Lack of reactivity of uterine arteries from patients with

obstetric hemorrhage

Sharon H. Nelson, PhD, and Maya S. Suresh, MD

Galveston, Texas

Obstetric hemorrhage may occur throughout pregnancy and the puerperium. The purpose of this studywas to investigate the reactivity of isolated, suffused uterine arteries from obstetric patients withuncontrollable uterine bleeding and to compare those blood vessels with uterine arteries from patientsundergoing cesarean hysterectomy for other medical reasons (control patients). The uterine arteries fromthe control patients (n = 9) responded with maximal or near-maximal constriction to norepinephrine(30 urnol/L, 3.6 ± 1 gm), potassium chloride (75 mrnol/L, 10.2 ± 3 gm), prostaglandin F2a (30 ILmol/L,1.8 ± 1 gm), and arginine vasopressin (1 IJomol/L, 18.8 ± 2.6 gm). In uterine arteries from five patientswith uncontrollable bleeding, the constrictor responses to the same drugs were markedly depressed:norepinephrine (30 urnol/L, 0.5 ± 0.2 gm), potassium chloride (75 mmol/L, 1.9 ± 0.8gm); prostaglandinF2• (30 urnol/L, 0 gm), and arginine vasopressin (1 urnol/L, 0.2 ± 0.05 gm). Uterine arteries from twopatients exhibited no constrictor responses to norepinephrine (30 IJomoI/L), potassium chloride (75mmoI/L), prostaglandin F2a (30 IJomoI/L), or arginine vasopressin (1 IJomoI/L). The impaired responses tothe vasoconstrictor drugs were not reversed by indomethacin (1 IJomoI/L), which is an inhibitor ofprostaglandin synthetase; methylene blue (10 IJomoI/L), which is a blocker of endothelium-derived relaxingfactor activation of guanylate cyclase; or propranolol (1 IJomoI/L), a l3-adrenergic receptor antagonist. Thelevels of adenosine 3' :5'-cyc"c rnonophosphate were not elevated in the uterine arteries from the patientswith obstetric hemorrhage.The impaired reactivity to the multiple vasoconstrictors implies that amechanism involved in constriction common to all of the constrictors is depressed or blocked.Furthermore, the depression or lack of reactivity of these isolated uterine arteries is not mediated byvasodilatory prostaglandins, endothelium-derived relaxing factor, l3-adrenergic receptors, or elevated levelsof adenosine 3' :5'-cyclic monophosphate. The results suggest that obstetric hemorrhage involves, in part,a lack of constrictor reactivity of the uterine vasculature. (AM J OasTET GVNECOL 1992;166:1436-43.)

Key words: Vascular smooth muscle, norepinephrine-induced constriction, potassiumchloride-induced constriction, obstetric hemorrhage, pregnancy

Obstetric hemorrhage may occur throughout preg­nancy and the puerperium. Postpartum hemostasis isgenerally thought to result from uterine muscle con­traction and retraction and from thrombosis and fi­brosis. Causes of the loss of this hemostasis, which re­sults in uncontrollable uterine bleeding, include uterineatony, trauma to the genital tract and adjacent struc­tures during labor and delivery, retained products ofconception, coagulation defects, and abnormal placen­tation. 1 The management in cases of hemorrhage in­cludes obstetric and pharmacologic maneuvers such asuterine massage, exploration and curettage of theuterus, surgical repair of traumatized tissue, and ad­ministration of ergots, oxytocin, prostaglandins, and

From the Department ofAnesthesiology,The University ofTexas Med­ical Branch.Supported byNational Institutes of Health grant HL 38876-01.Received for publication April 23, 1991; revised October30, 1991;accepted November 7,1991.Reprint requests: Sharon H. Nelson, PhD, Department of Anesthe­siology, The University of Texas Medical Branch, Galveston, TX77550.611134897

1436

other pressor drugs, as well as the administration ofelectrolytes, blood, and blood components.' If thesemeasures are unsuccessful, emergency hysterectomymay be required.

Although a number of factors are associated withobstetric hemorrhage, virtually nothing is known aboutthe reactivity of the uterine vasculature during uncon­trollable uterine bleeding of pregnant patients. Uncom­plicated (normal) pregnancy has been shown to havelittle or no effect on the responses of the human uterineartery to various constrictor agents.v " The possibilityexists that the uterine vasculature of pregnant patientswith uncontrollable uterine bleeding is depressed andthus does not constrict sufficiently for hemostasis. Thepurpose of this study was to investigate the reactivityof uterine arteries from patients with obstetric hem­orrhage who underwent cesarean hysterectomy and tocompare those blood vessels with uterine arteries frompatients who underwent cesarean hysterectomy forother medical reasons (control patients). The use ofuterine arteries obtained exclusively from pregnant pa­tients avoided confounding hemodynamic, humoral,

Volume 166Number 5

and neurogenic influences of the nonpregnant con­dition.

A number of mechanisms responsible for the uterinevasodilatation that occurs during normal pregnancyhave been investigated. For example, studies haveshown that pregnancy increases the release from theendothelium of vasodilator substances such as endo­thelium-derived relaxing factor in the uterine arteryfrom humans' and guinea pigs" or prostacyclin in theuterine artery from sheep." Other studies suggest analteration in calcium channels or intracellular calciumstores in the isolated uterine artery from the pregnanthuman." Drug-induced increases in levels of adenosine3': 5'-cyclic monophosphate (cAMP) are associated withrelaxation of smooth muscle, and elevated basal levelsof cAMP have been reported in uterine arteries frompregnant ewes.' Any of these mechanisms involved indecreasing uterine vascular tone during normal preg­nancy may play an even greater role in uterine arteriesfrom patients with obstetric hemorrhage.

Material and methods

Specimens of the ascending branch of the uterineartery were obtained from patients undergoing cesar­ean hysterectomy for various medical reasons. Use ofthe uterine arteries from patients undergoing hyster­ectomy was approved by the Institutional Review Boardfor the University of Texas Medical Branch, Galveston,Tex. Patients varied in age from 23 to 37 years. Allgestations were term. The patients received antacid(Bicitra, Willen Drug Co.) for aspiration prophylaxis.Anesthesia was induced with thiopental and maintainedwith nitrous oxide-oxygen and either halothane, en­flurane, or isoflurane.

Sections of the uterus that contained the uterine ar­teries were placed in gassed (95% oxygen and 5% car­bon dioxide) Krebs-bicarbonate solution immediatelyafter hysterectomy and before being taken to the lab­oratory. The preparation of isolated arterial rings andthe suffusion techniques used have been described indetail previously.8,9 In brief, rings 2 mm in length wereprepared from the uterine artery and mounted be­tween two tungsten wires in a 5 ml volume suffusionchamber. The diameter, measured by a dissecting mi­croscope, of the unstimulated uterine artery taken frompatients with uncomplicated pregnancy is not signifi­cantly different when compared with the uterine arteryfrom multiparous, nonpregnant patients. Changes inisometric tension were measured by means of a force­displacement transducer (Statham) connected to theupper mounting wire and recorded on a Gould re­corder (model 2400, Brush).

The mounted arteries were suffused continuouslywith gassed (95% oxygen and 5% carbon dioxide)Krebs-bicarbonate solution at a flow rate of 4 ml/min

Lack of reactivity of uterine arteries 1437

delivered by a peristaltic pump (Gilson Instruments).The Krebs-bicarbonate solution had the following com­position: sodium chloride, 119 mmol/L; sodium bicar­bonate, 25 mmol/L; potassium chloride, 3.6 mmol/L;magnesium sulfate, 1.2 mmol/L; potassium phos­phate monobasic, 1.2 mmol/L; calcium chloride, 2.5mmol/L; glucose, 11 mmol/L; ascorbic acid, 0.005mmol/L; and disodium ethylenediaminetetraacetate,0.03 mmol/L. The bath solution was maintained at370 C and had a pH of7.4. The drugs were adminsteredin the inflowing suffusion fluid.

The preparation was allowed a stabilization periodof 1 to 2 hours before the experiment began. Duringthat time, basal tension was adjusted repeatedly untilthe resting tension became stable at 1 gm, which wehave found to be the optimal resting tension for thehuman uterine arteries. Uterine arteries from patientshaving cesarean hysterectomy for medical reasonsother than obstetric hemorrhage (control patients, typeI) had reproducible constrictor responses to norepi­nephrine for several hours after the stabilizationperiod.

The drugs studied, including norepinephrine, se­rotonin, arginine vasopressin, prostaglandin F2a (PGF2a )

and potassium chloride have multiple mechanisms ofconstriction, such as activation of receptor-operatedand voltage-dependent calcium channels, as well as re­lease of intracellular calcium. The concentrations of thedrugs used in the studies were found to elicit 60% to100% of the maximum constriction produced by eachdrug in uterine arteries from normal pregnant andnonpregnant patients. If little or no constriction waselicited by the vasoconstrictor drug, the drug concen­tration was increased. Each concentration of the con­strictor drug was administered for at least 15 minutes,a time found to produce a stable constriction in uterinearteries from normal pregnant patients." Each drug waswashed out for 30 minutes before the subsequent ad­ministration of another drug.

The mechanisms for the impaired constrictor re­sponses were investigated with the following proce­dure. Norepinephrine (30 umol/L) was administeredfor 15 minutes. After the norepinephrine-induced con­striction returned to baseline or after 15 minutes, thearterial ring was suffused for 30 minutes with a givenconcentration of indomethacin, to inhibit prostaglan­din synthetase, with methylene blue, to prevent theeffect of endothelium-derived relaxing factor on sol­uble guanylate cyclase, or with propranolol, to antag­onize l3-adrenergic receptors. Norepinephrine, in thepresence of the antagonist or blocker, was again ad­ministered for 15 minutes.

To determine whether the impaired constrictor re­sponses were due to elevated basal levels of cAMP thatare associated with smooth muscle relaxation, basal

1438 Nelson and Suresh

Table I. Categories of patients on basis of reactivity of isolated uterine arteries

May 1992Am J Obstet Gynecol

Patient Estimated bloodNo. Diagnosis loss (ml) Coagulopathy Drugs

Associated clinicalphenomena

Type I (normal reactivity)03 84 Placenta accreta

0465

7472

7177

22

83

8380

8597

9693

85

84

FibroidsCervical carcinoma, mar­

ginal placenta previa

Cervical cancerSevere dysplasia

CarcinomaPlacenta accreta

Permanent sterilization

Cervical carcinoma

Levothyroxine so­dium (Syn­throid)

Terbutaline, folicacid

Repeat cesareansection

Amniocentesis

AmniocentesisRepeat cesarean

section, am­niocentesis

Chronic hyper­tension, tox­emia withpreviouspregnancy,repeat cesar­ean section,amniocentesis

Repeat cesareansection

Amniocentesis,anemia

Type II (weak reactivity)48 73

64 53

Repeat cesarean section

Uterine atony, possibleamniotic fluid em­bolism

900

10,000 + (DIC)

Oxytocin(Pitocin)

Methylergonovinemaleate (Meth­ergine), PGF2a ,

blood products,furosemide (La­six), gentamicin,dopamine,Pitocin

Amniocentesis,anemia

All patients were multiparous and had no evidence of hypoxia or sepsis. Coagulopathy was determined by prothrombin time,thromboplastin time, platelet levels, levels of fibrinogen, and fibrin degradation products. Drugs included medications other thanthose normally used. MAP, Lowest mean arterial pressure; DIG, disseminated intravascular coagulation.

levels (but not stimulated levels) of cAMP were deter­mined by radioimmunoassay.'?' 11 Stimulated levels ofcAMP were not determined because the de pessed con­tractility precluded a correlation between vasorelaxa­tion and stimulated levels of cAMP. Portions of arterieswere blotted, weighed, and frozen in liquid nitrogen.Frozen tissue was homogenized in 2 ml of 10% tri­chloroacetic acid (cold) with a Polytron tissue disruptorat a setting of 6 for 30 seconds. The supernatant waslyophilized and reconstituted in I ml of 50 mmol/Lsodium acetate, pH 4.75. Aliquots were acetylated andassayed for cAMP by radiommunoassay.

Because endogenous vasodilators and vasoconstric­tors may be released from the endothelium, the pres­ence of endothelium was determined by a silver staintechnique." Arterial rings proximal and distal to therings used for the pharmacologic studies were cut open,laid adventitia-side-down, and stained. Other ringswere stored in 10% formalin for subsequent paraffinembedding and staining with hematoxylin and eosin.

Norepinephrine-induced constriction was often as-

sociated with phasic constrictions (drug-induced oscil­latory activity). When norepinephrine-induced oscil­latory activity occurred, the lowest tension that devel­oped during the rhythmic activity was taken as theresponse (i.e., the amplitude of the rhythmic activitieswas not included in the measurements). Data are pre­sented as mean values ± SE.

All stock solutions and serial dilutions made in Krebs­bicarbonate solution were prepared shortly before thestart of each of the experiments. The following drugswere used and were obtained from Sigma ChemicalCompany, St. Louis: (- )-arterenol bitartrate, indo­methacin, and methylene blue, PGF2a , [ArgB]-vasopres­sin acetate, DL-propranolol hydrochloride, histaminediphosphate, and 5-hydroxytryptamine creatininesulfate.

ResultsThe patients were divided into three categories. The

control patients were considered type I. The hemor­rhaging patients were divided into types 11 and 111 on

Volume 166Number 5

Table I-Cont'd

Lack of reactivity of uterine arteries 1439

82 77Type III (no reactivity)

06

75

01

26

57

66

88

76

Retained products ofconception

Placenta accreta, second­ary vaginal and uterinelaceration

Uterine atony

Abnormal placental im­plantation

Uterine atony

History of in-travenousdrug abuseand hepatitisB, chronic in-flammation ofendometriumand myome-

PGF'a, blood trium,lym-Throm- products, ampi- phocyte infil-

bocyto- cil.li?, genta- tration of4000 penia rrucm myometrium

Methergine,PGF2a, bloodproducts, genta-micin, (Cleocin), Delivery atmagnesium home, breech

5000 + (DIC) sulfate birthMethergine,

PGE2a, ephed-rine, Pitressin, Repeat cesareanblood products, section,albumin, mag- chronic ane-nesium sulfate, mia, chronic

3200 + (DIC) hydralazine cervicitis

. Pitocin, PGF'a'2000 prostaglandin E,

Methergine, Amniocentesis,PGF'a, prosta- repeat cesar-glandin E" ean section,

1500 Pitocin anemia

the basis of reactivity of the isolated uterine artery.Types II and III were not associated with any particular

diagnosis, parity, presence of coagulopathies, or de­pressed mean arterial blood pressure (Table I).

The uterine arteries from the control patients (typeI, n = 9) responded with a maximum or near-maxi­

mum constriction to norepinephrine (30 urnol/L,3.6 ± 1 gm), potassium chloride (75 mmol/L, 10.2 ±3 gm), and arginine vasopressin (1 urnol/L, 18.8 ± 2.6gm). In uterine arteries from five patients with uncon­trollable bleeding (type II), the constrictor responses tothe same drugs were markedly depressed: norepi­nephrine (30 urnol/L, 0.5 ± 0.2 gm), potassium chlo­ride (75 mmol/L, 1.9 ± 0.8 gm), PGF2a (30 urnol/L,ogm) and arginine vasopressin (1 urnol/L, 0.2 ± 0.05gm). Uterine arteries from two patients (type III) ex­hibited no constrictor responses to norepinephrine (30urnol/L), potassium chloride (75 mmol/L), PGF2a (30urnol/L), or arginine vasopressin (1 urnol/L). Recordsof experiments that used uterine arteries from thethree categories of patients are shown in Fig. I. Theresults from the experiments are summarized in Fig.2. When the concentrations of the constrictor drugswere increased, little or no further increases in con­striction were observed. In preliminary studies thatused uterine arteries from patients with uncontrollableuterine bleeding various concentrations of the drugswere used but no responses were elicited except at

higher concentrations. The depressed responses to thevarious vasoconstrictors persisted, in spite of continualsuffusion and repeated administration of the drugs for8 to 12 hours.

Other constrictor agents (serotonin, 30 u.mol/L; cal­

cium chloride, 4 mmol/L; histamine, 30 umol/L) wereused in a few of the uterine arteries from patients withobstetric hemorrhage (types II and Ill). None of theseagents induced constriction. The impaired response tonorepinephrine was not reversed by indomethacin(1 u.mol/L), which is an inhibitor of prostaglandinsynthetase; by methylene blue (10 urnol/L). which isa blocker of endothelium-derived relaxing factoractivation of guanylate cyclase; or by propranolol(1 urnol/L), an antagonist of l3-adrenergic receptors(Fig. 3).

The basal concentrations of cAMP were 152.7 ± 1.4pmol/gm wet weight in uterine arteries (n = 3) frompatients with obstetric hemorrhage and 170.3 ± 4.5pmol/gm wet weight in uterine arteries (n = 3) fromcontrol patients.

The intimal surface of the arterial rings was exam­ined microscopically en face after application of a silverstain to ascertain the presence of endothelial cells. Al­most all the intimal surface of the rings was covered byendothelial cells as demonstrated by "black" cell bor­ders. In the tissues that were embedded, stained, andexamined at a magnification of x 40, there did not

1440 Nelson and Suresh

A. II III

~ji\4

9 ~ 13

9 9 210 min 15 min if-----1

0~

t L t tNE NE NE NE

(30 JIM) (30 JIM) (30 JIM) (100 JIM)

B. II III

145 min 1 14f---1

12 6 12

10 5 10

8 4 89 9 9

6 3 6

4 2 110 mini 4

2 ~2 10min

f-----1

0 0 0

t t tI KCIl5 mM

KCIl5 mM KCIl5 mM

c. II III

15 min,

:If :j~ "j16 ,5mini

9 9 98

0

t o t tADH 1 JIM ADH1J1M ADH1J1M

May 1992Am J Obstet Gyneco l

Fig. 1. Records of experiments showing effects of norepinephrine (NE), potassium chloride (KGI),and arginine vasopressin (ADH) in isolated uterine ar teries from the various pat ient types (I, II , andIII) . Concentrations of constrictor agents produced maximum or near-maximum constrictions incontrol patients (I) .

appea r to be an y differences in the endothelial andsmooth muscle layers of the art er ies from type I (con­trol) patients (n = 3) or type II (n = 1) or III (hem­orrhaging) patients (n = 2).

Comment

During obstetric hemorrhage the patient is oftengiven drugs to increase uterine tone. These drugs, in­cluding oxytocin , ergot alkaloids, vasopressin (Pitres­sin) , and PGF 2a , have vasoconstr ictor action . The pa­tients may also receive vasopressors such as ephedrineand phenylephrine. In thi s study isolated uterine ar­teries from patients experiencing obstetric hemorrhagedemonstrated depressed or lack of reactivity to someof th e same vasoconstrictor drugs used in the patient.

T o our knowledge thi s is the first report of such analtere d reactivity in the isolated uterine vasculature ofpa tients with obstetric hemorrhage compared withpregnant patients who undergo cesarean hysterectomyfor othe r medical reasons.

This study invol ved onl y uterine art er ies from preg­nant patients, which elim inates the variability in resultsthat may occur when pregnant and nonpregnant con­dition s are compared. We have previousl y shown thatthe norepinephrine or potassium chloride concentra­tion that produced 50% ofthe maximal contraction andthe maximal contractile respon se to norepinephrine orpotassium chloride are similar in uterine arteries frompregnant and nonpregnant patients ." However, otherstudies comparing pregnant and nonpregnant condi-

Volume HiliN umber 5

Lack of reactivity of uterine arteries 1441

5NE (30 ~M)

15KCI (75 mM)

9

4

3

2

oIII

13

11

9 9

4

2

oIII

IIIII

Vasopressin (1 ~M)23

21 .19

917

:111__IIIII

3

2

o

9

Fig. 2. Summary of results showing depressed constrictions induced by vasoconstrictors that actthroug h various mechanisms in uterine arteries from patients with obstetric hemorrhage. Data wereobtained from experiments similar to those shown in Fig. I . Each bar represents constrictor responsein grams (K) produced by vasoconstrictor and is mean ± SE of data from various patient types (I ,n = 9; II, n = 5: Ill. n = 2).

tion s sho w a var iability in the con strictor responses todrugs such as norep inephrine, phenylephrine, orPGF". in uterine arteries from pregnant sheep,":" rab­bits, I " and guinea pigs.HI

An effort was mad e to determine the mechanism forthe depressed rea ct ivity in the uterine arteries from thepatients exper iencing hemorrhage. Agents wer e usedthat have been shown to produce vasoco nstr iction bydifferent mechanisms such as activation of recep tor­op erated and/or voltage-de pe ndent calcium channelsas well as the release of calcium from intracellular sto r­age sites. These variou s compounds produced no oronly low constrictions. The results suggest that the de­pressed mechanism is com mon to all of th e constrictors.Thus pe rh ap s the contractile machinery itself is im­

paired .It is possible that these uterine arte ries fro m patients

experi en cing obste tr ic hemorrhage may have ele vatedlevels of an endoge no us vasodilator th at is cont inuo uslyreleased . We ha ve previou sly reported that endothe­lium -derived relaxin g facto r production or release isincreased in uterine a rteries from no rm al pregnant pa­tients when compared with that of nonpregnant pa ­tients.' However, in the ut erine arteries from pati en ts

with hemorrhage, methylene blu e (an inhibitor of theaction of endothelium-derived relaxing factor or nitricoxide on soluble guanylate cyclase)" did not increasethe baselin e or enhance the subsequent response tonorepinephrine to any grea ter extent th an it did innormal pregn ant uterin e a rte ries. Furthermore, in­domethacin , a cyclooxygen ase inhibitor, did not in­crease the baseline or increas e the constr ictor respon seto norepinephrine, which indicates th at an endogen ou svasodilator prostaglandin was not continuously re­leased or released with expos ure to norepinephrine.

In other blood vessels the ~2-adrenergic receptor hasbeen shown to mediate vasodilatation. Norepinephrinehas a low affinity for th e vasodilatory ~2-adrenergic

receptor but a high affinity for th e ~ l·adrenergic re­ceptor. Propranolol , a non selective ~-adrenergic recep­tor antagoni st , was used to determine whether therewas an increased affin ity of the ~2-adrenergic receptorfor norepinephrine or a cha nge in th e subtype from ~2

to ~ I in the uteri~e a rterie s that show ed depressed nor­epinephrine-induced constrictions. Such changeswould be reflected in a depressed norepinephrine-in­duced constriction. However, the response to norepi­nephrine was unaffected by the presen ce of propran-

1442 Nelson and Suresh May 1992Am J Obstet Gynecol

6NE (30 ).lM) + Indomethacin (1 ).lM)

6NE (30 ).lM) + Methylene Blue (10 ).lM)

I

5 5

4 4

9 3 9 3

2 2

II

III III0 0

5NE (30 ).lM) + Propranolol (1 ).lM)

I

9

4

3

2

oIII

Fig. 3. Lack of effect of inhibitors of various vasodilator mechanisms. Norepinephrine (30p.mol/L) was administered for 15 minutes in absence and presence of inhibitor. Inhibitor was presentfor 30 minutes before norepinephrine was readministered. Each bar represents constrictor responseto norepinephrine in grams (g) in presence of inhibitor (I, n = 9; 1I, n = 5; Ill, n = 2).

0101. Thus it appears that in the uterine arteries, as wellas in other blood vessels, norepinephrine is selectivefor the f31-adrenergic receptor and that there is nochange in the f3-adrenergic receptor subtype in theuterine arteries from patients with obstetric hemor­rhage.

Numerous studies have shown a good correlationbetween cAMP concentrations in blood vessels and thedegree of relaxation produced by stimulators of ade­nylate cyclase (e.g., isoproterenol), The basal levels ofcAMP were not elevated in the uterine arteries fromthe patients with hemorrhage compared with those ofuterine arteries from control pregnant patients. Henceour results exclude the possibilities that overstimulationof adenylate cyclase or dysfunctional adenylate cyclaseor its regulatory G proteins (Gi or Gs) may be involvedin the depression of contractility in the uterine arteries.

The mechanism for the depressed or lack of reactiv­ity of the uterine arteries from hemorrhaging patientsremains unclear. In a review of the charts of the patientsno risk factors could be found that were common toall of the hemorrhaging (types II and III) patients. Riskfactors included causes of obstetric hemorrhage, pres­ence of infection, development of hemorrhagic shock(volume of blood loss and fall in mean arterial pres-

sure), parity, drugs, and associated clinical phenomena.Histologically, there did not appear to be any differ­ences in the intimal or medial layers in the uterinearteries. Furthermore, the depressed or lack of reac­tivity could not be washed out from the isolated uterinevessels with time.

It is generally thought that obstetric hemorrhage isassociated with a number of causes such as uterineatony, abnormal placentation, trauma during labor anddelivery, and coagulation defects.' The results of thisstudy, which used isolated uterine arteries from pa­tients experiencing obstetric hemorrhage, suggest thatthese uterine arteries have lost their ability to constrictas a result of an unidentified mechanism. Failure ofuterine artery constriction may contribute to uncon­trollable obstetric hemorrhage.

REFERENCES

I. Cunningham FG, MacDonald PC, Gant NF, eds, Obstetrichemorrhage. In; Williams' obstetrics. 18th ed. Norwalk,Connecticut: Appleton & Lange, 1989:695-725.

2. Suresh MS, Nelson SH, Nelson TE, Steinsland OS. Preg­nancy: increased effect of verapamil in human uterinearteries. Eur] Pharmacol 1985;112:387-91.

3. Nelson SH, Suresh MS. Magnesium sulfate-induced re­laxation of uterine arteries from pregnant and nonpreg­nant patients. AM] OBSTET GYNECOL 1991;164:1344-50.

Volum e 166Number 5

4. Nelson SH, Steinsland OS. Endothelium-dependent in­crease in sensitivity to acet ylcholine (ACH) in human uter­ine arteries during preg nanc y. Blood Vessels 1987;24:218-19.

5. Weiner C, Martinez E, Zhu LK, Ghods i A, Chestnut D.In vitro release of endothelium-derived relaxing factorby acetylcholine is increased during the guinea pig preg­nancy. AMJ OBSTET GVNECOL 1989; 161:1599-1605.

6. Magness RR, Osei-Boaten K, Mitchell MD, Rosenfeld CR.In vitro prostacyclin production by ovine uterine and sys­temic arteries: effects of angiotensin II. J Clin In vest1985;76:2206-12.

7. Magness RR, Rosenfeld CR, Shaul PW. Role of endothe­lium in the control of cAMP and cGMP production byuterine artery (VA) vascular smoo th muscle (VSM) inovine pregnancy [Abstract]. In: Proceedings of the thirty­eighth annual meeting of the Society for Gynecologic In­vestigation, San Antonio, Texas , March 20-23, 1991. SanAntonio: Society for Gynecologic Investigation, 1991 :231.

8. Bevan JA, Osher JV. A direct method for recording ten­sion changes in the walls of small blood vessels in vitro .Agents Actions 1972;2:257-60. .

9. Nelson SH , Suresh MS. Comparison of nitroprusside andhydralazine in isolated uter ine arteries from pregn ant andnonpregnant pat ients. Anesthesiology 1988;68:63-9.

10. Brooker G, Harper JF, Te rasaki WL, Moylan RD. Radi o­immunoassay of cyclic AMP and cyclic GMP. In : BrookerG, Gr eengard P, Robison GA, eds. Advances in cyclic nu­cleotide rese arch. New York: Raven Press, 1979;10:1-33.

II . Nickols GA, Cline WH o Parathyroid hormone-inducedcha nges in cyclic nucleotide levels during relaxation ofthe rat aort a. Life Sci 1987;40:2351-9.

12. Furchgott RF, Zawadzki JV. The obligatory role of en­do thelium in the rela xation of art erial smooth muscle.Nature 1980 ;288:373-6.

Lack of reactivity of uterine arteries 1443

13. Magne ss RR, Rosenfeld CR. In vitro uterine arterial ton eand contractility in pregnant (P) and nonpregnant (NP)sheep [Abstract]. In : Proceed ings of the thirty-third an­nual meeting of the Society for Gyneco logic In vestigatio n,Toronto, Ontario, Canada, March 19-22, 1986. Toronto:Society for Gynecologic Investigation, 1986 :182.

14. Pope WF, Stormshak F. Effect of the ovine conceptus inin vitro responses of uterine arter ies to prostaglandin E.and norepinephrine . Bioi Reprod 1979;20:847-51.

15. Griendling KK, Fuller EO, Cox RH. Pregnancy-inducedcha nges in sheep ute rine and caro tid ar teries. AmJ Phys­iol 1985;248 :H658-65.

16. Ford SP, Weber LJ, Stormshak F. In vitro response ofovine and bovine uterine art eri es to prostaglandin F•• andperiarterial sympathetic nerve stimulation. Bioi Reprod1976; 15:58-65.

17. Annibale DJ, Rosenfeld CR, Kamm KE. Alterations invascular smooth muscle contractility during ovine preg­nancy. Am J Physiol i 989 ;256: H 1282-8.

18. Moisey DM, Tulenko T. Increased sensitivity to angio­ten sin in uterine arteries from pregnant rabbits. Am JPhysiol 1983;244:H335-40.

19. Weiner CP, Martinez E, Chestnut DH , Ghodsi A. Effe ctof pregnancy on uterine an d caro tid art ery response tonorepinephrine, epinephrine, and ph en ylephrine in ves­sels with documented functiona l endo thelium. AM J OB­STETGVNECOL 1989 ;161 :1605-10 .

20. Martin W, Villani GM, J othianandan D, Furchgott RF.Selective blockade of endothelium-dependent and gly­ceryl trinitrate-induced relaxation by hemoglobin and bymethylene blue in the rabbit aorta. J Pharmacol Exp Ther1985;232:708-16.