Stratified analysis of the association between periodontitis and female breast cancer based on age, comorbidities and level of urbanization: A population-based nested case-control study.

PURPOSE: To conduct stratified analysis of the association between periodontitis exposure and the risk of female breast cancer based on age, comorbidities and level of urbanization.
METHODS: Using claims data taken from the 1997-2013 Taiwanese National Health Insurance Research Database (NHIRD), we identified 60,756 newly-diagnosed female breast cancer patients during the period 2003-2013 from all beneficiaries. We then randomly selected 243,024 women without breast cancer matching (1:4) for age and the year of the index date during 1997-2013 from a one million representative population acting as the control group. A conditional logistic regression analysis was used to examine the association between periodontitis (ICD-9-CM codes 523.3-4) and the risk of breast cancer, shown as an odds ratio (OR) with a 95% confidence interval (CI) after adjustments for the Charlson Comorbidity Index (CCI) and level of urbanization. Subgroup analyses were conducted based on age, CCI and level of urbanization.
RESULTS: The mean ± standard deviation age was 53 ± 14 years. After adjusting for potential confounders, the risk of female breast cancer was found to be associated with a history of periodontitis (OR, 1.12; 95% CI, 1.10-1.14). Such an association was significantly different between patients aged < 65 years (OR, 1.09; 95% CI, 1.06-1.11) and patients aged ≥ 65 years (OR, 1.23; 95% CI, 1.18-1.28; p for interaction <0.001), as well as between patients where the CCI = 0 (OR, 1.17; 95% CI, 1.15-1.20) and patients with CCI > 0 (OR, 0.99; 95% CI, 0.96-1.03; p for interaction <0.001). The highest level of urbanization was also associated with the risk of breast cancer.
CONCLUSIONS: This population-based nested case-control study demonstrated that periodontitis was significantly associated with the risk of female breast cancer and such an association was modified by both age and comorbidities.

Introduction

Female breast cancer is one of the most common malignancies among women worldwide [1]. In Taiwan, the age-standardized incidence rate of female breast cancer per 100,000 persons was 44.5 and significantly increased during 2000–2006 [2], with the mean hospital treatment cost and length of stay also having increased [3]. Prior studies have shown that the age of menarche, pregnancy, breastfeeding, late menopause, obesity, alcohol use and lack of physical activity were risk factors for breast cancer [4, 5], and a recent study has demonstrated that BRCA mutation carriers experience more severe disease and poorer outcomes [6].

Periodontitis is a chronic oral inflammation condition which is altered by microbiota and the microenvironment [7]. Several reports have shown the association between periodontitis and various chronic inflammatory diseases, including coronary heart disease, stroke, pneumonia, chronic obstructive pulmonary disease, chronic kidney disease, and diabetes mellitus [7-10], which may be due to an altered immune cell function [11-13]. Because immune cell function is affected by periodontitis, this response may theoretically correlate with the development of cancer. Certain studies have demonstrated that periodontitis has also been revealed as having an association with an increased risk of several cancers, including esophageal cancer, head and neck cancer, as well as lung cancer [14-18].

In Taiwan, the prevalence of periodontitis also significantly increased from 11.5% to 19.59% during the period 1997 to 2013 [19]. Chung et al. [20] observed an increased risk of a number of cancers among chronic periodontitis patients, and the adjusted hazard ratio was 1.23 for breast cancer. The available data regarding the association between periodontitis and breast cancer is limited. Soder et al. [21] analyzed 3,273 patients and revealed that severe periodontal disease increased the risk of breast cancer. Shao et al. [22] demonstrated a meta-analysis study and found periodontal disease may be a risk of female breast cancer. However, no population-based study with a large sample size has investigated the relationship. The data from the Taiwanese Health Insurance Research Database (NHIRD) had facilitated nationwide, population-based longitudinal studies. Therefore, in this study we used data taken from the NHIRD to assess the relationship between periodontitis and female breast cancer.

Materials and methods

Ethics statement

The study was permitted by the Institutional Review Board (IRB) of Taichung Veterans General Hospital (IRB Number: CE17100B). The requirement for informed consent was waived given that personal information was anonymized.

Study design

This study was a nationwide, retrospective population-based nested case-control study.

Data source

Claim data from the NHIRD in Taiwan

The study data included the 1997–2013 administrative data from the Taiwanese NHIRD. The National Health Insurance (NHI) program currently covers over 99% of the Taiwanese population. The data found within the NHIRD includes medication prescription history, ambulatory care services, admission services and traditional medical services. Certain personal data and history data, such as body weight, body height, alcohol use, and smoking habits, are not available in the NHIRD. The National Health Research Institute (NHRI) manages the NHIRD and provides the database to researchers for research purposes after anonymization of personal information is assured. The study used several NHIRD datasets, including enrolment files, inpatient and outpatient 1997–2013 claims files and NHI catastrophic illness files.

Catastrophic illness files

The NHI registered patients with catastrophic or major diseasese, including cancer and several autoimmune diseases such as systemic lupus erythematosus and rheumatoid arthritis. The NHIRD distributed a package of NHIR catastrophic illness files which include claims of inpatient and outpatient claims and registry details of patients in the catastrophic illness registry. Patients with a catastrophic illness can apply a cartastrophic illness certificate (CIC). The Bureau of NHI (BNHI) validated the accuracy of these catastrophic illness diagnoses in patients who apply CIC by two or more specialists through review of related laboraty data, medical records, pathological data and imaging findings. Patients who meet the diagnositic criteria of a catastrophic illness are issued a CIC and free from co-payment. We utilized the catastrophic illness files in the NHIRD to select newly diagnosed female breast cancer patients and matched patients found during the period 2003 to 2013.

Longitudinal Health Insurance Database

In 2000, the NHRI randomly selected and enrolled one million representative individuals from the NHIRD, establishing the Longitudinal Health Insurance Database (LHID2000). Comprehensive information on enrolment and utilization associated this randomly selected cohort are available. We selected a matched non-breast cancer comparison cohort from the population found in the LHID2000. We used LHID2000 claims data from 2003 to 2013 for analysis of the comparison cohort.

Study subjects

The flowchart of study subjects inclusion was shown in Fig 1.

Fig 1

Flow chart of the study design.

Identification of female breast cancer patients from the entire Taiwanese population

Patients who had a CIC for breast cancer (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] code 174) were considered female breast cancer patients. Using the 1997–2013 NHIRD, we identified 87,738 newly diagnosed female breast cancer patients from 2003 to 2013. The first date of CIC application was defined as the index date.

Selection of female breast cancer patients and age-matched non-breast cancer female controls

Using claim data from the 1997–2013 LHID2000, we identified 421,288 females who never had a diagnosis of breast cancer (ICD-9-CM code 174). We matched female breast cancer patients with females without breast cancer at a 1:4 ratio for age and year of the index date and finally included 60,756 female breast cancer patients and 243,.24 matched controls in the study.

Definition of periodontitis exposure

In Taiwan, beneficiaries of NHI were allowed to receive dental scaling no more than twice per year. Dentists may code gingival or periodontal disease for ambulatory visits of patient who received regular dental scaling. Therefore, in the study, only patients who had at least one outpatient visit before the index date with a diagnosis of acute or chronic periodontitis (ICD-9-CM codes 523.3–4) and concurrently received antibiotic treatment, or periodontal treatment, or scaling more than twice a year by dentists, were identified as patients with periodontitis.

Potential confounders

Potential confounders included level of urbanization and CCI. The CCI, as adapted by Deyo et al. [23], was used to represent the general level of comorbidity medical conditions. The presence of comorbidity was defined as a patient having at least three ambulatory visits or one inpatient visit with a corresponding ICD-9-CM code within 1 year prior to the index date. The level of urbanization of patient’s residence was categorized into four clusters according to density of population (people/km2), ratio of elderly individuals aged > 65 years, population ratio of subjects with educational levels of college or above, number of physicians/105 subjects, and population ratio of agricultural workers [24]. The variable level of urbanization is an ecological level variable since it describes properties of the region.

Proxy measures for severity of periodontitis to assess dose-response relationship with female breast cancer risk

In order to measure the relationship between the severity of periodontitis and the risk of female breast cancer, the period between the last periodontitis visit date to breast cancer diagnosis date (i.e., 0–3 months, 3–6 months, 6–1 months, 1–3 years, > 3 years), the number of visits required for periodontitis according to the 25th, 50th and 75th centiles, and the cumulative cost of periodontitis-related visits according to the 25th, 50th and 75th centiles were incorporated into the analysis.

Sensitivity analysis

Sensitivity analyses were conducted by using various definitions of gingival and periodontal disease based on ICD-9-CM codes (i.e., gingival and periodontal disease: ICD-9-CM codes 523, chronic periodontitis: ICD-9-CM code 523.4, periodontal disease: ICD-9-CM codes 523–3–5) to test whether the results remained robust.

Subgroup analysis

We conducted subgroup analyses based on age (i.e., <65 years, ≥65 years), CCI (i.e., 0, ≥1) and level of urbanization to test whether the results remained robust in various subgroups and to examine whether or not these variables have modification effects.

Statistical analysis

We compared continuous variables including age and CCI using the Student’s t-test and the categorical variables including variables related to periodontitis history and level of urbanization by using the Pearson’s chi-square test between cases and controls. A multivariable conditional logistic regression with maximum likelihood estimation was used to examine the association between periodontitis exposure and the risk of breast cancer development after adjustment for potential confounders shown as adjusted odds ratio (aOR) with 95% confidence interval (CI). Potential confounders included CCI and the level of urbanization (i.e., Leve1 1 to 4) [24]. The significance of modification effect by each covariate on periodontitis exposure-associated female breast cancer risk was examined by calculating the P-value of the coefficient associate with the product of each indicator of the covariate and the indicator of periodontitis using the Wald test.

These statistical analyses were performed using SAS statistical software, version 9.3 (SAS Institute, Inc., Cary, NC, USA). A p value less than 0.05 was considered statistically significant.

Results

Table 1 demonstrated the demographic and clinical data of the matched female breast cancer patients and the non-breast cancer female controls. The mean ± SD age was 53±14 years in both groups. Female breast cancer patients had a higher proportion of having a history of periodontitis than the control group. Female breast cancer patients had a higher proportion of living in the more urbanized regions than controls. The period between the last periodontitis visit date to the index date was shorter in the female breast cancer patients than in the matched controls. The number of periodontitis-related visits and the cumulative cost of periodontitis related visits were higher in female breast cancer patients than in controls. Female breast cancer patients also had a higher CCI than controls.

Table 1

Demographic and clinical data of patients with breast cancer and non-breast cancer controls.

Variable	Female non-breast cancer patients (n = 243,024)	Female breast cancer patients (n = 60,756)	P-value
Age, years (mean ± SD)	53±14	53±14	1.000
Gingival and periodontal disease definitions by ICD-9-CM
Gingival and periodontal diseases (ICD-9-CM: 523)	106,081 (43.7)	28,358 (46.7)	<0.001
Acute or chronic periodontitis (ICD-9-CM: 523.3–4)	69,604 (28.6)	19,018 (31.3)	<0.001
Chronic periodontitis (ICD-9-CM: 523.4)	18,917 (7.8)	5,343 (8.8)	<0.001
Periodontitis (ICD-9-CM: 523.3–5)	94,477 (38.9)	25,357 (41.7)	<0.001
Last periodontitis visit date to breast cancer diagnosis date			<0.001
0–3 months	4,065 (1.7)	1,466 (2.4)
3–6 months	4,080 (1.7)	1,201 (2.0)
6–12 months	7,023 (2.9)	1,884 (3.1)
1–3 years	19,505 (8.0)	5,162 (8.5)
>3 years	34,931 (14.4)	9,305 (15.3)
Number of visits for periodontitis			<0.001
Q1, Q2 (1)	51,369 (21.1)	13,917 (22.9)
Q3 (2)	10,824 (4.5)	2,985 (4.9)
Q4 (>2)	7,411 (3.0)	2,116 (3.5)
Cumulative cost of periodontitis-related visits (US dollars)			<0.001
Q1 (0–13)	17,902 (7.4)	4,662 (7.7)
Q2 (14–16)	18,314 (7.5)	5,269 (8.7)
Q3 (17–43)	16,964 (7.0)	4,454 (7.3)
Q4 (>43)	16,424 (6.8)	4,633 (7.6)
Level of urbanization			<0.001
3 (least urbanization)	16,612 (6.8)	3,085 (5.1)
2	35,724 (14.7)	7,680 (12.6)
1	108,933 (44.8)	26,900 (44.3)
0 (most urbanization)	81,755 (33.6)	23,091 (38.0)
CCI	0.4±1.0	1.2±2.3	<0.001
CCI group
0	196,320 (80.8)	41,007 (67.5)	<0.001
≥1	46,704 (19.2)	19,749 (32.5)

Data are shown as number (percentage) unless specified otherwise.

Abbreviations: ICD-9-CM, International Classification of Diseases, Ninth Revision, Clinical Modification; Q, quartile; CCI, Charlson comorbidity index.

Table 2 revealed crude and adjusted OR with 95% CI for the associations of female breast cancer with periodontitis and other covariates. After adjusting for potential confounders, periodontitis was significantly associated with the risk of female breast cancer (aOR, 1.12; 95% CI, 1.09–1.14). CCI ≥1 and higher level of urbanization were also risk factors for female breast cancer.

Table 2

Crude and adjusted OR with 95% CI for the association between variables and female breast cancer using conditional logistic regression analyses.

Variable	Univariable analysis	Multivariable analysis*
Periodontitis	1.14 (1.12–1.17)	1.12 (1.09–1.14)
CCI group
0	1.00 (Reference)	1.00 (Reference)
≥1	2.31 (2.26–2.36)	2.32 (2.27–2.37)
Level of urbanization
3 (least urbanization)	1.00 (Reference)	1.00 (Reference)
2	1.16 (1.11–1.21)	1.17 (1.127–1.22)
1	1.34 (1.28–1.39)	1.36 (1.30–1.41)
0 (most urbanization)	1.53 (1.47–1.60)	1.57 (1.50–1.63)

*Adjusting for all variables.

Abbreviations: OR, odds ratio; CI, confidence interval; CCI, Charlson comorbidity index

As shown in Table 3, the association between periodontitis and breast cancer was consistent using various definitions of periodontitis, implying that acute and/or chronic periodontitis was associated with breast cancer.

Table 3

Sensitivity analysis for the association of female breast cancer with gingival and periodontal disease using various definitions based on ICD-9-CM coder.

Gingival and periodontal disease definition by ICD-9-CM code	Univariable	Multivariable*
Gingival and periodontal diseases (ICD-9-CM code 523)	1.14 (1.12–1.16)	1.11 (1.09–1.13)
Periodontitis (ICD9-CM codes 523.3–4)	1.14 (1.12–1.17)	1.12 (1.10–1.14)
Chronic periodontitis (ICD-9-CM code 523.4)	1.15 (1.11–1.18)	1.11 (1.07–1.15)
Periodontal disease (ICD-9-CM codes 523.3–5)	1.13 (1.11–1.16)	1.10 (1.08–1.12)

Statistical analyses were conducted using a conditional logistic regression model.

*Adjusting for Charlson comorbidity index group (0, ≥1) and level of urbanization.

Abbreviations: ICD-9-CM, International Classification of Diseases, Ninth Revision, Clinical Modification.

Table 4 shows the OR with 95% CI associated with variables for the risk of female breast cancer. After adjusting for age and, the risk of female breast cancer was associated with a history of periodontitis (OR, 1.12; 95% CI, 1.10–1.14). A short interval period between the last periodontitis visit date to the breast cancer diagnosis date (OR, 1.46; 95% CI, 1.37–1.55), more periodontitis visits (OR, 1.17; 95% CI, 1.11–1.23), a higher cumulated periodontitis-related cost (OR, 1.15; 95% CI, 1.11–1.19), and higher urbanization level were all associated with the risk of breast cancer.

Table 4

Crude and adjusted odds ratios with 95% confidence intervals for the association between history of periodontitis and female breast cancer by conditional logistic regression analyses.

History of periodontitis	Univariable analysis	Multivariable analysis*
Interval between the last periodontitis visit and the index date
No periodontitis	1.00 (Reference)	1.00 (Reference)
0–3 months	1.51 (1.42–1.60)	1.45 (1.36–1.54)
3–6 months	1.23 (1.15–1.31)	1.19 (1.12–1.27)
6–12 months	1.12 (1.07–1.18)	1.09 (1.04–1.15)
1–3 years	1.11 (1.07–1.14)	1.08 (1.04–1.11)
>3 years	1.11 (1.09–1.14)	1.10 (1.07–1.12)
Number of visits for periodontitis
No periodontitis	1.00 (Reference)	1.00 (Reference)
Q1, Q2 (1)	1.13 (1.11–1.16)	1.11 (1.09–1.13)
Q3 (2)	1.15 (1.11–1.20)	1.12 (1.07–1.17)
Q4 (>2)	1.20 (1.14–1.26)	1.17 (1.11–1.23)
Cumulative cost of periodontitis-related visits (US dollars)
No periodontitis	1.00 (Reference)	1.00 (Reference)
Q1 (0–13)	1.09 (1.05–1.12)	1.07 (1.03–1.11)
Q2 (14–16)	1.21 (1.17–1.25)	1.19 (1.15–1.23)
Q3 (17–43)	1.10 (1.06–1.14)	1.07 (1.03–1.10)
Q4 (>43)	1.18 (1.14–1.22)	1.15 (1.11–1.19)

*Adjusting for Charlson comorbidity index group (0, ≥1) and level of urbanization.

Abbreviations: Q, quartile.

Table 5 shows that the association was significantly different between patients < 65 years of age (OR, 1.09; 95% CI, 1.06–1.11) and patients ≥ 65 years of age (OR, 1.23; 95% CI, 1.18–1.28; p for interaction <0.001), as well as between patients without the CCI (OR, 1.17; 95% CI, 1.15–1.20) and patients with the CCI (OR, 0.99; 95% CI, 0.96–1.03; p for interaction <0.001). The supplementary figures showed the partial effect plots to summarize the statistical models.

Table 5

Stratified analysis of the association between periodontitis and female breast cancer based on age, comorbidities and level of urbanization: A population-based nested case-control study*.

Subgroup	OR (95% CI)	p value	p for interaction
Age group			<0.001
Age<65	1.09 (1.06–1.11)	<0.001
Age≥65	1.23 (1.18–1.28)	<0.001
CCI group			<0.001
0	1.17 (1.15–1.20)	<0.001
≥1	0.99 (0.96–1.03)	0.652
Level of urbanization group			0.727
0 (least urbanization)	1.12 (1.03–1.22)	0.009
1	1.15 (1.09–1.21)	<0.001
2	1.11 (1.08–1.14)	<0.001
3 (most urbanization)	1.09 (1.06–1.13)	<0.001

*Adjusted variables included age, CCI and level of urbanization with the exclusion of the subgroup variable.

Abbreviations: CCI, Charlson comorbidity index; OR, odds ratio; CI, confidence interval.

Discussion

This is a nationwide, population-based nested case-control study which assessed the association between a history of periodontitis and the risk of female breast cancer development in Asia, and we have demonstrated a significant association of female breast cancer with periodontitis. The results remained robust in patients aged <65 years or ≥65 years and patients with a CCI = 0. Consistent with our finding, Soder et al. [21] selected 3,273 patients aged between 30–40 years and found that chronic periodontitis indicated by missing molars seemed to be associated with breast cancer (odds ratio 2.36). Of note, a meta-analysis study [22] found a periodontal disease significantly increased breast cancer risk (RR, 1.22; 95% CI, 1.06–1.40). However, the risk of breast cancer development was not significantly increased among patients with periodontal disease and a history of periodontal therapy. Therefore, the author concluded a potential association between periodontitis and breast cancer. Freudenheim et al. [25] analyzed a cohort of 73,737 postmenopausal women and also revealed that periodontitis was associated with the risk of postmenopausal breast cancer (HR, 1.14; 95% CI, 1.03–1.26). On the contrary, Farhat et al. [26] revealed no association between periodontal disease and overall breast cancer risk (HR, 1.02; 95% CI, 0.94–1.10). However, in this study, no confounder was analyzed. In contrast, we analyzed the association between periodontitis and breast cancer adjusting for potential confounds, and we also found the important modification effects by age and comorbidities.

Another finding of our study was that the magnitude of the association between periodontitis and female breast cancer is strongest when the lag time of the last periodontitis-related visit was <3 months. However, it is possible that occult breast cancer might have occurred before the last periodontitis-related visit when the interval was less than 3 months. Therefore, we cannot exclude the possibility of reverse causality. Further studies are warranted to assess whether breast cancer is a risk factor for periodontitis. Additionally, we also discovered a greater risk of breast cancer in the patients who had more severe periodontitis, suggesting a dose-response relationship between periodontitis and female breast cancer. Consistent with this finding, Soder et al. [21] revealed that breast cancer was associated with severe periodontal disease (periodontal disease and/or missing molars), but not mild periodontal disease (periodontal disease without missing molars).

Another important finding was that the association between periodontitis and female breast cancer was significantly different on the basis of age and CCI. In subgroup analyses, the association between periodontitis and female breast cancer was greater in elderly patients 65 years or older than in younger patients. The reason for this may be that elderly patients may have had a longer history of periodontitis, reflecting a longer exposure to bacteria and greater inflammation, leading to more severe periodontitis.

We also found that periodontitis exposure was not associated female breast cancer in patients with CCI≥1. Cao et al. [27] revealed that overall survival and progression-free survival of breast cancer were impacted by d CCI, implying that some comorbidities may play a mechanistic role in the development and progression of female breast cancer. Therefore, the presence of comorbidities may be a competitive risk factor for breast cancer in patients with periodontitis, leading a negative association between periodontitis and female breast cancer in patients with comorbidities. Another explanation was that patients with comorbidities had a higher chance of breast cancer surveillance than those without comorbidities, leading to a detection bias.

Whether periodontitis may induce breast cancer or not still remains unclear. Oral cavity bacteria may play a role in the potential mechanism. Previous studies [28-30] have found that breast duct tissue was exposed to various bacteria, with bacteria also being found in breast tumors. The long-term bacteria stimulation and inflammation seems to lead to cancer formation. Periodontitis is a chronic inflammation condition and its associated systemic inflammation status may play a role as another mechanism. Certain studies [31-35] have revealed that periodontal disease increased systemic inflammation markers, including C-reactive protein (CRP), cytokines and chemokines. Chan et al. [31] reviewed the meta-analysis of several studies and found that circulating CRP, a low grade inflammation marker, was associated breast cancer development. Noack et al. [33] showed CRP levels increased in periodontitis patients. Hayashi et al. [34] demonstrated oral pathogens can induce and maintain a chronic state of inflammation at sites distant from oral infection. Elinav et al. [35] showed that microbials played an important role in inflammation-induced cancer and also affected cancer development. These inflammation markers had an impact on carcinogenesis and could explain the association between periodontitis and breast cancer.

We also found that a higher level of urbanization was associated with a greater risk of female breast cancer. This finding may be explained by differential dietary patterns and life stress levels among patients living in areas with various levels of urbanization. Zhang et al. [36] and Fei et al. [37] found that breast cancer shows an urban-rural disparity. Previous studies [38-40] established the association between breast cancer and socioeconomic status, and socioeconomic factors included diet, alcohol consumption, physical exercise, etc.

Some limitations were noted in this study. First, some confounders, including menarche, pregnancy, breastfeeding, menopause, body weight, body height, alcohol use, and smoking, were not made available in this study, and all these factors can affect the risk of breast cancer. Second, the severity of periodontitis may affect one’s inflammation status, which could lead to breast cancer. Another limitation is that some variables are continuous variables, but we stratified these variables as categorical variables. In this way, we can find out high-risk populations easier. In order to overcome these limitations, we enrolled large numbers of matched patients and stratified patients by their number of visits for periodontitis treatment and the cumulative cost of their periodontitis-related visits. A higher number of visits for periodontitis treatment, along with a higher cumulative cost for periodontitis-related visits, resulted in patients being considered to have more active and severe periodontitis. Third, although the regular audit by the BNHI had improved coding accuracy, the accuracy of diagnoses based on claims data is still an issue of concern. However, the non-differential misclassification bias related to periodontitis diagnosis always drives the bias towards the null. Therefore, the magnitude of the association between periodontitis and breast cancer could only be underestimated.

In conclusion, this population-based nested case-control study revealed periodontitis exposure is significantly associated with the risk of breast cancer. We also found this association is modified by both age and comorbidities. Further studies are warranted to clarify the mechanisms.

Partial effect plots to show the interaction between age and periodontitis on the probability of female breast cancer.

(TIF)

Click here for additional data file.

Partial effect plots to show the interaction between age group and periodontitis on the probability of female breast cancer.

(TIF)

Click here for additional data file.

Partial effect plots to show the interaction between CCI and periodontitis on the probability of female breast cancer.

(TIF)

Click here for additional data file.

Partial effect plots to show the interaction between CCI group and periodontitis on the probability of female breast cancer.

(TIF)

Click here for additional data file.

28 Apr 2022

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Antonis Valachis

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf  and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Thank you for stating the following in your Competing Interests section:

“NO authors have competing interests.”

Please complete your Competing Interests on the online submission form to state any Competing Interests. If you have no competing interests, please state "The authors have declared that no competing interests exist.", as detailed online in our guide for authors at http://journals.plos.org/plosone/s/submit-now

This information should be included in your cover letter; we will change the online submission form on your behalf.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: No

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: No

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: No

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Introduction

Page 4. Did the prevalence of periodontitis increase because of changes in definitions? This has often been the case in other countries.

Page 4. There are many other cohort studies on periodontitis and breast cancer that are not referenced (e.g. Farhat et al CEBP 2021, Jia et al CEBP 2020, and earlier meta-analysis of 11 studies: Shao et al 2018). Also, the Sfreddo et al study referenced is not a cohort study. Please edit the last paragraph of the introduction to reflect this. The discussion (first paragraph) also needs to be revised since the association between periodontal disease and breast cancer risk has previously been reported in Taiwan using the same dataset (PMID: 26280747); please explain the differences/overlap between the two studies in the discussion section.

Methods

This is not a traditional case-control study; based on the selection of all breast cancer cases during a 10-year period and selection of controls from a large cohort database for comparison, this would be better described as a retrospective population-based cohort study.

It appears that all incident breast cancer cases identified through the NHIRI between 2003 and 2013 were included in this analysis. It is unclear if the breast cancer cases that are not in the NHRI database would have detailed data on claims for periodontal disease or other co-morbidities and if these were also included. A figure would help between understand the selection process including exclusions that were made and available data available for each step.

Statistical analysis – please provide details on how the tests for interactions were conducted.

Would it be possible to differentiate different types of breast cancers? E.g. invasive vs non-invasive?

Also, given the strong association between urbanization and breast cancer, it would be useful to examine whether the periodontal disease and breast cancer is modified by this variable (this could be added to Table 5).

Discussion

The discussion could be improved. Prior results from existing studies on this topic are not provided (e.g., range of RRs to compare magnitude of association, quality of periodontal disease measurement). It would also be helpful to point out some limitations in prior studies, and what this study adds to the literature (given that there are over 10 papers on this topic). A better synthesis of the literature is needed.

Typos (tables and “CRT” in discussion) and English grammar need to be corrected throughout.

Reviewer #2: This study explored the effect of periodontitis on female breast cancer using a population based case control study. The paper is reasonably well written though I have some comments which the authors might consider to improve their study.

Introduction

Page 4 final paragraph – "small sample size" is somewhat relative – I would state the sample sizes of these studies

Page 5: "...while Sfreddo et al. [21] found there was a significant association between periodontitis and breast cancer" - Presumably this is a positive association? Please clarify in the text

Materials and methods

Page 7 "Potential confounders included level of urbanization". Presumably this is a city / regional level variable in the dataset?

"Sensitivity analyses were conducted based upon age (<65 years, ≥65 years) and CCI" – further elaboration required.

"We tested the differences in the continuous variables through use of the Student’s t-test and the categorical variables by using the Pearson’s χ2 test" What is the motivation for this and define the specific variables?

Logistic regression model – presumably maximum likelihood was used to estimate the model parameters, but it would be good to state this explicitly.

There is a strong case for modelling the dose response relationship as a continuous response. Arbitrarily creating categories is statistically inefficient and at worst can be misleading. I would advise modelling each explanatory variable of interest for the dose response as a continuous term. This can easily be achieved in SAS using a spline. A good choice would be a restricted cubic spline with the number of knots selected based on prior knowledge or beliefs (3-5 knots is often sensible – see Harrell F. (2015) Regression Modelling Strategies). The dose response adjusted for covariates can be presented as a smooth curve with confidence intervals. Odds ratios can still be obtained from this curve between specific points of interest. You might also consider modelling the CCI this way. Even if this approach is not adopted, the authors might consider mentioning this as a limitation.

Details of the interaction term (later referred to in subsequent sections) in the logistic regression models are missing. This is critical since the primary interest (according to the title of the paper) is effect modification necessitating interaction terms.

Perhaps the authors might consider making dataset and SAS code used available as SI

Results

Page 9 "After adjusting for age and index date"…this implies that age and index date were covariates in the regression model – this does not appear to be the case according to the statistical analysis section as it stands.

It would be nice to have some supportive figures (e.g. partial effect plots) summarizing the statistical models.

Discussion

First paragraph talks about several studies one by one – it would be preferable to integrate this as continuous prose and convey the message that your work is supportive of these previous studies.

"In this study, one important finding is that the risk of breast cancer is modified by both age and comorbidities" - this statement is redundant since it is a repetition of the first sentence of the paragraph.

Table 3 – It seems patient is the statistical unit as opposed to an adjustment variable as currently indicated at the foot of the table. Also urban should be "urbanization".

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

9 Jun 2022

Reviewer #1: Introduction

1.Page 4. Did the prevalence of periodontitis increase because of changes in definitions? This has often been the case in other countries.

Response: Thanks for your comment! Yes, the prevalence of periodontitis varied with changes in definitions based on ICD-9-CM codes (Table 1). Therefore, we conducted a sensitivity analysis for the association between periodontitis and breast cancer using various definitions of periodontitis and found consistent results with ORs from 1.10 to 1.12. Also, the non-differential misclassification bias related to periodontitis diagnosis always drives the bias towards the null. Therefore, the magnitude of the association between periodontitis and breast cancer could only be underestimated. We added a description regarding the influence of periodontitis definition on its association with breast cancer in the last paragraph of the Discussion section: ‘Third, although the Bureau of NHI audit had improved coding accuracy, the accuracy of diagnoses based on claims data is still an issue of concern. However, the non-differential misclassification bias related to periodontitis diagnosis always drives the bias towards the null. Therefore, the magnitude of the association between periodontitis and breast cancer could only be underestimated.’

2.Page 4. There are many other cohort studies on periodontitis and breast cancer that are not referenced (e.g. Farhat et al CEBP 2021, Jia et al CEBP 2020, and earlier meta-analysis of 11 studies: Shao et al 2018). Also, the Sfreddo et al study referenced is not a cohort study. Please edit the last paragraph of the introduction to reflect this. The discussion (first paragraph) also needs to be revised since the association between periodontal disease and breast cancer risk has previously been reported in Taiwan using the same dataset (PMID: 26280747); please explain the differences/overlap between the two studies in the discussion section.

Response: Shao et al [21] demonstrated a meta-analysis study and found periodontal disease may be a risk of female breast cancer. (page 4)

A meta-analysis study [21] found a potential association between periodontitis and breast cancer. (page 11)

In contrast, we analyzed the association between periodontitis and breast cancer in females, and we also found the important modification effects by age and comorbidities. (page 11)

3.Methods

This is not a traditional case-control study; based on the selection of all breast cancer cases during a 10-year period and selection of controls from a large cohort database for comparison, this would be better described as a retrospective population-based cohort study.

Response: Because we retrospectively selected the population of interested based on the outcome (i.e., incident female breast cancer patients as cases and female individuals without breast cancer as controls) and identified the history of periodontitis exposure before the index date, this study was a retrospective population-based case-control study. (page 5)

4.It appears that all incident breast cancer cases identified through the NHIRI between 2003 and 2013 were included in this analysis. It is unclear if the breast cancer cases that are not in the NHRI database would have detailed data on claims for periodontal disease or other co-morbidities and if these were also included. A figure would help between understand the selection process including exclusions that were made and available data available for each step.

Response: Taiwan's National Health Insurance (NHI) began in March 1, 1995 and covered 99.9% of Taiwan's residents by 2014. If the female breast cancer cases were not included in the NHRI database, data on claims for periodontal disease or other comorbidities should be unavailable. We added a figure to show the process of study subjects inclusion.

5.Statistical analysis – please provide details on how the tests for interactions were conducted.

Response: We had provided details on how to test the interaction effects in the statistical method subsection in the Methods section: ‘The significance of modification effect by each covariate on periodontitis exposure-associated female breast cancer risk was examined by calculating the P-value of the coefficient associate with the product of each indicator of the covariate and the indicator of periodontitis using the Wald test.’ (page 9)

6.Would it be possible to differentiate different types of breast cancers? E.g. invasive vs non-invasive?

Response: Female breast cancer patients were defined as those having a Catastrophic Illness Certificate (CIC) for breast cancer [International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) code 174]. This criteria only included the invasive breast cancer patients.

7.Also, given the strong association between urbanization and breast cancer, it would be useful to examine whether the periodontal disease and breast cancer is modified by this variable (this could be added to Table 5).

Response: We have examined the modification effect of urbanization on the association between periodontitis and female breast cancer and added the results to Table 5.

8.Discussion

The discussion could be improved. Prior results from existing studies on this topic are not provided (e.g., range of RRs to compare magnitude of association, quality of periodontal disease measurement). It would also be helpful to point out some limitations in prior studies, and what this study adds to the literature (given that there are over 10 papers on this topic). A better synthesis of the literature is needed.

Response: Soder et al. [20] selected 3,273 patients aged between 30-40 years and found that chronic periodontitis indicated by missing molars seemed to be associated with breast cancer (odds ratio 2.36). Of note, a meta-analysis study [21] found a periodontal disease significantly increased breast cancer risk (RR = 1.22, 95% CI: 1.06-1.40). However, the risk of breast cancer development was not significantly increased among patients with periodontal disease and a history of periodontal therapy. Therefore, the author concluded a potential association between periodontitis and breast cancer. Freudenheim et al. [24] analyzed a cohort of 73,737 postmenopausal women and also revealed that periodontitis was associated with the risk of postmenopausal breast cancer (HR 1.14, 95% CI: 1.03-1.26). On the contrary, Farhat et al. [25] revealed no association between periodontal disease and overall breast cancer risk (HR 1.02, 95% CI: 0.94-1.10). However, in this study, no confounder was analyzed. In contrast, we analyzed the association between periodontitis and breast cancer adjusting for potential confounds, and we also found the important modification effects by age and comorbidities. (page 11)

9.Typos (tables and “CRT” in discussion) and English grammar need to be corrected throughout.

Response: We corrected it. (page 13)

Reviewer #2: This study explored the effect of periodontitis on female breast cancer using a population based case control study. The paper is reasonably well written though I have some comments which the authors might consider to improve their study.

10.Introduction

Page 4 final paragraph – "small sample size" is somewhat relative – I would state the sample sizes of these studies

Response: Soder et al. [20] analyzed 3,273 patients and revealed that severe periodontal disease increased the risk of breast cancer. (page 4)

11.Page 5: "...while Sfreddo et al. [21] found there was a significant association between periodontitis and breast cancer" - Presumably this is a positive association? Please clarify in the text

Response: We clarified the results of the meta-analysis study by Sfreddo et al. [21]: ‘ Of note, a meta-analysis study [21] found a periodontal disease significantly increased breast cancer risk. However, the risk of breast cancer development was not significantly increased among patients with periodontal disease and a history of periodontal therapy. Therefore, the author concluded a potential association between periodontitis and breast cancer.’ (page 11)

12.Materials and methods

Page 7 "Potential confounders included level of urbanization". Presumably this is a city / regional level variable in the dataset?

Response: Thanks for your comment. The variable of level of urbanization is an individual level variable based on individual’s residence region in the dataset.

13."Sensitivity analyses were conducted based upon age (<65 years, ≥65 years) and CCI" – further elaboration required.

Response: We are sorry that this is a mistake. “Sensitivity” analyses should be replaced with “Subgroup” analyses. Conducted subgroup analyses and examined the modification effects by age group and CCI group on the association between periodontitis exposure and the risk of female breast cancer. We also added a description regarding the statistical method to assess the modification effect in the statistical method subsection in the Methods section: ‘The significance of modification effect by each covariate on periodontitis exposure-associated female breast cancer risk was examined by calculating the P-value of the coefficient associate with the product of each indicator of the covariate and the indicator of periodontitis using the Wald test.’(page 9)

14."We tested the differences in the continuous variables through use of the Student’s t-test and the categorical variables by using the Pearson’s χ2 test" What is the motivation for this and define the specific variables?

Response: We intended to compare the continuous variables and categorical variables between cases and controls. To clarify our motivation of this and define the specific variables, we revised our description as: ‘We compared continuous variables including age and CCI using the Student’s t-test and the categorical variables including variables related to periodontitis history and level of urbanization by using the Pearson’s chi-square test between cases and controls.’

15.Logistic regression model – presumably maximum likelihood was used to estimate the model parameters, but it would be good to state this explicitly.

Response: We revised the description regarding statistical analysis in the statistical analysis subsection in the Methods section as:’ A multivariable conditional logistic regression with maximum likelihood estimation was used to examine the association between periodontitis exposure and the risk of breast cancer development after adjustment for potential confounders shown as adjusted odds ratio (aOR) with 95% confidence interval (CI).’

16.There is a strong case for modelling the dose response relationship as a continuous response. Arbitrarily creating categories is statistically inefficient and at worst can be misleading. I would advise modelling each explanatory variable of interest for the dose response as a continuous term. This can easily be achieved in SAS using a spline. A good choice would be a restricted cubic spline with the number of knots selected based on prior knowledge or beliefs (3-5 knots is often sensible – see Harrell F. (2015) Regression Modelling Strategies). The dose response adjusted for covariates can be presented as a smooth curve with confidence intervals. Odds ratios can still be obtained from this curve between specific points of interest. You might also consider modelling the CCI this way. Even if this approach is not adopted, the authors might consider mentioning this as a limitation.

Response: We have modelled the dose response relationship of age and CCI for the interaction with periodontitis as a continuous response and showed the results in supplementary figure 1 & 3.

17.Details of the interaction term (later referred to in subsequent sections) in the logistic regression models are missing. This is critical since the primary interest (according to the title of the paper) is effect modification necessitating interaction terms.

Response: Details of the interaction term was described in the Statistical analysis subsection in the Methods section: ‘The significance of modification effect by each covariate on periodontitis exposure-associated female breast cancer risk was examined by calculating the P-value of the coefficient associate with the product of each indicator of the covariate and the indicator of periodontitis using the Wald test.’ (page 9)

18.Perhaps the authors might consider making dataset and SAS code used available as SI

Response: Thank you for your comment. We have provided our dataset and SAS code in the supplementary files.

19.Results

Page 9 "After adjusting for age and index date"…this implies that age and index date were covariates in the regression model – this does not appear to be the case according to the statistical analysis section as it stands.

Response: We are sorry to make a wrong description. Actually, age and year of the index date are matching variables, not adjustment variables. We corrected this description as: ‘After adjusting for CCI and level of urbanization, periodontitis was significantly associated with the risk of female breast cancer (aOR, 1.12; 95% CI, 1.09–1.14).’ (page 9)

‘

20.It would be nice to have some supportive figures (e.g. partial effect plots) summarizing the statistical models.

Response: Thank you for your suggestion. We have conducted partial effect plots to summarize the statistical models and showed them in the supplementary files. (page 10)

21.Discussion

First paragraph talks about several studies one by one – it would be preferable to integrate this as continuous prose and convey the message that your work is supportive of these previous studies.

Response: We had revised our first paragraph of the Discussion section: ‘This is the first nationwide, population-based case control study which assessed the association between a history of periodontitis and the risk of female breast cancer development in Asia, and we have demonstrated a significant association of female breast cancer with periodontitis. The results remained robust in patients aged <65 years or ≥65 years and patients with a CCI = 0. Consistent with our finding, Soder et al. [20] selected 3,273 patients aged between 30-40 years and found that chronic periodontitis indicated by missing molars seemed to be associated with breast cancer. Of note, a meta-analysis study [21] found a periodontal disease significantly increased breast cancer risk. However, the risk of breast cancer development was not significantly increased among patients with periodontal disease and a history of periodontal therapy. Therefore, the author concluded a potential association between periodontitis and breast cancer. Freudenheim et al. [24] analyzed a cohort of 73,737 postmenopausal women and also revealed that periodontitis was associated with the risk of postmenopausal breast cancer. On the contrary, Farhat et al. [25] revealed no association between periodontal disease and overall breast cancer risk (HR 1.02, 95% CI; 0.94–1.10). However, in this study, no confounder was analyzed. In contrast, we analyzed the association between periodontitis and breast cancer adjusting for potential confounds, and we also found the important modification effects by age and comorbidities.’ (page 11)

22."In this study, one important finding is that the risk of breast cancer is modified by both age and comorbidities" - this statement is redundant since it is a repetition of the first sentence of the paragraph.

Response: We deleted this statement. (page 11)

23.Table 3 – It seems patient is the statistical unit as opposed to an adjustment variable as currently indicated at the foot of the table. Also urban should be "urbanization".

Response: We corrected it. (Table 3)

Submitted filename: response to reviewers 20220609.docx

Click here for additional data file.

4 Jul 2022

9 Jul 2022

Reviewer #1:

1.The authors have not addressed the comment on how their study was different from the previous study published on periodontal disease and cancer using the same Taiwan database (and also fail to mention this previous study) "A population-based study on the associations between chronic periodontitis and the risk of cancer" Shiu-Dong Chung et al.

Response: Chung et al [20] observed an increased risk of a number of cancers among chronic periodontitis patients, and the adjusted hazard ratio was 1.23 for breast cancer. (page 4)

2.Since breast cancer is a female cancer, stating that they studied females (vs both sexes) makes little sense as obviously prior studies on breast cancer risk were in women only. Please mention this study in the Introduction and this study should not be referred to as the "first" case-control on periodontal disease that is population based (as it is not).

Response: We deleted ‘’first’’. (page 11)

3.This is not a retrospective case-control study. Retrospective case-control studies are studies that recruit cancer patients and ask about their past exposures; the exposure here is measured prior to the development of the disease. So, this is a retrospective cohort study, because the cohort exists and the analysis is conducted on a sample of the cohort retrospectively. Alternatively, you can call it a nested case-control study, which is nested in a cohort population.

Response: We rewrite as ‘’nested case-control study’’. (page 1, 2, 5,11,15)

4.Figure 1 – can you clarify what is meant by “unsured amount” given that half of the population is excluded for this reason (+ missing data) can you justify whether this was a reasonable exclusion and whether it might have caused some selection bias (in other words, how did the excluded population differ from the one included)?

Response: We are sorry to make a mistake here. Actually, we only excluded patients with missing data of residence region (a confounding factor in this study). Also, in 906,287 individuals from the 1997-2013 Longitudinal Health Insurance Database, we excluded excluded 458,873 men first and then excluded anther 26,126 women who had missing data of residence region. We have revised Figure 1.

5.Table 5. The title needs to be modified to clearly describe that the ORs presented are for the comparison of periodontal disease vs no periodontal disease stratified on age, CCI and urbanization. Many readers will misinterpret these results otherwise.

Response: We have revised the tile as ‘Stratified Analysis of the Association Between Periodontitis and Female Breast Cancer Based on Age, Comorbidities and Level of Urbanization: A Population-based Nested Case-control Study.’

6.I don’t think the figures add any value, especially the ones that are based on categorial comparisons.

Response: Thank you for your suggestion. The other reviewer requested us to plot these figures based on both categorical and continuous comparisons. However, we have put them in the supplemental materials.

7.Please edit the abstract for grammar "assess that periodontitis" change to "assess whether periodontitis"

Response: We revised the sentence as: ‘ To conduct stratified analysis of the association between periodontitis exposure and the risk of female breast cancer based on age, comorbidities and level of urbanization.’ (page 2)

Reviewer #2: I enjoyed re-reviewing this interesting study and was pleased that the authors considered many of my suggestions. I only have two minor comments:

8.Response to Reviewers - Reviewer #2: Point 12. The authors responded that "The variable of level of urbanization is an individual level variable based on individual’s residence region in the dataset." I believe this should be termed an ecological level variable since it describes properties of the region, and not each person.

Perhaps a very slight re-wording will suffice.

Response: Thank you for your suggestion. We added a description regarding the variable of level of urbanization ‘The variable level of urbanization is an ecological level variable since it describes properties of the region.’ (page 8)

9.The conclusion section looks a little peculiar with just one sentence. I would add a few more sentences or integrate into the discussion depending on the requirements of the journal.

Response: We revised the conclusion section and integrated it into the Discussion section as: ‘In conclusion, this population-based nested case-control study revealed periodontitis exposure is significantly associated with the risk of breast cancer. We also found this association is modified by both age and comorbidities. Further studies are warranted to clarify the mechanisms.’(page 15)

Submitted filename: response to reviewers.docx

Click here for additional data file.

12 Jul 2022

PONE-D-22-02756R2

Dear Dr Chen,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Antonis Valachis

Academic Editor

PLOS ONE

14 Jul 2022

PONE-D-22-02756R2

Dear Dr. Chen:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Assoc Prof Antonis Valachis

Academic Editor

PLOS ONE