An Updated BODE Index, New ADO Index May Improve Upon the Original Model

Pulmonary Reviews

Vol. 14, No. 10

October 2009

An Updated BODE Index, New ADO Index May Improve Upon the Original Model

With commentary by Stephen I. Rennard, MD

Key Point

Researchers have sought to improve the prognostic value of the BODE index by devising an updated version and by creating an alternative, simplified index.

The BODE index—a prognostic tool to determine COPD mortality risk that factors information on BMI, airflow obstruction, dyspnea, and exercise capacity (ie, six-minute walk distance [6MWD]) in a score ranging from 0 to 10—is a better predictor of outcomes than lung function alone; however, it correlated poorly with actual three-year mortality observed in two COPD cohorts, according to investigators reporting in the August 29 Lancet. They used these data to update the BODE index as well as to create a new index that takes into consideration age, dyspnea, and airflow obstruction (ADO). “The updated BODE index and the ADO index were similarly accurate in their risk prediction in an external validation cohort,” the investigators contended, adding that the simplified ADO index may be particularly suited to the primary care setting, where measurement of 6MWD is rarely performed.

“If we want to change the tremendous public health impact of COPD, we need to reach out to settings other than specialized pulmonary settings—such as primary care, where most COPD patients are treated,” lead author Milo A. Puhan, MD, PhD, from the Department of Epidemiology at Johns Hopkins Bloomberg School of Public Health, Baltimore, told Pulmonary Reviews. “The ADO index is in line with this strategy of knowledge transfer to areas where it is most needed.”

RECALIBRATION AND INNOVATION

Dr. Puhan and colleagues examined data from two different COPD populations: the Swiss Barmelweid cohort and the Spanish Phenotype and Course of COPD cohort. Patients in the Swiss cohort (n = 232) had long-standing and, on average, severe disease based on Global Initiative for Chronic Obstructive Lung Disease criteria. The Spanish cohort (n = 342) was composed of patients with a hospital stay (or an emergency department stay of at least 18 hours’ duration) for a first exacerbation. All patients were assessed for lung function, BMI, dyspnea, 6MWD, and three-year all-cause mortality.

On average, patients in the Swiss cohort were four years older and had more severe disease than those in the Spanish cohort. Physician-diagnosed cardiovascular disease was present in 38% of patients in the Swiss cohort versus 25% of patients in the Spanish cohort.

The investigators compared the observed three-year mortality risk in each cohort with the mortality risk predicted by the BODE index using calibration plots and goodness-of-fit statistics, and calculated the C statistic for the BODE score as a measure of discrimination.

It appeared that the BODE index significantly underpredicted three-year mortality risk in the Swiss cohort (mean predicted risk versus observed risk, 21.7% and 34.1%, respectively) and significantly overpredicted it in the Spanish cohort (16.7% vs 12.0%). Due to the poor calibration of the original BODE index, the investigators updated it based on results from the Swiss cohort. “We used the Swiss cohort because it had a high mortality incidence, yielding a more stable updated BODE index, and because the Swiss cohort was more similar to the original BODE study cohort than was the Spanish cohort,” they explained.

According to Dr. Puhan, “The four variables included in the updated BODE index are the same as in the original but the point system, which is on a scale of 0 to 15, better reflects the strength of the predictors. For example, 6MWD is a much stronger predictor of mortality than BMI, dyspnea, and lung function. Therefore, more points, meaning worse prognosis, must be assigned if 6MWD is low. For users, the big difference is that the score is now linked to explicit risks—for example, 10% for three-year mortality.”

The ADO risk index, which is scored on a scale of 0 to 10, replaces the 6MWD with age and does not take BMI into consideration. “When looking at the entire literature, BMI is associated with mortality in only about 30% of all studies,” Dr. Puhan noted. In the current study, age was shown to be the strongest predictor of three-year mortality in the Swiss cohort, followed by dyspnea and FEV₁, whereas there was no association between BMI and three-year mortality.

Validation of the both the updated BODE index and ADO index in the Spanish cohort showed good agreement between predicted and observed mortality risk.

THE ROLE OF PROGNOSIS

In an accompanying editorial, Holger Schünemann, MD, from the Department of Clinical Epidemiology and Biostatistics at McMaster University Health Sciences Centre, Hamilton, Ontario, asserted that the work by Dr. Puhan and colleagues have advanced COPD research. “However, prognostic information should not serve the sole purpose of knowing, but should improve important health outcomes.” Ideally, in addition to determining baseline risk and disease severity, prognostic tools also should balance the risks and benefits of disease management options and inform the discussion between clinicians and patients about treatment decisions, he added.

—Adriene Marshall

Stephen I. Rennard, MD, Larson Professor of Medicine, Chief of Pulmonary and Critical Care Medicine, Department of Internal Medicine, University of Nebraska Medical Center, Omaha, offers his perspective on this subject.

The description of the original BODE index by Celli and colleagues in 2004 has energized a branch of COPD epidemiology—namely determination of prognosis. The study by Puhan et al in the August 2009 Lancet provides two modifications. The first is a recalibration of the index that improves its predictive accuracy. The primary change was increasing the weighting placed on six-minute walk distance. The second was the creation of the ADO (age, dyspnea, airflow obstruction) index, which replaces BMI and six-minute walk distance with age.

One of the major motivating factors in the development of the BODE index was the heterogeneity of the disease. The underlying concept was that since COPD can affect different individuals in different ways, an index that accounted for multiple manifestations of the disease would provide a better prognostic measure. This concept is supported not only by the original study describing BODE, but also by a number of other studies—including the one by Puhan et al. The “recalibrated” BODE appears to be a better measure, and since it uses the same inputs as the original BODE, it represents no additional cost. If future research confirms it as a more accurate predictor, it will be an improvement over the original.

The ADO index may prove to be more easily applied in generalist offices; however, whether it will prove more useful than the BODE index is not yet known. In this context, it is a bit of a paradox that the single most heavily weighted parameter in the recalibrated BODE, the six-minute walk, is deleted in the ADO. Similarly, while BMI is reported to predict mortality in COPD in only 30% of reports, the effect is often quite strong. Moreover, for COPD it is not overweight but underweight that carries a bad prognosis. The BMI simply relates height to weight, but not all weight is the same. Many COPD patients have skeletal muscle depletion with a normal BMI due to increased fat mass. These individuals also have a poor prognosis. A revision of the BODE that replaces BMI with a measure of lean body mass may be a better choice. Whether the loss of inputs will limit the value of the ADO remains to be determined.

The utility of these prognostic indices will almost certainly depend on the purposes to which they are put. In the editorial that accompanied the current study, Dr. Schunemann reviewed potential uses for a prognostic index. Accurate prognosis will greatly inform the decision making required of patients and their health care providers. In addition, if an index can provide a validated surrogate that could be used in lieu of mortality, this would dramatically advance the development of treatments for COPD.

There are currently more than 60 novel therapies being evaluated for COPD. Based on recent experience, using mortality as an end point would require several thousand patients followed for a number of years—making the evaluation of these therapies very problematic. None of the indices are as yet validated for this purpose. The BODE index has been demonstrated to be sensitive to change. It is likely that the recalibrated version would as well, raising the possibility that it has potential in this regard.

The BODE index has been a major advance in our approach to COPD. It can be improved. Increasing the accuracy, simplification for integration into practice settings, and validation as surrogates for more difficult clinical outcomes will be important ongoing goals.

Return to table of contents