pH and concentration of bilirubin in feeding tube aspirates as predictors of tube placement

Background: Currently available bedside methods for determining feeding tub e placement often provide inconclusive results. Therefore, additional data are needed to assist nurses in making decisions regarding tube location. Objectives: To describe the usual concentration of bilirubin in aspirates f rom newly inserted feeding tubes and to determine the extent to which these measures can contribute to pH alone in correctly predicting feeding tube l ocation. Methods: Gastrointestinal samples for concurrent pH and bilirubin testing w ere obtained from adult, acutely ill patients with newly inserted small-bor e feeding tubes (nasogastric, n = 209; nasointestinal, n = 228) within 5 mi nutes of radiographs taken to determine tube location. Respiratory samples were tested (tracheobronchial, n = 126; pleural, n = 24). pH was measured w ith a pH meter, and bilirubin content was assayed spectrophotometrically. R esults from the pH and bilirubin tests were compared with tube location as determined by radiography. Results: Mean pH levels in the lung (7.73) and intestine (7.35) were signif icantly higher than the mean pH level in the stomach (3.90; p < .001 for ea ch comparison). Mean bilirubin levels in the lung (.08 mg/dl) and stomach ( 1.28 mg/dl) were significantly lower than the mean bilirubin level in the i ntestine (12.73 mg/dl; p < .001 for each). By visually inspecting distribut ion overlap and mean differences by tube site, results were dichotomized so that a combination of pH and bilirubin values could be used to develop a p redictive algorithm. A pH of >5 and a bilirubin value of <5 mg/dl correctly identified all respiratory cases, whereas a pH >5 coupled with a bilirubin level of greater than or equal to 5 mg/dl correctly identified three fourt hs of the intestinal cases. A pH of less than or equal to 5 coupled with a bilirubin value of <5 correctly identified more than two thirds of the gast ric cases. Conclusions: Preliminary laboratory-based data indicate that appropriate us e of the proposed algorithm could significantly reduce the number of x-rays needed to exclude respiratory placement and to distinguish between gastric and intestinal placement.