Gastric tubes are commonly used in hospitalized neonates, infants, and children to provide nutrition and medication (1), and to perform gastric decompression after intestinal obstruction or surgery (2). To ensure correct placement after insertion is important to reduce risks of misplacement and subsequent serious complications.
Although the prevalence of gastric tube placement errors in children is difficult to confirm owing to the differing definitions across studies (3), tube placement error rates have been reported between 20.9% and 43.5% in children (4). In addition, Quandt et al. (5) reported that 59% (179/303) of gastric tubes had been placed incorrectly in neonates. Gastric tube misplacement puts neonates, infants, and children at risk for a variety of complications. A misplaced gastric tube within the trachea or lungs could lead to adverse effects including respiratory distress, tracheal or pulmonary perforation and pneumothorax, empyema, pulmonary hemorrhage, chemical pneumonitis, and pneumonia (3). Gastric tubes misplaced into the esophagus increase the risk of apnea, bradycardia, oxygen desaturations, and aspiration because of the proximity of the trachea (1). Tubes placed unintentionally in the duodenum or jejunum can cause malabsorption, inadequate weight gain, dumping syndrome associated with abdominal pain and distension, hypoglycemia, and diarrhea (1,6).
There are several methods that have been used either individually or in combination to verify gastric tube placement in neonates, infants, and children. pH testing is commonly used as the first-line method to verify gastric positioning in both adult and pediatric populations. The National Patient Safety Agency (7) suggested a pH lower than 5.5 as an indicator of correct gastric tube position in children and infants, and this was supported by Meert et al.'s (8) study, which reported that gastric aspirates in critically ill infants often have pH values of 5.5 or less. Regarding the neonatal population, although the National Patient Safety Agency (9) and the Western Health and Social Care Trust (10) recommended using pH (with an upper pH limit of 5.5) to verify gastric placement, the ability of pH to determine gastric tube position has not yet been determined in the existing literature (11). Moreover, pH testing can be challenging or inaccurate because of difficulties in obtaining aspirate from some small-bore tubes (12) and the administration of antacids, H2 antagonists or proton pump inhibitors, which are likely to result in a stomach pH greater than 5.5 (13). Other methods to assess gastric tube position include color of aspirate (1), auscultation, carbon dioxide testing (14), ultrasound (15), testing of bilirubin (4), pepsin, or trypsin (4). However, there is no consensus on the diagnostic performance of these methods in the populations of neonates, infants, and children.
Although these methods have been used in clinical practice, radiological examination of the chest and abdomen is still considered the gold standard verification technique (4,16,17) and is recommended for verification of gastric tube position in pediatric patients by professional organizations (7,10,13,18,19). Given the importance of minimizing radiation exposure for vulnerable populations of neonates, infants, and children, as well as the expense and feasibility of transports to radiology departments, it is imperative to explore the diagnostic performance of these alternative methods so that health professionals can make an informed decision about the best methods to use other than radiology. The aim of the review was therefore to evaluate diagnostic performance of methods used to assess gastric tube placement verification in neonates, infants, and children (see Supplementary Strengths and limitations of the review, Supplemental Digital Content 1, http://links.lww.com/AJG/B260).
A systematic review and meta-analysis was conducted using the methods outlined in the Cochrane Handbook for Reviews of Diagnostic Test Accuracy (20) and in accordance with the Preferred Reporting Items for a Systematic Review and Meta-analysis of Diagnostic Test Accuracy Studies (21). This systematic review protocol was registered with the International Prospective Register of Systematic Reviews (PROSPERO registration number: CRD42016041354).
Criteria for considering studies
Studies on the diagnostic accuracy of gastric tube placement confirmation methods compared with the gold standard of x-ray visualization were included. Methods of gastric tube placement verification included were pH testing, color of aspirate, auscultation, carbon dioxide testing, ultrasound, bilirubin, pepsin, trypsin, separately or in combination with the above methods. Studies on fluoroscopically, endoscopically, electromagnetically, and ultrasound-guided tube placement and verification were excluded, in addition to studies in which diagnostic accuracy of tube position confirmation methods (e.g., specificity, or sensitivity) was not recorded or could not be calculated. Participants included neonates (including premature neonates), infants, and children receiving gastric tube placement in any care setting for any reason. The reference standard was x-ray of the chest or abdomen.
Different search strategies were developed according to search features of different databases (see Supplementary Search strategies, Supplemental Digital Content 2, http://links.lww.com/AJG/B257) with consultation and advice from a librarian. The following English database were searched: Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library, 2018), MEDLINE (OvidSP) (1946 to February 2018), EMBASE (1976 to February 2018), Web of Science, Science Citation Index Expanded, and Conference Proceedings Citation Index-Science (Web of Science Core Collection) (1990 to February 2018), and Cumulative Index to Nursing and Allied Health Literature (CINAHL) (via EBSCO) (1977 to February 2018). The search terms (subject key words and text words) were enteral nutrition, intubation, gastric tube, child, infant, neonate, and their appropriate truncation symbols. Three Chinese databases were also searched: Chinese WanFang Data, China Journal Net, and the Chinese Biomedical Literature Database, from inception to February 2018. No restrictions were placed on date of publications and language. Reference lists of all identified studies for inclusion were searched and screened to identify any further potential papers.
Selection of studies
Two authors (T.L. and Y.S.) independently screened the titles and abstracts of the articles identified in the search strategy. Full text of the articles that potentially met our inclusion criteria or where there was insufficient information to make a decision regarding inclusion were retrieved and assessed for inclusion independently by T.L. and Y.S. Disagreements regarding eligibility were resolved by the third person (D.H.).
The 2 reviewers (T.L. and Y.S.) independently extracted data on each included study including patient demographics, sample size, study country, study design, index tests, and methodological quality. Then, both reviewers extracted the following data to construct 2 × 2 contingency tables, i.e., true positives, true negatives, false positives, and false negatives.
Assessment of methodological quality
Risk of bias of included studies was assessed using the QUADAS-2 tool as outlined by Whiting et al. (22) and recommended by the Cochrane Diagnostic Test Accuracy Group. Assessment of methodological quality was carried out by 2 reviewers independently (T.L. and Y.S.); discrepancies were resolved by a third person (Y.L.). One modification was made specific to this review in the domain of patient selection (see Supplementary Differences between protocol and review, Supplemental Digital Content 3, http://links.lww.com/AJG/B258).
Statistical analysis and data synthesis
Study estimates of sensitivity and specificity for each index test were plotted, both in forest plots and the receiver operating characteristic (ROC) using Review Manager 5 software. To evaluate the summary sensitivity and specificity of each test, a meta-analysis using different models were planned. For the pH index test, different cutoff points chosen in the included studies were expected because of lack of validated cutoffs. Therefore, the hierarchical summary ROC model was used, which allows for the possibility of variation in thresholds between studies, while also accounting for variations within and between studies and any potential correlation between sensitivity and specificity. Results were entered from Stata/SE version 15 into Review Manager 5 to provide plots of the mean or summary point(s) and confidence region(s), superimposed on the study-specific estimates of sensitivity and specificity. For other index tests with recognized cutoff values or without threshold effects, fixed-effect meta-analysis in the absence of substantial heterogeneity and bivariate logit normal random-effects model for meta-analyses for studies with substantial heterogeneity were planned.
Investigations of heterogeneity
Heterogeneity was explored by visually examining the forest plots of sensitivities and specificities and the ROC plots for each index test. Exploration of potential sources of heterogeneity through subgroup analysis and meta-regression (including study country, study design, sample size, patient age) were planned.
The “leave-one-out” procedure was planned to assess the impact of each study on the meta-analysis results.
Results of search
Electronic searches resulted in identification of 11,889 references (English n = 896, Chinese n = 10,993) from the following database: Cochrane Central Register of Controlled Trials (CENTRAL) (n = 17), MEDLINE (OvidSP) (n = 483), EMBASE (n = 197), Web of Science (n = 9), CINAHL (via EBSCO) (n = 190), Chinese WanFang Data (n = 5,243), China Journal Net (n = 2,190), and the Chinese Biomedical Literature Database (n = 3,560). Ten further studies were identified from reference lists of the identified studies. After exclusion of duplicates, 6,920 references remained and a further 6,867 excluded after assessment of title or abstract, or both. Main reasons for exclusion were article type (reviews, case reports, or case series), animal research, studies concerning tube placement device and tube insertion or re-insertion procedures, and no x-ray reference standard. One potential study (23) was excluded as the data were insufficient to construct a 2 × 2 contingency table. Authors were contacted in an attempt to obtain data but with no reply. This resulted in 8 studies evaluating 9 index tests in 2 languages being included in this review (English n = 7, Chinese n = 1). Figure 1 shows the flow of references through the selection process.
Description of included studies
All included studies were observational studies (24–31) of gastric tube placement confirmed by various methods compared with x-ray as the reference standard with diagnostic accuracy of methods (e.g., specificity, or sensitivity) recorded or calculated. Overall, 911 neonates, infants, and children aged from birth to 18 years were enrolled in 4 different countries; United States (n = 5) and one each in Australia, China, and Turkey. A mean of 114 participants were enrolled per study (range: 7–404). Of the 8 included studies, 9 different methods were used for gastric tube placement confirmation. A summary of the included studies is shown in Table 1.
Methodological quality of included studies
Overall, the quality of the included studies was moderate, as illustrated in the QUADAS-2 results (Figures 2 and 3). No concerns regarding applicability of patient selection bias or interpretation of index test and reference test results were raised. However, in the domain of patient selection, high risk of bias for patient sampling was scored in 4 studies (24,26,27,31). In the domain of the index test, 1 study (31) performed index test interpretation after radiology confirmation; therefore, the risk of bias was rated as unclear because results of radiology could potentially have biased interpretation of the index test results. Most studies (25–29) were allocated an unclear risk of reference standard bias due to unclear explanations of whether reference standard results were interpreted with knowledge of index tests results. Two studies (25,29) had a high risk in the flow and timing domain of methodological quality. One study (25) reported an inappropriate interval (12 hours) between the index test and reference test. Another study (29) did not report the interval between the index and reference test and did not include all selected patients in reference standard confirmation and data analysis.
Six studies including 7 data sets with a total of 457 participants included assessment of the diagnostic accuracy of pH testing of aspirate for gastric tube position confirmation. The studies used different cutoff values (Figure 4) including pH ≤ 4 (25,29), pH ≤ 5 (26,28), and pH ≤ 6 (28,30,31). A meta-analysis revealed the summary sensitivity and specificity of 0.77 (95% CI 0.56–0.90) and 0.42 (95% CI 0.16–0.73) for cutoffs of ≤4, ≤5, and ≤ 6. Forest plot (Figure 4) and the ROC plot (Figure 5) showed a high degree of heterogeneity for diagnostic estimates, ranging from 0.38 to 0.97 for sensitivity and from 0.00 to 1.00 for specificity.
Color of aspirate
Two studies (28,31) explored the accuracy of color testing of aspirate. The estimates of sensitivity and specificity are 0.53 (95% CI 0.46–0.60) and 0.47 (95% CI 0.32–0.64) for color of white/green/tan and 0.93 (95% CI 0.80–0.98) and 0.71 (95% CI 0.29–0.96) for color of clear/colorless/off-white/green/brown (Figure 6).
Combined test of pH and color
Two studies (28,31) including 3 data sets explored the accuracy of combined test of pH and color of aspirate, of which Ellett et al. (28) first used combined test of pH ≤ 5 and color white/green/tan; then pH ≤ 6 and color white/green/tan; and Westhus (31) pH < 6 and color clear/colorless/off-white/green/brown. The estimates of sensitivity ranged between 0.34 and 0.71 and that of specificity between 0.40 and 1.00 (Figure 6).
Two studies (25,30) evaluated the accuracy of auscultation, reporting estimates of relatively high sensitivities of 0.84 (95% CI 0.64–0.95) and 1.00 (95% CI 0.86–1.00) and of specificities of 0.08 (95% CI 0.00–0.36) and 0.00 (95% CI 0.00–0.26) (Figure 6).
Carbon dioxide testing <15 mm Hg
Two studies (26,27) assessed the accuracy of carbon dioxide testing, which reported estimates of high sensitivity of 0.97 and 1.00, with specificity not estimated (Figure 6).
Ultrasound, bilirubin <5 mg/dL, trypsin <50 μg/mL, pepsin ≥20 μg/mL
One study (24) evaluated the accuracy of ultrasound (21 participants), showing a sensitivity of 1.00 (95% CI 0.84–1.00) and no specificity estimated (Figure 6). Ellett et al. (26) evaluated the accuracy of bilirubin testing (62 participants), showing a sensitivity of 0.97 (95% CI 0.88–1.00) and no specificity estimated (Figure 6). One study (31) (56 participants) assessed trypsin testing with a sensitivity of 0.90 (95% CI 0.78–0.97) and a specificity of 0.71 (95% CI 0.28–0.96), and pepsin testing with a sensitivity of 0.69 (95% CI 0.55–0.82) and a specificity of 0.71 (95% CI 0.29–0.96), respectively (Figure 6).
Investigations of heterogeneity
As seen in the forest plots (Figure 4), which displayed both sensitivity and specificity of pH index test, the between-study heterogeneities were substantial. Although attempts were made to investigate sources of heterogeneity of included studies, there were insufficient studies to allow subgroup analysis and meta-regression with covariates.
The leave-one-out analyses showed the stable summary diagnostic odds ratios of pH testing were in a range from 1.13 (95% CI 0.98–1.29) to 1.21 (95% CI 1.05–1.41), suggesting that no 1 study unduly influenced the results.
Summary of main results
The diagnostic performance of 9 methods of gastric tube placement detection were included in this review. Only pH testing was assessed in a sufficient number of studies for a meta-analysis, which revealed a moderate summary sensitivity and a low summary specificity for pH cutoffs ≤6. This finding suggests that pH ≤ 6 may not be sufficiently accurate to detect the gastric tube position, which does not support recommendations from multiple guidelines (7,9,10). However, because of the heterogeneity among studies, including 3 cutoff values, 2 types of instruments of pH meter and pH paper, different populations of fasting and fed participants, the conclusion needs to be considered with caution. Meanwhile, data were not separately pooled for the sensitivity and specificity of pH ≤ 4, pH ≤ 5, and pH ≤ 6 owing to the limited studies of each test. A comparison in diagnostic accuracy of pH testing with different cutoff values therefore could not be achieved.
Regarding other index tests that were not subjected to meta-analysis, although color of clear/colorless/off-white/green/brown testing and trypsin testing <50 μg/mL (31) showed good diagnostic performance, the findings should be used with caution because of the single study and wide confidence intervals of specificities. Ultrasound in 1 study (24) and bilirubin testing <5 mg/dL in another study (26) illustrated high sensitivities, suggesting these 2 index tests yield good diagnostic performance in predicting correct gastric placement. Future studies concerning these 2 methods are needed to support current findings.
Auscultation in 2 studies (25,30) demonstrated relatively high sensitivities, indicating its potential reliable role in detecting correct gastric tube placement, which conflicts with questions about the usefulness of this method in adult population (3,12).There are also guidelines (18,19) that have discouraged the use of auscultation in the pediatric population, but the recommendations are not evidence-based and present no grading of evidence level. In clinical settings, the vast majority of pediatric nurses check gastric tube placement by auscultation method (3,17), with a survey reporting 98% neonatal nurses used this method (11). It is hence suggested to take into reconsideration of its role in gastric confirmation, although the relatively small samples in this review may limit the strength of its reliability.
Implications for practice
The conclusion of the inability of pH ≤ 6 for detection of gastric tube position in neonates, infants, and children can only be drawn with caution because of the heterogeneity of studies. The paucity of data and methodological variations in studies make it difficult to arrive at any conclusions regarding the diagnostic test accuracy of pH ≤ 4 or 5 and other index tests in detection of gastric tube placement. Clinical practice related to the diagnostic tests used will continue to be dictated by local practices and preferences, availability of specific supplies for gastric tube insertion, and cost factors, until stronger evidence becomes available.
Implications for research
Well-designed studies to strengthen current evidence are recommended. These include (i) cross-sectional studies of comparison in pH values in different aspirates of respiratory tract, stomach, and intestine to determine the best cutoff value to be used in gastric tube position confirmation; (ii) diagnostic accuracy test of ultrasound performed by different professionals, auscultation, bilirubin testing to differentiate tube position among respiratory tract, esophagus, stomach, post-pyloric position, and (iii) diagnostic accuracy test of carbon dioxide testing to differentiate positioning in the gastrointestinal tract and respiratory tract.
It is further recommended that researchers use a study design that adheres to the Standards for Reporting of Diagnostic Accuracy recommendations and incorporate the QUADAS checklist, with particular focus on the domains of patient selection, reference standard, and flow and timing.
CONFLICTS OF INTEREST
Guarantor of the article: Tian Lin, MM.
Specific author contributions: T.L.: conceptualized and designed the study in partnership with Y.S. and D.H., conducted the literature search, collected data, carried out the initial analyses, and drafted the initial manuscript. D.H. and W.G.: made substantial contributions to all aspects of the writing of the manuscript, which included contribution to conception, design, analysis and interpretation of the article, and review and revision of the manuscript. Y.S., X.-L.L., and X.Q.Q.: participated in the literature search, screening, data collection, analysis, and review and revision of the manuscript. Y.-T.L. and K.C.: led the systematic review, conceptualized and designed the study, reviewed and revised the manuscript, and supervised and provided mentorship throughout all stages of the project and writing of the manuscript.
Financial support: None.
Potential competing interests: None.
WHAT IS KNOWN
- ✓ pH testing is the first-line method to verify gastric positioning in pediatric populations.
- ✓ Diagnostic performances of other methods to assess gastric tube position in neonates, infants, and children is unknown.
WHAT IS NEW HERE
- ✓ pH ≤ 6 may not be sufficiently accurate to detect gastric tube placement in neonates, infants, and children.
- ✓ Diagnostic performance of pH ≤ 4 or 5 and other methods cannot be determined in the light of current evidence.
- ✓ Clinical practice related to the diagnostic tests will continue to be dictated by local preferences.
1. Clifford P, Heimall L, Brittingham L, et al. Following the evidence: Enteral tube placement and verification in neonates and young children. J Perinat Neonatal Nurs 2015;29(2):149–61.
2. Phillips NM. Nasogastric tubes: An historical context. Medsurg Nurs 2006;15(2):84–8.
3. Farrington M, Lang S, Cullen L, et al. Nasogastric tube placement verification in pediatric and neonatal patients. Pediatr Nurs 2009;35(1):17–24.
4. Ellett MLC. What is known about methods of correctly placing gastric tubes in adults and children. Gastroenterol Nurs 2004;27(6):253–9.
5. Quandt D, Schraner T, Ulrich BH, et al. Malposition of feeding tubes in neonates: Is it an issue? J Pediatr Gastroenterol Nutr 2009;48(5):608–11.
6. Taylor SJ. Confirming nasogastric feeding tube position versus the need to feed. Intensive Crit Care Nurs 2013;29(2):59–69.
7. National Patient Safety Agency. Reducing the harm caused by misplaced naso and orogastric feeding tubes in babies under the care of neonatal units (http://www.nrls.npsa.nhs.uk/resources)
(2015). Accessed August 1, 2017.
8. Meert KL, Caverly M, Kelm LM, et al. The pH of feeding tube aspirates from critically ill infants. Am J Crit Care 2015;24(5):72–7.
9. National Patient Safety Agency. Reducing the Harm Caused by Misplaced Nasogastric Feeding Tubes in Adults, Children and Infants (https://www.nrls.npsa.nhs.uk/alerts
) (2011). Accessed August 1, 2017.
10. Western Health and Social Care Trust. Reducing harm caused by misplaced NG and OG feeding tubes (http://www.westerntrust.hscni.net/pdf/Reducing_Harm_Caused_by_Misplaced_NG_and_OG_Feeding_Tubes_-_April_2017.pdf)
(2017). Accessed September 9, 2018.
11. Parker LA, Withers JH, Talaga E. Comparison of neonatal nursing practices for determining feeding tube insertion length and verifying gastric placement with current best evidence. Adv Neonatal Care 2018;18(4):307–17.
12. Irving SY, Lyman B, Northington L, et al. Nasogastric tube placement and verification in children: Review of the current literature. Nutr Clin Pract 2014;29(3):267–76.
13. Guidelines and Audit Implementation Network. Guidelines for Caring for an Infant, Child, or Young Person Who Requires Enteral Feeding (https://www.rqia.org.uk/RQIA/files/4f/4f08bb34-7955-49ea-adf1-9de807d3da66.pdf
) (2015). Accessed July 22, 2018.
14. Gilbert RT, Burns SM. Increasing the safety of blind gastric tube placement in pediatric patients: The design and testing of a procedure using a carbon dioxide detection device. J Pediatr Nurs 2012;27(5):528–32.
15. Lin T, Gifford W, Lan Y, et al. Diagnostic accuracy of ultrasonography for detecting nasogastric tube (ngt) placement in adults: A systematic review and meta-analysis. Int J Nurs Stud 2017;71:80–8.
16. Cohen MD, Ellett MLC, Perkins SM, et al. Accurate localization of the position of the tip of a naso/orogastric tube in children: Where is the location of the gastro-esophageal junction? Pediatr Radiol 2011;41(10):1266–71.
17. Longo MA, Society of Pediatric Nurses (SPN) Clinical Practice Committee, SPN Research Committee. Best evidence: Nasogastric tube placement verification. J Pediatr Nurs 2011;26(4):373–6.
20. Deeks JJ, Bossuyt PM, Gatsonis C. Cochrane Handbook for Systematic Reviews of Diagnostic Test Accuracy Version 1.0. The Cochrane Collaboration (http://srdta.cochrane.org)
(2010). Accessed June 15, 2014.
21. McInnes MDF, Moher D, Thombs BD, et al. Preferred reporting items for a systematic review and meta-analysis of diagnostic test accuracy studies: The PRISMA-DTA Statement. JAMA 2018;319(4):388–96.
22. Whiting PF, Rutjes AW, Westwood ME, et al. QUADAS-2: A revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med 2011;155(8):529–36.
23. Metheny NA, Pawluszka A, Lulic M, et al. Testing placement of gastric feeding tubes in infants. Am J Crit Care 2017;26(6):466–73.
24. Atalay YO, Aydin R, Ertugrul O, et al. Does bedside sonography effectively identify nasogastric tube placements in pediatric critical care patients? Nutr Clin Pract 2016;31(6):805–9.
25. Ellett MLC, Beckstrand J. Examination of gavage tube placement in children. J Soc Pediatr Nurses 1999;4(2):51–60.
26. Ellett MLC, Croffie JM, Cohen MD, et al. Gastric tube placement in young children. Clin Nurs Res 2005;14(3):238–52.
27. Ellett MLC, Woodruff KA, Stewart DL. The use of carbon dioxide monitoring to determine orogastric tube placement in premature infants: A pilot study. Gastroenterol Nurs 2007;30(6):414–7.
28. Ellett MLC, Cohen MD, Croffie JM, et al. Comparing bedside methods of determining placement of gastric tubes in children. J Spec Pediatr Nurs 2013;19(1):68–79.
29. Stock A, Gilbertson H, Babl FE. Confirming nasogastric tube position in the emergency department: pH testing is reliable. Pediatr Emerg Care 2008;24(12):805–9.
30. Wang W, Zuo Z. The use of pH testing of aspirate to determine gastric tube placement in children. J Chin Nurs 2008;43(9):835–6.
31. Westhus N. Methods to test feeding tube placement in children. Am J Matern Child Nurs 2004;29(5):282–7.