Risk of thromboembolic events in patients with AMD examined

"Neovascular AMD affects older adults who are already at increased risk for cardiovascular and cerebrovascular events," noted Diana V. Do, MD, assistant professor of ophthalmology, Wilmer Eye Institute, Johns Hopkins University School of Medicine, Baltimore.

"Experience with intravenous administration of anti-VEGF therapy in cancer patients suggests that there is an increased rate of thromboembolic events in patients given VEGF inhibitors in combination with chemotherapy compared with patients given chemotherapy alone," Dr. Do said. "We do not know if the same risk applies to administration of intravitreal VEGF inhibitors."

"We believe ophthalmologists who care for patients with AMD should be informed about the risks of MI and CVA in those individuals," Dr. Do added. "We hope to examine additional databases to see if we can compare the incidence of MI or CVA between individuals with neovascular AMD and matched controls without AMD."

Although there is published information about incidence and mortality rates for MI and CVA in older adults, investigators at Wilmer Eye Institute and Genentech undertook their research based on the premise that the risk may differ in AMD patients depending on the number and seriousness of medical comorbidities. They performed a retrospective database analysis to determine rates of new and recurrent inpatient MI and stroke (ischemic or hemorrhagic) in patients with neovascular AMD with the subjects categorized by level of medical comorbidities.

A health-care utilization administrative database was used as the source to identify individuals ages 50 years and older who had been enrolled for at least 6 months with no inpatient claims for cardiovascular or cerebrovascular events, had an index neovascular AMD claim after that eligibility period, and at least two claims for neovascular AMD. The records of patients meeting those inclusion criteria were then reviewed for inpatient claims for MI or stroke after the index AMD claim. The study period ranged from 2002 to 2005.

CCI utilized

To determine the impact of comorbidities on the risk for the cardiovascular and cerebrovascular outcomes, the researchers adapted the Charlson Comorbidity Index (CCI), an existing, widely used tool for measuring the collective burden of comorbid conditions. The CCI assigned weights of 1 to 6 for medical comorbid conditions based on 1-year mortality rates occurring in a cohort of 604 inpatients. Scores for each condition a patient had were summed to give a total score that reflected the number and seriousness of the comorbid disease(s).

To rate comorbidity severity in their study, Dr. Do and colleagues reviewed the patients' records during the 6-month period prior to diagnosis of AMD for comorbidities and then categorized the patients into three groups based on an adapted CCI score: low (CCI = 0), moderate (CCI = 1 to 2), or high (CCI = 3+).

Their study included more than 7,000 patients, of whom about 60% were female. Comorbidity was low for about half of the patients and moderate for about one-third.

The analyses showed that incidence rates for MI, ischemic stroke, hemorrhagic stroke, and all events combined increased with increasing severity of comorbidity.

"We found that patients with a high CCI were twice as likely to have any of the events compared with those with a low CCI, and the differences between those groups were statistically significant," she said.

Within each comorbidity category, the rate of MI was also higher in patients ages 75 years and older compared with those younger than 75 years of age, although the differences were not statistically significant. In an analysis of the effect of age and comorbidity on stroke risk, older patients in the low and moderate CCI categories had significantly higher stroke rates than their younger counterparts. Among those with a high CCI, the stroke rate was higher in younger versus older patients, although the difference was not statistically significant.

"[The analysis] is based on a very large database that includes about 4 million enrollees aged 50 years and older and who represent a geographically diverse population. In addition, it is based on claims for medical care received and existing, available data," she explained. "However, it does not differentiate between index and recurrent events. In addition, because the database only captures what is billed, it may not include some fatal events. Furthermore, we relied on ICD-9 codes for identifying patients with neovascular AMD, CVA, and MIs. Those data were not validated, but we expect them to have reasonably high accuracy."