Background and study aims: Surveillance in Barrett's esophagus relies on the detection of dysplasia by histopathology. However, the natural history of this condition, particularly that of low-grade dysplasia (LGD) is poorly understood. This paper describes our experience of LGD over a period of 21 years.
Patients and methods: Between 1984 and January 1995, 357 patients with Barrett's esophagus without dysplasia were recruited for annual surveillance: 34 of these patients developed LGD during this period. This was a retrospective cohort study of this group in terms of survival and cancer outcomes >/= 8 years after the original diagnosis of LGD, comparing them with the patients who did not develop LGD over the same period, with a histopathological review of the original diagnoses of LGD. The outcomes of 356/357 (99.7 %) of the patients were established in December 2004.
Results: After 8 years, high-grade dysplasia (HGD) or cancer had developed in 9/34 patients with LGD (27 %) and in 16/322 controls (5 %). Cox's proportional hazards model revealed that the time from the first diagnosis of Barrett's esophagus to the first "event" of either HGD, esophageal cancer, or death did not show a statistically significant difference between the two groups. A further analysis treating death as "loss to follow-up" showed a significantly increased risk for the LGD group to progress to HGD or cancer (hazard ratio 5.9 [95 % confidence interval 2.6 - 13.4], P< 0.001). The histopathology review demonstrated a fair level of agreement between pathologists, with a kappa value of 0.48.
Conclusions: Patients diagnosed with LGD during surveillance of Barrett's esophagus are at a considerably increased risk of progressing to develop esophageal cancer over an 8-year period but most deaths are not cancer-related.