Background: Five billion people worldwide do not have access to safe, affordable surgical care. A significant proportion live in sub-Saharan Africa (SSA), where up to 50% of the population are children. There is limited literature on neonatal and paediatric surgery in SSA and children's surgery does not appear on any of the National Health Strategic Plans for the 47 independent countries across SSA.

Objectives: To form a collaboration of surgeons and allied health professionals involved in children's surgery across SSA and collectively undertake the largest prospective cohort study of paediatric surgery in this region.

Materials and Methods: Data will be collected via REDCap website on all patients with gastroschisis, anorectal malformation, appendicitis, inguinal hernia and intussusception, during a 1-month period of collaborators choice between October 2016 to April 2017, with a 30-day follow up until the end of May 2017. Estimated study population: 1450 patients from 50 institutions. Full ethical approval has been granted by the host centre; local ethical approval will be required at collaborating centres for participation. All collaborators will be co-authors.

Primary outcome will be in-hospital all-cause mortality. Secondary outcomes will include post-intervention complications. Data will be collected on institutional facilities, patient demographics, duration from condition onset to presentation, peri-operative resuscitation, intervention and outcome.

Differences in outcomes between SSA and benchmark data from high-income countries will be calculated using chi-squared analysis. Multi-level multivariate logistic regression analysis will be used to identify interventions and peri-operative factors associated with improved outcomes; p<0.05 will be deemed significant.

Outcome: Results will be used to advocate for enhanced children's surgical services in SSA. We shall identify context-appropriate interventions associated with improved outcome. The collaboration will help to enhance research capacity in the region.

Introduction

Research Collaboratives

PaedSurg Africa aims to recruit surgeons and allied health professionals undertaking neonatal and paediatric surgery (PS) across sub-Saharan Africa (SSA) - an area heavily neglected in global health prioritisation. Such research collaboratives are being increasingly utilised as a highly effective and efficient method of collecting large volume prospective data in a short period of time. A great recruitment incentive will be ethical group co-authorship of published results for all research collaborators.

Global Paediatric Surgery

In 2015, the Lancet Commission on Global Surgery (LCoGS) highlighted that 5 billion people worldwide do not have access to safe, affordable surgical care. The same year, the World Health Assembly (Resolution 68.15) incorporated emergency and essential surgery and anaesthesic care within 'Universal Health Coverage'. In concordance with this, plans have arisen to exponentially scale up access to surgical care in low- and middle-income countries (LMICs). Children's surgery has yet to be formally recognised within these plans despite children forming up to 50% of the population in LMICs.

SSA has the highest unmet need for surgical care in the world at 41 million cases per year (29% of the worlds unmet need). 9% of the global surgical burden of disease is attributable to congenital anomalies. It is estimated that 2.6 million children are born with a congenital anomaly in SSA each year.

Hence, in order for the world to achieve the LCoGS goal of 80% coverage of surgical care by 2030, there must be a focus on scaling up neonatal and PS care in SSA where a significant proportion of the burden of surgical disease lies.

To date there has been limited data published on neonatal and PS in SSA. Nwomeh et al undertook a systematic review and meta-analysis of neonatal surgery in SSA and identified just 13 prospective studies and 38 retrospective studies between 1995 - 2014. The majority of these were single institutional studies, many with limited numbers. The results highlight poor outcomes with mortality rates of over 50-100% for conditions such as gastroschisis, which has mortality rates consistently under 4% in high-income countries. Similarly, the limited literature on PS in SSA highlights significantly poorer outcomes for common conditions such as appendicitis, intussusception and inguinal hernia.

Surgery has largely been overlooked in global health prioritisation and funding, likely because of the perception of prohibitive cost and complexity of care. Conversely, recent health economic studies have amply demonstrated the overall cost-effectiveness of PS procedures; for instance investing in a paediatric inguinal hernia repair is similar to administering a tetanus vaccine or treating a patient with malaria in terms of disability-adjusted life years (DALYS)/ US$.

This study aims to collect prospective data on five common neonatal and PS conditions across SSA, all of which have low morbidity and mortality (M&M) in high income countries (HICs), but reported poorer outcomes in SSA. It will be the first neonatal and PS cohort study across SSA and will undoubtedly form one of the largest prospective data sets in this region of the world.

The five conditions to be studied are gastroschisis, anorectal malformation, appendicitis, inguinal hernia and intussusception.

Gastroschisis

The incidence of gastroschisis is increasing worldwide, as are the number of these patients presenting to hospital in SSA. In Pretoria, South Africa, there was a 35-fold increase in patients presenting with gastroschisis from the early 1980's to late 1990s. Centres such as Harare, Zimbabwe, report receiving up to 3-4 cases per week. The true incidence of gastroschisis is unknown in SSA, however studies suggest the incidence, at least in some regions, appears to be similar to HICs. With a prevalence of 5.4/10,000 births, we can estimate that 1440 cases will be born each month in SSA.

Mortality rates for gastroschisis have fallen dramatically in HICs from 60% in the 1960's to less than 4% today. This has resulted from improvements in neonatal resuscitation and peri-operative management, awareness of complications such as abdominal compartment syndrome, and the provision of parenteral nutrition (PN). Gastroschisis has been suggested as a bellwether procedure for neonatal surgery capacity in low resource settings because there are usually no co-morbidities, yet the condition tests all the key elements that go into the successful management of a newborn with a surgical condition.

The data available regarding gastroschisis in SSA suggests significantly poorer outcomes than in HIC. Mortality rates have been reported as: 60% Malawi, 57-75% Nigeria, 84-100% Zimbabwe, 100% Uganda, 100% Cote'd'Ivoire. Mortality in South Africa is lower at 12-43%. In an international survey of gastroschisis, two thirds of the 25 institutions in SSA stated their mortality rate was >75% and the remaining third stated it was between 50-75%.

There is great variation in the management of gastroschisis across the globe and even between surgeons within the same institution. Primary closure rates vary from 16%, particularly in centres where routine use of preformed silos (PFS) have been adopted, to 92%. The majority of staged closure in SSA is undertaken using a surgical silo; PFS are utilised less due to lack of availability and training. Some surgeons recommend primary palliative care in SSA. A randomised control trial of primary closure versus PFS has reported they have similar outcomes, however there is a trend towards fewer ventilator days with the latter. A meta-analysis highlighted that the studies with least selection bias show that PFS is associated with fewer ventilator days (P<0.0001), reduced time to first feed (p=0.04) and lower infection rates (p=0.03).

There are many challenges faced by a surgeon in SSA. Patients are often outborn with no antenatal diagnosis, and hence presentation is delayed resulting in hypothermia, sepsis, hypovolaemia and bowel compromise. There is a natural tendency for babies with gastroschisis to be born early; in Harare 43% were preterm and 72% were <2.5kg. Similarly in Durban 64% were preterm and 72% <2.5kg. Many centres lack a neonatal intensive care unit (NICU) and have limited anaesthetic resources. Exposure of the bowel to amniotic fluid results in dysmotility and problems with absorption of nutrients; in the UK the average duration of parenteral (PN) requirement in simple gastroschisis is 23 days and 51 days in complex gastroschisis (gastroschisis associated with bowel perforation, necrosis or atresia). PN was only available in 36% of SSA centres when surveyed. However, in Harare the median hospital stay was 4.5days for the 85% who died, suggesting death was related to inadequate resuscitation and support rather than lack of PN. A multi-centre Gastroschisis International (GiT) study showed septicaemia to be the leading cause of death.

Some centres such as the University of Nigeria Teaching Hospital have introduced a protocol to ensure efficient pre- and post-operative resuscitation and care, preservation of body heat, decompression of the stomach with a nasogastric tube, evacuation of meconium, respiratory support and parental nutrition. As a result their mortality rate has fallen from 65% to 35%. However, Hadley notes that in Durban, despite access to NICU and PN, the overall mortality is still 43%, with sepsis being the leading cause of death. They stress the importance of meticulous central line care and a need for improved antenatal care and delivery within a tertiary paediatric surgery centre. A protocol focussed on early and aggressive resuscitation, avoidance of compartment syndrome, appropriate IV access and proactive nutritional interventions may help SSA start to realise the improvements in outcome as seen in HIC over the last 55-years. Implementing such interventions should have a knock on effect to strengthen the neonatal healthcare system in SSA and hence help to improve outcomes for all newborns with surgical conditions.

Anorectal Malformation

The birth prevalence of anorectal malformations (ARMs) has been reported as 1:1,500 - 1:5,000 in South Africa. In institutional publications from around SSA, caseload varies from 1-3 cases/ month. ARMs are one of the most common neonatal emergencies presenting to hospitals in SSA: 9.5% surgical neonates in Tanzania, 13.4% congenital malformations in Nigeria.

Mortality in HIC has fallen from 23% in the 1940's to less than 3% today. Mortality has been documented at 18.5 - 20% in Nigeria, with most deaths related to overwhelming sepsis, respiratory insufficiency and cyanotic heart disease. Unlike gastroschisis, 50% cases are associated with other anomalies: VACTERL (vertebral, anorectal, cardiac, tracheoesophageal, renal and limb). Other challenges include late presentation (2-5 days), low birth weight (38% < 2.5kg) and lack of adequate NICU facilities.

Management is dependent on the type of anomaly with low ARM typically being managed with a primary anoplasty and higher anomalies requiring a primary colostomy followed by a posterior sagittal anorectoplasty (PSARP) within the first year of life, and subsequent colostomy closure. Divided sigmoid colostomy is preferred to avoid overspill of stool from the proximal limb into the distal pouch, however some centres in SSA undertake loop colostomies. Lukong et al highlighted their improvement in outcome following a change in practice from a transverse loop colostomy to divided sigmoid colostomy with delayed PSARP. In some centres local anaesthesia is utilised to undertake the colostomy, reducing the risks associated with general anaesthesia where resources are limited.

Udefiagbon et al, highlight the problems with a colostomy, particularly in a low income setting: skin excoriation, wound infection, sepsis, prolapse, fluid and electrolyte losses and poor acceptance by care givers. They present impressive outcomes for primary PSARP undertaken during the first week of life. Authors advocate the benefits of reduced infection rate due to the sterile nature of the meconium and avoidance of a prolonged 3-stage procedure, which is more expensive and often unfeasible for African families. However, this practice requires safe neonatal anaesthesia, skilled surgeons and good standards of peri-operative care. Olivieri et al have described a modified PSARP technique utilised in Eritrea to reduce the risk of perianal wound infection where they leave a 3cm rectal stump, which is trimmed after 2 weeks. Poenaru et al advocate performing a wider anoplasty, which is less prone to stenosis, as regular dilatation and following-up can be challenging in this context.

This study will capture the primary peri-operative management and outcomes for patients presenting with ARM. Longer-term morbidity including faecal incontinence, anal stenosis, constipation, urinary and ejaculatory problems and outcomes of subsequent surgery are beyond the remit of this study.

Appendicitis

Acute appendicitis is one of the commonest causes for acute abdomen in children worldwide. In Ethiopia appendicitis accounts for 12% of emergency paediatric surgery and in the Congo it constitutes 30% of paediatric visceral surgery. Mortality from appendicitis is up to 4% in some SSA settings and morbidity high with wound infection rates up to 60% and wound dehiscence in up to 25%. Conversely, mortality from appendicitis in HICs is 0.04% and only 0.006% in children aged 9-19years. The LCoGS has listed laparotomy as one of three bellwether procedures that must be safe and available at all institutions providing first level care.

There are a number of challenges for managing children with acute appendicitis in the SSA setting. Children present late with 25-67% perforation rates; re-operation for intra-abdominal sepsis is as high as 40%. Differential diagnoses in SSA are more varied. In America, 82% of peritonitis in children is secondary to appendicitis whereas in SSA children present with peritonitis secondary to typhoid perforation, tuberculosis and ascariasis. There is often a lack of appropriate antibiotics, paediatric intensive care, blood products and parental nutrition. HIV may impact on the spectrum and severity of surgical infection in African children.

Inguinal Hernia

Paediatric inguinal hernia repair (PIHR) is the most commonly performed operation worldwide. The incidence is up to 5% of term infants and 30% of preterm. It is widely considered a basic and essential procedure that should be available to all and yet is largely unavailable in SSA. Complicated inguinal hernias pose a threat to the life of the child and considerable morbidity, which could be avoided by timely management of what is otherwise a straightforward procedure. Infants are at greatest risk of delay; Zamakhshary found that waiting over 2 weeks for a PIHR doubled the risk of incarceration in this age group.

In Ein's study of 6361 paediatric inguinal hernia repairs in Canada there were no deaths, recurrence rate of 1.2%, wound infection rate of 1.2% and testicular atrophy rate of 0.3%. In this study, 12% presented as incarcerated, but only 1% required emergency surgery for an irreducible hernia and 2/6361 required a bowel resection. Ergogan reported similar outcomes in Turkey.

In SSA, delays in presentation and lack of surgical capacity often result in a higher proportion of complicated paediatric hernias. In Nigeria, outcomes have been reported as: mortality 2.4%, intestinal resection 7%, testicular gangrene 15% and wound infection 15%. Another centre reported that in their cohort of infants presenting <42days of age, 52% were incarcerated, 36% required a bowel resection and 18% underwent an orchidectomy for testicular infarction. Strangulated hernias constitute 22% of the emergency visceral surgery in the Democratic Republic of the Congo. A randomised controlled trial in Nigeria reported a reduction in infection rate following PIHR from 4.8% to 0% with the use of prophylactic IV gentamycin at induction.

Intussusception

Intussusception is a leading cause of intestinal obstruction (IO) in children. A study from Nigeria reported 29% of IO from intussusception, 22% from ARM and 17% from obstructed PIH. The incidence is South Africa is comparable to HICs at 56/100,000 < 1-year compared to 33-49/100,000 < 1-year in the US. Published studies suggest institutions in SSA see at least one case/ month.

In HIC mortality is low at 0.1% in Europe and 0.4% in the US. Mortality rates in SSA vary widely and have been reported at 0 - 55%. The average mortality rate across SSA is 9.4%. There are a number of challenges to managing a child with intussusception in SSA. They often present late; 92% patients in Ile-Ife, Nigeria, present >24 hours. In Kenya the main duration of symptoms at presentation was 5 days (1-14). It was noted that the median delay for those who died (6.4%) was 5-days versus 3-days in those who survived. Surgical site infections following laparotomy for intussusception have been reported as high as 37.5%.

Management varies widely in SSA. In HICs most centres undertake air enema reduction (AER) as primary management in cases without peritonitis, perforation or non-responsive shock. This has been shown to be 1.48 times more effective than hydrostatic reduction (61% and 44% respectively). Centres in SSA with access to AER report good success rates; in a study in Ghana AER was successful in 59-67% cases and their mortality rate was 2%. They note that AER costs 20% of the fee for surgical management. A study from Nigeria reports offering hydrostatic reduction to 40% of cases, where it is successful in 64%. They also highlight the significant cost savings.

Ahmed et al in Egypt describe their simple AER equipment with a pressure release value at 120mmHg. They report an 88.2% success rate with no complications. Wiersma and Hadley advocate the use of AER in the operating theatre, which has improved their reduction rate to 53% from 22%.

Aim:

To compare outcomes of five common neonatal and paediatric surgical conditions between SSA and HICs.

Objectives:

1. To undertake the first multi-centre prospective cohort study across SSA to compare outcomes of common paediatric surgical conditions with benchmark data from HICs.
2. To identify context appropriate interventions and peri-operative factors associated with an improved outcome.
3. To form a research collaboration of paediatric surgeons and allied health professionals across SSA and help to enhance research capacity.
4. To raise awareness and provide advocacy for neonatal and paediatric surgical care within global health prioritisation, planning and funding.

Methodology

Authorship:

PaedSurg Africa constitutes a network of surgeons and allied health professionals who work with children requiring surgery across sub-Saharan Africa. This methodology is based on an equal partnership model previously described in the Lancet and utilised by a number of national and international collaboratives.

Publishing journals will be asked to make all co-authors PubMed citable. Articles will be published under 'PaedSurg Africa Research Collaboration'. At the end of the article co-authors will be listed under the following headings:

• Lead investigators: collaborators who have contributed to the study protocol, data analysis and write-up of the manuscript.
• Country leads: collaborators who have recruited several sites in their country to contribute to the study.
• Local investigators: collaborators who have gained ethical approval for the protocol in their centre and collected data at their site including patient identification and data completion with follow-up of mortality and complications.

The study invites up to three local investigators per institution per month of data collection.

Collaborator Recruitment:

Primary recruitment will be through the lead investigators' personal contacts, which span many centres and countries across SSA, and the Pan-African Paediatric Surgical Association (PAPSA) mailing list. Research collaborators were recruited at the Global Initiative for Children's Surgery (GICS) Conference hosted by the Royal College of Surgeons of England, May 2016. A further recruitment drive will be undertaken at the PAPSA Conference in Nigeria (September 2016) where the research proposal will be presented, investigators actively sought, and REDCap sign-up, app download and project set-up achieved.

Data Collection Tool:

Prospective data will be collected over a 1-month period utilising the free, user-friendly, secure database, REDCap. The tool includes a Smartphone app that allows offline data collection. Data collection sheets that can be printed for written data collection and later uploaded will also be provided. Data collection will be open for any continuous 1-month period between October 2016 to April 2017, to optimise uptake. A pilot study will be run in August/ September 2016 at five institutions in SSA.

Conditions Studied:

The five conditions selected are common congenital or acquired neonatal and paediatric surgical conditions with low morbidity and mortality in HICs. Consequently surgical interventions for these conditions carry high avertable DALYs and cost-effectiveness ratios, as saving the life of a neonate or child with minimal long-term disability can permit a lifetime of labour and income for their family and their country's economy. Limited, mostly retrospective, individual institutional studies suggest significantly poorer outcomes of these conditions in SSA.

Validation:

At 5% of collaborating centres, one additional research collaborator will be asked to identify patients and upload the study data independent to the other research collaborator(s) at their centre. This data will be collected on a separate REDCap validation database and the inputted data will be cross-checked with that entered into the main database. Only patients with over 90% of data inputted will be included in the study. Data can be initially uploaded and completed at a later date prior to submission if any data is missing.

Questionnaire Data:

A short questionnaire will be undertaken by all research collaborators at the time of project sign up regarding the facilities and resources available at their institution.

Estimated Population:

Estimated patient numbers per centre during a 1-month study period are: 1-2 gastroschisis, 1-2 anorectal malformations, 11 appendicitis, 14 inguinal hernias, 1 intussusception (29 patients per centre). We aim to include a minimum of 50 centres, which would generate 1450 patients. Estimates were calculated using the mean number of patients presenting per month to all institutions from SSA who have published data on these conditions.

Data Analysis:

Power calculations were undertaken to determine the minimum detectable differences in mortality between SSA and benchmark data from HICs. These confirm that the study is adequately powered to detect a difference in mortality of each of the five conditions between SSA and HICs at a 5% significance level and 80% power. Outcomes in SSA compared to HICs will be presented as relative risks and significant differences determined using Chi-Squared analysis.

Multi-level, multivariate logistic regression analysis will be utilised to identify context appropriate interventions and peri-operative factors associated with improved outcomes. For example: does availability of air enema reduction significantly reduce mortality from intussusception in the SSA context and if so, by how much? Results will be presented as odds ratios. Data will be adjusted for confounding factors such as delay in presentation. P<0.05 will be deemed significant.

Ethics:

Full ethical approval for the project has been granted by the host centre, Kings College London Research Ethics Committee. Research collaborators in SSA will be required to receive approval for the project at their own centres according to local ethical regulations and provide evidence of this in order to submit data.

If no formal ethics or audit committee exists, collaborators must provide written consent from the Director of the Hospital or Head of Surgical Department. All data will be anonymous. Data will not be identifiable at individual surgeon, institution or country level.

Outcome:

The results from this study can be used to advocate for enhanced neonatal and paediatric surgical care in SSA, at an institutional level, country level via 'National Health Strategic Plans' and at an international level. This is vital to ensure neonatal and paediatric surgery are appropriately prioritised over the coming 15-years as access to surgical care in low- and middle-income countries is scaled up following publication of the LCoGS and the WHA resolution 68:15. Participation in this international project will help to enhance research capacity amongst surgeons, anaesthetists and allied healthcare professionals across SSA. Formation of the PaedSurg Africa Research Collaboration will provide the infrastructure for future research projects and interventional studies to help improve outcome.

Inclusion Criteria:

All patients under 16-years of age presenting primarily with one of the five study conditions during the data collection period. This includes those managed conservatively without surgery or intervention and those offered palliative care, except patients with an inguinal hernia who must receive surgery to be included.

Exclusion Criteria:

Patients re-presenting with one of the five study conditions with a complication from a previous intervention or requiring further surgical intervention.

Collaborating centres were recruited using convenience sampling with snowballing. Centres include both tertiary paediatric surgery centres and district hospitals.

Following publication, the full anonymous data set will be made publicly available.