Article Text
PDF
Abstract
Bilious vomiting in a neonate may be a sign of intestinal obstruction often resulting in transfer requests to surgical centres. The aim of this study was to assess the use of clinical findings at referral in predicting outcomes and to determine how often such patients have a time-critical surgical condition (eg, volvulus, where a delay in treatment is likely to compromise gut viability).
Methods 4-year data and outcomes of all term newborns aged ≤7 days with bilious vomiting transferred by a regional transfer service were analysed. Specificity, sensitivity, likelihood ratios, correlations, prior and posterior probability of clinical findings in predicting newborns with surgical diagnosis were calculated.
Results Of 163 neonates with bilious vomiting, 75 (46%) had a surgical diagnosis and 23 (14.1%) had a time-critical surgical condition. The diagnosis of a surgical condition in neonates with bilious vomiting was significantly associated with abdominal distension (χ2=5.17, p=0.023), abdominal tenderness (χ2=5.90, p=0.015) and abnormal abdominal X-ray findings (χ2=5.68, p=0.017) but not with palpation findings of a soft as compared with a tense abdomen (χ2=3.21, p=0.073). Abnormal abdominal X-ray, abdominal distension and tenderness had 97%, 74% and 62% sensitivity, respectively, with regard to association with an underlying surgical diagnosis. Normal abdominal X-ray reduced the posterior probability of surgical diagnosis from 50% to 16%. Overall, clinical findings at referral did not differentiate between infants with or without surgical or time-critical condition.
Conclusions We recommend that term neonates with bilious vomiting referred for transfer are prioritised as time critical.
- Paediatric Practice
- Neonatology
- Paediatric Surgery
Statistics from Altmetric.com
What is already known on this topic
-
In a neonatal surgical centre, a surgical cause is found in 26%–51% of neonates with bilious vomiting.
-
Bilious vomiting in a neonate is a recognised presentation of intestinal obstruction.
What this study adds
-
This study provided clinicians with a clearer idea of the likelihood of a surgical diagnosis based on clinical and abdominal X-ray findings.
-
There is no association between clinical signs or abdominal X-ray findings of neonates with time-critical surgical pathologies and those without.
-
Therefore, it is recommended that the transfer of term neonates referred with bilious vomiting be prioritised.
Background
Bilious vomiting in a neonate may be a presenting sign of intestinal obstruction and can be easily missed.1 Most neonates developing bilious vomiting are not born in neonatal surgical centres and are transferred to such facilities.2 These infants are often transferred beyond their local networks.3 Initial detection, assessment and management are usually performed by paediatric and neonatal staff in non-surgical centres.4 If not managed in a timely manner, some of the underlying causes may result in gut compromise associated with death or long-term sequelae such as short gut syndrome.
Most neonatal units in the UK have access to specialised transfer services to facilitate movement of these infants into surgical centres. These transfer requests are triaged and prioritised accordingly. Currently, such referrals are not categorised nationally as time critical, although individual services may choose to label them as such and respond accordingly.
Previous studies involving neonates with bilious vomiting have mainly looked at the outcome of these infants from a large neonatal unit5 or the tertiary neonatal surgical unit perspective.6–8 The burden of neonates with bilious vomiting on neonatal transfer services (NTSs) and the outcome of these transfers are not known. Furthermore, it is unclear whether all neonates referred for transfer with bilious vomiting need to be seen and transferred to a tertiary neonatal surgical centre and whether these should be nationally categorised as time critical transfers.
The aim of this study was to investigate the use of clinical findings and investigations available at the time of transfer to predict the outcome of neonates with bilious vomiting and to identify those with time critical pathologies in a cohort of babies transferred by a dedicated, regional NTS.
Methods
Term neonates (37 weeks and above) with bilious vomiting transferred at ≤7 days of age between 1 January, 2007, and 31 December, 2010, by the London NTS were identified from the transfer team database.
The clinical and radiological findings were recorded as assessed by the referring team or the neonatal transfer team. These included the presence or absence of abdominal distension, abdominal tenderness, abdominal palpation findings of either a soft or tense abdomen and the abdominal X-ray (AXR) findings of being normal or abnormal. The post-transfer outcomes were obtained from surgical centres as part of routine follow-up. Data were collected on patient demographics, blood gas parameters, radiological findings and diagnostic outcomes from neonatal surgical units. Barts and the London clinical effectiveness unit approved the study.
Neonates were classified to have a surgical diagnosis if they required surgery or had a condition that was likely to require surgery. They were ascribed to have a time-critical condition if they had a surgical diagnosis that may result in compromise of bowel viability (eg, malrotation and volvulus) or had developed postnatal gut perforation. The time from referral to transfer team mobilisation was defined as dispatch time. The period from referral to arrival of the transfer team at the referring unit was considered as response time.
Data were summarised as median and ranges. Association between categorical variables was analysed using χ2 test. Sensitivity, specificity, predictive values and likelihood ratios for both clinical signs and AXR findings were calculated for neonates with surgical diagnoses.9 To assist decision making, positive likelihood ratio (LR+), negative likelihood ratio (LR−) and prior probability of the disease were used to calculate posterior probability for surgical diagnosis and results were visually displayed using ‘Fagan's Nomogram’.10
Results
Three thousand eight hundred and thirty-one neonates were transferred by NTS between the periods of 1 January 2007 to 31 December 2010. Of these, 1085 were full-term neonates aged ≤7 days who were transferred for various reasons including those with bilious vomiting. A total of 163 term infants with bilious vomiting were transferred to six neonatal surgical centres for assessment and management of suspected intestinal obstruction. The median (range) gestation, birth weight and age at transfer were 39.9 (37–42) weeks, 3.3 (1.8–5.0) kg and 35.3 (2.3–166) h, respectively.
Of 163 neonates with bilious vomiting, 75 (46%) had a surgical diagnosis confirmed and 23 (14.1%) had a time-critical surgical condition (table 1). Information on the presence or absence of clinical findings and abnormality of AXR findings were available for abdominal distension (148/163), abdominal tenderness (68/163), abdominal palpation findings (117/163) and abnormality of AXR (143/163) neonates. The prior probability of a surgical diagnosis was 47% for abdominal distension, 50% for abdominal tenderness, 39% for abdominal palpation findings and 47% for abnormal AXR findings. The diagnosis of a surgical condition in neonates with bilious vomiting was significantly associated with abdominal distension (χ2=5.17, p=0.023), abdominal tenderness (χ2=5.90, p=0.015) and abnormal AXR findings (χ2=5.68, p=0.017) but not with palpation findings of a soft as compared with a tense abdomen (χ2=3.21, p=0.073).
List of major outcome diagnosis
Table 2 shows the sensitivity, specificity, positive predictive value, negative predictive value, LR+ and LR− of abdominal distension, tenderness, abdominal palpation findings (soft or tense) and AXR appearance for the diagnosis of a surgical condition. Abnormal AXR findings, presence of abdominal distension and a tender abdomen had the sensitivity to detect babies with surgical diagnoses in 97%, 74% and 62%, respectively. The effect of presence or absence of clinical signs and abnormal AXR findings on the prior and posterior probability of a surgical diagnosis was evaluated with a Fagan's nomogram. The presence of firm or tense abdomen increases the probability of a surgical diagnosis from 63% to 71%, a tender abdomen only increases the probability from 64% to 66%. A normal AXR reduces the probability of surgical diagnosis from 50% to 16%.
Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), positive likelihood ratio (LR+) and negative LR (LR−) of abdominal distension, tenderness, feel of abdomen (soft versus firm or tense) and abdominal X-ray (AXR) for a diagnosis of surgical condition
The combination of clinical findings and AXR findings for prediction of a surgical diagnosis was as follows: there were five neonates with bilious vomiting who did not have abdominal distension and AXR findings were considered normal. None of these neonates had a surgical diagnosis (LR− of 0 (95% CI 0 to 2.6)) and had a posterior probability of 0% (95% CI 0% to 60%). A combination of other clinical and AXR findings only marginally altered the posterior probabilities of a surgical diagnosis (table 3).
Combination of clinical findings and abdominal X-ray (AXR) findings and their effect on likelihood ratios (LR), prior probabilities and posterior probabilities in newborn with bilious vomiting
Dispatch and response times
Median (IQR) dispatch time for transfer team was 31 (16–106) min, whereas response time was 70 (50–131) min. The dispatch times were longer than 60 min in 34% of transfers of babies with bilious vomiting.
Twenty-three (14.1%) infants with bilious vomiting were found to have a surgical diagnosis warranting a time critical transfer response. The principal diagnoses in these neonates were malrotation (7), malrotation with volvulus (6), volvulus (5), volvulus with atresias (2) and intestinal perforation (3). There were no significant associations between neonates who were found to have time critical pathologies and abdominal distension, palpation findings, abdominal tenderness and abnormal AXR findings when compared with those who did not have time critical pathologies. Median (IQR) response time for these neonates was 31 (20–120) min and was similar to other neonates with bilious vomiting who were considered to not have conditions requiring time-critical transfers. Eight of these 23 neonates had a response time of >60 min.
Malrotation and volvulus versus other neonates with bilious vomiting in terms of signs and AXR findings
There were 20 newborns that were diagnosed to have malrotation, volvulus or both. The proportion of neonates with these diagnoses that had abnormal findings on clinical examination or abnormal AXR findings was not significantly different from those neonates with bilious vomiting due to other causes.
Discussion
This is the largest reported series looking at the outcome of neonates with bilious vomiting. In our study of term infants referred for transfer to any one of the six surgical centres in the London region, we have shown that bilious vomiting is a clinical sign of underlying surgical pathology in 46% of cases and time critical pathologies in 14% of cases. The proportion of babies requiring surgical interventions was similar to a UK study reported by Godbole (38%) and Kao (51.1%) from Taiwan, whereas a North-American study from a tertiary neonatal unit and an Australian study have reported lower rates of 20% and 26.6%, respectively. An Iranian study by Alehossein et al11 reported a series of 23 babies admitted over an 18-month period to a children's hospital with bilious vomiting where 18 (78.3%) had a surgical diagnosis. These differences in likelihood of surgical diagnosis in neonates with bilious vomiting may reflect variation in practices and thresholds for further investigation and referral.
We found a significant association between the presence of surgical diagnoses and abdominal distension, tenderness and abnormal X-ray findings. Although abnormal X-ray findings were associated with surgical pathology, there was no association with abdominal distension and abdominal tenderness in the study by Kao.6 Although the presence of abdominal distension or tenderness accompanied by abnormal AXR findings increased the likelihood of an underlying surgical cause, these findings were not diagnostic. In our study, abnormal AXR findings had a sensitivity of 97% and specificity of only 14% for a surgical diagnosis. One previous study from a neonatal surgical unit had reported the use of AXR findings for detection of surgical conditions and had a similar sensitivity of 96% and a higher specificity of 90%. The lower specificity of AXR findings in our study may reflect the interpretation of AXR findings by a large number of junior staff from the referring hospital and transfer team rather than by the specialist surgical team as in the previous study.8
The absence of clinical signs and a normal AXR examination reduces the likelihood of a surgical diagnosis. No neonates with bilious vomiting in this study were found to have a surgical diagnosis if there was no abdominal distension with normal AXR findings. This finding needs to be interpreted with caution because there were only five infants in this group. A previous study by Godbole and Stringer8 did report one infant with malrotation who had neither abdominal signs nor abnormalities on a plain abdominal radiograph.
One of the limitations of this study is that the denominator data are based on cases of bilious vomiting referred for transfer. Neonates presenting with bile-stained vomiting who were not referred for transfer as well as those born in surgical centres are not represented in our dataset. There is still some evidence to suggest that not all cases of neonates presenting with bilious vomiting, particularly a single episode are referred for a surgical evaluation.9
We also do not have information relating to the presence of abnormal antenatal ultrasound findings that may have suggested intestinal pathology. This also may affect clinical decision making and may influence the threshold for referral.
There was no association between clinical signs and AXR findings with time-critical pathologies: malrotation or volvulus. Clinical presentation at the time of referral and transfer did not allow differentiation between surgical and non-surgical pathology and in particular surgical conditions with increased risk of gut compromise such as gut malrotation and volvulus and those without. Therefore, all term neonates with bilious vomiting referred for transfer should be subject to a time-critical team dispatch on the basis of a potential surgical emergency and moved by the transfer team urgently for a specialist surgical opinion.
The Neonatal Transport Group (UK) has devised a dataset of nationally agreed time critical diagnoses for benchmarking purposes. Neonates referred for urgent transfer who fall into one of these categories are expected to be subject to a team dispatch of <1 h. A transfer service's responsiveness to the most critically ill babies is measured in this way, compared between teams and can be used in performance discussions with commissioning bodies. At present, neonates referred for transfer with bilious vomiting are not ascribed ‘time-critical status’ and should be considered for this.
Footnotes
-
Contributors SM and AS designed data collection tools, monitored data collection for the whole trial, wrote the statistical analysis plan, analysed the data, drafted and revised the paper. They are the guarantors. NR, SK and HCW analysed the data, drafted and revised the paper. PS and BB initiated the project, designed data collection tools and collected data.
-
Competing interests None.
-
Provenance and peer review Not commissioned; externally peer reviewed.