Diagnostic aids in the differentiation of pyloric stenosis from severe gastroesophageal reflux during early infancy: the utility of serum bicarbonate and serum chloride.

This study evaluated whether serum bicarbonate levels, serum chloride levels, and other diagnostic criteria could be used to differentiate pyloric stenosis (PS) from severe gastroesophageal reflux (GER) during early infancy. The investigation was a retrospective, case-control study conducted in the emergency department of a large, academic children's hospital. Cases were 75 consecutive infants with PS confirmed in the operating room. Controls were 75 consecutive infants 12 weeks of age or younger with the diagnosis of GER whose serum electrolytes had been examined. Projectile vomiting was sensitive (0.93) but not specific (0.39) for PS. The mean serum bicarbonate level was 27.2 mmol/L for PS patients and 22.3 mmol/L for GER patients (P < .00001), and the mean serum chloride level was 95.7 mmol/L and 103.6 mmol/L for PS patients and GER patients, respectively (P < .00001). Serum bicarbonate levels of > or =29 mmol/L and serum chloride levels of < or =98 mmol/L had high positive predictive values (0.96 and 0.97, respectively) and were specific (0.99 for both) but not very sensitive (0.36 and 0.50, respectively) in identifying patients with PS. Only one patient would have been misclassified (false positive) as having PS using either of these cutoff values. These laboratory tests can also help discriminate between PS and GER when the history and physical examination fail to do so. For example, of the 20 patients with PS who did not have a pyloric mass palpated, 3 (15%) had serum bicarbonate levels of > or =29 mmol/L, and 6 (30%) had serum chloride levels of < or =98 mmol/L. In conclusion, the serum bicarbonate or serum chloride level offers a useful additional diagnostic tool in the evaluation of children presenting during early infancy with vomiting of uncertain etiology.