Article Text

Neonatal early onset sepsis in Middle Eastern countries: a systematic review
2. Heather B Blunt3,
3. Zhongze Li4,
4. Tyler Hartman2
1. 1 Pediatric Infectious Diseases, Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio, USA
2. 2 Department of Pediatrics, Dartmouth-Hitchcock Medical Center, Lebanon, New Hampshire, USA
3. 3 Biomedical Libraries, Dartmouth College, Hanover, New Hampshire, USA
4. 4 Biostatistics Shared Resource, Dartmouth-Hitchcock Norris Cotton Cancer Center, Lebanon, New Hampshire, USA
1. Correspondence to Dr Nadim Khalil, Pediatric Infectious Diseases, Cincinnati Children's Hospital Medical Center, Cincinnati, OH 45229, USA; nadim.khalil{at}cchmc.org

## Abstract

Background Early onset neonatal sepsis (EOS) accounts for a significant portion of neonatal mortality, which accounts for 46% of global under five child mortality.

Objective This systematic review studies the bacterial aetiology of EOS in the Middle East, susceptibility patterns to recommended empirical antibiotic therapy and whether this differs between high-income and middle-income countries in the region.

Methods Articles were collected from Medline, Web of Science, the Cochrane Library and Index Medicus for the Eastern Mediterranean Region. The articles included in our systematic review met the following criteria: published after January 2000, data relevant to the Middle East, data specific for early onset sepsis, no language restriction. Data on aetiology and susceptibility were extracted from prospective and retrospective studies. Risk of bias was assessed using the Newcastle-Ottawa Scale. This study focused on EOS but does include data regarding neonatal late-onset sepsis antibiotic susceptibility. The data regarding coagulase-negative Staphylococcus species were excluded from final analysis, as possible contaminants. The protocol for this systematic review was registered on PROSPERO: CRD42017060662.

Results 33 articles from 10 countries were included in the analysis. There were 2215 cases of culture-positive EOS, excluding coagulase-negative Staphylococcus. In middle-income countries, Klebsiella species (26%), Staphylococcus aureus (17%) and Escherichia coli (16%) were the most common pathogens, in contrast to group B Streptococcus (26%), E. coli (24%) and Klebsiella (9%) in high-income countries. Overall susceptibility to ampicillin/gentamicin and third-generation cephalosporin were 40% and 37%, respectively, in middle-income countries versus 93% and 91%, respectively, in high-income countries.

Conclusions EOS in middle-income countries was more likely to be due to Gram-negative pathogens and less likely to be susceptible to empirical antibiotic therapy. This has important public health implications regarding neonatal mortality in the Middle East region.

• neonatology
• microbiology
• infectious diseases
• epidemiology

## Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

## Introduction

Neonatal sepsis remains a major cause of morbidity and mortality in neonates despite medical advancements.1 According to the latest WHO report, child mortality below age 5 years has decreased by more than 50% over the last 15 years.2 3 However, 2.6 million newborn still die in the first month of life.3 Of all neonatal deaths, 25%–45% occur within the first 24 hours of life1 and 77.7% occur in the first week of life.4 A recent systematic review has shown that 30% of under-five mortality occurs within the first 3 days of life and that 47.4% of neonatal sepsis occurs within the first week of life.4

Early onset neonatal sepsis (EOS) is thought to be due to intra-amniotic infections and exposure to microorganisms while passing through the birth canal. EOS is generally defined as infections occurring in the first week of life,5 even though some definitions limit it to the first 72 hours of life.6 This is in contrast to neonatal late-onset neonatal sepsis (LOS) that occurs after the first week of life5 and is usually community or nosocomial in origin.7

According to the current WHO guidelines,8 empirical treatment with intravenous ampicillin and gentamycin is still recommended for management of neonatal sepsis. Unfortunately, with the current increase in rates of antibiotic therapy,9 this empirical antibiotic therapy might not be adequate in low-income and middle-income countries. To decrease the burden of EOS, intrapartum prophylaxis for group B Streptococcus (GBS) was initiated in developed countries in the 1990s with successful decrease in frequency of EOS caused by GBS.10 11 However, GBS does not appear to be a common cause of EOS sepsis in developing countries.9 12 13

The previous systematic reviews on neonatal sepsis in developing countries either included articles from 1980,13 did not include data specific to EOS9 or did not include articles on the Middle East.9 12 13 This study focuses on EOS because with proper screening and prevention, the incidence and mortality due to EOS may be decreased.14 The countries in the Middle East provide an opportunity to compare the bacterial aetiology of EOS and their susceptibility between countries with different economies and development in a relatively small geographical area.

## Methods

### Search strategy

A review of published articles in Medline (PubMed), Web of Science, the Cochrane Library and the Index Medicus for the Eastern Mediterranean Region (IMEMR), part of the WHO database, was conducted. Although there is a lack of a uniform definition for the Middle East, most resources15–18 agree on the following countries being part of the Middle East: Bahrain, Egypt, Iraq, Iran, Israel, Jordan, Kuwait, Lebanon, Oman, Palestine, Qatar, Saudi Arabia, the Syrian Arab Republic (Syria), the United Arab Emirates and Yemen.

The search criteria for PubMed is included in table 1. Search criteria for other databases can be found in online supplemental data 1. The search was conducted on 30 April 2017 and again on 4 December 2017. We also reviewed articles from reference lists and from Google Scholar. The protocol for this systematic review was registered on international prospective register of systematic reviews (PROSPERO) on 12 April 2017: CRD42017060662. This systematic review was conducted according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.

### Supplemental material

Table 1

Search terms for Medline (PubMed)

### Inclusion criteria

The articles included in our systematic review met the following criteria: published after January 2000 (to better represent current trends), data relevant to studied countries, data specific for EOS, no language restriction and no study-type restriction. The research results were compiled and uploaded into Rayyan.org (a web-based app for systematic reviews)19 for analysis. The studies were then reviewed by two independent investigators. Studies that fit the above inclusion criteria or had abstracts with incomplete information were selected for full article review. Any discrepancy was resolved by the two investigators after the initial screen and again after full article review.

### Data extraction and quality assessment

One investigator extracted the data into a Microsoft Excel datasheet. When available, data extracted included the country, year during which study took place, study design, definition of EOS, number of live births, number of newborns suspected to have EOS, number of newborns admitted during the period, number of live births during the period, number of positive blood cultures, rate of EOS, and the frequency and proportion of various pathogens.

When reported, data regarding antibiotic susceptibility were obtained for each separate pathogen, by extracting the number of blood cultures on which susceptibility testing was performed as well as the number of organisms susceptible/resistant to each of ampicillin, gentamicin, third-generation cephalosporins, amikacin, piperacillin/tazobactam, carbapenems, vancomycin, ciprofloxacin and trimethoprim-sulfamethoxazole (available in online supplemental data 2). When available, data for cefotaxime and imipenem were selected over ceftriaxone or meropenem for the categories of third-generation cephalosporin and carbapenems, respectively. In addition to the 6 articles with both carbapenems susceptibilities, 12 articles reported susceptibility to imipenem only in contrast to three studies with meropenem-only data. As current WHO guidelines recommend ampicillin plus gentamicin as empirical therapy for EOS, an estimated combined susceptibility was calculated (online supplemental data 3).

### Supplemental material

The data extracted on aetiology is specific to EOS only. However, as several articles combined EOS and LOS susceptibility data, the susceptibility data in this statistical analysis contains information regarding LOS.

Articles that met inclusion criteria where assessed for quality using the Newcastle-Ottawa Scale quality assessment.20 The studies were evaluated based on selection of study groups, comparability of the groups and ascertainment of the outcome. Included studies could obtain a maximum of nine stars, with higher scores indicating a lower risk of bias (online supplemental data 4).

### Analysis

Data were compiled by country and per United Nations income classification21 into middle-income ($1006–$12,235) or high-income categories ($12 236 or more). None of the studied countries met the UN criteria for low-income country (less than$1006). Analysis was also conducted based on per capita gross national income (GNI) from the latest 2016 World Bank report.22

The data regarding coagulase-negative Staphylococcus species (CONS) was extracted but was excluded from final analysis as it is usually regarded as a contaminant in this population.7 13 In addition, most of the articles did not clarify if CONS were classified as true pathogen or a contaminant and lacked details regarding the presence of central lines.

Logistic regression modelling was used to determine the association of per capita GNI with the odds of occurrence of a particular organism or antibiotic susceptibility and presented as ORs alongside 95% CIs and p values. To compare differences between middle-income and high-income countries, we presented frequencies and percentages of organisms and used Fisher’s exact test to assess significant differences. Two-sided significance level was set at 5%. All statistical analyses were conducted in SAS V.9.4 (SAS Institute, Cary, North Carolina, USA).

## Results

Our search of the databases yielded 3878 citations: 1618 from Medline, 1688 from Web of Science, 210 from Cochrane Library and 362 from IMEMR (figure 1). We also identified seven articles from other sources.23–29 Of these, 1199 were duplicate articles that were removed. A screen of the titles and abstracts produced 67 possible articles. After a full review of the 67 articles, 33 articles were included in the analysis. Thirty-two articles were excluded because they either did not meet inclusion criteria, combined the aetiology of EOS with those of LOS or had incomplete EOS aetiology data available. In addition, two articles,30 31 cited in IMEMR, were also excluded as they could not be retrieved even after the IMEMR database, the authors and the publishing journals were contacted.

Figure 1

Overview of literature search results identified through article references and Google Scholar.

### Description of studies

The 33 articles identified were divided as follows: Egypt, Iran and Israel each had seven articles, Turkey had four articles, Jordan and Iraq had two articles each and Kuwait, Syria and Saudi Arabia had one article each (table 2). One multicountry study by Hammoud et al,32 containing data from Kuwait, United Arab Emirates and Saudi Arabia was included only in the high-income versus middle-income analysis. There were six countries without any available data: Bahrain, Lebanon, Palestine, Oman, Qatar and Yemen.

Table 2

Included studies for aetiology of early onset neonatal sepsis with data from the latest WHO report, 2016

The rates of EOS was calculated using the number of EOS with positive blood cultures and number of live births reported during the study period. The rate of EOS, calculated for 12 articles with a total of 620 306 live births, ranged from 0.6 to 15.7 (mean 3.07, CI 0.5 to 5.7) cases per 1000 live births (table 2). Twenty-two studies were conducted in neonatal intensive care units (NICU), four were in both NICU and nursery, one on neonatal ward,33 while six did not clarify location. EOS was defined as infections occurring in the first 3 days of life (18 out of 33 articles) while 7 articles defined EOS in the first week of life. There were three articles that did not clarify their definition of EOS.

Twenty-two articles were retrospective cohort studies while 11 articles were prospective cohort studies. Regarding blood culture techniques, only 11 articles clarified the blood collection process while 21 articles included details on blood culture growth in the microbiology laboratory and 27 articles reported information regarding the susceptibility of organism causing EOS. Results of the Newcastle-Ottawa Scale quality assessment are available in online supplemental data 4.

### Aetiology

In the studied articles, there were a total of 2832 positive blood cultures that fulfilled the definition of EOS. After excluding CONS, there were a total of 2215 identified organisms in the positive blood cultures, 1182 (53%) were in middle-income countries and 1033 (47%) were in high-income countries. The number of positive blood cultures per country varied from 12 to 807 (mean 85.8, CI 57.2 to 114.5) (figure 2, table 3 and online supplemental data 5).

### Supplemental material

Figure 2

Percentage of microorganismsa, per country, associated with neonatal early onset sepsis in the Middle East for middle-income and high-income countries. a, total positive cultures per country excluding cons; b, article by Hammoud et al containing data from Kuwait, United Arab Emirates and Saudi Arabia. GBS, group B Streptococcus.

E. coli (20%) was the most common organism recovered in EOS in the Middle East (range 9%–33%; mean 19%; CI 13% to 24%). It was most prevalent in Israel (28%), Jordan (27%) and Saudi Arabia (3%). E.coli was the third most common organism in all middle-income countries (16%) and the second most common organism in all high-income countries (24%). The odds of having E. coli increased with each $1000 increase in GNI per capita (OR 1.02; CI 1 to 1.03; p=0.03). That is, E.coli was more common in countries with higher per capita GNI. Klebsiella species (18%) were the second most common causative organisms of culture-positive EOS in the Middle East (range 0%–36%; mean 17%; CI 9% to 25%). Klebsiella was most prevalent in Egypt (36%), Iran (26%) and Syria (25%). Klebsiella were also found to be the most common causative organisms of EOS in middle-income countries (26%) and the third most causative agent in all high-income countries (9%). The odds of having Klebsiella decreased with each$1000 increase in per capita GNI (OR 0.96; CI 0.94 to 0.98; p<0.0001). That is, Klebsiella was more common in countries with lower per capita GNI.

GBS (14%) was the third most common causative organism of culture-positive EOS in the Middle East (range 0%–64%; mean 15%; CI 1% to 29%). GBS was most prevalent in Kuwait (23%) and in the study by Hammoud et al (64%). It was the second most causative agent in Israel (22%). GBS was also found to be the most common causative organism of EOS in high-income countries (26%) but the sixth most causative agent in all middle-income countries (4%). The odds of having GBS increased with each $1000 increase in per capita GNI (OR 1.06; CI 1.03 to 1.1; p<0.0001). Staph aureus (SA) (11%) was the fourth most common causative organism of culture-positive EOS in the Middle East (range 2%–41%; mean 16%; CI 6% to 26%). SA was the most prevalent in Iraq (41%) and Turkey (39%). SA was also the second most prevalent organism causing EOS in all middle-income countries (17%) and fourth most causative agent in high-income countries combined (4%). The odds of having SA decreased with each$1000 increase in per capita GNI (OR 0.96; CI 0.9 to 1; p=0.05).

Enterobacter species (8%) were the fifth most common causative organisms of culture-positive EOS in the Middle East (range 0%–21%; mean 7%; CI 2% to 12%). Enterobacter species were also the fourth most prevalent organisms causing EOS in all middle-income countries (12%) but much less common in high-income countries combined (3%). The odds of having Enterobacter decreased with each $1000 increase in per capita GNI (OR 0.95; CI 0.93 to 0.98; p=0.001). Gram negatives (E.coli, Klebsiella, Enterobacter, Pseudomonas, Acinetobacter and other Gram negatives) were more common as causative agents of culture-positive EOS in the Middle East (62%). Gram negatives were also more prevalent in middle-income countries (71% vs 51%). The odds of having gram negatives decreased with each$1000 increase in per capita GNI (OR 0.98; CI 0.95 to 1; p=0.024). Gram positives were more prevalent in high-income countries (46% vs 27%). The odds of having Gram positives increased with each \$1000 increase in per capita GNI (OR 1.03; CI 1 to 1.05; p=0.025).

Table 3

Distribution and association of the occurrence of microorganisms with per capita gross national income (GNI) in the Middle East

### Susceptibility

Twenty-seven out of the 33 articles contained data regarding antibiotic susceptibility. Of those, 21 articles combined susceptibility results for EOS with LOS and 20 articles were from middle-income countries.

Ampicillin susceptibility per country varied from 0% in Syria to 83% in the study by Hammoud et al (mean 29%, CI 10% to 48%) (table 4). Ampicillin susceptibility in all middle-income countries was 11% vs 60% in high-income countries combined (OR 1.08; CI 1.05 to 1.1; p<0.0001). GBS susceptibility to ampicillin varied from 17% in Jordan to 100% in all high-income countries (mean 69%, CI 16% to 122%). Overall GBS susceptibility was 25% in middle-income countries combined versus 100% in high-income countries combined (p<0.0001) (table 5).

Table 4

Overall antibiotic susceptibility for treating neonatal sepsis in the Middle East by country and by income classification

Table 5

Susceptibility of most common organisms causing neonatal sepsis in the Middle East to select antibiotic therapy per income classification (p values calculated using Fisher’s exact test)

Gentamicin susceptibility per country varied from 5% in Syria to 100% in the study by Hammoud et al (mean 57%, CI 33% to 81%). Gentamicin susceptibility in all middle-income countries combined was 33% vs 88% in high-income countries combined (OR 1.1; CI 1.04 to 1.2; p=0.002).

At the time of this systematic review, the WHO recommended regimen for suspected EOS was intravenous/intramuscular ampicillin and gentamicin. Combined estimated ampicillin plus gentamicin susceptibility per country varied from 9% in Syria to 100% in the study by Hammoud et al (mean 60%, CI 36% to 83%). Combined ampicillin/gentamicin susceptibility in all middle-income countries combined was 40% vs 93% in high-income countries combined (OR 1.1; CI 1.04 to 1.2; p<0.001).

Cefotaxime or ceftriaxone susceptibility was reported by most articles as third-generation cephalosporins are occasionally used as an alternative for gentamicin. The overall susceptibility per country to third-generation cephalosporins varied from 10% in Syria to 96% in Israel (mean 57%, CI 33% to 80%). Third-generation cephalosporin susceptibility in all middle-income countries was 37% vs 91% in high-income countries combined (OR 1.1; CI 1 to 1.3; p=0.031).

Susceptibility to some broad-spectrum antibiotics was occasionally reported. Susceptibility to amikacin was 56% in middle-income countries versus 96% in high-income countries (OR 1.2; CI 1 to 1.3; p=0.05). Susceptibility to piperacillin/tazobactam was 28% in all middle-income countries versus 42% in high-income countries without statistically significant difference (OR 1; CI 1 to 1.1; p=0.214). The susceptibility to carbapenems and vancomycin was higher when reported, 83% and 93%, respectively, in middle-income countries combined versus 89% and 100% in high-income countries combined but not statistically different (p=0.886 and p=0.292, respectively).

## Discussion

The aim of this systematic review was to better understand which organisms are most likely to cause EOS in the Middle East. This area was selected as it is a relatively small geographical area (76% the size of the USA) but with unequal distribution of resources to control for possible geographical variations. Previously, Candida non-albicans has emerged in Europe and Asia as the primary cause of neonatal candidemia in contrast to Candida albicans in North America, despite a similar income level.34 This is the first study of its kind looking at EOS in the Middle East and at association of aetiology with level of income. When possible, data were combined for descriptive purposes and to expand results and analysis.

The rate of EOS, calculated from 12 articles with a total of 620 306 live births, varied from 0.6 (30) to 15.7 (39) cases per 1000 live births and with higher rates of EOS in countries with lower income. EOS in high-income countries was more likely due to GBS (26%) and E. coli (24%) with an equal distribution of Gram positives (46%) and Gram negatives (51%). This is similar to EOS in developed countries such as the USA where GBS was found to be the most common cause of EOS.35

On the other hand, EOS in countries with lower GNI was more likely to be secondary to Gram negatives (71%), Klebsiella (26%) and SA (17%). This is comparable to data from developing nations,13 where Klebsiella are the most common organisms. The divergence in aetiology between high-income and middle-income countries can be due to several factors such the lack of prenatal care, more rural populations, difficulty in accessing healthcare services, poor water infrastructures and increased maternal intrapartum infections.36 37

The results found in this systematic review could indicate that initial empirical therapy of EOS with either ampicillin plus gentamicin (estimated susceptibility 40%) or third-generation cephalosporin (37%) is less likely to be effective in countries with lower GNI. This is in contrast to bacteria in high-income countries which were more likely to be susceptible to initial empirical antibiotic therapy with ampicillin plus gentamicin (estimated susceptibility 93%) or third-generation generation cephalosporin (91%). The discrepancy in antibacterial resistance is partly due to differences in organisms, but may be the result of a lack of antibiotic regulation, antibiotic overuse,38 increased antibiotic use in agriculture,39 and transmission to humans via ingestion and the environment.40 41

Our systematic review had some limitations. While there were 2215 culture-positive EOS in our study, this is a small percentage of EOS occurring in the region and might not be a true representation of bacterial distribution in the Middle East. The cultures were mostly collected from NICUs and could possibly exclude infections occurring in areas with less access to intensive care.35 There were six countries without any microbiological results at the time of this systematic review. Data from Egypt, Iran and Israel account for 68% of reported EOS infections and data from Israel account for 78% of data from high-income countries.

There was also heterogeneity regarding the study designs, study setting, study methodology, blood culture collection and processing. In addition, patient characteristics were not always available and the exact number of newborns that were evaluated for EOS is not clear. Seven articles defined EOS as infections occurring in the first week of life and another three articles did not specify a definition for EOS. This can introduce infections acquired from the community or the hospital as opposed to infections acquired during delivery and skewed bacterial distribution.

Only six articles contained susceptibility data for EOS only while the others had incomplete or combined information with LOS. This can lead to falsely elevated rates of resistance and decreased susceptibility to recommended empirical antibiotic therapy. Not all articles reported data regarding susceptibility making some of the sample sizes small and data analysis challenging when comparing middle-income countries to high-income countries. There was also variability between the articles in reporting the susceptibilities to ampicillin, gentamicin or third-generation cephalosporins which can skew overall results. In addition, break point and minimal inhibitory concentration laboratory reference used in the articles were not always available or well defined.

There were two articles30 31 from Egypt and Jordan, that could not be located despite multiple attempts. It is unclear what impact the missing data could have on our results as the study size and scope could not be determined from the articles’ titles. The data from Hammoud et al combined results from three different high-income countries. These data could not be included in logistic regression analysis but were included in Fisher’s exact test analysis. This discrepancy was noted when comparing GBS susceptibility to ampicillin and rates of MRSA between high-income and middle-income countries.

Per capita GNI was extracted from latest World Bank 201622 data except for Syria as the last data available were from 2007. There was only one study from Syria fitting our inclusion criteria, published in 2006 prior to the civil war that has affected the country. It is not clear at this time how the civil war will affect the bacterial aetiology for EOS in Syria. Finally, the large number of comparisons performed in this study could have introduced bias.

The current WHO guidelines recommend intravenous/intramuscular ampicillin and gentamicin for treatment of EOS. However, there is concern of an increase in antibiotic resistance in developing countries which could lead to inadequate treatment of serious bacterial infection and possible increase in neonatal mortality secondary to sepsis. Physicians should be alert to a potential increase in antibiotic resistance in middle-income countries as well. Intrapartum prophylaxis, which has worked wonderfully for GBS in developed countries, might not be as effective in low-income or middle-income countries because of possible bacterial heterogeneity and increased resistance. Treatment of EOS, in particular in middle-income Middle Eastern countries, should be tailored using WHO guidelines as well as local antibiogram and culture results. The results of this systematic review could be used to help guide antimicrobial therapy when microbiological data are unavailable or when experiencing empirical therapy failure: no clinical improvement within 24–48 hours of antibiotic initiation or failure to sterilise the bloodstream. However, clinicians should be aware that there is an increased risk of colonisation or infection with extended spectrum β-lactamasesproducing Enterobacteriaceae when cephalosporins are routinely used in a neonatal unit.42

Currently, there is a continued need to improved control and regulation to prevent the increasing bacterial resistance to broad spectrum antibiotics. There is also benefit in continuing to monitor susceptibility for possible change in current antibiotic guideline.

## Acknowledgments

The authors thank Dima Abu Izzeddin for support and assistance with manuscript editing.

1. 1.
2. 2.
3. 3.
4. 4.
5. 5.
6. 6.
7. 7.
8. 8.
9. 9.
10. 10.
11. 11.
12. 12.
13. 13.
14. 14.
15. 15.
16. 16.
17. 17.
18. 18.
19. 19.
20. 20.
21. 21.
22. 22.
23. 23.
24. 24.
25. 24.
26. 26.
27. 27.
28. 28.
29. 29.
30. 30.
31. 31.
32. 32.
33. 33.
34. 34.
35. 35.
36. 36.
37. 37.
38. 38.
39. 39.
40. 40.
41. 41.
42. 42.
43. 43.
44. 44.
45. 45.
46. 46.
47. 47.
48. 48.
49. 49.
50. 50.
51. 51.
52. 52.
53. 53.
54. 54.
55. 55.
56. 56.
57. 57.
58. 58.
59. 59.
60. 60.
61. 61.
62. 62.
63. 63.
64. 64.
65. 65.
66. 66.
67. 67.

• ## Supplementary Data

This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

## Footnotes

• Contributors All authors were involved with the study design and concept. NK extracted the data, completed figures and tables. HBB performed the literature search and retrieved the articles. Articles were reviewed independently by NK and TH. Quality assessment was conducted by NK. Statistical analysis was carried out by ZL. All authors contributed to the writing of the manuscript.

• Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

• Competing interests None declared.

• Patient consent for publication Not required.

• Provenance and peer review Not commissioned; externally peer reviewed.

• Data availability statement All data relevant to the study are included in the article or uploaded as supplementary information.