Reproducibility of the neurocardiovascular responses to common laboratory-based sympathoexcitatory stimuli in young adults

Abstract

The magnitude of blood pressure (BP) and muscle sympathetic nerve activity (MSNA) responses to laboratory stressors is commonly used to compare neurocardiovascular responsiveness between groups and conditions. However, no studies have rigorously examined the reproducibility of BP and MSNA responsiveness. Here, we assess the within-visit reproducibility of BP (finger photoplethysmography) and MSNA (microneurography) responses to isometric handgrip (HG) and postexercise ischemia (PEI) in young healthy adults (n = 30). In a subset (n = 21), we also examined the between-visit reproducibility of responsiveness to HG, PEI, and the cold pressor test (CPT). Intraclass correlation coefficients (ICCs) were used as a primary reproducibility measure (e.g., ICC >0.75 is considered very good). Within a visit, the increase in mean arterial pressure during HG [ICC = 0.85 (0.69–0.93); P < 0.001] and PEI [ICC = 0.85 (0.69–0.93); P < 0.001] demonstrated very good reproducibility. Furthermore, the between-visit reproducibility of the pressor response to HG [ICC = 0.85 (0.62–0.94); P < 0.001], PEI [ICC = 0.84 (CI = 0.58–0.94); P < 0.001], and the CPT [ICC = 0.89 (0.72–0.95) P < 0.001]) were also very good. However, there was greater variability in both the within- [HG: ICC = 0.58 (−0.22–0.85), P = 0.001; PEI: ICC = 0.33 (−0.24–0.69), P = 0.042] and between-visit reproducibility of MSNA responsiveness [HG: ICC = 0.87 (0.53–0.96), P = 0.001; PEI: ICC = 0.24 (−0.62–0.78), P = 0.27; CPT: ICC = 0.77 (0.29–0.93), P = 0.007]. The magnitude of the BP response to several standard laboratory stimuli was very good, whereas the variability of the MSNA response to these perturbations was generally less consistent, particularly during PEI. These data provide novel insight for both study design and data interpretation when comparing neurocardiovascular responsiveness between different conditions, groups, or studies, as well as before and after interventions/treatments.

NEW & NOTEWORTHY The magnitude of the increases in blood pressure and muscle sympathetic nerve activity in response to sympathoexcitatory stimuli such as static handgrip, postexercise ischemia, and the cold pressor test are commonly used to assess neurocardiovascular responsiveness. However, limited studies have comprehensively examined the reproducibility of these responses. We demonstrate that the reproducibility of the pressor response to these perturbations was very good within an individual, whereas the reproducibility of the MSNA response was less consistent.

INTRODUCTION

Assessing blood pressure (BP) and muscle sympathetic nerve activity (MSNA) in response to acute laboratory stressors yields important insight regarding neural cardiovascular control beyond which can be gleaned under basal conditions alone (6). The magnitude of the pressor and sympathetic responses to such perturbations is commonly used to 1) identify conditions associated with increased cardiovascular risk, 2) compare responsiveness between distinct groups, 3) examine mechanisms of aberrant regulation in disease, and 4) evaluate the efficacy of intervention strategies targeting neurocardiovascular reactivity (10, 13, 22, 26, 27, 34). Isometric handgrip (HG) and the cold pressor test (CPT) are classic sympathoexcitatory stimuli regularly used to evoke concurrent increases in BP and sympathetic outflow (30, 45). Longitudinal studies indicate that abnormally large increases in BP during HG or the CPT increase the risk of a cardiovascular event and are predictive of the development of future cardiovascular disease (2, 43). However, despite the widespread utilization of these laboratory stressors and the compelling evidence that MSNA at rest is reproducible over time (41), the within-person reproducibility of the BP and MSNA responsiveness to these stressors remains unclear. This is an important gap in knowledge, particularly given the National Institutes of Health’s renewed calls for rigor and reproducibility in biomedical research (8, 20).

There is evidence demonstrating substantial interindividual heterogeneity in the magnitude of the BP and MSNA responses to moderate-intensity static HG and isolated muscle metaboreflex activation using postexercise ischemia (PEI) (i.e., “responders” and “nonresponders”) (11, 13, 25). However, limited studies have assessed the intraindividual reproducibility of BP responses, and in those that have, reproducibility has been an ancillary outcome (4, 12, 14, 15, 17, 38). To our knowledge, only one study has prospectively examined the reproducibility of the pressor response to sympathoexcitatory stimuli (33). In this study, BP was measured in eight men during isometric HG performed at 40% of maximal voluntary contraction (MVC) and the CPT; each participant repeated six distinct trials of each stimulus within the experimental visit. The variability of BP responsiveness, assessed as a coefficient of variation, was ∼22.2% for HG and ∼17.2% for the CPT (33). Based on this, the authors concluded that the reproducibility of the BP response to acute sympathoexcitatory stimuli is minimal. However, in addition to the limited feasibility and applicability of performing six trials of each perturbation within a single study visit, this investigation also did not include an assessment of the reproducibility of these responses between experimental visits. Moreover, to our knowledge, limited studies have attempted to examine the reproducibility of the MSNA response to these perturbations.

With this background in mind, the primary purpose of the present investigation was to examine the within-visit reproducibility of the BP and MSNA responses to moderate-intensity isometric HG (30% MVC) and isolated metaboreflex activation during PEI in young healthy adults. In addition, we also examined the between-visit reproducibility of responsiveness to these stimuli and a CPT in a subset of the participants. The results of this study will help to define the consistency of BP and MSNA responsiveness to acute sympathoexcitatory stimuli. These data will have important implications for the assessment and interpretation of neurocardiovascular regulation when designing studies and particularly when making comparisons between groups or between conditions across experimental visits.

METHODS

Participants

The Institutional Review Boards at The Pennsylvania State University and The University of Texas at Arlington approved all experimental procedures. Verbal and written consents were obtained voluntarily from all the subjects before participation and in accordance with the guidelines set forth by the Declaration of Helsinki. Thirty young healthy adults (7 women) participated (age: 23 ± 1 yr, body mass index: 25 ± 1 kg/m², resting BP: 119 ± 2/76 ± 1 mmHg). All the participants completed a medical health history and the International Physical Activity Questionnaire (28). Participants were nonobese (body mass index <30 kg/m²), did not use tobacco products, and were not taking any prescription medications. All women were taking hormonal contraceptives and were tested during the low-hormone phase. Women on contraceptives that did not include a low-hormone phase were excluded from participation. A urine pregnancy test confirmed the absence of pregnancy before experimental visits.

Experimental Measurements

Beat-to-beat arterial BP was measured using finger photoplethysmography (NOVA, Finapres Medical Systems, Amsterdam, The Netherlands), obtained from the nondominant hand positioned at heart level. Automated brachial artery BP (Connex Spot Monitor, Welch Allyn, Skaneateles Falls, NY) was measured intermittently during rest periods throughout the protocol and was used to verify absolute Finometer BP measurements. Heart rate (HR) was measured via single-lead electrocardiogram (NOVA). Respiratory movements were monitored using a strain gauge pneumograph placed over the abdomen (Pneumotrace II, UFI, Morro Bay, CA) to ensure that the subjects did not inadvertently perform breath-holds or Valsalva maneuvers during the protocols.

Multiunit postganglionic MSNA was recorded using standard microneurographic techniques, as previously described (18, 19, 44). Briefly, a tungsten microelectrode was inserted into muscle nerve fascicles of the peroneal nerve near the fibular head; a reference electrode was inserted 2–3 cm away. The contralateral leg was used for microneurographic recordings in the participants who completed a second experimental visit within 30 days. Because there is some evidence for seasonal variation in MSNA at rest and in response to sympathoexcitatory perturbations (9), each individual who completed multiple experimental visits was tested within the same season (i.e., within 1 mo). Neural signals were amplified, bandpass filtered (700–2,000 Hz), rectified, and integrated (absolute value; 0.1 s time constant decay) to obtain mean voltage neurograms (n = 20, Neuro Amp EX, ADInstruments, Colorado Springs, CO; n = 10, Nerve Traffic Analyzer 662C-4, University of Iowa Bioengineering, Iowa City, IA). MSNA recordings were identified by the presence of spontaneously occurring bursts with characteristic pulse synchronicity, responsiveness to an end-expiratory breath hold or Valsalva maneuver, and lack of response to arousal stimuli or skin stimulation (44). We decided a priori to proceed with protocols only if MSNA recordings were technically excellent, defined by a signal-to-noise ratio greater than 3:1 and stable baselines for at least 10 min (13).

Experimental Protocol

Before each experimental session, the participants were instructed to abstain from caffeine for 12 h and alcoholic beverages and strenuous physical activity for 24 h. Upon arrival to the laboratory for each visit, the subjects were positioned supine in a hospital bed. The MVC of the dominant hand was determined by squeezing a HG device (ADInstruments) at maximal effort three to four times. The highest force production was used as the MVC, which was subsequently used to calculate the relative work rate of 30% for the HG experimental protocol. During HG, the subjects were provided with visual feedback of force production. After determination of the relative work rate, a familiarization HG trial was conducted to allow the subjects to practice maintaining the target force of 30% MVC (∼60 s) and to experience the rapidity of cuff inflation for PEI (∼20 s).

After subject instrumentation and a satisfactory MSNA recording was obtained, the subjects rested quietly in the supine position for 10 min. Thereafter, two HG + PEI trials were completed; each trial was separated by at least 15 min to allow BP, HR, and MSNA to return to baseline values. Two min of baseline immediately preceded each HG trial. Participants then performed isometric HG at 30% MVC for 2 min. With 5 s remaining in HG, an occlusion cuff placed on the upper arm of the exercising limb was rapidly inflated to suprasystolic BP (≥220 mmHg). The occlusion cuff remained inflated for 3 min and 15 s following the completion of exercise (i.e., PEI). PEI was used to isolate activation of the skeletal muscle metaboreflex from central command and the muscle mechanoreflex (1, 19, 29). The additional 15 s of PEI was incorporated to account for the robust and transient initial decrease in BP and MSNA that occurs when exercise is terminated. Ratings of perceived exertion were obtained at the end of exercise using the standard 6–20 Borg scale (3), to examine whether the same peak level of voluntary effort was attained during each trial. Ratings of perceived pain were obtained at the end of PEI using a modified 0–15 Borg scale (3). All the subjects were able to complete the full 2 min of HG during all trials. Nerve recordings during both HG trials were maintained in 22 (3 women) out of 30 participants on the first visit. In a subset of subjects, the CPT was also performed, as described in detail below.

To assess the reproducibility of BP and MSNA responsiveness between experimental visits, a subset of participants (n = 21, 7 women) completed a second study visit (visits were separated by at least 1 wk). In addition to the two HG + PEI trials described above, these individuals also completed a CPT. Following 2 min of baseline, a laboratory team member passively moved the participant’s hand into a bucket containing an ice slurry for 2 min. Ratings of perceived pain were obtained at the end of CPT using a modified 0–15 Borg scale (3). Nerve recordings were maintained in 13 participants (3 women) during HG + PEI and the CPT during both visits.

Data and Statistical Analysis

Data were recorded at 1,000 Hz (PowerLab and LabChart; ADInstruments). The MSNA signal was calibrated by assigning the voltage of the average of the three largest bursts during the baseline, preceding each perturbation the value of 100 arbitrary units (AU), and all other bursts within the trial were normalized with respect to this value. MSNA was analyzed using a custom-designed LabVIEW program (16, 18), which generated synchronized beat-by-beat data of all recorded variables gated by the R-wave of the electrocardiogram. All neurograms were also evaluated manually to confirm assignment of MSNA bursts from the automated analysis and adjustments were made when necessary. For completeness, MSNA was quantified using standard measures, including burst frequency (bursts/min), burst incidence (bursts/100 heartbeats), and total activity (AU/min). However, burst frequency was used as the primary outcome. For HG and PEI trials, BP, HR, and MSNA were calculated as mean values over an initial 1 min baseline, then during the last 30 s (BP and HR) or the last min (MSNA) of HG and during the full 3 min of PEI. For the CPT, all variables were analyzed during a 1 min baseline and during the last 30 s (BP and HR) or the last min (MSNA) of the CPT. One minute segments were used for the analysis of MSNA, as this is more standard and shorter sampling durations demonstrate poorer reliability (31).

Paired t tests were used to analyze group mean data during each perturbation (IBM SPSS Statistics v25). Because the degree of reproducibility can be somewhat subjective, we provide multiple indices of reproducibility in an effort to be comprehensive and transparent. Intraclass correlation coefficients (ICCs; two-way mixed effects, average measures, and absolute agreement) were used as the primary index of the reproducibility of the magnitude of responsiveness to each perturbation, consistent with the approach of recent studies (21, 31). Descriptive characterization of the relative quality of reproducibility was based on a range of ICCs, i.e., poor (<0.39), moderate/fair (0.40–0.59), good (0.60–0.74), very good (0.75–0.89), and excellent (0.90–1.00); this modified descriptive scale is based on that originally put forth by Cicchetti (7). Data are also presented in Bland–Altman plots (along with the calculated upper and lower limits, bias, and correlation coefficients), which present the difference in responsiveness between trials/visits within an individual against the mean value (36). In addition, we calculated the typical error of the measurement (TEM; i.e., within-person standard error of the measurement) (23) and the coefficient of variation (CV) as additional indices of reproducibility. The CV for the HR responses during PEI was not calculated, because the mean of the responses approached zero, resulting in inconclusive results.

Between-visit reproducibility of responsiveness to HG and PEI was calculated for each participant for both trials (i.e., visit 1 trial 1 versus visit 2 trial 1 and, separately, visit 1 trial 2 versus visit 2 trial 2). The results were qualitatively similar for these comparisons, and, therefore, only data for trial 1 are presented (visit 1 trial 1 versus visit 2 trial 1). All data were statistically evaluated in the full data set and, separately, in only men. Because the results and interpretation based on the data in men are nearly identical to those obtained from analyses including both sexes, these secondary analyses are not presented.

Group data are reported as means ± standard deviation. All data were normally distributed (D’Agostino–Pearson test and Shapiro–Wilk test). Reproducibility indices are also reported as the mean and 95% confidence intervals (CIs), and TEM is reported as both an absolute value and a percentage. Significance was set at P < 0.05.

RESULTS

BP Reactivity to HG and PEI

Baseline.

There were no differences in baseline mean arterial pressure (MAP) between trials on either visit 1 [trial 1: 86 ± 6 mmHg versus trial 2: 85 ± 5 mmHg, P = 0.34; ICC = 0.83 (0.64–0.92), P < 0.001; TEM = 3.7 (2.9–5.0; 4.3%); CV = 3.6% (2.9%–4.8%)] or visit 2 [trial 1: 85 ± 6 mmHg versus trial 2: 86 ± 6 mmHg, P = 0.41; ICC = 0.91 (0.79–0.95), P < 0.001; TEM = 2.4 (1.8–3.5; 2.9%); CV = 2.8% (2.1%–4.0%)]. Between visits, there were no differences in baseline MAP [P = 0.26; ICC = 0.57 (−0.41 to 0.82), P = 0.033; TEM = 4.9 (3.7–7.1; 5.8%); CV = 7.0% (5.4%–10.1%)]. Mean data and reproducibility indices for systolic and diastolic BP and HR during baseline are presented in Table 1.

Table 1.

Cardiovascular and sympathetic variables during baseline

	SBP, mmHg	DBP, mmHg	HR, beats/min	Burst Incidence, bursts/100 beats	Total Activity, AU/min
Visit 1, Trial 1	118 ± 9	70 ± 7	58 ± 6	22 ± 10	863 ± 353
Visit 1, Trial 2	119 ± 9	68 ± 6	59 ± 7	21 ± 9	853 ± 292
Paired t test	0.64	0.11	0.87	0.83	0.86
ICC (95% CI)	0.85 (0.68–0.93)†	0.83 (0.64–0.92)†	0.86 (0.70–0.93)†	0.86 (0.64–0.94)†	0.85 (0.62–0.94)†
TEM (%; 95% CI)	5.3 (4.5; 4.2–7.1)	4.0 (5.8; 3.2–5.4)	3.9 (29.8; 3.1–5.2)	6.5 (29.8; 5.0–9.4)	232.7 (27.1; 178.0–335.9)
CV (95% CI)	3.7 (2.9–5.0)	4.8 (3.8–6.5)	5.6 (4.5–7.5)	22.1 (16.9–31.9)	19.7 (15.1–28.4)
Bias (95% CI)	0.6 ± 6.3 (−11.8–12.9)	−1.5 ± 4.7 (−10.6–7.7)	0.1 ± 4.6 (−8.9–9.2)	−0.3 ± 6.8 (−13.6–13.0)	−9.6 ± 239 (−477–458)
r (95% CI)	−0.02 (−0.4–0.4)	−0.1 (−0.4–0.3)	0.1 (−0.2–0.5)	−0.2 (−0.6–0.2)	−0.3 (−0.6–0.2)
Visit 2, Trial 1	117 ± 3	69 ± 6	56 ± 6	25 ± 11	879 ± 107
Visit 2, Trial 2	118 ± 2	69 ± 6	58 ± 7	21 ± 12	863 ± 125
Paired t test	0.29	0.61	0.11	0.24	0.84
ICC (95% CI)	0.92 (0.80–0.97)†	0.92 (0.80–0.97)†	0.86 (0.66–0.94)†	0.78 (0.33–0.93)†	0.89 (0.62–0.97)†
TEM (%; 95% CI)	3.5 (3.0; 2.7–5.1)	2.4 (3.5; 1.8–3.5)	3.2 (5.6; 2.4–4.6)	7.5 (32.9; 5.3–12.8)	163.4 (18.8; 115.8–278.0)
CV (95% CI)	2.9 (2.2–4.2)	3.5 (2.7–5.1)	2.8 (2.1–4.0)	5.3 (3.8–9.0)	24.1 (17.1–41.0)
Bias (95% CI)	1.1 ± 4.8 (−8.3–10.6)	0.4 ± 3.4 (−6.3–7.0)	1.5 ± 4.3 (−6.8–9.9)	−3.4 ± 9.7 (−22.4–15.7)	−16.1 ± 276 (−557–524)
r (95% CI)	−0.2 (−0.6–0.2)	−0.2 (−0.6–0.3)	0.3 (−0.2–0.6)	0.1 (−0.4–0.6)	0.2 (−0.4–0.7)
Between Visits Trial 1
Paired t test	0.77	0.11	0.12	0.61	0.93
ICC (95% CI)	0.51 (−0.25–0.80)†	0.65 (0.18–0.86)†	0.72 (0.34–0.89)†	0.27 (−1.87–0.80)	0.44 (−1.46–0.86)
TEM (%; 95% CI)	6.5 (5.8; 5.0–9.4)	4.7 (6.4; 3.6–6.8)	4.0 (8.1; 3.1–5.8)	9.5 (46.1; 6.8–15.7)	279.1 (31.3; 200.3–460.9)
CV (95% CI)	6.1 (4.7–8.8)	6.3 (4.8–9.1)	6.8 (5.2–9.8)	45.2 (32.4–74.6)	40.8 (29.3–67.4)
Bias (95% CI)	−0.7 ± 10.2 (−20.7–19.3)	−2.2 ± 6.2 (−14.4–9.9)	−2.1 ± 5.7 (−13.2–9.1)	2.3 ± 15.4 (−27.9–32.5)	−13.5 ± 516 (−1,024–997)
r (95% CI)	0.1 (−0.3–0.5)	0.1 (−0.4–0.5)	−0.2 (−0.6–0.3)	−0.1 (−0.7–0.5)	−0.03 (−0.6–0.6)

Values are mean ± SD.

^†ICC P < 0.05.

SBP, systolic blood pressure; DBP, diastolic blood pressure; HR, heart rate; AU, arbitrary units; ICC, intraclass correlation coefficient [poor (0–0.39), moderate (0.40–0.59), good (0.60–0.74), very good (0.75–0.89), excellent (0.90–1.00)]; CI, confidence interval; TEM, typical error of the measurement; CV, coefficient of variation; r, Pearson’s correlation coefficient for Bland–Altman plot.

HG.

The group mean increase in MAP during HG was not different between trials on either visit 1 (trial 1: ∆22 ± 9 mmHg versus trial 2: ∆24 ± 8 mmHg; P = 0.21) or visit 2 (trial 1: ∆22 ± 12 mmHg versus trial 2: ∆22 ± 8 mmHg; P = 0.98). Furthermore, the HG-induced increase in MAP demonstrated very good within-visit reproducibility [Fig. 1, A and B; visit 1: ICC = 0.85 (0.69–0.93), P < 0.001; TEM = 5.5 (4.4–7.4; 23.8%); CV = 19.3% (15.4%–25.9%); visit 2: ICC = 0.79 (0.47–0.93), P < 0.001; TEM = 6.2 (4.7–8.9; 28.0%); CV = 28.1% (21.5–40.6%)]. Likewise, the between-visit reproducibility of the MAP response to HG was also very good [Fig. 1C; ICC = 0.85 (0.62–0.94), P < 0.001; TEM = 5.8 (4.4–8.4; 26.3%); CV = 26.3% (20.1%–38.0%)]. Mean data and reproducibility indices for additional cardiovascular outcome variables during HG are presented in Table 2.

Fig. 1.

Bland–Altman plots for the within-visit reproducibility (n = 30; 7 women) of the increase in mean arterial pressure (MAP) during static handgrip (HG; A and B) and postexercise ischemia (PEI; D and E), as well as the between-visit reproducibility (n = 21; 7 women) of the MAP response to HG (C) and PEI (F). The mean bias of the plot is represented by the solid line (with standard deviation and Pearson correlation). The thick dashed lines represent the 95% confidence interval of the bias. The dashed line at y = 0 represents perfect reproducibility. The mean and 95% confidence intervals are presented; *P < 0.05. ICC, intraclass correlation coefficient.

Table 2.

Cardiovascular and sympathetic responses to HG

	∆SBP, mmHg	∆DBP, mmHg	∆HR, beats/min	∆Burst Incidence, bursts/100 heartbeats	∆Total Activity, AU/min
Visit 1, Trial 1	27 ± 13	20 ± 8	17 ± 8	16 ± 8	1572 ± 972
Visit 1, Trial 2	29 ± 13	23 ± 8	21 ± 8.0*	26 ± 10*	2411 ± 1375*
Paired t test	0.35	0.10	<0.001	<0.001	0.002
ICC (95% CI)	0.81 (0.61–0.91)†	0.74 (0.46–0.87)†	0.80 (0.41–0.92)†	0.16 (−0.35–0.57)	0.65 (0.04–0.87)†
TEM (%; 95% CI)	8.8 (31.7; 7.0–11.8)	6.3 (29.1; 5.0–8.5)	5.8 (30.9; 4.6–7.8)	16.8 (80.2; 12.8–24.2)	1428.1 (70.3; 1092.1–2061.3)
CV (95% CI)	25.9 (20.6–34.8)	23.1 (18.4–31.1)	20.7 (16.5–27.8)	25.9 (19.8–37.4)	23.1 (17.7–33.3)
Bias (95% CI)	1.8 ± 10.1 (−18.1–21.7)	2.2 ± 7.0 (−11.6–15.9)	4.0 ± 5.5 (−6.9–12.8)	10.5 ± 11.4 (−11.9–32.9)	839 ± 1,071 (−1,259–2,938)
r (95% CI)	−0.03 (−0.4–0.3)	−0.04 (−0.4–0.3)	0.07 (−0.3–0.4)	0.2 (−0.2–0.6)	0.4 (−0.03–0.7)
Visit 2, Trial 1	25 ± 13	22 ± 11	15 ± 9	17 ± 12	1664 ± 1285
Visit 2, Trial 2	24 ± 10	22 ± 7	18 ± 10*	24 ± 13*	1996 ± 915
Paired t test	0.75	0.98	0.04	0.04	0.24
ICC (95% CI)	0.67 (0.16–0.87)†	0.84 (0.59–0.93)†	0.89 (0.70–0.96)†	0.71 (0.10–0.91)†	0.81 (0.40–0.94)†
TEM (%; 95% CI)	8.4 (34.2; 6.4–12.1)	5.2 (23.9; 4.0–7.5)	4.6 (27.0; 3.5–6.6)	9.3 (44.8; 6.6–15.8)	663.3 (36.5; 470.1–1128.5)
CV (95% CI)	34.1 (26.1–49.2)	23.9 (18.3–34.5)	24.2 (18.5–34.9)	37.0 (26.2–63.0)	34.4 (24.4–58.5)
Bias (95% CI)	−0.9 ± 11.9 (−24.2–22.5)	0.04 ± 7.4 (−14.4–14.5)	2.8 ± 5.8 (−8.5–14.1)	7.1 ± 10.8 (−14.1–28.3)	303 ± 883 (−1,429–2,034)
r (95% CI)	−0.3 (−0.6–0.2)	−0.6 (−0.8–0.2)	0.2 (−0.2–0.6)	0.1 (−0.5–0.6)	−0.4 (−0.8–0.1)
Between visits Trial 1
Paired t test	0.51	0.45	0.88	0.53	0.99
ICC (95% CI)	0.84 (0.61–0.94)†	0.79 (0.49–0.92)†	0.84 (0.59–0.93)†	0.66 (−0.19–0.90)†	0.95 (0.80–0.99)†
TEM (%; 95% CI)	7.1 (27.4; 5.4–10.2)	6.0 (28.5; 4.6–8.7)	4.6 (29.7; 3.5–6.6)	6.3 (43.5; 4.5–10.4)	364.8 (24.6; 261.8–602.4)
CV (95% CI)	27.6 (21.1–39.8)	28.1 (21.5–40.6)	29.7 (22.7–42.9)	44.6 (32.0–73.7)	27.0 (19.4–44.6)
Bias (95% CI)	−1.5 ± 10.1 (−21.3–18.3)	1.42 ± 8.4 (−14.9–17.8)	0.23 ± 6.5 (−12.4–12.9)	1.71 ± 9.2 (−16.3–19.7)	2.3 ± 565 (−1,105–1,110)
r (95% CI)	−0.1 (−0.5–0.4)	0.4 (−0.1–0.7)	0.2 (−0.3–0.6)	0.4 (−0.2–0.8)	0.3 (−0.3–0.8)

Values are means ± SD.

^*P < 0.05 vs. trial 1;

^†ICC P < 0.05.

AU, arbitrary units; CI, confidence interval; CV, coefficient of variation; DBP, diastolic blood pressure; HR, heart rate; ICC, intraclass correlation coefficient [poor (0–0.39), moderate (0.40–0.59), good (0.60–0.74), very good (0.75–0.89), excellent (0.90–1.00)]; TEM, typical error of the measurement; r, Pearson’s correlation coefficient for Bland–Altman plot; SBP, systolic blood pressure.

PEI.

The group mean increase in MAP during PEI was not different between trials at either visit 1 (trial 1: ∆15 ± 8 mmHg versus trial 2: ∆17 ± 7 mmHg; P = 0.09) or visit 2 (trial 1: ∆15 ± 9 mmHg versus trial 2: ∆14 ± 6 mmHg; P = 0.55). More importantly, the within-visit reproducibility ranged from good to very good [Fig. 1, D–E; visit 1: ICC = 0.85 (0.69–0.93), P < 0.001; TEM = 4.8 (3.8–6.5; 29.9%); CV = 23.5% (18.7%–31.6%); visit 2: ICC = 0.62 (0.04–0.85), P = 0.02; TEM = 5.6 (4.3–8.1; 38.7%); CV = 39.4% (30.1%–56.9%)], and the between-visit reproducibility was very good [Fig. 1F; ICC = 0.84 (0.58–0.94), P < 0.001; TEM = 4.5 (3.4–6.5; 30.5%); CV = 31.3% (23.9–45.2%)]. Mean data and reproducibility indices for additional cardiovascular outcome variables are presented in Table 3.

Table 3.

Cardiovascular and sympathetic responses to PEI

	∆SBP, mmHg	∆DBP, mmHg	∆HR, beats/min	∆Burst Incidence, bursts/100 heartbeats	∆Total Activity, AU/min
Visit 1, Trial 1	20 ± 12	13 ± 7	1 ± 6	23 ± 12	1,483 ± 751
Visit 1, Trial 2	22 ± 10	14 ± 6	1 ± 6	25 ± 11	1,864 ± 886*
Paired t test	0.14	0.16	0.78	0.17	0.02
ICC (95% CI)	0.89 (0.76–0.95)†	0.89 (0.77–0.95)†	0.70 (0.37–0.86)†	0.87 (0.69–0.95)†	0.77 (0.39–0.91)†
TEM (%; 95% CI)	6.0 (28.9; 4.8–8.1)	3.5 (25.0; 2.8–4.7)	5.0 (502.3; 4.0–6.7)	7.8 (32.9; 6.0–11.3)	672.3 (41.0; 514.1–970.4)
CV (95% CI)	22.6 (18.0–30.4)	19.9 (15.8–26.7)	***	22.6 (17.3–32.6)	28.9 (22.1–41.7)
Bias (95% CI)	1.9 ± 6.8 (−11.5–15.3)	1.05 ± 4.0 (−6.7–8.8)	0.4 ± 5.8 (−11.0–11.8)	2.4 ± 7.6 (−12.6–17.3)	381 ± 645 (−884–1,645)
r (95% CI)	−0.2 (−0.5–0.2)	−0.3 (−0.6–0.1)	−0.1 (−0.4–0.3)	−0.1 (−0.5–0.4)	0.2 (−0.2–0.6)
Visit 2, Trial 1	19 ± 10	13 ± 8	2 ± 5	10 ± 21	686 ± 916
Visit 2, Trial 2	17 ± 9	13 ± 6	2 ± 6	16 ± 10	762 ± 455
Paired t test	0.163	0.59	0.80	0.34	0.77
ICC (95% CI)	0.48 (−0.36–0.80)	0.72 (0.28–0.89)†	0.76 (0.38–0.91)†	0.05 (−2.13–0.71)	0.39 (−1.22–0.82)
TEM (%; 95% CI)	8.1 (4.2; 6.2–11.7)	4.5 (35.2; 3.4–6.5)	3.6 (160.6; 2.8–5.2)	16.5 (131.3; 11.7–28.1)	637.0 (88.0; 451.5–1083.8)
CV (95% CI)	34.4 (26.3–49.7)	35.8 (27.4–51.7)	***	126.2 (89.4–214.7)	87.6 (62.1–149.0)
Bias (95% CI)	−1.3 ± 11.7 (−24.1–21.6)	−0.8 ± 6.6 (−13.6–12.0)	−0.3 ± 5.2 (−10.4–9.8)	6.2 ± 22.5 (−27.9–50.2)	75.9 ± 897 (−1683–1835)
r (95% CI)	−0.1 (−0.5–0.3)	−0.4 (−0.7–0.01)	0.3 (−0.2–0.6)	−0.6 (−0.9–0.1)	−0.6 (−0.9–0.1)
Between visits Trial 1
Paired t test	0.84	0.57	0.16	0.13	0.09
ICC (95% CI)	0.79 (0.47–0.92)†	0.82 (0.55–0.93)†	0.54 (−0.11–0.81)†	0.37 (−0.78–0.81)	0.31 (−0.82–0.79)
TEM (%; 95% CI)	6.5 (34.9; 5.0–9.4)	4.1 (31.8; 3.1–5.9)	5.0 (407.5; 3.8–7.1)	16.3 (117.3; 11.7–26.9)	454.7 (49.3; 326.3–750.9)
CV (95% CI)	35.9 (27.5–51.8)	32.4 (24.8–46.8)	***	109.9 (78.9–181.5)	84.1 (60.4–138.9)
Bias (95% CI)	−0.4 ± 9.5 (−19.1–18.2)	0.8 ± 5.9 (−10.8–12.4)	2.3 ± 6.9 (−11.3–15.8)	−10.2 ± 21.6 (−52.6–32.2)	−626 ± 1,096 (−2,774–1,522)
r (95% CI)	−0.2 (−0.6–0.3)	0.2 (−0.3–0.6)	−0.3 (−0.7–0.1)	0.5 (−0.1–0.8)	0.3 (−0.2–0.7)

Values are means ± SD.

^*P < 0.05 vs. trial 1;

^†ICC P < 0.05.

AU, arbitrary units; CI, confidence interval; CV, coefficient of variation; DBP, diastolic blood pressure; HR, heart rate; ICC, intraclass correlation coefficient [poor (0–0.39), moderate (0.40–0.59), good (0.60–0.74), very good (0.75–0.89), excellent (0.90–1.00)]; PEI, postexercise ischemia; r, Pearson’s correlation coefficient for Bland–Altman plot; SBP, systolic blood pressure; TEM, typical error of the measurement. ***, not calculated.

MSNA Reactivity to HG and PEI

Baseline.

There were no differences in baseline burst frequency between trials within either visits [visit 1: trial 1: 13 ± 6 mmHg versus trial 2: 13 ± 5 mmHg, P = 0.83; ICC = 0.78 (0.48–0.91), P = 0.001; TEM = 4.7 (3.6–6.8; 36.4%); CV = 20.5% (15.7%–29.6%); visit 2: trial 1: 14 ± 7 mmHg versus trial 2: 13 ± 7 mmHg, P = 0.31; ICC = 0.89 (0.65–0.97), P < 0.001; TEM = 3.3 (2.4–5.4; 25.3%); CV = 24% (17.2%–39.6%)]. There also were no differences in baseline burst frequency between visits (P = 0.69). However, there was slightly more variability in the reproducibility of between-visit baseline values [ICC = 0.50 (−0.48 to 0.86), P = 0.043; TEM = 6.0 (4.3–9.9; 44.3%); CV = 34.9% (25.0%–57.6%)]. Mean data and reproducibility indices for additional metrics of MSNA during baseline are presented in Table 1.

HG.

The increase in burst frequency during HG was greater during the second trial for both visit 1 (trial 1: ∆16 ± 7 versus trial 2: ∆24 ± 9 bursts/min; P < 0.001) and visit 2 (trial 1: ∆ 17 ± 11 versus trial 2: ∆ 22 ± 9 bursts/min; P = 0.04). As such, the reproducibility of the increase in burst frequency during HG was less than for MAP [Fig. 2, A and B; visit 1: ICC = 0.58 (−0.22 to 0.85), P = 0.001; TEM = 10.2 (7.8–14.7; 50.4%); CV = 24.8% (19.0–35.8%); visit 2: ICC = 0.85 (0.45–0.96), P < 0.001; TEM = 5.7 (4.0–9.7; 29.3%); CV = 24.2% (17.2%–41.2%)]. Interestingly, the degree to which the MSNA response to HG is reproducible appears to be independent of the quality of the nerve recording. Original recordings of MSNA during HG from four subjects are presented in Fig. 3. Notably, even some individuals in whom high-quality nerve recordings were obtained demonstrated highly variable MSNA responsiveness (e.g., Subjects 3 and 4). Between visits, there was very good reproducibility [Fig. 2C; ICC = 0.87 (0.53–0.96), P = 0.001; TEM = 4.6 (3.3–7.6; 29.4%); CV = 30.4% (21.8–50.2%)], with no differences in the mean increase in burst frequency during HG (P = 0.81). Qualitatively similar results were obtained for the reproducibility of burst incidence and total activity during HG (Table 2).

Fig. 2.

Bland–Altman plots for the within-visit reproducibility (n = 22; 3 women) of the increase in muscle sympathetic nerve activity burst frequency during handgrip (A and B), as well as the between-visit reproducibility (n = 13; 3 women; C). The mean bias of the plot is represented by the solid line (with standard deviation and Pearson correlation). The thick dashed lines represent the 95% confidence interval of the bias. The dashed line at y = 0 represents perfect reproducibility. The mean and 95% confidence intervals are presented; *P < 0.05. HG, handgrip; ICC, intraclass correlation coefficient.

Fig. 3.

Original recordings of muscle sympathetic nerve activity (MSNA) during the last 30 s of static handgrip (HG) during a single experimental visit from four individuals. The relative degree of reproducibility of MSNA appears to be independent of the quality of the microneurographic signal. That is, even some individuals in whom lesser-quality nerve recordings were obtained demonstrated adequate reproducibility (Subject 2). Conversely, some individuals with high-quality nerve recordings demonstrated highly variable MSNA responsiveness to HG (Subjects 3 and 4).

PEI.

The group mean increase in burst frequency during PEI was not statistically different between trials on either visit 1 (trial 1: ∆13 ± 6 mmHg versus trial 2: ∆16 ± 5 mmHg; P = 0.07) or visit 2 (trial 1: ∆8 ± 5 mmHg versus trial 2: ∆9 ± 4 mmHg; P = 0.85). However, similar to HG, there was greater variability of the MSNA response, with the within-visit reproducibility ranging from poor to moderate [Fig. 4, A and B; visit 1: ICC = 0.33 (−0.24–0.69), P = 0.042; TEM = 12.0 (9.2–17.3; 60.6%); CV = 29.0% (22.2%–41.9%); visit 2: ICC = 0.43 (−1.24–0.84), P = 0.20; TEM = 3.5 (2.4–6.2; 41.6%); CV = 43.4% (0.3%–76.7%)]. Similarly, the between-visit reproducibility was generally poor [Fig. 4C; ICC = 0.24 (−0.62–0.78), P = 0.27; TEM = 4.6 (3.3–7.6; 48.3%); CV = 47.5% (34.1%–78.4%)]. Original recordings of MSNA during PEI from two subjects are presented in Fig. 5 and highlight the relative lack of consistency in MSNA responsiveness in some individuals, despite high-quality MSNA recordings (e.g., Subject 2). Mean data and reproducibility indices for additional metrics for MSNA are presented in Table 3. Interestingly, during visit 1, the reproducibility of both the mean burst incidence and total activity response to PEI was very good, despite relatively large individual variability (Table 3). However, similar to burst frequency, the within-visit reproducibility for visit 2, as well as the between-visit reproducibility, was generally poor (Table 3).

Fig. 4.

Bland–Altman plots for the within-visit reproducibility (n = 22; 3 women) of the increase in muscle sympathetic nerve activity burst frequency during postexercise ischemia (PEI; A and B), as well as the between-visit reproducibility (n = 13; 3 women; C). The mean bias of the plot is represented by the solid line (with standard deviation and Pearson correlation). The thick dashed lines represent the 95% confidence interval of the bias. The dashed line at y = 0 represents perfect reproducibility. The mean and 95% confidence intervals are presented; *P < 0.05. ICC, intraclass correlation coefficient.

Fig. 5.

Original recordings of muscle sympathetic nerve activity (MSNA) during 30 s of postexercise ischemia (PEI) during a single experimental visit from two individuals. The relative degree of reproducibility of MSNA appears to be independent of the quality of the microneurographic signal, because even some individuals in whom very high-quality nerve recordings were obtained demonstrated high variable MSNA responsiveness to PEI (Subject 2).

BP and MSNA Reactivity to the CPT

There were no differences in MAP during baseline of the CPT (visit 1: 85 ± 7 mmHg versus visit 2: 87 ± 7 mmHg; P = 0.15), and the reproducibility of baseline MAP between visits was very good [ICC = 0.86 (0.65–0.94), P < 0.001; TEM = 3.6 (2.8–5.2; 4.2%); CV = 4.1% (3.1%–5.9%)]. There were no differences in the group mean increase in MAP in response to the CPT (visit 1: ∆21 ± 12 mmHg versus visit 2: ∆22 ± 13 mmHg; P = 0.57), and the between-visit reproducibility of the MAP response was very good [Fig. 6A; ICC = 0.89 (0.72–0.95), P < 0.001; TEM = 5.7 (4.4–8.2; 27.1%); CV = 26.9% (20.6%–38.8%)].

Fig. 6.

Bland–Altman plots for the between-visit reproducibility of the increase in mean arterial pressure (MAP; A; n = 30; 7 women) and muscle sympathetic nerve activity burst frequency (B; n = 13; 3 women) during the cold pressor test. The mean bias of the plot is represented by the solid line (with standard deviation and Pearson correlation). The thick dashed lines represent the 95% confidence interval of the bias. The dashed line at y = 0 represents perfect reproducibility. The mean and 95% confidence intervals are presented; *P < 0.05. ICC, intraclass correlation coefficient.

There were no differences in MSNA burst frequency during baseline of the CPT (visit 1: 14 ± 6 versus visit 2: 12 ± 7 bursts/min; P = 0.46), and the reproducibility of baseline MSNA between visits was very good [ICC = 0.77 (0.27–0.93), P = 0.009; TEM = 4.3 (3.1–7.1; 33.5%); CV = 32.7% (23.5%–54.0%)]. There were no differences in the group mean increase in burst frequency during the CPT between visits (visit 1: ∆23 ± 9 versus visit 2: ∆21 ± 10; P = 0.24). Furthermore, the within-individual increase in burst frequency in response to the CPT demonstrated very good reproducibility [Fig. 6B; ICC = 0.77 (0.29–0.93), P = 0.007; TEM = 6.1 (4.4–10.1; 28.0%); CV = 26.3% (18.9%–43.4%)]; however, the reproducibility of both burst incidence and total activity during the CPT was more variable (Table 4).

Table 4.

Cardiovascular and sympathetic variables during baseline and in response to the CPT

	SBP, mmHg	DBP, mmHg	HR, beats/min	Burst Incidence, busts/100 beats	Total Activity, AU/min
Baseline
Visit 1	119 ± 10	69 ± 7	58 ± 7	23 ± 10	844 ± 190
Visit 2	120 ± 10	71 ± 7*	59 ± 7	21 ± 12	716 ± 355
Paired t test	0.62	0.03	0.32	0.38	0.19
ICC (95% CI)	0.73 (0.34–0.89) †	0.88 (0.68–0.95) †	0.82 (0.57–0.93) †	0.75 (0.21–0.92) †	0.59 (−0.33–0.89)
TEM (%; 95% CI)	6.7 (5.6; 5.1–9.7)	3.3 (4.7; 2.5–4.8)	3.7 (6.4; 2.8–5.3)	7.3 (32.9; 5.2–12.1)	212.8 (27.3; 152.7–351.4)
CV (95% CI)	5.6 (4.3–8.1)	4.3 (3.3–6.2)	6.3 (4.8–9.1)	31.7 (19.6–45.1)	27.3
Bias (95% CI)	0.8 ± 9.7 (−18.2–19.8)	2.2 ± 4.2 (−6.2–10.5)	1.2 ± 5.2 (−9.1–11.4)	−2.5 ± 10 (−22.0–17.0)	−128 ± 301 (−718–462)
r (95% CI)	−0.03 (−0.5–0.4)	0.1 (−0.3–0.5)	0.03 (−0.5–0.5)	0.3 (−0.3–0.7)	0.6 (0.03–0.8)

	∆SBP, mmHg	∆DBP, mmHg	∆HR, beats/min	∆Burst Incidence, busts/100 beats	∆Total Activity, AU/min
CPT
Visit 1	26 ± 16	18 ± 11	7 ± 9	34 ± 14	2,682 ± 1,098
Visit 2	27 ± 19	19 ± 10	5 ± 8	29 ± 14	2,027 ± 1,452
Paired t test	0.76	0.48	0.07	0.29	0.12
ICC (95% CI)	0.89 (0.72–0.95) †	0.87 (0.67–0.95) †	0.89 (0.72–0.96) †	0.49 (−0.60–0.84)	0.59 (−0.22–0.88)
TEM (%; 95% CI)	8.0 (30.6; 6.1–11.5)	5.3 (28.4; 4.1–7.6)	3.9 (69.7; 3.0–5.6)	11.9 (38.0; 8.5–19.7)	1022.1 (43.4; 733.5–1687.9)
CV (95% CI)	30.5 (23.3–44.0)	28.1 (21.5–40.6)	63.7 (48.7–91.9)	36.2 (26.0–59.8)	40.4 (29.0–66.7)
Bias (95% CI)	0.76 ± 11.3 (−21.5–23.0)	1.2 ± 7.3 (−13.2–15.5)	−2.1 ± 5.1 (−12.2–7.9)	−4.9 ± 16.0 (−36.4–26.5)	−655 ± 1,346 (−3,294–1,984)
r (95% CI)	0.2 (−0.2–0.6)	−0.04 (−0.5–0.4)	−0.2 (−0.6–0.3)	−0.4 (−0.6–0.5)	−0.3 (−0.3–0.7)

Values are means ± SD.

^*P < 0.05 vs. trial 1;

^†ICC P < 0.05.

AU, arbitrary units; CI, confidence interval; CPT, cold pressor test; CV, coefficient of variation; DBP, diastolic blood pressure; HR, heart rate; ICC, intraclass correlation coefficient [poor (0–0.39), moderate (0.40–0.59), good (0.60–0.74), very good (0.75–0.89), excellent (0.90–1.00)]; r, Pearson’s correlation coefficient for Bland–Altman plot; SBP, systolic blood pressure; TEM, typical error of the measurement.

Additional Factors to Consider for the Reproducibility of BP and MSNA Reactivity

MVC was not different between visits (visit 1: 401 ± 137 N versus visit 2: 382 ± 124 N; P = 0.08) and demonstrated excellent reproducibility [ICC = 0.96 (0.91–0.99), P < 0.001; TEM = 35.0 (26.8–50.5; 8.9%); CV = 8.5% (6.5%–12.3%)]. Force production during the last 30 s (visit 1: 28.2 ± 1.9 versus 27.6 ± 2.1%MVC, P = 0.90; visit 2: 28.1 ± 2.2 versus 28.1 ± 2.3%MVC, P = 0.90) and the last minute of HG (visit 1: 27.9 ± 1.8 versus 27.6 ± 2.1%MVC, P = 0.18; visit 2: 28.1 ± 2.3 versus 28.1 ± 2.2%MVC, P = 0.86) was not different between trials on either visit and demonstrated very good-to-excellent reproducibility [visit 1: ICC = 0.90 (0.79–0.95), P < 0.01; TEM = 1.0 (0.8–1.3; 3.7%); CV = 3.0% (2.4%–4.0%); visit 2: ICC = 0.97 (0.93–0.99), P < 0.01; TEM = 0.5 (0.4–0.7; 2.0%); CV = 1.9% (1.5%–2.6%)]. During both visits, the rating of perceived exertion at the end of HG was significantly higher in trial 2 compared with trial 1 (visit 1: trial 1: 14 ± 3 versus trial 2: 16 ± 3 arbitrary units, P < 0.001; visit 2: trial 1: 14 ± 3 versus trial 2: 16 ± 3 arbitrary units, P < 0.001), but the reproducibility of perceived effort was very good-to-excellent [visit 1: ICC = 0.86 (0.29–0.95), P < 0.001; TEM = 1.7 (1.4–2.3; 11.5%); CV = 6.8% (5.4%–9.1%); visit 2: ICC = 0.94 (0.24–0.99), P < 0.001; TEM = 1.1 (0.8–1.6; 7.6%); CV = 4.6% (3.5%–6.6%)]. Between visits, there were no differences in the rating of perceived effort during HG (P = 0.40), and the reproducibility was very good [ICC = 0.82 (0.54–0.93), P < 0.001; TEM = 1.6 (1.2–2.3; 10.7%); CV = 10.6% (8.1%–15.3%)].

There were no differences in the rating of perceived pain during PEI between trials at visit 1 (trial 1: 6 ± 2 versus trial 2: 6 ± 2 arbitrary units; P = 0.40), though these ratings were slightly higher following the second trial at visit 2 (trial 1: 5 ± 2 versus trial 2: 6 ± 2 arbitrary units; P = 0.03). However, this index demonstrated very good-to-excellent reproducibility at both visits [visit 1: ICC = 0.90 (0.77–0.95), P < 0.001; TEM = 1.4 (1.1–1.9; 23.6%); CV = 15.2% (12.1%–20.4%); visit 2: ICC = 0.80 (0.47–0.92), P < 0.001; TEM = 1.5 (1.1–2.2; 26.3%); CV = 23.2% (17.7%–33.5%)]. Between visits, there were no differences in the rating of perceived pain during PEI (P = 0.30) and the reproducibility was very good [ICC = 0.76 (0.40–0.90), P < 0.001; TEM = 1.6 (1.2–2.3; 27.0%); CV = 21.5% (16.4%–31.0%)]. The rating of perceived pain during the CPT was not different between visits (visit 1: 7 ± 2 versus visit 2: 7 ± 2 arbitrary units; P = 0.70) and had good reproducibility [ICC = 0.74 (0.32–0.90), P = 0.004; TEM = 1.2 (0.9–1.7; 16.5%); CV = 17.2% (13.2%–24.8%)].

DISCUSSION

There is an increasing recognition of the importance of considering individual responses when interpreting group data and renewed calls for reporting individual data in published manuscripts in an effort to enhance the rigor, reproducibility, and transparency of biomedical research (46). Although inherent variability in the magnitude of neurocardiovascular responsiveness to a variety of sympathoexcitatory stimuli is evident in healthy adults (11, 25), an occurrence that may be magnified in pathological conditions (6, 10, 12, 13, 22), the degree to which this variability is consistent within and between experimental visits is not well defined. In the present investigation, the reproducibility of the magnitude of the BP response to standard laboratory stimuli (isometric HG, PEI, and the CPT) was very good. However, the variability of the MSNA response of both burst occurrence and total activity to these perturbations was generally less consistent, particularly during PEI.

Based on several earlier studies (4, 12), the prevailing notion is that the cardiovascular responses to both static and dynamic HG are fairly reproducible. In this regard, in 1974, Ewing et al. (12) reported that the between-subject variability in the BP response during moderate-intensity static HG performed at multiple experimental visits was seven times greater than the within-subject variability. In addition, a majority of studies that have examined BP responsiveness to the CPT have concluded adequate reproducibility (17, 35, 38, 47), though this is not a universal finding (14, 33). In the current study, the overall reproducibility of BP responsiveness to HG, PEI, and the CPT was considered very good both within and between experimental visits. Indeed, the majority of subjects had a highly reproducible BP response (<4 mmHg). However, for each perturbation, there were always a few subjects (∼2–4) with substantial variability in the magnitude of BP responsiveness (>10 mmHg). Notably, these individuals were unique for each stimulus and for each visit (e.g., those with the largest variability in responsiveness to HG during visit 1 were not the same as those with the largest variability in responsiveness to HG during visit 2). Interestingly, removal of these individuals from statistical analyses yielded marked improvements in the relative degree of reproducibility. For example, the ICC for the HG-induced increase in MAP increased from 0.79 (0.47–0.91) to 0.91 (0.81–0.96), providing further support for the conclusion that BP responsiveness to sympathoexcitatory perturbations is largely reproducible. Nevertheless, accounting for this heterogeneity in the magnitude of BP responsiveness is an important consideration for study design and data interpretation, and our data suggest that differences of less than ∼5–6 mmHg in the BP response (i.e., the typical error of the measurement) between groups or conditions should be interpreted with some caution.

To date, few, if any, studies have comprehensively assessed the reproducibility of the MSNA response to static HG or PEI. However, multiple studies have reported marked individual heterogeneity in MSNA responsiveness to static HG (e.g., responders and nonresponders) (11, 13, 25). In this study, we also observed a relatively wide range (∼∆8–40 bursts/min) in the magnitude of the MSNA responses to HG between subjects. More importantly, despite the inclusion of a familiarization trial, the majority of participants exhibited greater increases in MSNA during the second HG trial at each visit, leading to the reproducibility of MSNA responsiveness to HG being more variable within an experimental visit, with a typical error of the measurement of up to 10 bursts/min noted during visit 1. This increased variability was consistent for all indices of MSNA. To account for this, statistical analyses to determine reproducibility between visits were made separately for each trial (i.e., visit 1 trial 1 versus visit 2 trial 1). The increased variability in MSNA responsiveness to HG appears limited to multiple assessments made within a single experimental visit, because it was highly reproducible between visits (i.e., visit 1 trial 1 versus visit 2 trial 1), with a typical error of the measurement of ∼5 bursts/min. In addition, the reproducibility of the increase in total activity during HG was excellent, further strengthening the overall conclusion. Similar to HG responses, many individuals also demonstrated greater MSNA responses during the second PEI trial within a visit. Surprisingly, unlike HG, the reproducibility of the MSNA response to PEI was poor-to-moderate, both within and between visits.

The reason(s) for the increased variability in MSNA responsiveness to static HG and PEI is not readily apparent. The degree of variability of MSNA appears to be independent of the quality of the microneurographic signal. That is, even some individuals in whom very high-quality nerve recordings were obtained demonstrated highly variable MSNA responsiveness to HG and PEI. Although speculative, it is possible that the sympathetic nervous system is sensitized by an initial bout of static HG, such that greater increases in MSNA are elicited during subsequent trials within a single experimental visit. Given the evidence for a threshold for metaboreflex activation (39), it is conceivable that, in some individuals, static HG performed at 30% MVC is slightly below the metaboreflex threshold. Indeed, 2 min of static HG at 30% MVC does not always evoke a robust MSNA and BP response (22, 26, 37). However, in these individuals, performing a repeated bout of HG, perhaps now crosses the metaboreflex threshold such that larger neural adjustments are evoked during the second trial. In this regard, metaboreflex sensitization may result from the first HG bout and contribute to a heightened MSNA response during the subsequent trial. Additionally, some individuals may rely on an increased contribution of central command to maintain the target force during the second HG trial (5). Interestingly, the greater MSNA response to the second HG and PEI trial does not appear to translate to a heightened pressor response, as BP responses were similar between trials and, overall, were more reproducible than MSNA. This may have to do with changes in the transduction of MSNA to a pressor response and/or the contribution of cardiac output to the BP response (24, 42). In terms of the former, Shoemaker et al. (40) have reported a dissociation between MSNA and leg vascular resistance during PEI.

To our knowledge, only two microneurography studies have reported the reproducibility of the MSNA response to the CPT (14, 38); one concluded that MSNA responsiveness was reproducible (albeit simply based on the lack of difference in the mean) (38), whereas the other reported a high degree of variability around the line of the identity (similar to the presentation of our data, though without a direct statistical comparison) (14). In the current study, similar to the BP response, the reproducibility of the MSNA burst frequency response to the CPT was very good. That is, the majority of participants displayed differences in the magnitude of the MSNA response to the CPT that were within the calculated typical error of the measurement of ∼6 bursts/min, with only 2 individuals demonstrating differences >10 bursts/min. However, of note, more variability was observed for both MSNA burst incidence and total activity.

Collectively, our data suggest important and notable differences in the degree of the reproducibility of MSNA responsiveness to standard laboratory sympathoexcitatory stimuli within and between experimental visits. These conclusions are generally consistent for all indices of MSNA. However, because burst incidence necessarily depends on the reproducibility of heart rate responses, and considering the inherent limitations of burst amplitude analyses and comparisons of total activity, we relied upon burst frequency as our primary index by which to assess and interpret the reproducibility of MSNA responsiveness. Within a single visit, the typical error of the measurement of the magnitude of the MSNA response to HG and PEI was as high as ∼10–12 bursts/min; however, interestingly, this variability was reduced between visits (∼5–6 bursts/min). This is an important consideration when performing multiple bouts of HG and PEI during a single visit, whether by study design or because of an experimental issue. For example, if the bout of HG and PEI has to be repeated for some subjects within a single visit because of the loss of the nerve recording or a subject-related matter (e.g., coughing/sneezing, tension), but in other subjects it is not repeated, there may be greater responses in the subjects who repeated the trial that is not being considered in data analysis or interpretation. Overall, this would have significant implications for studies making comparisons between groups or treatments and may be an underappreciated source of variability.

Reasons for the greater MSNA variability are not clear but several possibilities warrant consideration and discussion. Because absolute baseline values may influence the magnitude of responsiveness to a perturbation (i.e., calculation of a “delta” value), a minimum of 15 min of quiet rest separated all sympathoexcitatory maneuvers to ensure that BP and MSNA returned to resting baseline values. Perhaps surprisingly, the individuals having the greatest differences in baseline values between trials were not those who demonstrated the greatest variability in responsiveness, suggesting that baseline variability is not a contributing factor. A second possibility is differences in MVC, and therefore absolute workload, between study visits. To account for this potential confounding source of variation, MVC was assessed at each experimental visit and was reproducible. Further, the participant having the least (<1% difference) and the greatest (26% difference) variability in MVC between visits each demonstrated highly reproducible BP and MSNA responses to HG. Thus, MVC appears unlikely to explain the variability in responsiveness to HG and PEI. Similarly, the potential for differences in perceived effort and perceived pain to influence the variability of responsiveness was also not likely. The water temperature of the ice bath used for the CPT was not measured, which may be a potential source of error. However, to limit this possibility, the amount of ice added was consistent for all studies. More importantly, the variability of the MSNA response to the CPT was less than that for HG and PEI. Lastly, although frequency domain or action potential recruitment analyses of MSNA were beyond the scope of the present study, it will be of interest in future studies to consider how variability in these parameters may potentially contribute to the degree of reproducibility of MSNA outflow during sympathoexcitatory perturbations.

In conclusion, the present investigation suggests that the reproducibility of the pressor response to several standard laboratory sympathoexcitatory stimuli is very good. However, there is more variability in the magnitude of the MSNA response to these perturbations, despite attempts to control for potential known sources of heterogeneity. Collectively, these data provide novel insight for both study design and data interpretation when comparing neurocardiovascular responsiveness between different conditions, groups, or studies, as well as before and after interventions and treatments.

GRANTS

This work was supported by TL1 TR002016 and NASA Pennsylvania Space Grant Consortium (to G. A. Dillon), NIH) R01 HL093238 (to L. M. Alexander), NIH K99 HL133414 (to J. L. Greaney), and NIH PO1 HL137630-01A1 (to P. J. Fadel).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

L.M.A., P.J.F., and J.L.G. conceived and designed research; G.A.D., Z.S.L., and J.L.G. performed experiments; G.A.D. and Z.S.L. analyzed data; G.A.D., Z.S.L., L.M.A., L.C.V., J.W., P.J.F., and J.L.G. interpreted results of experiments; G.A.D. prepared figures; G.A.D. and J.L.G. drafted manuscript; G.A.D., Z.S.L., L.M.A., L.C.V., P.J.F., and J.L.G. edited and revised manuscript; G.A.D., Z.S.L., L.M.A., L.C.V., J.W., P.J.F., and J.L.G. approved final version of manuscript.