- Email: [email protected]
Experimental toxemia in the pregnant primate
Experimental toxemia in the pregnant primate DENIS
CAVANAGH,
PAPINENI CHENG
S. C.
TIMOTHY
M.D.
RAO,
TSAI, C.
PH.D.,
B.V.Sc.
M.D.
O’CONNOR,
M.B.,
B.CH.
St. Louis, Missouri
In order to develop a model for the study of eclamptogenic toxemia, a series of experiments were carried out on 31 female baboons. In Group 1, consisting of IO animals, metal clips were placed around the uterine arteries in order to partially occlude them, and the ouarian vessels were transected. The animals were subsequently mated. Nine developed hypertension and proteinuria, and one aborted. The renal lesions in these animals were indistinguishable f rom those described in human toxemia. Group 2 consisted of three of the 10 baboons from Group I, which became pregnant a second time. They again developed hypertension and proteinuria. In Group 3, three baboons at 100 days of gestation were treated as in Group 1 with similar results. Groups 4 and 5 served as pregnant (3) and nonpregnant (15) controls. It is concluded that a toxemia model has been developed in a subhuman @mate. This model will prove useful in the further study of eclamptogenic toxemia. (AM. J. OBSTET. GYNECOL. 128: 75, 1977.)
ECLAMPTOGENIC TOXEMIA complicates approximately 6 per cent of all pregnancies in this country and remains a leading cause of maternal death.’ In addition, it has been linked to intrauterine growth retardation and to behavioral disorders in surviving children.’ Although most likely related to nutritional deficiency, the exact mechanism of pathogenesis is unknown, and the treatment remains symptomatic. An extensive review of the literature on the pathogenesis of toxemia was made in 1974 in association with a preliminary report,3 so this will not be repeated here. The present study represents an expansion of a multidisciplinary program designed to evaluate the concept that the “trigger mechansim” in the pathogehesis of toxemia is uteroplacental ischemia and that, by reducing the blood supply to the uterus, a toxemia animal model can be developed in the subhuman primate. Baboons were used in this study because their reproductive physi-
ology is remarkably similar to that of the human subject, and this certainly would make the findings more meaningful. The average weight of the animals used was approximately 14 kilograms.
Method and material A total of 31 female baboons (pap;0 anztbis) were divided arbitrarily into five groups, with three animals which had repeat pregnancies being used twice and each study animal serving as its own control. Pregnant study group (Group 1) This group consisted of 10 baboons. All of the baboons were sedated with an intramuscular injection of ketamine hydrochloride (Ketaset) in a dose of 5 mg. per kilogram of body weight. This was supplemented by intravenous administration of pentobarbital sodium as required to induce surgical anesthesia. Laparotomy was performed, and the uterine blood vessels were identified and isolated. Metal clips were placed around the uterine arteries, so that the uterine artery flow would be little changed after the procedure. Ovarian blood vessels were transected and ligated bilaterally, so that the main collateral blood supply of the uterus was interrupted. With the ovarian arteries ligated, almost all of the blood which supplied the uterus had to flow through the “clipped” uterine arteries, and the tnetal clips would not allow hypertrophy during pregnancy. Thus, this group of study animals was predisposed to
From the Departments of Obstetrics-Gynecolog and Pathology, St. Louis University School of Medicine. Supported in part by United States Public Health Seruice Grants FR 5388-08 and FR 5388-09 and by National Foundation-March of Dimes Grant I-335. Presented at the Eighty-seventh Annual Meeting of the American Association of Obstetricians and Gynecologists, Hot Springs, Virginia, September 9-11, 1976. Reprint requests:Dr. Denis Cavanagh, Department of Obstetrics and Gynecology, College of Medicine, University of South Flon’dn, Tampa, Florida 33620.
75
76 Cavanagh et al.
Fig. 1. Radiograph showing pregnant baboon near term with metal clips in place around the uterine arteries.
uteroplacental ischemia by preconceptional clipping. These animals were then observed for the crythematous perineal swelling indicative of ovulation. When this was noted, the female animal was transferred from her cage to that of her male consort and remained there for a period of approximately four days. The female animal was then returned to her cage and observed throughout the next menstrual cycle. If evidence of ovulation occurred, the mating process was repeated. This course was followed until all of the animals were pregnant. An x-ray of the abdomen in late pregnancy was used to confirm that the metal clips were still in position (Fig. 1). Repeat pregnancies (Group 2). After cesarean section, some of the baboons of Group 1 were killed so that we could obtain the tissues for various histopathologic studies. Following wound healing and rethe remaining baboons were allowed to cuperation, become pregnant. Three of these animals became pregnant for the second time and carried the fetus to term. Antepartum study group (Group 3). Three baboons were used in this group in which the experimental technique was similar to that of Group 1, with the main difference being that these baboons were at the gesta-
tional age of approximately 100 days. All three animals carried the pregnancy to term and Iverc delivered by cesarean section. Pregnant control group (Group 4). This group consisted of three baboons which underwent a “sham” operation in which the ovarian arteries were ligated and transected and metal clips were placed around the uterine arteries and then removed. All of these animals became pregnant and went to term. Nonpregnant control group (Group 5). This group consisted of 1.i baboons which had no dissection around the uterine or ovarian arteries. None was mated, and these animals ivcre used to provide baseline nonpregnant control data. In addition to gross maternal and fetal outcome, four methods were used to evaluate the effect of the experimental technique on pregnancy outcome. Hemodynamic studies. The animals were anesthetized with ketamine hydrochloride supplemented by sodium pcntobarbital. ‘They were then placed on an operating table. Pressures were measured by Statham transducers, and a Sanborn recorder (7700 series) was used for recording aortic pressure. Aortic pressure was measured b! passing a polyethylene catheter along the femoral arter! into the aorta to the level of the renal arteries. A poi~ethylcne catheter bias passed along the femoral vein into the inferior vena tava so that the tip came to rest just proximal to the right atrium. This venous catheter was used for the injection of (Zardiogreen (indoc.);anine green) tar the cardiac output estimation. Estimation of cardiac output was accomplished by the dye-dilution method. Dilution curves from the arterial catheter were recorded by continuous sampling througtr a c,uvette densitome& (Gilson Medical Electronics. 1nc.), which ~vas connected to a directreading cardiac- output computer (Lexington Instruments (;orp.) and a direct strip chart recorder. ‘The computer was calibrated and programmed to read out cardiac output directly f-rom the densitometer and reject all cur\‘es not arithmetically valid. Total peripheral resistance was calculated as described by Guyton.’ Standard elec.trocardiographic leatfs were used to rpcord cardiac rate and rhythm. Urinalysis. Urine teas obtained by catheterizing the bladder, and the levels of protein, sugar, and pH were determined by Labstis reagent strips (Ames <:ompany).
Histopathologic studies. The kidneys were removed from three mothers in Group 1, and biopsies were vbtained from the remaining pregnant animals except those in Group 2, at or immediately after delivery. Tissues were fixed in 10 per cent formalin fixative and
Volume Number
Experimental
128 1
Fig. 2. Photomicrograph
of renal biopsy specimen showing glomerular enlargement of the glomerular capillary lumina due to cytoplasmic swelling is indistinguishable from that found in human eclamptogenic
larity, with reduction lial cells. This lesion
Table
I. Hemodynamic
parameters
in animals
Systolic pressure (mm. Hg) Diastolic pressure (mm. Hg) Mean aortic pressure (mm. Hg) Cardiac output (mUmin.) Peripheral resistance (dynes sec./cm.‘)
Table
II. Results
of urinalysis
Protein Sugar
(mg./
embedded with
in
paraffin,
hematoxylin
Schiff
and
Small
and
and
5 p sections
eosin
and
blocks
of renal and
cortex
embedded
tions
were
stained
with
and
viewed
with
a Phillips
were acetate
electron
3
Group
4 Group 5 (controls)
(pregnant
clipping)
controls)
160 100 120 1.913 5,307
120 75 90 2,000 3,510
*
stained
periodic
fixed
acid-
in 4 per Extra-thin
and
Group 3 (antepartum clipping) Negative Negative
7.0
were
with
in epon.
uranyl
Pwgnancy)
Trace to 100 Negative
in
hematoxylin.
glutaraldehyde
Group (antepartum
croup 2 (repeat Pw-cy)
30 to 300 Negative (except one baboon) 5.5
PH
and hypercelluof the endothetoxemia.
136 82 100 1,635 5,375
studied
clipping) 100 ml.)
2
(repeat
155 107 123 1,400 7,256
Group I (preconceptional Urinalysis
Group
167 107 127 1,479 7,946
in animals
77
studied
Group I (preconceptional clipping)
Hemodynamics
toxemia
lead
microscope.
cent sec-
citrate
to 300
Group 4 (pregnant control}
(controls)
Negative Negative
Negative Negative
7.0
6.0
Immunofiuorescence goat
anti-human
IgE,
albumin,
and
Behring
used
for
reactivity
(Meloy
was
direct
6.0
G (IgG), (C3
Laboratories,
Diagnostics, the
5
study. Fluorescein-conjugated
immunoglobulin complements
fibrinogen
Group
C4),
IgA, and
Springfield,
Somerville,
immunofluorescent
confirmed
and
by
New
IgM, fibrin/
Virginia, Jersey)
studies.
immunoelectrophoresis
were Cross-
78 Cavanagh et al.
May
1, 1977
Am. J. Obsfet. Gynecol.
Fig. 3. Electron micrograph of the kidney of a “toxemic” baboon showing the endothclial cell nucleus (EN) and narrowing of the glomerular capillar) lumina (CL) as a result of cytoplasmic swelling and subendothelial deposition (SED) of fibrillary StrucTures. This is similar to the pictur-e seen in human eclamptogenic toxemia. (Original magnification x6.960.) and radial immunodiffusion between human immunoglobulins and complements and those of the baboon. Tissues were rapidly frozen in liquid nitrogen. Cryostat sections, 6 p thick, were fixed in ethanol: ether (1: 1) for 10 minutes. They were then placed in 95 per cent ethanol for 20 minutes. They were washed off in buffered saline for five minutes. and this washing was repeated. They were then incubated with fluorescein-conjugated antisera, washed in phosphate-buffered saline, and mounted on a cover slip with buffered glycerin. Reactivity of tissues with antisera to human immunoglobulins, complement components C3 and C4. fibrin/fibrinogen were studied. The resulting immunofluorescence was quantitated on a scale of 0 to 4+. The sections were photographed with a Leitz. ortholux fluorescence microscope .
pregnant animals. Sign&rant differences were observed during the third trimester of pregnancy between Groups 1.2, 3, and the “sham’‘-operated animals (Group 4). l‘he findings in 15 nonpregnant baboons (Group 5) were obtained and compared with those in the four groups of pregnant animals. Hemodynamic studies. The results of these are summarized in Table I. In Group 1 the animals became progressively hypertensive during the third trimester of pregnancy. ‘l‘he mean blood pressure value prior to pregnancy teas 117168 mm. Hg, with a mean
Three
Results All 10 baboons in Group 1 became pregnant but one aborted. No abortions occurred in the other groups of
arterial
pressure
of‘ 84 mm.
Hg.
Near
term,
the
mean blood pressure value was 1671107 mm. Hg. with a mean arterial pressure of- 127 mm. Hg. The mean values for cardiac output and total peripheral resistance were 1,479 ml. per minute and 7,946 dynes sec./ cm.S, respectively. of
pregnant
tact around Group
the
for
nine
the
baboons
second
the uterine
2. Near
term,
time.
arteries, they
had
from
Group
1 became
clips inand were designated
with
the
a mean
metal
blood
pressure
Volume Number
Experimental toxemia
128 1
79
Fig. 4. Electron
micrograph of the kidney of a “toxemic” baboon showing mesangial cell nucleus (MN) and markedly increased glomerular mesangial matrices (MM) with two large, fine, granular, electron-dense deposits (EDD). (Original magnification x 16,770.)
of 155/107 mm. Hg, with a mean arterial pressure of 123 mm. Hg and a mean cardiac output of 1,400 ml. per minute. The mean total peripheral resistance was slightly less than that of Group 1. Three baboons were used in Group 3 ‘with the experimental technique being identical to that used in Group 1, except that these baboons were at approximately 100 days of gestation when operated upon. All
three carried the pregnancy to term and were delivered by cesarean section. The mean blood pressure at the time of application of the metal clips was 97/57 mm. Hg, with a mean arterial pressure of 70 mm. Hg. At the time of cesarean section, the mean pressure was 160/ 100 mm. Hg, with a mean arterial pressure of 120 mm. Hg. (All of these animals developed hypertension and, one developed severe hypertension [2 1O/ 130 mm.
80
Cavanagh
et al.
Fig. 5. Electron
micrograph of a renal biopsy taken from a normal pregnant control baboon. This shows the capillary lumen (CL), basement membrane (BM), endothelial cell nucleus (EN). mesangial matrix (MM). and mesangial cell nucleus (MNI. (Original magnification x9.570.)
W
and3.
disc
output
+ proteinuria.)
peril Dheral resistance Gl -0up 4 consisted undo et-went
In this
was 1.9 13 ml. per
a “sham”
of 5,307 of three
group,
minute,
1 ‘he mean
operation
blood
mean
car-
mm.
a mean
total
representing
dynes sec./cm.” baboons. These in
artel =ies we :re ligated and transected, clips were 1Aaced around the uterine remc -wed.
the with
pressure
which and arteries at term
the
Hg
with
a mean
arterial
no significant
animals
ping. The mean cardiac ute with a mean total
ovarian
dynes
the metal and then was 120/75
Group
pressure rise
output peripheral
from
of 90 ml the
tin:
was 2,000 ml. resistance
-. Hg,
IC ( If
clip-
PC :r minof 3.4 10
sec.lcm.’ five
Urinalysis. 31. Proteinuria
consisted of The findings was
not
15 nonpregnant are summarized
detected
in any
c( mt rols. in Table of th ea inimals
Volume Number
Experimental toxemia
128 1
Fig. 6. Immunofluorescent photomicrograph showing extensive deposition of fibrin/fibrinogen the glomerular mesangium. and capillary vessel walls. This lesion is indistinguishable from found in human eclamitogenic toxemia. i x250.) during the first or second trimester of pregnancy. It was apparent only during the third trimester in Groups 1, 2, and 3. In Group 1, tested at term, the values ranged from 30 to 300 mg. per 100 ml.; in Group 2, a trace to 100 mg. per 100 ml.; in Group 3, negative to 300 mg. per 100 ml. Urinary glucose was absent in all of the animals except one in Group 1. The pH values ranged from 5.5 to 7.0. Histopathologic studies. Renal biopsies were obtained at term from five animals in Group 1 and from all other animals except those in Group 2. Light microscopy. The glomeruli in the biopsy specimens appeared equaliy affected in pregnant Groups 1 and 3 with hypertension and proteinuria. They were somewhat enlarged and hypercellular. The glomerular capillary walls were thickened, but the basement membrane proper was not demonstrably thickened even when special stains were used. The glomerular capillary lumina were reduced by cytoplasmic swelling of the endothelial cells (Fig. 2). Polymorphonuclear leukocytes were sparse or absent within the glomeruli. The epithelial cells and tubules were within normal limits when compared with those in biopsies from the pregnant and nonpregnat control groups. Electron microscopy. Renal biopsies were studied from animals in all groups except Group 2. The glomerular basement membrane was not thickened on electron
81
in that
microscopic study. The cytoplasm of the endothelial cells was swollen. The subendothelial spaces were widened, and depositions of loose fibrillary material were also found. As a result of these factors, there was narrowing of the glomerular capillary lumina (Fig. 3). Epithelial cells showed only focal fusion of foot processes especially over the paramesangial areas. The mesangial cells were hyperplastic, and the mesangial matrices were increased. There were fine granular electron-dense deposits found over the mesangial region in the animals with severe hypertension and proteinuria (Fig. 4). For purposes of comparison, an electron micrograph of a kidney from a normal pregnant baboon is shown in Fig. 5. Immunofluorrscencu studirs. Depositions of IgM, C3, C4, and fibrin/fibrinogen were found in the animals with hypertension and proteinuria in Groups 1 and 3. The distribution was similar to that described in human toxemia of pregnancy. The deposits were located in the glomerular mesangium and extended into the peripheral capillary loops, which give the impression of a “wire-loop” appearance. They were smooth, homogeneous, and evenly stained. Fine granular deposits were occasionally found. Rarely, fibrin thrombi were noted in the glomerular capillary Iumina. Fig. 6 shows extensive deposition of fibrin/fibrinogen in the glomerular mesangium and capillary vessel walls. Fine granular staining is evident in the renal glomeruli. Fig.
82
Cavanagh et al. Am
Fig. 7. Immunofluorescent photomicrograph similar to that for fibrin/fibrinogen. This eclamptogenic toxemia. (x250.)
Fig. 8. Immunofluorescent kidney of the “toxemic” (X100.)
showing extensive lesion is indistinguishable
deposition from
May 1, 1977 J. Obstet. G~nmx~l.
of IgM in a pattern that found in human
photomicrograph demonstrating the presence of complement (C4) in the baboon in a distribution pattern similar to that of IgM and fibrin/fibrinogen.
Volume Number
Experimental toxemia
128 I
7 shows that tion to that presence of tern similar
the IgM deposition is similar in distribuof fibrin/fibrinogen. Fig. 8 shows the complement (C4) with a distribution patto that of IgM and fibrin/fibrinogen.
Comment Ciinical eclamptogenic toxemia usually occurs after the twentieth week of pregnancy. It is characterized by hypertension (mean arterial pressure rise greater than 20 mm. Hg), edema, proteinuria, reduced renal artery flow with a reduced glomerular filtration rate, some evidence of disseminated intravascular coagulation, and distinctive renal glomerular changes. In the previously reported pilot study,3 we reported the development of hypertension and proteinuria with a reduction in renal artery flow in pregnant study animals which had been “clipped” prior to conception. In addition, fibrin/fibrinogen deposits were confirmed in the pregnant study group with immunofluorescence studies, although no evidence of disseminated intravascular coagulation could be found in four animals in which repeated coagulation studies were performed. The serum uric acid was elevated in the pregnant study
83
group as compared with the pregnant controlf group. Two obvious limitations in the pilot study vere the small number of animals in the study group (Qve) and our failure to demonstrate the typical histopathologic renal lesion of “toxemia” on light or electron microscopy in the only animal studied. In the present study, in baboons predisposed to uterine ischemia by preconceptional operation or in which uterine ischemia is produced during pregnancy, we have confirmed that the animals develop hypertension and proteinuria. In addition, with light microscopy, electron microscopy, and immunofluorescence studies, the renal lesion is indistinguishable from that found in women with eclamptogenic toxemia.5-‘0 After eight years of work on this animal model, we feel justified in concluding that although refinements are required a primate experimental model is now available for studying the pathogenesis of and evaluating new methods of treatment for toxemia of pregnatlcy. The availability of such a model could bring great dividends in health care for mothers and babies afflicted by this serious and common complication of pregnancy.
REFERENCE’S Statistics of the United States, Washington, D. C., 1976, vol. 24, No. 13, Department of Health, Education and Welfare.
1. Vital
2. Pasamanick, B.: The epidemiology of reproductive casualty, in Conference on Obstetric Factors in Child Development, Atlanta, Emory University School of Medicine. 3. Cavanagh, D., Rao, P. S., Tung, K. S. K., and Gaston, L.: AM. J. OBSTET. GYNECOL. 120: 183, 1974. 4. Guyton, A. C.: Textbook of Medical Physiology, ed. 5, Philadelphia, 1976, W. B. Saunders Company.
Discussion
DR. FRANK C. GREISS, Winstom-Salem,
North Carolina. In his continuing efforts to develop an experimental model of pregnancy toxemia, Dr. Cavanagh has amplified his initial report of 1974 and has added new data to demonstrate the similarity of the renal concomitant of human pre-eclampsia-eclampsia to that of his model. I am not a nephrologist and shall confine my comments to the cardi&ascular aspects of the report. Dr. Cavanagh has accumulated experience with 14 baboons whose uterine arteries were clipped prior to conception and with three baboons whose uterine blood supply was compromised at 100 days of gestation. It would appear that two of the three classic signs of toxemia, hypertension and albuminuria, have been induced. One would like assurance that environmental
and
mental
groups,
dietary
factors
and I would
were
similar
in the
be particularly
experi-
interested
5. Spargo, B., McCartney, C. P., and Winemiller, R.: Arch. Pathol. 68: 593, 1959. 6. McCartney, C. P.: Circulation (Suppl. 2) 30: 37, 1964. 7. Morris. R. H. Vassali. P.. Belier. F. K.. and McCluskev. ,j R. ’ T.: Ob$tet. Gynecol. i4:’ 32, 1964. 8. Altchek, A.: J. A. M. A. 175: 791, 1961. 9. Altchek, A., Albright, N. L., and Sommers, S. C.: Obstet. Gynecol. 31: 595, 1968. 10. Petrucco, 0. M., Thomson, N. M., Lawrence, J. R., and Waldon, M. W.: Br. Med. J. 1: 473, 1974.
in the patterns of weight gain, edema, and blood volume changes in the baboons. An obvious criticism of these experiments is that observations were made under anesthesia. Such methodology is understandable when one works with large subhuman primates, but observed responses may be variably altered. In view of these factors, mean values of blood pressure levels and albuminuria may be deceptive, and a breakdown of individual changes would be helpful to permit an estimate of the range of responses. Our prime attention should be directed to the thesis being tested, that uteroplacental ischemia induces pregnancy toxemia. The pregnancy experiences of at least three women who had preconceptual bilateral internal iliac and ovarian artery ligations have been reported.’ None developed toxemia of pregnancy. It is evident that sufficient collateral circulation developed
84
Cavanagh et al.
to support normal pregnancy. Perhaps similar collaterals developed in Dr. Cavanagh’s baboons or perhaps the baboon lacks the capacity to develop adequate collateral circulation. Dr. Cavanagh cannot presume that uterine artery clipping per se reduces uteroplacental blood flow. This presumption must be documented by objective How determinations in future experiments! ‘l‘aking the opposite tack, one could infer that the experimental method presently employed may represent “overkill” since compromise of the spiral arteries supplying the intervillous space would, in itself, be adequate to limit placental blood How. The observations of DeWolf and associates’. ’ are pertinent in this regard. They have observed replacement of the walls of spiral arterioles by trophoblast cells in normal pregnancies. These changes extend into the myometrial segments of the spiral arteries, creating widely dilated blood vessels conducive to high volume How. In preeclamptic patients, these “physiologic” changes do not extend to myometrial levels, and placental vascular resistance may be increased thereby. Thus, a mechanism which might limit placental blood flow has been observed, one which would have its inception during the time of definitive placentation. Dr. Cavanagh has a unique opportunity to evaluate further such vascular changes in his “toxemic” baboons. I have enjoyed reviewing Dr. Cavanagh’s paper, and I would exhort him to continue his work in an area of such obvious future potential. REFERENCES
Mengert, W. F., Burchell, R. C., Blumstein, R. W., and Daskal, J. L.: Pregnancy after bilateral ligation of the internal iliac and ovarian arteries, Ob. Gynecol. 34: 664, 1969. DeWoIf. F., DeWolf-Peeters, C., and Brosens, I.: Ultrastructure of the spiral arteries in the human placental bed at the end of normal pregnancy. AM. 1. OBSTET. GYNECOL. 117: 833,
1973.
1
-
-
DeWolf. F.. Robertson. W. G.. and Brosens, I.: The ultrastructure of acute atherosis in hypertensive pregnancy, .4~. J. OBSTET.
GYNECOI..
123:
164, 1975.
DR. CHARLESE. FLOWERS,JR., Birmingham, Alabama. I sincerely congratulate Dr. Cavanagh and his co-workers for a thoughtfully conceived and superbly executed study that significantly adds to our knowledge of the hypertensive complications of pregnancy. It is my belief that they have created an experimental model that reproduces the clinical syndrome we observe in the young primiparous patient with hypertension and without pre-existing vascular or glomerular pathology. I am particularly impressed with the electron microscopic studies of the glomerular capillary narrowing the subendothelial deposition of fibrin which is indistinguishable by immunofluorescence from the renal lesion seen in eclamptogenic toxemia in the human subject.
It is the young, short, poorly nourished, primiparous patient who so frequently develops hypertension in the first pregnancy but remains symptom-free in subsequent pregnancies. These studlcs would indicate that she has a uterine vascular system that is inadequatc~ to respond to the requirements of gestation. Hcmcvct-. the uterus, having become hvpertrophied in the hlitial pregnancy, is able to adjust acleqtiately to stibscclttent pregnancies. The uterine ischemia is associated with placental release of thromboplasdn. \\.hich initiates subclinical disseminated intravascular c,oagulation that is so mimimal it cannot be detected by present laboratory mcrhods. Within the kidney, fibrin deposition. ca$llarv narrowing, and increased glomerular mesangial matrices occur and produce an increase in renal r&tam r- al1t1 a reduction in renal blood How and cardiac output. These events pathologically alter angiotensin and aldosterone production. The degree of thext, alte.l.alic,m dictate the severity of the disease. its response to bed rest, and the ability of the kidney to rcco\er ccnnpletely. It may {bell be that some (;t our primiparous patients who develop hypertension in prcgnancl have such extensive fibrin deposition and renal danlage that hypertension of renal origin develops in suhsrquent pregnancies. This may be the reason for- tlic development of hypertension in Group ‘i’ of the esppt imental subjects. Although the explanation which I have ol’f~rcd places the major emphasis upon rrnal reaction IO glomerular fibrin deposition, it does not preclude the additional possibility that uterine ischemia results in the production of vasopressor substances and/or prostaglandins from the decidua and the placenta. Hopefully, this experimental model will eventually elucidate more of the pathophysiology of this disease and provide answers to some of the controversial issues of treatment such as the effect of magnesium s111f;tte, antihypertensive clrugs, diuretics. and cpidural anesthesia upon uterine blood How. Dr. Cavanagh did not indicate the effects of the induced hypertension upon the placenta and thr vital organs of the mother or whether these lesions are rcversible as thcv are in the human subject. I would he interested in these findings as uell as in hearing a of his present interpretation of the summary pathophysiology of the various tvpes of hypertension in pregnancy. DR. CAVANAGH(Closing). Regarding environmental and dietary factors, the environmental and dietary factors were identical in all groups of animals studied. With regard to weight gain, patterns of we-ight gain, and the weight gain in the toxemic versus the normal pregnant control animals, it is my impression that the toxemic baboons gained more weight, but this did not reach statistical significance. It is extremely difficult to
Volume Number
128 1
find curves on baboons, and so we cannot really give an unequivocal answer that weight gain was increased with the toxemic group at this time, but I suspect it is. With regard to edema, that is hard to assess in the baboon, and, if you ever, try to detect edema in a pregnant baboon, you will, I think, agree with me. However, we should try to find some way of assessing edema. I think these are very pertinent comments, especially Dr. Greiss’ question, did we perform blood volume studies in the baboons? I regret to say that the answer is no. But I think that would be an extremely pertinent study to perform in these toxemic animals. The problem was that we were determining cardiac output and also wanted to estimate total peripheral resistance, and we were afraid that the two methods would conflict. But blood volume studies are planned for the future, and I think the suggestion is excellent. With regard to anesthesia, there certainly is no way we are going to work with unanesthetized baboons without a restraining chair. We do not happen to have one, so all of the studies were done on anesthetized animals, as Dr. Greiss mentioned. However, the animals were lightly anesthetized, and, if anything, I would expect light anesthesia to reduce, not increase, blood pressure. Furthermore, the control animals also were anesthetized, and so I think we are covered there. Dr. Greiss also mentions that pregnancy experiences with at least three women who had preconceptional bilateral internal iliac and ovarian artery ligation have been reported, and this is true. There is no doubt that some women, and some baboons, have better collateral circulation than others. However, I would like to point out that toxemia of pregnancy has been reported in
Experimental toxemia
85
association with hypogastric aplasia and that hypertension and proteinuria have been found in pregnant women with stenosis of the aorta below the level of the renal arteries. The next point was about refinement in the model, and I again think that is an extremely important point. I think we should try to refine the model, and it certainly would be an improvement if, with interpretive arteriography or flow rates, we could assess accurately uterine blood flow, but I think that we are being quite reasonable in postulating that if the uterine arteries are clipped and the ovarian arteries are transected, especially in a pregnant animal, uteroplacental ischemia is produced. As a matter of fact, very recently, and not in this series, we thought we were very clever and decided that we would clip the uterine arteries and transect the ovarian arteries at 120 days of gestation. All three fetuses died in utero, and one mother developed gangrene of the uterus, so I believe there is no doubt that we are creating uteroplacental ischemia. This brings me to his next question regarding “overkill.” Does our experiment represent overkill? I think Dr. Greiss is correct and that there is probably a more intelligent way to experiment than our present method. The model is still relatively crude, but I think with suggestions such as we have had from Dr. Greiss this morning we will be able to improve it. With regard to Dr. Flowers’ comment on the effects of induced hypertension on the placenta, we are in the process of studying a considerable number of removed organs and biopsies from the animals. I do not have the answer to this at the present time, but we should have that in the near future.
CAVANAGH,
PAPINENI CHENG
S. C.
TIMOTHY
M.D.
RAO,
TSAI, C.
PH.D.,
B.V.Sc.
M.D.
O’CONNOR,
M.B.,
B.CH.
St. Louis, Missouri
In order to develop a model for the study of eclamptogenic toxemia, a series of experiments were carried out on 31 female baboons. In Group 1, consisting of IO animals, metal clips were placed around the uterine arteries in order to partially occlude them, and the ouarian vessels were transected. The animals were subsequently mated. Nine developed hypertension and proteinuria, and one aborted. The renal lesions in these animals were indistinguishable f rom those described in human toxemia. Group 2 consisted of three of the 10 baboons from Group I, which became pregnant a second time. They again developed hypertension and proteinuria. In Group 3, three baboons at 100 days of gestation were treated as in Group 1 with similar results. Groups 4 and 5 served as pregnant (3) and nonpregnant (15) controls. It is concluded that a toxemia model has been developed in a subhuman @mate. This model will prove useful in the further study of eclamptogenic toxemia. (AM. J. OBSTET. GYNECOL. 128: 75, 1977.)
ECLAMPTOGENIC TOXEMIA complicates approximately 6 per cent of all pregnancies in this country and remains a leading cause of maternal death.’ In addition, it has been linked to intrauterine growth retardation and to behavioral disorders in surviving children.’ Although most likely related to nutritional deficiency, the exact mechanism of pathogenesis is unknown, and the treatment remains symptomatic. An extensive review of the literature on the pathogenesis of toxemia was made in 1974 in association with a preliminary report,3 so this will not be repeated here. The present study represents an expansion of a multidisciplinary program designed to evaluate the concept that the “trigger mechansim” in the pathogehesis of toxemia is uteroplacental ischemia and that, by reducing the blood supply to the uterus, a toxemia animal model can be developed in the subhuman primate. Baboons were used in this study because their reproductive physi-
ology is remarkably similar to that of the human subject, and this certainly would make the findings more meaningful. The average weight of the animals used was approximately 14 kilograms.
Method and material A total of 31 female baboons (pap;0 anztbis) were divided arbitrarily into five groups, with three animals which had repeat pregnancies being used twice and each study animal serving as its own control. Pregnant study group (Group 1) This group consisted of 10 baboons. All of the baboons were sedated with an intramuscular injection of ketamine hydrochloride (Ketaset) in a dose of 5 mg. per kilogram of body weight. This was supplemented by intravenous administration of pentobarbital sodium as required to induce surgical anesthesia. Laparotomy was performed, and the uterine blood vessels were identified and isolated. Metal clips were placed around the uterine arteries, so that the uterine artery flow would be little changed after the procedure. Ovarian blood vessels were transected and ligated bilaterally, so that the main collateral blood supply of the uterus was interrupted. With the ovarian arteries ligated, almost all of the blood which supplied the uterus had to flow through the “clipped” uterine arteries, and the tnetal clips would not allow hypertrophy during pregnancy. Thus, this group of study animals was predisposed to
From the Departments of Obstetrics-Gynecolog and Pathology, St. Louis University School of Medicine. Supported in part by United States Public Health Seruice Grants FR 5388-08 and FR 5388-09 and by National Foundation-March of Dimes Grant I-335. Presented at the Eighty-seventh Annual Meeting of the American Association of Obstetricians and Gynecologists, Hot Springs, Virginia, September 9-11, 1976. Reprint requests:Dr. Denis Cavanagh, Department of Obstetrics and Gynecology, College of Medicine, University of South Flon’dn, Tampa, Florida 33620.
75
76 Cavanagh et al.
Fig. 1. Radiograph showing pregnant baboon near term with metal clips in place around the uterine arteries.
uteroplacental ischemia by preconceptional clipping. These animals were then observed for the crythematous perineal swelling indicative of ovulation. When this was noted, the female animal was transferred from her cage to that of her male consort and remained there for a period of approximately four days. The female animal was then returned to her cage and observed throughout the next menstrual cycle. If evidence of ovulation occurred, the mating process was repeated. This course was followed until all of the animals were pregnant. An x-ray of the abdomen in late pregnancy was used to confirm that the metal clips were still in position (Fig. 1). Repeat pregnancies (Group 2). After cesarean section, some of the baboons of Group 1 were killed so that we could obtain the tissues for various histopathologic studies. Following wound healing and rethe remaining baboons were allowed to cuperation, become pregnant. Three of these animals became pregnant for the second time and carried the fetus to term. Antepartum study group (Group 3). Three baboons were used in this group in which the experimental technique was similar to that of Group 1, with the main difference being that these baboons were at the gesta-
tional age of approximately 100 days. All three animals carried the pregnancy to term and Iverc delivered by cesarean section. Pregnant control group (Group 4). This group consisted of three baboons which underwent a “sham” operation in which the ovarian arteries were ligated and transected and metal clips were placed around the uterine arteries and then removed. All of these animals became pregnant and went to term. Nonpregnant control group (Group 5). This group consisted of 1.i baboons which had no dissection around the uterine or ovarian arteries. None was mated, and these animals ivcre used to provide baseline nonpregnant control data. In addition to gross maternal and fetal outcome, four methods were used to evaluate the effect of the experimental technique on pregnancy outcome. Hemodynamic studies. The animals were anesthetized with ketamine hydrochloride supplemented by sodium pcntobarbital. ‘They were then placed on an operating table. Pressures were measured by Statham transducers, and a Sanborn recorder (7700 series) was used for recording aortic pressure. Aortic pressure was measured b! passing a polyethylene catheter along the femoral arter! into the aorta to the level of the renal arteries. A poi~ethylcne catheter bias passed along the femoral vein into the inferior vena tava so that the tip came to rest just proximal to the right atrium. This venous catheter was used for the injection of (Zardiogreen (indoc.);anine green) tar the cardiac output estimation. Estimation of cardiac output was accomplished by the dye-dilution method. Dilution curves from the arterial catheter were recorded by continuous sampling througtr a c,uvette densitome& (Gilson Medical Electronics. 1nc.), which ~vas connected to a directreading cardiac- output computer (Lexington Instruments (;orp.) and a direct strip chart recorder. ‘The computer was calibrated and programmed to read out cardiac output directly f-rom the densitometer and reject all cur\‘es not arithmetically valid. Total peripheral resistance was calculated as described by Guyton.’ Standard elec.trocardiographic leatfs were used to rpcord cardiac rate and rhythm. Urinalysis. Urine teas obtained by catheterizing the bladder, and the levels of protein, sugar, and pH were determined by Labstis reagent strips (Ames <:ompany).
Histopathologic studies. The kidneys were removed from three mothers in Group 1, and biopsies were vbtained from the remaining pregnant animals except those in Group 2, at or immediately after delivery. Tissues were fixed in 10 per cent formalin fixative and
Volume Number
Experimental
128 1
Fig. 2. Photomicrograph
of renal biopsy specimen showing glomerular enlargement of the glomerular capillary lumina due to cytoplasmic swelling is indistinguishable from that found in human eclamptogenic
larity, with reduction lial cells. This lesion
Table
I. Hemodynamic
parameters
in animals
Systolic pressure (mm. Hg) Diastolic pressure (mm. Hg) Mean aortic pressure (mm. Hg) Cardiac output (mUmin.) Peripheral resistance (dynes sec./cm.‘)
Table
II. Results
of urinalysis
Protein Sugar
(mg./
embedded with
in
paraffin,
hematoxylin
Schiff
and
Small
and
and
5 p sections
eosin
and
blocks
of renal and
cortex
embedded
tions
were
stained
with
and
viewed
with
a Phillips
were acetate
electron
3
Group
4 Group 5 (controls)
(pregnant
clipping)
controls)
160 100 120 1.913 5,307
120 75 90 2,000 3,510
*
stained
periodic
fixed
acid-
in 4 per Extra-thin
and
Group 3 (antepartum clipping) Negative Negative
7.0
were
with
in epon.
uranyl
Pwgnancy)
Trace to 100 Negative
in
hematoxylin.
glutaraldehyde
Group (antepartum
croup 2 (repeat Pw-cy)
30 to 300 Negative (except one baboon) 5.5
PH
and hypercelluof the endothetoxemia.
136 82 100 1,635 5,375
studied
clipping) 100 ml.)
2
(repeat
155 107 123 1,400 7,256
Group I (preconceptional Urinalysis
Group
167 107 127 1,479 7,946
in animals
77
studied
Group I (preconceptional clipping)
Hemodynamics
toxemia
lead
microscope.
cent sec-
citrate
to 300
Group 4 (pregnant control}
(controls)
Negative Negative
Negative Negative
7.0
6.0
Immunofiuorescence goat
anti-human
IgE,
albumin,
and
Behring
used
for
reactivity
(Meloy
was
direct
6.0
G (IgG), (C3
Laboratories,
Diagnostics, the
5
study. Fluorescein-conjugated
immunoglobulin complements
fibrinogen
Group
C4),
IgA, and
Springfield,
Somerville,
immunofluorescent
confirmed
and
by
New
IgM, fibrin/
Virginia, Jersey)
studies.
immunoelectrophoresis
were Cross-
78 Cavanagh et al.
May
1, 1977
Am. J. Obsfet. Gynecol.
Fig. 3. Electron micrograph of the kidney of a “toxemic” baboon showing the endothclial cell nucleus (EN) and narrowing of the glomerular capillar) lumina (CL) as a result of cytoplasmic swelling and subendothelial deposition (SED) of fibrillary StrucTures. This is similar to the pictur-e seen in human eclamptogenic toxemia. (Original magnification x6.960.) and radial immunodiffusion between human immunoglobulins and complements and those of the baboon. Tissues were rapidly frozen in liquid nitrogen. Cryostat sections, 6 p thick, were fixed in ethanol: ether (1: 1) for 10 minutes. They were then placed in 95 per cent ethanol for 20 minutes. They were washed off in buffered saline for five minutes. and this washing was repeated. They were then incubated with fluorescein-conjugated antisera, washed in phosphate-buffered saline, and mounted on a cover slip with buffered glycerin. Reactivity of tissues with antisera to human immunoglobulins, complement components C3 and C4. fibrin/fibrinogen were studied. The resulting immunofluorescence was quantitated on a scale of 0 to 4+. The sections were photographed with a Leitz. ortholux fluorescence microscope .
pregnant animals. Sign&rant differences were observed during the third trimester of pregnancy between Groups 1.2, 3, and the “sham’‘-operated animals (Group 4). l‘he findings in 15 nonpregnant baboons (Group 5) were obtained and compared with those in the four groups of pregnant animals. Hemodynamic studies. The results of these are summarized in Table I. In Group 1 the animals became progressively hypertensive during the third trimester of pregnancy. ‘l‘he mean blood pressure value prior to pregnancy teas 117168 mm. Hg, with a mean
Three
Results All 10 baboons in Group 1 became pregnant but one aborted. No abortions occurred in the other groups of
arterial
pressure
of‘ 84 mm.
Hg.
Near
term,
the
mean blood pressure value was 1671107 mm. Hg. with a mean arterial pressure of- 127 mm. Hg. The mean values for cardiac output and total peripheral resistance were 1,479 ml. per minute and 7,946 dynes sec./ cm.S, respectively. of
pregnant
tact around Group
the
for
nine
the
baboons
second
the uterine
2. Near
term,
time.
arteries, they
had
from
Group
1 became
clips inand were designated
with
the
a mean
metal
blood
pressure
Volume Number
Experimental toxemia
128 1
79
Fig. 4. Electron
micrograph of the kidney of a “toxemic” baboon showing mesangial cell nucleus (MN) and markedly increased glomerular mesangial matrices (MM) with two large, fine, granular, electron-dense deposits (EDD). (Original magnification x 16,770.)
of 155/107 mm. Hg, with a mean arterial pressure of 123 mm. Hg and a mean cardiac output of 1,400 ml. per minute. The mean total peripheral resistance was slightly less than that of Group 1. Three baboons were used in Group 3 ‘with the experimental technique being identical to that used in Group 1, except that these baboons were at approximately 100 days of gestation when operated upon. All
three carried the pregnancy to term and were delivered by cesarean section. The mean blood pressure at the time of application of the metal clips was 97/57 mm. Hg, with a mean arterial pressure of 70 mm. Hg. At the time of cesarean section, the mean pressure was 160/ 100 mm. Hg, with a mean arterial pressure of 120 mm. Hg. (All of these animals developed hypertension and, one developed severe hypertension [2 1O/ 130 mm.
80
Cavanagh
et al.
Fig. 5. Electron
micrograph of a renal biopsy taken from a normal pregnant control baboon. This shows the capillary lumen (CL), basement membrane (BM), endothelial cell nucleus (EN). mesangial matrix (MM). and mesangial cell nucleus (MNI. (Original magnification x9.570.)
W
and3.
disc
output
+ proteinuria.)
peril Dheral resistance Gl -0up 4 consisted undo et-went
In this
was 1.9 13 ml. per
a “sham”
of 5,307 of three
group,
minute,
1 ‘he mean
operation
blood
mean
car-
mm.
a mean
total
representing
dynes sec./cm.” baboons. These in
artel =ies we :re ligated and transected, clips were 1Aaced around the uterine remc -wed.
the with
pressure
which and arteries at term
the
Hg
with
a mean
arterial
no significant
animals
ping. The mean cardiac ute with a mean total
ovarian
dynes
the metal and then was 120/75
Group
pressure rise
output peripheral
from
of 90 ml the
tin:
was 2,000 ml. resistance
-. Hg,
IC ( If
clip-
PC :r minof 3.4 10
sec.lcm.’ five
Urinalysis. 31. Proteinuria
consisted of The findings was
not
15 nonpregnant are summarized
detected
in any
c( mt rols. in Table of th ea inimals
Volume Number
Experimental toxemia
128 1
Fig. 6. Immunofluorescent photomicrograph showing extensive deposition of fibrin/fibrinogen the glomerular mesangium. and capillary vessel walls. This lesion is indistinguishable from found in human eclamitogenic toxemia. i x250.) during the first or second trimester of pregnancy. It was apparent only during the third trimester in Groups 1, 2, and 3. In Group 1, tested at term, the values ranged from 30 to 300 mg. per 100 ml.; in Group 2, a trace to 100 mg. per 100 ml.; in Group 3, negative to 300 mg. per 100 ml. Urinary glucose was absent in all of the animals except one in Group 1. The pH values ranged from 5.5 to 7.0. Histopathologic studies. Renal biopsies were obtained at term from five animals in Group 1 and from all other animals except those in Group 2. Light microscopy. The glomeruli in the biopsy specimens appeared equaliy affected in pregnant Groups 1 and 3 with hypertension and proteinuria. They were somewhat enlarged and hypercellular. The glomerular capillary walls were thickened, but the basement membrane proper was not demonstrably thickened even when special stains were used. The glomerular capillary lumina were reduced by cytoplasmic swelling of the endothelial cells (Fig. 2). Polymorphonuclear leukocytes were sparse or absent within the glomeruli. The epithelial cells and tubules were within normal limits when compared with those in biopsies from the pregnant and nonpregnat control groups. Electron microscopy. Renal biopsies were studied from animals in all groups except Group 2. The glomerular basement membrane was not thickened on electron
81
in that
microscopic study. The cytoplasm of the endothelial cells was swollen. The subendothelial spaces were widened, and depositions of loose fibrillary material were also found. As a result of these factors, there was narrowing of the glomerular capillary lumina (Fig. 3). Epithelial cells showed only focal fusion of foot processes especially over the paramesangial areas. The mesangial cells were hyperplastic, and the mesangial matrices were increased. There were fine granular electron-dense deposits found over the mesangial region in the animals with severe hypertension and proteinuria (Fig. 4). For purposes of comparison, an electron micrograph of a kidney from a normal pregnant baboon is shown in Fig. 5. Immunofluorrscencu studirs. Depositions of IgM, C3, C4, and fibrin/fibrinogen were found in the animals with hypertension and proteinuria in Groups 1 and 3. The distribution was similar to that described in human toxemia of pregnancy. The deposits were located in the glomerular mesangium and extended into the peripheral capillary loops, which give the impression of a “wire-loop” appearance. They were smooth, homogeneous, and evenly stained. Fine granular deposits were occasionally found. Rarely, fibrin thrombi were noted in the glomerular capillary Iumina. Fig. 6 shows extensive deposition of fibrin/fibrinogen in the glomerular mesangium and capillary vessel walls. Fine granular staining is evident in the renal glomeruli. Fig.
82
Cavanagh et al. Am
Fig. 7. Immunofluorescent photomicrograph similar to that for fibrin/fibrinogen. This eclamptogenic toxemia. (x250.)
Fig. 8. Immunofluorescent kidney of the “toxemic” (X100.)
showing extensive lesion is indistinguishable
deposition from
May 1, 1977 J. Obstet. G~nmx~l.
of IgM in a pattern that found in human
photomicrograph demonstrating the presence of complement (C4) in the baboon in a distribution pattern similar to that of IgM and fibrin/fibrinogen.
Volume Number
Experimental toxemia
128 I
7 shows that tion to that presence of tern similar
the IgM deposition is similar in distribuof fibrin/fibrinogen. Fig. 8 shows the complement (C4) with a distribution patto that of IgM and fibrin/fibrinogen.
Comment Ciinical eclamptogenic toxemia usually occurs after the twentieth week of pregnancy. It is characterized by hypertension (mean arterial pressure rise greater than 20 mm. Hg), edema, proteinuria, reduced renal artery flow with a reduced glomerular filtration rate, some evidence of disseminated intravascular coagulation, and distinctive renal glomerular changes. In the previously reported pilot study,3 we reported the development of hypertension and proteinuria with a reduction in renal artery flow in pregnant study animals which had been “clipped” prior to conception. In addition, fibrin/fibrinogen deposits were confirmed in the pregnant study group with immunofluorescence studies, although no evidence of disseminated intravascular coagulation could be found in four animals in which repeated coagulation studies were performed. The serum uric acid was elevated in the pregnant study
83
group as compared with the pregnant controlf group. Two obvious limitations in the pilot study vere the small number of animals in the study group (Qve) and our failure to demonstrate the typical histopathologic renal lesion of “toxemia” on light or electron microscopy in the only animal studied. In the present study, in baboons predisposed to uterine ischemia by preconceptional operation or in which uterine ischemia is produced during pregnancy, we have confirmed that the animals develop hypertension and proteinuria. In addition, with light microscopy, electron microscopy, and immunofluorescence studies, the renal lesion is indistinguishable from that found in women with eclamptogenic toxemia.5-‘0 After eight years of work on this animal model, we feel justified in concluding that although refinements are required a primate experimental model is now available for studying the pathogenesis of and evaluating new methods of treatment for toxemia of pregnatlcy. The availability of such a model could bring great dividends in health care for mothers and babies afflicted by this serious and common complication of pregnancy.
REFERENCE’S Statistics of the United States, Washington, D. C., 1976, vol. 24, No. 13, Department of Health, Education and Welfare.
1. Vital
2. Pasamanick, B.: The epidemiology of reproductive casualty, in Conference on Obstetric Factors in Child Development, Atlanta, Emory University School of Medicine. 3. Cavanagh, D., Rao, P. S., Tung, K. S. K., and Gaston, L.: AM. J. OBSTET. GYNECOL. 120: 183, 1974. 4. Guyton, A. C.: Textbook of Medical Physiology, ed. 5, Philadelphia, 1976, W. B. Saunders Company.
Discussion
DR. FRANK C. GREISS, Winstom-Salem,
North Carolina. In his continuing efforts to develop an experimental model of pregnancy toxemia, Dr. Cavanagh has amplified his initial report of 1974 and has added new data to demonstrate the similarity of the renal concomitant of human pre-eclampsia-eclampsia to that of his model. I am not a nephrologist and shall confine my comments to the cardi&ascular aspects of the report. Dr. Cavanagh has accumulated experience with 14 baboons whose uterine arteries were clipped prior to conception and with three baboons whose uterine blood supply was compromised at 100 days of gestation. It would appear that two of the three classic signs of toxemia, hypertension and albuminuria, have been induced. One would like assurance that environmental
and
mental
groups,
dietary
factors
and I would
were
similar
in the
be particularly
experi-
interested
5. Spargo, B., McCartney, C. P., and Winemiller, R.: Arch. Pathol. 68: 593, 1959. 6. McCartney, C. P.: Circulation (Suppl. 2) 30: 37, 1964. 7. Morris. R. H. Vassali. P.. Belier. F. K.. and McCluskev. ,j R. ’ T.: Ob$tet. Gynecol. i4:’ 32, 1964. 8. Altchek, A.: J. A. M. A. 175: 791, 1961. 9. Altchek, A., Albright, N. L., and Sommers, S. C.: Obstet. Gynecol. 31: 595, 1968. 10. Petrucco, 0. M., Thomson, N. M., Lawrence, J. R., and Waldon, M. W.: Br. Med. J. 1: 473, 1974.
in the patterns of weight gain, edema, and blood volume changes in the baboons. An obvious criticism of these experiments is that observations were made under anesthesia. Such methodology is understandable when one works with large subhuman primates, but observed responses may be variably altered. In view of these factors, mean values of blood pressure levels and albuminuria may be deceptive, and a breakdown of individual changes would be helpful to permit an estimate of the range of responses. Our prime attention should be directed to the thesis being tested, that uteroplacental ischemia induces pregnancy toxemia. The pregnancy experiences of at least three women who had preconceptual bilateral internal iliac and ovarian artery ligations have been reported.’ None developed toxemia of pregnancy. It is evident that sufficient collateral circulation developed
84
Cavanagh et al.
to support normal pregnancy. Perhaps similar collaterals developed in Dr. Cavanagh’s baboons or perhaps the baboon lacks the capacity to develop adequate collateral circulation. Dr. Cavanagh cannot presume that uterine artery clipping per se reduces uteroplacental blood flow. This presumption must be documented by objective How determinations in future experiments! ‘l‘aking the opposite tack, one could infer that the experimental method presently employed may represent “overkill” since compromise of the spiral arteries supplying the intervillous space would, in itself, be adequate to limit placental blood How. The observations of DeWolf and associates’. ’ are pertinent in this regard. They have observed replacement of the walls of spiral arterioles by trophoblast cells in normal pregnancies. These changes extend into the myometrial segments of the spiral arteries, creating widely dilated blood vessels conducive to high volume How. In preeclamptic patients, these “physiologic” changes do not extend to myometrial levels, and placental vascular resistance may be increased thereby. Thus, a mechanism which might limit placental blood flow has been observed, one which would have its inception during the time of definitive placentation. Dr. Cavanagh has a unique opportunity to evaluate further such vascular changes in his “toxemic” baboons. I have enjoyed reviewing Dr. Cavanagh’s paper, and I would exhort him to continue his work in an area of such obvious future potential. REFERENCES
Mengert, W. F., Burchell, R. C., Blumstein, R. W., and Daskal, J. L.: Pregnancy after bilateral ligation of the internal iliac and ovarian arteries, Ob. Gynecol. 34: 664, 1969. DeWoIf. F., DeWolf-Peeters, C., and Brosens, I.: Ultrastructure of the spiral arteries in the human placental bed at the end of normal pregnancy. AM. 1. OBSTET. GYNECOL. 117: 833,
1973.
1
-
-
DeWolf. F.. Robertson. W. G.. and Brosens, I.: The ultrastructure of acute atherosis in hypertensive pregnancy, .4~. J. OBSTET.
GYNECOI..
123:
164, 1975.
DR. CHARLESE. FLOWERS,JR., Birmingham, Alabama. I sincerely congratulate Dr. Cavanagh and his co-workers for a thoughtfully conceived and superbly executed study that significantly adds to our knowledge of the hypertensive complications of pregnancy. It is my belief that they have created an experimental model that reproduces the clinical syndrome we observe in the young primiparous patient with hypertension and without pre-existing vascular or glomerular pathology. I am particularly impressed with the electron microscopic studies of the glomerular capillary narrowing the subendothelial deposition of fibrin which is indistinguishable by immunofluorescence from the renal lesion seen in eclamptogenic toxemia in the human subject.
It is the young, short, poorly nourished, primiparous patient who so frequently develops hypertension in the first pregnancy but remains symptom-free in subsequent pregnancies. These studlcs would indicate that she has a uterine vascular system that is inadequatc~ to respond to the requirements of gestation. Hcmcvct-. the uterus, having become hvpertrophied in the hlitial pregnancy, is able to adjust acleqtiately to stibscclttent pregnancies. The uterine ischemia is associated with placental release of thromboplasdn. \\.hich initiates subclinical disseminated intravascular c,oagulation that is so mimimal it cannot be detected by present laboratory mcrhods. Within the kidney, fibrin deposition. ca$llarv narrowing, and increased glomerular mesangial matrices occur and produce an increase in renal r&tam r- al1t1 a reduction in renal blood How and cardiac output. These events pathologically alter angiotensin and aldosterone production. The degree of thext, alte.l.alic,m dictate the severity of the disease. its response to bed rest, and the ability of the kidney to rcco\er ccnnpletely. It may {bell be that some (;t our primiparous patients who develop hypertension in prcgnancl have such extensive fibrin deposition and renal danlage that hypertension of renal origin develops in suhsrquent pregnancies. This may be the reason for- tlic development of hypertension in Group ‘i’ of the esppt imental subjects. Although the explanation which I have ol’f~rcd places the major emphasis upon rrnal reaction IO glomerular fibrin deposition, it does not preclude the additional possibility that uterine ischemia results in the production of vasopressor substances and/or prostaglandins from the decidua and the placenta. Hopefully, this experimental model will eventually elucidate more of the pathophysiology of this disease and provide answers to some of the controversial issues of treatment such as the effect of magnesium s111f;tte, antihypertensive clrugs, diuretics. and cpidural anesthesia upon uterine blood How. Dr. Cavanagh did not indicate the effects of the induced hypertension upon the placenta and thr vital organs of the mother or whether these lesions are rcversible as thcv are in the human subject. I would he interested in these findings as uell as in hearing a of his present interpretation of the summary pathophysiology of the various tvpes of hypertension in pregnancy. DR. CAVANAGH(Closing). Regarding environmental and dietary factors, the environmental and dietary factors were identical in all groups of animals studied. With regard to weight gain, patterns of we-ight gain, and the weight gain in the toxemic versus the normal pregnant control animals, it is my impression that the toxemic baboons gained more weight, but this did not reach statistical significance. It is extremely difficult to
Volume Number
128 1
find curves on baboons, and so we cannot really give an unequivocal answer that weight gain was increased with the toxemic group at this time, but I suspect it is. With regard to edema, that is hard to assess in the baboon, and, if you ever, try to detect edema in a pregnant baboon, you will, I think, agree with me. However, we should try to find some way of assessing edema. I think these are very pertinent comments, especially Dr. Greiss’ question, did we perform blood volume studies in the baboons? I regret to say that the answer is no. But I think that would be an extremely pertinent study to perform in these toxemic animals. The problem was that we were determining cardiac output and also wanted to estimate total peripheral resistance, and we were afraid that the two methods would conflict. But blood volume studies are planned for the future, and I think the suggestion is excellent. With regard to anesthesia, there certainly is no way we are going to work with unanesthetized baboons without a restraining chair. We do not happen to have one, so all of the studies were done on anesthetized animals, as Dr. Greiss mentioned. However, the animals were lightly anesthetized, and, if anything, I would expect light anesthesia to reduce, not increase, blood pressure. Furthermore, the control animals also were anesthetized, and so I think we are covered there. Dr. Greiss also mentions that pregnancy experiences with at least three women who had preconceptional bilateral internal iliac and ovarian artery ligation have been reported, and this is true. There is no doubt that some women, and some baboons, have better collateral circulation than others. However, I would like to point out that toxemia of pregnancy has been reported in
Experimental toxemia
85
association with hypogastric aplasia and that hypertension and proteinuria have been found in pregnant women with stenosis of the aorta below the level of the renal arteries. The next point was about refinement in the model, and I again think that is an extremely important point. I think we should try to refine the model, and it certainly would be an improvement if, with interpretive arteriography or flow rates, we could assess accurately uterine blood flow, but I think that we are being quite reasonable in postulating that if the uterine arteries are clipped and the ovarian arteries are transected, especially in a pregnant animal, uteroplacental ischemia is produced. As a matter of fact, very recently, and not in this series, we thought we were very clever and decided that we would clip the uterine arteries and transect the ovarian arteries at 120 days of gestation. All three fetuses died in utero, and one mother developed gangrene of the uterus, so I believe there is no doubt that we are creating uteroplacental ischemia. This brings me to his next question regarding “overkill.” Does our experiment represent overkill? I think Dr. Greiss is correct and that there is probably a more intelligent way to experiment than our present method. The model is still relatively crude, but I think with suggestions such as we have had from Dr. Greiss this morning we will be able to improve it. With regard to Dr. Flowers’ comment on the effects of induced hypertension on the placenta, we are in the process of studying a considerable number of removed organs and biopsies from the animals. I do not have the answer to this at the present time, but we should have that in the near future.