinical features were observed between the groups. Table 5 details the presence of AITD antibodies among patients with AIG, with thyroid peroxidase antibodies being the most common, found in 75.3% of cases.

Table 4 . Comparison of Autoimmune Gastritis Patients with and without Autoimmune Thyroid Disease.

Variable	With AITD (n=68)	Without AITD (n=38)	p-value
Age, yr	50.9±12.0	58.0±11.5	0.003*
Sex			0.023*
Female	55 (80.9)	23 (60.5)
Male	13 (19.1)	15 (39.5)
Histopathology of corpus†
Atrophy			0.009*
Normal	5 (7.9)	1 (2.9)
Mild	36 (57.1)	15 (42.9)
Moderate	22 (34.9)	11 (31.4)
Marked	0	8 (22.9)
Intestinal metaplasia	39 (61.9)	26 (74.3)	0.214
Gastric antibody
PCA positive	64 (94.1)	37 (97.4)	0.652
IFA positive	24 (35.3)	12 (31.6)	0.699
Hypergastrinemia‡	61 (92.4)	33 (89.2)	0.719
Serum pepsinogen§
Decreased PG-I	48 (94.1)	23 (85.2)	0.227
Decreased PG-I/PG-II	32 (64.0)	17 (63.0)	0.928
PG-I levels, ng/mL	13.9 (7.1–56.9)	13.2 (5.6–60.9)	0.643
PG-II levels, ng/mL	10.2 (6.2–14.1)	10.2 (6.6–12.7)	0.969
PG-I/PG-II	2.4 (0.9–3.7)	1.7 (0.4–5.7)	0.661
Anemia
PAⅡ	6 (15.0)	5 (21.7)	0.511
IDA¶	9 (21.4)	2 (8.7)	0.302
Current HP infection	16 (26.7)	11 (32.4)	0.558

Data are presented as mean±SD, number (%), or median (interquartile range)..

AITD, autoimmune thyroid disease; PCA, parietal cell antibodies; IFA, intrinsic factor antibodies; PG, pepsinogen; PA, pernicious anemia; IDA, iron deficiency anemia; HP, Helicobacter pylori..

*p<0.05; ^†n=98, missing in 5 patients with AITD and 3 patients without AITD; ^‡n=103, missing in 2 patients with AITD and 1 patients without AITD; ^§n=77, missing in 18 patients with AITD and 11 patients without AITD; ^Ⅱn=63, missing in 28 patients with AITD and 15 patients without AITD; ^¶n=65, missing in 26 patients with AITD and 15 patients without AITD..

Table 5 . Antibodies in Patients with Autoimmune Gastritis.

Variable	No. (%)
TPOAb (n=73)	55 (75.3)
TgAb (n=72)	43 (59.7)
TRAb (n=58)	9 (15.5)
TSAb (n=17)	8 (47.1)

TPOAb, thyroid peroxidase antibody; TgAb, anti-thyroglobulin antibody; TRAb, thyrotropin receptor antibody; TSAb, thyroid stimulating antibody..

This study provided insights into the clinical characteristics of patients with AIG in China and explored differences across age, sex, and the presence/absence of AITD comorbidity.

First, we compared clinical characteristics between younger and older patients. Our study observed a notably higher positivity rate of IFA among elderly patients. PA, a serious complication of AIG, is primarily attributed to these antibodies, which hinder vitamin B₁₂ absorption.¹⁶ The high prevalence of PA among older patients likely correlates with their increased rates of IFA positivity. Moreover, the elevated IFA positivity in the elderly might signify heightened autoimmune activity in this demographic, possibly influenced by age-related changes in immune function. Further studies are warranted to elucidate the precise mechanisms linking IFA and older patients with AIG. Interestingly, older patients also exhibited higher positivity rates for PCA compared to younger patients, contrasting findings from prior studies reporting lower antibody prevalence in older individuals, as noted by Conti et al.¹⁷ This discrepancy likely arises from differences in diagnostic criteria, as our study included patients with at least one positive antibody marker. According to Conti’s research, 20% of elderly patients are seronegative for PCA at the time of AIG diagnosis.¹⁷ Thus, conducting IFA testing in elderly patients may increase diagnostic accuracy for AIG.

Younger patients showed a higher prevalence of IDA and comorbid autoimmune diseases. Notably, patients with IDA were approximately 15 years younger on average compared to those with PA, suggesting that AIG may predispose to iron deficiency many years before vitamin B₁₂ stores are depleted. The incidence of IDA is relatively high in young people, which may be related to various factors. First, AIG results in decreased stomach acid production, which is necessary for iron absorption.^18,19 Second, young women are prone to iron deficiency due to menstruation.^10,20 Third, the time for iron stores depletion in the body occurs earlier than that for vitamin B₁₂.¹⁰ Additionally, it is noteworthy that all IDA cases in this study were female. Menstrual blood creates an increased need for iron, yet poor iron absorption leads to a heightened vulnerability to IDA in young women.

Second, there was an apparent female predominance in our study cohort. This higher prevalence among females could potentially be attributed to sex-specific hormonal influences, immune responses, organ vulnerability, and differences in gut microbiota composition.^12,21 Our findings revealed increased incidences of pseudo-pyloric metaplasia and AITD among female patients, consistent with previous findings by Lahner et al.¹² However, contrary to their study, we did not observe a statistically significant age difference between male and female patients. Interestingly, ECL cell hyperplasia, which is linked to G-NET development, was notably associated with female patients in our cohort, consistent with the finding that G-NET predominantly affects female patients. In our study, we observed no statistically significant differences in gastrin levels between males and females. Despite this, females exhibited a consistently higher proliferation rate of ECL cells compared to males. These observations suggest a potential predisposition of female AIG patients towards G-NET development. Moreover, the incidence rate of polyps was higher in females compared to males, although this difference lacked statistical significance (p=0.058), possibly due to the small sample size. A previous study found that females exhibited a slightly higher prevalence of polyps (p=0.06).²² Despite being benign lesions, polyps have a non-negligible risk of neoplastic progression.²³ Consequently, it is advisable to consider endoscopic surveillance for women diagnosed with AIG.

Third, in our study, the prevalence of autoimmune comorbidity in AIG was 39.9%, with AITD (64.2%) being the most common. We conducted comparisons between patients with and without AITD. The higher prevalence of autoimmune thyroiditis among females and younger individuals aligns with findings from other studies.^17,24 Within AIG patients with concurrent AITD, there is a noticeable female predominance, which aligns with the overall characteristics of autoimmune illnesses. The following factors were found to be risk factors for AITD in a study by Lahner et al.²⁴ on 319 patients with AIG: female sex, presence of PCA, and presence of metaplastic atrophy. An important distinction in our study is that AIG patients with co-existing AITD exhibited milder atrophic changes; however, their study did not compare the extent of atrophy. This suggests that AIG patients with concurrent AITD are more likely to be diagnosed in the early stages of the disease. Analysis of thyroid antibodies found that thyroid peroxidase antibodies were most frequently elevated, followed by anti-thyroglobulin antibody. These two antibodies are commonly associated with Hashimoto’s thyroiditis, consistent with previous study²⁵ indicating Hashimoto’s thyroiditis as the predominant type of AITD in AIG.⁹ A study has shown that normal gastric acid secretion is necessary for the absorption of orally administered thyroid hormones.²⁶ Therefore, in patients with AIG combined with hypothyroidism, it is important to adjust the dosage of thyroid hormone medications accordingly. Hence, these findings underscore the importance of screening for additional autoimmune diseases, particularly AITD, following a diagnosis of AIG.

This study has several limitations. First, its retrospective and single-center design may limit the generalizability of the findings. Second, the unavoidable selection bias of a retrospective study may weaken the study’s representativeness. Third, the complications of G-NET, gastric cancer, and others have not been valuably evaluated due to the limited sample and low incidence rate. To provide a more thorough understanding of AIG’s clinical presentation and related risks, future research should include patients from multiple healthcare settings and large samples.

In conclusion, our findings indicate a notable prevalence of PA among older patients with AIG, likely attributable to their high rates of positive IFA. Conversely, younger women with AIG show a higher incidence of IDA. The most prevalent autoimmune comorbidity observed was AITD, particularly among young women, who also exhibited milder gastric atrophy. These findings will contribute to a better management strategy for patients with AIG.

This work was supported by the Natural Science Foundation of Fujian Province (2023J011616) and Investigator-Initiated Clinical Research Fund Project of The First Affiliated Hospital of Xiamen University (XMFHIIT-2023SL060).

No potential conflict of interest relevant to this article was reported.

Study concept and design: W.F.H. Data acquisition: X.W., Y.D. Data analysis and interpretation: C.J.L., Z.X. Drafting of the manuscript: X.W. Critical revision of the manuscript for important intellectual content: W.F.H., J.H.W. Statistical analysis: J.Y.Z., N.W. Obtained funding: W.F.H. Administrative, technical, or material support; study supervision: W.F.H., J.Y. Approval of final manuscript: all authors.

The datasets generated during and/or analyzed during the current study are not publicly available due to including personal medical and life information. But the datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.