The need for enteral nutrition via nasogastric (NG) feeding tubes is very common in a wide variety of patients. Patients requiring NG feeding tubes often lack the protective airway reflexes that allow clinicians to recognize endotracheal placement of the NG tube by virtue of their illness. In the unconscious or obtunded patient, responses such as profound coughing, gag response, and difficulty with phonation are commonly absent. Endotracheal placement of NG feeding tubes can result in significant morbidity. The complication rate of feeding tube placement has been reported to be 4.4%,¹ and includes pneumothorax, atelectasis, pleural effusion, bronchopleural fistula, hydrothorax, empyema, mediastinitis, pneumonitis, esophageal perforation, and pneumonia.² The presence of a cuffed endotracheal tube does not protect against the tracheal placement of NG feeding tubes. The small diameter and relatively stiff tip of the NG feeding tubes allow them to pass the endotracheal tube cuff without resistance. In order to minimize these complications, a strategy for determining proper esophageal placement of NG feeding tubes is crucial.

A two-step X-ray technique has been the standard of care to safely confirm NG tube placement. First described in 1989,1 the technique involves inserting a radio-opaque NG tube to a length of approximately 30cm. A plain radiograph of the chest is then taken to ensure the tube is visualized in the midline distal to the carina, signifying that the tube has been placed in the esophagus. If this is not the case, or if the tube is seen on X-ray extending into either mainstem bronchus, the tube is repositioned and the X-ray is repeated. When esophageal placement is confirmed, it is advanced into the stomach and a second X-ray is taken prior to feeding tube use. While safe, this technique requires a great deal of clinician effort and time, averaging 170 minutes.³ It also incurs the financial costs and radiation exposure of two X-rays.

In an effort to streamline the safe insertion of NG feeding tubes, investigators began using end-tidal carbon dioxide (CO2) detection in place of the first X-ray described above.^3,4 Electronic capnographs that use infrared absorption spectrometry, developed to confirm endotracheal tube placement, have been used to confirm non-tracheal (esophageal) placement of feeding tubes. In this technique, the NG feeding tube is advanced 30cm as before, and gas sampling from the NG tube is used to rule out endotracheal placement by noting the absence of CO₂.

In their study, Kindopp et al.3 found that end-tidal capnography correctly identified all endotracheal and esophageal placements of NG feeding tubes when compared with the two-step X-ray technique. Electronic capnography simplifies safe NG tube insertion and saves the clinician valuable time, increasing the likelihood that the practitioner will make the effort of ensuring esophageal placement before advancing the NG tube completely. Electronic capnographs are expensive, however, and are not always available on hospital wards and critical care units, where NG tubes are inserted.

Colorimetric end-tidal CO2 devices use sulfonephthalein-impregnated pH-sensitive filter paper as an indicator that changes from purple to yellow in the presence of CO₂.⁵ They were developed to confirm endotracheal tube placement and they have no adaptors for feeding tubes and no sampling stream to ensure a flow of gases through the feeding tube. They are single-use and disposable and can be easily stocked anywhere in the hospital. Several investigators have developed apparatus that allow the colorimetric CO2 detector to sample gas from the NG feeding tube.

In 1998, Thomas et al. described 10 patients for whom colorimetric endtidal CO2 detection was used to document proper feeding tube placement.⁶ In 2002, a study of 53 patients used a disposable tracheostomy tube inner cannula with a 15mm snap-lock connector to attach the feeding tube to the colorimeter. This study reported 100% accuracy in determining tube placement.2 While impressively accurate, the sample sizes of these studies were modest.

In 2005,⁷ we developed a method for using colorimetric CO2 detectors to confirm feeding tube placement. Two types of adaptors were developed (see Figure 1) using materials readily available in most intensive care units (ICUs).

Wall suction was used to ensure gas sampling through the feeding tube.

The adaptor from a 3.5-French pediatric endotracheal tube connects directly to the intake of the colorimeter and forms a snug fit into the feeding tube. For institutions without pediatric endotracheal tubes, the body of a 3cc syringe has a Luer-Lok to connect to the feeding tube while the adaptor from a 7-French endotracheal tube connects to the intake port of the colorimeter and fits into the body of the syringe. Low wall suction can be briefly applied to the outflow of the colorimeter by using a 15mm nipple tip adaptor to ensure gas sampling.

In a study to evaluate this device, 93 procedures were analyzed. We first briefly inserted the NG feeding tube with the colorimetric CO2 detector into the endotracheal tubes of 10 ICU patients who were receiving mechanical ventilation and evaluated whether the device could detect CO2. The colorimetric CO2 detector correctly identified the presence of CO2 in all 10 patients. Following confirmation that the apparatus could accurately identify endotracheal placement of the NG feeding tube, we evaluated its accuracy in 93 patients who required the placement of a feeding tube in our ICU. We found it to be 98% accurate compared with the two-step X-ray technique, with a Kappa of 0.86 (excellent agreement) and a standard error of 0.10. In only one instance, the NG feeding tube was placed in the trachea and the colorimetric CO2 detector failed to detect CO2. In retrospect, we had been re-using the CO2 detector, and it had been overused to the point of failure. We advise that the colorimetric CO2 detectors be used only once. When not re-used, we found these devices to be extremely accurate in determining the placement of NG feeding tubes.

In 2006, Burns⁸ compared capnographic and colorimetric end-tidal CO2 detection in a variety of ICU patients using different types of NG and orogastric tubes. Colorimetric devices detected CO2 in all patients where capnography revealed CO2. They concluded that capnographic and colorimetric CO2 detection were equivalent for the confirmation of esophageal placement of the nasogastric or orogastric feeding tubes in medical ICU patients.

Both capnography and colorimetric CO2 detection seem to perform favorably compared with a gold standard of the two-step X-ray technique. Kindopp3 showed that capnographic placement confirmation saves an average of 97 minutes of clinician time per NG tube insertion compared with the two-step radiographic technique. More importantly, by being easier and less cumbersome, this safe alternative is more likely to be used at the bedside.

In light of its accuracy, readily available materials, and significant time savings, colorimetric CO2 detection should be considered for the routine practice of confirming NG feeding tube placement. Future research to evaluate improved compliance and cost savings will likely demonstrate significant advantages over current practices and lead to wider adoption of this simple, safe, time-saving technique.