Subject Category: Esophagus

Am J Gastroenterol 2012; 107:1826–1832; doi:10.1038/ajg.2012.342; published online 23 October 2012

Symptom Reports Are Not Reliable During Ambulatory Reflux Monitoring

Robert T Kavitt MD1, Tina Higginbotham MPA1, James C Slaughter DrPH2, Dilan Patel3, Elif Saritas Yuksel MD1, Zurabi Lominadze1, Anas Abou-Ismail MD1, Trisha Pasricha1, C Gaelyn Garrett MD4, David Hagaman MD5 and Michael F Vaezi MD, PhD, MSc (Epi)1

  1. 1Division of Gastroenterology, Hepatology and Nutrition, Vanderbilt University Medical Center, Nashville, Tennessee, USA
  2. 2Department of Biostatistics, Vanderbilt University Medical Center, Nashville, Tennessee, USA
  3. 3University of Tennessee Graduate School of Medicine, Knoxville, Tennessee, USA
  4. 4Vanderbilt University Medical Center, Voice Center, Nashville, Tennessee, USA
  5. 5Vanderbilt University Medical Center, Allergy, Sinus and Asthma Program, Nashville, Tennessee, USA

Correspondence: Michael F. VaeziMD, PhD, MSc (Epi), Division of Gastroenterology, Hepatology, and Nutrition, Vanderbilt University Medical Center, 1660 TVC, 1301-22nd Avenue South, Nashville, Tennessee 37232-5280, USA. E-mail:

Received 30 April 2012; Accepted 2 August 2012
Advance online publication 23 October 2012





Patient reporting of symptom events during ambulatory reflux monitoring is commonly performed with little data regarding its accuracy. We employed a novel time-synchronized ambulatory audio recording of symptom events simultaneously with prolonged pH/impedance monitoring to assess temporal accuracy of patient-reported symptoms.



An acoustic monitoring system was employed to detect cough events via tracheal and chest wall sounds and it was temporally synchronized with an ambulatory impedance/pH monitoring system. Patients were instructed to record their symptoms in the usual manner. Six separate observers independently listened to the 24-h audio recordings and logged the exact timing of each cough event. Patients were blinded to study design and the audio reviewers were blinded to their own reports and those of patients and other reviewers. Concurrence of audio recordings and patient-reported symptoms were tested for three separate time thresholds: 1, 2, and 5min.



The median (interquartile range (IQR)) number of cough events by audio detection was significantly (P<0.001) higher than those reported by patients: 216 (90–275) and 34 (22–60), respectively. There was significantly (P<0.001) higher agreement among the audio recording listeners (substantial to almost perfect agreement; kappa=0.77–0.82) than between the audio recording and patient-reported symptoms (slight to fair agreement; kappa=0.13–0.27). Patients did not report 91, 82, and 71% of audible cough events based on 1-, 2-, and 5-min concordance time windows, respectively.



We found that patients do not report the majority of their symptoms during ambulatory reflux monitoring even within a 5-min time window of the true event and advise caution in clinical decision-making based solely on symptom indices.