Factors related to reversal of prediabetes in patients from a cardiovascular risk program during 2019 - 2023

Abstract

Introduction

Type 2 diabetes mellitus (T2DM) impacts the development of cardiovascular disease. Prediabetes, called intermediate hyperglycemia, is diagnosed with glucose levels that do not meet the criteria for T2DM, but are elevated. Lifestyle changes achieve a reversal of up to 58%, structured within an adequate follow-up in cardiovascular risk programs.

Methodology

Longitudinal, descriptive study, with analytical scope, retrospective collection of information from a cohort of adults with a diagnosis of prediabetes; in quarterly follow-up by a cardiovascular risk program of a III level institution between 2019 - 2023. Univariate and bivariate analysis and logistic regression were performed to define the correlation between reversion to normoglycemia and associated factors.

Results

Mean age was 69.5 years (SD 10.1), 50.8% women, median follow-up 366 days (IQR 249 - 569).

Performing physical activity > 150 minutes per week (validated by International Physical Activity Questionnaire - IPAQ), increases 4.15 times the chance of reversing prediabetes (p< 0.001); while decreasing the chance of reversal, having an HbA1c ≥ 6.0 to 86% (p= 0.014) and BMI ≥ 25 to 75% (p= 0.005).

Conclusions

Follow-up in structured cardiovascular risk programs allows estimating factors related to prediabetes reversion and emphasizing strategies to reduce comorbidities. Population-based studies are required to expand the results.

Background

Type 2 diabetes mellitus (T2DM) is a condition that impacts the risk of cardiovascular disease and implies a high public health burden [1–3]. T2DM was the eighth leading cause of death and disability combined worldwide, with nearly 460 million people in all countries and age groups living there in 2019 [4, 5], with similar statistics in Colombia between 20–79 years of age and projected 5,014,000 diagnosed by 2045 [6].

Prediabetes or intermediate hyperglycemia [7, 8], is defined as elevated glycemia levels not meeting diagnostic criteria for diabetes. According the American Diabetes Association (ADA) [9] the diagnosis of prediabetes can be made with altered fasting glycemia, impaired oral glucose tolerance and/or elevated levels of glycated hemoglobin (HbA1c) of 5.7% to 6.4%. Oral intolerance following consumption of a 75-gr glucose load is defined as glycemia between 140–199; furthermore, an altered fasting glycemia is established between 100–125 mg/dL [7, 10].

Prediabetes prevalence ranges between 27–49%, according to age, gender, ethnicity, geographic location, lifestyle and socioeconomic level [3, 11]. Progression rate from prediabetes to T2DM is estimated at 5–10% per year [12–15], while 70% will develop diabetes in their lifetime, with a higher risk in overweight or obese people [12, 14, 16, 17]. Prediabetes has been proveen, an association with cardiovascular disease (CVD) [18, 19], risk of death from all causes [14, 16, 20, 21] and microvascular complications in up to 25%, as diabetic neuropathy [11], diabetic retinopathy in 8% [22] and also has been associated with chronic kidney disease [23].

Prediabetic state is considered a decisive moment for progression to T2DM [20]. Evidence shows that lifestyle changes can delay or even reverse the risk of becoming diabetic [14, 21, 24]. Diagnosis of prediabetes can be reduced by up to 58% with timely lifestyle interventions [15], decreased burden of disease and increased life expectancy [14]. Reduction of body mass index (BMI) and physical activity of at least 150 minutes per week have decreased T2DM incidence by more than 50% [14, 15, 17, 19]. Multidisciplinary cardiovascular risk programs aim these interventions to delay complications related to cardiovascular disease [13, 17, 19, 20]. Consequently, the proposal includes encouraging lifestyle changes, such as: a diet low in simple carbohydrates, physical activity for a minimum of 150 minutes, cessation of smoking and alcoholic beverages [25]. Management options are available in prediabetes, timely interventions at an earlier stage, are important to prevent progression to T2DM [4].

Results of previous studies [26, 27], strategies to identify individuals predisposed to T2DM have been developed. In 2003, one study was published, generated a predictive model, which resulted FINDRISC (Finnish Diabetes Risk Score) [28], a score that identified most relevant factors, validated in several populations [29, 30], today is the most recommended instrument in T2DM screening guidelines [31]. Monitoring of prediabetes in structured cardiovascular risk programs is one of these interventions, estimating factors related to disease progression or reversal, allows us to emphasize these aspects and motivate continuity of follow-up to evaluate impact of these strategies for reduction of comorbidities [5, 19, 32], including strategies for early assessment of insulin resistance as a predictor of non-regression, using HOMA-IR and, more recently, the glucose/triglyceride index [33–35].

Knowing sociodemographic characteristics of patients diagnosed with prediabetes, included in cardiovascular risk programs with implementation of lifestyle changes, establishing factors related to non-reversion of prediabetes and progression to diabetes, would allow developing policies on prevention and health promotion, which help to reduce the burden of disease [1, 26]. Therefore, the objective of this study is to determine the factors related to the reversion of prediabetes in prediabetic adults participating in a cardiovascular risk program at a referral center in Cali, Colombia, between 2019 and 2023.

Methodology

Study design

Longitudinal and analytical scope with retrospective collection of information from a cohort of participants.

Study population

Eighteen years and older persons with prediabetes diagnosis, enrolled in cardiovascular risk program of hight complexity health care institution between 2019 and 2023.

Selection criteria

Inclusion criteria

Records of adults of both sexes enrolled cardiovascular risk program of DIME neurocardiovascular clinic will be included, with laboratory reports for metabolic assessment and anthropometric measurements, at baseline and after 6 months of follow-up.

Exclusion criteria

Records of persons who received hypoglycemic treatment or drugs for weight reduction during the follow-up period.

Study area

The selected population was taken from a Hight complexity health care institution, located in Cali - Valle (Colombia), in which patients with chronic pathologies are treated in outpatient and inpatient care.

Cardiovascular risk program includes: quarterly follow-up with assessment by an internist, psychologist, physiotherapist, nutritionist, to reinforce adherence to lifestyle changes (diet with reduction of simple carbohydrates, increased consumption of fiber and PUFAs, in an individualized manner by Nutritionist of the program; physical activity greater than 150 minutes per week (validated by physiotherapist through the International Physical Activity Questionnaire (IPAQ)), avoidance of alcohol and tobacco consumption, supervision and guidance for a weight loss goal of 7% in overweight or obese participants, data evaluated during the quarterly interview and recorded in institutional medical records).

In addition, quarterly control of fasting glucose, HbA1c, total cholesterol, triglycerides (TG), HDL (high density cholesterol), LDL (low density cholesterol), renal function and blood pressure, anthropometric measurements with calculation of BMI. Calculation of the glucose/triglyceride index (as a marker for prediabetes diagnosis, in absence of data to calculate HOMA-IR, considering its relationship with diagnosis of insulin resistance) [36, 37].

$$\text{Glucose/Triglyceride Index = Ln ((Triglyceride (mg/dL)}\times \text{Glucose (mg/dL)) /2)}$$

Clinical definitions

• Prediabetes diagnosis and classification: patients with fasting glucose intolerance, defined by fasting plasma glucose ≥ 100 mg/dl or HbA1c ≥ 5.7% or both [7, 8, 22, 26, 29, 30, 38].

• Diabetes diagnosis [31]: At follow-up time, diagnosis of T2DM was established in those with two repeated fasting plasma glucose tests > 125 mg/dl or HbA1c > 6.4%.

• Reversion to normoglycaemia: Patients with return to normoglycaemia were defined as those with fasting plasma glucose < 100 mg/dl and HbA1c < 5.7%.

Sample

A total of 130 available patient records were analyzed; therefore, neither sample calculation nor sampling strategies were used. Being the total number of records evaluated, a power greater than 99.9% was calculated.

Data collection techniques

An electronic database under custody of DIME Neurocardiovascular’s Clinical, cardiovascular risk program was used.

Source of information

Database of medical history records containing general data, dates of care, assessments by cardiovascular risk team, identification with demographic data, and serum laboratory reports (fasting glucose, HbA1c, total cholesterol, TG, HDL, LDL) and anthropometric measurements (weight and BMI), blood pressure numbers, performed on the study participants.

Data management and statistical analysis

All data were collected in Microsoft Office Excel® 2022 and analyzed using the statistical package Stata 17 TM (Stata Corp, Collage Station, TX, USA). Characteristics of the study population were summarized using descriptive statistics. To determine distribution of numerical variables, univariate analysis was performed. Normality hypothesis was tested using Kolmogorov-Smirnov test, rejecting normality null hypothesis and assuming that variables have a distribution different from normal when the p value was less than 0.05.

For those variables with normal or parametric distribution, data were summarized with their mean and standard deviation (SD), and for those with non-normal distribution, median and interquartile range (IQR) were used. Categorical variables were summarized as proportions in a frequency table.

Population was divided into patients with reversion from prediabetes to normoglycemia and patients with progression to T2DM and possible factors related to admission to Cardiovascular Risk Program were evaluated by bivariate analysis with data obtained at admission to the Cardiovascular Risk Program for each outcome. Comparison between the groups was performed taking p < 0.05 as significant. Fisher or Chi2 test was used for categorical variables and T-test or U-Mann-Withney for continuous variables, with a significant value of p < 0.05 (OR: odds ratio; 95% confidence interval (CI)).

Some continuous variables were categorized, as reported in literature, as ranges that behaved as risk factors for progression to T2DM, in previous population-based studies [4, 5, 39] and could have pathophysiological implications limiting reversion to normoglycemia. Variables categorized were age (range < 60 years - ≥ 60 years), schooling (range ≤ 11 years - > 11 years), BMI (range < 25 kg/m2 - ≥ 25 kg/m), fasting glucose (range < 110 mg/dl- ≥ 110 mg/dl), HbA1c (range < 6.0% - ≥ 6.0%).

For control of the possible biases detected (information and confounding), data were analyzed using descriptive statistics with subsequent bivariate and multivariate analysis; significant variables were included in a logistic regression in order to establish a probable association. Initial or saturated model was constructed with those variables that in bivariate analysis reported p-values less than or equal to 0.25 and through the backwards strategy, according to the statistical likelihood test, the most parsimonious model was chosen. We considered it significant if variable had a p-value < 0.05 (OR: adjusted odds ratio; 95% CI). To determine performance of the most parsimonious model, ROC (receiver operating characteristic curve) analyses was used.

Results

Clinical-demographic characteristics

During this evaluation period, 183 patients were attended in DIME Cardiovascular Risk Program, of 137 with a diagnosis of prediabetes; 5.4% (n= 7) were excluded due to incomplete data (n= 2), follow-up time of less than six months (n= 3), and failure to comply with quarterly check-ups by the multidisciplinary team (n= 2). Records of 130 patients who met the inclusion criteria were analyzed (Fig. 1).

Fig. 1.

Flow chart of prediabetic population evaluated to determine factors related to reversal of prediabetes and progression to type 2 diabetes mellitus. T2DM: Type 2 Diabetes Melllitus

Table 1, summarizes clinical-demographic characteristics; mean age was 69.5 years (SD 10.1), 50.8% were women, 74.6% had more than high school education. Regarding clinical characteristics, median HbA1c was 5.9% (IQR 5.7–6.2), BMI presented a mean of 26.7 (SD 3.5), 57.7% of patients reported more than 150 minutes of physical activity per week, 80.7% were hypertensive and 47.7% were overweight. Median follow-up time was 366 days (IQR 249–569).

Table 1.

Clinical-demographic characteristics of population evaluated

Characteristics	Description	Summary measure
		n: 130	%
Age/Years	Median	69.5	SD 10.1
Gender	Male	64	49.2
	Female	66	50.8
Education	Primary	7	5.4
	High school	26	20
	Technician	17	13.1
	University	58	44.6
	Postgraduate	22	16.9
Weight/Kg	Median	74	SD 12.3
BMI/Kg/m2	Mean	26.7	SD 3.5
HbA1c/%	Median	5.9	IQR 5.7–6.2
Fasting glucose/mg/dL	Median	101.8	IQR 97–107
Triglycerides/mg/dL	Median	125	IQR 95–161
Glucose/Triglycerides Index	Median	6.51	IQR 4.8–8.7
Smoking	Yes	6	4.6
	Not	124	95.4
Physical activity greater than 150 minutes per week	Yes	75	57.7
	Not	55	42.3
Arterial hypertension	Yes	105	80.8
	Not	25	19.2
Previous CVDs	Yes	50	38.5
	Not	80	61.5
Dyslipidemia	Yes	92	70.8
	Not	38	29.2
Obesity	Yes	25	19.2
	Not	105	80.8
Overweight	Yes	62	47.7
	Not	68	52.3
OSAHS	Yes	7	5.4
	Not	123	94.6
Hypothyroidism	Yes	18	13.8
	Not	112	86.2
Hepatic steatosis	Yes	12	9.2
	Not	118	90.8
Follow-up time/Days	Median	366	IQR 249–569

IQR Interquartile range, SD Standard deviation, BMI body mass index, HbA1c Glycated hemoglobin, CVDs cardio-cerebrovascular disease, OSAH Obstructive Sleep Apnea Hypopnea Syndrome

Analysis according reversion to normoglycemia

In bivariate analysis (Table 2), 21.5% reversion to normoglycemia was evident. Age ≥ 60 years and BMI ≥ 25 were found to decrease probability of reversion by 59% and 67%, respectively. Another related factor decreased chance of reversion was HbA1c ≥ 6.0% (74% less chance of reversion relative to participants with HbA1c < 6.0%).

Table 2.

Bivariate analysis according reversion to normoglycemia

Characteristics at baseline	Description	Total	Reversion to normo-glycemia n= 28	No reversion			OR	CI 95%	P Value
				Prediabetes permanence n= 84	Progression to T2MD n= 18	Total n= 102
Age/Years	≥ 60	94	16	63	15	78	0.41	0.17–0.98	0.026
	< 60	36	12	21	3	24
Gender	Male	64	13	41	10	51	0.86	0.37–2.00	0.372
	Female	66	15	43	8	51
Education/years of training	> 11 years	33	19	68	10	78	0.65	0.31–1.71	0.183
	< 11 yearsa	97	9	16	8	24
BMI/Kg/m2	≥ 25	83	12	58	13	71	0.33	0.13–0.78	0.005
	< 25	47	16	26	5	31
HbA1c/%	≥ 6.0	51	5	37	9	46	0.26	0.08–0.72	0.004
	< 6.0	79	23	47	9	56
Fasting glucose/mg/dL	≥ 110	20	2	14	4	18	0.36	0.05–1.47	0.089
	< 110	110	26	70	14	84
Triglycerides/mg/dL	≥ 150	46	9	28	9	37	0.83	0.32–2.01	0.350
	< 150	84	19	56	9	65
Glucose/Triglycerides Index	Median (IQR)	6.89 (4.8–8.7)	5.33 (4.1–6.5)	6.2 (5.1–7.3)	7.4 (6.3–8.4)	6.9 (6.5–8.2)	0.34	0.12–0.57	0.031
Smoking	Yes	6	1	3	2	5	0.72	0.02–5.47	0.421
	Not	124	27	81	16	97
Physical activity > 150 minutes per week	Yes	75	22	52	1	53	3.36	1.29–9.75	0.005
	Not	55	6	32	17	49
Arterial hypertension	Yes	105	23	68	14	82	1.01	0.39–3.67	0.432
	Not	25	5	16	4	20
Previous CVDs	Yes	50	8	30	12	42	0.57	0.21–1.41	0.117
	Not	80	20	54	6	60
Dyslipidemia	Yes	92	18	59	15	74	0.68	0.28–1.71	0.202
	Not	38	10	25	3	28
Obesity	Yes	25	5	15	5	20	0.89	0.27–2.56	0.432
	Not	105	23	69	13	82
Overweight	Yes	62	13	43	6	49	0.94	0.39–2.19	0.442
	Not	68	15	41	12	53
OSAHS	Yes	7	1	5	1	6	0.59	0.02–4.26	0.353
	Not	123	27	79	17	96
Hypothyroidism	Yes	18	1	16	1	17	0.19	0.01–1.11	0.035
	Not	112	27	68	17	85
Hepatic steatosis	Yes	12	2	8	2	10	0.70	0.11–3.14	0.361
	Not	118	26	76	16	92
Follow-up time/Days	Median (IQR)	366 (249–569)	369 (234–541)	364 (248–564)	413 (242–588)	364 (248–564)	1	0.99–1	0.550

BMI body mass index, HbA1c Glycated hemoglobin, CVDs cardio-cerebrovascular disease, OSAH Obstructive Sleep Apnea Hypopnea Syndrome, IQR Interquartile range

^aEducation < 11 years: refers to primary - high school

It was also observed that performing physical activity > 150 minutes/week increases 3.36 times the chance of reversion to normoglycemia.

Factors maintaining significance after adjustment in the multivariate analysis (Table 3), and which continued decreased chance of reversion, were BMI ≥ 25 and HbA1c ≥ 6.0. Attention should be drawn to the fact that when adjusted, physical activity was statistically significant, increasing opportunity of reversion to normoglycemia by 4.15 times. Performance of the proposed model to explain prediabetes reversion through the set of variables, reported that 77% of data were correctly classified according ROC (Fig. 2).

Table 3.

Multivariate analysis for reversion to normoglycemia

Characteristics at baseline	Description	Total	Reversion to normo-glycemia n= 28	No reversion			ORa	CI 95%	P Value
				Prediabetes permanence n= 84	Progression to T2MD n= 18	Total n= 102
BMI/Kg/m2	≥ 25	83	12	58	13	71	0.25	0.12–0.54	0.003
	< 25	47	16	26	5	31
HbA1c/%	≥ 6.0	51	5	37	9	46	0.14	0.11–0.63	0.014
	< 6.0	79	23	47	9	56
Physical activity > 150 minutes per week	Yes	75	22	52	1	53	4.15	1.62–5.01	< 0.001
	Not	55	6	32	17	49

^aOR Adjusted Odds Ratio, BMI body mass index, HbA1c Glycated hemoglobin

Fig. 2.

ROC for multivariate analysis. AUC: area under curve

Discussion

The present study is pioneering in southwestern Colombia and presents factors related to the reversion of prediabetes in the prediabetic population who underwent multidisciplinary follow-up in a structured cardiovascular risk program. Patients in this cohort received periodic assessments by specialists in Internal Medicine, Nutrition, Psychology, and Physiotherapy, and underwent clinical and metabolic follow-up, including the assessment of insulin resistance using the glucose/triglyceride index. Participants in this study received evaluations similar to those reported by Galavis et al. [3], Who found in 14 studies that participants who received group education from healthcare professionals in a follow-up program were 33% less likely to develop diabetes than control participants.

As described by Echouffo et al [40], strategies implemented for the detection and management of prediabetes through follow-up programs, educational interventions, multidisciplinary follow-up, weight loss plans, nutritional recommendations, and physical activity have been shown to reduce the burden of disease if followed over time. These strategies are fundamental within primary care plans and have significant implications in the health-disease process, making them an essential pillar or first-line approach [2]. Prediabetes is one of preventable and therefore controllable pathologies, in which non-pharmacological initiatives are cost-effective, achieving public health relevance [41–43]. Factors related to prediabetes and non-pharmacological management were measured in individuals of this cohort, in a program that implements interventions aimed at health promotion and disease prevention. The follow-up needs to be extended, evaluating related outcomes and measuring quality of life, in order to determine the impact of programs like this one, which could be standardized within treatment strategies.

Individuals aged 60 or older had a lower chance of reversion in this study compared to those younger than 60 years of age. However, this has been documented before, with a 22% reversion rate at 12 years for those older than 60 years, and a reversion rate of 2.2/100 person-years, associated with lower blood pressure, weight, and no history of heart disease [44]; in turn, the percentage decreases to 13% in those over 70 years of age, with follow-up using HbA1c [20].

Bennasar et al. conducted a cohort study with 22,293 Spanish workers diagnosed with prediabetes based on fasting glucose, with a mean age of 45 years, and a five-year follow-up. They observed a 40% reversal with lifestyle changes, similar to what was found in this cohort, where both BMI and higher HbA1c were associated with greater persistence of prediabetes [45]. In this study, patients whose care process was guided toward lifestyle modifications showed that, after six months of follow-up (median time 366 days, IQR 249–569), 21.5% of patients achieved normoglycemia. This highlights the need to extend the follow-up period and increase the sample size to continue implementing measures that allow for further evaluation, improvement, and reinforcement of the program.

Recent reports suggest that the reversal of prediabetes may prevent cardiovascular outcomes and death [32]; however, evidence is lacking to establish this. So far, reports of reversion to normoglycemia are highly heterogeneous, ranging from 18% to 59% at five years and decreasing to a maximum of 42% during 6–11 years of follow-up in one Cochrane meta-analysis [8]. One of the strongest reasons for divergence in statistics is likely the method of measurement. Vistisen et al. found a higher percentage of prediabetes reversal at five years using less precise methods, such as fasting glucose and a two-hour post-glucose load tolerance test (45% and 37%, respectively), compared to using HbA1c as a reversal parameter (17%), a reliable marker of glycemic behavior over the preceding three months, which has been shown to be a better predictor of cardiometabolic risk [21]. Consistent with these data, this study showed statistical significance in the multivariate model, indicating that an HbA1c ≥ 6.0 is less likely to revert to normoglycemia.

In addition, it was shown that a higher glucose/triglyceride index was associated with a lower likelihood of reversion to prediabetes. The glucose/triglyceride index is known to be a useful marker for determining insulin resistance, which could predict the risk of progression to T2DM [36, 37, 46]. This could represent a cost-effective strategy, given the availability of lipid profile and fasting glucose monitoring programs, which would enable the calculation of the glucose/triglyceride index to predict risks and implement strategies for their reduction, even in the absence of HOMA-IR. Further studies with larger sample sizes could validate these findings and facilitate the standardization of its use.

A recent meta-analysis of 27 studies examining lifestyle modification found an increased probability of reversing prediabetes compared to the control group, with a relative risk (RR) of 1.76 (95% CI 1.41–2.19) and a risk difference of 18% [12]. Cohort studies have highlighted the importance of measuring exercise intensity weekly and conducting follow-up through physiotherapy and sports medicine, using validated scales such as the IPAQ [27, 32, 47, 48]. Identifying factors that prevent the reversal of prediabetes and addressing modifiable ones could enhance the impact of interventions. In this cohort, we found that engaging in > 150 minutes of physical activity per week increases the likelihood of reversion by 3.36 times, compared to those who did not engage in such activity.

A systematic review of 103 studies [43] in middle-aged adults found that the cumulative incidence of diabetes after 6 years of follow-up was 17% among those with prediabetes based on HbA1c levels of 5.7%–6.4%, and 22% among those with fasting glucose levels of 100–125 mg/dL. In this population, 13.8% progressed to T2DM, with a median follow-up time of 413 days (IQR 242–588). In addition to hyperglycemia in the intermediate range, patients in this study, at baseline, had several risk factors for developing prediabetes and its progression to T2DM. For example, 72% were at least 60 years old, 67% were overweight or obese, 81% were receiving antihypertensive treatment, and 42% did not meet the recommended physical activity guidelines of at least 150 minutes per week. These findings suggest that controlling modifiable risk factors is crucial in this population, as emphasized in the literature [10, 42, 49].

Limitations of the study include the retrospective nature of the data analyzed and the relatively short follow-up period of less than two years. However, strengths of the study include the inclusion of all available records for review and the control of confounding factors through both bivariate and multivariate statistical analyses.

Conclusions

Studies investigating factors related to prediabetes and its reversal can contribute to the development of strategies from a public health perspective aimed at reducing the disease burden. One such strategy involves monitoring through cardiovascular risk programs, which offer a structured approach to track and implement plans that educate individuals about healthy eating, physical activity (≥ 150 minutes per week), weight management, reduction in HbA1c levels, and the promotion of health through primary care. The use of the glucose/triglyceride index as a marker of insulin resistance could reduce costs, enhance metabolic monitoring, and improve control, potentially decreasing unfavorable outcomes.

Authors’ contributions

Conceptualization: WARM, MECV, AMSS, LFT. Methodology: RP, AMSS, YBM, DMSM. Data curation: WARM, MECV, AMSS, LFT. Formal analysis: WARM, MECV, AMSS, LFT. Writing-original draft: WARM, MECV, AMSS, LFT, RP, DMSM, YBM. Writing-review and editing: WARM, MECV, AMSS, LFT, RP, DMSM, YBM. All authors read and approved the final version.

Funding

The authors declare that they have not received funding for this research.

Data availability

The authors confirm that all data underlying the findings are fully available without restriction. All relevant data is in the present manuscript.

Declarations

Ethics approval and consent to participate

This research was approved by the ethics committee of DIME Neurocardiovascular’s Clinic, and by Research Committee of the Universidad Libre (Act number 012 of February 23, 2023) as a research without risk according to the classification of research risk declared in Resolution 008430 of 1993 of the Colombian Ministry of Social Protection. The approval of the institutional ethics committee included approval for the use of the data and according to the guidelines of the Helsinki declaration.

Consent for publication

All authors assert that there are no undisclosed conflicts of interest (both personal and institutional) regarding specific financial interests that are relevant to the work conducted or reported in this manuscript.

Competing interests

The authors declare no competing interests.