Propofol and midazolam are used extensively for both general anesthesia and sedation, and small-dose propofol and midazolam sedation is used in dental practice. Hemodynamic changes, including changes in heart rate (HR) and arterial blood pressure (BP), have been reported during conscious sedation with propofol (1,2) and midazolam (3). One potential mechanism is that propofol and midazolam affect the HR and BP via changes in the cardiac autonomic nervous system (ANS), an important neural control system for maintaining cardiovascular stability. Previous studies have evaluated the effects of propofol and midazolam on cardiac ANS functions by means of HR variability (HRV) analysis, a noninvasive and widely used technique (4,5). HRV analysis can provide important clinical information on the effect of anesthesia on the ANS and central nervous system because cyclical variation in HR is mediated by central neural mechanisms and by baroreceptors and chemoreceptors (6). The sympathetic and parasympathetic components of HRV are active over different frequency ranges. The low-frequency component (LF, 0.05–0.15 Hz) is influenced by both cardiac sympathetic and parasympathetic activity (5), and the high-frequency component (HF, >0.15 Hz) originates from cardiac parasympathetic activity (7). Therefore, the LF/HF ratio reflects dominance of cardiac sympathetic activity (8,9). In addition, entropy has been reported to reflect parasympathetic modulation of HR (10).
There has been controversy as to whether propofol increases sympathetic or parasympathetic activity during general anesthesia (11–13) and there are no reports on the effects of propofol sedation on cardiac ANS activity. It has not been clarified that midazolam activates either sympathetic or parasympathetic activity during general anesthesia (9,14) and that it increases sympathetic activity during sedation (15). Further, it is unknown whether sympathetic or parasympathetic activity is more affected during conscious sedation with propofol and midazolam. Two specific hypotheses were tested by HRV analysis in this study: 1) propofol induces predominance of parasympathetic activity, leading to decreased HR and BP, and 2) midazolam induces predominance of sympathetic activity, leading to increased HR and decreased BP.
With the approval of the Tokyo Medical and Dental University Ethics Committee and informed consent, 30 dental implantation patients (ASA physical status I–II; age 30–62 yr) were recruited for the study. Patients who had severe ischemic heart disease, congestive heart failure, diabetes mellitus, or other disorders known to affect ANS function were excluded. Five patients were taking an antihistamine for allergy. No patients received premedication. The patients were randomly divided into a propofol group and a midazolam group. Each patient was requested to take only liquids and a light soft meal within 2 h before sedation. The patients rested supine on a dental chair for 2 min while standard monitors were applied. BP, HR, electrocardiogram (ECG), and peripheral oxygen saturation (Spo2) by pulse oximetry (BP-608 Evolution; Colin Medical Technology; Tokyo, Japan) were noninvasively and continuously monitored. Respiratory rate (RR) and body temperature were measured at baseline and during the recovery period. During surgery RR was monitored at 5-min intervals by an anesthesiologist. The Bispectral Index (BIS) value was monitored with a BIS® monitor (Model A2000; Aspect Medical Systems, Natick, MA) and BIS Sensor strips. The impedance of each electrode was maintained below 5 KΩ.
An anesthesiologist used the modified Observer’s Assessment of Alertness/Sedation Score (OAA/S) (Table 1) to assess sedation depth (16). We defined OAA/S 3 as a conscious sedation state and OAA/S 5 as baseline and recovery period from sedation. The anesthesiologist was blinded as to the BIS value so that sedation depth would be assessed only by the OAA/S clinically. The control data or baseline data were recorded for 2 min before sedation with the patient lying quietly and breathing spontaneously.
A 22-gauge catheter was inserted into a forearm vein for drug administration. Propofol (Diprifusor™, AstraZeneca Pharmaceuticals, UK) was infused at an initial target effect-site concentration of 1 μg/mL and increased in increments of 0.5 μg/mL after the OAA/S assessment every minute until OAA/S 3 was reached. Midazolam was administered at a rate of 0.5 mg over 30 s each time, and the dose was increased after the OAA/S assessment every minute to a total bolus of 0.075 mg/kg until OAA/S 3 was reached. The sedation period data were recorded for 2 min beginning 1 min after reaching OAA/S 3. Lidocaine (2%) with epinephrine (1:80,000 or 1:160,000) was used as the local anesthetic to suppress pain associated with the dental procedures after the sedation period. The surgical procedure was started 5 min after infiltration with local anesthetic, provided vital signs were stable (i.e., within 20% of baseline values). An OAA/S of 3 to 4 was maintained in both groups during the surgery by additional drug titration. Both anesthetics were discontinued when the last suture was completed. While the patients were left undisturbed, the OAA/S score was determined every minute until achieving the OAA/S 5, and recovery data were recorded for 2 min. The BIS value and HRV data were continuously recorded on separate personal computers by an investigator. Data were analyzed and averaged at 2-min intervals at baseline, during sedation, and during the recovery period.
ECG electrodes (leads I, II, and V5) were attached for HRV analysis, and the ECG signals were fed to a memory HR monitor (LRR-03™; GMS, Tokyo, Japan). The fast peaks of R waves on the ECG were detected, and the R-R interval (RRI) was continuously monitored. The data were transferred to an online computer loaded with HRV analysis software (TARAWA/WIN; Suwa Trust, Tokyo, Japan). For real-time analysis, the RRI data were obtained at 2-ms sampling intervals and analyzed by the maximum entropy method at high resolution with the “MemCalc” computer program (17–19). “MemCalc” executes a linearized version of the nonlinear least squares method for fitting analysis in the time domain combined with the maximum entropy method or spectral analysis in the frequency domain. It provides reliable analyses of HRV over a minimum interval of 30 s and recognizes the abnormal RRI of premature beats or artifacts, including noise, and removes it automatically. The power of the RRI (ms2) with LF (0.04–0.15 Hz) and HF (0.15–0.5 Hz) bands was calculated. Total spectral power (TP, 0.04–0.5 Hz) was obtained by addition of LF and HF. To confirm normal distribution, LF and HF (proportional power) were calculated as percentages of TP (%LF = LF/TP × 100 and %HF = HF/TP × 100), and the LF/HF ratio was also assessed. “MemCalc” was also used to calculate entropy from a pulse time series of 4 RRI. HRV was expressed as randomness of the pulse interval. Thus, entropy was expressed from 0 (regular interval pulse series, no variability) to 100% (maximal randomness, e.g., noise).
The sample size of 12 patients in each group was determined by a power analysis (α = 0.05; β = 0.20) based on our pilot data on the assumption that a 10% difference between groups in percentage change in HRV parameters relative to baseline would be important clinically, and 15 patients were actually included in each group in the study. The data for descriptive (categorical) variables were subjected to the χ2 test. Total spectral power data and entropy were analyzed by a nonparametric method (Mann-Whitney U-test). Repeated-measures analysis followed by Bonferroni correction was performed for between-group comparison of hemodynamic variables and percentage changes in HRV data and LF/HF ratio. Statistical significance was determined as P < 0.05.
There were no significant differences in demographic data or duration of sedation and surgery between the two groups (Table 2). The median dose of local anesthetic was 5.4 (range, 3.6–10.8) mL in the propofol group and 5.4 (range, 3.6–9.5) mL in the midazolam group. BIS values at baseline and during recovery were comparable in both groups (Table 3) but were significantly different during sedation (propofol 95% confidence interval, 69–75; midazolam 95% confidence interval, 74–79), even though all patients were consciously sedated at OAA/S 3, underwent surgery uneventfully, and required 4 ± 1 min after surgery to reach OAA/S 5 with propofol and 5 ± 2 min with midazolam.
Baseline HR and BP were comparable in the propofol and midazolam groups (Table 3), but a statistically significant difference in HR was noted during sedation and the recovery period. Both groups showed comparable decreases in BP during sedation. During the recovery period, BP returned to baseline with midazolam but remained below baseline with propofol. There were no changes in RR, Spo2, or body temperature between the baseline and the recovery periods in either group. During the sedation period, the RR (breaths/min) was 16 ± 2 in the propofol group and 16 ± 3 in the midazolam group, and Spo2 was 97 ± 2% in both groups.
TP, LF and HF power, and entropy decreased significantly during the sedation period in both groups (Table 4). During the recovery period, HF and entropy were significantly higher than the baseline values in the propofol group but below the baseline values in the midazolam group (Table 4). During the sedation period, LF/HF decreased significantly in the propofol group but did not change significantly in the midazolam group. During recovery, LF/HF remained below baseline in the propofol group but increased in the midazolam group. The differences between the groups in both the sedation and recovery period were statistically significant. The entropy changes in both groups were directionally similar to the changes in HF.
The results of this study demonstrated that IV conscious sedations with propofol and midazolam differed in terms of changes in HR and HRV during the sedation and recovery periods.
Propofol induced significant decreases in TP, LF, HF, entropy, and LF/HF ratio with no change in HR, indicating predominance of parasympathetic activity during sedation (OAA/S 3). The decreased BP with no change in HR indicates that propofol attenuates the baroreflex reaction (20). According to Deutschman et al. (11), propofol (2.5 mg/kg) reduces TP, LF, and HF without any change in HR. Our findings are consistent with theirs. By contrast, Kanaya et al. (12) reported that continuous infusion of propofol at a rate of 300 μg · kg−1 · min−1 reduced cardiac parasympathetic tone based on a decrease in entropy and HF with no significant changes in LF, LF/HF, and HR at BIS values of 80 and 60. However, HRV was measured during spontaneous respiration (RR unknown) and intermittent positive pressure ventilation was performed occasionally. Changes in RR and tidal volume influence HRV, and their study failed to demonstrate a clinical conscious level despite BIS monitoring. The absolute power of HRV used in their study is inconsistent with the proportional power used in ours. Scheffer et al. (13) observed a decrease in HF with no significant change in LF after 2.5 mg/kg of propofol, suggesting predominance of sympathetic action. In their study, HRV was analyzed in the unconscious state during spontaneous respiration with assisted artificial ventilation. Although HF is affected by tidal volume and RR, they did not discuss their effect, nor did they maintain patients on propofol infusion before further intervention, as we and Deutschman et al. did.
During the sedation period, midazolam caused a significant decrease in TP, LF, HF and entropy, with a significant increase in HR and no change in LF/HF, suggesting decreased parasympathetic activity and unchanged sympathetic activity. The increase in HR with decrease in BP during sedation means that baroreflex activity is compensated. However, there has been controversy as to whether midazolam attenuates (21) or compensates (22) the negative feedback mechanism in the regulation of HR by baroreflex activity. Galletly et al. (15) demonstrated a vagolytic effect of midazolam (0.1 mg/kg) that caused a decrease in HF and BP, and our findings are consistent with theirs. Komatsu et al. (14) found that midazolam (0.3 mg/kg) decreased normalized unit (nu) LF while increasing the nuHF power component of HRV without affecting HR, thereby resulting in sympathetic depression. Increased RR after midazolam could not be excluded as a cause for the change in their HRV data. Previous reports (9,14–15) and our findings suggest that midazolam enhances the dominance of parasympathetic action at very large doses but induces dominance of sympathetic action at small doses.
In the recovery period (OAA/S 5), propofol induced significant increases in HF and entropy and significant decreases in LF and LF/HF, resulting in parasympathetic activation and sympathetic depression. The decrease in HR was probably attributable to dominance of parasympathetic action. The decrease in HR without change in BP cannot be interpreted as activation of a baroreflex. In the recovery period midazolam was associated with a significant increase in LF/HF ratio and significantly decreased HF and entropy, indicating predominant sympathetic action. It is difficult to explain the absence of any change in HR during the recovery period despite the dominance of sympathetic action. It may be attributable to attenuated baroreflex activity because BP recovered to the baseline level. Although the OAA/S score and BIS value returned to their baseline levels 4 minutes and 5 minutes after discontinuation of propofol and midazolam infusion, respectively, cardiac ANS activity in the recovery period differed from baseline. This activity is attributable to the remaining plasma drug action because of the context-sensitive half-time (23). The parasympathetic activation after discontinuation of propofol infusion is probably attributable to delayed recovery of sympathetic activity because the inhibition of sympathetic activity rate in the sedation period was more than that of parasympathetic activity. The predominant sympathetic activity was probably caused by delayed recovery of parasympathetic activity after discontinuation of midazolam infusion because parasympathetic activity decreased without any decrease in sympathetic activity in the sedation period.
Finally, our study may be criticized for lack of blindness to drug and data analysis, influence of the respiratory pattern, and use of epinephrine and lidocaine. The evaluation of sedation depth was independently conducted by an anesthesiologist, and the data recording and analyses were done by an investigator. Therefore, the potential of observer and recording bias would be reduced. Differences in BIS during sedation between groups may give rise to different sedation levels. However, BIS has drug-specific characteristics, and it is still impossible to rely on a single BIS value to assess depth of sedation. Because the BIS values for anesthetics and sedative drugs were found to be widely dispersed (24), BIS monitoring may not have been an appropriate means of assessing depth of conscious sedation. We assessed depth of sedation clinically based on the OAA/S in our research. The possibility remains that changes in respiratory pattern during sedation may alter the power of HRV. HF values vary with tidal volume and RR if RR is <8 bpm (5). RR remained at 13–19 bpm in both groups, meaning that HF was unaffected by RR. Unfortunately, tidal volume and minute volume were not monitored in this study. Epinephrine is likely to have little effect on HRV, and the effects of small concentrations of lidocaine (the median dose in both groups was 5.4 mL, or 108 mg) are unclear. However, our data were collected 1 hour after infiltration in both groups. In addition, the mean plasma epinephrine concentration peak occurs 3 minutes after oral mucosal injection and returns to baseline 5 minutes after injection (25).
Our report showed that HRV analysis is a noninvasive method that is applicable to assessment of changes in sympathovagal regulation that are associated with hemodynamic changes. Our findings imply that conscious sedation with propofol and midazolam enhances dominance of parasympathetic and sympathetic activity, respectively. This finding should be considered during conscious sedation, especially in patients at risk of cardiovascular complications.
1. Dorrington KL. Asystole with convulsion following a subanaesthetic dose of propofol plus fentanyl. Anaesthesia 1989;44:658–9.
2. Cillo JE. Propofol anesthesia for outpatient oral and maxillofacial surgery. Oral Surg Oral Med Oral Pathol 1999;87:530–8.
3. Van Der Bijl P, Roelofse JA, Joubert JJ, et al. Comparison of various physiologic and psychomotor parameters in patients sedated with intravenous lorazepam, diazepam, or midazolam during oral surgery: J Oral Maxillofac Surg 1991;49:672–8.
4. Akselrod S, Gordon D, Ubel FA, et al. Power spectral analysis of HR fluctuation: a quantitative probe of beat-to-beat cardiovascular control. Science 1981;213:220–2.
5. Pomeranz B, Macaulay RJ, Caudill MA, et al. Assessment of autonomic function in human by HR spectral analysis. Am J Physiol 1985;248(1Pt 2):H151–3.
6. Malliani A, Pagani M, Lombardi F, Cerutti S. Cardiovascular neural regulation explored in the frequency domain. Circulation 1991;84:482–92.
7. Kunze DL. Reflex discharge patterns of cardiac vagal efferent fibres. J Physiol (Lond) 1972;222:1–15.
8. Huang HH, Chan HL, Lin PL, et al. Time frequency spectral analysis of HR variability during induction of general anaesthesia. Br J Anaesth 1997;79:754–8.
9. Nishiyama T, Misawa K, Yokoyama T, Hanaoka K. Effects of combining midazolam and barbiturate on the response to tracheal intubation: changes in autonomic nervous system. J Clin Anesth 2002;14:344–8.
10. Palazzolo JA, Estafanous FG, Murray PA. Entropy measures of heart rate variation in conscious dogs. Am J Physiol 1998;274:H1099–105.
11. Deutschman CS, Harris AP, Fleisher LA. Changes in heart rate variability under propofol anaesthesia. A possible explanation for propofol-induced bradycardia. Anesth Analg 1994;79:373–7.
12. Kanaya N, Hirata N, Kurosawa S, et al. Differential effects of propofol and sevoflurane on heart rate variability. Anesthesiology 2003;98:34–40.
13. Scheffer GJ, Ten-Voorde BJ, Karemaker JM, et al. Effects of thiopentone, etomidate and propofol on beat-to-beat cardiovascular signals in man. Anaesthesia 1993;48:849–55.
14. Komatsu T, Singh PK, Kimura T, et al. Differential effects of ketamine and midazolam on HR variability. Can J Anaesth 1995;42:1003–9.
15. Galletly DC, Williams TB, Robinson BJ. Periodic cardiovascular and ventilatory activity during midazolam sedation. Br J Anaesth 1996;76:503–7.
16. Chernik DA, Gillings D, Laine H, et al. Validity and reliability of the observer’s assessment of alertness/sedation scale: study with intravenous midazolam. J Clin Psychopharmacol 1990;10:244–51.
17. Ohtomo N, Tanaka Y. New method of time series analysis and “MemCalc.” In: Saito K, ed. A recent advance in time series analysis by maximum entropy method. Sapporo: Hokkaido University Press, 1994;11–29.
18. Ohtomo N, Terachi S, Tanaka Y, et al. New method of time series analysis and its application to Wolf’s sunspot number data. Jpn J Appl Physiol 1994;33:2821–31.
19. Sawada Y, Ohtomo N, Tanaka Y, et al. New technique for time series analysis combining the maximum entropy method and non-linear least squares method: its value in heart rate variability analysis. Med Bio Engl Comput 1997;35:318–22.
20. Ebert TJ, Muzi M. Propofol and autonomic reflex function in humans. Anesth Analg 1994;78:369–7.
21. Mary J, Gauzit R, Lefevre P, et al. Effects of diazepam and midazolam on baroreflex control of HR and on sympathetic activity in human. Anesth Analg 1986;65:113–9.
22. Reves JG, Fragen RJ, Vinik HR, Greenblatt DJ. Midazolam: pharmacology and uses. Anesthesiology 1985;62:310–24.
23. Reves JG, Glass PSA, Lubarsky DA. Nonbarbiturate intravenous anesthetics. In: Miller RD, eds. Anesthesia. New York: Churchill Livingstone, 2000;228–72.
24. Ibrahim AE, Taraday JK, Kharasch ED. Bispectral Index monitoring during sedation with sevoflurane, midazolam, and propofol. Anesthesiology 2001;95:1151–9.
© 2005 International Anesthesia Research Society
25. Homma Y, Ichinohe T, Kaneko Y. Oral mucosal blood flow, plasma epinephrine and haemodynamic responses after injection of lidocaine with epinephrine during midazolam sedation and isoflurane anaesthesia. Br J Anaesth 1999;82:570–4.