Detecting gastroparesis: 3 hour versus 4 hour gastric emptying study

Objectives A recommended gastric emptying protocol (J Nucl Med Technol. 2008;36:44-54) requires imaging at 0, 1, 2, & 4 hrs. after ingestion of a standardized radiolabeled meal. The purpose of this retrospective investigation was to determine if the study could be terminated at 3 hrs. without adversely affecting test accuracy.

Methods The study group consisted of 109 patients, 34 men & 75 women (mean age 54±20 yrs.) who underwent a standard 4 hr. gastric emptying protocol. At the time of the test, no patient was taking medication that could affect test results. For purposes of the study, gastroparesis was defined as > 10% gastric retention at 4 hrs. Receiver Operating Characteristic (ROC) analysis was used to determine the gastric retention value at 3 hrs. that best discriminated between patients with & without gastroparesis.

Results ROC analysis indicated that a gastric retention value of > 25% at 3 hrs. most accurately predicted gastroparesis (ROC area = 95%±3%). By this criterion, there was “good agreement” (kappa = 0.77) between 3 & 4 hr. results for differentiating normal from abnormal gastric emptying, with no significant difference (McNemar Δ = 2.8%, p=0.45). Specificity (98%) of the 3 hr. test for detecting gastroparesis was high; sensitivity (75%) was lower. Positive predictive value of the 3 hr. test was 88%; the negative predictive value was 95%.

Conclusions More than 25% gastric retention at 3 hrs. accurately diagnoses gastroparesis & the study can be terminated; 25% or less gastric retention at 3 hrs., however, does not exclude gastroparesis, & the test must continue through 4 hrs.