Purpose: The efficacy of surveillance esophagoduodenoscopy (EGD) for esophageal adenocarcinoma (EAC) is controversial. We aimed to examine the effect of EGD at least 1 year prior to the diagnosis of EAC on survival after the diagnosis of EAC among United States veterans with gastroesophageal reflux (GER).
Methods: The national administrative database of the Veterans Administration was accessed, and patients diagnosed with EAC from 1995 through 2003 who had a prior diagnosis of GER were identified. Data were collected for EGDs performed up to 5 years prior to cancer diagnosis. Electronic medical records were abstracted for confirmation of EAC, date of diagnosis, stage at diagnosis, therapy, and date of death.
Results: 155 confirmed cases of EAC were identified. Mean age at diagnosis was 68.9; 154 were male, and 3.5% were African American. Cancer mortality was associated with increasing stage at diagnosis (HR 1.74, 95% CI 1.45–2.10), Charlson comorbidity index (HR 1.18, 95% CI 1.03–1.35), and age (HR 1.02, 95% CI 1.00–1.04). Patients with a history of EGD at least 1 year prior to diagnosis of EAC (N = 25) were diagnosed at earlier stages than those without prior EGD (stages I, II, III, IV: 28%, 44%, 12%, 16% vs. 13%, 36%, 23%, 28%, p= 0.02), and were more likely to undergo surgical resection, controlling for age and comorbidity (OR 2.1, 95% CI 0.8 – 5.4). Patients with prior EGD appeared to have better short-term cancer survival than those without prior EGD, but there was no survival advantage after 6 years of follow-up (figure). Controlling for age and comorbidities, or examining all-cause mortality yielded similar results.
Conclusions: Surveillance for esophageal adenocarcinoma improved the stage at diagnosis, and the likelihood of receiving surgical resection, but did not alter long-term cancer-associated or all-cause survival among this cohort. This is likely due to the poor long-term survival of even early stage EAC. Lead-time effects may have biased prior short-term studies toward finding a benefit from endoscopic surveillance. [figure 1]