| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
From the *Department of Anaesthesia, Clinical Sciences Institute, National Centre for Biomedical Engineering Sciences, National University of Ireland, and Public Analysts Laboratory, Newcastle, Galway, Ireland.
Address correspondence and reprint requests to Prof. John G. Laffey, MD, MA, BSc, FCARCSI, Department of Anaesthesia, Clinical Sciences Institute, National University of Ireland, Galway, Ireland. Address e-mail to john.laffey{at}nuigalway.ie.
Abstract
BACKGROUND: The effects and mechanisms of action of volatile anesthetics on the feto-placental vasculature are not known. We aimed to quantify the vasoactive effects of sevoflurane and determine the role of nitric oxide (NO) and of vasoactive eicosanoids in mediating these effects in isolated human chorionic plate arterial rings.
METHODS: Quadruplicate ex vivo human chorionic plate arterial rings were used in all studies. Series 1 quantified the vasodilation produced by sevoflurane in rings preconstricted with the thromboxane analog U46619. Series 2A–C examined the role of NO in sevoflurane-mediated vasodilation. In separate experiments, we examined the potential for the nonspecific NO inhibitors, L-NAME, L-nMMA, and the inactive D-NAME, to modulate the vasodilation produced by sevoflurane. Series 2D determined whether sevoflurane altered vascular smooth muscle sensitivity to exogenous NO. Series 3A–D examined the role of vasoactive eicosanoids in sevoflurane-mediated vasodilation. In separate experimental series, we examined whether the nonspecific cyclooxygenase inhibitor, indomethacin, or the 5-lipoxygenase inhibitor, nordihydroguaiaretic acid, modulated sevoflurane-mediated vasodilation.
RESULTS: Sevoflurane produced dose-dependent vasodilation of preconstricted chorionic plate arterial rings, with mean ring vasodilation increasing from 15 ± 7% at 2% sevoflurane to 67 ± 17% (mean ± sd) at 8% sevoflurane. Blockade of NO synthase did not attenuate the vasodilator effects of sevoflurane. Sevoflurane did not alter smooth muscle sensitivity to NO. Indomethacin augmented sevoflurane vasodilation at 10–5 M, but not at 10–6 M. Conversely, nordihydroguaiaretic acid attenuated sevoflurane-mediated vasodilation at 3 x 10–6 M but not at 3 x 10–7 M.
CONCLUSIONS: Sevoflurane was a vasodilator in the feto-placental vasculature in this in vitro model. Sevoflurane-mediated vasodilation is NO and cyclooxygenase-independent and appears to be mediated in part via a lipoxygenase generated vasodilator eicosanoid.
A substantial number of patients require general anesthesia during pregnancy. Indications include the need to perform nonobstetric surgery during pregnancy and general anesthesia for cesarean delivery. In future years, the need for general anesthesia during pregnancy is likely to increase due to the continuing development of in utero fetal surgical techniques.1 The use of sevoflurane has gained widespread acceptance in obstetric anesthesia2–4 and has been demonstrated to have an excellent recovery profile in this population.3,4
The effects and mechanisms of action of volatile anesthetics such as sevoflurane on the feto-placental vasculature are not known and are potentially important. Volatile anesthetics have been demonstrated to cause vasodilation in most, but not all, systemic vascular beds, with the precise mechanism contributing to the vasoactive effects specific to the vascular system studied. The potential mechanisms underlying the vasodilatory effect in this circulation include alterations in nitric oxide (NO) release or sensitivity5–9 and/or alterations in vasoactive eicosanoid activity.10
We therefore hypothesized that sevoflurane would vasodilate the feto-placental circulation by a mechanism involving alterations in NO release or sensitivity and/or by modulation of vasoactive eicosanoid activity.
METHODS
After ethical committee approval, and written informed patient consent, term placentae were obtained after both vaginal delivery and elective or emergency cesarean delivery under neuraxial anesthesia from healthy parturients. None of the patients from whom the samples were taken received general anesthesia for delivery. Exclusion criteria included preexisting hypertension, intrauterine growth retardation, preeclampsia, and patients with pregnancy-induced hypertension, hepatitis, and human immunodeficiency virus infection.
All studies, conducted in nine separate series of experiments, followed a randomized, controlled, paired design. Umbilical arteries and their branches were identified as they spread out onto the chorionic plate of the placenta. Samples of the second-generation (second-order) chorionic plate arteries were taken within 120 min of delivery and placed directly into ice-cold pyrogen-free physiologic saline solution (PSS) (122.6 mM NaCl, 5.4 mM KCl, 20 mM NaHCO3, 0.8 mM MgSO4, 0.9 mM Na2HPO4, 2.4 mM CaCl2, and 5.5 mM glucose).
Four chorionic rings, each 3 mm in length, from a single artery from each placenta, were isolated, mounted in PSS at 37°C, and equilibrated with 95% O2–5% CO2 in tissue baths (10-mL capacity) throughout all experiments. Samples of the solution were intermittently analyzed for Po2, Pco2, and pH, using an automated blood–gas analyzer (Radiometer ABL 500). Rings were threaded onto a horizontally fixed platinum surgical wire (300-µm diameter). A second hook, connected to an isometric force transducer, was then passed through the lumen of the ring. Isometric tension was recorded as a function of time using a transducer system (Grass FT03, Quincy, MA).
Baseline Interventions
After a 60-min equilibration period, an optimal pretension of 2 g for each chorionic plate arterial ring was used. These pretensions were determined in preliminary experiments in our laboratory to elicit active tensions of 80%–100% of the maximum in each ring. Priming contractions were induced by exposing the rings three times to 80 mM KCl solution (iso-osmotically substituted for NaCl), and pretension reestablished at the end of each 5 min exposure by rinsing with PSS.
Experimental Design and Interventions
Preconstriction of rings. In each series, after the baseline interventions, the rings were submaximally preconstricted. In series 1, 2A–C and 3A–D, the rings were submaximally preconstricted with the thromboxane analog U46619 (9,11-dideoxy-11alpha, 9alpha-epoxymethanoprostagladin F2, 5 x 10–6 M). This concentration of U46619 was demonstrated in preliminary experiments to reliably produce a contraction of approximately 70% of the maximum amplitude obtainable. In series 2D, prostaglandin F2
(PGF2)was used to preconstrict the rings, as described below. Once a stable plateau contractile response was obtained, the rings were allowed to remain at the plateau tension for 30 min.
Series 1 examined the direct effects of sevoflurane on preconstricted chorionic plate arterial rings. After preconstriction with U46619, two rings from each placenta were randomly assigned to receive sevoflurane (2%–8%) in the perfusing gas, cumulatively attained by increasing sevoflurane concentration in 2% increments at 30 min intervals, with the remaining two rings serving as controls. The effects of sevoflurane at each sevoflurane increment were measured by calculating the mean amplitude of the ring tension over the final 5 min of each 30 min interval.
Series 2 examined the role of NO release or sensitivity in mediating the vasodilatory effect of sevoflurane in preconstricted chorionic plate arterial rings. Series 2A examined the potential for 
-nitro-l-arginine methyl ester (L-NAME), a nonspecific inhibitor of NO synthase (NOS), to attenuate the vasodilation produced by sevoflurane. After U46619 preconstriction, two rings from each placenta were randomly assigned to receive sevoflurane in the perfusing gas, with the remaining two rings serving as controls. L-NAME (1 x 10–3 M) was first added to one sevoflurane exposed bath and one control bath and allowed to equilibrate for 20 min, and vehicle added to the remaining baths. This resulted in four groups: (1) control with vehicle (control); (2) control with L-NAME (L-NAME); (3) sevoflurane with vehicle (SEVO); (4) sevoflurane with L-NAME (SEVO + L-NAME). Increasing perfusate sevoflurane concentrations from 2% to 8% were then cumulatively attained at 30 min intervals. Using this methodology, Series 2B examined the effect of -nitro-d-arginine methyl ester (D-NAME), the inactive analog of L-NAME, whereas Series 2C examined the potential for NGMonomethyl L-arginine (L-nMMA), a second nonspecific NOS inhibitor, to attenuate the vasodilation produced by sevoflurane. These experiments were performed in order to verify that NOS inhibition did not play any role in sevoflurane-mediated vasodilation.
Series 2D examined the potential for sevoflurane to alter smooth muscle responsiveness to NO. This was examined by determining the dose–response relationship to the NO donor, sodium nitroprusside (SNP), for concentrations ranging from 10–10 to 10–7 M in half-log increments. PGF2 was used to produce vasoconstriction in this experiment, because in preliminary experiments in our laboratory, it produced a contraction that was more reproducibly dose–responsive than that obtained with U46619. Two rings from each placenta were randomly assigned to receive 8% sevoflurane in the perfusing gas (95% O2–5% CO2) or perfusing gas alone. After 40 min, a 6 point PGF2 cumulative concentration response curve (0.25–5.0 x 10–6 M) was obtained and the dose that produced a 50% contraction (EC50c) determined for each bath, in the presence and absence of sevoflurane as appropriate. Baseline tension was restored by rinsing with PSS. After submaximal ring contraction with the EC50c concentration of PGF2, the effect of sevoflurane on NO-mediated vasodilation was assessed using a five point cumulative concentration response curve to SNP.
Series 3 examined the role of vasoactive eicosanoids in mediating the vasodilatory effect of sevoflurane in preconstricted chorionic plate arterial rings. Series 3A examined effect of indomethacin (1 x 10x6 M), a nonspecific inhibitor of cyclooxygenase, on sevoflurane-mediated vasodilation. Series 3B examined the effect of higher concentrations of indomethacin (1 x 10–5 M). In each series, after U46619 preconstriction, two rings were randomly assigned to receive sevoflurane, with the remaining two rings serving as controls. Indomethacin was first added to one sevoflurane-exposed bath and one control bath and vehicle added to the remaining baths. Increasing perfusate sevoflurane concentrations from 2% to 8% were then cumulatively attained at 30 min intervals.
Series 3C examined the effect of nordihydroguaiaretic acid (NDGA, 3 x 10–7 M), an inhibitor of lipoxygenase, whereas Series 3D examined effect of higher concentrations of NDGA (3 x 10–6 M) on sevoflurane-mediated vasodilation. After U46619 preconstriction, two rings were randomly assigned to receive sevoflurane, with the remaining two rings serving as controls. NDGA (3 x 10–6 M) was first added to one sevoflurane-exposed bath and one control bath and vehicle added to the remaining baths. Increasing perfusate sevoflurane concentrations from 2% to 8% were then cumulatively attained at 30 min intervals.
End Protocol Contractility Assessment
For each series, at the end of the experiment, the volatile anesthetic was discontinued and pretension was reestablished by rinsing with PSS. Where necessary, 10–7 M SNP was added to baths to produce full vasodilation, and rinsed with PSS. The contractile response to 80 mM KCl was then reassessed to assess performance over the course of the experiment. A final exclusion criterion was applied at this point, with data from all four rings excluded from analysis if the maximum final KCl-induced contraction in controls was <90% of the initial KCl response.
Determination of Bath Sevoflurane Concentrations
The sevoflurane was introduced into the tissue baths in the perfusing gas via a sevoflurane vaporizer (Abbott Inc., Dublin, Ireland). The vapor content of anesthetic in the carrier gas was continuously measured by means of an in-line agent monitor (Datex Capnomac Ultima®, Helsinki, Finland). The perfusing gas was continuously sampled downstream from the vaporizer, at a point just proximal to the entry of the gas into the tissue baths. Volatile anesthetic concentrations increased within 10 min to within 10% of the value dialed on the vaporizer in all experiments.
In addition, in a separate series of experiments, the bath concentrations of sevoflurane were determined using gas chromatography-mass spectrometry, for sevoflurane concentrations in the perfused gas of 0%–8%, as previously described.11,12 Samples were taken in duplicate, at 30 and 60 min after alteration of the sevoflurane vaporizer setting. These were compared with a standard curve generated by adding known volumes of sevoflurane to a known volume of PSS. Bath concentrations of sevoflurane increased to more than 90% of the predicted value based on its partition coefficient at 30 min. No further change was detected in bath sevoflurane concentrations at 60 min, confirming the existence of steady-state conditions at 30 min.
Chemicals
U46619, PGF2, L-NAME, D-NAME, LnMMA, SNP, indomethacin, and NDGA and all salts were all purchased from Sigma Aldrich (Poole, Dorset, UK). The volatile anesthetic agents were purchased from Abbott Ireland (Dublin, Ireland).
Statistical Analysis
Data are presented as means ± sd, with vasodilation by sevoflurane and SNP expressed as a percentage of the submaximal contraction elicited with U46619 or PG F2. The EC50r value, which represents the concentration of sevoflurane or SNP required to dilate the preconstricted rings by 50%, is calculated for each ring, in each experimental series. First, the slope and intercept of the line connecting the sevoflurane values immediately below and above 50% is determined. The EC50r value is calculated by the following formula: EC50r = 10X where X = (50 – Intercept)/ Slope. Comparison between control and test rings was made using two-way repeated measures analysis of variance, with group as a between subjects factor and percentage sevoflurane as a repeated measures factor. Between group analyses were restricted to comparisons relevant to our a priori hypotheses, and were made using Student's t testing with corrections for multiple comparisons. The null hypothesis was rejected for P < 0.05.
RESULTS
Chorionic plate arterial rings were obtained from 80 normal term placentae for these studies. Stable and comparable gas tensions were maintained throughout all experiments, and comparable baseline levels of contractile responses were observed in all series. Postintervention responses to potassium chloride were not different compared to baseline for any rings studied, and therefore no rings were excluded.
Series 1—Effect of Sevoflurane
Sevoflurane dose-dependently dilated chorionic arterial rings (n = 10 placentae; 20 rings per group) after U46619 preconstriction. Mean vasodilation increased from 15 ± 7% at 2% sevoflurane to 67 ± 17% (±sd) with 8% sevoflurane [Figure 1]. The concentration of sevoflurane that produced a 50% relaxation (EC50r) was 5.8 ± 2.0%. Sevoflurane resulted in significant vasodilation over control conditions at each concentration studied (P < 0.001).
|
Series 2—Role of NO
Series 2A examined the potential for the NOS inhibitor L-NAME to modulate sevoflurane-mediated vasodilation (n = 10 placentae; n = 10 rings per group). There was minimal vasodilation in the rings exposed to control conditions or L-NAME alone. L-NAME did not alter did not alter the response of the rings to sevoflurane. There was no significant difference in the maximal dilation with 8% sevoflurane (68 ± 14% vs 52 ± 12%, P = 0.1) [Figure 2, Panel A], or in the EC50r (6.2 ± 1.2% vs 7.2 ± 1.4%, P = 0.1) in the presence and absence of L-NAME, respectively.
|
Series 2B examined the potential for the inactive isomer D-NAME to modulate sevoflurane-mediated vasodilation (n = 10 placentae; n = 10 rings per group). There was minimal vasodilation in the rings exposed to control conditions or D-NAME alone. D-NAME did not modulate sevoflurane-mediated vasodilation. There was no difference in maximal dilation with 8% sevoflurane (64 ± 17% vs 68 ± 30%, P = 0.8), or in the EC50r (6.6 ± 1.0% vs 6.4 ± 0.8%, P = 0.8) in the presence and absence of D-NAME, respectively.
Series 2C was performed in order to verify that inhibition of NOS, using a second NOS inhibitor, L-nMMA, did not modulate sevoflurane-mediated vasodilation (n = 10 placentae; n = 10 rings per group). There was minimal vasodilation in the rings exposed to control conditions or L-nMMA alone. L-nMMA did not modulate sevoflurane-mediated vasodilation. There was no difference in maximal dilation with 8% sevoflurane (73 ± 22% vs 71 ± 24%, P = 0.9), or in the EC50r (6.2 ± 2.0% vs 6.4 ± 1.6%, P = 0.8) in the presence and absence of L-nMMA, respectively.
Series 2D examined the potential for sevoflurane to alter vessel sensitivity to exogenous NO (n = 10 placentae; n = 10 rings per group). SNP dose-dependently dilated preconstricted rings, with a maximal vasodilation of 100% at 10–7 M SNP [Figure 2, Panel B]. Sevoflurane did not alter the response of the rings to SNP, and there was no difference in the EC50r in the presence and absence of sevoflurane (3 ± 3 x 10–9 M vs 2 ± 2 x 10–9 M, P = 0.4) respectively.
Series 3—Role of Vasoactive Eicosanoids
Series 3A and B examined the role of cyclooxygenase-derived prostanoids to modulate sevoflurane-mediated vasodilation. In series 3A, pretreatment with 1 x 10–6 M indomethacin (n = 5 placentae; n = 5 rings per group) did not increase vasodilation compared with control conditions, and did not alter sevoflurane-mediated vasodilation. In series 3B, pretreatment with 1 x 10–5 M indomethacin (n = 10 placentae; n = 10 rings per group) did not increase vasodilation compared with control conditions (17 ± 8% vs 12 ± 10% respectively, P = 0.3). Indomethacin 1 x 10–5 M significantly augmented sevoflurane-mediated vasodilation [Figure 3, Panel A]. Indomethacin pretreatment significantly decreased the EC50r for sevoflurane (3.8 ± 1.6% vs 7.1 ± 1.0%, P < 0.001).
|
Series 3C and D examined the role of lipoxygenase-derived eicosanoids to modulate sevoflurane-mediated vasodilation. In series 3C, pretreatment with 3 x 10–7 M NDGA (n = 5 placentae; n = 5 rings per group) did not alter sevoflurane-mediated vasodilation. In series 3D, 3 x 10–6 M NDGA significantly attenuated sevoflurane-mediated vasodilation (n = 10 placentae; n = 10 rings per group) [Figure 3, Panel B], and significantly increased the EC50r for sevoflurane (5.9 ± 1.0% vs 7.1 ± 1.2%, P < 0.04).
DISCUSSION
General anesthesia is required in many patients during pregnancy for a variety of indications. Volatile anesthetics are used during general anesthesia for cesarean delivery to reduce awareness, notwithstanding their uterine-relaxant effects.13 The need for general anesthesia during pregnancy is likely to increase in future years due to the emergence of in utero fetal surgical techniques, such as the ex utero intrapartum treatment procedure.1 The successful use of sevoflurane to provide anesthesia for the ex utero intrapartum treatment procedure has been reported.14
Several studies demonstrate an association between surgical procedures during pregnancy and adverse neonatal outcome.15–18 However, difficulties in separating the contributions of maternal disease, surgical interventions and anesthetic techniques mean that it is unclear whether the anesthetic contributes to these adverse outcomes. Nevertheless, there is the potential for volatile anesthetics to adversely affect feto-placental blood flow and contribute to adverse neonatal outcome.19,20 This could occur either by a direct effect on placental vasomotor tone, such as by increasing feto-placental vascular resistance, or by diverting blood flow to other vascular beds in the absence of an effect on the feto-placental vasoactivity. These concerns are supported by the experimental finding, in the pregnant ewe, that isoflurane rapidly crossed the placenta, appeared in the fetal circulation within 2 min, reaching levels near that seen in the maternal circulation, and produced a significant fetal acidosis consistent with the development of fetal hypoxia.20 However, there are no data to our knowledge characterizing the direct effects of volatile anesthetics on feto-placental vascular tone.
Our findings demonstrate, for the first time, that sevoflurane produces dose-dependent vasodilation in the feto-placental circulation in vitro. Sevoflurane produced vasodilation of preconstricted chorionic plate arterial rings, with mean ring vasodilation of 15% with 2% sevoflurane, which increased to 67% with 8% sevoflurane. While concentrations of 2% sevoflurane are commonly used in patients undergoing cesarean delivery under general anesthesia,3,4 and 4% to 5% sevoflurane has been used in in utero fetal surgery,1 caution must be exercised in extrapolating from these findings in an in vitro model to the clinical context.
Volatile anesthetics have been demonstrated to vasodilate via NO-dependent mechanisms in other vascular beds.10 In addition, NO is intimately involved in regulation of basal vascular tone in the feto-placental circulation,21 whereas alterations in NO may contribute to the pathogenesis of key disorders of pregnancy.22,23 These findings raised the potential that sevoflurane might produce vasodilation via either an increase in endothelial NO production or by increasing vascular smooth muscle sensitivity to NO. Our findings demonstrate that sevoflurane did not produce vasodilation by either mechanism. Blockade of human chorionic plate arterial ring NO production by L-NAME did not attenuate sevoflurane-induced vasodilation of preconstricted rings. The series using D-NAME, the inactive isomer of L-NAME, and a second NOS inhibitor, L-nMMA, confirm that alterations in NO production did not contribute to the vasoactive effects of NO. In series 2D, in rings precontracted to the same degree, preincubation with 8% sevoflurane did not alter the degree of relaxation produced by SNP, an NO donor. Therefore, sevoflurane does not alter ring sensitivity to NO.
Vasoactive prostanoids are also intimately involved in regulation of placental vascular tone, and feto-placental endothelial cells have been demonstrated to release the vasodilator prostaglandin prostacyclin (PGI2).21 Alterations in the balance between prostanoid vasoconstrictors and vasodilators, with increased placental production of thromboxane and decreased prostacyclin production, are central to the pathogenesis of vascular endothelial dysfunction in preeclampsia.24 Furthermore, volatile anesthetics have been demonstrated to produce diverse effects via eicosanoid-dependent mechanisms.25
Our findings demonstrate that sevoflurane did not produce vasodilation via a production of a cyclooxygenase-generated prostanoid. In fact, indomethacin-induced inhibition of cyclooxygenase augmented sevoflurane-mediated vasodilation. This suggests that sevoflurane activates the production of a vasoconstrictor prostanoid, perhaps thromboxane A2 or PGF2. Alternatively, sevoflurane may reduce prostacyclin production and release in the feto-placental circulation, as has been demonstrated in human umbilical vein endothelial cells.26 However, this is less likely, given that inhibition of cyclooxygenase produces increased sevoflurane-mediated vasodilation.
In contrast, inhibition of the 5-lipoxygenase enzyme attenuated sevoflurane-mediated vasodilation. This suggests that sevoflurane releases a lipoxygenase generated vasodilator, which mediates, in part, its vasodilator effect. 5-lipoxygenase is a key enzyme in the metabolism of arachidonic acid, catalyzing both the oxygenation of arachidonic acid to (5S)-hydroperoxy-6,8,11,14-eicosatetraenoic acid and its subsequent conversion to leukotrienes. Leukotrienes are potent mediators of numerous biological processes, including chemotaxis, vascular permeability, and alterations in vasomotor tone. Taken together, these finding demonstrate that sevoflurane may result in the generation of eicosanoids with both vasoconstrictor and vasodilator effects, and that the net effect of these eicosanoids may be to vasodilate the feto-placental vasculature.
There are some limitations to this study. First, these studies were conducted in second-order chorionic plate arteries. Characterization of the effects of sevoflurane on smaller vessels in the feto-placental circulation (i.e., placental resistance arteries) is also required. Second, chorionic plate arterial ring samples were obtained from healthy parturients. Characterization of the effects of sevoflurane in the setting of compromised uteroplacental circulation, in which fetal hypoxia may be more frequent, would add further useful information. Third, NDGA, which was used to determine the role of 5-lipoxygenase in mediating the vasorelaxant effect of sevoflurane, has less specific effects, such as tyrosine kinase inhibition. However, these effects are only seen at concentrations 10–100 times that used in these studies.27 Fourth, our findings may not apply to other inhaled anesthetics, especially in a quantitative manner. The effects of other volatile anesthetics will need to be separately determined. Finally, there are limitations in extrapolating from in vitro experiments to the in vivo situation. The degree of vessel preconstruction used in these experiments may not reflect the usual clinical condition. In addition, there was significant variability among experimental series. Although this is not unexpected in a biologic model, it does limit the power of the studies, and makes verification of the results in preclinical models essential.
In conclusion, our findings demonstrate that sevoflurane exerts a dose-dependent vasodilatory effect in the feto-placental vasculature in this isolated placental vessel model. These effects are mediated via a mechanism that is NO-independent and which is mediated, at least in part, via lipoxygenase-generated vasoconstrictor prostanoids. Further study is required to determine the clinical significance of these findings.
ACKNOWLEDGMENTS
The authors extend their gratitude to the staff of the delivery suite at Galway University Hospital, for their help and cooperation in obtaining placental tissue.
Footnotes
Accepted for publication March 5, 2008.
Supported by the Millennium Fund, NUI Galway, Ireland, and the Yamanouchi European Foundation.
The first two authors contributed equally to this paper.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
