Limits...
The Relation Between Global Longitudinal Strain and Serum Natriuretic Peptide Is More Strict Than That Found Between the Latter and Left Ventricular Ejection Fraction: A Retrospective Study in Chronic Heart Failure.

De Vecchis R, Baldi C, Di Biase G - J Clin Med Res (2015)

Bottom Line: As regards the two determinations, the one echocardiographic and the other laboratory-based, the former should have been done not more than 24 hours before or after the latter.However, the C statistics for GLS were significantly higher than for LVEF (area under the curve (AUC): 0.949 (GLS) vs. 0.730 (LVEF); P = 0.0030).Thus, in both CHF with preserved and reduced LVEF, GLS is more accurate compared with LVEF in predicting increased levels of NT-proBNP.

View Article: PubMed Central - PubMed

Affiliation: Cardiology Unit, Presidio Sanitario Intermedio "Elena d'Aosta", ASL Napoli 1 Centro, Napoli, Italy.

ABSTRACT

Background: In chronic heart failure (CHF), the finding of elevated levels of the N-terminal fragment of the pro B-type natriuretic peptide (NT-proBNP) is a marker of pathological increase in myocardial ventricular wall stress and detrimental rise in ventricular filling pressures. However, the ensemble of data concerning the relationship between longitudinal deformation indices and NT-proBNP is still rather vague and approximate.

Methods: We carried out a retrospective study that involved 118 patients with CHF admitted to our clinic for CHF outpatients. For inclusion in the study, the CHF patients were required to have undergone at least a determination of global longitudinal strain (GLS) by means of speckle tracking echocardiography and to have practiced at least a determination of NT-proBNP. As regards the two determinations, the one echocardiographic and the other laboratory-based, the former should have been done not more than 24 hours before or after the latter.

Results: Correlation between log (NT-proBNP) and GLS was highly significant (r = 0.8386; P < 0.0001). The observed correlation between log (NT-proBNP) and left ventricular ejection fraction (LVEF) was also significant, but explained a smaller magnitude of the variance (r = -0.5465; P < 0.0001). In multiple linear regression analysis, GLS was shown to be the strongest independent predictor of log (NT-proBNP), within a parsimonious model including age, body mass index, estimated glomerular filtration rate, left atrial volume index, and LVEF (β (regression coefficient) = 305, rpartial = 0.7076; P < 0.0001). By using the median value of NT-proBNP (299.5 pg/mL) as a discriminating value for identifying relatively low (i.e., below the median) and relatively high (i.e., above the median) levels of NT-proBNP, GLS was associated with the upper quartiles, whereas LVEF was associated with lower quartiles of NT-proBNP. However, the C statistics for GLS were significantly higher than for LVEF (area under the curve (AUC): 0.949 (GLS) vs. 0.730 (LVEF); P = 0.0030).

Conclusions: In CHF patients, GLS shows a stronger association with NT-proBNP levels with respect to LVEF. Thus, in both CHF with preserved and reduced LVEF, GLS is more accurate compared with LVEF in predicting increased levels of NT-proBNP.

No MeSH data available.


Related in: MedlinePlus

In this ROC plot, there is the representation of the diagnostic performance (AUC = 0.949) of GLS as a predictor of relatively high values (> 299.5 pg/mL) of NT-proBNP among 50 CHF patients. GLS: global longitudinal strain; ROC: receiver operating characteristic; NT-proBNP: N-terminal fragment of the pro B-type natriuretic peptide; pg: pictograms.
© Copyright Policy - open access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4625819&req=5

Figure 5: In this ROC plot, there is the representation of the diagnostic performance (AUC = 0.949) of GLS as a predictor of relatively high values (> 299.5 pg/mL) of NT-proBNP among 50 CHF patients. GLS: global longitudinal strain; ROC: receiver operating characteristic; NT-proBNP: N-terminal fragment of the pro B-type natriuretic peptide; pg: pictograms.

Mentions: However, the C-statistics for GLS were significantly higher than for LVEF (AUC: 0.949 (GLS) vs. 0.730 (LVEF); P = 0.0030) (Fig. 5-7). In other words, the ability of GLS to predict the association with a relatively high level of NT-proBNP, i.e., located above its 50th percentile, was compared with that exhibited by LVEF. In this way, the value of GLS ≥ -17.7% was shown to have the best diagnostic accuracy (sensitivity = 92%; specificity = 96%; positive likelihood ratio (+LR) = 23; negative likelihood ratio (-LR) = 0.083) in predicting the presence of relatively high (i.e., above the median) levels of NT-proBNP, with an AUC of 0.949. Similarly, among all values of LVEF detected in the patient population, an LVEF ≤ 56% could achieve the best diagnostic accuracy (sensitivity = 88%; specificity = 56 %; +LR = 2; -LR = 0.21) in the detection of relatively high levels (i.e., above the median) of NT-proBNP, although showing an AUC of 0.730, namely, significantly lower than that exhibited by GLS (P = 0.0030) (Fig. 7). As already specified in the previous section (“Echocardiography”), all echocardiographic measurements concerning the strain were made by only one experienced operator who per customary practice was almost always kept blinded to clinical and biochemical data concerning the patients. For this experienced sonographer, Bland-Altman analysis demonstrated a good intra-observer agreement with a small non-significant bias for GLS. In fact, mean difference ± 2 standard deviations for GLS was 0.30±0.7% with a percentage of error of 3.1% calculated from the repeated examination of 20 successive individuals.


The Relation Between Global Longitudinal Strain and Serum Natriuretic Peptide Is More Strict Than That Found Between the Latter and Left Ventricular Ejection Fraction: A Retrospective Study in Chronic Heart Failure.

De Vecchis R, Baldi C, Di Biase G - J Clin Med Res (2015)

In this ROC plot, there is the representation of the diagnostic performance (AUC = 0.949) of GLS as a predictor of relatively high values (> 299.5 pg/mL) of NT-proBNP among 50 CHF patients. GLS: global longitudinal strain; ROC: receiver operating characteristic; NT-proBNP: N-terminal fragment of the pro B-type natriuretic peptide; pg: pictograms.
© Copyright Policy - open access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4625819&req=5

Figure 5: In this ROC plot, there is the representation of the diagnostic performance (AUC = 0.949) of GLS as a predictor of relatively high values (> 299.5 pg/mL) of NT-proBNP among 50 CHF patients. GLS: global longitudinal strain; ROC: receiver operating characteristic; NT-proBNP: N-terminal fragment of the pro B-type natriuretic peptide; pg: pictograms.
Mentions: However, the C-statistics for GLS were significantly higher than for LVEF (AUC: 0.949 (GLS) vs. 0.730 (LVEF); P = 0.0030) (Fig. 5-7). In other words, the ability of GLS to predict the association with a relatively high level of NT-proBNP, i.e., located above its 50th percentile, was compared with that exhibited by LVEF. In this way, the value of GLS ≥ -17.7% was shown to have the best diagnostic accuracy (sensitivity = 92%; specificity = 96%; positive likelihood ratio (+LR) = 23; negative likelihood ratio (-LR) = 0.083) in predicting the presence of relatively high (i.e., above the median) levels of NT-proBNP, with an AUC of 0.949. Similarly, among all values of LVEF detected in the patient population, an LVEF ≤ 56% could achieve the best diagnostic accuracy (sensitivity = 88%; specificity = 56 %; +LR = 2; -LR = 0.21) in the detection of relatively high levels (i.e., above the median) of NT-proBNP, although showing an AUC of 0.730, namely, significantly lower than that exhibited by GLS (P = 0.0030) (Fig. 7). As already specified in the previous section (“Echocardiography”), all echocardiographic measurements concerning the strain were made by only one experienced operator who per customary practice was almost always kept blinded to clinical and biochemical data concerning the patients. For this experienced sonographer, Bland-Altman analysis demonstrated a good intra-observer agreement with a small non-significant bias for GLS. In fact, mean difference ± 2 standard deviations for GLS was 0.30±0.7% with a percentage of error of 3.1% calculated from the repeated examination of 20 successive individuals.

Bottom Line: As regards the two determinations, the one echocardiographic and the other laboratory-based, the former should have been done not more than 24 hours before or after the latter.However, the C statistics for GLS were significantly higher than for LVEF (area under the curve (AUC): 0.949 (GLS) vs. 0.730 (LVEF); P = 0.0030).Thus, in both CHF with preserved and reduced LVEF, GLS is more accurate compared with LVEF in predicting increased levels of NT-proBNP.

View Article: PubMed Central - PubMed

Affiliation: Cardiology Unit, Presidio Sanitario Intermedio "Elena d'Aosta", ASL Napoli 1 Centro, Napoli, Italy.

ABSTRACT

Background: In chronic heart failure (CHF), the finding of elevated levels of the N-terminal fragment of the pro B-type natriuretic peptide (NT-proBNP) is a marker of pathological increase in myocardial ventricular wall stress and detrimental rise in ventricular filling pressures. However, the ensemble of data concerning the relationship between longitudinal deformation indices and NT-proBNP is still rather vague and approximate.

Methods: We carried out a retrospective study that involved 118 patients with CHF admitted to our clinic for CHF outpatients. For inclusion in the study, the CHF patients were required to have undergone at least a determination of global longitudinal strain (GLS) by means of speckle tracking echocardiography and to have practiced at least a determination of NT-proBNP. As regards the two determinations, the one echocardiographic and the other laboratory-based, the former should have been done not more than 24 hours before or after the latter.

Results: Correlation between log (NT-proBNP) and GLS was highly significant (r = 0.8386; P < 0.0001). The observed correlation between log (NT-proBNP) and left ventricular ejection fraction (LVEF) was also significant, but explained a smaller magnitude of the variance (r = -0.5465; P < 0.0001). In multiple linear regression analysis, GLS was shown to be the strongest independent predictor of log (NT-proBNP), within a parsimonious model including age, body mass index, estimated glomerular filtration rate, left atrial volume index, and LVEF (β (regression coefficient) = 305, rpartial = 0.7076; P < 0.0001). By using the median value of NT-proBNP (299.5 pg/mL) as a discriminating value for identifying relatively low (i.e., below the median) and relatively high (i.e., above the median) levels of NT-proBNP, GLS was associated with the upper quartiles, whereas LVEF was associated with lower quartiles of NT-proBNP. However, the C statistics for GLS were significantly higher than for LVEF (area under the curve (AUC): 0.949 (GLS) vs. 0.730 (LVEF); P = 0.0030).

Conclusions: In CHF patients, GLS shows a stronger association with NT-proBNP levels with respect to LVEF. Thus, in both CHF with preserved and reduced LVEF, GLS is more accurate compared with LVEF in predicting increased levels of NT-proBNP.

No MeSH data available.


Related in: MedlinePlus