Sex Disparities in Prehospital and In-Hospital Delays to Intravenous Thrombolysis in Egyptian Stroke Patients

Abstract

Introduction and Aim of the Work

Numerous studies have shown contradictory data regarding sex differences in pre- and in-hospital delays of intravenous thrombolysis (IVT) for acute ischemic stroke (AIS). Our objective was to investigate these differences among Egyptian patients presenting with AIS.

Methods

This prospective, single-center, open-label cohort study was conducted from September 2022 to December 2023. Patients with IVT-eligible AIS, who did not receive IVT, were included in the study. Patients were recruited from the largest tertiary stroke center in the south of Egypt. All participants underwent a full neurological assessment including NIHSS, laboratory workup, and computed tomography or magnetic resonance imaging. The patients were subjected to a structured questionnaire designed to determine the parameters and time metrics associated with prehospital and in-hospital delays.

Results

A total of 897 patients, 41.8% of whom were females, were included in our study. General demographic characteristics were similar between both groups. A total of 70.9% of female patients and 65.1% of male patients arrived at stroke-ready hospital outside the therapeutic time window. Female patients experienced a significantly longer duration from symptom onset to hospital arrival (600 ± 618 min) compared to male patients (480 ± 438 min). Stroke misdiagnosis was nonsignificantly higher among the female group (58.9%). The time spent in the hospital did not significantly differ between males and females.

Conclusion

The study results indicate a significant disparity in the management of female stroke patients in both prehospital and in-hospital settings, with the most pronounced delays occurring during prehospital care.

Plain Language Summary

Sex differences in stroke care are well documented in high-income countries, but limited data exist from low- and middle-income countries (LMICs), particularly from Africa. This is the first study from Egypt and Africa to examine sex differences in both prehospital and in-hospital delays among patients with acute ischemic stroke who were eligible for intravenous thrombolysis (IVT) but did not receive it due to time delays. We studied 897 stroke patients eligible for IVT, but who did not receive it at the largest tertiary stroke center in southern Egypt between September 2022 and December 2023. Of these patients, 41.8% were women. The results showed that women experienced significantly longer delays from symptom onset to hospital arrival (average 600 min) compared to men (480 min). A greater proportion of women also arrived too late to receive IVT. Misdiagnosis was more common among women and was a key factor contributing to the delay. No major sex differences were found in in-hospital processes. The use of emergency medical services (EMS) was associated with shorter delays for both sexes. These findings highlight the urgent need for public health campaigns to improve stroke recognition among women and for training healthcare providers to better identify atypical stroke presentations in female patients. Promoting greater use of EMS may also help reduce prehospital delays. The study highlights the importance of addressing sex-specific barriers in stroke care and calls for further research into targeted interventions to improve timely access to treatment for women in LMIC settings.

Introduction

Egypt has one of the highest incidences of stroke in the Middle East region. Epidemiological studies have described the south of Egypt, the Assiut Governorate, as having the highest crude prevalence rate of stroke at 963 per 100,000 individuals [1, 2]. Time is a key determining factor for acute ischemic stroke (AIS) treatment and clinical outcome. Therefore, the identification of the factors contributing to delays AIS management is crucial for optimizing pathways. In many cases, the patients’ level of education, financial circumstances, and cultural environment influences how they interpret stroke symptoms and how quickly they seek help. These everyday realities can play a major role in timely arrival to the hospital for acute reperfusion treatments. Evidence from the Assiut region shows that challenges like illiteracy and limited socioeconomic resources not only increase stroke burden but may also slow patients’ access to stroke care [1]. There is inconsistent information about gender inequality in AIS management. These gender discrepancies in acute stroke care are more frequent but not confined to low- and middle-income countries [3]. Some studies also from high-income countries reported a lower number of women receiving optimal prehospital management, acute stroke care, and intravenous thrombolysis (IVT) in comparison to men [4–6]. This discrepancy could be attributed to many factors like inadequate stroke knowledge among women [7], delayed contact with emergency medical services (EMS) [8], and frequent underdiagnosis of stroke among women in comparison to men [9]. So far, studies on gender disparities in prehospital and in-hospital access to acute stroke management among Egyptian and North African patients are limited and focused mainly on in-hospital factors and geographical variations [10, 11]. In this study, our objective was to investigate the disparities in prehospital and in-hospital delays between male and female Egyptian patients who were eligible for IVT for AIS but did not receive the treatment.

Materials and Methods

Study Design

From September 1, 2022, to December 31, 2023, we conducted a prospective, single-center, cohort study at the Neurology, Psychiatry, and Neurosurgery Hospital, Assiut University, located in South Egypt. This hospital has the largest tertiary stroke center in the south of Egypt, with a total capacity of 358 beds. It serves approximately 1,200 ischemic stroke patients annually from both urban and rural areas in southern Egypt.

Patients’ Selection

Consecutive AIS patients were recruited to investigate the reasons for not receiving IVT. We included all patients who did not receive reperfusion despite presenting (either via EMS or other means) with an IVT-eligible AIS. IVT-eligible AIS is defined as patients (≥18 years) with a confirmed diagnosis of AIS who presented within a guideline-recommended time window for IVT of 4.5 h after symptom onset and who had no documented contraindications to IVT [12].

The enrolled patients fulfilled the following inclusion criteria: a diagnosis of AIS, age 18 years or older, and not being treated with IVT being eligible. Patients were excluded from participation if they had a history of transient ischemic attack, hemorrhagic stroke, or contraindications to IVT [12] (Fig. 1).

Fig. 1.

Study methodology.

Data Collection

Each patient underwent detailed history taking, full neurological assessment and initial NIHSS, and mRS assessments. Computed tomography imaging or magnetic resonance imaging of the brain as well as urgent laboratory investigations were performed for all patients in the emergency department. Coagulation and vasculitis blood analysis profiles were also performed for all patients.

A full cardiac assessment was performed in all patients. Patients with suspected cardioembolic stroke also underwent a 12-lead ECG and transesophageal echocardiography. The Trial of ORG 10172 in Acute Stroke Treatment (TOAST) classification system was used to classify stroke subtypes. Demographic data were collected for all participants. Information about the preliminary diagnosis, the ability of physicians or relatives to recognize the patient’s symptoms as a stroke, and the final AIS diagnosis were documented. The type of first help-seeking behavior, and which patients or their relatives were notified were also recorded. For all participants, the following time metrics were noted for later analysis: prehospital metrics such as time of symptom onset, onset to alarm time (OAT), and time of ambulance response (time from emergency call to arrival on the scene).

Prehospital delay was defined as the interval from symptom onset (or last known well) to hospital arrival, expressed in minutes. All analyses of prehospital delay were restricted to patients, who arrived outside the thrombolysis time window, defined as >4.5 h from onset to arrival. Within this restricted cohort, we modeled prehospital delay as a continuous outcome to identify associated factors.

Onset to door time (ODT) was also calculated, which is the time from symptom onset to stroke-ready hospital door in minutes. In addition, we also recorded the type and pathway of stroke patients or means of transport used by their relatives to the hospital. In-hospital metrics were also recorded such as the time spent from hospital door to neurological examination, time spent in imaging (calculated from the time of sending the patient to imaging until imaging results), time spent until laboratory results, and time spent from arrival at the hospital to stroke unit management. All data including time metrics were obtained from our hospital’s registry, which incorporates a study-specific structured questionnaire. Clinical details including diagnoses, comorbidities, and examination findings were extracted from the patients’ medical records by the treating clinicians and entered into the registry, which served as the unified source from which all study data were subsequently retrieved.

Statistical Analysis

We investigated the possibility of sex variations in pre- and in-hospital delays. Data obtained from this study were inputted into an IBM-compatible computer. Descriptive statistics (mean [SD], numbers, and percentages) were calculated using the software package for Social Sciences (SPSS Inc., Chicago, IL, USA) for Windows, Version 25. Frequencies were noted and associations were determined using the Pearson’s chi-squared test (χ² test). Additionally, results were analyzed using an independent sample t-test that did not assume equal variances. The significance level was set at p < 0.05. A multiple logistic regression analysis was performed to identify the factors associated with prehospital delay in women with stroke. A multiple linear regression analysis was performed to identify predictors of prehospital delay among women with stroke. For each variable, unstandardized coefficients (B), standard errors (SEs), standardized beta values (β), p values, and 95% confidence intervals (CI) were reported. Statistical significance was set at p < 0.05. All model assumptions, including linearity, independence, homoscedasticity, and normality of residuals were tested and satisfied.

Results

Demographics

A total of 897 patients, 375 (41.8%) females, and 522 (58.2%) males, were included in our study. The mean age for females was 60 ± 14 years and the mean for males was 61 ± 14 years of age (Table 1). A total of 24.3% of our female group and 23% of our male group were young adults (<50 years of age). Hypertension and smoking were statistically more prevalent among males, while AF and valvular heart disease were significantly more prevalent among females. Dyslipidemia, ischemic heart disease, and diabetes mellitus were more evident among males without any significant difference. A total of 70.9% of female patients and 65.1% of male patients arrived at stroke-ready hospital outside the therapeutic time window for IVT (later than 4.5 h after symptom onset) (Table 1).

Table 1.

General demographic characteristics of the study population

​	Female (n = 375, 41.8%)	Males (n = 522, 58.2%)	p value
Age (mean±SD), years	60±14	61±14	0.79
Age ≤50 years, n (%)	91 (24.3)	120 (23)	0.65
Age >50 years, n (%)	284 (75.7)	402 (77)	​
NIHSS at admission (mean±SD)	9.8±4	9.7±4	0.72
Risk factors distribution, n (%)
Diabetes mellitus	73 (19.5)	111 (21.3)	0.51
Hypertension	100 (26.7)	176 (33.7)	0.02
Ischemic heart Disease	19 (5.1)	36 (6.9)	0.26
Atrial fibrillation and valvular heart disease	53 (14.1)	33 (6.3)	0.001
Dyslipidemia	96 (25.6)	161 (30.8)	0.08
Smoking	14 (3.7)	288 (55.2)	0.001
Vasculitis	9 (2.4)	7 (1.3)	0.24
Patients arrived at the hospital within the thrombolysis time window, n (%)	109 (29.1)	182 (34.9)	0.06
Patients arrived at the hospital outside the thrombolysis time window, n (%)	266 (70.9)	340 (65.1)	0.07

Data are described as number and percentages or mean ± standard deviation.

Pre- and In-Hospital Assessments

Both sexes tended to most often contact either EMS or relatives and friends to seek help when they experienced their strokes. There were no significant differences regarding help-seeking behaviors between both sexes (Table 2). The stroke diagnosis was more frequently overlooked by the first examiners in female patients (58.9%), though this difference was not statistically significant (Fig. 2).

Table 2.

Type of first help-seeking behaviors among study population

Different types of contact for help	Female (n = 375, 41.8%), n (%)	Males (n = 522, 58.2%), n (%)	p value
Ambulance	79 (21.8)	109 (22.4)	0.46
General practitioner	51 (14)	56 (11.5)
Internal medicine physicians	66 (18.2)	82 (16.8)
Neurologists	37 (10.2)	69 (14.2)
Relatives, friends, neighbors	88 (24.2)	114 (23.4)
ENT (ear, nose, throat physicians)	18 (5)	17 (3.5)
Ophthalmologists	7 (1.9)	12 (2.5)
Other medical specialties	16 (4.4)	22 (4.5)
Traditional healers	1 (0.3)	6 (1.2)

Patients contacted physicians in their private clinics, hospital outpatient clinics, and emergency rooms in non-stroke-ready hospitals.

Fig. 2.

Accuracy of stroke diagnosis. Fewer females than males received a correct stroke diagnosis and were misjudged as having other diseases. The difference was not significant.

Stroke mimics like epilepsy, migraine, and functional disorders were more significantly diagnosed for females than males. Misdiagnoses like cardiovascular disease and drug toxicity were more frequently observed in male group (Table 3; Fig. 3).

Table 3.

Accuracy of stroke diagnosis variations among both sexes

Initial diagnosis of the patients	Female (n = 375, 41.8%)	Males (n = 522, 58.2%)	p value
Recognize incident as stroke, n (%)	154 (41.1)	248 (47.6)	0.052
Recognize the incident as non-stroke-related, n (%)	221 (58.9)	273 (52.4)
Common stroke misdiagnosis, n (%)	​	​	0.001
Cardiovascular disorder	19 (5.1)	43 (8.2)	​
Diabetic coma	23 (6.1)	31 (5.9)	​
Ear, nose, and throat disorders	21 (5.6)	39 (7.5)	​
Ophthalmic disorder	9 (2.4)	18 (3.4)	​
Gastrointestinal disorders	2 (0.5)	4 (0.8)	​
Coma for further investigations	78 (21)	99 (19)	​
Polydrug toxicity	1 (0.3)	9 (1.7)	​
Functional disorder	27 (7.2)	7 (1.3)	​
Migraine	13 (3.5)	6 (1.1)	​
Epilepsy	15 (4)	6 (1.1)	​
Other diagnosis	13 (3.5)	11 (2.1)	​

Fig. 3.

Stroke misdiagnosis. The most common misdiagnosis was unclear coma. Significantly more females got misdiagnosed with functional disorders, migraine, or epilepsy, while cardiovascular diseases and intoxications were more common in males.

Time Metrics

The symptom onset to alarm time (OAT) was similar between both sexes (Fig. 4). However, subgroup analysis showed that the OAT was nonsignificantly longer in the older female group (>50 years of age) than in the older male group, while young women showed a tendency to have a shorter OAT than young men. The time of ambulance response (from notification to arrival at the patient’s site) was also similar between females and males (33 ± 11 minutes compared to 33 ± 25) minutes. The time spent from symptom onset to stroke-ready hospital arrival (onset to door time [ODT]) was significantly longer among female patients (600 ± 618 min compared to 480 ± 438 min among men, p value: 0.001). Further subgroup analysis revealed that the observed significant differences in ODT were primarily attributed to older patients. Specifically, the ODT for older female patients was 618 ± 624 min, significantly longer than the male group’s ODT of 450 ± 360 min (Table 4, Fig. 4). When examining in-hospital metrics, such as the time spent from ER arrival to being examined by a neurologist and the time spent in imaging, there was minimal disparity between males and females (Table 4; Fig. 4). Additionally, the time spent on laboratory investigations and the time spent from hospital arrival to stroke unit management exhibited a nonsignificant, slightly longer duration for females (Table 4; Fig. 4).

Fig. 4.

Stroke care time metrics. The time from symptoms onset to stroke-ready hospital arrival (onset to door time) was significantly longer for female patients compared to males. There was a trend toward longer in-hospital times for females without reaching significance.

Table 4.

Time metrics

​	Female (n = 375, 41.8%)	Males (n = 522, 58.2%)	p value
Symptom onset to alarm time (OAT) (mean±SD), min	198±252	198±156	0.80
Time of ambulance response (time from emergency call to on scene arrival (mean±SD), min	32±11	33±25	0.71
Time from symptoms onset to stroke-ready hospital arrival (onset to door, ODT) (mean±SD), min	600±618	480±438	0.001
Time from stroke-ready hospital arrival – meeting neurologist (mean±SD), min	15±4	15±8	0.79
Time spent for hospital laboratory investigations (mean±SD), min	107±69	98±66	0.32
Time spent in imagining (mean±SD), min	70±25	69±29	0.61
Time from hospital arrival – stroke unit management (mean±SD), min	87±27	83±29	0.30

Factors Affecting Prehospital Delay among Women

Several factors affected prehospital delay in women with stroke (Table 5). The most significant contributor was misdiagnosis (B = 4.201, p < 0.001), which led to considerable delays in hospital arrival. A history of cardiac disease including ischemic heart disease and atrial fibrillation was also linked to prolonged delays (B = 5.542, p = 0.010), possibly due to symptom overlap complicating early assessment. In contrast, ambulance transport significantly reduced prehospital delay (B = −3.052, p = 0.006), emphasizing the importance of EMS in improving response time. Additionally, women with diabetes experienced shorter delays (B = −2.811, p = 0.013). Other factors, including patient's age, hypertension, and initial evaluation by a neurologist rather than emergency physicians from referring non-stroke-ready hospitals, did not have a significant impact on prehospital delay. The NIHSS score showed a borderline effect (B = −0.184, p = 0.052), suggesting that more severe strokes may lead to shorter delays, though the relationship was not conclusive (Table 5).

Table 5.

Linear regression analysis for predictors of prehospital delay among women with stroke

Variable	Unstandardized regression coefficient (B value)	Standard error	β	p value	95% CI (lower–upper)
Stroke misdiagnosis	4.20	0.93	0.25	0.00	2.36 to 6.04
Known history of cardiac troubles	5.54	2.14	0.13	0.01	1.32 to 9.77
Ambulance transport	−3.05	1.12	−0.15	0.01	−5.23 to −0.87
Diabetes	−2.81	1.13	−0.14	0.01	−5.04 to −0.59
NIHSS score	−0.18	0.09	−0.11	0.05	−0.37 to 0.00
Hypertension	−1.37	1.06	−0.07	0.20	−3.47 to 0.72
Age	0.79	1.05	0.04	0.45	−1.27 to 2.87
Referred by non-neurologists vs. neurologists from non-stroke-ready hospitals	−0.71	1.59	−0.02	0.66	−3.84 to 2.42

Bold values are those below <0.05, which indicates statistical significance.

DM, diabetes mellitus; NIHSS, National Institutes of Health Stroke Scale; SE, standard error; β, standardized coefficient; CI, confidence interval.

Discussion

Our study has shown a lower number (41.8%) of recruited female stroke patients than males, with a ratio of 1.4:1. Our results were consistent with many Egyptian studies that found stroke to be more prevalent in males than females [1, 2, 10]. However, the lower prevalence rate at the hospital level in our results may also reflect a real neglect of stroke symptoms among female patients.

Both genders showed almost matched mean age for stroke with a similar percentage of stroke in patients younger than 50 years of age. However, in general, our stroke patients showed a younger age than stroke patients in high-income countries (HIC) [13, 14], but were similar to many Egyptian and Middle Eastern studies [15–17]. This younger age group in our study may be attributed to poor control and follow-up of modifiable-risk stroke factors in low- and middle-income countries in comparison to HIC.

Hypertension, dyslipidemia, diabetes, AF, and valvular heart disease are the most common risk factors among the female group. Our female stroke patients showed a significantly higher prevalence of AF and valvular heart disease. This may be due to a higher frequency of rheumatic heart disease (RHD) leading to valvular atrial fibrillation. this may contribute to the younger stroke onset observed in women compared to men, although the age difference was not statistically significant according to the Egyptian Ministry of Health’s 2018 report, Egypt has nearly 300,000 RHD patients aged between 5 and 15 years annually. RHD is more prevalent among females than males [18]. The male group showed a significantly higher prevalence of smoking and hypertension, and a nonsignificantly higher prevalence of dyslipidemia, ischemic heart disease, and diabetes. The higher prevalence of hypertension among the male group is consistent with many publications which have shown that men have a higher incidence of uncontrolled hypertension than women [19].

Our stroke patients tended to alarm relatives or neighbors, ambulances, general practitioners (GPs), internal medicine doctors, neurologists, and other medical specialties, respectively. Our results showed no significant differences between either group. Two Egyptian studies [10, 20] identified that only 20–26% of stroke patients in Egypt used ambulances to reach the hospital, a finding consistent with our data. In line with our results, data coming from the Netherlands showed that patients also alerted relatives or friends, GPs, alongside ambulances, respectively [21].

Among female stroke patients, 58.9% were initially misdiagnosed by the first examiner, a rate that was not significantly higher than the 52.4% observed in male patients. Women were more likely to be misdiagnosed as stroke mimics and having functional disorders, while males were more frequently misdiagnosed as having cardiovascular diseases and polydrug toxicity.

The higher percentage of misdiagnosis among females is well documented in the literature. This phenomenon can be attributed to the more frequent presentation of non-localizing and atypical stroke symptoms in females, coupled with the limitations of diagnostic tools in emergency settings and examinations conducted by non-neurologists or non-stroke specialists. Additionally, patients often sought medical advice directly from various specialists in private hospitals after their symptoms were initially misjudged [9, 22–24].

Time spent from symptom onset to hospital arrival (onset to door) was significantly higher in the female group, resulting in a delayed hospital arrival that affected potential IVT treatment options. This delay was more frequent in females (70.9%) than in males (65.1%). Our results were consistent with many studies conducted in Africa, the Middle East, and Western countries, which also showed that women have a longer onset to door times [25–28]. In contrast to our findings, one Egyptian study showed that women had shorter onset to door times and prehospital time metrics compared to men [29]. These different results may be explained by the fact that all patients in that study were recruited from a specialized stroke unit in the capital of Egypt, where only urban patients receive the best acute stroke services. Therefore, the results may not be representative.

Misdiagnosis was the most significant factor contributing to prehospital delay, emphasizing the need for improved stroke recognition especially for women, who more frequently present with atypical symptoms than men [25]. Another factor that may contribute to prehospital delay is whether stroke onset was witnessed or not. If the stroke onset was witnessed, it was most commonly identified by the patient’s partner or family members. Although a higher proportion of women in our sample were not employed, the reported help-seeking behaviors did not demonstrate meaningful sex differences. Nonetheless, the availability of family members at the time of symptom onset, and broader social support dynamics may influence the timeliness of alarm activation. These contextual factors were not directly assessed in this study and warrant further investigation in research specifically designed to evaluate their contribution to prehospital delay. Patients with a history of cardiac diseases also experienced significantly longer delay, possibly due to the stroke being misdiagnosed as a cardiac event or EMS personnel prioritizing the exclusion of cardiac issues, thereby delaying stroke care. In contrast, ambulance transport and diabetes were associated with significantly shorter delays, highlighting the importance of EMS utilization and healthcare accessibility. These findings emphasize the need for targeted interventions, such as enhancing stroke diagnostic accuracy and promoting emergency ambulance use, to minimize delays in stroke care.

In our study, patients who were presented within the therapeutic time window of both genders were treated almost equally in terms of ER management, imaging, and laboratory protocol. This means there was no obvious in-hospital delay.

The study results reflect a clear discrepancy between prehospital and in-hospital management of female stroke patients. The most significant delay was the time from symptom onset to hospital arrival. This delay implies a longer on scene management time and a profound lack of knowledge and awareness regarding stroke presentations among female patients in our study population.

Our findings have important implications for clinical practice in Egypt and in similar low- and middle-income settings. The substantial prehospital delay driven largely by misdiagnosis, limited EMS activation, and reliance on informal help-seeking methods highlights the urgent need to strengthen public stroke awareness. Targeted community education campaigns emphasizing recognition of stroke symptoms and the importance of calling EMS may help reduce delays, particularly among women who more frequently present with atypical symptoms. At the provider level, expanding stroke training for non-neurologists, GPs, and outpatient clinicians is critical, as many patients initially seek care in these settings. Implementing standardized triage tools, such as simple stroke screening checklists, in outpatient and private practice environments could improve early recognition and timely referral. From a system's perspective, improving EMS availability and public trust is essential. Developing coordinated stroke networks and expanding the number of stroke-ready hospitals would facilitate faster patient sending and earlier treatment. Beyond Egypt, our findings underscore the need for integrated prehospital systems, public education, and targeted provider training within national stroke strategies.

Our study has several limitations. First, it was conducted in Assiut Governorate, which is located in the southern area of Egypt and has a distinct geoculture compared to the northern part of the country. Additionally, our study did not collect specific educational, socioeconomic indicators of the patients or their next of kin as measurable variables for assessing prehospital delay.

Conclusion

Data from our study showed a significant delay from the onset of symptoms to hospital arrival among female stroke patients. These time disadvantages reduce their chances of an excellent outcome. Future work should explore how educational level, socioeconomic factors, and caregiving responsibilities contribute to these sex-based differences in prehospital delay. Interventions such as public awareness campaigns emphasizing atypical stroke symptoms in women, enhanced stroke training for non-neurologists, and the use of simple screening tools in outpatient and primary care settings may help reduce diagnostic delays. Increasing public awareness and adapting prehospital management to account for sex-specific presentation patterns are essential steps toward minimizing these disparities.

Statement of Ethics

The study protocol was approved by the Ethical Committee of the Faculty of Medicine, Assiut University, Egypt (IRB:042023300151), and followed the principles described in the Helsinki Declaration. Written informed consent was taken from the participants or their family members before their participation.

Conflict of Interest Statement

The authors have no conflicts of interest to declare.

Funding Sources

No funding was received for the conduct of the study or analysis of the results.

Author Contributions

A.N. and F.A.-A.: designed the research idea, collected data, performed analysis, and drafted the manuscript. E.K.: collected data and critical revision of the manuscript. S.W.: contributed data and analysis tools and drafted the manuscript. A.E.A.: collected data. V.C.: contributed data and analysis tools and critical revision of the manuscript.

Funding Statement

No funding was received for the conduct of the study or analysis of the results.

Data Availability Statement

The datasets analyzed during the current work are not publicly available due to institutional data sharing regulations. Data are available from the corresponding author upon request and with permission of the relevant Institutional Review Board.