OBJECTIVE: To establish if agreement exists between central venous pressure (CVP) and pulmonary capillary wedge pressure (PCWP) measurements in severe hypertension in pregnancy as analyzed by tests of bias, precision, and 95% limits of agreement. METHODS: In a prospective study, CVP and PCWP data in 30 patients were collected by means of a pulmonary artery catheter from initiation of therapy until delivery. Patients with a diastolic blood pressure of more than 110 mm Hg were included. Correlation and agreement between CVP and PCWP before and after treatment were evaluated. RESULTS: The correlation coefficient (r) for CVP-PCWP data in 30 untreated patients was r = 0.64 (p = 0.0002) and for 256 pairs of posttreatment data, it was r = 0.53 (p < 0.0001). Linear regression and correlation for each individual patient in 29 patients with more than 3 measurements showed a significant correlation (p < 0.05) in 19 patients (66%). Correlation was poor (p > 0.05) in 10 patients (34%). The mean difference between PCWP and CVP was 3.5 +/- 2.6 mm Hg (limits of agreement: -1.6 to 8.7) in untreated patients. The mean difference between PCWP and CVP for 256 pairs of data derived posttreatment was 4.9 +/- 3.8 mm Hg (limits of agreement: -2.7 to 12. 5). CONCLUSION: Invasive measurements of CVP and PCWP were found to agree poorly. Until a reliable noninvasive method is available to measure left ventricular preload, PCWP is the measurement of choice when invasive hemodynamic monitoring is necessary in patients with severe preeclampsia.
OBJECTIVE: To establish if agreement exists between central venous pressure (CVP) and pulmonary capillary wedge pressure (PCWP) measurements in severe hypertension in pregnancy as analyzed by tests of bias, precision, and 95% limits of agreement. METHODS: In a prospective study, CVP and PCWP data in 30 patients were collected by means of a pulmonary artery catheter from initiation of therapy until delivery. Patients with a diastolic blood pressure of more than 110 mm Hg were included. Correlation and agreement between CVP and PCWP before and after treatment were evaluated. RESULTS: The correlation coefficient (r) for CVP-PCWP data in 30 untreated patients was r = 0.64 (p = 0.0002) and for 256 pairs of posttreatment data, it was r = 0.53 (p < 0.0001). Linear regression and correlation for each individual patient in 29 patients with more than 3 measurements showed a significant correlation (p < 0.05) in 19 patients (66%). Correlation was poor (p > 0.05) in 10 patients (34%). The mean difference between PCWP and CVP was 3.5 +/- 2.6 mm Hg (limits of agreement: -1.6 to 8.7) in untreated patients. The mean difference between PCWP and CVP for 256 pairs of data derived posttreatment was 4.9 +/- 3.8 mm Hg (limits of agreement: -2.7 to 12. 5). CONCLUSION: Invasive measurements of CVP and PCWP were found to agree poorly. Until a reliable noninvasive method is available to measure left ventricular preload, PCWP is the measurement of choice when invasive hemodynamic monitoring is necessary in patients with severe preeclampsia.