Keywords Quality and logistical aspects - Quality management - Performance and complications
- Endoscopy Upper GI Tract - Diagnosis and imaging (inc chromoendoscopy, NBI, iSCAN,
FICE, CLE)
Introduction
Nasogastric (NG) tubes are commonly used in clinical practice, but blind insertion
can lead to complications with an average malposition rate of 1.9% [1 ]. Various methods have been developed to confirm NG tubes position, with x-ray being
the most accurate method, but it lacks real-time information and requires radiation
[2 ]. Notably, the National Patient Safety Agency of the United Kingdom reported 21 deaths
and 79 cases of injuries resulting from misplaced NG tubes between 2005 and 2010,
with 45% of these incidents attributed to misinterpretation of x-ray results [3 ].
The CORTRAK2 enteral access system, which uses a reusable electromagnetic sensing
device to guide NG tubes placement, had a reported assisted success rate of 82.6%
to 85% [4 ]
[5 ]
[6 ]. However, the US Food and Drug Administration (FDA) recalled the product in 2022
due to safety concerns, including injuries and deaths [7 ]. Another Integrated Real-Time Imaging System (IRIS) incorporates a miniature video
camera at its tip to enable real-time visualization during insertion, resulting in
reported assisted success rates of 86% to 97.8% [8 ]
[9 ]
[10 ]. However, the higher cost of this technology limits its use.
We developed a video-assisted system for NG tube placement guidance, consisting of
a cable with a mini camera that can be inserted into an NG tube. Our system combines
the advantages of both CORTRAK2 and IRIS systems by providing real-time visualization
during tube insertion and is potentially reusable. This study aimed to evaluate the
efficacy of this system in ensuring proper NG tube placement.
Patients and methods
Study design and patient selection
This prospective study was conducted in an Intensive Care Unit (ICU) and Gastroenterology
Ward in our hospital. Physicians selected and invited the patients, who were aged
20 years and older and required NG tubes for enteral feeding or decompression, to
participate in this study. Informed consent was obtained before enrollment. Patients
who could not undergo x-ray for tube position confirmation, those who were hemodynamically
unstable (mean arterial pressure ≤ 65 mm Hg), and those with skull base fractures
were excluded. This study was approved by the Institutional Review Board of our hospital
(B-BR-111–006) and the Taiwan FDA (no. 1110601678) and registered on ClinicalTrials.gov
(NCT05486286).
Design of video-assisted system for nasogastric tube placement
The video-assisted system comprised several components, as shown in [Fig. 1 ]. The three-way connector is attached to the rear end of the NG tube, allowing for
camera probe insertion. If needed, one port of the three-way connector can be connected
to an air inflator to improve visibility by introducing air into the NG tube ([Fig. 2 ]). The camera probe is then connected to the image processing box, which is connected
to the image display for real-time viewing. The camera probe has the following technical
specifications: resolution (400 × 400 pixels), field of view (120°), four light-emitting
diodes for illumination, maximum diameter (3.0 mm), and wire diameter (2.1 mm).
Fig. 1 The video-assisted nasogastric (NG) tube placement assisting system. The components
of this system include an image display facilitated by an Android phone, an image
processing box, a camera probe, a three-way connector with its cap, and the NG tube
itself.
Fig. 2 Setup of the 3-way connector. Illustration of the 3-way connector attached to the
rear end of the NG tube, allowing for camera probe insertion and optional air insufflation
to enhance visibility.
Feeding tube insertion and follow-up
This study used 15F feeding tubes with an outer diameter of 4.5 mm and an inner diameter
of 3.5 mm (Freka, Bad Homburg, Germany). Patients were advised to maintain a nothing-by-mouth
status for at least 4 hours. Patient position during NG tube insertion is sitting
upright or semi-upright. If misplacement into the trachea was detected during NG tube
placement, the feeding tube was withdrawn and reinserted into the stomach. A carbon
dioxide (CO2 ) insufflator (Olympus, Tokyo, Japan) was used whenever necessary to facilitate feeding
tube placement or confirm its position [11 ]. After placing the NG tube, the position of the NG tube was reconfirmed by chest
x-ray imaging for every patient. The operator completed the case report form and a
questionnaire. The following variables were recorded: time to complete tube placement,
time to visually confirm the stomach position, time to take x-ray, number of attempts,
vocal cord/trachea visualization, need for air insufflation, patient level of consciousness,
oxygen device (endotracheal tube or tracheostomy, with or without mechanical ventilation),
sedation status, and purpose of tube placement. A study nurse monitored patient condition
for the next 7 days and recorded any observed serious complications.
Outcome measurements and definitions
The primary outcome of this study was the technical success rate for placing the NG
tube in the stomach using this video-assisted system, as confirmed by x-ray examination.
This outcome serves as the main measure of the feasibility of the system. Secondary
outcomes focused on assessing usability of the system. We designed a questionnaire
to measure operator response using a Likert-type scale [12 ]: (1) This system helps facilitate NG tube placement; (2) This system helps identify
the safety landmark (not entering the vocal cord/trachea); (3) This system helps identify
the gastric mucosa; (4) This system provides good image visibility; and (5) Removing
the camera probe is easy. The operator assessed usability of this video-assisted system
to capture their subjective experience during the procedure.
Statistical analysis
Because this was a first-in-human pilot study, no sample size was calculated. Continuous
variables are presented as means with standard deviations for normally distributed
variables and as medians with interquartile ranges (IQRs) for skewed distributions.
For categorical variables and outcomes, we calculated counts and percentages. If patients
had missing data or loss of follow-up, they were excluded from the final analysis.
Results
This study enrolled 30 patients from August 2022 to May 2023, of whom 21 were admitted
to the ICU. Mean age was 74.3 ± 13.6 years. Median Glasgow Coma Scale (GCS) score
was E3/Vt/M5. Most patients had an endotracheal tube or tracheostomy (63.3%), indicating
the need for respiratory support. More than half of the patients (53.3%) were mechanically
ventilated ([Table 1 ]).
Table 1 Baseline characteristics of the 30 enrolled patients.
Baseline characteristics
E, eye opening; V, verbal response; M, motor response; T, tracheostomy; NG, nasogastric;
SD, standard deviation.
Age (mean ± SD)
74.3 ± 13.6
Gender (male: female)
18: 12
Conscious level, median (Q1, Q3)
E (median, IQR)
3 (3, 4)
V (median)
T (T, 2)
M (median)
5 (4, 6)
Oxygen device, N (%)
No need for oxygen device
11 (36.7%)
Endotracheal tube
11 (36.7%)
Tracheostomy
8 (26.7%)
Indication for NG tube placement, N (%)
For nutrition support
29 (96.7%)
For gastrointestinal tract decompression
1 (3.3%)
Outcomes
The primary outcome of successful placement into the stomach was achieved in 29 patients
(96.7%). In the first case, placement could not be achieved due to a tortuous lower
third esophagus, and the NG tube was placed with the assistance of esophagogastroduodenoscopy.
For all patients except this one, follow-up was completed and data were collected.
[Table 2 ] shows procedure results. Median time to complete tube placement was 3.8 minutes
(IQR, 1.9–5.3). Median time to take an x-ray for position confirmation was 71 minutes
(IQR, 44.5–151.5). Thirteen patients needed air insufflation during the procedure
to improve visibility of the gastric mucosa. Four patients required a second NG tube
placement attempt, and two underwent tracheal visualization. No serious complications
were observed during 7-day post-procedure follow-up, such as aspiration pneumonia,
pneumothorax, hollow organ perforation, or death.
Table 2 Nasogastric tube placement-related outcomes and technique information from the 30
enrolled patients.
N (%) or median (IQR)
*Serious complications included aspiration pneumonia, pneumothorax, hollow organ perforation,
and even mortality during the 7-day post-procedure follow-up.
IQR, interquartile range.
Successful placement into the stomach
29 (96.7%)
Serious complications*
0
Technique information
Time for completing tube placement, minutes
3.8 (1.9–5.3)
Time to visual confirmation of stomach position, minutes
1.5 (0.9–3.1)
Time until taking X-ray, minutes
71 (44.5–151.5)
Deep of tube placement, cm
60 (60–65)
Need for air insufflation for position confirmation
13 (43.3%)
Patients requiring a second nasogastric insertion attempt
4 (13.3%)
Patients with visualization of the trachea
2 (6.7%)
As shown in [Table 3 ], the operators consistently rated the video-assisted system usability favorably
(> 4 points) for all questions. However, the rating regarding image visibility was
slightly lower (mean score: 4.0 points), although it still provided valuable insights.
The most common problem was that oral secretions could interfere with the visual field,
which required air insufflation in 13 patients to ensure clear visualization of the
gastric mucosa.
Table 3 Operator Likert scale and mean scores.
Likert scale: from (1) strongly disagree to (5) strongly agree
Mean (ranges)
This system helps facilitate nasogastric tube placement
4.3 (4–5)
This system helps identify the safety landmark (not entering the vocal cord/trachea)
4.6 (3–5)
This system helps identify the gastric mucosa
4.8 (3–5)
This system provides good image visibility
4.0 (2–5)
Removing the camera probe is easy
4.7 (4–5)
Technical challenge
We trimmed the NG tube opaque tip in our initial case to improve visibility ([Fig. 3 ]
a,b ). However, this created an open end that allowed secretions from the nasal cavity
or gastrointestinal tract to obstruct the visual field. In the subsequent 29 cases,
we kept the tip of the NG tube untrimmed and positioned the camera probe between the
tip and the side hole. Notably, the camera was not directly attached to the opaque
tip, enabling the surroundings to be seen through the transparent tube wall ([Fig. 3 ]
c-f , [Video 1 ]). However, saliva or other secretions could still enter the tube through the side
hole. We occasionally had to perform repetitive back-and-forth movements of the camera
wire to improve the visual field, which was inconvenient. This technical challenge
warrants future attention and resolution.
Fig. 3 Various aspects of a trimmed and untrimmed nasogastric (NG) tube, along with images
captured by the camera probe. a The trimmed tip of the NG tube.
b Images obtained from the camera probe while employing the
trimmed tip of the NG tube. c The untrimmed tip of the NG tube,
with the camera probe inserted closely adjacent to the tip. d
Images captured by the camera probe from the perspective of position c . e The untrimmed tip of the NG tube, with the
camera probe inserted between the side hole and the tip. f
Images obtained from the camera probe at the position of e .
With the NG tube placement assistance system, the camera can visualize the surrounding
anatomy through the transparent tube wall, from entry into the nasal cavity through
the oropharynx, hypopharynx, esophagus, esophagogastric junction, and into the stomach.Video
1
Discussion
We developed an NG tube placement-assisting system with a success rate of 96.7% in
this pilot study. Operator feedback indicated the system was easy to use. Real-time
visualization during NG tube placement prevented tracheal misplacement, thereby diminishing
risk of complications such as aspiration pneumonia or pneumothorax. Furthermore, the
simple external structure of the camera probe suggests the possibility of reuse following
proper certification, making it potentially suitable for widespread application in
real-world scenarios. In addition, median time to completing NG tube placement using
this system was 3.8 minutes, which is comparable to or slightly shorter than previous
reports about the IRIS system [8 ]
[10 ]
[13 ].
The traditional method of placing these tubes without visual guidance may result in
misplacement and subsequent serious complications, such as aspiration pneumonia, pneumothorax,
hollow organ perforation, and even death [14 ]. Although x-ray confirmation remains the gold standard, misinterpretation can still
lead to adverse events [4 ]. Scientists and medical companies have been working on new inventions, such as CORTRAK2
or IRIS, to address this issue. CORTRAK2 used an alternative method by inserting an
electromagnetic sensing cable into the NG tube to display its relative path and guide
insertion [4 ]
[5 ]
[6 ]. One noteworthy advantage of this method is reusability of the sensing cable, which
is compatible with standard feeding tubes. However, the device faced significant adverse
outcomes related to misinterpretation of the position based on depiction of the tube
form, leading to a recall by the US FDA in 2022 [7 ]. In contrast, IRIS used real-time imaging to guide insertion and demonstrated promising
performance in various studies [8 ]
[9 ]
[10 ]
[15 ]. However, its integration of the video camera at the NG tube tip resulted in high
cost, limiting its widespread adoption, particularly under Taiwanese Public Health
Insurance.
To address these challenges, we developed our video-assisted NG tube placement system
that combines the advantages of both approaches: real-time imaging to guide insertion
and potential reusability of the camera probe. Integration of the video camera at
the NG tube tip in IRIS resulted in high costs and limited its widespread use. To
reduce overall costs, our system consists of an image processing box and a camera
probe, both of which are reusable. The camera probe may be more prone to breakage
after repeated use, but it can be replaced separately, helping to minimize expenses.
This design is intended to be cost-effective, particularly for high-risk patients.
Although exact pricing is difficult to determine before market release and may vary
by country, we estimate the system cost to be approximately USD 3,500, based on current
conditions in Taiwan.
Nonetheless, our system faced challenges. Because this study used a commercial NG
tube with a nontransparent tip, the camera probe must be placed a little distal to
the side hole away from the tip end so that surrounding structures can be visualized
through the transparent tube wall. Secretion bubbles from the NG side hole may affect
the visual field, necessitating air insufflation, camera probe cleaning, or gastrointestinal
content suctioning during NG tube insertion. This problem may be solved by further
designing a dedicated NG tube. In addition, using an endoscopic CO2 insufflator unit for air insufflation is inconvenient in clinical practice. Therefore,
there is a need to develop more convenient methods, such as insufflator bulbs similar
to those used in the IRIS system.
Limitations
This study has several limitations. First, due to its status as a first-in-human study,
it featured a relatively small sample size and lacked comparative groups. Given the
relatively low incidence of serious complications associated with NG insertion, this
pilot study was not designed to assess the potential of the device to reduce complication
rates. A larger-scale study is required to address this question, and to provide more
robust and generalizable results. Second, there is a potential for selection bias,
because both the patient/family and the physician may have had specific considerations
for case enrollment involving a totally new device. This is reflected in the final
cohort, in which the enrolled patients had relatively higher GSC scores. Third, this
study used a relatively large-bore NG tube (15F) due to its standard use in our hospital,
despite the fact that smaller-bore NG tubes are more comfortable for patients. To
address this, we have developed and tested a second-generation camera probe in vitro
that is compatible with 10F and 12F NG tubes. Finally, in real-world scenarios, NG
tube placement is often performed by general physicians. However, all investigators
in this study were gastroenterology specialists. Therefore, general physicians may
have to learn to identify the esophageal and gastric mucosae from the trachea. Nevertheless,
based on past reports about the IRIS system, there is reason to believe that all physicians
could quickly learn to identify the correct anatomical structures.
Conclusions
In conclusion, the video-assisted NG tube placement system provides real-time visualization
of the gastrointestinal tract during tube insertion. This capability enhances the
success rate for confirming feeding tube placement, potentially minimizing severe
complications. With ongoing design improvements, this method has the potential to
evolve into a more practical approach.
Bibliographical Record
