We appreciate the opportunity to respond to Dr. Spinzi and colleagues. We completely agree that the benefits and risks of diagnostic and therapeutic interventions should be weighed by the treating physician and explained to the patient; doing such is part of our moral obligation. However, a careful examination of the erroneous extrapolation of population-based assumptions about radiation effects to the patients we studied (that is, those with compensated cirrhosis), as well as a more holistic view of the patient and his or her alternatives (no surveillance or endoscopic surveillance), strongly argue for the benefit of computed tomographic (CT) imaging in variceal screening.
The median survival of patients with compensated cirrhosis is estimated to be 7–10 years from the time of diagnosis, with the development of varices reducing median survival of 4–7 years.1 Thus, the life expectancy of these patients is much less than the decade or more it takes to even begin to accumulate even a tiny theoretical risk of radiation–induced malignancy.2 Additionally, the risk for the annual development of esophageal varices is estimated to be 4%–12%, with the 2-year risk of bleeding from large esophageal varices being about 25%.1 Each episode of variceal bleeding is associated with a 6-week mortality of about 20%. These estimates for mortality and significant morbidity far outweigh the minimal risks of surveillance CT, which is overstated in Dr. Spinzi's letter. Consider, for example, that even accepting the linear no-threshold model for radiation-induced cancer risk (which is open to some debate),3 the risk of radiation-induced cancer falls dramatically after 35 years of age,4 Most patients with cirrhosis who require screening for varices are older. Moreover, in older patients, the greatest risks of cancer induction by radiation are lung cancer and leukemia, and neither the lung nor the bone marrow are among the organs most exposed during an abdominal CT scan.2 Dr. Spinzi and colleagues incorrectly reference the National Academy of Sciences, which does support a linear nonthreshold model for radiation-induced malignancy, but states that at “doses at 100 mSv or less, statistical limitations make it difficult to evaluate cancer risk in humans.”5 The dose used for biphase liver CT to screen for varices is 15 mSv. Moreover, the BEIR (Biological Effects of Ionizing Radiation) risk model referenced is a lifetime risk model based on populations exposed to whole-body radiation, not older individuals with limited exposure to the abdominal cavity.5
Finally, the risks and benefits of screening CT must be balanced against the alternatives, such as no surveillance (previously discussed) or screening endoscopy. Despite guidelines by multiple bodies, including the American Association for the Study of Liver Diseases, The Baveno Consensus panel, and the American College of Gastroenterology, a recent survey of community gastroenterologists found that only 54% perform screening endoscopy in patients with cirrhosis.6 CT scanning would, of course, be beneficial to patients who do not undergo endoscopic screening. Routine endoscopic screening carries risks of conscious sedation. The risks of diagnostic endoscopy are low and estimated to be 1.1% in patients with cirrhosis.7 These risks exceed the very small risk of radiation-induced malignancy from a CT scan. Most important, patients preferred CT screening to endoscopic screening, and when cost-effectiveness is taken into account, CT screening is the preferred modality.8 Unlike screening endoscopy, CT has the advantage of imaging beyond the confines of the esophageal lumen and can identify hepatocellular carcinomas and periesophageal varices, which can portent recurrent bleeding.9 In our study, 6% of patients had a hepatocellular carcinoma or other neoplasm not seen on screening ultrasound, demonstrating another advantage of CT over endoscopy.
Given multiple studies that demonstrate the benefit of medical and endoscopic therapy in patients with large varices, the age of patients with compensated cirrhosis who would presumably undergo CT screening, the poor compliance with routine endoscopic surveillance, the ability of CT to detect clinically significant pathologies beyond the esophagus, and the limited and small radiation dose to the patients, we feel the benefit-risk ratio is largely in favor of CT in these patients.