COPD Scoring System May Predict Near-Term Mortality

Karla Harby

September 19, 2007 (Stockholm) — A scoring system that uses readily obtainable clinical findings to predict all-cause, near-term morality appears viable in chronic obstructive pulmonary disease (COPD), according to a population-based study in Scotland, said principal investigator Stuart Schembri, MD, from the Thoracic Unit of Ninewells Hospital, Dundee, United Kingdom. He added that his "most surprising" finding was the observation that vaccination against influenza and pneumococcal infection appeared protective in these patients.
Previous studies have identified variables that appear to be associated with mortality in COPD, including age, smoking history, and forced expiratory volume in 1 second (FEV1), Dr. Schembri told attendees here at the European Respiratory Society (ERS) 17th Annual Congress, but the goal of this study was to combine many of these variables into one model that would predict patients' all-cause mortality during the 5-year study period.
The investigators identified 6945 individuals for screening who lived in Dundee from January 2001 through December 2005. The researchers excluded 38 persons who left the area during this time, as well as another 16 persons who had never smoked, from any further analysis. They defined COPD as an FEV1 of less than 80% and an obstructive ratio (FEV1/ forced vital capacity [FVC]) of less than 70%. A total of 3327 patients were diagnosed with COPD, they were followed up every year. During this 5-year period, 550 COPD patients died.
At every clinic visit, the investigators assessed and measured a long list of variables — so many variables, in fact, that the data were entered into a computer program for tracking and analysis, Dr. Schembri said. In all there were more than 10,000 patient interviews, and missing data were handled by creating a value of "unknown = zero."
Dr. Schembri explained that nonsignificant variables were dropped to create a reduced model of only significant predictors. The researchers discovered that a relatively simple scoring system was just as accurate as an area-under-the-curve model, so they abandoned their model in favor of the scoring system. Under this system, a negative numerical score for a variable is protective, whereas positive numbers suggest increased risk of mortality within 5 years.
Among his numerous findings, Dr. Schembri said that being obese was found to be protective (score of −2), but not being morbidly obese (the positive score varies with body mass index [BMI]). Being a current smoker was assigned a score of 1, whereas having smoked only slightly in one's lifetime was assigned a score of −5 (strongly protective). The sum of such scores can then be compared with the cumulative risk algorithm, a graph the investigators created that predicts the patient's likelihood of death within 1, 2, 3, 4, and 5 years, given the patient's score. The data can also be entered into a computer program that solves an equation to provide results.
Dr. Schembri noted that although the population survey approach is a strength of this study, as a consequence of so many patient interviews, quite a few data are missing. For example, the investigators are missing about 3% of the BMI data. In addition, all of these data were generated among whites living in Scotland, and research suggests that different racial and ethnic groups experience different mortality in this disease, he said.
The cochair of the session was John R. Hurst, MD, MRCP, from the Academic Unit of Respiratory Medicine, Royal Free and University College Medical School, London, United Kingdom. Asked whether vaccination may simply be a proxy for more attentive care overall, Dr. Hurst told Medscape Pulmonary Medicine, "I think that's possible, but there is already existing data that vaccination is protective and can reduce hospitalization and mortality." He added that some patients will decline vaccination, saying they react badly to vaccines, whereas others may lack access to healthcare or may have healthcare providers who are unaware of the value of vaccination among these patients.
"We don't have a widely applicable score at the moment," Dr. Hurst added. "If [Dr. Schembri] and his team can replicate these results in different populations, we could use that by showing patients what their risk of hospitalization is, and potentially of death."
This work was financed by unrestricted research grants from Pfizer and Boehringer Ingelheim. This abstract was awarded the Best Abstract prize among all those presented by British junior doctors at the ERS meeting, an award financed by GlaxoSmithKline, Dr. Hurst told Medscape. This abstract was selected by independent physicians, and the award is completely unrelated to the financing of the research, he added.
European Respiratory Society 17th Annual Congress: Abstract 4248. Presented September 19, 2007.