Abstract
Objective:
This cross-sectional study investigates the prevalence and clinical outcomes of upper gastrointestinal bleeding (UGIB) among patients at a tertiary care center. It focuses on esophageal varices and their association with hepatitis C and B infections.
Methods:
A total of 383 patients aged 18 and above who presented with UGIB were included. Data were collected from the hospital’s endoscopy department and analyzed using Statistical Package for the Social Sciences (SPSS) version 25 and R package. Key variables included age, gender, endoscopic findings, and the presence of esophageal varices. Fisher’s exact test was used to evaluate the association between hepatitis infections and variceal bleeding. The statistical significance cut-off was P <0.05.
Results:
The mean age of the 383 patients was 56 ± 11 years, with a male predominance (55.4%). Esophageal varices were the most common finding (78.3%) and were strongly associated with hepatitis C (P < 0.001), with 70.7% of hepatitis C patients presenting with varices. Hepatitis B had a much weaker association, with only 1.7% of patients with varices also testing positive for hepatitis B. Acid peptic disease, present in 11% of cases, was more common in females. The statistical analysis confirmed a significant relationship between hepatitis C and esophageal varices. The study highlights the need for routine variceal screening and hepatitis C management in high-risk populations.
Conclusion:
Esophageal varices are the leading cause of UGIB in this population, particularly in patients with chronic hepatitis C. The findings underscore the importance of early intervention and management of hepatitis C to reduce the incidence and mortality associated with variceal bleeding in chronic liver disease patients.
Keywords: endoscopy, hepatitis C, tertiary care centers, upper gastrointestinal bleeding, variceal bleeding
Introduction
Upper gastrointestinal bleeding (UGIB) refers to bleeding from the esophagus to the duodenojejunal flexure, primarily supplied by the celiac trunk. Key symptoms, such as hematemesis and melena, typically indicate bleeding from a source proximal to the ligament of Treitz[1]. While UGIB is more common in older populations, it is also prevalent in those with liver disease or bleeding disorders. Endoscopy is crucial for diagnosing UGIB, providing direct visualization of the gastrointestinal tract. In underdeveloped regions, esophageal varices are often the leading cause of UGIB[2]. A major cause of cirrhosis-related variceal bleeding is the direct consequence of portal hypertension. The primary cause of portal hypertension in cirrhosis is an increase in intrahepatic vascular resistance, which occurs due to massive structural changes associated with fibrosis and increased vascular tone in the hepatic microcirculation. As portal hypertension progresses, the formation of collateral vessels and arterial vasodilation leads to increased blood flow to the portal circulation. This eventually results in the development of a hyperdynamic circulatory syndrome, contributing to the formation of esophageal varices or ascites[3]. Other common causes of upper gastrointestinal bleeding include gastroesophageal varices, Mallory-Weiss tears, malignancies, and portal hypertensive gastropathy (PHG). Common symptoms of UGIB include melena, hematemesis, nausea, and abdominal pain[1,4].
HIGHLIGHTS
Esophageal varices are a leading cause of upper gastrointestinal bleeding, especially in patients with chronic liver disease and portal hypertension, with hepatitis C being a significant risk factor.
This study demonstrates a strong association between hepatitis C and variceal bleeding in a Pakistani cohort, with 70.7% of hepatitis C patients presenting with varices, while hepatitis B had a much weaker association (1.7%).
Routine screening for esophageal varices and early management of hepatitis C are crucial to reducing UGIB incidence and improving outcomes in high-risk populations.
Globally, acid peptic disease, particularly peptic ulcers, remains the most common cause of UGIB[5]. Research indicates that nearly half of UGIB cases in Europe are attributed to peptic ulcers, primarily involving non-variceal ulcers[6]. Similarly, in Iran, peptic ulcers are a leading cause of UGIB, with a significant portion attributed to duodenal ulcers[7]. This study investigates the endoscopic findings of UGIB patients at a tertiary care center in Punjab, a region with the highest incidence of hepatitis B and C in Pakistan[8,9]. Hepatitis B and C are major causes of cirrhosis, and the development of esophageal varices is a common and potentially life-threatening complication in these patients. Nearly 5% of the global population is chronically infected with HBV, and 75% of those infections are concentrated in Asia[10]. The high prevalence of hepatitis B and C in this region, coupled with the associated risk of variceal bleeding, makes it a high-risk population. This underscores the importance of exploring the relationship between these infections and UGIB. This study aims to raise awareness of the need for early detection, routine screening for varices, and better management of hepatitis in populations at high risk for these complications.
Methods:
Study design and population
This study was designed as a cross-sectional investigation conducted at a tertiary care hospital in Punjab, Pakistan. The Ethical Review Committee of the local medical university approved this study. Patient data were obtained from the endoscopy registers maintained by the hospital’s endoscopy department between October 2022 and September 2023. The registers documented key demographic details, clinical presentations, and endoscopic findings, including esophageal varices, other UGIB causes, and outcomes of various management strategies. To minimize selection bias, a systematic sampling method was employed. Rather than focusing on data from a specific ward or subgroup of patients, the study included all consecutive cases recorded in the endoscopy registers during the study period. This approach ensured that the sample was representative of the hospital’s patient population and reduced the potential for overrepresentation or underrepresentation of particular patient groups. By relying on comprehensive and hospital-wide records, we aimed to capture a diverse range of cases, avoiding biases introduced by convenience sampling. The work has also been reported in line with the STROCSS criteria[11].
The study enrolled 383 patients aged 18 and above who presented to the endoscopy department with symptoms of UGIB and were willing to undergo endoscopy. Patients were initially screened for hepatitis B and C using enzyme-linked immunosorbent assay to detect antibodies. Polymerase chain reaction testing was subsequently performed to confirm active infection by measuring the viral load for both hepatitis B and C. Patients with contraindications to endoscopy (e.g., suspected bowel perforation, severe shock), traumatic injuries, known bleeding disorders, recent surgeries, or those unwilling to participate were excluded. All eligible participants were recruited, and their data were anonymized to ensure privacy.
Statistical analysis
Data are reported as mean ± standard deviation (SD) or number (percentage). Quantitative data, such as age and duration of symptoms, were analyzed using descriptive statistics (means and SDs). In contrast, categorical variables, such as gender distribution and the presence of esophageal varices, were analyzed using frequencies and percentages. Fisher’s exact test was employed to assess the association between hepatitis status (both hepatitis B and C) and the presence of esophageal varices. Statistical significance was set at P <0.05, and a 95% confidence interval (CI) was utilized for the analysis.
SPSS version 25 and the R package were used for all data analyses. Fisher’s exact test was also conducted to assess the correlation between hepatitis C and the occurrence of esophageal varices and examine the distribution of endoscopic findings across the patient population.
Results
Among the 383 patients in the study, the mean age was 56 years, with a standard deviation of ±11 years, indicating a predominantly middle-aged population with some variability. The most vulnerable age range for UGIB is 50–59 years (see Fig. 1). The gender distribution slightly favored males, with 171 (55.4%) male and 212 (44.6%) female. These demographic insights provide a comprehensive understanding of the patient population under study.
Figure 1.
Depicts the association between esophageal varices and acid peptic disease in different age groups.
In this cohort of patients with UGIB, 78.3% presented with esophageal varices, a common complication of portal hypertension. Acid peptic disease, observed in 11% of patients, was the second most prevalent finding, though much less common than variceal presentations. At the same time, fundal varices, a rare condition, were seen in just 1% of patients as an isolated cause of UGIB. PHG was identified in 3.4% of patients as an isolated cause of UGIB. Other miscellaneous findings accounted for 2.9% of the total cases. Unremarkable findings were noted in 3.4% of cases (see Table 1). Detailed endoscopic evaluations revealed that most esophageal varices cases (60.3%) were associated with PHG (see supplementary Table 4, available at: http://links.lww.com/MS9/A807). Additionally, 19.2% of patients with esophageal varices also had fundal varices. Gastritis and gastric ulcers were less commonly associated with esophageal varices. In contrast, acid peptic disease was frequently associated with both gastritis (53.1%) and gastric ulcers (18.8%), indicating the coexistence of multiple peptic pathologies.
Table 1.
illustrates the frequency and percentage distribution of various endoscopic findings
| Endoscopic findings | Frequency (percentage) | Male | Female |
|---|---|---|---|
| Esophageal varices | 300 (78.3 %) | 172 | 128 |
| Acid peptic disease | 42 (11.0 %) | 20 | 22 |
| Fundal varices | 4 (1.0 %) | 3 | 1 |
| PHG | 13 (3.4 %) | 7 | 6 |
| Others | 11 (2.9 %) | 6 | 5 |
| Unremarkable | 13 (3.4 %) | 4 | 9 |
| Total | 383 (100 %) | 212 | 171 |
The study found that esophageal varices were more prevalent in male patients 172 (57.3%) compared to female patients 128 (42.7%). However, acid peptic disease was slightly more common in females 22 (52.4%) than in males 20 (47.6%). These findings suggest a possible predisposition of males to esophageal varices, whereas acid peptic disease does not show a strong gender preference.
Our analysis demonstrated a strong and statistically significant association between hepatitis C infection and the presence of esophageal varices. Specifically, 70% of patients with Hepatitis C had esophageal varices (see Table 2) and 70% of those with esophageal varices were also hepatitis C positive (P < 0.001). Fisher’s exact test (see Table 3) confirmed this association with a P value of <0.001, indicating a robust correlation. Detailed methodology for Fisher’s exact test and odds ratio calculation is provided in the supplemental material.
Table 2.
The hepatitis status of these individuals is shown based on their endoscopic findings
| Endoscopic findings | Hepatitis C negative (n, %) | Hepatitis B positive (n, %) | Hepatitis C positive (n, %) | Hepatitis B and C positive (n, %) | P value |
|---|---|---|---|---|---|
| Esophageal varices | 70 (23.3%) | 5 (1.7%) | 212 (70.7%) | 13 (4.3) | <0.001 |
| Acid peptic disease | 21 (50.0%) | 0 (0%) | 17 (40.5%) | 4 (9.5%) | |
| Fundal varices | 1 (25.0%) | 0 (0%) | 3 (75.0%) | 0 (0%) | |
| PHG | 4 (30.8%) | 0 (0%) | 8 (61.5%) | 1 (7.7%) | |
| Rare findings | 4 (36.4%) | 0 (0.0%) | 7 (63.6%) | 0 (0.0%) | |
| Unremarkable | 8 (61.5) | 0 (0.0%) | 5 (38.5%) | 0 (0.0%) |
Table 3.
Contingency table (2 × 2) for Fisher’s exact test evaluating the association between hepatitis C and esophageal varices
| Hepatitis C status | Esophageal varices present | Esophageal varices absent | Total |
|---|---|---|---|
| Hepatitis C positive | 212 | 40 | 252 |
| Hepatitis C negative | 70 | 38 | 108 |
| Total | 282 | 78 | 360 |
Fisher’s exact test was performed on this contingency table using a two-sided hypothesis. The corresponding P value was 0.0000817. The odds ratio is 2.87 95% CI: 1.65–4.9995. Patients with hepatitis C are nearly three times more likely to have esophageal varices than those without hepatitis C.
The odds ratio suggests that hepatitis C-positive patients are approximately three times more likely to develop esophageal varices compared to those without hepatitis C. In contrast, hepatitis B showed a much weaker association with esophageal varices (1.7%), indicating that hepatitis C is a more significant predictor of esophageal variceal bleeding. Conditions such as acid peptic disease showed different patterns, with no hepatitis B alone cases but 9.5% having both hepatitis B and C. These findings highlight the importance of hepatitis C in the risk assessment and management of esophageal varices.
Discussion
The present study provides valuable insights into the clinical characteristics, endoscopic findings, and prognosis of 383 patients with UGIB at a tertiary care center in Pakistan. The findings confirm that esophageal varices are a significant manifestation of UGIB, especially in patients with chronic liver disease (CLD) and portal hypertension[2,6]. These findings concord with the observations in similar studies worldwide and highlight the need for screening and intervention for varices in the high-risk populations.
In our cohort, esophageal varices were present in 78.3% of patients, making them the most common cause of UGIB, particularly in individuals with CLD. This aligns with the findings of Raj et al and Syed et al, who reported a high prevalence of variceal bleeding in patients with liver disease, especially those with hepatitis C[2,12]. Other studies, such as those by Ahmed et al and Pinto et al, corroborate these results, showing that portosystemic shunting and portal hypertension are major contributing factors to variceal formation and bleeding in patients with cirrhosis[6,13]. A significant observation in our study is the strong association between hepatitis C and esophageal varices. Among patients with esophageal varices, 70.7% had hepatitis C, highlighting the significant risk posed by hepatitis C in variceal formation. These findings align with those of Raj et al and Pinto et al, who also identified hepatitis C as a critical risk factor for varices and UGIB[2,6].
In contrast, hepatitis B was less strongly associated with variceal bleeding, with only 1.7% of patients with esophageal varices having hepatitis B alone (Table 2). It supports the findings of Custovic et al, which demonstrated that hepatitis C is a more aggressive predictor of variceal bleeding compared to hepatitis B[14,15]. These results underscore the importance of timely identification and treatment of hepatitis C in populations at risk of UGIB, especially in regions like Pakistan and other South Asian countries, where hepatitis C remains a significant public health concern[2,12].
Variceal bleeding has been associated with a mortality rate of up to 13% in studies by Pinto et al and Syed et al[6,12]. While our study did not examine mortality, predictors such as the need for blood transfusions and elevated creatinine levels (>1.5 mg/dL) have been linked to increased mortality in UGIB patients. Cleveland et al also found that anemia and renal disease elevate mortality risk[16]. These findings highlight the importance of prompt intervention, including endoscopy, vasopressin analogues, and proton pump inhibitors (PPIs), in managing variceal UGIB.
The high prevalence of esophageal varices exists in our group of patients, and these have different impacts on the further management of UGIB. Thus, a great emphasis should be placed on early endoscopic evaluation and prevention of variceal eradication in order to decrease the incidence of life-threatening UGIB in patients with CLD and hepatitis C. It is in concordance with American College of Gastroenterology guidelines in which routine variceal monitoring in patients with cirrhosis and portal hypertension is recommended[16,17]. Furthermore, the association between hepatitis C and variceal bleeding emphasizes the role of antiviral therapies in managing CLD patients. With the introduction of new direct-acting antivirals (DAAs) for the management of hepatitis C infection, there is a chance to decrease the rate of variceal bleeding by eradicating the cause behind cirrhosis. As per Hsieh et al, those patients who achieved sustained virological response (SVR) with DAAs have been shown to have a lower risk of developing varices than non-SVR patients, emphasizing the importance of multidisciplinary care involving both gastroenterologists and hepatologists[18].
Acid peptic disease was observed in 11% of patients in our study, which is lower than previously reported studies. This reduction in non-variceal UGIB may be due to the widespread use of PPIs, as noted by Custovic et al[14]. Interestingly, acid peptic disease showed a slight female preponderance (Table 2), possibly linked to hormonal or lifestyle factors, though this finding warrants further research[6,15]. Despite advances in modern medication reducing the incidence and mortality of peptic ulcer disease, it remains the leading cause of UGIB globally[5]. In contrast, in developing countries like Pakistan, India and Uganda, hepatitis C-induced variceal hemorrhage is the primary cause of UGIB[2,19,20].
While this study provides valuable insights, several limitations should be considered. The cross-sectional design limits the ability to establish causal relationships between risk factors and outcomes. Cross-sectional studies are best utilized to compare the prevalence of esophageal varices in hepatitis-infected patients with CLD at a specific point in time[14,17]. Consequently, the findings may not be generalizable to younger or older populations or those without liver disease. The study sample, predominantly middle-aged male patients, further restricts generalizability to populations with differing demographics or socioeconomic backgrounds. This limitation emphasizes the need for future research to include more diverse patient populations, as suggested by Wang et al, who highlighted the role of demographic factors, such as age, gender, and socioeconomic status, in influencing UGIB outcomes[21]. The strong association observed between hepatitis C infection and esophageal varices (odds ratio: 2.87) underscores the importance of screening management. However, the cross-sectional design prevents temporal relationships from being established, necessitating cautious interpretation of the odds ratio. Longitudinal studies are needed to clarify the timeline between hepatitis status and endoscopic findings.
Further research should also evaluate the long-term effects of modern treatments, such as DAAs for hepatitis C, in reducing UGIB in high-risk groups[2,21]. Integrating screening of esophageal varices into routine care for patients with CLD and portal hypertension is strongly recommended, particularly as varices are closely linked to hepatitis C. Proactive screening for hepatitis C in these patients could significantly improve outcomes. By expanding future studies to include broader age ranges, genders, and geographical locations, these findings can be validated, paving the way for more effective treatment and management strategies.
Conclusion
This study highlights the significant prevalence of esophageal varices as the leading cause of UGIB in patients with CLD, particularly those infected with hepatitis C. The strong association between hepatitis C and variceal bleeding emphasizes the need for routine screening and proactive management in at-risk populations. Early detection and treatment of hepatitis C, along with prompt endoscopic interventions, could substantially reduce the incidence and mortality related to variceal bleeding. The findings underscore the importance of integrating gastroenterological and hepatological care for better patient outcomes in regions where CLD and hepatitis infections are prevalent.
Acknowledgements
The authors would like to express their sincere gratitude to the entire staff of the Department of Medicine at Allied Hospital 2, Faisalabad, Pakistan, for their assistance and support.
Footnotes
Sponsorships or competing interests that may be relevant to content are disclosed at the end of this article.
Supplemental Digital Content is available for this article. Direct URL citations are provided in the HTML and PDF versions of this article on the journal’s website, www.lww.com/annals-of-medicine-and-surgery.
Published online 22 April 2025
Contributor Information
Masab Ali, Email: masabali17@gmail.com.
Haris Naveed, Email: haris99qadir@gmail.com.
Muhammad Husnain Ahmad, Email: husnainahmad601@gmail.com.
Ateeq Ur Rehman Sheikh, Email: decent.ateeq99@gmail.com.
Umer Hussain, Email: mumerhussain775@gmail.com.
Sana Javeriya, Email: sanajaveriyaresearch@gmail.com.
Kanchan Chaudhary, Email: kanchanmalhi421@gmail.com.
Hafiz Amjad Hussain, Email: dramjadpulmo@gmail.com.
Sarmad Naeem, Email: Dr.sarmadmirza@gmail.com.
Ethical approval
This study was approved by the Ethical Review Committee of Faisalabad Medical University. All methods were performed according to the relevant guidelines and regulations.
Consent
Written informed consent was obtained from the patient for publication of this case report and accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal on request.
Sources of funding
The authors did not receive any funding for this work.
Author contributions
M.A.: conceptualization, methodology, software, data curation, investigation, validation, formal analysis, supervision, project administration, resources, writing – original draft, writing – review & editing; H.N.: validation, visualization, writing – original draft; A.U.R.S.: software, data curation, validation, resources, writing – review & editing; U.H.: methodology, software, data curation, investigation, formal analysis, visualization, resources, writing – review & editing; H.A.H.: data curation, validation, formal analysis, resources; S.N.: visualization; M.H.A.: validation, formal analysis, visualization; K.C.: validation, visualization, writing – original draft; S.J.: formal analysis, validation, visualization, writing – review & editing.
Conflicts of interest disclosure
The authors declare no conflicts of interest.
Research registration unique identifying number (UIN)
Unique identifying number (UIN): 46ERC/FMU/2023-2024/504.
Guarantor
Masab Ali and Ateeq Ur Rehman Sheikh.
Provenance and peer review
Not commissioned, externally peer-reviewed.
Data availability statement
The data are available upon reasonable request from the corresponding author.
References
- [1].Weledji EP. Acute upper gastrointestinal bleeding: a review. Surg Pract Sci 2020;1:100004. [Google Scholar]
- [2].Raj A, Kaeley N, Prasad H, et al. Prospective observational study on clinical and epidemiological profile of adult patients presenting to the emergency department with suspected upper gastrointestinal bleed. BMC Emerg Med 2023;23:107. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [3].Iwakiri Y. Pathophysiology of portal hypertension. Clin Liver Dis 2014;18:281–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [4].Costable NJ, Greenwald DA. Upper gastrointestinal bleeding. In: Pitchumoni CS, Dharmarajan T, eds. Geriatric Gastroenterology. Cham. Springer; 2021. doi: 10.1007/978-3-030-30192-7_47. [DOI] [PubMed] [Google Scholar]
- [5].Lanas A, Chan FKL. Peptic ulcer disease. Lancet 2017;390:613–24. [DOI] [PubMed] [Google Scholar]
- [6].Pinto C, Parra P, Magna J, et al. Hemorragia digestiva alta variceal y no variceal: mortalidad intrahospitalaria y características clínicas en un hospital universitario (2015-2017). Revista Médica de Chile 2020;148:288–94. [DOI] [PubMed] [Google Scholar]
- [7].Sharifian A, Tavakoli E, Ashtari S, et al. Endoscopic findings in upper gastrointestinal bleeding patients at Tehran’s Taleghani Hospital, Iran. Govaresh. 2017;21:260–65. [Google Scholar]
- [8].Mehmood S, Raza H, Abid F, et al. National prevalence rate of hepatitis B and C in Pakistan and its risk factors. J Public Health (Berl) 2020;28:751–64. [Google Scholar]
- [9].Ahmed B, Ali T, Qureshi H, et al. Population-attributable estimates for risk factors associated with hepatitis B and C: policy implications for Pakistan and other South Asian countries. Hepatol Int 2013;7:500–07. [DOI] [PubMed] [Google Scholar]
- [10].El-Serag HB. Epidemiology of viral hepatitis and hepatocellular carcinoma. Gastroenterology 2012;142:1264–1273.e1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [11].Mathew G, Agha R. for the STROCSS Group. STROCSS 2021: strengthening the reporting of cohort, cross-sectional and case-control studies in surgery. Int J Surg 2021;96:106165. [DOI] [PubMed] [Google Scholar]
- [12].Ali SS, Butt N, Altaf HH, et al. The etiology and outcome of upper gastrointestinal bleeding in patients presenting to tertiary care hospital. Karachi 2022;16:646–48. [Google Scholar]
- [13].Ahmed A, Ali L, Shehbaz L, et al. The prevalence of acute upper gastrointestinal bleeding and the factors causing hemorrhage as observed at a tertiary health care centre in Karachi, Pakistan. J Surg Pak 2017;33:36–40. [Google Scholar]
- [14].Custovic N, Husic-Selimovic A, Srsen N, et al. Comparison of Glasgow-Blatchford score and Rockall score in patients with upper gastrointestinal bleeding. Med Arch 2020;74:270–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [15].Manko M, Bello A, Mohammed M, et al. Demographic profile and endoscopic findings among patients with upper gastrointestinal bleeding in Ahmadu Bello University Teaching Hospital, Zaria, North-Western Nigeria. Niger J Clin Pract 2020;23:1163. [DOI] [PubMed] [Google Scholar]
- [16].Cleveland PW, Woodward TA, De Melo SW, et al. 511l: The safety of duodenal endoscopic mucosal resection (EMR) in non-ampullary lesions: a study of relative risk. Gastrointest Endosc 2010;71:AB124. [Google Scholar]
- [17].Jemilohun AC, Akande KO, Ngubor TD, et al. Endoscopic findings in patients with upper gastrointestinal bleeding in Ogun State, Nigeria. Cureus 2022;14:e23637. doi: 10.7759/cureus.23637. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [18].Hsieh YY, Chen WM, Chang KC, et al. Direct-acting antivirals reduce the de novo development of esophageal varices in patients with hepatitis C virus related liver cirrhosis. Viruses 2023;15:252. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [19].Ahmed T, Kamal M, Riaz R, et al. Variceal bleeding is leading cause of upper GI bleed: a study from northern part of Pakistan. Pak J Med Health Sci 2021;15:1837–39. [Google Scholar]
- [20].Alema ON, Martin DO, Okello TR. Endoscopic findings in upper gastrointestinal bleeding patients at Lacor Hospital, Northern Uganda. Afr Health Sci 2012;12:518–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [21].Wang J, Cui Y, Wang J, et al. Clinical epidemiological characteristics and change trend of upper gastrointestinal bleeding over the past 15 years. Zhonghua Wei Chang Wai Ke Za Zhi. 2017;20:425–31. [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data are available upon reasonable request from the corresponding author.