Sendaway capillary NT-proBNP in pulmonary hypertension

Abstract

Background: N-terminal pro-B-type natriuretic peptide (NT-proBNP) is a biomarker of cardiac ventricular wall stress that is incorporated into pulmonary hypertension (PH) risk stratification models. Sendaway sampling may enable patients to perform NT-proBNP tests remotely. This UK-wide study aimed to assess the agreement of sendaway NT-proBNP with standard venous NT-proBNP and to assess the effect of delayed processing.

Methods: Reference venous NT-proBNP was collected from PH patients. Samples for capillary and venous sendaway tests were collected contemporaneously, mailed to a reference laboratory and processed at 3 and 7 days using a Roche Cobas e411 device. Differences in paired measurements were analysed with Passing-Bablok regression, percentage difference plots and the % difference in risk strata.

Results: 113 patients were included in the study. 13% of day 3 capillary samples were insufficient. Day 3 capillary samples were not equivalent to reference samples (Passing Bablok analysis slope of 0.91 (95% CI 0.88 to 0.93) and intercept of 6.0 (95% CI 0.2 to 15.9)). The relative median difference was -7% and there were acceptable limits of agreement. Day 3 capillary NT-proBNP accurately risk stratified patients in 93.5% of cases. By comparison, day 3 venous results accurately risk stratified patients in 90.1% of cases and were equivalent by Passing-Bablok regression. Delayed sampling of sendaway tests led to an unacceptable level of agreement and systematically underestimated NT-proBNP.

Conclusions: Sendaway NT-proBNP sampling may provide an objective measure of right ventricular strain for virtual PH clinics. Results must be interpreted with caution in cases of delayed sampling.

Keywords: Primary Pulmonary Hypertension.

Figure 1

Analysis comparing reference and sendaway day 3 capillary NT-proBNP (pg/mL) using (A) percentage difference plot demonstrating a median difference of −7%, (B) Passing-Bablok regression showing strong correlation (r=0.99), with a slope of 0.91 (95% CI 0.88 to 0.93) and intercept of 6.0 (95% CI 0.2 to 15.9) and (C) Bland-Altman plot showing a bias of −89.9 with limits of agreement −1029 to 851. NT-proBNP, N-terminal pro-B-type natriuretic peptide.

Figure 2

Percentage of patients as classified by four-strata risk based on reference and study NT-proBNP results. NT-proBNP, N-terminal pro-B-type natriuretic peptide.

Figure 3

Analysis comparing reference and sendaway day 3 venous NT-proBNP (pg/mL) using (A) percentage difference plot demonstrating a median difference of 0%, (B) Passing-Bablok regression showing strong correlation (r=0.99), with a slope of 1.0 (95% CI 0.9 to 1.0) and intercept of 0.0 (95% CI −5.4 to 6.1) and (C) Bland-Altman plot showing a bias +76.6 with limits of agreement −897.5 to 1051. NT-proBNP, N-terminal pro-B-type natriuretic peptide.

Figure 4

Analysis comparing reference and sendaway day 7 capillary NT-proBNP (pg/mL) using (A) percentage difference plot demonstrating a median difference of −28%, (B) Passing-Bablok regression showing strong correlation (r=0.98), with a slope of 0.75 (95% CI 0.73 to 0.79) and intercept of 0.89 (95% CI −5.2 to 7.09) and (C) Bland-Altman plot showing a bias of −392.4 with limits of agreement −1999 to 1214.

Figure 5

Analysis comparing reference and sendaway day 7 venous NT-proBNP (pg/mL) using (A) percentage difference plot demonstrating a median difference of −18.4%, (B) Passing-Bablok regression showing strong correlation (r=0.98), with a slope of 0.83 (95% CI 0.78 to 0.86) and intercept of 0.24 (95% CI −9.6 to 8.0) and (C) Bland Altman plot showing a bias of −227 with limits of agreement −1488 to 1034. NT-proBNP, N-terminal pro-B-type natriuretic peptide.