Introduction
Intestinal atresia (IA) is a congenital malformation of the gastrointestinal tract, occurring in approximately 3.4 out of every 10 000 live births.1 As one of the leading causes of neonatal intestinal obstruction in sub-Saharan Africa, IA represents a critical surgical emergency, necessitating prompt intervention. Despite its amenability to emergent surgical treatment, the condition continues to pose significant challenges in resource-limited settings, underscoring the importance of understanding its impact and outcomes within these environments.2–4
The notably higher postoperative mortality rate for IA in sub-Saharan Africa (21.7–41%) compared to well-resourced settings (about 5%) can be attributed to several interrelated factors.5–7 First, delays in seeking, reaching, and receiving appropriate surgical care are prevalent in this region, resulting in emergent interventions typically being performed within 4 to 10 days after birth, significantly later than the 24 to 30 hours in well-resourced settings.8,9 These delays compound the risk of complications and diminish the effectiveness of surgical outcomes. Furthermore, the prevalence of associated anomalies, particularly cardiac defects, and the greater incidence of low birth weight further contribute to increased vulnerability and poor prognosis among affected neonates.10,11 Collectively, these systemic barriers and clinical challenges create a context in which mortality remains high for neonates with IA in sub-Saharan Africa.
Pathological classification as delineated by the Grosfeld system also plays a pivotal role in determining patient outcomes.12 In this scheme, IA is categorized according to its anatomical location—duodenal (DA), jejunoileal (JIA), or colonic atresia (CA)—and is further subclassified based on morphological features. Four principal types of IA are recognized: Type I, involving a mucosal atresia with the bowel wall remaining intact; Type II, characterized by 2 blind ends connected by a fibrous cord; Type IIIA, where bowel segments are completely separated by a mesenteric defect; and Type IIIB, described as the “apple-peel” or “Christmas tree” deformity, which involves extensive loss of bowel.12 Type IV is defined by the presence of multiple atresias, often resulting in considerable morbidity.
For DA atresia, mortality rates in the literature range from 3% to 5%, largely influenced by the presence and severity of other congenital anomalies.13 JIA atresia demonstrates a distinction between noncomplex types—Types I, II, and IIIA, which are associated with perioperative mortality rates of less than 1%—and complex forms, specifically Types IIIB and IV, where mortality in high-income settings may approach 30% to 70%.9,14 The predominant contributors to postoperative mortality in IA include anastomotic leak, stricture formation, and both functional and anatomical short bowel syndrome, a consequence of underlying intestinal dysmotility or the necessity for extensive bowel resection, particularly in complex JIA atresia. In this context, the availability of neonatal intensive care, anesthesia support, parenteral nutrition, and specialized intestinal rehabilitation is critical to improving outcomes.14
As Tanzania’s primary referral center, Muhimbili National Hospital (MNH) has historically served as the nation’s sole public institution up to 2020 offering comprehensive pediatric surgical care.15 Neonates identified with IA through clinical and radiological assessment are taken to the operating theater for exploratory laparotomy performed by pediatric surgeons, with anesthesia provided by anesthesia specialists. Intraoperative management is tailored to the pathology encountered, focusing on restoring intestinal continuity and maximizing bowel preservation by excising diseased segments and creating an intestinal anastomosis. After surgery, newborns are transferred to the neonatal intensive care unit, where a multidisciplinary team of neonatologists and pediatric surgeons oversees their recovery and nutritional needs. Discharge is considered once the infants can tolerate age-appropriate volumes of breast milk.
Despite the critical impact of IA on neonatal morbidity and mortality, there is a notable absence of published data on their clinical presentation and outcomes within our setting. This knowledge gap hinders the ability to implement evidence-based, context-specific interventions that aim to improve survival and care quality for affected neonates. Addressing this void is essential in identifying the unique challenges and risk factors faced in our environment, thereby enabling the development of targeted strategies that can meaningfully reduce postoperative mortality and enhance neonatal outcomes. This study aimed to assess postoperative mortality in IA at MNH, Tanzania from 2017 to 2022) and to identify clinical and pathological factors linked to mortality. These insights can guide targeted interventions for high-risk neonates and help reduce postoperative deaths in this setting.
Methods
Study design and settings
The study employed a retrospective chart review design, focusing on neonates with IA who were treated at MNH between January 1, 2017 and January 1, 2022. MNH, a tertiary-level referral center in Tanzania and one of only 4 hospitals in the country with pediatric surgical services, served as the setting for this investigation.
Eligibility criteria
Eligibility for inclusion required an intraoperative confirmation of IA, while cases with missing postoperative data or those associated with gastroschisis were excluded due to the latter’s significantly higher mortality related to the coexistence of 2 major malformations.
Data sources
Researchers gathered baseline data from patient records, including referral status (whether the neonate was born at the tertiary center or referred from a lower-level facility), mode of delivery, and initial provisional diagnoses. Gestational age and birth weight were documented as continuous variables. The time from birth to surgical consult and the interval from consult to surgery were carefully noted to provide a complete picture of any delays in care.
Operative notes were examined to determine the anatomical site and type of atresia based on the Grosfeld classification, alongside any additional gastrointestinal anomalies and the specifics of surgical management. Postoperative in-hospital mortality was defined as any death occurring after laparotomy during the same hospitalization; outcomes were recorded as either discharged alive or death during admission.
Statistical methods and analysis
For the analysis, frequency distributions, means, medians, 95% confidence intervals, and interquartile ranges were calculated. The proportion of postoperative deaths was established, and associations with clinical factors were assessed using the chi-square test, with significance set at a P value less than .05. All analyses were conducted with SPSS software, version 26.
Ethical considerations
Ethical approval for the study was secured from the MNH Ethics and Research Committee (reference number MNH/TRCU/Perm/2022/019). As this was a retrospective study, no patient identifiers were collected and a waiver of informed consent was granted. The research adhered to the principles of the Helsinki Declaration, and no external funding was sought for its completion.
Results
A total of 66 neonates underwent surgical exploration for suspected IA between January 2017 and January 2022. Of these, 5 were found intraoperatively to have necrotizing enterocolitis, 3 had intestinal stenosis, 2 presented with intestinal malrotation with Ladd’s bands only, and 1 had complicated meconium ileus; these cases were excluded from further analysis. Thus, 55 neonates met the study’s inclusion criteria. The male-to-female ratio among these cases was 1:2.4. Most infants (83%) were delivered vaginally, with a mean gestational age at birth of 35.67 weeks (95% CI 34.79–36.55 weeks) and a mean birth weight of 2359 g (95% CI 2206–2511 g). Prior to pediatric surgical consultation, provisional diagnoses included neonatal intestinal obstruction in 60% of cases, early onset neonatal sepsis in 29.1%, prematurity in 7.3%, and low birth weight in 3.6%. Of the 55 neonates enrolled, 23.6% (13/55) were born at the referral facility, while 76.4% (42/55) were referred from lower-level health care centers. The vast majority, 93% (51/55), originated from the Dar es Salaam region. The median interval from birth to surgical consultation for the clinical and radiological diagnosis of IA was 3 days, followed by a further 4 days before operative intervention. Consequently, the mean time from birth to laparotomy was 10 days (95% CI 8.32–11.7 days), with no neonates undergoing surgery within 24 hours of birth (Table 1). Before surgery, echocardiograms were performed in 11% (6/55) of the neonates. Of these, half showed normal findings while the other half revealed minor cardiac anomalies that were not clinically significant. All patients underwent routine blood tests, including complete blood counts and serum electrolytes. Electrolyte imbalances were identified in 23% (13/55) of the neonates, and anemia was present in 27.3% (15/55). Importantly, these abnormalities were corrected prior to surgical intervention.
Of the 55 neonates analyzed, intraoperative findings revealed that 5.5% (3/55) of atresias were located in the colon, 29% (16/55) in the duodenum, and 65.5% (36/55) in the JIA segment. All cases of CA were classified as Grosfeld Type II, characterized by blind-ending segments connected via a fibrous cord, and all cases underwent stoma creation.
Within the DA atresia group, 62.5% (10/16) were Grosfeld Type I lesions, denoting mucosal atresia with continuity of the bowel wall; notably, 2 of these cases had concurrent Ladd’s bands and annular pancreas (12.5% of all DA atresias). Type II atresia accounted for 18.8% (3/16), defined by 2 blind ends linked by a fibrous cord; among these, one case also exhibited Ladd’s bands with intestinal malrotation, though without midgut volvulus. The remaining 18.8% (3/16) were classified as Type III DA atresia, featuring bowel segments separated by a mesenteric defect. Of these, one had Ladd’s bands and intestinal malrotation without midgut volvulus, and another—representing 6% (1/16) of all DA atresias—presented with a distal single JIA atresia.
Among neonates with Type I DA atresia, surgical management was evenly divided: half (5/10) underwent duodenotomy with web excision, while the remaining half received a diamond-shaped duodeno-duodenostomy. For Type II lesions, 2 infants were managed with diamond duodeno-duodenostomy, and one underwent a duodeno-jejunostomy. All cases classified as Type III DA atresia were treated with duodeno-jejunostomy (Table 2).
Within the cohort of neonates diagnosed with JIA atresia, the distribution of lesion types was as follows: Type I, characterized by mucosal atresia with an intact bowel wall, was observed in 30.5% (11/36) of the cases. Type II, defined by 2 blind ends connected by a fibrous cord, accounted for 22% (8/36). Type IIIA, marked by completely separated bowel segments due to a mesenteric defect, represented 11% (4/36). Type IIIB, the so-called “apple-peel” or “Christmas tree” deformity, comprised 25% (9/36), while Type IV, involving multiple JIA atresias, was identified in 11% (4/36) of subjects. Malrotation was present in 36% of JIA atresia cases, distributed as follows: 2 cases in Type I, 3 in Type II, 4 in Type IIIA, and 2 in Type IV, none of which were associated with midgut volvulus. Within Type IIIB, malrotation was noted in 2 cases, one of which also exhibited midgut volvulus with concurrent intestinal perforation.
The surgical procedures performed for various types of JIA atresia include resection with end-to-end anastomosis (61%), enterotomy with web excision (22%), and diamond entero-enterostomy (5.5%). For neonates presenting with noncomplex JIA atresia—namely Types I, II, and IIIA—the standard surgical interventions included resection with end-to-end anastomosis, enterotomy with web excision, or a diamond-shaped entero-enterostomy. In cases classified as complex JIA atresia, particularly Type IIIB, 8 out of 9 infants underwent tapering of the proximal atretic segment followed by end-to-end anastomosis. The remaining patient in this subgroup, who presented with volvulus and intestinal perforation, required multiple resections and subsequent anastomoses. With Type IV atresia, 4 infants underwent resection of several atretic segments with end-to-end anastomosis, and one case required multiple anastomoses between affected segments. It is important to note that the final length of the bowel after surgical repair was not documented (Table 3).
The overall postoperative mortality rate in this cohort was 36.4% (20 out of 55 neonates). Stratified by type, mortality was 18% (3/16) among those with DA atresia, 44% (16/36) for JIA atresia, and 33% (1/3) for CA. Comparison between survivors and nonsurvivors revealed no statistically significant differences in gestational age at birth or time from birth to surgery. The mean gestational age at birth was 36.1 weeks (95% CI 34.9–37.3) in survivors vs 34.9 weeks (95% CI 33.5–36.3) in those who died; the mean time from birth to laparotomy was 9.2 days (95% CI 7.1–11.3) among survivors and 11.5 days (95% CI 8.5–14.5) among nonsurvivors. Notably, the mean birth weight was significantly higher in those who survived—2.39 kg (95% CI 2.21–2.57)—compared to those who died, who had a mean birth weight of 2.31 kg (95% CI 2.01–2.61) (P = .033). Among survivors, disturbances in electrolyte balance were identified in 38.5% (5/13), whereas these abnormalities were present in 61.5% (8/13) of neonates who died, a difference that reached statistical significance (P = .031).
When examining mortality by the location of the atresia, no statistically significant association was found. However, within the DA atresia group, mortality rates differed according to Grosfeld classification, with 0% (0/10) in Type I, 33% (1/3) in Type II, and 67% (2/3) in Type III, though this was not statistically significant (P = .38). Among neonates with JIA atresia, mortality was 25% (5/20) for noncomplex types and 64.7% for complex types, representing a statistically significant difference (P = 0.015). The presence of additional gastrointestinal anomalies did not have a statistically significant impact: for the nonsurvivors, 53% (8/15) had at least one gastrointestinal anomaly, compared to 46.7% (7/15) among the survivors (P = .331) (Table 4). Upon conducting binary logistic regression analysis of the factors linked to mortality, it was found that only complex-type JIA atresia demonstrated a statistically significant association with mortality (P = .014).
For those who survived, the length of the postoperative hospital stay required to achieve full tolerance of oral feeds ranged from 5 to 65 days, with a mean of 32 days (95% CI 27–36). One infant required a return to theater for wound dehiscence.
Discussion
The findings from this study at MNH reveal a postoperative mortality rate of 36.4% among neonates with IA, a figure that aligns with the higher spectrum of rates reported for sub-Saharan Africa (21.7–41%).2,6,7 This outcome is in marked contrast to the much lower mortality (around 5%) typically observed in high-income countries, and it highlights the persistent disparity in neonatal surgical care outcomes between resource-limited and well-resourced settings.5,6 The predominance of JIA atresias (65.5%) within the cohort is consistent with epidemiologic patterns seen globally.16 When outcomes are stratified by type, the data show the lowest mortality in Type I DA atresia (0%), intermediate rates in other DA types, and the highest mortality in complex JIA atresias (Types IIIB and IV), where mortality reached nearly 65%. These findings are in agreement with published literature that indicates complex lesions are associated with the gravest prognosis due to the technical challenges of surgery and the extensive perioperative care required.9,13,14
In examining factors associated with mortality, this study reinforces the significance of lesion complexity, low birth weight, and electrolyte imbalances prior to surgery. Complex JIA atresias proved to be the most lethal, underscoring the difficulties of managing multifocal or anatomically challenging lesions in neonates. The mean birth weight of survivors was significantly higher than that of nonsurvivors, supporting the well-established link between low birth weight and adverse surgical outcomes in this population.8,10 Electrolyte disturbances were also significantly more common among neonates who died, suggesting that preoperative correction and stabilization may be critical modifiable factors in improving survival. Notably, while there was a trend for higher mortality with longer delays from birth to surgical intervention, neither gestational age at birth nor time to surgery reached statistical significance in this cohort. The presence of additional gastrointestinal anomalies did not significantly affect survival in this group.
These findings suggest that targeted efforts to optimize preoperative care, most importantly correcting metabolic derangements and improving the nutritional status of neonates, could be beneficial. Additionally, reducing surgical delays through streamlined referral and diagnostic processes may offer further gains, even if these did not achieve statistical significance in this study. Targeted improvements in both surgical technique and postoperative management for complex JIA atresia hold promise for significantly enhancing survival and overall outcomes in this high-risk neonatal population.
The internal validity of the study is supported by clear inclusion criteria, systematic exclusion of confounding pathologies, and standardized classification of IA based on intraoperative findings. However, as a retrospective analysis, the study is inevitably limited by reliance on the completeness and accuracy of medical records. The lack of data on certain postoperative variables, such as final bowel length, as well as the modest sample size, may have affected the ability to detect some associations or fully describe longer-term outcomes.
In terms of generalizability, the study’s results are most applicable to tertiary referral centers in sub-Saharan Africa, where similar patient populations, health care infrastructure, and resource constraints exist. However, the findings may be less applicable to rural areas, other regions with different health care delivery systems, or to centers with greater access to specialized neonatal surgical and intensive care services.
Limitations of this study primarily relate to its retrospective design and the absence of some follow-up and postoperative outcome data, which may limit understanding of the full spectrum of morbidity and longer-term consequences of treatment. Furthermore, some potentially relevant variables, such as the quality and timing of nutritional support, were not assessed.
This study offers a critical perspective on neonatal IA outcomes within a major Tanzanian referral hospital, uniquely highlighting the significant challenges faced in resource-limited settings, challenges that are often overlooked in existing literature from high-income countries. The research fills a gap in global understanding by thoroughly stratifying atresia types, identifying lesion complexity, low birth weight, and preoperative electrolyte imbalances as key mortality factors, and documenting the stark contrast in survival rates compared to wealthier settings. Based on the findings, it is recommended that efforts be focused on early identification and referral of affected neonates, as well as establishing standardized protocols for preoperative optimization, particularly correction of electrolyte imbalances and support of nutritional status. Where possible, the interval from birth to surgery should be minimized. Finally, prospective data collection and research are needed to more precisely identify modifiable risk factors and to track long-term outcomes in this vulnerable population. These steps may collectively help narrow the outcome gap between resource-limited and high-income settings for neonates with IA.
Ethical Approval
Ethical approval for the study was secured from the MNH Ethics and Research Committee (reference number MNH/TRCU/Perm/2022/019).
Informed Consent
As this was a retrospective study, no patient identifiers were collected, and a waiver of informed consent was granted. The research adhered to the principles of the Helsinki Declaration.
Data Availability
Deidentified data are available on request
Conflict of Interest
None declared.
Funding
No external funding was sought for the completion of this research.
Acknowledgments
We would like to thank Dr Theresia W. Mbuli and Mr Kai M.N. for their invaluable contributions to this work.