Comparison of 2011 and 2007 Global Initiative for Chronic Obstructive Lung Disease guidelines for predicting mortality and hospitalization

Am J Respir Crit Care Med. 2013 Jul 1;188(1):51-9. doi: 10.1164/rccm.201212-2276OC.


Rationale: The GOLD (Global Initiative for Chronic Obstructive Lung Disease) 2011 update on chronic obstructive pulmonary disease (COPD) bases disease classification on level of dyspnea, exacerbation history, and FEV1, whereas the previous GOLD categorized disease severity according to FEV1 only. Information on how the new classification predicts long-term hospitalizations and mortality is limited.

Objectives: To examine how GOLD 2011 predicts hospitalizations and mortality over an 8-year period and to assess differences in predictive ability between GOLD 2011 and GOLD 2007.

Methods: In the GenKOLS study, 912 patients with COPD (FEV1/FVC < 0.7 and FEV1 < 80% predicted) aged 40 to 91 years were clinically examined. Patients answered questionnaires and performed lung function testing in 2003-2005. The population was followed for 8 years regarding hospitalizations (all-cause, respiratory) and mortality (all-cause, respiratory, cardiovascular). We performed logistic regression and receiver operating curve analyses for GOLD 2007 and GOLD 2011 with estimations of area under the curve (AUC) to compare the different classifications.

Measurements and main results: Twenty percent of patients were classified as GOLD 2011 group A (mild), 30% as group B, 6% group as C, and 44% as group D (very severe). Patients in GOLD 2011 group D had odds ratios of 4.1 (95% confidence interval [CI], 2.5-6.7), 9.6 (95% CI, 3.4-27.0), and 3.0 (95% CI, 0.7-13.2) relative to group A for all-cause, respiratory, and cardiovascular mortality, respectively, and 3.8 (95% CI, 2.4-5.9) and 13.0 (95% CI, 6.6-25.6) for all-cause and respiratory hospitalizations, respectively. Associations were similar also for GOLD 2007. The adjusted AUC values for GOLD 2011 and GOLD 2007 were 0.82/0.82 for respiratory mortality (P = 0.87) and 0.77/0.76 for respiratory hospitalizations (P = 0.51).

Conclusions: The predictive ability of GOLD 2011 did not differ significantly from GOLD 2007 in terms of hospitalizations and mortality.

Publication types

  • Comparative Study

MeSH terms

  • Adult
  • Aged
  • Aged, 80 and over
  • Area Under Curve
  • Cohort Studies
  • Female
  • Follow-Up Studies
  • Hospitalization / statistics & numerical data*
  • Humans
  • Internationality*
  • Length of Stay / statistics & numerical data
  • Male
  • Middle Aged
  • Odds Ratio
  • Population Surveillance / methods*
  • Predictive Value of Tests
  • Pulmonary Disease, Chronic Obstructive / diagnosis*
  • Pulmonary Disease, Chronic Obstructive / epidemiology*
  • ROC Curve
  • Respiratory Function Tests / methods
  • Respiratory Function Tests / statistics & numerical data
  • Severity of Illness Index
  • Surveys and Questionnaires
  • Survival Rate