The incidence of intraoperative gastric tube malposition verified by Point-of-Care Ultrasound

Abstract

BACKGROUND

Over a million gastric tubes are placed yearly for varying medical reasons including gastric decompression. In the operating room (OR), this is performed blindly, and position is confirmed by auscultation, aspiration, or palpation by a surgeon. Despite the known risks of malpositioned gastric tubes, there is limited data in anesthesia literature about the incidence of intraoperative malpositioned gastric tubes. In this study, we use Point-of-Care ultrasonography (POCUS) to confirm gastric tube placement in the OR.

METHODS

Prospective observational study with a total of 149 subjects, all over 18 years of age, undergoing surgery with general endotracheal anesthesia and intraoperative blind placement of a gastric tube by an anesthesia provider. The primary objective of this study is to determine the incidence of malposition of blindly placed gastric tubes.

RESULTS

In our analysis, we found that out of 149 patients 110 patients were successfully visualized; the incidence of malposition was 0.14 [95% CI: 0.08–0.21]. We did not find age, Body Mass Index, or sex to be associated with predisposing patients to intraoperative malposition of gastric tube. However, increasing years of experience of anesthesia provider correlated with higher malposition rates.

CONCLUSIONS

In summary, we demonstrated that the incidence of malposition of blindly gastric tubes was 14%. Given the attendant risks of malpositioned gastric tubes, this data should inform decision algorithms for the blind placement of gastric tubes.

Over a million gastric tubes are placed yearly for varying medical reasons including gastric decompression to decrease aspiration risk.¹ In the operating room (OR), this is performed blindly, and position is confirmed by auscultation, aspiration, or palpation by a surgeon and by X-ray. The latter is the gold standard for confirming tube position.² However, the risk of X-ray exposure outweighs the risk of a misplaced tube in the OR since the tube is used for gastric decompression and not for feeding. This leaves only simple and often crude methods of verification available to the anesthesia provider. Many of these methods like the inspection of the tube aspirate or the auscultation method (instilling air by syringe into the tube and listening to gurgling by stethoscope over the stomach) have been proven to be highly inaccurate in determining tube position.^3–5 In addition, other methods like the pH test do not have enough specificity to differentiate between respiratory and intestinal placements.^{6, 7} Likewise, the Nose-Ear-Xiphoid (NEX) method, which has been historically favoured as a preferred method, has recently been found to have high rates of misplacement in adults and been recommended that such method not be used to determine the length of tube insertion.^8,9 With all these limitations there is interest in developing more reliable methods for confirming correct placement of gastric tubes in the OR.

Here we demonstrate the use of Point-Of-Care Ultrasonography (POCUS) as an alternative method for determining gastric tube placement. While use of ultrasonography to determine gastric tube placement in other settings such as intensive Care Unit and emergency room has been described,^10–12 this is the first study of its kind done on gastric tubes in the operating room. the operating room environment and nuanced skill set of anesthesia providers may contribute to a different malposition rate. POCUS refers to bedside diagnosis and treatment of medical conditions using portable ultrasound guidance methods. Use of POCUS has increased in recent years, with potential to revolutionize healthcare by improving patient safety through increase in efficacy of procedures and decrease in complication rates.^13–15

The primary focus of this study will be to estimate the incidence of malpositioned tubes intraoperatively by anesthesia providers using POCUS. Our secondary objective will consider factors associated with the occurrence of malposition. These variables will include experience of the provider and patient demographics.

Materials and methods

The protocol for this study was approved by the Memorial sloan Kettering cancer center (MSKCC) Institutional Review Board (IRB) (IRB #17–653) and the requirement for written informed consent was waived by the IRB. The study was registered in ClinicalTrials.gov prior to patient enrollment with identifier: NCT03430908, Principal Investigator: Luis Tollinche, registration date February 13, 2018 and carried at a single site (MsKcc) starting in February 2018 and is on-going. This is a prospective observational study and adheres to the EQUATOR/STROBE guidelines/statement.

Study population

All patients 18 years or older undergoing general endotracheal anesthesia with intraoperative plan for blind gastric tube insertion by an anesthesia provider were eligible. Exclusion criteria included patients who had a gastric tube inserted under direct visualization with the use of a Mac or Miller or any video laryngoscope; patients in whom visualization of gastric tube by ultrasound was not possible due to body habitus or other anomalous anatomy; patients with preanesthesia in-situ gastric tube; and patients whose gastric tube was inserted by a clinician not part of the anesthesia team. A total of 149 patients were recruited. all patients for whom the gastric tube was visualized under ultrasonography (either as correct placement or misplacement) were included in the incidence of malposition analysis.

Study design

Real-time ultrasonography was performed to confirm gastric tube position, but providers who were placing the gastric tube were blinded to the ultrasound. Performance of ultrasonography during gastric tube placement was conducted by a board-certified radiologist specializing in diagnostic ultrasonography. The stomach was examined in a collapsed state after an overnight fast. No water distention or peristaltic agents were used. The long axis of the stomach was first determined by the operator using the scanning probe and the hand was immobilized to remain perpendicular to the stomach’s long axis. Adjustments were made to the ultrasound to improve the image quality. The scanning probe was moved along the axis of the stomach and a series of images were directly recorded on tape. The gastroesophageal junction was identified and the operator notified the anesthesia provider that the insertion of the gastric tube could commence.

All ultrasound studies were saved for audit and quality assurance. Correct placement of the gastric tube was defined as the terminal end of the tube lying in the body of the stomach. The GE junction can be visualized by ultrasonography at the substernal region using the acoustic window of the left lobe of the liver in most patients. a gastric tube that is placed in the esophagus and not in the stomach does not function as intended and is considered a malposition (Figure 1). Other studies have attempted to image the stomach after placement of gastric tube was complete and not in “real-time” during placement. Previous studies that perform ultrasonography after positioning and not in real time are flawed because it is necessary to visualize the tip of the tube passing the GE junction to confirm correct placement. This is the rationale for visualizing the gastric tube in real time as it passes the GE junction.¹⁶

Figure 1.—

Longitudinal view representation of the substernal region. A) Transverse view of GE junction (arrow) without gastric tube; B) transverse view of gastric tube in place at the GR junction (descending arrow, red in the online version).

Malposition was defined as: 1) gastric tube not seen entering the esophagus (e.g. if the tube coils in the mouth and never enters the esophagus); 2) gastric tube seen entering the tracheo pulmonary tract; or 3) gastric tube seen entering the esophagus but not visualized entering the stomach through the GE junction. Each malposition was recorded as a single event and documented. Additional provider and patient demographics were recorded as well. All provider and patient demographics were extracted from the Electronic Medical Record (EMR). Each patient had a maximum of one malposition event even if they satisfied multiple malposition types described above. Data was analyzed considering malposition as a dichotomous (yes/no) variable.

Statistical analysis

Characteristics of patients are summarized by status (all patients, correct position, malposition and, non-visualizable by ultrasound). Among patients where the blindly placed gastric tube placement can be visualized using the ultrasound, the incidence of malposition is summarized as the proportion of malposition cases, along with exact 95% confidence interval. Demographic results are presented using descriptive statistics. sample size for our feasibility endpoint was based on convenience, and not based on a priori calculation.

In the absence of reliable reporting of the incidence of malposition in anesthesia literature, the primary objective of this prospective observational study is to estimate the incidence of malposition of gastric tubes in an operative setting. As this is a summary study without specific hypothesis testing, no formal power calculation was necessary or was conducted. Instead, sample size calculation was performed based on expected margin of error. Without reliable prior estimation of the sample proportion, the sample size calculation was based on upon the sample proportion of 50%. This hypothesized proportion is the most conservative choice and provides the most conservative sample size. Using the 50% incidence estimate, the sample size needed to achieve 5% margin of error at 95% confidence level is 384. if the true incidence is 20%, the margin of error given 384 patients is 4%; if the true incidence is 5%, the margin of error is 2.2%. Given the actual incidence observed throughout this study a sample size of 149 has yielded a narrow confidence interval that allows us to make our conclusion.

All summary statistics were calculated using Stata 15.0 (Stata Corp, College Station, TX, USA) and visualized with Tableau 10.5 (Tableau, Seattle, WA, USA).

Results

A total of 149 patients were enrolled with an age median and interquartile range (IQR) of 61 (52–70) and Body Mass Index (BMI) median and IQR of 28.0 (24.1–31.6) Of all patients enrolled, 54.4% were male. Surgical procedures included colorectal, urology, GMT, gynecological, neurosurgery, thoracic, and HPB. Most patients underwent colorectal and urology surgeries (Table I).

Table I.—

Demographics of all patients included in the analysis. Age and BMI are presented as median and interquartile range (Q1-Q3). Sex and surgical service are presented as a proportion of the total.

Demographic data	All patients N.=149	correct position N.=95	Malposition N.=15	Non-visualizable N.=39
Age	61 (52–70)	61 (52–71)	61 (49–73)	62 (52–70)
Sex
Male	81 (54%)	40 (42%)	8 (53%)	33 (85%)
Female	68 (46%)	55 (58%)	7 (47%)	6 (15%)
BMI	28.0 (24.1–31.6)	27.8 (23.9–31.3)	26.9 (22.1–30.5)	29.2 (25.5–32.9)
Surgical service
Colorectal	40 (26.9%)	26 (27.4%)	3 (20.0%)	11 (28.2%)
Colorectal/urology	1 (0.7%)	1 (1.1%)	0	0
Gastric mixed tumor	11 (7.4%)	6 (6.3%)	0	5 (12.8%)
GMT/colorectal	1 (0.7%)	0	1 (6.6%)	0
Gynecological	17 (11.4%)	13 (1.4%)	2 (13.3%)	2 (5.1%)
Gynecological/colorectal	1 (0.7%)	1 (1.1%)	0	0
Hepatobiliary	33 (22.2%)	18 (18.9%)	4 (26.7%)	11 (28.2%)
Neurosurgery	1 (0.7%)	1 (1.1%)	0	0
Thoracic	6 (4.0%)	4 (4.2%)	1 (6.6%)	1 (2.6%)
Urology	38 (25.5%)	25 (26.3%)	4 (26.7%)	9 (23.1%)

Successful visualization of the GT placement by ultrasound failed in 39 patients and were considered non-visualizable. All of these were excluded from incidence of malposition analysis but indicated a proportion of 0.26 [95% CI: 0.20–0.31]. The demographics show that most of non-visualizable patients were male (85%); with an age and BMI median and IQR of 62 (56–70) and 29.2 (25.5–32.9) respectively (Table I). Reasons for non-visualization were recorded for each instance and categorized (Table II). The most common reason was failure to visualize the gastroesophageal junction.

Table II.—

Reasons for failure to visualize among non-visualizable patients.

Breakdown of non-visualizable cases	Count
Unable to visualize GE junction on ultrasound	16
Poor image quality affecting visualization	3
Machine malfunction	1

Incidence of GT malposition

Of all patients enrolled successful visualization was accomplished in 95 patients. A proportion of 0.86 [95% CI: 0.78–0.92] had correct GT placement after blind insertion as monitored by the real-time ultrasonography. This translates to an incidence of malposition of 0.14 [95% CI: 0.08–0.21]. Study midpoint analysis of all cases shows there is no malposition variability across patient volume overtime: the first half of the cohort had an incidence of malposition of 0.13 [95% ci: 0.08–0.26] and the second half an incidence of malposition of 0.14 [95% CI: 0.06–0.24].

The results show that patients with correct position vs. malposition when stratified by sex are proportionally similar: males accounted for 42% of correct position cases and 53% of malpositions. In addition, there is no significant difference in the age or BMI of patients with correct position vs. malposition (age P value =0.66; BMI P value =0.64, two-tailed t-test P<0.05). Incidence of malposition was also evaluated based on years of experience of the anesthesia provider (Figure 2). the results show that those with 10 to 20 years and 20 or more years of experience had higher rates of malposition (20% each) while those with three to 10 years had the lowest rates of malposition (4%) followed by those with one year or less (7%).

Figure 2.—

Incidence of malposition based on years of experience of anesthesia provider.

Discussion

Routinely, gastric tube (GT) insertion is performed blindly which introduces potential for serious harm to the patient. the most frequent malposition involves placement into the respiratory tract, possibly resulting in ‘aspiration by proxy’ a complication which has the potential to be fatal.^{17, 18} Other serious complications include perforation of the esophagus or stomach, obstruction of nasal breathing, as well as sinusitis and epistaxis among other.^{5, 17} It is the blind method that gives this insertion potential for serious harm. The ipsilateral piriform sinus and arytenoid cartilages offer the greatest paths of resistance in the hypopharynx. The tube’s preference for low resistance tracks and the shared alimentary and respiratory pathways contribute to the morbidity associated with placement.¹ A literature search of malpositioned gastric tube case studies resulted in 21 case reports out of 71 that focused on the GT placement in the intraoperative setting and performed by anesthesia providers.¹ The types of complications reported were as follows: perforation, oral mucosal injury, pulmonary misplacements, coiling and breathing tube constriction, lung collapse, and entanglements with other tubes. Six of these resulted in surgical interventions to repair the iatrogenic injury caused by the tube. Nine of the reported malpositions were in the pulmonary tree and of those, two went unrecognized intraoperatively and were discovered malpositioned after the patient left the OR. Two cases studies reviewed in the literature used the auscultation method as a verification method and determined the tube to be placed correctly although both were misplaced into the pulmonary tree.

In medical literature, the rates of malposition range from 0.2–46%.^{2, 19–21} This broad range is reflective of the differences of how malposition is defined. Many studies with the lower rates only looked at pulmonary malposition while the higher rates included other areas of malposition such as esophageal placement (where the terminal tip of the gastric tube is lying in the esophagus).²¹ Additionally, most of the data is disseminated in medical literature from a cohort of patients that may not translate to the intraoperative setting. There are differences in provider and patient characteristics that should be taken to account. These characteristics include provider experience, the operating room environment and specific patient characteristics.

Serious complications associated with GT malposition may increase recovery times and morbidity rates.^{5, 17} Specific to anesthesia practice, types of complications reported in the literature include perforation, oral mucosal injury, pulmonary misplacements, coiling and breathing tube constriction, lung collapse, and entanglements with other tubes.¹ With over a million GT insertions performed every year in the US a conservative estimate of 10% malpositions translates to over 100,000 patients at risk for serious complications related to GT placement.¹ Hence, it is imperative that we understand and document the incidence of malpositioned gastric tubes by anesthesia providers. If an incidence rate of malpositioned tubes exists then we need to raise provider awareness, determine predisposing patient risk factors, and ultimately develop strategies for verifying correct intraoperative placement.

Ultrasound has been studied as a verification method with promising results. However, there is limited data in anesthesia literature about the incidence of intraoperative malpositioned gastric tubes. Brun et al. identified a sensitivity and specificity rate of 99% and 100% when using the ultrasound in the emergency room setting.²² Kim et al. conferred a positive predictive value of 97% when using an ultrasound for determining tube position.²³ Vigneau et al. reported a sensitivity rate of 97% when using ultrasound as a verification method in a medical Intensive Care Unit.²⁴

The introduction of POCUS is revolutionizing healthcare. The use of ultrasound as a diagnostic tool started in 1941 and intraoperative echocardiography was first described in 1972. Nonetheless, the concept of POCUS is still considered a recent development.²⁵ Its profound impact stems from improving diagnostic and therapeutic capabilities of various procedures previously limited due to the lack of visualization. POCUS has become an integral part of anesthesia practice due to the trend of increasing use for previously established as well as novel diagnostic and therapeutic purposes in the perioperative care of patients.

Our results show that 14% of blinded GT insertions are malpositioned. We did not find age, sex, or BMI to be associated with malposition. We also found that anesthesia providers with more years of experience (10 to 20 years and 20 years or more) had higher rates of malposition (40% combined). Most GT insertions were done by providers with one to three years of experience and this malposition rate was 15%. These finding are counterintuitive as one might expect more experienced providers to have lower rates of malposition because of refined technique and improved skill over time. However, it is plausible that more senior anesthesiologists with extensive experience may demonstrate overconfidence during placement and fail to verify correct placement. It is also likely that experienced providers at teaching institutions may in fact perform fewer procedures over time as trainees are given these opportunities to develop their skills. As a result, senior clinicians may experience a loss of skill and technique due to infrequency of performing the procedure.

Limitations of the study

One limitation of our study is that given the methodology, the Hawthorne effect is likely present. Even though clinicians placing the gastric tube were blinded to the ultrasound results of the gastric tube position, the clinicians were aware that they were being observed during placement. It is reasonable to assume that an observed clinician will perform differently from their counterpart who is not being observed during gastric tube placement. it follows, too, that an individual will attempt to place the tube more diligently or introduce new placement techniques when they are being observed. Given this known observation effect, it is likely that clinicians were indeed more earnest and cautious in placing the gastric tube than occurs in routine care. Our malposition rate of 14% is possibly an underestimate of the actual malposition rate when clinicians are not being observed. This information further supports the need to reconsider the practice of blind placement of gastric tubes. Future studies should attempt to mitigate the effect of observation on placement techniques.

Conclusions

In this study, we determined the rate of gastric tube malposition during placement in the intraoperative setting is 14%. There is no consensus on malposition rates in the medical literature and given the morbidity and mortality associated with blindly placed gastric tubes, clarification of this controversy is paramount. Our data should yield improvement in patient care because our study highlights the rate of malposition of blindly placed gastric tubes. This data will inform gastric tube placement decisions and challenge clinicians to critically consider whether gastric tube insertion is indicated. Importantly, this study will also enable clinicians to recognize risk factors for malposition. Ultimately, these results may be used to guide and modify department algorithms for gastric tube placement and help create new protocols to enhance point-of-care placement verification techniques.

What is known

• Gastric tube malpositioning has the potential to be fatal in addition to many other potentially serious complications which may increase recovery time and morbiditiy.

• Current establisheds methods for determining gastric tube placement suffer from many limitations including being too invasive as well as lacking accuracy and specificity.

What is new

• The use of POCUS for real time determination of gastric tube placement demonstrates an incidence of 14% rate of malpositioning among anesthesia providers.

• Demographic characteristics of sex, age, and BMI do not appear to be correlated to increased risk of gastric tube malposition.