Advice

Evidence review

Evidence review

There are 2 versions of CORTRAK EAS (the original, predecessor version, referred to as CORTRAK 1 EAS for differentiation, and CORTRAK 2 EAS) included in the evidence. Because both versions work in the same way, studies which used either version were included in this briefing (referred to collectively as CORTRAK EAS unless explicitly stated otherwise).

Clinical and technical evidence

Regulatory bodies

A search of the Medicines and Healthcare Products Regulatory Agency website revealed no manufacturer Field Safety Notices or Medical Device Alerts for this device. There were 68 adverse events identified from a search of the US Food and Drug Administration (FDA) database: Manufacturer and User Device Facility Experience (MAUDE) from 2007 to present. Fourteen events were device malfunctions and did not result in patient harm. Forty-one events involved incorrect tube placement which caused patient harm including pneumothorax and, in 5 cases, death.

Clinical evidence

A literature search revealed 48 journal articles that reported on CORTRAK EAS (1 or 2). Studies were included if they investigated the efficacy (successful tube placement, time to placement, the proportion of patients needing imaging to confirm placement/resource utilisation, accuracy for confirming tube placement) and safety of the device. Retrospective studies and studies with fewer than 30 patients were excluded. Consequently, 7 studies (3 randomised controlled trials and 4 prospective cohort studies) have been included in this briefing.

Evidence on guiding post-pyloric placement in place of existing insertion methods

Holzinger et al. (2011) was a single-centre randomised controlled trial set in Austria (NCT00500851). The aim of the study was to compare the success rate of NJ feeding tube placement using CORTRAK 1 EAS with that of the endoscopic technique. Patients (n=66) were adults (55±18 years old) in intensive care (ICU) who did not tolerate NG feeding. The primary outcome was successful placement of the NJ feeding tube. Secondary outcomes included placement times, rates of nose bleeds, ICU survival and hospital survival. Neither the difference in placement success rates nor the difference in placement time between CORTRAK 1 EAS and the endoscopic technique were significant. However, CORTRAK 1 EAS resulted in correct placement at the first attempt more often than the endoscopic method (relative risk 2.44, 95% confidence interval [CI] 1.24 to 4.78, p=0.009). Comparisons between the rates of nose bleeds, ICU survival and hospital survival showed no statistically significant differences between the 2 groups. The authors concluded that using CORTRAK 1 EAS to place NJ feeding tubes in an ICU adult population is as fast, safe, and successful as using the endoscopic technique.

Gray et al. (2007) was a single-centre, prospective, observational study set in the USA. Its main aim was to evaluate the safety (absence of serious adverse events such as lung intubation) of CORTRAK 1 EAS in post-pyloric feeding tube placement compared to blind placement. Patients (n=101) were adults in ICU who needed post-pyloric feeding tubes. Secondary outcomes included a comparison of the rate of accurate placement (as confirmed by X-ray), number of X-rays needed to confirm placement, time from clinician request to start of enteral feeding (reported as 2 separate outcomes: time from clinician request to tube placement and time from placement to initiation of enteral feeding). For the primary outcome, no complications or adverse events were reported in the intervention or control groups. In terms of secondary outcomes, the difference in the success rate of feeding tube placement between the CORTRAK 1 EAS and blind placement groups was not statistically significant but the CORTRAK 1 EAS group needed 50% fewer abdominal X-rays to confirm feeding tube placement (p=0.0001). The time between clinician request and start of feeding was 66% lower in the CORTRAK 1 EAS group than in the blind placement group (p=0.0032). The time between clinician request and actual post-pyloric placement was 48% lower in the CORTRAK 1 EAS group than in the control group (p=0.0059); however, the median time between tube placement and initiation of feeding was 4.5 hours in the control group and 4 hours in the study group (not statistically significant). The authors concluded that using CORTRAK 1 EAS for post-pyloric feeding tube placement avoided serious adverse events and resulted in fewer X-rays and more timely initiation of enteral feedings compared with the blind placement technique.

October and Hardart (2009) was a single-centre prospective cohort trial with serial control groups set in the US. The aim of the study was to evaluate the effectiveness of post-pyloric feeding tube placement using CORTRAK 1 EAS compared with blind placement. Patients (n=107; 1 month to 25 years old) were recruited from a paediatric ICU. The primary outcome of the study was the success rate of post-pyloric tube placement as determined by abdominal X-ray. The secondary outcomes included the median time to successful placement (as confirmed by the radiologist review of the abdominal X-ray) and the total number of abdominal X-rays conducted. The study found that the difference in the success rate between the CORTRAK 1 EAS group and the blind placement group was significant (82.0% versus 39.0%, p<0.0001) as was the median time to successful placement (including confirmation; 1.7 versus 21.0 hours, p<0.0001). The unusually long 'time to successful placement' shown for blind placement is due to the definition of this outcome in the paper and the fact that blind placement procedure might require repeated placement attempts and X-rays to confirm successful placement. Significantly fewer abdominal X-rays (p<0.0001) were needed in the CORTRAK group than in the blind placement group. The authors concluded that CORTRAK 1 EAS is a safe and effective technique for post-pyloric feeding tube placement.

Kline et al. (2011) was a single-centre randomised clinical trial set in the USA. The aim of the study was to compare the time to successful post-pyloric feeding tube placement using CORTRAK 1 EAS with that of blind placement. Patients (n=49) were children (neonates to 17 years old) in a paediatric ICU needing feeding tube placement. The primary outcome of the study was the time to successful placement. The secondary outcome was success rate as confirmed by X-ray. The time to successful placement was significantly longer in the CORTRAK 1 EAS group than in the blind placement group (hazard ratio 2.1; 95% CI 1.4 to 4.1, p<0.03). The difference between the success rates with CORTRAK 1 EAS and the blind placement technique was not statistically significant (p=0.49). The authors concluded that CORTRAK 1 EAS does not shorten the time needed by experienced practitioners to place post-pyloric feeding tubes in children.

Viana et al. (2011) was a single-centre randomised controlled trial set in Brazil. The aim of the study was to evaluate the success rate of post-pyloric feeding tube placement using CORTRAK 1 EAS with that of blind placement. Patients (n=37) were adults (67.3±14.2 years) in a general ICU of a tertiary hospital. The secondary outcome was time to successful placement. The difference in success rates between CORTRAK 1 EAS and blind placement was significant (p<0.001), and CORTRAK 1 EAS took significantly less time than blind placement (p<0.001). The authors concluded that CORTRAK 1 EAS was a faster and more effective method of placing post-pyloric feeding tubes than blind placement.

Evidence on confirming post-pyloric placement in place of existing methods

Powers et al. (2011) was a multicentre, prospective cohort study set in the USA. The aim of this study was to compare the accuracy of CORTRAK 1 EAS with abdominal X-rays for confirmation of placement of post-pyloric feeding tubes at the bedside. Patients (n=194) included children and adults (12 days to 102 years old) needing feeding tube placement and were recruited from tertiary referral centres including an ICU, a general adult ward and a general paediatric ward. The primary outcome was agreement between the CORTRAK 1 EAS signal reading and X-ray image on potentially 3 separate occasions: directly following CORTRAK 1 EAS-guided placement, after contrast was injected through the feeding tube and during a final radiographic reading by an independent radiographer. Secondary outcomes included successful placement in the small intestine, percentage of placements where real-time tracing using CORTRAK 1 EAS showed airway placement and the process was halted, median time for feeding tube placement using CORTRAK 1 EAS, and safety. The percentage of agreement between CORTRAK 1 EAS interpretation and the X-rays was 86.9% for the first X-ray, 97.4% for the second X-ray and 99.5% for the third X-ray. During the study, 191 feeding tubes (98.4%) were successfully placed in the small intestine at the bedside. In 7.5% of cases, the tube's advancement was stopped after CORTRAK 1 EAS tracing on the screen demonstrated placement in the airway. Median time for feeding tube placement using CORTRAK 1 EAS was 12 minutes (range 1 to 52 minutes). No complications associated with the use of CORTRAK 1 EAS were identified. The authors concluded that CORTRAK 1 EAS can accurately confirm placement of post-pyloric tubes when compared with abdominal X-rays and that CORTRAK 1 EAS appears to be safe to use.

Evidence on guiding nasogastric placement in place of existing methods

Taylor et al. (2014) carried out a single-centre, prospective study set in the UK. Although the stated aim of this study was to determine the success rate of CORTRAK EAS in confirming NG feeding tube placement compared with pH testing or X-ray, the study provides evidence on guiding NG feeding tube placement. The study included 113 adults (median age 53 years) in ICU who needed a new or replacement NG tube. A total of 127 tube placements using the CORTRAK EAS were included in the analysis. Overall, CORTRAK EAS guided placement and confirmation of placement (with aspiration of fluid with a pH≤5.0, or X-ray) took a median of 6.4 minutes (interquartile range 4 to 10.4). In 7% of patients, the CORTRAK EAS trace deviated significantly to the left or right, suggesting placement in the left or right main bronchus. All tubes were withdrawn without complication, demonstrating that CORTRAK EAS enabled users to view the path of the feeding tube in real time which enabled them to avoid incorrect tube placement before trauma occurred. There were no reported tube misplacements.

Evidence on confirming nasogastric placement in place of existing methods

No relevant evidence was identified.

Recent and ongoing studies

Two ongoing studies on CORTRAK EAS were identified in the Netherlands Trial Register (NTR) and 1 was notified by the manufacturer.

  • CORRECT trial (NTR4286): a parallel group randomised controlled trial which aims to compare the success rate of duodenal feeding tube placement using the CORTRAK EAS with the endoscopic technique. Patients had achalasia or dysphagia and needed a duodenal feeding tube. The trial had a planned starting date in December 2013 and its planned closing date was September 2014. The current trial status is unknown.

  • CORE trial (NTR4420): a parallel-group, multicentre, non-inferiority randomised controlled trial which aims to evaluate the effectiveness of nasoenteral feeding placement with CORTRAK EAS compared with endoscopic placement. Participants were surgical patients admitted to gastrointestinal wards in 5 hospitals requiring nasoenteral feeding. The trial had a planned starting date in October 2013 and its planned closing date was March 2015. Data collection has been completed.

Costs and resource consequences

The British Association for Parenteral and Enteral Nutrition estimates that as of 2011, between 28% and 34% of those admitted to hospital in the UK were at medium or high risk of malnutrition. NG feeding is the most common method used, and an estimated 271,000 NG tubes are supplied to the NHS annually (Macmillan Cancer Support 2013; Great Ormond Street Hospital 2014, National Patient Safety Agency Quarterly Data Summary 2008). This indicates the potential use of the CORTRAK EAS for placing NG tubes in the NHS.

According to the manufacturer, the device is being used at 34 NHS centres and 1 private centre in the UK. In addition, the device is being used at 6 centres in Ireland.

If CORTRAK 2 EAS were adopted, there would be no need to change the way current services are organised or delivered. No other additional facilities or technologies are needed alongside the technology.

The systematic review identified 2 conference abstracts and 4 published studies that provide some evidence concerning resource consequences. All 6 are non-randomised with sample sizes below 100 patients. These studies provide some evidence of cost savings for the NHS based on savings in staff time and potential reduction in need for X-ray confirmation, endoscopic tube placement and intra-hospital transports. All costs were adjusted for inflation and, where relevant, converted to pounds sterling.

Three studies based in the US compared resource use between nasoenteral feeding using CORTRAK EAS and conventional blind bedside placement (sometimes including prokinetic medication). Brown (2012) suggested a relative cost savings for CORTRAK EAS system of around £94 per placement; Gray (2007) suggested cost savings of £115 per placement and October (2009) suggested savings of £41 per placement.

The Taylor (2010) study was based at Frenchay Hospital in Bristol and compared feeding with CORTRAK 1 EAS with NG feeding plus prokinetics. They report cost savings equivalent to £143 per patient (originally reported in US dollars).

Windle (2010) used medical, dietetic and nursing records for 2 sites in the Mid Yorkshire NHS Trust to provide a trust perspective costing estimate for CORTRAK 1 EAS. They estimate a cost of £122 per tube insertion attempt, inclusive of a wide range of resource use.

An abstract by Sharma (2013) describes the experience of a single UK ICU. They report that 2 inadvertent lung placements confirmed by CORTRAK EAS led to mistrust of the device and an increase in X-ray-led confirmation. This unnecessary use of X-rays and radiographer time while using the CORTRAK EAS created an average overspend of £86 per patient.

Strengths and limitations of the evidence

Two studies included in this briefing are randomised controlled trials and 4 are prospective cohort studies. Taylor et al. (2014) was the only study conducted in the UK and so these results may be more relevant to the NHS. It was also the only study of the use of CORTRAK EAS in NG tube placement. Although the authors concluded that CORTRAK EAS may be considered a standalone method of confirming NG tube position, this may be misleading because the study did not actually observe or confirm any tube misplacements.

All but 1 of the studies (Power et al. 2011) recruited patients from ICUs. However, according to the manufacturer, CORTRAK EAS can be used in other settings including acute assessment units and outpatient settings. This focus on critically ill populations introduces a significant source of bias. Many of the ICU patients are mechanically ventilated, sedated and having antihypotensive agents. These factors increase the risk of tube misplacement and also increase the likelihood that enteral tube placement will be needed.

The 2 randomised controlled trials (Holzinger et al. 2011 and Kline et al. 2011) and 1 prospective cohort study (October and Hardart 2009) had sample size calculations for their primary outcome of success rate and time needed for accurate placement. The third randomised controlled trial by Viana et al. (2011) was stopped early due to a lack of device supplies, but according to the authors the study was shown to have sufficient statistical power to confirm the study hypothesis. Gray et al. (2007) performed retrospective sample size calculations and stated that a larger sample size would have strengthened the power of their study. A small sample size reduces the probability of detecting a difference between groups where such a difference exists (type II error), and will also increase the likelihood that a statistically significant finding is actually a false positive. The 2 remaining prospective studies (Powers et al. 2011; Taylor et al. 2014) did not report sample size calculations. Due to the relative low frequency of harm associated with any tube insertion method and the rarity of direct harm in existing placement checks, it is unclear whether studies had enough power to compare the safety of CORTRAK EAS with comparator methods.

Most studies used appropriate comparators although it should be noted that comparators vary among countries and clinical settings. In the UK the gold standard for the placement of post-pyloric feeding tubes is the endoscopic technique, which has success rates above 90% (Byrne and Fang 2006; Wiggins and DeLegge 2006); endoscopy was the comparator in the Holzinger et al. (2011) study. However, Gray et al. (2007), October and Hardart (2009), Kline et al. (2011) and Viana et al. (2011) used blind placement as a comparator noting that this was the conventional technique for post-pyloric tube placement (Viana et al. 2011; October and Hardart 2009), the hospitals' standard clinical practice (Kline et al. 2011) or what is traditionally advocated (Gray et al. 2007). X-rays and pH measurement are standard methods for confirmation of the tube's position and were used in studies where confirmation was necessary (Gray et al. 2007; Powers et al. 2011; October and Hardart 2009; Taylor et al. 2014).

Four studies with evidence on CORTRAK-EAS used the same tube type for both the intervention and comparator groups. Only 1 study (Holzinger et al. 2011) used a different type of tube for the endoscopic placement control group (a 150 cm double lumen jejunal tube [Freka Trelumina, manufactured by Fresenius Kabi]) which may have biased the results. The operators could not be blinded to the intervention and control groups in any of the randomised controlled trials. Although this may introduce performance bias, this limitation is common in studies involving medical devices.

In medical device procedures, another source of potential bias is the training provided to use the device and the resulting proficiency with the procedure. Holzinger et al. (2011) stated that endoscopy was done by experienced gastroenterologists, whereas the CORTRAK EAS-guided procedure was done by a single ICU staff member with limited experience (only 3 CORTRAK EAS-guided tube placements) before study initiation. Kline et al (2011) reported that all 3 practitioners placing the post-pyloric feeding tubes had experience of blind placement, and although they had received training in using the CORTRAK EAS they each had the opportunity to place only 1 electromagnetically guided post-pyloric tube before the start of the study. Gray et al. (2007) reported that blind tube placements were done by experienced clinical personnel who had also been trained in using CORTRAK EAS by both an outside expert and a manufacturer representative. October and Hardart (2009) was the only study to report an extensive training phase during which the group of CORTRAK EAS operators was trained in the use of the device over a 2-week period followed by a 6-week practice period. Powers et al. stated that tubes were placed by an investigator experienced in the use of CORTRAK EAS and that X-rays were read and verified by 2 radiologists. Taylor et al. (2014) did not report whether operators had undergone prior training or their level of experience.

In the October and Hardart (2009) study there was a large difference in the time to successful placement between CORTRAK and the blind placement technique (1.7 hours compared with 21 hours respectively). Time to successful placement was defined as the time between the operator removing the tube from stock to the confirmatory abdominal X-ray being done. In the case of CORTRAK placements, the shorter time is partly attributable to the real-time imaging that the technology can afford and the fact that the operator would be able to make appropriate adjustments before obtaining the confirmatory abdominal X-ray. The blind placement technique, on the other hand, does not allow for real-time assessment of tube placement and thus might require repeated attempts and multiple X-rays before the final confirmatory X-ray can be taken. This could account for the large difference in 'time to successful placement' outcome.

Lastly, the manufacturer funded the time and equipment for the Taylor et al. (2014) study and the lead author had served on a CORPAK MedSystems consultation committee in 2007. This had the potential for introducing bias in the reporting of outcomes.