Comparison of intra-arterial and manual auscultation of blood pressure during submaximal exercise in humans

Abstract

Blood pressure (BP) is a key measure of cardiovascular function, and accurate measurement is important to ensure proper clinical evaluation, diagnosis, and treatment. We compared intra-arterial (direct) and cuff auscultation (manual) measurement techniques at rest and during 2 levels of submaximal constant-load exercise (9 min at 40% and 75% maximum watts). Sixty-four adults (aged 29.0 ± 0.7 years; 48% male; height, 173.7 ± 1.2 cm; mass, 73.0 ± 1.7 kg; body mass index, 24.1 ± 0.4 kg·m⁻²; body surface area, 1.87 ± 0.03 m²) participated in the study. At rest, low, and moderate intensity, direct measurement demonstrated higher systolic BP (SBP) and diastolic BP (DBP) (bias for SBP: 22, 31, and 27 mm Hg and for DBP: 5, 7, and 17 mm Hg; rest, low-, and moderate-intensity, respectively; p < 0.01). At rest, the correlation and agreement between the 2 methods was modest (SBP: r = 0.56, bias = +22.1 mm Hg; DBP: r = 0.53, bias = +4.9 mm Hg; p < 0.001). There was good correlation and agreement with SBP at low and moderate intensity; however, DBP demonstrated a weaker relationship (SBP: r = 0.74 and 0.74, bias = +30.2 and +26.8 mm Hg; DBP: r = 0.39 and 0.28, bias = +7.1 and +13.4 mm Hg; for low and moderate intensity, respectively; p < 0.001). Further, manual measurement demonstrated a greater slope from rest to moderate exercise for the relationship between pulse pressure (PP) and cardiac output (13.6 ± 0.4 vs 12.3 ± 0.4, p = 0.03). As exercise intensity increases, manual DBP tends to bias low compared with direct DBP, which, when combined with parallel increases in SBP, leads to no differences in PP between methods at moderate exercise. Because PP is used to calculate other cardiovascular parameters (mean arterial pressure, systemic vascular resistance), measurement technique and exercise intensity should be considered when using cardiovascular variables as outcome measures.

Introduction

Accurate measurement of blood pressure (BP) at rest and during exercise is critical for accurate evaluation, data collection, diagnosis, and patient safety in the clinical and research settings. In the clinical setting, BP is commonly used for diagnostic purposes and thus it is essential that the most accurate and reliable method be used (Rasmussen et al. 1985).

Despite standardization to minimize interobserver variability, manual cuff auscultation (manual) BP is known to be susceptible to large degrees of intraobserver variability. Disparities between manual and intra-arterial assessment (direct) of BP become further exaggerated for both maximal static and dynamic modes of exercise (Gould et al. 1985; Sagiv et al. 1999; White et al. 1990). Although intra-arterial BP assessment is considered the gold standard and is highly recommended for research purposes, its invasiveness, cost, and risk makes this method less practical, especially in the clinical setting. To date, much of the research examining the accuracy of indirect (manual) BP measurement has been performed in unhealthy patients (hypertensive, heart failure) at rest and during maximal exercise. The purpose of this study was to compare the measurement of BP in healthy adults by manual auscultation of the brachial artery using a sphygmomanometer, which remains the most common method of measuring BP, with direct intra-arterial measurement of BP through the radial artery at rest and during constant-load submaximal exercise. We hypothesized that the manual method for BP measurement would provide a similar trend from rest to constant-load exercise when compared with direct measurement and, moreover, that the agreement between manual and direct measurement would be closer at low-intensity exercise, but the disparity between methods would be exaggerated with higher-intensity exercise, even during constant load.

Materials and methods

Participant recruitment

Sixty-four young healthy male (n = 31) and female (n = 33) participants were recruited to participate in this study. The protocol was reviewed and approved by the Mayo Clinic Institutional Review Board, all participants gave written informed consent prior to participation, and all aspects of the study conformed to the Declaration of Helsinki and Health Insurance Portability and Accountability Act (HIPAA) guidelines. All participants were recruited from the surrounding community and were healthy non-smokers with no evidence of cardiovascular, pulmonary, or musculoskeletal disease and were not pregnant or currently taking any medications (except for birth control).

Overview of protocol

The data set created by this study, where the objective was to explore the cardiovascular consequences of exercise according to β-2 adrenergic receptor genotype, has been published elsewhere (Snyder et al. 2006). The data were grouped according to β-2 receptor genotypes to investigate differences in cardiovascular function at rest and during exercise; as such, the cardiovascular parameters in this previous manuscript were presented under stratification, but never as a whole, as presented in this manuscript.

During visit 1, the participants reported to the Mayo Clinic − Clinical Research Unit (CRU) to provide a blood sample for a complete blood count (to rule out anemia) and, for women, to undergo a pregnancy test. After results were reported, the participants underwent baseline pulmonary function testing according to the American Thoracic Society (ATS) standards (ATS 1995) (Elite Series Plethysmography, Medical Graphics, St. Paul, Minn., USA) and an incremental exercise test on an upright electronically braked cycle ergometry to volitional fatigue (Corival, Lode Medical Technology, Netherlands). As with visits 2 and 3 (see below), prior to exercise testing, the participants were fitted with a nose clip and a standard mouthpiece attached to a PreVent Pneumotach (Medical Graphics), which were worn throughout the testing procedure. For maximal exercise testing, all participants were encouraged verbally to continue the exercise protocol to maximal exertion, identified as a rating of perceived exertion ≥17 (Borg scale = 6−20), a respiratory exchange ratio of ≥1.10, and (or) an inability to continue pedaling. This baseline exercise study served as an initial familiarization session, was used to determine work intensities for subsequent sessions, and acted as a screening study to rule out myocardial ischemia, arrhythmias, and pulmonary abnormalities. The participants returned to the CRU for subsequent exercise testing at visits 2 and 3.

Visit 2 consisted of an incremental exercise test, similar to that in visit 1, but with the additional measurement of cardiac output ($\dot{Q}$) using a previously validated open-circuit acetylene uptake method during each stage of exercise (Johnson et al. 2000). This session served as further familiarization with the measurements to be made on the final study day and allowed for confirmation of workloads used during visit 3.

Data collection

For visit 3, the participants were instrumented with a 5-cm 20-gauge Teflon-coated indwelling catheter (FA-04020, Arrow International Inc., Reading, Pa., USA) placed in the radial artery with the strongest palpable pulse after local anesthesia with 2% lidocaine. This catheter was used for the direct measurement of arterial BP using a SpaceLab 512D patient monitor (SpaceLabs Inc., Hillsboro, Ore., USA). Analog signals from the patient monitor were recorded on a beat-by-beat basis on a digital oscilloscope (ADInstrument PowerLab 16/30 CV) for later off-line analysis. Both the SpaceLab 512D patient monitor and the PowerLab digital oscilloscope were calibrated to manufacturers’ specifications. On the opposite arm, the participants were fitted with a standard arm cuff (Critikon DURA-CUF, GE Healthcare, Finland) suited to the circumference of their arm, and a standard stethoscope (Littman Classic II SE, 3M, St. Paul, Minn., USA) was used for manual measurement of BP. During manual measurement, the technician listened for the first and fifth Korotkoff sounds to determine the SBP and DBP, respectively. The same technician conducted all manual measurements for all participants throughout the study. After 30 min of quiet rest, baseline measurements of $\dot{Q}$, heart rate (HR), stroke volume (SV), and BP were made 10 min, 5 min, and 1 min prior to the start of exercise and were averaged to gain a single resting measurement (Fig. 1). The participants then exercised for 9 min at ~40% (Low) and 9 min at ~75% (Mod) of their peak workload achieved during the initial exercise visits while measurements were repeated every 2−3 min. Nine minutes of exercise were performed because pilot data suggested that this was an adequate time frame for obtaining 3 sets of measures and it brought the participants close to exhaustion during the higher workload.

Fig. 1.

Timeline of measurements at rest and during exercise.

Measurement of BP

The intra-arterial BP was recorded continuously while manual BP was taken at 3 specific time points (every 2−3 min) during each stage of the protocol. Off-line analysis of the intra-arterial pressure recording allowed us to match the timing of the manual BP measurements with the intra-arterial pressure recordings. The intra-arterial BP measurements were averaged for 5 beats before and after the specific time point of when the manual BP measurement was taken. All 3 measurements (for both intra-arterial and manual) were then averaged for each stage of exercise (as reported in Table 1).

Table 1.

Cardiovascular measures at rest and during steady-state exercise.

	Rest	Low intensity	Moderate intensity
$\dot{V}{\text{O}}_2$ (mL·kg−1·min−1)	3.75±0.10	18.18±0.43†	31.76±0.85†
${\dot{Q}}_{\text{I}}$ (L·min−1·m−2)	3.3±0.1	5.7±0.1†	7.9±0.2†
SVI (mL·beats−1·m−2)	41.8±2.0	43.3±1.1*	43.6±1.1
HR (beats·min−1)	81.0±1.5	133.1±1.8†	182.7±1.3†
SOT (mL·min−1·m−2)	624.7±23.0	1150.1±40.0†	1619.9±37.0†
PaO2 (mm Hg)	100.7±1.39	97.70±1.25	98.56±1.3
CaO2 (mL·100 mL−1)	19.14±0.23	19.86±0.32	20.44±0.33
Epinephrine	95.8±6.7	100.9±6.70	318.0±40.9†
Norepinephrine	297.9±10.8	630.1±28.9†	2252.8±151.9†

Note: Data are presented as means ± SE. $\dot{V}{\text{O}}_2$, volume of oxygen consumption; ${\dot{Q}}_{\text{I}}$, cardiac index; SV_I, stroke volume index; HR, heart rate; SOT, systemic oxygen transport; PaO₂, partial pressure of arterial oxygen; CaO₂, arterial content of oxygen.

^*p < 0.05.

^†p < 0.01 vs rest.

Measurement of ventilation and gas exchange

Oxygen uptake ($\dot{V}{\text{O}}_2$) and carbon dioxide production were measured continuously during all exercise tests and stages using a metabolic measurement system (MedGraphics CPX/D, Medical Graphics) interfaced with a mass spectrometer (MGA 1100, Marquette Electronics, Milwaukee, Wis., USA). Manual volume calibration was performed with a 3-L syringe and gas calibration was performed with manufacturer-recommended gases of known concentration. All calibration procedures were conducted immediately prior to each testing protocol. This system has been validated against classic “Douglas bag” collection techniques, and stability is verified laboratory personnel by regular testing at standard exercise intensities by (Proctor and Beck 1996).

Assessment of cardiovascular function

$\dot{Q}$ was assessed using an 8- to 10-breath open-circuit acetylene wash-in technique as described previously (Johnson et al. 2000). Briefly, the pneumotachograph was connected to a non-rebreathing Y valve (Hans Rudolph, Kansas City, Mo., USA) with the inspiratory port connected to a pneumatic switching valve (Hans Rudolph); this allowed for rapid switching from room air to the test gas mixture (a large reservoir containing 0.7% C₂H₂; 21% O₂; 9% He; and balance, N₂). Gases were sampled using the mass spectrometer, which was integrated with custom analysis software for the assessment of $\dot{Q}$ using our previously described iterative technique, which has been validated against the direct Fick method (Johnson et al. 2000). HR was measured continuously by standard 12-lead electrocardiogram (GE Case, GE Medical Systems, Milwaukee, Wis., USA). SV was calculated by dividing $\dot{Q}$ by HR. Both $\dot{Q}$ and SV were standardized to body surface area (BSA) to obtain the $\dot{Q}$ and SV indexes (${\dot{Q}}_{\text{I}}$ and SV_I, respectively).

For both the intra-arterial and the manual method, mean arterial pressure (MAP) was calculated using the equation MAP = DBP × 1/3(SBP − DBP), where DBP is diastolic BP and SBP is systolic BP, and the difference between the 2 is pulse pressure (PP). Systemic vascular resistance (SVR) was calculated as the systemic MAP divided by $\dot{Q}$ and multiplied by 80 (conversion from Woods units). SVR was then divided by BSA to obtain the SVR index (SVR_I). The intra-arterial catheter also provided access for the collection of arterial blood for the measurement of arterial catecholamines (see below for catecholamines) as well as arterial blood gases and hemoglobin content (partial pressure of arterial oxygen (PaO₂) and oxygen saturation (SaO₂)). Arterial oxygen content (CaO₂) was calculated using the following equation:

$$\text{Ca}{\text{O}}_2=\left(1.34\times \text{Hgb}\times \text{Sa}{\text{O}}_2\right)+\left(\text{Pa}{\text{O}}_2\times 0.0031\right)$$

Systemic oxygen transport (SOT) was calculated as CaO₂ multiplied by $\dot{Q}$.

Measurement of catecholamines

Epinephrine (EPI) and norepinephrine (NE) were assessed according to methods developed in the Mayo Clinic CRU immunochemical core laboratory and the methods of Sealey (Sealey 1991) For EPI, our laboratory intra-assay coefficients of variation (CVs) are 12.2% and 3.6% at 13.8 and 242 pg·mL⁻¹. Interassay CVs are 8.5% and 6.3% at 179 and 390 pg·mL⁻¹.

Statistical analyses

Statistical analyses and graphic presentation were accomplished using JMP (version 8.0, Cary, N.C., USA) and GraphPad Prizm (version 4.0, San Diego, Calif., USA). Paired Student’s t tests were used to determine differences between the 2 measurement techniques for the average of all measures made during each specific stage of the protocol (i.e., rest, low intensity, and moderate intensity). Analysis of variance (ANOVA) with Tukey’s HSD post hoc analysis was used to determine differences at rest and during exercise for ${\dot{Q}}_{\text{I}}$, HR, SV_I, SBP, DBP, MAP, and SVR_I between direct and manual measurement techniques for all measurements made throughout all stages of the protocol (Fig. 2). All data were found to have homogeneity of variance prior to the ANOVA using the Levene’s test for equality of variance. Pearson correlation coefficients were used to determine the relationship between measurement techniques and Bland–Altman plots were constructed to assess agreement between the measurement techniques (Bland and Altman 1986). Statistical significance was set at an α level of 0.05 for all analyses. Data are presented as means ± SE unless otherwise indicated.

Fig. 2.

(a–e) Hemodynamic variables at rest and the response to low- and moderate-intensity exercise (40% and 75% maximal workload). The dashed line represents the manual measurement and the solid line represents the direct measurement. *, p < 0.05; †, p < 0.01.

Results

Participant characteristics

There were 64 participants (48% male) with a mean age of 29.0 ± 0.7 years (range, 18−32 years). The average height was 173.7 ± 1.2 cm, with an average mass of 73.0 ± 1.7 kg. The mean body mass index of the group was 24.1 ± 0.4 kg·m⁻². The participants had a mean BSA of 1.87 ± 0.03 m².

Cardiovascular and catecholamine response to exercise

Table 2 highlights the cardiovascular and catecholamine response from rest to exercise. As expected, $\dot{V}{\text{O}}_2,{\dot{Q}}_{\text{I}}$, SV_I, HR, and SOT all increased from rest to low-intensity to moderate-intensity exercise. In addition, both EPI and NE increased from rest to low-intensity and moderate-intensity exercise.

Table 2.

Blood pressure measures at rest and during steady-state exercise.

	Average	Manual	Direct	p
Rest
SBP (mm Hg)	123.6±1.6	112.6±1.3	134.6±2.0	<0.01
DBP (mm Hg)	77.0±1.3	74.4±1.1	79.6±1.5	<0.01
MAP (mm Hg)	91.0±1.2	85.9±1.0	96.1±1.5	<0.01
SVRI ((dyn·s)·cm−5·m−2)	2301.8±77.1	2170.7±70.4	2432.9±83.9	<0.01
PP (mm Hg)	46.5±1.3	38.1±1.0	54.9±1.7	<0.01
Low intensity
SBP (mm Hg)	159.5±3.0	144.1±2.4	174.9±2.7	<0.01
DBP (mm Hg)	74.9±1.4	71.3±1.4	78.5±1.4	<0.01
MAP (mm Hg)	100.3±1.6	93.2±1.3	107.4±1.8	<0.01
SVRI ((dyn·s)·cm−5·m−2)	1413.1±27.6	1314.6±25.0	1511.7±30.2	<0.01
PP (mm Hg)	84.6±2.7	72.8±2.5	96.4±2.9	<0.01
Moderate intensity
SBP (mm Hg)	183.1±3.6	169.7±3.1	196.4±4.2	<0.01
DBP (mm Hg)	73.4±1.9	65.1±2.4	81.8±1.4	<0.01
MAP (mm Hg)	106.3±1.9	96.5±1.8	116.2±1.9	<0.01
SVRI ((dyn·s)·cm−5·m−2)	1102.7±24.8	1004.5±25.1	1200.8±24.6	<0.01
PP (mm Hg)	109.7±4.0	104.6±4.3	114.7±3.7	<0.01

Note: Data are presented as means ± SE. p value = Student’s t test of manual vs direct. SBP, systolic blood pressure; DBP, diastolic blood pressure; MAP, mean arterial pressure; SVR_I, systemic vascular resistance index; PP, pulse pressure.

Difference between manual and direct measurement of BP at rest

As shown in Table 1 and Fig. 2, where Table 1 presents the average of the values at each of the 3 time points for each intensity as depicted in Fig. 2, at rest the manual and direct measurement techniques demonstrated a significant difference of 22 mm Hg (p < 0.01) for SBP being higher when using the direct measurement technique. DBP was also elevated using the direct technique, with a difference of 5 mm Hg (p < 0.01). These differences resulted in higher MAP, SVR_I, and PP with the direct methodology. The difference in MAP was 10 mm Hg (p < 0.01), SVR_I showed a difference of 262 (dyn·s)·cm⁻⁵·m⁻² (p < 0.01), and PP had a difference of 17 mm Hg (p < 0.01). Figs. 3a and 3b demonstrate the correlation between the manual and direct methods for measurement of SBP (Fig. 3a) and DBP (Fig. 3b). Despite the consistent underestimation by the manual technique for both the SBP and DBP, these figures show a moderately good correlation for SBP (r = 0.56, p < 0.001) and DBP (r = 0.53, p < 0.001). Similarly, despite the significant differences between the 2 measurement techniques, when using the Bland–Altman plots (Figs. 4a and 4b) to examine the agreeability between the 2 measurement techniques, there appears to be good agreement for both the SBP (Fig. 4a) and DBP (Fig. 4b), with a bias of +22.1 mm Hg for SBP and +4.9 mm Hg for DBP.

Fig. 3.

(a–f) Relationship between direct and manual blood pressure (BP) measurement techniques at rest and during low- and moderate-intensity exercise (40% and 75% maximal workload). The diagonal line is the true line of identity representing what would be expected if the measurement techniques were providing a 1:1 relationship.

Fig. 4.

(a–f) Assessment of agreement and bias between the direct and manual measurement techniques to measure blood pressure (BP) at rest and during low- and moderate-intensity exercise (40% and 75% maximal workload). The dashed lines above and below the solid line, identifying the mean, represent 2 SDs from the mean (similar to 95% confidence interval of 1.96 SD).

Difference between manual and direct measurement of BP at low intensity

Table 1 and Fig. 2 also show the hemodynamic variables at low-intensity exercise. SBP remained higher when using the direct method, with a mean difference of 31 mm Hg (p < 0.01); DBP was also higher with the direct method, demonstrating a mean difference of 7 mm Hg (p < 0.01). Again, because of these differences, the MAP, SVR_I, and PP were also significantly elevated using the direct method compared with the manual technique. These differences included a mean difference of 14 mm Hg (p < 0.01) for MAP, 137 (dyn·s)·cm⁻⁵·m⁻² (p < 0.01) for SVR_I, and 24 mm Hg (p < 0.01) for PP. Figs. 3c and 3d demonstrate a strong correlation between the 2 methods for the measurement of SBP (Fig. 3c) at low-intensity exercise (r = 0.74, p < 0.001), similar to the resting data; however, in contrast and despite the significant correlation, the DBP (Fig. 3d) demonstrates only a moderate correlation between the 2 methods (r = 0.39, p < 0.001). When examining the Bland–Altman plots (Figs. 4c and 4d), it appears that at low-intensity exercise, the relatively good agreement between the 2 techniques for SBP is preserved (Fig. 4c), whereas the DBP (Fig. 4d) demonstrates good agreement at the lower pressures with a trend to bias toward a lower pressure using the manual method at the higher overall pressures. The level of bias between the systems was +30.2 mm Hg for SBP and +7.1 mm Hg for DBP.

Difference between manual and direct measurement of BP at moderate intensity

Also seen in Table 1 and Fig. 2 are the comparisons between manual and direct measurement techniques at moderate intensity. Similar to the rest and low-intensity data, the SBP, DBP, and mean arterial BP were all significantly elevated when measured using the direct method compared with the manual technique. These differences included 27 mm Hg (p < 0.01) for SBP and 17 mm Hg (p < 0.01) for DBP. Again, these differences led to increased MAP (mean difference of 20 mm Hg, p < 0.01), SVR_I (mean difference of 196 (dyn·s)·cm⁻⁵·m⁻², p < 0.01), and PP (mean difference of 10 mm Hg, p < 0.01). Again, similar to rest and the low-intensity exercise level, Figs. 3e and 3f demonstrate a strong correlation between the 2 measurement techniques for SBP (Fig. 3e) (r = 0.74, p < 0.001), whereas the DBP (Fig. 3f) again demonstrates a moderate but significant correlation (r = 0.28, p < 0.001). The Bland–Altman plots (Figs. 4e and 4f) during moderate-intensity exercise show that at higher pressures, the manual method begins to demonstrate greater systematic error despite a lower mean bias when compared with low-intensity exercise (+26.8 mm Hg). This is in contrast to the DBP, which shows a consistent underestimation of pressure across the average pressure range for the group. This is in agreement with the nearly double mean bias at this higher intensity when compared with the low intensity (+13.4 mm Hg).

Finally, with increasing constant-load exercise intensity, manual measurement demonstrated a greater slope for the relationship between PP and $\dot{Q}$ (13.6 ± 0.4 vs 12.3 ± 0.4, p = 0.03). These changes resulted in a narrowing of the difference in PP between techniques during moderate intensity.

Discussion

The purpose of this study was to examine the differences between 2 commonly used methods for the assessment of BP, at rest and during 2 levels of submaximal constant-load exercise. Our results demonstrate the following: (i) Manual auscultation consistently underestimates BP when compared with direct intra-arterial measurement at rest and during submaximal exercise. This is consistent with previous findings in that the differences between methods for SBP and DBP we observed were similar to or greater than others (Sagiv et al. 1999; Gould et al. 1985; White et al. 1990). Similar to the results presented by Sagiv et al. (1999), during maximal exercise, the correlation between manual and direct measurement of BP was high at rest for both SBP and DBP, but was significantly weakened for DBP at peak exercise (Sagiv et al. 1999). (ii) During moderate-intensity constant-load exercise, DBP measured via manual auscultation deviates downward compared with the intra-arterial measurement, resulting in an upward shift in PP towards the intra-arterial-derived PP. This is an interesting finding considering the importance of PP in the calculation of MAP and SVR.

Clinical exercise testing provides a means for clinicians to challenge the cardiac and other physiologic systems and to potentially identify problems that are not evident while in the resting condition. This early identification of individuals at risk allows for preventative medical strategies to be implemented earlier. Therefore, it is increasingly important that the most accurate and reliable method be used to measure BP both at rest and during exercise.

When examining the manual method, we recognized both benefits and limitations (Kroeker and Wood 1955; Rowell et al. 1968; Van Bergen et al. 1954). The primary benefits in using the manual technique are its lower cost and noninvasive nature, which causes less apprehension on the part of the patient or research participant. Also, by positioning the cuff on the upper arm and measuring the BP at the brachial artery, one can minimize the differences in BP that occur because of location within circulation; at this position there will be less wave reflection and thus less wave summation, leading to BP measurements that more closely reflect central aortic BP (Bruner et al. 1981).

However, benefits often come with limitations, which in the case of manual BP measurement include terminal digit bias (most noticeable during exercise) (Harrison et al. 2008). These limitations were noted in this study; as the intensity of exercise increased, the difference between the DBP measured between methods was magnified, such that no true relationship between the direct and manual DBP was measured. Terminal digit bias is the rounding off of the last digit to the nearest 5 or 10 mm Hg because of the inability to auscultate the fifth Korotkoff sound. We hypothesize that the disparity we observed resulted from the inability of the technician to auscultate the fifth Korotkoff sound, which has been demonstrated previously to be caused by confounding factors including ambient noise, arm motion during exercise, and changing pressure-flow relationships within the artery (Harrison et al. 2008; White et al. 1990). Specifically, our finding of a weakening reliability of manually assessed DBP as exercise intensity increased has clinical relevance because DBP is an important contributor to MAP. Increases in MAP reflect increases in DBP and cardiac afterload, which is an important clinical measure for diagnosing and assessing the severity of serious disorders (heart failure, diabetes, myocardial infarction, pulmonary hypertension, etc.).

In contrast to the manual method of BP measurement, the direct methodology has its own set of benefits and limitations. A primary benefit of the direct method is the significant reduction in human error during measurement, a result of the direct, indwelling catheter placement. This placement also provides the ability to measure BP beat by beat, which provides a much more consistent record of BP changes over time. Because of the beat-by-beat direct intra-arterial measurement, there is no terminal digit bias associated with the values measured. Unfortunately, despite these significant advantages, there are also considerable limitations to direct measurement. One major limitation, apart from cost, is the potential inability of the direct recording system to provide an adequate frequency response, which may ultimately result in an artificially elevated BP response (Wood 1956). However, with recent advancements in biomedical technologies, this frequency response limitation has largely been minimized. The error caused by catheter placement, in contrast, often cannot be avoided (Van Bergen et al. 1954); this error results when the catheter is placed facing directly upstream, leading to falsely elevated BP because of the augmented kinetic energy associated with the perpendicular force of the blood flow on the measurement system (Van Bergen et al. 1954). In some cases, this limitation can be overcome by adjustment to the catheter and adequately securing the catheter to the skin.

Limitations

The cohort used in this study was a young population (aged 18−32 years); as such, the findings on the disparities between direct and manual measurement of BP may not be applicable to an older population. Because aging can increase cardiac muscle and vessel wall thickness, and because manual BP measurement is the primary method used clinically in the elderly population, further research is warranted to determine if our findings hold as age increases. Additionally, these measurements were performed during cycle ergometery only, so the findings may not be applicable to all exercise testing modalities. Another limitation is that we had only 1 technician measuring BP manually, and because of the differences in manual BP assessment among individual technicians, it is likely that the discrepancies between direct and manual will also be different based on the technician performing the manual assessment. The last point that must be made is that the sphygmomanometer and catheter were placed in different arms and different locations, and the placement of the catheter itself can alter blood flow and change pressure.

Conclusion

In summary, there are a number of benefits and limitations to both the manual or direct methods for the measurement of BP. These benefits and limitations must be taken into full consideration when choosing a method that will be most suitable for the setting in which the BP must be acquired. Our results suggest that the manual method tracks a similar pattern of BP when compared with the direct method, from rest to low and moderate submaximal constant-load exercise; however, the manual method systematically underestimates the direct method at all time points. Similarly, when approaching higher workloads, the manually acquired DBP is much lower than that acquired by the direct method, which results in a higher PP, ultimately leading to similar PPs between the methods. This similar PP at higher workloads suggests that either method of BP acquisition may be suitable for the assessment of myocardial work and oxygen consumption; however, MAP and SVR remained different at this intensity.