Diagnostic accuracy of ultrasonography for detecting nasogastric tube (NGT) placement in adults: A systematic review and meta analysis

1. Review
Diagnostic accuracy of ultrasonography for detecting nasogastric tube
(NGT) placement in adults: A systematic review and meta analysis
Lin Tiana
, Gifford Wendyb,1
, Lan Yutaoa
, Qin Xiuqunc
, Liu Xuelianc
, Wang Juana
,
Yang Bipinga
, You Tianhuia,
*, Chen Kena,
*
a
School of Nursing, Guangdong Pharmaceutical University, 283 Jianghai Avenue, Haizhu Distrct, Guangzhou, China
b
School of Nursing, Faculty of Health Sciences, University of Ottawa, Ottawa, Ontario, Canada
c
Nursing Department, Third Affiliated Hospital of Sun Yat-sen University, 600 Tianhe Road, Tianhe District, Guangzhou, China
A R T I C L E I N F O
Article history:
Received 26 May 2016
Received in revised form 1 March 2017
Accepted 8 March 2017
Keywords:
Nasogastric tube
Ultrasonography
Adult
Systematic review
A B S T R A C T
Objective: To review the evidence on diagnostic accuracy of ultrasonography for detecting correct
nasogastric tube (NGT) placement in adults compared to X-ray as the reference standard.
Methods: This is a systematic review and meta-analysis of observational studies, searched in the literature
between 1961 and 2015. We included studies which compared the diagnostic accuracy of
ultrasonography detection for NGT placement with X-ray in adult patients who were undergoing
NGT placement for any reason in any care setting. We searched published studies in the following
electronic databases: Cochrane Central Register of Controlled Trials, MEDLINE, EMBASE, CINAHL, Web of
Science, WanFang Data, China Journal Net, and the Chinese Biomedical Literature Database. Both English
and non-English-language articles were retrieved. Risk of bias was assessed using a standard procedure
according to the Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) criteria.
Results: We included five studies involving a total of 420 adult patients undergoing nasogastric tube
placement: three trials were undertaken in mechanically ventilated patients in prehospital settings and
two involved participants with comma or severe trauma in emergency room or intensive care unit (ICU).
Pooled results showed that ultrasonography had a sensitivity of 0.93 (95% CI 0.87 to 0.97), and specificity
of 0.97 (95% CI 0. 23 to 1.00), suggesting that diagnostic performance of ultrasound is useful to confirm
correct NGT placement, but not optimal to detect incorrect NGT position. This was confirmed through a
summary receiver operator characteristics (SROC) curve that showed the area under the curve was 0.96
(95% CI 0.94 to 0.98).
Discussion: The main limitation of the review is the relatively moderate level of heterogeneity of included
studies which may partially undermine the reliability and reproducibility of results. The insufficient
studies included did not allow identification of possible sources of heterogeneity and exploration of
reporting bias. Due to heterogeneity of studies, the diagnostic performance of ultrasonography could only
be drawn cautiously. Physicians and nurses should perform routine X-ray if visualization of NGT is not
possible. More well designed studies exploring ultrasound as a diagnostic tool for accuracy of NGT
placement are needed to strengthen the current evidence.
© 2017 Elsevier Ltd. All rights reserved.
What is already known about the topic?
There is a recognized risk that nasogastric tubes (NGTs) can be
misplaced into the trachea, resulting in severe pulmonary
complications.
Ultrasonography is proposed as an accurate diagnostic device to
determine NGT position, and provides a promising alternative to
X-rays to confirm NGT placement.
There is no consensus on the role of ultrasonography in detecting
NGT placement in routine clinical practice.
What this paper adds
Ultrasonography may be considered clinically useful to confirm
correct NGT placement.
There is insufficient evidence to suggest ultrasonography as a
diagnostic tool for incorrect NGT placement.
* Corresponding authors.
E-mail addresses: youth888cn@aliyun.com (T. You), chenkenck@126.com
(K. Chen).
1
First co-author.
http://dx.doi.org/10.1016/j.ijnurstu.2017.03.005
0020-7489/© 2017 Elsevier Ltd. All rights reserved.
International Journal of Nursing Studies 71 (2017) 80–88
Contents lists available at ScienceDirect
International Journal of Nursing Studies
journal homepage: www.elsevier.com/ijns

2. 1. Background
The insertion of a nasogastric tube (NGT) is the passage of a tube
via the nostril into the stomach. Insertion of an NGT is a
complicated procedure that requires skills and expertise by
practitioners (Fletcher, 2011) as a misplaced NGT within the
respiratory tract could lead to severe complications that include
pneumonia, pneumothorax (Lyske, 2011), empyema, pulmonary
hemorrhage and death (Miller, 2011). Between 2005 and 2010 in
the UK, 21 deaths and 79 cases of harm relating to feeding through
misplaced NGTs were reported (National Patient Safety Agency,
2011).
The key to safety with NGTs is verification of the correct
position in the stomach when the tube is initially passed and
before each administration of fluid (National Patient Safety Agency,
2011). A wide range of methods are used to verify NGT positioning
that include: examining the appearance of aspirate and pH testing
(Turgay and Khorshid, 2010); observing for signs of respiratory
distress (Fernandez et al., 2010); observing for bubbling when
placing the tip of the tube under water, air insufflated through the
NGT in combination with epigastric ausculation for whooshing
sounds (Simons and Abdallah, 2012); testing biochemical markers
(concentrations of bilirubin, pepsin or trypsin) (Fernandez et al.,
2010); and monitoring respiratory gases (e.g. capnography or
capnometry)(Chau et al., 2011). Although these tests are widely
known, they are not officially recommended as stand-alone
measures for NGT verification. Current guidelines recommend a
combination of aspirate testing and radiological confirmation
(National Patient Safety Agency, 2011). Aspirate pH readings
between 1 and 5.5 are considered indicative of gastric placement
and a reliable method for excluding placement in the pulmonary
system (National Patient Safety Agency, 2011); however studies
have found that pH confirmation could not be obtained for 44% of
tubes (Taylor et al., 2014), and the ability to obtain gastric aspirate
was only achievable 48.2% intubations (Boeykensa et al., 2014).
Concurrently, radiography for direct visualization is widely
accepted as the gold standard for determining tube location,
particularly following initial tube placement. (National Patient
Safety Agency, 2011). However repeated radiographic confirmation
is not practical and can interfere with medical and feeding
regimens for patients while increasing exposure to radiation and
medical cost (Chan et al., 2012). Methods with high accuracy and
clinical usefulness are needed to be alternatives to radiography in
NGT placement verification. One recent study has found that
electromagnetic (EM) traces and real-time 3-dimensional locali-
zation of NGT with anatomic landmarks accurately confirmed tube
position 100% of the time compared with X-ray (Taylor et al., 2014).
However, whether EM trace is proposed to be an alternative
method for NGT placement verification should be tested and re-
evaluated in future studies (Taylor et al., 2014), and cost-
effectiveness of this method has not yet been studied or
established.
Recent studies have also found that ultrasonography provides
good diagnostic accuracy to confirm NGT placement (Brun et al.,
2012, 2014; Chenaitia et al., 2012). These studies mainly focused on
NGT placement for decompression of the stomach in a prehospital
situation, for example, out-of-hospital resuscitation in settings
where X-rays are not available. Ultrasound could therefore provide
a promising alternative to X-ray in confirming NGT placement.
Despite studies addressing NGT placement with ultrasonography,
diagnostic accuracy has not yet been evaluated and there is no
consensus on the role of ultrasnography in routine clinical practice.
Thus, the aim of this systematic review is to synthesize the best
available evidence concerning the diagnostic accuracy of ultraso-
nography for determining NGT placement after insertion in adults.
2. Methods
We undertook a systematic review and meta analysis of
observational studies using the methods outlined in the Cochrane
Handbook for Systematic Reviews of Interventions (Higgins and
Green, 2011) and the Cochrane Handbook for Reviews of Diagnostic
Test Accuracy (Deeks et al., 2010).
2.1. Criteria for considering studies for this review
2.1.1. Types of studies
Studies on the diagnostic accuracy of NGT placement confir-
mation by ultrasound compared with X-ray visualization were
included. We considered controlled diagnostic test accuracy
studies (prospective cohort study, cross-sectional study, case-
control study) of ultrasound against X-ray visualization.
Uncontrolled reports (case series, case reports) were excluded,
in addition to studies on ultrasound-guided NGT placement and
verification, or gastrostomy and enteric tube placement. Studies in
which diagnostic accuracy of ultrasound (e.g. specificity, or
sensitivity) was not recorded or could not be calculated were also
excluded.
2.1.2. Participants
Adult patients receiving NGT placement in any care setting for
any reason. If not made explicit in the studies, we assumed
participants were adults unless identified as children.
2.1.3. Index tests
Ultrasonographic confirmation of NGT placement. We included
all studies regardless of where ultrasound test was performed (e.g.
bedside or X-ray department) or who performed the test (e.g.
ultrasonographer, physician or nurse), or whether the ultrasound
technique was enhanced by injection of saline and air into the NGT.
2.1.4. Target conditions
The target conditions included all which required proper NGT
placement.
2.1.5. Reference standards
The reference standard was X-ray of the chest or abdomen (X-
ray visualization).
2.2. Search methods for identification of studies
2.2.1. Electronic searches
With consultations and advice from academics familiar with
systematic reviews we developed search strategies and performed
literature searching. We developed different searching strategies
according to searching features of different databases, referring
and following the Cochrane Handbook for Systematic Reviews of
Interventions. These strategies were applicable to our review, so
we used it as a guider, making modifications to meet our review
purpose (see Supplement Data). The search terms used in both
English and Chinese were intubation, gastrointestinal, enteral
nutrition, nasogastric tube, feeding tube, Ryles tube, fine bore tube,
ultrasonography, ultrasound, sonography. Chinese search terms
were based on the terminology used in China. No restrictions were
placed on date of publications and language, and each database
was searched as far back as possible. We systematically searched
the following databases: Cochrane Central Register of Controlled
Trials (CENTRAL) (The Cochrane Library, 2015), MEDLINE (via
PubMed) (1966 to February 2015), EMBASE (OvidSP) (1974 to
February 2015), Web of Science (1965 to February 2015). We also
included CINAHL(via EBSCO) (1961 to February 2015), Cumulative
Index to Nursing and Allied Health Literature, because NGT
T. Lin et al. / International Journal of Nursing Studies 71 (2017) 80–88 81

3. placement is mainly performed by nursing staff in clinical settings.
And three other Chinese databases, e.g. WanFang Data, China
Journal Net, and the Chinese Biomedical Literature Database were
also searched from inception to February 2015 for the purpose of
searching as far range as possible.
2.2.2. Searching other resources
We screened reference lists within relevant trials to identify any
further potential papers worthy of review.
2.3. Data collection and analysis
2.3.1. Selection of studies
Results of the search were exported into Endnote software after
duplicates removed (n = 6816). Two authors (KC and TL) indepen-
dently screened the titles and abstracts of the articles identified in
the search strategy. Full text copies 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 relevance independently by KC and TL. Disagree-
ments regarding eligibility were resolved by consensus throughout
the selection process.
2.3.2. Data extraction and management
The two reviewers (KC and TL) independently extracted data on
study characteristics, patient demographics, sample size, test
methods, methodological quality, sensitivity, and specificity. Then,
both reviewers extracted data to construct a 2 2 contingency
table.
2.3.3. Assessment of methodological quality
We assessed risk of bias of included studies using QUADAS-2
tool as outlined by Whiting et al. (2011) and recommended by the
Cochrane Diagnostic Test Accuracy Group (Wisniewski, 2012). A
description of how the study addressed each item in the tool with a
judgment of ‘low,’ ‘high’ or ‘unclear’ and an overall rating was given
for each domain. We considered studies which were classified as
‘low risk of bias’ and ‘low concern’ in all the domains as studies
with high methodological quality. Assessment of methodological
quality was carried out by two reviewers independently; a final
decision was made by agreement.
2.3.4. Statistical analysis and data synthesis
Data synthesis was performed using the methods recom-
mended by the working group of the Cochrane Collaboration on
systematic reviews of diagnostic test accuracy (Deeks et al., 2010).
Forest plot was used to display the number of true positives (TP),
true negatives (TN), false positives (FP) and false negatives (FN), as
well as sensitivity and specificity, with their 95% confidence
intervals (CI), for all included studies. We also used summary
receiver operating characteristic (SROC) plots to display the results
of individual studies in a ROC space with each study being plotted
as a single sensitivity-specificity point.
For the meta-analysis, we used a bivariate random-effects
model to determine summary estimates of sensitivity and
specificity with 95% confidence and prediction regions (Reitsma
et al., 2005). Clinical utility of ultrasonography using likelihood
ratios were evaluated to enable the calculation of post-test
probability (based on the Bayes’ theorem) by means of the Fagan’s
nomogram (Deeks and Altman, 2004). Statistical analyses using
both Review Manager 5 software and the Metandi and Midas
Records identified through database
searching
(n = 10102)
No additional records identified
through other sources (i.e. reference
searching)
Records after duplicated removed
(n = 6816)
Records screened
(n = 6816)
Records excluded
(n = 6733)
(Two studies could not be retrieved)
Full-text articles assessed
for eligibility
(n = 83)
Full-text articles excluded (n = 78)
reasons:
Review (n =2)
Case report (n = 1)
Editorial (n = 1)
Not a diagnostic accuracy study (n =26)
Animal research (n =7)
Studies concerning ultrasound-guided
tube placement (n =24)
Studies in neonates or children (n =16)
Insufficient data (n = 1) (Specificity,
could not be calculated for true negative
number and false positive number was
zero at the same time)
Studies included for
quantitative synthesis
(n = 5)
Fig. 1. Study flow diagram.
82 T. Lin et al. / International Journal of Nursing Studies 71 (2017) 80–88

4. programs for the STATA software version 14.0 (Stata Corporation,
College Station, TX) was conducted.
2.4. Investigations of heterogeneity
Heterogeneity was investigated through visual examination of
forest plots of sensitivity and specificity and the result of I2
statistics. Although we planned to explore potential sources of
heterogeneity through subgroup analysis and meta-regression
(including: sample size, country, setting, whether tests were
carried out prior to index test with saline or air injection, different
areas for visualization, e.g. stomach area or stomach, neck and
esophagus area), insufficient studies did not allow us to conduct
the analysis.
2.5. Sensitivity analyses
We used the ‘leave-one-out’ procedure to perform sensitivity
analysis to examine the extent to which results are altered as a
result of changes in data; namely, a single study in the meta-
analysis was deleted each time to reflect the influence of the
individual data set to the pooled result.
2.6. Assessment of reporting bias
We intended to use a funnel plot to explore reporting bias, but
this was not conducted due to a lack of sufficient studies.
3. Results
3.1. Description of studies
We identified 10,102 references through electronic searches
(English n = 10,037, Chinese n = 64, French n = 1) from the following
database: Cochrane Central Register of Controlled Trials (CENTRAL)
(no reference), MEDLINE (via PubMed) (9277 references), EMBASE
(OvidSP) (613 references), Web of Science (133 references), CINAHL
(via EBSCO) (15 references), WanFang Data (6 references), China
Journal Net (2 references), and the Chinese Biomedical Literature
Database (56 references). No articles were identified through
scanning reference lists of the identified studies. After exclusion of
duplicates, 6816 references remained. Fig. 1 shows the flow of
references through the selection process.
Eighty five full text articles were retrieved for assessment. The
main reasons full text articles were excluded were study design or
non-research, a non-diagnostic accuracy study, animal studies,
studies concerning ultrasound-guided tube placement, studies in
neonates or children or insufficient data (Fig. 1). Two potential
studies that showed high relevance from the abstract could not be
retrieved and were therefore excluded. Another potential study
(Vigneau et al., 2005) was excluded because the specificity could
not be calculated for the true negative number and false positive
number were zero at the same time. This resulted in five studies in
three languages being included in this review (English n = 3,
Chinese n = 1, French n = 1).
All included studies were published after 2012 and were
prospective observational studies (Brun et al., 2012, 2014;
Chenaitia et al., 2012; Kim et al., 2012; Wang and Zhang, 2013).
These studies were all diagnostic accuracy studies of NGT
placement confirmed by ultrasound compared with X-ray as the
reference standard with diagnostic accuracy of ultrasound (e.g.
specificity, or sensitivity) recorded. Overall, 420 patients were
enrolled in three different countries (3 in France; 1 in China; 1 in
Korea) with a mean of 84 patients enrolled per study (range: 32 to
130). Of the five studies included, Brun et al. (2012), Brun et al.
(2014), Chenaitia et al. (2012) were conducted in prehospital
settings in France with a sample of 96, 32, and 130 mechanically
ventilated patients respectively; Kim et al. (2012) was conducted in
an emergency center in Korea and had a sample of 47 emergency
room patients with low consciousness; Wang and Zhang (2013)
included 115 trauma patients in an intensive care unit in China.
Prior to ultrasonography (index test), two bedside methods were
used to determine NGT placement: auscultation or insufflation of
air through a syringe coupled with epigastric auscultation and
gastric aspirate (Brun et al., 2014; Chenaitia et al., 2012). The
ultrasound was conducted by emergency room physicians, ICU
physicians, and emergency medicine specialists. In one study
(Wang and Zhang, 2013), confirmation relied on visualization of
the NGT in the stomach, while the four other studies (Brun et al.,
2012, 2014; Chenaitia et al., 2012; Kim et al., 2012) confirmed
placement by injecting small amount of saline or air into the NGT
under ultrasound. Saline or air injection under ultrasound showed
dynamic fogging in the stomach which verified NGT position when
visualization was not possible. The areas for visualization of the
NGT were different in included studies: two studies (Brun et al.,
2012; Chenaitia et al., 2012) placed an ultrasound probe on the
surface of stomach area and three studies (Brun et al., 2014; Kim
et al., 2012; Wang and Zhang, 2013) placed probe on the surface of
stomach, neck and esophagus area. Summary of the included
studies are summarized in Table 1.
3.2. Methodological quality of included studies
Overall, the quality of the included studies was good, as
illustrated in the QUADAS-2 results (Figs. 2 and 3). No concerns
regarding applicability or patient selection bias, or interpretation
of reference test results were raised. However, in two studies (Brun
et al., 2014; Chenaitia et al., 2012) the ultrasound index test was
Table 1
Summary of included studies (n = 5).
Author
(year)
Country Sample Setting Tests prior to index
test
Index test
(with saline or air injection into
NGT)
Index test
(areas for visualization)
Brun et al. (2012) France 96 mechanically
ventilated patients
prehospital NO YES stomach
Brun et al. (2014) France 32 mechanically
ventilated patients
prehospital YES YES stomach, neck and
esophagus
Chenaitia et al.
(2012)
France 130 mechanically
ventilated patients
prehospital YES YES stomach
Kim et al. (2012) Korea 47 emergency room patients
with low
consciousness
emergency
center
NO YES stomach, neck and
esophagus
Wang and Zhang
(2013)
China 115 critical traumatic patients ICU NO NO stomach, neck and
esophagus
T. Lin et al. / International Journal of Nursing Studies 71 (2017) 80–88 83

5. performed after the NGT position was determined by auscultation
and gastric aspirate tests: accordingly we deemed the risk of bias in
these cases as unclear since results of these two methods could
potentially have bias interpretation of the ultrasound results. Three
studies (Brun et al., 2012, 2014; Chenaitia et al., 2012) presented an
unclear risk of reference standard bias due to an unclear
explanation of whether or not reference standard results were
interpreted with knowledge of index tests results. In most studies
(Brun et al., 2012, 2014; Chenaitia et al., 2012; Kim et al., 2012) the
interval between the index and reference test was unreported, and
one study (Kim et al., 2012) did not include all selected patients in
the analysis resulting in all studies being unclear or high risk of bias
in the flow and timing domain of methodological quality.
3.3. Estimates of the diagnostic accuracy of ultrasonography for
detecting NGT placement
The sensitivity and specificity of each single study are shown in
Fig. 4. The SROC curve along with the summary point 95%
confidence interval(CI) and prediction regions are illustrated in
Fig. 5.Results included: pooled sensitivity 0.93(95% CI 0.87 to 0.97),
specificity 0.97(95% CI 0.23 to 1.00), positive likelihood ratio (PLR)
28.60(95% CI 0.32 to 2527.92), and negative likelihood ratio (NLR)
0.07(95% CI 0.03 to 0.15). Both summary sensitivity and specificity
value were high, but between-study heterogeneity was moderate
as visually assessable through the forest plot (Fig. 4) and predictive
ellipse (Fig. 5).
The Fagan plot (Fig. 6) shows that ultrasonography may be
clinically informative
because it increases the previous probability of being classified
as N+ from 92% (the average prevalence of N+ cases) to 100% when
positive, and the same probability is lowered to 44% when
negative. The likelihood ratio (LR) scattergram (Fig. 7) shows that
the summary point of positive and negative LR is located in the left
upper quadrant, suggesting that ultrasonography accuracy is
optimal not only for NGT position confirmation but also for
exclusion.
3.4. Investigations of heterogeneity
As seen in the forest plot (Fig. 4) which displayed both
sensitivity and specificity of all included studies, the between-
study heterogeneity was moderate. The result of I2
statistics also
indicated heterogeneity (I2
= 55.20, 95% CI = 0.00–100.00). Sub-
group analysis and meta-regression with covariates to formally
investigate potential sources of heterogeneity were not possible
because of the limited number of included study.
3.5. Sensitivity analyses
Fig. 8 showed the result of sensitivity analysis, suggesting the
relatively stable results of the meta-analysis.
3.6. Reporting bias
We did not explore reporting bias due to insufficient studies
included.
Fig. 2. Risk of bias and applicability concerns summary: review authors’ judgments
about each domain for each included study.
Fig. 3. Risk of bias and applicability concerns graph: review authors’ judgments about each domain presented as percentages across included studies.
Fig. 4. Forest plots of ultrasonography sensitivity and specificity for NGT placement and using X-ray as a reference standard. TP = true positive; FP = false positive; FN = false
negative; TN = true negative. Values between brackets are the 95% CIs of sensitivity and specificity. The figure shows the estimated sensitivity and specificity of the study (blue
square) and its 95% CI (black horizontal line).
84 T. Lin et al. / International Journal of Nursing Studies 71 (2017) 80–88

6. 4. Discussion
4.1. Summary of findings
Ultrasonography summary sensitivity was 0.93 (95% CI 0.87 to
0.97), suggesting that ultrasound yields a good diagnostic
performance in predicting correct NGT placement. This finding
is strengthened by the results of the Bayesian analysis (Fig. 6)
which demonstrated that ultrasonography also performed better
than a ‘smart observer,’ that is, one who knows the rate of correct
NGT placement and thus would assign the probability value of
100% to patients undergoing NGT placement. Even though the
results are encouraging, three out of five included studies (Brun
et al., 2012, 2014; Chenaitia et al., 2012) were performed in
prehospital settings, and most of the participants had a full
stomach that could have facilitated the ultrasound visualization of
the NGT. Still, four studies (Brun et al., 2012, 2014; Chenaitia et al.,
2012; Kim et al., 2012) used small amount of saline or air injected
into the NGT showing dynamic fogging in the stomach to make
ultrasound diagnosis more efficient to verify NGT position if tube
visualization was not possible. In this review, we did not explore
whether saline or air injection was an effective way to strengthen
diagnostic accuracy of ultrasonography for correct NGT placement.
Although it could aid in diagnostic accuracy, safety of injections
must be considered in future studies as saline injections into a
misplaced NGT could lead to severe complications (Wang and
Zhang, 2013). Several studies (Chenaitia et al., 2012; Kim et al.,
2012; Vigneau et al., 2005) reported that, the failure for
ultrasonography to confirm correct NGT placement was due to
gas interposition, which remains a major limitation of ultrasound.
Wang and Zhang (2013) suggested placing probes on the surface of
neck and esophagus area to visualize the route of the NGT
placement and rule out the possibility of interference from the gas
in the digestive tract, thus increases the diagnostic performance of
ultrasound to verify correct NGT placement. However, the
reliability of this method should be further studied.
Ultrasonography summary specificity was 0.97 (95% CI 0. 23 to
1.00) with extremely high value of specificity, while the 95%
confidence intervals of the summary sensitivity did, in fact, cross
the 0.50 value. This result makes it difficult to directly analyze and
interpret whether ultrasonography examination is useful to
correctly detect incorrect NGT position. We suggest this was
partly due to the relatively low number of cases of incorrect
insertions of the NGT in the included five studies and recommend
that performing ultrasound to confirm incorrect NGT placement
should be used cautiously. Nevertheless, the likelihood ratio (LR)
scattergram (Fig. 7) shows the summary point of positive and
0.0
0.5
1.0
Sensitivity
0.0
0.5
1.0
Specificity
Observed Data
Summary Operating Point
SENS = 0.93 [0.87 - 0.97]
SPEC = 0.97 [0.23 - 1.00]
SROC Curve
AUC = 0.96 [0.94 - 0.98]
95% Confidence Ellipse
95% Prediction Ellipse
SROC with Confidence and Predictive Ellipses
Fig. 5. Summary plots of 5 studies investigating the diagnostic ability of ultrasonography to detect NGT position. The solid circles correspond to the summary estimates of
sensitivity and specificity and are shown with 95% confidence regions (dotted lines) and 95% prediction regions (dashed lines).
0.001
0.002
0.005
0.01
0.02
0.05
0.1
0.2
0.5
1
2
5
10
20
50
100
200
500
1000
Likelihood Ratio
0.1
0.2
0.3
0.5
0.7
1
2
3
5
7
10
20
30
40
50
60
70
80
90
93
95
97
98
99
99.3
99.5
99.7
99.8
99.9
Post-test
Probability
(%)
0.1
0.2
0.3
0.5
0.7
1
2
3
5
7
10
20
30
40
50
60
70
80
90
93
95
97
98
99
99.3
99.5
99.7
99.8
99.9
Pre-test
Probability
(%)
Prior Prob (%) = 92
LR_Positive = 29
Post_Prob_Pos (%) = 100
LR_Negative = 0.07
Post_Prob_Neg (%) = 44
Fagan's Nomogram
Fig. 6. Fagan plot estimating how much the result of ultrasonography changes the
probability of a NGT position, considering a given pre-test probability.
T. Lin et al. / International Journal of Nursing Studies 71 (2017) 80–88 85

7. negative LR (located in the left upper quadrant) suggesting that
ultrasound is optimal for confirming correct NGT positioning and
also for excluding an incorrect NGT position. The contradiction
between the two results cannot be fully explained, and more high
quality studies are needed to add to future meta-analysis with
regard to ultrasonography diagnostic accuracy.
We planned to explore the factors that influenced diagnostic
accuracy of ultrasonography, including whether tests were carried
out prior to index tests, different areas for visualization (stomach
alone or stomach, neck and esophagus areas), the size of NGT, and
ultrasound performers. We failed however to explore these factors
because the information needed was unavailable or limited in
studies.
In two studies (Brun et al., 2014; Chenaitia et al., 2012),
ultrasonography was performed after NGT position was deter-
mined by auscultation method and gastric aspirate test. Although
we could not identify the effects of the two prior methods from the
results, we suggest that the tests carried out prior to index test
would influence the judgment of ultrasound performer. Therefore,
we suggest future studies avoid this method unless the person
performing the ultrasound is blinded to results of prior index tests.
1
10
100
Positive
Likelihood
Ratio
0.1 1
Negative Likelihood Ratio
LUQ: Exclusion Confirmation
LRP10, LRN0.1
RUQ: Confirmation Only
LRP10, LRN0.1
LLQ: Exclusion Only
LRP10, LRN0.1
RLQ: No Exclusion or Confirmation
LRP10, LRN0.1
Summary LRP LRN for Index Test
With % Confidence Intervals
Likelihood Ratio Scattergram
Fig. 7. Ultrasonography ability for NGT position confirmation. Likelihood ratio (LR) scattergram defining quadrants of informativeness based on desirable thresholds (positive
LR 10, negative LR 0.1): left upper quadrant (test suitable both for diagnosis exclusion and confirmation), right upper (confirmation only), left lower (exclusion only), right
lower (neither confirmation nor exclusion).
1.05 1.77
1.17 2.69 4.08
Brun 2012 (2012)
Brun 2014 (2014)
Chenaitia 2012 (2012)
Kim 2012 (2012)
wang 2013 (2013)
Lower CI Limit Estimate Upper CI Limit
Meta-analysis estimates, given named study is omitted
Fig. 8. Sensitivity analyses in meta-analysis of ultrasonography for NGT placement confirmation.
86 T. Lin et al. / International Journal of Nursing Studies 71 (2017) 80–88

8. Two different ultrasound procedures were used to determine
NGT placement in the five included studies: 1) placing a probe on
the surface of stomach area to visualize the end point of the NGT,
and 2) placing probes on the surface of stomach, neck and
esophagus area to visualize the route of the NGT placement. There
were no differences found between these two procedures for
detecting NGT placement in this review. Differences in ultrasound
procedure for NGT placement is a topic in need of further
investigation.
4.2. Strengths and weaknesses of the review
The main strength of this review is that a comprehensive search
was conducted with no restriction placed on language of
publication, guaranteeing good representation of the studies
worldwide. Moreover, we provided conventional meta-analysis
results such as summary estimates of diagnostic performance in
addition to Bayesian analysis including Fagan plots and likelihood
ratio matrices which offers further information of clinical use. The
main limitation of this review is the relatively moderate level of
heterogeneity of included studies which may partially undermine
the reliability and reproducibility of results. The data available in
the literature did not allow identification of possible sources of
heterogeneity.
4.3. Applicability of findings to the review question
This review is applicable to nurses and other health care
providers who are caring for patients who require accurate NGT
placement. As anticipated, heterogeneity was a problem and we
could not confirm consistent sources of heterogeneity. Therefore,
factors that could influence the performance of ultrasound as a
diagnostic tool could not be suggested.
5. Conclusions
5.1. Implications for practice
Our findings partly support the use of ultrasound to confirm
NGT placement. Ultrasonography diagnostic performance may be
considered clinically useful to confirm correct NGT placement,
although there is insufficient evidence to suggest ultrasound as a
diagnostic tool for incorrect NGT placement. Physicians and nurses
should be aware that if visualization of NGT is not possible through
ultrasound, X-ray must then be performed to confirm correct or
incorrect placement. The heterogeneity of the evidence currently
available warrants some caution in interpreting the present
results. Overall, we observed heterogeneity and its sources need
to be clarified before any definitive conclusion can be drawn and
the use of ultrasonography can be proposed in a routine clinical
setting.
5.2. Implications for research
The valid but suboptimal trials of ultrasonography for
confirming NGT position prompts further investigation. Techno-
logical advancements such as the use of ultrasound while injecting
small amount of saline or air into the NGT may be the optimal
confirmation of NGT placement if proved to be safe. Other factors
could be investigated including the size of NGT and full or empty
stomach when considering the diagnostic performance of ultra-
sound. It will also be important to compare the diagnostic
performance of different tools (e.g., carbon dioxide, bilirubin
testing, pH testing) and to investigate the diagnostic potential of
combining different tools in order to optimize NGT placement.
Lastly, conclusions of this study needs regular review as the quality
of ultrasonography improves.
Conflict of interest
None declared.
Funding
This work was supported by the Department of Education of
Guangdong Province, China [grant numbers 2015JGXM-MS37
2015QTLXXM37]
Ethical approval
The systematic review was not subject to ethical review.
Appendix A. Supplementary data
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.ijnurstu.2017.
03.005.
References
Boeykensa, K., Steemanb, E., Duysburgha, I., 2014. Reliability of pH measurement
and the auscultatory method to confirm the position of a nasogastric tube. Int. J.
Nurs. Stud. 51 (11), 1427–1433.
Brun, P.M., Chenaitia, H., Bessereau, J., Leyral, J., Barberis, C., Pradel-Thierry, A.L.,
Stephan, J., Benner, P., Querellou, E., Topin, F., 2012. Ultrasound evaluation of the
nasogastric tube position in prehospital. Ann. Fr. Anesth. Reanim. 31 (5), 416–
420.
Brun, P.M., Chenaitia, H., Lablanche, C., Pradel, A.L., Deniel, C., Bessereau, J., Melaine,
R., 2014. 2-point ultrasonography to confirm correct position of the gastric tube
in prehospital setting. Mil. Med. 179 (9), 959–963.
Chan, E.Y., Ng, I.H., Tan, S.L., Jabin, K., Lee, L.N., Ang, C.C., 2012. Nasogastric feeding
practices: a survey using clinical scenarios. Int. J. Nurs. Stud. 49 (3), 310–319.
Chau, J.P., Lo, S.H., Thompson, D.R., Fernandez, R., Griffiths, R., 2011. Use of end-tidal
carbon dioxide to determine correct placement of nasogastric tube: a meta
analysis. Int. J. Nurs. Stud. 48 (4), 513–521.
Chenaitia, H., Brun, P.M., Querellou, E., Leyral, J., Bessereau, J., Aimé, C., Bouaziz, R.,
Georges, A., Louis, F., 2012. Ultrasound to confirm gastric tube placement in
prehospital management. Resuscitation 83 (4), 447–451.
Deeks, J.J., Altman, D.G., 2004. Diagnostic tests 4: likelihood ratios. Br. Med. J. 329
(7458), 168–169.
Deeks, J.J., Bossuyt, P.M., Gatsonis, C., 2010. Cochrane Handbook for Systematic
Reviews of Diagnostic Test Accuracy Version 1.0The Cochrane Collaboration. .
(Accessed 13 January 2015) http://srdta.cochrane.org/.
Fernandez, R.S., Chau, J.P., Thompson, D.R., Griffiths, R., Lo, H.S., 2010. Accuracy of
biochemical markers for predicting nasogastric tube placement in adults-A
systematic review of diagnostic studies. Int. J. Nurs. Stud. 47 (8), 1037–1046.
Fletcher, J., 2011. Nutrition: safe practice in adult enteral tube feeding. Br. J. Nurs. 20
(19), 1234–1239.
Higgins, J.P.T., Green, S., 2011. 2011. Cochrane Handbook for Systematic Reviews of
Interventions Version 5.1.0The Cochrane Collaboration. . (Accessed 13 January
2015) http://www.cochrane-handbook.org/.
Kim, H.M., So, B.H., Jeong, W.J., Choi, S.M., Park, K.N., 2012. The effectiveness of
ultrasonography in verifying the placement of a nasogastric tube in patients
with low consciousness at an emergency center. Scand. J. Trauma Resusc.
Emerg. Med. 20 (1), 38–46.
Lyske, 2011. A Rare Complication of Nasogastric Tube Insertion. BMJ Case Reports.
(Bcr. 08. 2011.4606).
Miller, S.L., 2011. Capnometry vs. pH testing in nasogastric tube placement.
Gastrointestinal Nurs. 29 (2), 30–33.
National Patient Safety Agency, 2011. Reducing the Harm Caused by Misplaced
Nasogastric Feeding Tubes in Adults, Children and Infants. Patient Safety Alert
002NPSA, London. . (Accessed 10 January 2015) http://www.nrls.npsa.nhs.uk/
resources/type/alerts/?entryid45=129640q=0%C2%ACReducing+the+Harm
+Caused+by+Misplaced+Nasogastric+Feeding+Tubes+in+Adults%2c+Children
+and+Infants%C2%AC/.
Reitsma, J.B., Glas, A.S., Rutjes, A.W., Scholten, R.J., Bossuyt, P.M., Zwinderman, A.H.,
2005. Bivariate analysis of sensitivity and specificity produces informative
summary measures in diagnostic reviews. J. Clin. Epidemiol. 58 (10), 982–990.
Simons, S.R., Abdallah, L.M., 2012. Bedside assessment of enteral tube placement:
aligning practice with evidence. Am. J. Nurs. 112 (2), 40–46.
Taylor, S., Allan, K., McWilliam, H., Manara, A., Brown, J., Toher, D., Rayner, W., 2014.
Confirming nasogastric tube position with electromagnetic tracking versus pH
or X-ray and tube radioopacity. Br. J. Nurs. 23 (7), 352–358.
T. Lin et al. / International Journal of Nursing Studies 71 (2017) 80–88 87

9. Turgay, A.S., Khorshid, L., 2010. Effectiveness of the auscultatory and pH methods in
predicting feeding tube placement. J. Clin. Nurs. 19 (11–12), 1553–1559.
Vigneau, C., Baudel, J.L., Guidet, B., Offenstadt, G., Maury, E., 2005. Sonography as an
alternative to radiography for nasogastric feeding tube location. Intensive Care
Med. 31 (11), 1570–1572.
Wang, Xiaolan, Zhang, Mao, 2013. Primary Study on Rapid Location of Feeding Tube
by Ultrasound in Major Trauma Patients. Dissertation. Zhejiang University
College of Medicine.
Whiting, P.F., Rutjes, A.W., Westwood, M.E., Mallett, S., Deeks, J.J., Reitsma, J.B.,
Leeflang, M.M., Sterne, J.A., Bossuyt, P.M., 2011. QUADAS-2: a revised tool for the
quality assessment of diagnostic accuracy studies. Ann. Intern. Med. 155 (8),
529–536.
Wisniewski, S., 2012. Quality Assessment in DTA Reviews. . February 2012 update.
Accessed at (accessed 11.02.2015) http://srdta.cochrane.org/news/quality-
assessment-dta-reviews-%E2%80%93-february-2012-update/.
88 T. Lin et al. / International Journal of Nursing Studies 71 (2017) 80–88