Original Article
Assessment of Central Blood Pressure in Patients With Type 2 Diabetes: A Comparison Between Sphygmocor and Invasively Measured Values Esben Laugesen,1,2 Niklas B. Rossen,1,3 Christian D. Peters,4 Michael Mæng,5 Eva Ebbehøj,1 Søren T. Knudsen,1 Klavs W. Hansen,3 Hans E. Bøtker,5 and Per L. Poulsen1
methods In 34 patients with type 2 diabetes we estimated ascending aortic SBP and DBP using the SphygmoCor device and compared these data with invasively recorded data. The validity of the transfer function was assessed by calibrating with invasively recorded DBP and mean BP (MBP). The influence of noninvasive calibration strategies was assessed by calibrating with brachial oscillometric SBP+DBP vs. DBP+MBP using a form factor (ff ) of 0.33 and 0.40, respectively. results When calibrating with invasive BP, the difference between estimated and invasively measured ascending aortic SBP and DBP was
−2.3 ± 5.6/1.0 ± 0.9 mm Hg. When calibrating with oscillometric brachial BPs, the differences were −9.6 ± 8.1/14.1 ± 6.2 mm Hg (calibration with SBP and DBP), −8.3 ± 11.7/13.9 ± 6.1 mm Hg (DBP and MBP; ff = 0.33), and 1.9 ± 12.2/14.1 ± 6.2 mm Hg (DBP and MBP; ff = 0.40), respectively. Calibration with the average of 3 brachial BPs did not improve accuracy.
conclusions The SphygmoCor transfer function seems valid in patients with type 2 diabetes. Noninvasive calibration with DBP and MBP (ff = 0.40) enables accurate estimation of mean ascending aortic SBP at the group level. However, the wide limits of agreement indicate limited accuracy in the individual patient. clinical trials registration Clinical Trials No. NCT01538290. Keywords: applanation tonometry; blood pressure; blood pressure monitoring; central blood pressure; hypertension; invasive blood pressure monitoring; transfer function; type 2 diabetes; validation. doi:10.1093/ajh/hpt195
Patients with type 2 diabetes (T2DM) have significantly higher cardiovascular morbidity and mortality than nondiabetics.1,2 Treatment of hypertension in diabetic patients significantly reduces cardiovascular morbidity and mortality,3 although the optimal target brachial blood pressure (BP) is currently debated.4 However, the brachial systolic BP (SBP), which is the classical risk stratification parameter, does not fully reflect the aortic SBP to which the heart, brain and kidneys are exposed because of the phenomenon of pulse pressure amplification.5 As the pulse wave progresses from the conduit arteries to the brachial artery, increasing vessel stiffness and the effect of reflected waves leads to a steeper systolic upstroke and peaked pressure waves to the effect that SBP is increased, whereas diastolic BP (DBP) and mean BP
(MBP) remain fairly constant (i.e., within a few mm Hg).5,6 Thus, from a physiological point of view, risk stratification based on central SBP could be more appropriate. The SphygmoCor device (AtCor Medical, Sydney, Australia) is widely used for noninvasive assessment of central BP. With the SphygmoCor, the central waveform is synthesized from the radial waveform by a generalized transfer function. Studies in mainly nondiabetic patients have demonstrated that calibration with intra-arterial pressures enables precise derivation of the ascending aortic SBP and DBP.7–9 Because the extent and distribution of age-related arterial stiffening in diabetic patients differs from that in nondiabetics,10 it has previously been proposed that a diabetes-specific transfer function might be pertinent.11 However, the necessity of this remains
Correspondence: Esben Laugesen ([email protected]).
1Department of Endocrinology and Internal Medicine, Aarhus University Hospital, Aarhus, Denmark; 2The Danish Diabetes Academy, Odense, Denmark; 3Department of Medicine, Diagnostic Center, Silkeborg Regional Hospital, Silkeborg, Denmark; 4Department of Renal Medicine, Aarhus University Hospital Skejby, Aarhus, Denmark; 5Department of Cardiology, Aarhus University Hospital Skejby, Aarhus, Denmark.
Initially submitted July 11, 2013; date of first revision July 30, 2013; accepted for publication September 17, 2013; online publication December 4, 2013.
© American Journal of Hypertension, Ltd 2013. All rights reserved. For Permissions, please email: [email protected]
American Journal of Hypertension 27(2) February 2014 169
Downloaded from http://ajh.oxfordjournals.org/ at New York University on October 19, 2014
background The SphygmoCor is used for noninvasive assessment of ascending aortic blood pressure (BP). However, the validity of the SphygmoCor transfer function has not been tested in an exclusively type 2 diabetic patient sample. Calibration with systolic (SBP) and diastolic (DBP) brachial BP has previously been associated with substantial imprecision of central BP estimates. We hypothesized that different noninvasive calibration strategies might improve the accuracy of the estimated ascending aortic BPs.
Laugesen et al.
METHODS Study sample
Thirty-four consecutive patients with T2DM referred to elective coronary angiography (CAG) at the Department of Cardiology, Aarhus University Hospital Skejby, Denmark, were included in the study. Inclusion criteria were diagnosis of T2DM and age >18 years with a palpable left radial artery. Exclusion criteria were atrial fibrillation/other cardiac arrhythmias or diagnosis of subclavian or brachial artery stenosis. Power analysis showed that inclusion of 32 patients would confer 80% power to detect a systematic 5 mm Hg difference between estimated and invasively measured BP with a significance level of 0.5 given an SD of the difference of 10 mm Hg. From November 2011 to March 2012, 54 consecutive patients were invited to participate in the study. Two patients declined participation. Eight were not examined because of laboratory time constraints (arrival of acute patients), 6 were excluded because of cardiac arrhythmias developed during the CAG (atrial fibrillation/ frequent ventricular extrasystoles), and in 4 patients data were not obtained because of technical problems. Thus data from 34 patients were available for analysis. The study was approved by the Research Ethics Committee of Central Denmark Region and by the Danish Data Protection Agency. All patients gave their written informed consent. BP data collection
The invasive ascending aortic pressures were obtained with a fluid filled 6F Boston Scientific Expo Angiographic catheter (Boston Scientific, Natick, MA) attached to a NAMIC transducer (Navilyst Medical, Marlborough, MA). The catheters were 100 cm long with an internal diameter of 1.4 mm. Transducers were placed at the midaxillary line and 170 American Journal of Hypertension 27(2) February 2014
calibrated to zero before each examination. Catheters were inserted through a femoral sheath into the ascending aorta and were flushed every 2 minutes. Brachial DBP and SBP were measured at the left brachial artery with an oscillometric Riester Champion N automatic blood pressure monitor (Riester GmbH, Jungingen, Germany) with the patient in the supine position on the operating table. The Riester Champion N is technically identical to the Microlife 3BTO-A blood pressure monitor, which achieved grade A for both systolic and diastolic pressure when validated according to the British Hypertension Society protocol.17 The circumference of the left upper arm was assessed using a measuring tape. When the circumference exceeded 32 cm, the normal cuff size (for circumferences 22–32 cm (bladder size 24 × 13 cm)) was exchanged with a large cuff (for circumferences 32–48 cm (bladder size 35 × 13.5 cm)). Three calibration BPs were recorded for all patients. The radial artery pressure waveform was noninvasively recorded using a high-fidelity Millar tonometer (Millar Instruments, Houston, TX) and SphygmoCor equipment and software version 8.2 (AtCor Medical, Sydney, Australia). Recording time was 10 seconds. Only high quality data was accepted; median operator index was 95. Immediately after this recording, 10 seconds of invasively measured central SBP, DBP, and MBP were recorded and printed. SBP and DBP were determined as the average of the pressure peaks during the recording period, and MAP was determined as area under the curve by the recording equipment software (Philips Xper Physiomonitoring 5; Philips, Amsterdam, the Netherlands) and Siemens Axiom Sensis XP (Siemens, Munich, Germany). This procedure was performed twice; 2 datasets were thus available for all patients. To minimize procedure time, the total procedure was conducted as follows: When the CAG procedure was finished, the catheter was placed in the ascending aorta. After ensuring that the pressure curve was stable, the first calibration BP was measured at the left brachial artery. Next, the radial waveforms were obtained with the applanation tonometer, and immediately after this, the invasive BP from the preceeding 10 seconds was recorded and printed. The same sequence was used for recording of the second calibration BP, radial pulse wave sequence, and invasive BP. Finally, the third calibration BP was recorded. Post-procedure calibration of the radial pressure wave
For the invasive calibration study, the first invasive DBP and MBP averaged over the 10 seconds were used to calibrate the first radial pressure wave recording. Based on the radial waveform, the aortic waveform was then synthesized by the transfer function in the SphygmoCor. The second radial waveform was subsequently calibrated using the second invasive BP data, and the aortic waveform was synthesized. Data from both the first and the second waveforms and the mean of the 2 recordings are reported. For the noninvasive validation study, the radial waveform recordings were each subsequently calibrated using (i) the brachial BP recorded immediately before each tonometry and (ii) the mean of the 3 brachial BPs. The radial waveforms
Downloaded from http://ajh.oxfordjournals.org/ at New York University on October 19, 2014
unresolved, as the validity of the SphygmoCor transfer function has not previously been tested specifically in an exclusively type 2 diabetic patient sample. The SphygmoCor is most often calibrated with noninvasive brachial SBP and DBP. Several studies have shown that this may introduce substantial imprecision in the estimates of central BP.12–15 Recent studies by Weber et al.9 and Pucci et al.16 based on oscillometric BP data indicated that calibration with brachial DBP and MBP could improve precision. However, the majority of oscillometric BP devices do not provide information on MBP. Instead, MBP can be approximated from DBP and SBP and is usually calculated using a form factor (ff) of 0.33 (i.e., as DBP + 0.33(SBP−DBP)) or as recently suggested using an ff of 0.4.19–21 Whether calibration with DBP and an approximated MBP could improve accuracy also remains to be tested by comparison with invasive data. Hence, in this study our aims were, in a sample of patients with T2DM, (i) to test the validity of the SphygmoCor transfer function when calibrated with invasively measured ascending aortic BP and (ii) to test the impact of different noninvasive calibration methods for the accuracy of the synthesized ascending aortic BP.
SphygmoCor Validation in Type 2 Diabetes
were calibrated with (i) DBP and SBP; (ii) DBP and MBP (ff = 0.33), MBP defined as DBP + 0.33(SBP−DBP); and (iii) DBP and MBP (ff = 0.4), MBP defined as DBP + 0.4(SBP− DBP). The following data are reported: (1st invasive − 1st synthesized aortic BP) calibrated with both 1 and the mean of 3 BPs, and ((1st invasive − 1st synthesized aortic BP) + (2nd invasive − 2nd synthesized aortic BP))/2, with each radial wave calibrated with the mean of 3 BPs.
synthesized and invasively measured ascending aortic SBP and DBP were assessed by paired t tests. Agreement between the BPs was assessed by the approach described by Bland and Altman.18 A 2-tailed P
Assessment of Central Blood Pressure in Patients With Type 2 Diabetes: A Comparison Between Sphygmocor and Invasively Measured Values Esben Laugesen,1,2 Niklas B. Rossen,1,3 Christian D. Peters,4 Michael Mæng,5 Eva Ebbehøj,1 Søren T. Knudsen,1 Klavs W. Hansen,3 Hans E. Bøtker,5 and Per L. Poulsen1
methods In 34 patients with type 2 diabetes we estimated ascending aortic SBP and DBP using the SphygmoCor device and compared these data with invasively recorded data. The validity of the transfer function was assessed by calibrating with invasively recorded DBP and mean BP (MBP). The influence of noninvasive calibration strategies was assessed by calibrating with brachial oscillometric SBP+DBP vs. DBP+MBP using a form factor (ff ) of 0.33 and 0.40, respectively. results When calibrating with invasive BP, the difference between estimated and invasively measured ascending aortic SBP and DBP was
−2.3 ± 5.6/1.0 ± 0.9 mm Hg. When calibrating with oscillometric brachial BPs, the differences were −9.6 ± 8.1/14.1 ± 6.2 mm Hg (calibration with SBP and DBP), −8.3 ± 11.7/13.9 ± 6.1 mm Hg (DBP and MBP; ff = 0.33), and 1.9 ± 12.2/14.1 ± 6.2 mm Hg (DBP and MBP; ff = 0.40), respectively. Calibration with the average of 3 brachial BPs did not improve accuracy.
conclusions The SphygmoCor transfer function seems valid in patients with type 2 diabetes. Noninvasive calibration with DBP and MBP (ff = 0.40) enables accurate estimation of mean ascending aortic SBP at the group level. However, the wide limits of agreement indicate limited accuracy in the individual patient. clinical trials registration Clinical Trials No. NCT01538290. Keywords: applanation tonometry; blood pressure; blood pressure monitoring; central blood pressure; hypertension; invasive blood pressure monitoring; transfer function; type 2 diabetes; validation. doi:10.1093/ajh/hpt195
Patients with type 2 diabetes (T2DM) have significantly higher cardiovascular morbidity and mortality than nondiabetics.1,2 Treatment of hypertension in diabetic patients significantly reduces cardiovascular morbidity and mortality,3 although the optimal target brachial blood pressure (BP) is currently debated.4 However, the brachial systolic BP (SBP), which is the classical risk stratification parameter, does not fully reflect the aortic SBP to which the heart, brain and kidneys are exposed because of the phenomenon of pulse pressure amplification.5 As the pulse wave progresses from the conduit arteries to the brachial artery, increasing vessel stiffness and the effect of reflected waves leads to a steeper systolic upstroke and peaked pressure waves to the effect that SBP is increased, whereas diastolic BP (DBP) and mean BP
(MBP) remain fairly constant (i.e., within a few mm Hg).5,6 Thus, from a physiological point of view, risk stratification based on central SBP could be more appropriate. The SphygmoCor device (AtCor Medical, Sydney, Australia) is widely used for noninvasive assessment of central BP. With the SphygmoCor, the central waveform is synthesized from the radial waveform by a generalized transfer function. Studies in mainly nondiabetic patients have demonstrated that calibration with intra-arterial pressures enables precise derivation of the ascending aortic SBP and DBP.7–9 Because the extent and distribution of age-related arterial stiffening in diabetic patients differs from that in nondiabetics,10 it has previously been proposed that a diabetes-specific transfer function might be pertinent.11 However, the necessity of this remains
Correspondence: Esben Laugesen ([email protected]).
1Department of Endocrinology and Internal Medicine, Aarhus University Hospital, Aarhus, Denmark; 2The Danish Diabetes Academy, Odense, Denmark; 3Department of Medicine, Diagnostic Center, Silkeborg Regional Hospital, Silkeborg, Denmark; 4Department of Renal Medicine, Aarhus University Hospital Skejby, Aarhus, Denmark; 5Department of Cardiology, Aarhus University Hospital Skejby, Aarhus, Denmark.
Initially submitted July 11, 2013; date of first revision July 30, 2013; accepted for publication September 17, 2013; online publication December 4, 2013.
© American Journal of Hypertension, Ltd 2013. All rights reserved. For Permissions, please email: [email protected]
American Journal of Hypertension 27(2) February 2014 169
Downloaded from http://ajh.oxfordjournals.org/ at New York University on October 19, 2014
background The SphygmoCor is used for noninvasive assessment of ascending aortic blood pressure (BP). However, the validity of the SphygmoCor transfer function has not been tested in an exclusively type 2 diabetic patient sample. Calibration with systolic (SBP) and diastolic (DBP) brachial BP has previously been associated with substantial imprecision of central BP estimates. We hypothesized that different noninvasive calibration strategies might improve the accuracy of the estimated ascending aortic BPs.
Laugesen et al.
METHODS Study sample
Thirty-four consecutive patients with T2DM referred to elective coronary angiography (CAG) at the Department of Cardiology, Aarhus University Hospital Skejby, Denmark, were included in the study. Inclusion criteria were diagnosis of T2DM and age >18 years with a palpable left radial artery. Exclusion criteria were atrial fibrillation/other cardiac arrhythmias or diagnosis of subclavian or brachial artery stenosis. Power analysis showed that inclusion of 32 patients would confer 80% power to detect a systematic 5 mm Hg difference between estimated and invasively measured BP with a significance level of 0.5 given an SD of the difference of 10 mm Hg. From November 2011 to March 2012, 54 consecutive patients were invited to participate in the study. Two patients declined participation. Eight were not examined because of laboratory time constraints (arrival of acute patients), 6 were excluded because of cardiac arrhythmias developed during the CAG (atrial fibrillation/ frequent ventricular extrasystoles), and in 4 patients data were not obtained because of technical problems. Thus data from 34 patients were available for analysis. The study was approved by the Research Ethics Committee of Central Denmark Region and by the Danish Data Protection Agency. All patients gave their written informed consent. BP data collection
The invasive ascending aortic pressures were obtained with a fluid filled 6F Boston Scientific Expo Angiographic catheter (Boston Scientific, Natick, MA) attached to a NAMIC transducer (Navilyst Medical, Marlborough, MA). The catheters were 100 cm long with an internal diameter of 1.4 mm. Transducers were placed at the midaxillary line and 170 American Journal of Hypertension 27(2) February 2014
calibrated to zero before each examination. Catheters were inserted through a femoral sheath into the ascending aorta and were flushed every 2 minutes. Brachial DBP and SBP were measured at the left brachial artery with an oscillometric Riester Champion N automatic blood pressure monitor (Riester GmbH, Jungingen, Germany) with the patient in the supine position on the operating table. The Riester Champion N is technically identical to the Microlife 3BTO-A blood pressure monitor, which achieved grade A for both systolic and diastolic pressure when validated according to the British Hypertension Society protocol.17 The circumference of the left upper arm was assessed using a measuring tape. When the circumference exceeded 32 cm, the normal cuff size (for circumferences 22–32 cm (bladder size 24 × 13 cm)) was exchanged with a large cuff (for circumferences 32–48 cm (bladder size 35 × 13.5 cm)). Three calibration BPs were recorded for all patients. The radial artery pressure waveform was noninvasively recorded using a high-fidelity Millar tonometer (Millar Instruments, Houston, TX) and SphygmoCor equipment and software version 8.2 (AtCor Medical, Sydney, Australia). Recording time was 10 seconds. Only high quality data was accepted; median operator index was 95. Immediately after this recording, 10 seconds of invasively measured central SBP, DBP, and MBP were recorded and printed. SBP and DBP were determined as the average of the pressure peaks during the recording period, and MAP was determined as area under the curve by the recording equipment software (Philips Xper Physiomonitoring 5; Philips, Amsterdam, the Netherlands) and Siemens Axiom Sensis XP (Siemens, Munich, Germany). This procedure was performed twice; 2 datasets were thus available for all patients. To minimize procedure time, the total procedure was conducted as follows: When the CAG procedure was finished, the catheter was placed in the ascending aorta. After ensuring that the pressure curve was stable, the first calibration BP was measured at the left brachial artery. Next, the radial waveforms were obtained with the applanation tonometer, and immediately after this, the invasive BP from the preceeding 10 seconds was recorded and printed. The same sequence was used for recording of the second calibration BP, radial pulse wave sequence, and invasive BP. Finally, the third calibration BP was recorded. Post-procedure calibration of the radial pressure wave
For the invasive calibration study, the first invasive DBP and MBP averaged over the 10 seconds were used to calibrate the first radial pressure wave recording. Based on the radial waveform, the aortic waveform was then synthesized by the transfer function in the SphygmoCor. The second radial waveform was subsequently calibrated using the second invasive BP data, and the aortic waveform was synthesized. Data from both the first and the second waveforms and the mean of the 2 recordings are reported. For the noninvasive validation study, the radial waveform recordings were each subsequently calibrated using (i) the brachial BP recorded immediately before each tonometry and (ii) the mean of the 3 brachial BPs. The radial waveforms
Downloaded from http://ajh.oxfordjournals.org/ at New York University on October 19, 2014
unresolved, as the validity of the SphygmoCor transfer function has not previously been tested specifically in an exclusively type 2 diabetic patient sample. The SphygmoCor is most often calibrated with noninvasive brachial SBP and DBP. Several studies have shown that this may introduce substantial imprecision in the estimates of central BP.12–15 Recent studies by Weber et al.9 and Pucci et al.16 based on oscillometric BP data indicated that calibration with brachial DBP and MBP could improve precision. However, the majority of oscillometric BP devices do not provide information on MBP. Instead, MBP can be approximated from DBP and SBP and is usually calculated using a form factor (ff) of 0.33 (i.e., as DBP + 0.33(SBP−DBP)) or as recently suggested using an ff of 0.4.19–21 Whether calibration with DBP and an approximated MBP could improve accuracy also remains to be tested by comparison with invasive data. Hence, in this study our aims were, in a sample of patients with T2DM, (i) to test the validity of the SphygmoCor transfer function when calibrated with invasively measured ascending aortic BP and (ii) to test the impact of different noninvasive calibration methods for the accuracy of the synthesized ascending aortic BP.
SphygmoCor Validation in Type 2 Diabetes
were calibrated with (i) DBP and SBP; (ii) DBP and MBP (ff = 0.33), MBP defined as DBP + 0.33(SBP−DBP); and (iii) DBP and MBP (ff = 0.4), MBP defined as DBP + 0.4(SBP− DBP). The following data are reported: (1st invasive − 1st synthesized aortic BP) calibrated with both 1 and the mean of 3 BPs, and ((1st invasive − 1st synthesized aortic BP) + (2nd invasive − 2nd synthesized aortic BP))/2, with each radial wave calibrated with the mean of 3 BPs.
synthesized and invasively measured ascending aortic SBP and DBP were assessed by paired t tests. Agreement between the BPs was assessed by the approach described by Bland and Altman.18 A 2-tailed P
