Choice of Anesthesia Regimen Significantly Impacts Shunting and Diastolic Ventricular Function During Cardiac Catheterization

doi:10.21203/rs.3.rs-2086136/v1

Download PDF

Research Article

Choice of Anesthesia Regimen Significantly Impacts Shunting and Diastolic Ventricular Function During Cardiac Catheterization

https://doi.org/10.21203/rs.3.rs-2086136/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background

Most pediatric cardiac catheterizations are performed under general anesthesia, but few studies have examined how choice of anesthesia agent impacts hemodynamic data in pediatric patients. Our objective was to compare hemodynamics obtained under inhaled versus total intravenous anesthesia, and to determine if one more closely reflects the awake state.

Methods

Pediatric patients requiring percutaneous closure of a patent ductus arteriosus (PDA) or surveillance endomyocardial biopsy after orthotopic heart transplant (OHT) at Monroe Carell Jr. Children’s Hospital at Vanderbilt were enrolled. Each participant underwent hemodynamic assessment under inhaled volatile anesthesia (sevoflurane), and then under total intravenous anesthesia (dexmedetomidine and propofol) after a brief washout period. Paired data were analyzed using the Wilcoxon-signed rank test. Linear regression was used to analyze correlation between continuous variables.

Results

Twenty-five patients (13 OHT, 12 PDA) aged 6 months to 18 years were studied. Compared to baseline pre-procedure non-invasive blood pressure, systolic blood pressure was significantly lower under inhaled anesthesia versus intravenous anesthesia. Systemic vascular resistance was also significantly lower under inhaled anesthesia, while pulmonary vascular resistance was similar in both phases. In patients with a PDA, left-to-right shunting increased significantly under intravenous anesthesia versus inhaled. Similarly, in transplant patients, left ventricular end diastolic pressure (LVEDP) increased following transition to intravenous anesthesia, with a positive correlation between the increase observed and the initial LVEDP under inhaled anesthesia.

Conclusions

Hemodynamics were impacted by the mechanism of general anesthesia. An intravenous anesthesia regimen may more closely reflect the awake, pre-catheterization state compared to inhaled volatile anesthesia.

Anesthesia

Catheterization

Congenital Heart Disease

Obtaining accurate hemodynamic data is vitally important during catheterization of pediatric patients with congenital heart disease (CHD). These data inform surgical planning and critical care management in fragile patients. While practice varies among institutions, in the previous two decades, pediatric catheterization labs have moved away from operator directed sedation and toward general anesthesia delivered by an anesthesiologist [1,2]. This has followed the trend of more complex and longer interventional procedures being performed as alternatives to surgery.

Anesthetic agents are known to affect myocardial performance and vascular resistance [3]. This may affect whether hemodynamic data obtained at the time of cardiac catheterization can be interpreted to accurately reflect the physiology of an awake state. Given differential effects on pulmonary vascular resistance (PVRi) and systemic vascular resistance (SVRi), the choice of anesthetic agent has the potential to result in dramatic changes in hemodynamic measurements for CHD patients, especially those with shunting lesions.

Prior studies in pediatric patients have examined the hemodynamic effect of intravenous agents administered at a conscious sedation level or given via rapid bolus [4-6], but to our knowledge, no prior studies have compared the hemodynamic effects of two widely employed strategies for general anesthesia – inhaled volatile anesthesia (IA) versus total intravenous anesthesia (TIVA). Since most pediatric catheterizations are now performed under general anesthesia, understanding how anesthetic approach affects hemodynamics is critically important.

We conducted a prospective study of patients presenting to the pediatric catheterization lab to determine whether measured hemodynamic markers differed while under IA (sevoflurane) versus TIVA (dexmedetomidine and propofol) in the same patient. We sought to determine which anesthetic regimen produced hemodynamic measurements that most closely reflected the patients’ resting state and determine the degree of change in the hemodynamic state.

Study population.

Patients aged 6 months to 18 years presenting to the pediatric catheterization lab for percutaneous closure of a PDA or endomyocardial biopsy after OHT under general anesthesia were eligible for enrollment. The institutional review board at Vanderbilt University Medical Center approved the study. Exclusion criteria included subjects actively treated for pulmonary hypertension, allergy, adverse reaction, or medical contraindication to receiving sevoflurane, propofol, or dexmedetomidine, and need for inotropic or vasopressor support. Eligible subjects were approached prior to catheterization and informed consent was obtained. Demographic data including pre-procedure medications and diagnoses were recorded.

Anesthesia procedures.

Induction of anesthesia was performed with inhaled sevoflurane with additional intravenous administration of rocuronium 0.7-1 mg/kg and fentanyl 1-2mcg/kg following peripheral intravenous (PIV) catheter placement. Endotracheal intubation was then performed. In the first phase of hemodynamic measurements, general anesthesia was maintained using 0.75 – 1.5 minimum alveolar concentration of sevoflurane (IA) to sustain a bispectral index (BIS) measurement of 40 – 60. Percent of inhaled oxygen was maintained between 21 – 30% and end-tidal CO2 was maintained between 35 – 40 mmHg throughout the procedure. A full right and left heart catheterization without angiography was then performed and hemodynamic measurements obtained.

A transition period was then performed between IA and TIVA during which 1) a dose of IV midazolam (0.05 – 0.1 mg/kg) was administered, 2) followed by an IV dexmedetomidine bolus of 0.5 mcg/kg over five minutes and then continuous infusion at 1.0 – 1.5 mcg/kg/hr, 3) initiation of a propofol infusion at a rate of 100 – 150 mcg/kg/min (without bolus), and finally 4) sevoflurane discontinued and exhaled percent of sevoflurane monitored until less than 0.2% prior to proceeding with the second set of hemodynamics. During the transition period, BIS was monitored and maintained within the range of 40 – 60. Bolus doses of propofol could be administered as needed to maintain adequate depth of anesthesia during both the transition period and throughout the use of TIVA. The length of transition time for each patient was recorded. Repeat left and right heart catheterization was then performed under TIVA.

Hemodynamics.

Baseline non-invasive assessment of vital signs prior to anesthesia induction were obtained including blood pressure, heart rate, and height and weight. After induction of IA pressure measurements were obtained in the right atrium, right ventricle, a branch pulmonary artery, a pulmonary capillary wedge pressure (PCWP), descending aorta, and the left ventricle (LVEDP). Oxygen saturations were measured in the superior vena cava (mixed venous saturation), a branch pulmonary artery, and the descending aorta. For the purpose of calculations, pulmonary venous saturation was assumed to be the same as the aortic saturation. These measurements were used to calculate both systemic and pulmonary blood flow using the Fick method, SVRi, PVRi, and the ratio of pulmonary to systemic blood flow (Qp/Qs).

Statistical analysis.

The primary exposure in this study was IA versus TIVA. The primary outcome was the hemodynamic data obtained during each phase of anesthesia, which included blood pressure (mmHg), heart rate (bpm), SVRi (Wood Units x m²), PVRi (Wood Units x m²), LVEDP (mmHg) and Qp/Qs. Paired data were analyzed using the Wilcoxon-signed rank test. Linear regression was used to analyze correlation between continuous variables. All data analysis was performed using Stata IC 15 (Statacorp, College Station, Texas).

Study Population.

We enrolled 25 patients including 13 OHT patients and 12 PDA patients (table 1).

Table 1: Demographics and baseline characteristics.

Demographics	N = 25
Diagnosis	OHT (N = 13)	PDA (N = 12)
Age	10.3 (8.7 – 13.1)	2.4 (1.3 – 3.7)
% Female	54%	75%
Baseline systolic BP	113 (107 – 127)	108 (105 – 112)
Baseline heart rate	103 (101 – 113)	119 (98 – 126)

In the OHT patients, 54% were female and the median age was 10.3 years (IQR 6.6 – 15.3). 91% of OHT patients were on antihypertensive therapy. As per the policy at our institution, OHT patients took their medications as scheduled on the morning of their catheterization. In the PDA patients 75% were female and the median age was 2.0 years (1.0 – 3.8). None of the PDA patients were on outpatient medications prior to catheterization.

Hemodynamics.

For all patients, systolic blood pressure (SBP) was significantly lower while receiving IA compared to baseline [70 (66 – 78) vs 108 (106 – 114) mmHg, p<0.001]. However, no difference was observed in SBP while receiving TIVA from baseline [109 (99 – 116) vs 108 (106 – 114) mmHg, p=0.61] (figure 1).

Median SVRi was significantly lower under IA compared with TIVA [11.1 (8.1 – 14.3) vs 21.0 (17.2 – 22.2) Wood Units x m², p<0.001], while PVRi did not differ between IA and TIVA [1.2 (0.7 – 1.7) vs 1.3 (1.0 – 1.9) Wood units x m², p=0.058] (figure 2). The transition period between IA and TIVA ranged from 9 to 20 minutes with a median transition period of 13 minutes.

In PDA patients, the degree of left-to-right shunt (Qp/Qs) measured under IA was significantly lower than when measured under TIVA (p=0.003). In addition, there was a positive correlation between the magnitude of Qp/Qs measured under IA and the increase in Qp/Qs under TIVA (R²=0.4, p=0.03) (figure 3).

In OHT patients, the LVEDP was significantly lower under IA than under TIVA (p=0.002), and similarly to Qp/Qs, there was a positive correlation between the LVEDP measured under IA and the increase in LVEDP after switching to TIVA (R²=0.39, p=0.03) (figure 4).

For all patients, there was a strong relationship between age and SVRi under IA (R²=0.57, p<0.001). In addition, there was a trend toward larger increases in SVRi between anesthesia phases at younger ages, though this did not reach statistical significance. The relationship between age and SVRi persisted but was less prominent under TIVA. Age-dependent changes in PVRi were much smaller and weakly correlated with age under both anesthesia protocols. There was no observed association between age and PVRi change between anesthesia phases (figure 5).

Hemodynamic measurements obtained during cardiac catheterization differed significantly in the same patient depending on the type of general anesthetic administered. Nearly every PDA patient showed an increase in degree of left to right shunting under TIVA as compared with IA (figure 3a). The only patients in whom there was not an increase in shunt were those who had very small PDAs with no measurable shunting in either condition. There were no patients with a Qp/Qs ≥ 2 under IA, while 5 of 12 patients had Qp/Qs ≥ 2 under TIVA. While these data would not necessarily alter the decision to close a PDA, such differences could alter management decisions in other populations such as shunt-dependent or other single ventricle patients.

The large difference in shunting in the PDA cohort can be attributed to significantly higher SVRi under TIVA versus IA, while PVRi was not significantly different between anesthesia phases. We would expect that similar changes in shunting could be seen in a variety of CHD lesions where the degree of shunting is determined by the relationship of SVR to PVR. Such lesions would include non-restrictive ventricular septal defects and atrioventricular septal defects, patients (including single ventricle patients) palliated with Blalock-Taussig-Thomas or Sano shunts and other forms of complex CHD. We observed a strong correlation between the magnitude of left-to-right shunt under IA and the increase in shunt observed after changing to TIVA (figure 3b). This demonstrates that the larger the shunt, the more it may be underestimated by obtaining data under IA. In a clinical context, this may result in underappreciation of pulmonary over-circulation in patients with certain forms of CHD.

Similarly, every OHT patient showed an increase in LVEDP when transitioned from IA to TIVA (figure 4a). There was a strong correlation between higher LVEDP under IA and larger increases in LVEDP between phases (figure 4b). We suspect this is caused by exposure to increased SVR under TIVA as compared with IA. The relationship between SVR and LVEDP becomes steeper in a failing ventricle [7]. The transition between anesthetic agents therefore constitutes a dynamic assessment of left ventricular diastolic function for these patients.

Systolic blood pressure under IA was much lower than baseline (figure 1). In contrast, systolic blood pressure under a TIVA regimen of dexmedetomidine and propofol closely matched baseline non-invasive systolic pressures obtained prior to anesthesia. SVRi under IA was much lower than under TIVA, mirroring the change in systolic blood pressure. Based on the well-known ability of sevoflurane to reduce SVRi [8], we believe that systolic blood pressure is a reasonable surrogate for baseline SVRi. These data suggest that the SVRi under TIVA may more closely reflect the awake hemodynamic state. PVRi was not significantly different when measured under IA versus TIVA. Thus, the mode of anesthesia disproportionately affected SVRi over PVRi. Since SBP under TIVA closely matched SBP prior to anesthesia, it can be argued that the higher Qp/Qs and LVEDP obtained under TIVA more closely reflects the awake physiologic state.

We also observed a trend toward an inverse correlation between the change in SVRi from one anesthesia phase to the other and the age of the patient, though this did not reach statistical significance (figure 5c). Smaller IA-induced changes with increasing age imply that the degree of Qp/Qs underestimation under IA may be greatest in the youngest patients. We did not enroll any patients younger than six months old in our study, so it is not clear if this association extends to newborns. However, it is often the youngest shunted patients with CHD that struggle the most with “over-circulation” of the pulmonary bed and associated compromise of systemic perfusion. Thus, the choice of anesthetic agent is of particular importance in shunted single ventricle patients.

A strength of our study is measuring hemodynamic data in the same patient under different anesthesia conditions, allowing each subject to serve as his or her own control. We started with inhaled sevoflurane as the first phase because of the ability to rapidly eliminate it from the body during the transition phase and to document this by direct measurement of the exhaled percentage of sevoflurane. The two most common inhaled anesthetic anesthesia agents used in our pediatric catheterization lab are sevoflurane and isoflurane. Sevoflurane has a shorter half-life and so was chosen over isoflurane for this study. Isoflurane has been shown to increase PVR in some patient populations [3] and it could be useful to repeat these experiments with isoflurane to determine its effects relative to sevoflurane in pediatric patients.

One limitation is that we would like to have studied some patients in the TIVA phase first followed by the IA phase to determine if the impact on hemodynamics was similar. In some patients, we observed a transient spike in systolic blood pressure during the dexmedetomidine bolus, as has been previously described [5]. It is possible that some of this increase in pressure may have represented a rebound effect from turning off the inhaled sevoflurane. This effect did not persist longer than 10 minutes, and blood pressure plateaued by the time the exhaled sevoflurane approached zero. Starting with TIVA would have required a much longer transition period to allow for the IV agents to be eliminated. The prolonged anesthesia time that would have been required for this approach represented an unacceptable added risk.

The combination of dexmedetomidine and propofol was chosen in close collaboration with our pediatric cardiac anesthesiologists. Both agents are used routinely in our catheterization laboratory. Dexmedetomidine has been shown to have little effect on SVR and PVR [5,9]. Propofol is known to decrease SVR and PVR [6,3] but was chosen for our study due to the ability to rapidly titrate the dose. Interestingly, the combination of these two agents showed markedly higher SVR and similar PVR to inhaled sevoflurane.

A well-established clinical approach to management of the low SVR state associated with IA is to increase SVR using intermittent bolus dosing of a vasoconstrictor agents such as ephedrine or phenylephrine [10]. This approach is complicated by the need to establish “stable” hemodynamics during all phases of data collection in the catheterization lab. Our data suggest that IA results in relatively significant hypotension during pediatric cardiac catheterization procedures. This hypotension and the need to respond to it by adding vasoactive agents may induce a significant bias to the hemodynamic data obtained. Understanding the age-dependent effects of IA on SVR and blood pressure and the availability of IV alternatives that may have less effect on hemodynamics may help obtain data that more reflect the awake state.

Total intravenous anesthesia may help maintain resting state hemodynamics and inhaled anesthesia significantly lowers systolic blood pressure from baseline. The mechanism of general anesthesia may result in significant hemodynamic changes in pediatric patients undergoing cardiac catheterization. Furthermore, IA leads to underestimation of Qp/Qs and LVEDP as compared with a total intravenous regimen. The magnitude of underestimation of these variables may be age-dependent, and particularly in younger and sicker patients may be of a magnitude that is sufficient to affect clinical decision making. Therefore, future study of anesthetic impact of hemodynamics in neonates and in single-ventricle physiology will be of great importance to our approach to patient care and planning.

The authors have no financial or non-financial conflicts of interest to disclose.

Lam JE, Lin EP, Alexy R, Aronson LA (2015) Anesthesia and the pediatric cardiac catheterization suite: a review. Paediatric anaesthesia 25 (2):127-134. https://doi.org/10.1111/pan.12551
O'Byrne ML, Millenson ME, Steven JM, Gillespie MJ, Dori Y, Glatz AC, Rome JJ (2019) Operator-Directed Procedural Sedation in the Congenital Cardiac Catheterization Laboratory. JACC Cardiovasc Interv 12 (9):835-843. https://doi.org/10.1016/j.jcin.2019.01.224
Pilkington SA, Taboada D, Martinez G (2015) Pulmonary hypertension and its management in patients undergoing non-cardiac surgery. Anaesthesia 70 (1):56-70. https://doi.org/10.1111/anae.12831
Oklu E, Bulutcu FS, Yalcin Y, Ozbek U, Cakali E, Bayindir O (2003) Which anesthetic agent alters the hemodynamic status during pediatric catheterization? Comparison of propofol versus ketamine. J Cardiothorac Vasc Anesth 17 (6):686-690. https://doi.org/10.1053/j.jvca.2003.09.009
Jooste EH, Muhly WT, Ibinson JW, Suresh T, Damian D, Phadke A, Callahan P, Miller S, Feingold B, Lichtenstein SE, Cain JG, Chrysostomou C, Davis PJ (2010) Acute hemodynamic changes after rapid intravenous bolus dosing of dexmedetomidine in pediatric heart transplant patients undergoing routine cardiac catheterization. Anesthesia and analgesia 111 (6):1490-1496. https://doi.org/10.1213/ane.0b013e3181f7e2ab
Akin A, Esmaoglu A, Guler G, Demircioglu R, Narin N, Boyaci A (2005) Propofol and propofol-ketamine in pediatric patients undergoing cardiac catheterization. Pediatric cardiology 26 (5):553-557. https://doi.org/10.1007/s00246-004-0707-4
Schulman DS, Herman BA, Edwards TD, Ziady G, Uretsky BF (1993) Diastolic dysfunction in cardiac transplant recipients: an important role in the response to increased afterload. Am Heart J 125 (2 Pt 1):435-442. https://doi.org/10.1016/0002-8703(93)90023-3
Rodig G, Keyl C, Wiesner G, Philipp A, Hobbhahn J (1996) Effects of sevoflurane and isoflurane on systemic vascular resistance: use of cardiopulmonary bypass as a study model. Br J Anaesth 76 (1):9-12. https://doi.org/10.1093/bja/76.1.9
Chaitanya G, Arivazhagan A, Sinha S, Madhusudan Reddy KR, Thennarasu K, Bharath RD, Rao MB, Chandramouli BA, Satishchandra P (2014) Hemodynamic effects of dexmedetomidine during intra-operative electrocorticography for epilepsy surgery. J Neurosci Rural Pract 5 (Suppl 1):S17-21. https://doi.org/10.4103%2F0976-3147.145190
Lonjaret L, Lairez O, Minville V, Geeraerts T (2014) Optimal perioperative management of arterial blood pressure. Integr Blood Press Control 7:49-59. https://doi.org/10.2147%2FIBPC.S45292

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Choice of Anesthesia Regimen Significantly Impacts Shunting and Diastolic Ventricular Function During Cardiac Catheterization

Status:

Version 1

Abstract

Figures

Introduction

Methods

Results

Discussion

Conclusions

Declarations

References

Additional Declarations

Status:

Version 1