This article is more than 5 years old. Information may no longer be current.
Mathematic diagnostic model predicted likelihood of acute HF presence
Click Here to Manage Email Alerts
A mathematic diagnostic prediction model incorporating clinical assessment and N-terminal pro b-type natriuretic peptide can predict the likelihood of the presence of acute HF.
The researchers included 483 patients from the IMPROVE-CHF trial in the analysis. Patients were classified according to their probability for acute HF, with ≤20% representing low probability (n=163), 21% to 79% with intermediate probability (n=184) and ≥80% with a high probability (n=136). Patients were compared with blind adjudicated acute HF diagnosis in 573 patients from the PRIDE trial.
The likelihood ratio for acute HF with N-terminal pro b-type natriuretic peptide (NT-proBNP) was 0.11 (95% CI, 0.06-0.19) for cut-point values <300 pg/mL. The ratio increased to 3.43 (95% CI, 2.34-5.03) for values 2,700 pg/mL to 8,099 pg/mL and 12.80 (95% CI, 5.21-31.45) for values ≥8,100 pg/mL. The researchers used age, pretest probability for acute HF and log NT-proBNP as variables and reported that external validation when compared with the PRIDE trial population confirmed discrimination (c=0.97) and reclassification improvement (P<.001). The model appropriately reclassified 44% of patients of intermediate probability to either high or low probability for acute HF, with <2% reclassified in an inappropriate direction.
“A diagnostic prediction model for acute HF that uses clinical assessment and NT-proBNP value has been derived and externally validated to appropriately direct the physician in a significant number of indeterminate cases,” the researchers concluded. “Further studies of implementation, cost and impact analyses will help define the model’s general utility across all levels of clinical certainty and may foster similar analyses of other categorical diagnostic tests.”
In an accompanying editorial, Kenneth Dickstein, MD, PhD, a professor of medicine at Stavanger University Hospital in Rogaland, Norway, praised the model but also warned of the limitations, such as the exclusion of certain patients with common ACS, as well as patients with infection, obstructive airway disease or moderate renal dysfunction.
“For giving us a glimpse of the inevitable future, Steinhart et al deserve credit,” Dickstein wrote. “The article demonstrates that the model has excellent diagnostic accuracy, especially in cases of intermediate clinical probability. In these patients, the model appropriately reclassified the likelihood to either low or high probability of acute HF with negligible inappropriate redirection.”
For more information:
- Steinhart B. J Am Coll Cardiol. 2009;54:1515-1521.
This is an excellently done study in which researchers employed a sophisticated model to answer a very simple question. That question is: “Which patients, based on their pretest probability of HF, benefit from obtaining a natriuretic peptide value for diagnosis?” The answer, not unexpectedly, is that those who benefit are those with an intermediate probability of disease. There may be other reasons to obtain baseline levels in those with high or low pretest probabilities, but for diagnosis, this study’s results supported the Baysian nature of clinical testing.
– Allan S. Jaffe, MD
Cardiology Today Editorial Board member