Introduction: Characteristics of patients undergoing Upper GI endoscopy (UGIE) in Kumasi, Ghana are largely unknown. This paper reviews the work of three endoscopy units in Kumasi.

Methods: A review of the records of patients undergoing diagnostic UGIE in the three centers from October 2006 to December 2011 was undertaken.

Results: 3110 completed UGIE were performed over the period. In 80% of the patients the primary indication for UGIE was dyspepsia occurring without any other symptom. In 50% of patients UGIE findings were normal. Peptic ulcer disease, the most common positive finding, was diagnosed in 27.4% of patients. The odds ratio (O.R) of yielding a positive endoscopic finding for patients presenting with recurrent vomiting, dyspepsia associated with weight loss and UGI bleeding were 3.87 (95% C.I: 2.23-6.69), 1.72 (95% C.I: 1.03-2.87) and 1.81 (95% C.I: 1.03-3.16) respectively. Dyspepsia without any other symptom, on the other hand, yielded a positive endoscopic finding with O.R of 0.67 (95% CI: 0.57-0.80). Alarm symptoms (UGI bleeding, recurrent vomiting, dysphagia or weight loss associated with dyspepsia) yielded a positive endoscopic finding with an O.R of 2.34 (95% C.I: 1.74-3.13).

Conclusion: Most patients in Kumasi underwent UGIE because of dyspepsia in the absence of any other symptom. These patients were more likely to have normal endoscopic findings. The opposite was true for those presenting with alarm symptoms. Our results suggest that initial UGIE may be preferentially offered to patients presenting with alarm symptoms especially in resource-poor settings such as ours.