Case reports indicate that nitroglycerin (TNG) may be useful in achieving acute relaxation of the uterus in various emergent (1–8) and nonemergent (9,10) 1 situations. Unfortunately, there are no randomized prospective clinical trials or dose-response studies evaluating the effect of TNG on the uterus during the peripartum period. Indeed, given the potential for TNG to decrease the maternal blood pressure and compromise uterine blood flow, its use might well be questioned without adequate data supporting its administration for uterine relaxation.
No in vivo study has ever demonstrated that TNG decreases the force of uterine contraction, yet clinically the drug produces a perceptible decrease in uterine tone readily appreciated by the obstetrician. An increase in uterine compliance rather than a decrease in contractile force might produce similar changes in the resistance palpable to the obstetrician.
These experiments were undertaken to elucidate the effect of TNG on the peripartum uterus by measuring active tension and compliance separately.
Approval was obtained before the start of our study from the appropriate institutional animal care and use committee. Three experiments were designed to evaluate the effect of TNG on the uterus. Two in vivo experiments, conducted on laboring ewes and rabbits at 2 h postpartum, were used to evaluate the effect of TNG on the uterine contractile force. In vitro experiments using uterine tissue harvested from pregnant rabbits at term were used to assess the effect of TNG on compliance. Both in vivo experiments failed to demonstrate any effect of TNG on uterine contraction, whereas terbutaline arrested uterine contractions completely and halothane reduced contractions in a dose-dependent fashion.
Experiment 1: In Vivo–Postpartum Rabbits
Near term primigravida New Zealand white rabbits (n = 4) were observed every 2 h until delivery. After the entire litter was delivered, mothers were anesthetized with halothane (1%) in oxygen. A 5F catheter (Cook #C-PMS-501J-RSC; Cook Inc, Bloomington, IN) was placed in the maternal femoral artery and vein, and a balloon catheter, fashioned from a finger cot and polyvinyl chloride tubing, was inserted into the uterine cavity through a small flank incision and hysterotomy near the fundus of one uterine horn. Earlier attempts to insert the uterine catheter transvaginally in other animals proved unreliable. Once instrumented, rabbits were allowed to recover from anesthesia for at least 3 h in a box that minimally restricted their movement. Thereafter, an IV infusion of syntocin, 0.3 μU · kg−1 · h−1, was started. Once hemodynamics were stable and uterine contractions were rhythmic (approximately 1 contraction/min) for 5 min, increasing doses of TNG, from 1.0 to 5 mg/kg, were administered through the femoral venous catheter. Maternal blood and intracavitary uterine pressures were recorded throughout. TNG doses were administered at the end of contractions to exclude the possibility that the uterus was refractory to the effect of TNG because of a reduction in uterine blood flow (11). After the last dose of TNG was administered, hemodynamics were allowed to return to baseline. Thereafter, 250 μg of terbutaline was administered as a positive control to ensure that a cessation of uterine contraction could be detected. We further elected to confirm that our preparation was sensitive enough to detect decreases in intracavitary pressure as well as tocolysis. Accordingly, two additional animals were instrumented as previously described and rhythmic uterine contractions induced with syntocin while the animals remained lightly anesthetized at an end-tidal halothane concentration of approximately 0.5%. The administration of TNG again failed to elicit any detectable response in uterine contraction, whereas increasing the end-tidal concentration of halothane to 3.0% reduced and finally arrested uterine contraction. Contractions returned to baseline as the concentration of halothane was decreased back to 0.5%.
Experiment 2: In Vivo–Laboring Sheep
Four time-dated primigravida ewes with singleton pregnancies were anesthetized with halothane at 128 to 132 days of gestation (term = 141 days). Using aseptic technique, a midventral abdominal incision was made and a hysterotomy was performed for placement of a uterine catheter. Polyvinyl chloride catheters were also placed in the maternal femoral artery and vein and, together with the uterine catheters, tunneled subcutaneously, and externalized through a small flank incision. The laparotomy was closed, and the ewe was given 2 GU of penicillin and allowed to recover for 7 days.
At 135 to 139 days of gestation, the ewes were brought to the laboratory in pairs. Maternal blood pressure, heart rate, and uterine pressure were monitored, and after a 1-h control period, an infusion of syntocin, 0.3 μU · kg−1 · h−1, was started. As soon as uterine contractions were well established (occurring more often than one 1 every 2–3 min), increasing doses of TNG (1, 2, 5, 10, 50, 100, 500, and 1000 μg/kg and 5 mg/kg) were administered via the femoral vein. In two sheep, an additional dose of 15 mg/kg was given. After the sheep recovered from the last dose of TNG, 250 μg of terbutaline was administered IV as a positive control to ensure that a cessation in uterine contractions could be detected.
Experiment 3: In Vitro–Term Rabbits
Primigravida New Zealand white rabbits (n = 5) at term (30 to 31 day gestation), were anesthetized with halothane in oxygen, and a total abdominal hysterectomy was performed. The uterus was removed immediately after ligating the blood supply. Once removed, an incision was made along the lateral wall of each uterine horn extending from the fundus to the cervix. The offspring and placenta were removed, and then the entire uterus was immersed in oxygenated, cool (25°C), Krebs-Henseleit Buffer (KHB). The interval between clamping the blood supply and immersion in buffer never exceeded 5 min.
Within 10 min of its immersion in buffer, two ribbons of tissue 2- to 3-mm wide were excised from the cut edges of one uterine horn. These ribbons were then cut into 2-cm lengths. The in situ anatomic orientation of the muscle strips was maintained at all times. A staple was affixed to the fundal end of each strip and the staple laid across the top part of a Y-shaped tension transducer in such a way that the distal end of the strip hung freely between the two ends of the Y mount. Each strip was allowed to relax in warmed (37°C), aerated (95% oxygen/5% carbon dioxide) KHB. Once relaxed, the distal end of the strip was impaled on a post exactly 10-mm from the staple. The strips were again allowed to relax and incubate for another 30 min.
Mounted strips were placed in fresh buffer, and the tissue was pretensioned to 1.0 g and then allowed to incubate for another 30 min. During this time, uterine strips invariably relaxed and the tension was essentially zero by the end of this 30-min period. Any slack remaining in the tissue at this time was removed by stretching until the tension was just barely perceptible (tension ≤ 0.1 g). The strips were not otherwise pretensioned further. Each strip was then maximally contracted with potassium chloride KCl (40 mM). (40 mM KCl represents a supramaximal stimulus for contraction.) Once maximum tension was achieved, the strips were washed in calcium-free KHB-containing sodium glycerol tetra-acetate (EGTA 0.008M) for 30 min.
Only strips that generated a maximal tension of greater than 5 g were used in this study. A matched pair of muscle strips, defined as two strips that were adjacent in situ that generated similar tension (within 1 g ± 10%) was then selected for study. After incubating in calcium-free KHB with EGTA for 30 min, the tissue was repeatedly contracted with 40 mM KCl and washed in calcium-free KHB with EGTA to deplete calcium stores. This cycle of contracting and washing the tissue was performed every 15 min until the tissue no longer contracted (tension < 0.5 g).
Paired strips were then randomized by coin toss for incubation for 30 min with TNG, 10−5 M, or an equal volume of buffer without TNG. We chose a dose of 10 μM of TNG because this concentration has been shown to produce an effect both on blood pressure and the uterus. 2, 3 Any laxity in the tissue strips at the end of the incubation period was removed by stretching the tissue until tension was barely perceptible. After the length of each strip was measured (L0), they were instantaneously stretched to 133% of L0. (The tissue was stretched from L0 to (1.33 L0) to correspond to a fairly linear portion of the compliance curve.) The time-tension curve was recorded and integrated over 10 min. The areas under the curve for the TNG-treated and control strips were compared by using a paired Student’s t-tests.
Once this experiment was complete, any excess tissue below the fixed post was discarded. The strips were removed from the tension transducers, and the staple was trimmed from the strip with fine scissors. The strips were desiccated overnight, and their dry weights were determined the following day. This exact experimental protocol was repeated with 10−4 M S-nitroso-N-acetylpenicillamine (SNAP) instead of 10−5 M TNG in muscle strips from another group of rabbits.
Experiment 1: In Vivo–Postpartum Rabbits
Blood pressure reductions were quite marked at the higher doses of TNG in both awake (n = 4) and lightly anesthetized (n = 2) rabbits (Figure 1). Doses of TNG larger than 10 μg/kg were sufficient to reduce the mean arterial pressure by at least 30% in the anesthetized animals and by approximately 50% in the awake animals. Baseline uterine pressure in rabbits ranged from 12 to 25 mm Hg. TNG did not affect the frequency of uterine contractions or the uterine pressure in lightly anesthetized or awake rabbits. Figure 2 shows a representative recording from one animal. In contrast, increasing the end-tidal halothane concentration from 0.5% to 3% in two anesthetized rabbits caused a rapid decline and then a cessation of uterine contractions (Figure 2). Contractions spontaneously resumed when the end-tidal concentration was returned to 0.5%. Because there never was any detectable uterine response to TNG, we have not reported a tension for each dose of drug.
Experiment 2: In Vivo–Laboring Sheep
The administration of TNG resulted in a reduction in blood pressure in excess of 30% at doses of 5 mg/kg. These declines in blood pressure were accompanied with a reflex increase in heart rate (Figure 3). Despite doses of drug sufficient to reduce the blood pressure, neither the uterine pressure nor the frequency of contraction were affected by TNG. Figure 4 is a representative tracing from one of the sheep experiments. This specific recording is shown because a dose of 5 mg/kg may have produced some minimal transient effect. There was virtually no effect at any of the doses of TNG in any of the other animals, and repeated doses of 5 mg/kg and 15 mg/kg failed to elicit a subsequent decline in tension in this animal. However, a single dose of terbutaline 250 μg IV reliably abolished uterine contractions in all the animals for nearly 20 min (Figure 4) but had no appreciable effect on blood pressure.
Experiment 3: In Vitro–Term Rabbits
Time-tension curves were generated by stretching the uterine tissue to 133% of L0 (resting length). Analog data were recorded from the output of the force transducer and converted to digital data by hand or with a scanner to generate the time-tension curves. The area under the curve was integrated over 10 min.
The dry weight of uterine smooth muscle and L0 were not significantly different for TNG-treated and control strips of muscle, indicating that the mass of tissue and its density were similar in the groups. Because matched strips of tissue used in these experiments had to be adjacent to each other, the muscle fiber orientation in the strips and, therefore, the main vector of force relative to the transducer were similar. Thus, an equivalent maximal force of contraction (response to a supramaximal dose of KCl) should occur with an equivalent muscle mass.
TNG significantly reduced the area under the compliance curve as compared with control strips (P < 0.05) indicating that TNG-treated strips were more compliant than control strips (Table 1). Similar treatment of noncontractile uterine strips with 1 × 10−4 M SNAP failed to produce a significant reduction in the area under the curve.
Rabbits and sheep were chosen because most previous work related to this topic has been in rodent and ovine models. Our initial experiments were conducted in postpartum rabbits because a uterine response to the nitrosovasodilators in general (12), and TNG in particular (1) has been described in postpartum women. Although phasic contractions could be sustained immediately postpartum with a small-dose infusion of syntocin, TNG failed to produce any detectable response in uterine contractions, whereas terbutaline arrested them completely. It occurred to us that, while uterine contractions might be detectable grossly, our preparation might not be sensitive enough to detect small decreases in the force of contractions attributable to TNG. Nonetheless, our preparation was sensitive to changes in the force of contraction because halothane rapidly and reversibly decreased the force of contraction as previously reported in a dose-dependent manner (13) and, at an end-tidal concentration of 3%, completely eliminated contractions.
Because postpartum rabbits might be resistant to TNG, we examined the in vivo effect of TNG on induced uterine contraction by using term pregnant sheep. Kumar et al. (14) had previously demonstrated a questionable uterine response to another nitrosovasodilator, amyl nitrate, in women who had labor induced.
TNG significantly reduced the blood pressure in rabbits and sheep; however, it failed to decrease the frequency of contraction or the intracavitary uterine pressure in either the postpartum rabbit or the laboring sheep. Although the number of animals studied was small, we could not justify continuing these experiments when no response at all was seen at any dose. The absence of a response to TNG in sheep and rabbits does not eliminate the possibility that there may be an effect in humans, considering the many differences in the species involved. However, studies with other nitrosovasodilators failed to demonstrate a decrease in either the force or frequency of contraction in laboring women.
Active tension is the force generated by the contractile elements of a muscle, whereas compliance is the change in tension with respect to length. Because smooth muscle normally contracts when it is stretched (15), the total tension measured at the ends of a muscle is the sum of the active and passive tensions. It is not possible to study passive compliance in preparations of smooth muscle that can generate active tension because stretching the tissue itself is a stimulus for contraction, which obscures the measurement of passive tension. Passive compliance of smooth muscle can only be evaluated if the muscle remains at rest (16,17). Therefore, we studied the effect of TNG on uterine compliance in noncontractile tissue by depleting it of calcium.
Previous in vitro preparations used to evaluate the effect of TNG on the uterus examined contractile force exclusively and have not differentiated between changes in compliance and active tension (18).2,3 Our in vitro experiments assessed the effect of TNG on the compliance of uterine tissue independent of the contractile force of the uterine smooth muscle. These experiments demonstrate that TNG significantly increased the compliance of the gravid rabbit uterus, although higher concentrations of a more potent nitric oxide (NO) donor, SNAP, did not.
The effect of compliance on the tension measured at the ends of a muscle caused by the series and parallel elastic elements is well known. Reductions in tension as measured with a tension transducer, only implies that the force of contraction is reduced if the compliance of the tissue does not change (19). If TNG alters passive compliance, then changes in measured tension may not be attributable to changes in active tension.
Segal and Datta (18) reported that the tension measured in uterine rings harvested from term pregnant rats may be reduced, in a dose-dependent manner, by treatment with TNG. This occurred only in the presence of the placenta, which suggests that a factor derived from the placenta retards the degradation or enhances the production of NO, thus increasing its concentration and relaxing the uterine smooth muscle. This would be similar to the mechanism of action of TNG in vascular tissue and consistent with the lack of any effect on uterine contraction in the postpartum rabbit. However, why term pregnant sheep with intact placentae should also fail to respond to the drug is not clear, although it is not unusual to find that the in vivo response to a drug differs from its effect in vitro. A species difference or the known resistance of the uterus to NO at term (20) and in labor (20,21) may also contribute to this apparent conflict.
However, studies have yet to confirm that placental tissue can increase the effective concentration of NO liberated by TNG or that the effect of TNG on the uterus is even NO-mediated. Some forms of nonvascular smooth muscle are resistant to the effect of TNG, even though they may demonstrate a vigorous response to NO (22), whereas other types of smooth muscle respond to TNG in a manner that does not involve the NO/cyclic guanosine monophosphate (cGMP) cascade (23). In any tissue containing nonvascular smooth muscle, the relationship between TNG and NO may not be straightforward. Thus, it may be premature to conclude that TNG relaxes the uterus by a NO/cGMP-mediated mechanism. Diamond and Marshall (23) have shown that TNG relaxes rat uterus but does not increase cGMP levels. SNAP is a potent spontaneous donor of NO, which should relax the smooth muscle even in the absence of the placenta if the action of TNG was NO-mediated.
In every case in which TNG has been reported to be effective, the obstetrician has perceived its effect clinically. This has involved the application of an external force by the obstetrician to perform a version (7), dilate the cervix (9), 1 or facilitate delivery (8). In short, the obstetrician subjectively senses a change in the compliance of the uterus as an external force is applied. Decreases in tension perceived by the clinician may not represent muscle relaxation. TNG neither alters the frequency of contraction nor the active tension in the uterine smooth muscle in sheep or rabbits. Clearly, species differences preclude discounting such an effect in humans, although other nitrosovasodilators have failed to produce uterine relaxation in laboring women (14). However, TNG does increase uterine compliance, and that may be the reason that uterine muscle appears to “relax” after its administration. Because SNAP had no effect on uterine compliance, TNG appears to increase uterine compliance through mechanisms that are independent of the NO/cGMP cascade. TNG increases uterine compliance, apparently through mechanisms that do not involve the generation of active tension, because increased compliance is evident in calcium-depleted noncontractile tissue.
1Cauldwell CB, Rosen MA, Hanison MR. The use of nitroglycerin for uterine relaxation during fetal surgery [abstract]. Anesthesiology 1995;83:A929.
2Chen B, Hou X, Kwan W, Lee C. Comparison of uterine relaxation between nitroglycerin and ritodrine in the rat [abstract]. Anesthesiology 1993;79:A995.
3Shin Y, Wang K, Kim Y, et al. Effect of nitroglycerin on the pregnant uterine muscle in vitro. A dose response study [abstract]. Anesth Analg 1994;78:S391.
1. Peng AT, Gorman RS, Shulman SM, et al. Intravenous nitroglycerin for uterine relaxation in the postpartum patient with retained placenta [letter]. Anesthesiology 1989; 71:172–3.
2. DeSimone CA, Norris MC, Leighton BL. Intravenous nitroglycerin aids manual extraction of a retained placenta [letter]. Anesthesiology 1990; 73:787.
3. Rolbin SH, Hew EM, Bernstein A. Uterine relaxation can be life-saving [letter]. Can J Anaesth 1991; 38:939–40.
4. Mayer DC, Weeks SK. Antepartum uterine relaxation with nitroglycerin at Caesarean delivery. Can J Anaesth 1992; 39:166–9.
5. Altabef KM, Spencer JT, Zinberg S. Intravenous nitroglycerin for uterine relaxation of an inverted uterus. Am J Obstet Gynecol 1992; 166:237–8.
6. Abouleish AE, Corn SB. Intravenous nitroglycerin for intrapartum external version of the second twin [letter]. Anesth Analg 1994; 78:508–9.
7. Greenspoon JS, Kovacic A. Breech extraction facilitated by glyceryl trinitrate sublingual spray [letter]. Lancet 1991; 338:124–5.
8. Chan ASH, Ananthanarayan C, Rolbin SH. Alternating nitroglycerin and syntonin to facilitate uterine exploration and removal of an adherent placenta. Can J Anaesth 1995; 42:335–7.
9. Lees C, Campbell S, Jauniaux E, et al. Arrest of preterm labour and prolongation of gestation with glyceryl trinitrate, a nitric oxide donor. Lancet 1994; 343:1325–6.
10. Clavin DK, Bayhi DA, Nolan TE, et al. Comparison of intravenous magnesium sulfate and nitroglycerin for preterm labor: preliminary data [abstract]. Am J Obstet Gynecol 1996; 174:307.
11. Greiss F Jr : Uterine blood flow during labor. Clin Obstet Gynecol 1968; 11:96.
12. Barnes F. Hour-glass contraction of the uterus treated with nitrite of amyl. BMJ 1892; 18:377.
13. Vasicka A, Kretchmer H. Effect of conduction and inhalation anesthesia on uterine contractions: experimental study of the influence of anesthesia on intraamniotic pressures. Am J Obstet Gynecol 1961; 82:600–11.
14. Kumar D, Zourlas PA, Barnes AC. In vivo effect of amyl nitrite on human pregnant uterine contractility. Am J Obstet Gynecol 1965; 91:1061–8.
15. Bulbring E. Smooth muscle potential recorded in the taenia coli of the guinea pig. J Physiol 1954; 123:55–6.
16. Dean GE, Cargill RS III, Macarak E, et al. Active and passive compliance of the fetal bovine bladder. J Urol 1997; 158:1094–9.
17. Kim P, Sundt TM Jr, Vanhoutte PM. Alterations of mechanical properties in canine basilar arteries after subarachnoid hemorrhage. J Neurosurg 1989; 71:430–6.
18. Segal S, Datta S. Placental tissue markedly enhances uterine relaxation by nitroglycerin [abstract]. Anesthesiology 1996; 85:A879.
19. Pollack GH. Maximum velocity as an index of contractility in cardiac muscle. A critical evaluation. Circ Res 1970; 26:111–27.
20. Izumi H, Yallampalli C, Garfield RE. Gestational changes in L-arginine-induced relaxation of pregnant rat and human myometrial smooth muscle. Am J Obstet Gynecol 1993; 169:1327–37.
21. Yallampalli C, Izumi H, Byam-Smith M, Garfield RE. An L -arginine-nitric oxide-cyclic guanosine monophosphate system exists in the uterus and inhibits contractility during pregnancy. Am J Obstet Gynecol 1994; 170:175–85.
22. Hussain AS, Poklewska-Koziel M, Brien JF, et al. Nitric oxide formation from glyceryl trinitrate by rabbit aortic strip: detection by rabbit taenia coli concurrent with vasorelaxation. Br J Pharmacol 1994; 113:1508–14.
© 2000 International Anesthesia Research Society
23. Diamond J, Marshall JM. A comparison of the effects of various smooth muscle relaxants on the electrical and mechanical activity of rat uterus. J Pharmacol Exp Ther 1969; 168:21–30.