- Email: [email protected]
Low-dose aspirin
Clinical Articles Low-dose aspirin I. Effect on angiotensin II pressor responses and blood prostaglandin
concentrations in pregnant women sensitive to angiotensin II Bernard Spitz, MD, Ronald R. Magness, PhD, Susan M. Cox, MD, Charles E. L. Brown, MD, Charles R. Rosenfeld, MD, and Norman F. Gant, MD Dallas, Texas Decreased incidence of proteinuric hypertension after low-dose aspirin therapy is hypothesized to be a consequence of selective thromboxane A2 inhibition and sparing of prostacyclin. This study was designed to ascertain if low-dose aspirin therapy (81 mg/day for 1 week) alters vascular refractoriness to angiotensin II and the prostacyclin/thromboxane A2 ratio in pregnant women sensitive to angiotensin II (n = 17). Low-dose aspirin increased the effective pressor dose of angiotensin II from 5.9 ± 2.4 to 10.2 ± 5.5 ng/kg/min (p < 0.01, mean ± SD). Platelet-derived serum thromboxane 8 2 (a metabolite of thromboxane A2 ), a measure of therapy compliance, decreased from 1804 ± 1771 to 132 ± 206 pg/ml (p < 0.01). Plasma thromboxane 8 2 decreased from 130 ± 107 to 19 ± 12 pg/ml (p < 0.01). Inhibition was not selective because 6-keto-prostaglandin F'" (a metabolite of prostacyclin) also decreased from 243 ± 90 to 163 ± 90 pg/ml (p = 0.039) and prostaglandin E2 was reduced from 155 ± 67 to 95 ± 40 pg/ml (p = 0.014). Decreases in thromboxane 8 2 , however, were significantly greater (75% ± 19%) than decreases in 6-keto-prostaglandin F, (21% ± 33%) or prostaglandin E2 (29% ± 36%); thus the 6-keto-prostaglandin F'"/thromboxane 8 2 ratio increased from 3.1 ± 2.0 to 12.4 ± 9.9 (p < 0.01 ). Although low-dose aspirin increases the effective pressor dose of angiotensin II, it does not return to normal pregnancy values. This observation is consistent with the hypothesis that this represents only a partial selective prostaglandin inhibition. (AM J OssTET GYNECOL 1988;159:1035-43.) 0
Key words: Low-dose aspirin, preeclampsia, angiotensin II, pressor responses, prostaglandins, blood pressure control, baroreceptor
Gestational proteinuric hypertension, or pregnancyinduced hypertension, may be associated with an imbalance between prostacyclin and thromboxane A 2 • These eicosanoids have potent and opposing effects on vascular reactivity and platelet aggregation. 13 It has been suggested that this imbalance is a contributing factor in the loss of refractoriness to the pressor effects of angiotensin II and platelet aggregation, which characterize this syndrome.4 5 Because thromboxane A 2 is From the Departments of Obstetrics and Gynecology and Pediatrics, The University of Texas Southwestern Medical Center at Dallas. Supported by National Institutes of Health grants HL34150, HD08783, and HD24971, and by American Heart Association Texas Affiliate 86R-071. Received October 13, 1987; revised April 28, 1988; accepted May 22, 1988. Reprint requests: Norman F. Gant, MD, Professor of Obstetrics and Gynecology, The University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd., Dallas, TX 75235.
released by circulating platelets,' it is logical to consider the use of antiplatelet agents to prevent or treat pregnancy-induced hypertension. Two groups of investigators"· 7 have shown a decreased incidence of pregnancy-induced hypertension in women given cyclooxygenase and/or plateletinhibiting drugs. In a controlled trial by Beaufils et al.," aspirin (150 mg/day) combined with dipyridamole (300 mg/day) was tested in women at the end of the first trimester of pregnancy. They were selected to participate because their previous obstetric histories indicated they were likely to develop hypertension. This antiplatelet regimen conferred a clear benefit in terms of the occurrence of gestational protein uric hypertension, intrauterine growth retardation, and perinatal survival. In a subsequent report by Wallenburg et al.," only primigravid women were selected. In these randomized placebo-controlled and double-blind studies, treatment 1035
t\ovember l 988
1036 Spitz et al. Am
J Obstet Gynecol
Table I. Accuracy and precision of plasma prostaglandin assays Concentration (pg I ml)
I
25
Prostaglandin added
5
Thromboxane B2 6-keto-prostaglandin F1. Prostaglandin E2
6.6 ± 4.2 2.2 ± 1.5
24.2 ± 6.1 27.4 ± 2.9
6.4 ± 8.3
23.0 ± 9.8
Values are means ± SD (n
=
I
100
Correlation coefficient
Buffer blank (pg/ml)
50.5 ± 17.4 40.2 ± 18.6
104.8 ± 2.9 114.l ± 60.3
0.999 0.987
7.6 1.6
49.6 ± 10.5
113.0 ±
7.3
0.997
5.3
50
I
6).
with aspirin (60 mg/day) was associated with a significantly lower incidence of pregnancy-induced hypertension. In none of these studies were the concentrations of circulating prostanoids reported. Thus the hypothesis that restoration of angiotensin II refractoriness and improved clinical outcome were consequences of selective inhibition of thromboxane A 2 and sparing of prostacyclin remains unproved. The aim of this study was to determine the effects of low-dose aspirin on changes in vascular reactivity to angiotensin II and whether such changes could be explained by alterations in circulating prostacyclin and thromboxane A2 • Material and methods
In this study we determined the effect of 81 mg of aspirin per day for 1 week beginning at approximately 32 weeks' gestation on (1) the minimum amount of angiotensin II per kilogram per minute required to increase diastolic blood pressure 20 mm Hg (the effective pressor dose) in 22 women sensitive to angiotensin II (effective pressor dose< 10 ng/kg/min), (2) the concentration of thromboxane A 2 in serum, and (3) the ratio of plasma prostacyclin/thromboxane A 2 • We obtained angiotensin II pressor responses in 25 nonsensitive women who served as controls, to provide current data in regard to the effective pressor dose of angiotensin II in nonsensitive women and to permit us to determine whether aspirin returned the angiotensin II pressor response to normal pregnancy values. Sensitivity to angiotensin II is defined as a pressor dose< 10 ng/kg/min, and refractoriness is defined as a pressor dose:;,,, 10 ng/kg/min.4 Nonsensitive women were not treated with aspirin because they were refractory to angiotensin II. The selection of women sensitive to angiotensin II was based on the observations of Gant et al.4 that 50% to 90% of such patients will ultimately have pregnancy-induced hypertension. This study was approved by the Institutional Review Board of The University of Texas Southwestern Medical Center at Dallas, and all study subjects gave informed written consents. To locate women sensitive to angiotensin II, angiotensin II sensitivity tests were performed in a high-risk population recruited from two sources. First, women
admitted to Parkland Memorial Hospital (Dallas, Texas) and the High-Risk Antepartum Care Unit with early onset of pregnancy-induced hypertension (who subsequently became normotensive) were studied. Second, women with positive "roll-over tests" were selected from the antepartum clinics. This test has been used to identify women in whom hypertension is likely to develop later in pregnancy, 10 and it was useful in helping to discover women sensitive to angiotensin II who were candidates for this study. Women were excluded if they were known to have taken drugs such as aspirin 11 or theophylline 12 that could influence either test. Patients with a history of chronic hypertension also were excluded from this study. Diets were unrestricted. The angiotensin II sensitivity test was performed as previously described.4 Women were placed in the left lateral recumbent position in a quiet room. Blood pressure and heart rate were automatically measured from the right arm every 4 minutes by means of an oscillometer (Spacelabs 512D; E. R. Squibb & Sons, Princeton, N.J.). After stabilization of the diastolic blood pressure for at least 15 minutes, an intravenous infusion was started in the left arm. This was connected to an infusion pump with a syringe containing angiotensin II amide (1000 ng/ml in 0.9% sodium chloride, CibaGiegy Pharmaceuticals). Infusion of angiotensin II was started at a rate of approximately 1.5 ng/kg/min and was gradually increased at 4-minute intervals. The test was stopped after the effective pressor dose response was observed. This procedure was repeated after 5 minutes; results in all studies were identical to the first pressor dose response. Data from the two studies were averaged and used as a single observation for statistical analysis. Before the angiotensin II infusion was started, blood was collected for analysis of serum and plasma prostaglandin concentrations. Specific methods for the measurement of platelet-derived serum thromboxane B, (the stable metabolite of thromboxane A 2 ), plasma thromboxane B 2 , 6-keto-prostaglandin F 1• (the stable metabolite of prostacyclin), and prostaglandin E 2 are described below. Prostaglandin assays. Serum concentrations of thromboxane B, reflective of platelet-derived prostanoid were obtained from samples after clotting (45 min-
Low-dose aspirin and angiotensin II presser response
Volume 159 Number 5
20
c ·e
i
20
NULLIPAROUS PAROUS
18
ALL
16
15
c
-
:!
Cl
~
UJ
5
Cl
E
....
0
c
0
0:: 0
UJ
....0:: 0.. ....>
VI VI
0::
a...
UJ
>
i==
10
8
j::
u
u.. u..
12
VI
0
10
Ill Ill
UJ
14
Cl
Ill
0::
- - - NULLIPAROUS - - - - - PAROUS
0. ..le
Cl
1037
....'-'
5
LL.
LL. ....
UJ
6
A 4
"' ';: 0
z
Fig. I. Effective pressor doses in nonsensitive (A) and sensitive women before (B) and after (C) aspirin treatment. (A > B, p < 0.001; B < C, p = 0.037; C
utes at room temperature) and were produced in quantities of 1 to 10 ng/ml; thus a direct assay was performed on the serum. To measure thromboxane B2 , we used standard (up to 500 pg) and unknown amounts of thromboxane B2 mixed in 0.1 moll L phosphatebuffered saline solution (pH = 7.4) with 0.1 % polyvinylpyrrolidone. Duplicate samples of serum (0.025 to 0.lml) were assayed by the addition of 0.1 ml antithromboxane B2 antiserum ( 1 : 16,000 titer) and 0.1 ml 3 H-thromboxane B2 .(New England Nuclear; 12,000 dpm) and were incubated for 12 to 18 hours at 4° C. Antibody-bound thromboxane B2 was separated with dextran-coated charcoal. To measure plasma prostaglandins, the following methods were used. Blood samples (6.0 ml) were drawn into chilled syringes, transferred immediately to chilled tubes containing ethylenediaminetetraacetic acid (4.2 mg) and aspirin (0.3 mg), centrifuged, and stored at - 20° C. Measurement of plasma prostaglandins was by the methods of Campbell and Ojeda. 13 To determine recovery, the appropriate tritiated prostaglandin (3000 dpm) was added to 2 ml acidified plasma (pH = 3.0; glacial acetic acid) and extracted on individual Bond Elut C-18 cartridges (Analytichem International, Harbor City, Calif.). Bond Elut C-18 cartridges, which had
2 BEFORE ASPIRIN
AFTER ASPIRIN
0 Fig. 2. Effective pressor dose in women sensitive to angiotensin II before and after aspirin treatment. (Paired ttest: p < 0.001.)
been washed first with IO ml 95% ethanol and then with IO ml distilled water, were slowly loaded with acidified samples. They were then washed with 5 ml distilled water and 5 ml petroleum ether and were eluted with 5 ml ethyl acetate. The extract was dried and purified by silicic acid chromatography that used solvents of increasing polarity (e.g., toluene : ethyl acetate : methanol). The first fraction, which contained free fatty acids and prostaglandins A2 and B2 , was discarded. The second fraction contained thromboxane B2 , 6-keto-prostaglandin F 1a, and prostaglandin E 2 ; the eluate was dried under vacuum (37° C), reconstituted in phosphate-buffered saline solution, and assayed for thromboxane B2 , 6-keto-prostaglandin F 1a, and prostaglandin E 2 by radioimmunoassay. The assay for thromboxane B2 was the same as that described above for the serum. The 6-keto-prostaglandin F 1a assay procedure used standard (up to 200 pg) and unknown quantities of 6-keto-prostaglandin F 1a (0.1 ml) mixed in 0.1 mol/L phosphate-buffered saline solution (pH = 7.4) with 0.1 % polyvinylpyrrolidone prepared in duplicate aliquots. Antiserum (0.1 ml, 1: 4,000 titer) and 0.1 ml 3 H-6-keto-prostaglandin F 1a (New England Nuclear; 12,000 dpm) were added and the tubes were incubated at 4° C for 12 to 18 hours. Bound ligand and
November 1988
1038 Spitz et al.
J Obstet Gynecol
Am
Table II. Experimental data in nonsensitive, sensitive and aspirin-treated sensitive women Sensitive before aspirin Nonsensitive (nulliparous)
I
Nulliparous
I
Parous
75.6 ± 14.4 (22) 22.5 ± 5.4 (22) 6.3 ± 2.3 (22)
71.4 ± 12.3 (16) 20.8 ± 3.6 (16) 6.1 ± 2.4 (16)
86.9 ± 14.7 (6) 27.3 ± 6.9 (6) 6.9 ± 2.0 (6)
103.8 ± 9.1 (25)
109.9 ± 12.3 (22)
108.8 ± 10.7 (16)
112.8 ± 16.7 (6)
56.0 ± 8.9 (25)
61.1 ± 9.5 (22)
60.0 ± 9.3 (16)
64.3 ± 10.2 (6)
90.6 ± 10.4 (16)
74.2 ± 9.2 (15)
73.2 ± 9.9 (12)
78.3 ± 4.7 (3)
78.0 ± 9.7 (16)
66.4 ± 7.8 (15)
65.5 ± 7.0 (12)
70.3 ± 11.7 (3)
31.3 ± 2.7 (25)
32.7 ± 2.1 (22)
32.7 ± 2.3 (16)
32.8 ± 1.5 (6)
76.7 ± 15.7 (25) 20.3 ± 3.1 (25) 20.1 ± 6.1 (25)
Weight (kg) Age (yr) Effective pressor dose (ng/kg/min) Systolic blood pressure (mm Hg) Diastolic blood pressure (mm Hg) Baseline heart rate (beats/ min) Heart rate during effective pressor dose (beats/min) Gestational age (wk)
All
Values are means ± SD. Numbers in parentheses are numbers of observations.
Table III. Prostaglandin (pg I ml) concentrations in women sensitive to angiotensin II before and after aspirin treatment (81 mg) Sensitive after aspirin
Sensitive before aspirin
Serum Thromboxane B2 Plasma Thromboxane B2 Plasma 6-Ketoprostaglandin F,a Plasma Prostaglandin E2
I
I
I
I
Parous
All
1804 ± 1771 (14)
718 ± 428 (6)
113 ± 180 (16)
132 ± 206 (12)
57 ± 25 (4)
111 ± 98 (16)
130 ± 107 (12)
56 ± 29 (4)
19 ± 10 (16)
19 ± 12 (12)
21 ± 8 (4)
231 ± 83 (16)
244 ± 90 (12)
196 ± 50 (4)
172 ± 80'(16)
163 ± 90 (12)
199 ± 29 (4)
152 ± 58 (16)
155 ± 67 (12)
144 ± 15 (4)
100 ± 52 (16)
95 ± 40 (12)
114 ± 87 (4)
All
1478 ± 1567 (20)
Nulliparous
Nulliparous
Parous
Values are means ± SD. Numbers in parentheses are numbers of observations.
free ligand were separated with dextran-coated charcoal mixed in the assay buffer. For the prostaglandin E2 assay we used standard (up to 200 pg) and unknown quantities of prostaglandin E2 mixed in the assay buffer. Duplicate samples were mixed with 0.1 ml of antisera ( 1: 4,000 titer) and 0. 1 ml 3 H-prostaglandin E2 (New England Nuclear; 12,000 dpm), incubated 12 to 18 hours, and separated with dextran-coated charcoal. The antibodies were the gift of William B. Campbell, PhD, Department of Pharmacology, Southwestern Medical School, Dallas, Tex. Cross-reactivities of the thromboxane B2 antibody with prostaglandins A2 , E., and F2a and with 6-keto-prostaglandin F,a are <0.001 % and that with prostaglandin D2 is 0.003%. Crossreactivities of the 6-keto-prostaglandin F,a antibody with prostaglandins A2 , D2 , E" E 2 , F,a, and F2 a and with 6,15-diketo-prostaglandin F1a, and thromboxane B 2 are 0.20%, 0.88%, 1.14%, 0.59%, 0.88%, 0.62%, 2.8%, and
0.001 %, respectively. Cross-reactivities of the prostaglandin E2 antibody with prostaglandins A 2 , B 2 , D2 , E,, F 1a, and F2a, and with 15-keto-prostaglandin E 2 and 13,14-dihydro-15-keto-prostaglandin E2 are 0.30%, 0.58%, 0.1%, 14.0%, 0.01%, 0.70%, 0.10%, and 0.15%, respectively. Recoveries of thromboxane B2 , 6-ketoprostaglandin Fla• and prostaglandin E2 were consistent for each compound at 86.4% ± 3.0%, 63.4% ± 1.1 %, and 76.1 % ± 9.5%, respectively. Values were adjusted according to individual internal recoveries within each sample for each hormone. The assay coefficients of variation for plasma thromboxane B2 , plasma 6-keto-prostaglandin F 1a, and prostaglandin E2 in our laboratories were 10.2%, 9.8%, and 9.4%, respectively. To reduce variation, all plasma samples were extracted, chromatographed, and assayed in a single assay. When known amounts of these three prostaglandins were added to human plasma, the re-
Low-dose aspirin and angiotensin II pressor response
Volume 159 Number 5
1039
Sensitive after aspirin All
Nulliparous
Parous
10.2 ± 5.5 (17)
10.0 ± 5.8 (13)
10.9 ± 5.2 (4)
112.5 ± 10.0 (18)
111.7 ± 8.9 (13)
114.4 ± 13.6(5)
62.7 ± 8.6 (18)
61.9 ± 9.7 (13)
65.0 ± 4.6 (5)
83.0 ± 13.7 (13)
79.5 ± 13.l (10)
95.0 ± 9.0 (3)
74.3 ± 11.1(13)
71.6 ± 11.2(10)
83.6 ± 2.8 (3)
covered amounts measured after extraction and silicic acid chromatography are as shown in Table I. Statistical analysis. All data are expressed as the mean ± 1 SD. Comparisons between groups were performed by the Student t test for paired and unpaired data as appropriate. In case of unequal variances, Welch's approximation to the t test was used. Results
We studied two groups that consisted of 25 nulliparous women nonsensitive to angiotensin II and 22 women who were sensitive to angiotensin II (Table II). The 22 sensitive participants comprised 16 nulliparous and 6 parous women. The parous women were heavier than the nulliparous women (86.9 ± 14.7 kg versus 71.4 ± 12.3 kg, p = 0.02), but they did not significantly differ in age, gestational age, systolic or diastolic baseline blood pressures, effective pressor dose of angiotensin II, baseline heart rate, or heart rate during the administration of angiotensin II. Nonsensitive and sensitive nulliparous women differed in effective pressor doses (20. l ± 6.1 versus 6.1±2.4 ng/kg/min, p<0.001) and in baseline heart rates (90.6 ± 10.4 versus 73.2 ± 9.9 beats/min, p < 0.001). Heart rates decreased 12.6 ± 9.1 beats/min (13.6% ± 9.6%) during the effective pressor doses of angiotensin II (a reflection of baroreceptor reflex function) in the nonsensitive group and 7.8 ± 6.9 beats/min (10.2% ± 8.0%) in the sensitive women (p = 0.07). When comparisons involved the combination of sensitive nulliparous and parous patients, similar differences between sensitive and nonsensitive women were observed. Repeat studies could not be obtained in five of the aspirin-treated sensitive women (three nulliparous and two parous) because of delivery or withdrawal from the study. As illustrated in Table II and Fig. 1 and 2, 1
week of aspirin treatment (81 mg/day) increased the effective pressor dose of angiotensin II from 6.1 ± 2 .4 to 10.0 ± 5.8 ng/kg/min in the nulliparous group (p = 0.037). In the 17 women studied before and after aspirin treatment, the effective pressor dose of angiotensin II increased from 5.9 ± 2.4 to 10.2 ± 5.5 ng/kg/min (p < 0.001). Although the effective pressor dose of angiotensin II increased after aspirin treatment, the values observed remained less than those recorded in the nonsensitive participants (p < 0.001). One week of aspirin treatment did not significantly alter systolic blood pressure, diastolic blood pressure, baseline heart rate, or heart rate during the effective pressor dose of angiotensin (Table II). As shown in Table III, serum thromboxane B2 in nulliparous sensitive patients fell from 1804 ± 1771 to 132 ± 206 pg/ml (p < 0.001). These values of platelet-derived thromboxane B2 are a measure of therapy compliance. Only in one woman was the thromboxane B2 inhibition doubtful, i.e., 5867 pg/ml before aspirin and 766 pg/ml after aspirin. Her angiotensin II refractoriness improved very little; the effective pressor dose increased from 3.8 ng/kg/min before aspirin to 5.2 ng/kg/min after treatment. As further illustrated in Table III, plasma thromboxane B 2 in nulliparous sensitive women also decreased after 1 week of aspirin treatment from 130 ± 107 to 19 ± 12 pg/ml (p = 0.004), but plasma 6-keto-prostaglandin F 10 also fell from 244 ± 90 to 163 ± 90 pg/ml (p = 0.039) and prostaglandin £ 2 from 155 ± 67 to 95 ± 40 pg/ml (p = 0.014). As illustrated in Fig. 3, the inhibition of plasma thromboxane B 2 (75% ± 19%) was more pronounced (p < 0.001) than for either 6-keto-prostaglandin F 10 (21 % ± 33%) or prostaglandin £ 2 (29% ± 36%). Thus the 6-keto-prostaglandin F 1)thromboxane B2 ratio (Fig. 4) increased from 3.1 ± 2.0 to 12.3 ± 9.9 (p = 0.002).
1040
November 1988 Am .J Obstet Gynecol
Spitz et al.
i
.2s 1. 0
0
-
PLASMA THROHBOXANE 8 2
~
PL ASHA 6-kETO-PGF 1•
~
~
31
NULLIPAROUS PAROUS ALL
-nr
PLASMA PROSU.GLANDIN E2
~
SERUM THROHBOXANE 8 2
~
~
-25%
-50%
-75%
-100%
Fig. 3. Relative inhibition of serum and plasma thromboxane B2 and plasma 6-keto-prostaglandin F 1• and prostaglandin E2 after I week of 81 mg aspirin treatment. Values are illustrated as means ± 1 SD.
When compared with sens1t1ve nulliparous participants (Table III), sensitive parous women before aspirin therapy had lower concentrations of serum thromboxane B2 (1804 ± 1771 versus 718 ± 428 pg/ml, p = 0.047) and plasma thromboxane B2 (130 ± 107 versus 56 ± 29 pg/ml, p;< 0.05). It should be pointed out that not all subjects responded to low-dose aspirin therapy by becoming nonsensitive to infused angiotensin II. Specifically, 9 of 17 (52.9%) women increased their pressor responses to ;;;,10 ng/kg/min; however, only 6 (35.3% of the 17) participants reached doses within 1 SD (> 14 ng/kg/min) of those observed in the nonsensitive patients (20. l ± 6.1 ng/kg/min). Thus, despite becoming more refractory to infused angiotensin II, 8 of 17 (47.1%) women remained sensitive (<10 ng/kg/min) to angiotensin II (Fig. 2). Comment
There is evidence that various hemodynamic changes characteristic of normal pregnancy may depend on a delicate balance between the production of vasodilator and vasoconstrictor prostaglandins and that prostacyclin and thromboxane A 2 could be involved in this phenomenon. 1. 2 · 11 Whereas normal pregnancy may be characterized by increased prostacyclin production relative to thromboxane A 2 , in pregnancy-induced hypertension, the production of these two eicosanoids ap-
pears to be tilted in favor of thromboxane A2 • Enhanced sensitivity to angiotensin II could be an early, if not the first, sign of such an imbalance. 2· 1 · 11 In recent reports the beneficial effect of low-dose aspirin in the prevention of gestational proteinuric hypertension has been attributed to suppression of thromboxane A 2 synthesis with sparing of prostacyclin8 • 9 ; however, this has not been clearly demonstrated. In this study 1 week of aspirin treatment with 81 mg/ day significantly inhibited thromboxane A, production; thus the 6-keto-prostaglandin F 1.lthromboxane B, ratio increased. However, this inhibition was not selective for thromboxane A,, and plasma concentrations of prostacyclin and prostaglandin E, were reduced significantly, 21 % ± 33% and 29% ± 36%, respectively. Although it is not known whether this represents a reduction in prostacyclin production by vessels, comparable doses of aspirin and/ or their cumulative effects have been reported to inhibit vascular prostacyclin production. 15• 16 Moreover, there is considerable in vitro evidence that prostacyclin is the primary prostanoid produced by arteries. 1•· 17 Although prostaglandin E, also can be of vascular origin, 17 its primary source in pregnancy is most likely the uteroplacental unit. 18 In contrast, thromboxane A, is primarily produced by platelets5 · 15- 16 but also may be produced by the placenta.' Thus our data support the view that platelet, vascular, and uteroplacental prostanoid production are significantly altered by low-dose aspirin. Nevertheless, the effect of aspirin treatment on the effective pressor dose of angiotensin II in this study reflects only a partial selective prostaglandin inhibition. While the effect of low-dose aspirin was to cause the women sensitive to angiotensin II to become more refractory to infused angiotensin II, they did not become as refractory to infused angiotensin II as normal pregnant women. Their pressor dose was 10.2 ± 5.5 ng/kg/min as compared with 20.l ± 6.1 ng/kg/min, respectively. Moreover, 4 7% (8 of 17) of these women did not increase their effective pressor dose of angiotensin II> 10 ng/kg/min. On further visual inspection of the pressor responses to angiotensin II depicted in Fig. 2, there appear to be at least two subgroups of women within the sensitive group that received aspirin; one of these is pointed out above. Six others of these women (35%) exhibited an effective pressor dose > 14 ng/kg/min, a value that is within 1 SD of the group of women nonsensitive to angiotensin II. When we further examined the data from this subgroup of women, they were found to have no significant decrease in their plasma levels of prostacyclin after aspirin treatment and only a 65% decrease in plasma thromboxane B,. In contrast, in the seven women whose effective pressor dose of angiotensin II remained <7 ng/kg/min after low-dose aspirin, there was nearly a 25% inhibition of plasma 6-keto-prostaglandin F 1• and an 80% inhibition
Volume 159 Number 5
of thromboxane B2 • Thus there appears to be a substantial pharmacologic difference between these two subgroups of women relative to their prostanoid production, and there are substantial differences in their subsequent vascular responses to infused angiotensin II. It should be pointed out that the two subgroups of women noted previously are relatively small in number, and they were recognized only after examination of Fig. 2. Nevertheless, these observations provide provocative ideas with regard to the aforementioned differences and the heterogeneity of women with pregnancy-induced hypertension. It is possible that the differences seen reflect a difference in the dose of aspirin when corrected for weight. This is not likely, because the mean body weight of the women with less prostaglandin inhibition and more refractoriness to angiotensin II was less, i.e., 69 ± 13 kg as compared with 87 ± 14 kg. It also is not explained by a difference in age or race. However, the differences observed are consistent with the hypothesis that pregnancy-induced hypertension may be a heterogeneous disease with two or three subgroups that may be separated by their pharmacologic and physiologic responses to aspirin and infused angiotensin II, respectively. For example, in the subgroup with increased refractoriness to angiotensin II after aspirin treatment (i.e., >14 ng/kg/min), we observed that prostacyclin production was not inhibited by aspirin, whereas thromboxane A2 was inhibited. Thus not only did they maintain plasma prostacyclin values relative to plasma thromboxane A 2 (which resulted in an increased prostacyclin/thromboxane A 2 ratio), but their vessels also may have retained the capacity to produce additional prostacyclin when stimulated by infused angiotensin II. This concept is supported by our recent observations in vitro, 11 and in vivo (Magness RR, et al., unpublished observations) that angiotensin II causes increased vascular and uteroplacental production of prostacyclin in ovine pregnancy. In contrast, the subgroup of women whose effective pressor dose of angiotensin II increased little or not at all (i.e., remained <7 ng/kg/min) had a nearly 25% decrease in basal plasma 6-keto-prostaglandin F 10 in addition to an 80% decrease in basal plasma thromboxane B2 • Thus their vessels and platelets appear to be very sensitive to prostaglandin inhibition by aspirin treatment. This also suggests that their vasculature might be very different with regard to prostanoid production and that after aspirin treatment their vessels may have lost the capacity to respond to infused angiotensin II by producing additional prostacyclin to offset or antagonize the vasoconstrictor effects of this octapeptide. Nevertheless, there was a greater decrease in thromboxane A 2 than in prostacyclin and.their ratio was increased; however, there was no change in the effective pressor dose of angiotensin II. Although our participant numbers
Low-dose aspirin and angiotensin II presser response
1041
AFTER ASPIRIN
20
N
a:I
x I-
u:-"'
t..:I 0...
0
I-
UJ
10
:lo:: I
'°0
t:= C(
IX
BEFORE ASPIRIN
0
+
Fig. 4. Ratio of 6-keto-prostaglandin F 10 /thromboxane B, before and after 1 week of 81 mg aspirin treatment (p = 0.002). Values are expressed as means ± 1 SD.
are too small to be conclusive, these observations are exciting and subsequently may help to explain why pregnancy-induced hypertension may recur in some women after the first pregnancy; i.e., there may be a primary vascular defect in prostanoid production among these women that could possibly lead to transient forms of hypertension with each pregnancy. These observations also are of interest with regard to the interpretation of the plasma prostacyclin/thromboxane A 2 ratio. While our overall group of women sensitive to angiotensin II had a substantial and significant increase in their prostacyclin/thromboxane A, ratio after low-dose aspirin, there was greater variability in this ratio after treatment. The likelihood that there were two subgroups of women accounts for this. Although both subgroups had decreases in plasma thromboxane B2 levels, one subgroup appeared to have a much greater fall in plasma thromboxane B2 than the other. Thus the ratio was misleading in the prediction of the participants' responses to therapy as reflected by the angiotensin II pressor response. Therefore the ratio should be used and interpreted with a great deal of caution and careful examination of the actual changes in prostanoids. It has been suggested that the prevention of thromboxane A, domination, without alteration of prostacyclin, might decrease the incidence and severity of
1042 Spitz et al.
pregnancy-induced hypertension. 6 • 7 This study was neither randomized nor blinded; therefore it cannot be viewed as a clinical trial that would address this issue. Nevertheless, it is of interest that of the six women sensitive to angiotensin II who became as refractory after aspirin therapy as the nonsensitive participants, only one was found to have pregnancy-induced hypertension at the end of pregnancy and she did not have proteinuria. On the other hand, of the subgroup that remained relatively sensitive to infused angiotensin II (<7 ng/kg/min), all seven had pregnancy-induced hypertension and six of them had proteinuria. These observations serve as further support that there were at least two populations of women who by chance randomly entered our study. They can be separated pharmacologically, physiologically, and clinically. Whether low-dose aspirin will aid in further delineation of these two groups or whether even lower doses of aspirin will yield similar results remains to be ascertained. It also is of interest to determine whether the sensitive subgroup of gravidas were the patients who failed to respond to aspirin treatment in other studies, regardless of dose. 8 · 9 Another observation made in these studies relates to the differences in heart rate observed in the women either not sensitive or sensitive to angiotensin II. Seligman 19 reported that during normotensive pregnancy, baroreceptor function is set at a high level of sensitivity, i.e., it is more responsive to afterload changes than it is in women with pregnancy-induced hypertension. In this study we observed that decreases in heart rate during angiotensin II administration were 12.6 ± 9.1 beats/min in women not sensitive to angiotensin II as compared with 7.6 ± 6.9 beats/min in the group sensitive to angiotensin II (p = 0.07). This observation is consistent with that of Seligman. 19 The increase in pressor dose of angiotensin II after aspirin treatment could not be explained by a change in baroreceptor function because the decrease in heart rate in this group was unchanged by aspirin therapy (Table II). Our studies in sheep, animals that have been shown to be strikingly similar to women in their response to angiotensin II, support the hypothesis that in normal pregnancy the reduced pressor response to angiotensin II is not due in any large part to altered cardiac responses but reflects a real reduction in vascular responsiveness. 20 The same may be said of women. A surprising finding in our study was that the basal heart rate values at equal blood pressures were lower in women sensitive to angiotensin II than in women not sensitive to angiotensin II. If the stroke volume is not increased (and little is known about this at this stage of the disease), it appears that women sensitive to angiotensin II may face increased peripheral vascular re-
November 1988 Am J Obstet Gynecol
sistance and maintain a normal blood pressure at least in part by decreasing their cardiac output. What role, if any, prostaglandins have in this heart rate alteration, and presumably vascular resistance, remains to be determined. We thank Laurie Daniels for her technical assistance in the preparation of this manuscript and the house officers in obstetrics and gynecology at Parkland Memorial Hospital for their assistance in the clinical management of the women described in this study. REFERENCES 1. Fitzgerald DJ, Entman SS, Mulloy K, Fitzgerald GA. Decreased prostacyclin biosynthesis preceding the clinical manifestation of pregnancy-induced hypertension. Circulation l 987;75:956. 2. Goodman RP, Killam AP, Brash AR, Branch RA. Prostacyclin production during pregnancy: comparison of production during normal pregnancy and pregnancy complicated by hypertension. AM J OBSTET GYNECOL 1982;142:817. 3. Walsh SW. Preeclampsia: an imbalance in placental prostacyclin and thromboxane production. AM J 0BSTET GYNECOL 1985; 152:335. 4. Gant NF, Daley GL, Chand S, Whalley PJ, MacDonald PC. A study of angiotensin II pressor response throughout primigravid pregnancy. J Clin Invest 1973;52:2682. 5. Fitzgerald DJ, Mayo G, Catella F, Entman SS, Fitzgerald GA. Increased thromboxane biosynthesis in normal pregnancy is mainly derived from platelets. AM J OBSTET GYNECOL 1987; 157:325. 6. Collins E, Turner G. Maternal effects of regular salicylate ingestion in pregnancy. Lancet 1975;2:335. 7. Crandon AJ, Isherwood DM. Effect of aspirin on incidence of pre-eclampsia. Lancet 1979; 1: 1356. 8. Beaufils M, Uzan S, Donsimoni R, Colau JC. Prevention of pre-eclampsia by early antiplatelet therapy. Lancet 1985; 1:840. 9. Walleriburg HCS, Dekker GA, Makovitz JW, Rotmans P. Low-dose aspirin prevents pregnancy-induced hypertension and preeclampsia in angiotensin-sensitive primigravidae. Lancet 1986; 1: 1. 10. Gant NF, Chand S, Worley RJ, Whalley PJ, Crosby UD, MacDonald PC. A clinical test useful for predicting the development of acute hypertension in pregnancy. AM J 0BSTET GYNECOL 1974;120:1. 11. Everett RB, Worley RJ, MacDonald PC, Gant NF. Effect of prostaglandin synthetase inhibitors on pressor response to angiotensin II in human pregnancy. J Clin Endocrinol Metab 1978;46:1007. 12. Everett RB, Worley RJ, MacDonald PC, Gant NF. Oral administration of theophylline to modify pressor responsiveness to angiotensin II in women with pregnancyinduced hypertension. AM J OBSTET GYNECOL 1978; 132:359. 13. Campbell WB, Ojeda SR. Measurements of prostaglandins by radioimmunoassay. Methods Enzymol. 1987; 141:323. 14. 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 Invest 1985;76:2206. 15. Preston FE, Whipps S,Jackson CA, French AJ, Wyld PJ, Stoddard CJ. Inhibition of prostacyclin and platelet thromboxane A 2 after low-dose aspirin. N Engl J Med 1981;304:76. 16. Hanley SP, Bevan J, Cockbill SR, Heptinstall S. Differ-
Volume 159 Number 5
ential inhibition by low-dose aspirin of human venous prostacyclin synthesis and platelet thromboxane synthesis. Lancet 1981; 1:969. 17. Satoh H, Hosono M, Satoh S. Distinctive effect of angiotensin II on prostaglandin production in dog renal and femoral arteries. Prostaglandins 1984;27:807. 18. Mitchell MD, Flint APF. Prostaglandin production by intra-uterine tissues from periparturient sheep: use of a superfusion technique. J Endocrinol 1978;76: 111.
Low-dose aspirin and angiotensin II presser response
19. Seligman SA. Baroreceptor reflex function in preeclampsia. J Obstet Gynaecol Br Commonw 1971:78:413. 20. Naden RP, Gant NF, Rosenfeld CR. The pressor response to angiotensin II: The roles of peripheral and cardiac responses in pregnant and non pregnant sheep. AM J OBSTET GYNECOL 1984;148:450.
Preeclampsia: A microvesicular fat disease of the liver? Hisanori Minakami, MD," Naoko Oka, MD," Tadashi Sato, MD," Taro Tamada, MD," Yoshikazu Yasuda, MD,b and Norio Hirota, MD< Tochigi-ken, Japan To study the interrelationship between preeclampsia, the syndrome of hemolysis, elevated liver enzymes, and low platelet count, and acute fatty liver of pregnancy, 41 liver specimens from 41 preeclamptic women with and without liver dysfunction were examined for the amount of fat deposited in hepatocytes. All 41 specimens stained with oil red 0 on frozen sections showed a significant amount of microvesicular fat droplets in varying degrees. In contrast, only 11 of the 41 stained conventionally (with hematoxylin-eosin) showed significant fatty infiltration. The density of hepatocellular fat correlated positvely with plasma urate concentration and negatively with the platelet count. These findings suggest that preeclampsia may be one of several microvesicular fatty diseases of the liver and that acute fatty liver of pregnancy may be the most severe form. (AM J OBSTET GYNECOL 1988;159:1043-7.)
Key words: Preeclampsia, fatty liver, hyperuricemia, microvesicular fat disease Acute fatty liver of pregnancy is an uncommon but potentially fatal disorder that may complicate the third trimester.' Acute fatty liver of pregnancy is considered an independent clinical entity whose definitive diagnosis requires histologic examination of the liver.2 A microscopic feature of acute fatty liver of pregnancy is microvesicular fat vacuolation of the hepatocytes.2 Recently, the incidence of acute fatty liver of pregnancy has apparently increased, mainly because of a greater awareness and early recognition of this disorder, particularly of milder cases. 3 The syndrome of hemolysis, elevated liver enzymes, and low platelet count in the third trimester is thought to be a form of preeclampsia, even though it sometimes lacks the traditional signs of this disorder.4· 5 Furthermore, it is reported that traditional preeclampsia complicates liver dysfunction in a considerable number of cases. 6 Since the three clinical conditions-acute fatty liver of preg-
From the Departments of Obstetrics and Gynecology," Surgery,' and Pathology,' ]ichi Medical School. Received for publication December 14, 1987; revised May 13, 1988; accepted May 19, 1988. Reprint requests: Hisanori Minakami, MD, Jichi Medical School, Minamikawachi-machi, I'ochigi,-ken, 329-04 japan.
nancy, the syndrome of hemolysis, elevated liver enzymes, and low platelet count, and preeclampsia with liver dysfunction-share many clinical and laboratory abnormalities, it is possible that they form a disease spectrum. However, there has been no evidence to confirm this hypothesis. To resolve the issue, we performed liver biopsies or autopsies on preeclamptic patients with and without liver dysfunction, in patients who fulfilled the criteria of hemolysis, elevated liver enzymes, and low platelet count syndrome described by Weinstein,' and in patients suspected of having acute fatty liver of pregnancy.
Patients and methods We examined liver specimens from three patients who died of preeclampsia complicated by disseminated intravascular coagulation and hepatic failure in the third trimester and from 38 preeclamptic patients from whom specimens were obtained within 14 days of delivery. Preeclampsia was defined as a systolic or diastolic blood pressure ;;;, 140 or 90 mm Hg on two occasions, recorded 24 hours apart, with the presence of proteinuria (>500 mg/day) in a patient who had been normotensive during the first 20 weeks of pregnancy. Patients ranged in age from 21 to 43 years. All had re-
1043
concentrations in pregnant women sensitive to angiotensin II Bernard Spitz, MD, Ronald R. Magness, PhD, Susan M. Cox, MD, Charles E. L. Brown, MD, Charles R. Rosenfeld, MD, and Norman F. Gant, MD Dallas, Texas Decreased incidence of proteinuric hypertension after low-dose aspirin therapy is hypothesized to be a consequence of selective thromboxane A2 inhibition and sparing of prostacyclin. This study was designed to ascertain if low-dose aspirin therapy (81 mg/day for 1 week) alters vascular refractoriness to angiotensin II and the prostacyclin/thromboxane A2 ratio in pregnant women sensitive to angiotensin II (n = 17). Low-dose aspirin increased the effective pressor dose of angiotensin II from 5.9 ± 2.4 to 10.2 ± 5.5 ng/kg/min (p < 0.01, mean ± SD). Platelet-derived serum thromboxane 8 2 (a metabolite of thromboxane A2 ), a measure of therapy compliance, decreased from 1804 ± 1771 to 132 ± 206 pg/ml (p < 0.01). Plasma thromboxane 8 2 decreased from 130 ± 107 to 19 ± 12 pg/ml (p < 0.01). Inhibition was not selective because 6-keto-prostaglandin F'" (a metabolite of prostacyclin) also decreased from 243 ± 90 to 163 ± 90 pg/ml (p = 0.039) and prostaglandin E2 was reduced from 155 ± 67 to 95 ± 40 pg/ml (p = 0.014). Decreases in thromboxane 8 2 , however, were significantly greater (75% ± 19%) than decreases in 6-keto-prostaglandin F, (21% ± 33%) or prostaglandin E2 (29% ± 36%); thus the 6-keto-prostaglandin F'"/thromboxane 8 2 ratio increased from 3.1 ± 2.0 to 12.4 ± 9.9 (p < 0.01 ). Although low-dose aspirin increases the effective pressor dose of angiotensin II, it does not return to normal pregnancy values. This observation is consistent with the hypothesis that this represents only a partial selective prostaglandin inhibition. (AM J OssTET GYNECOL 1988;159:1035-43.) 0
Key words: Low-dose aspirin, preeclampsia, angiotensin II, pressor responses, prostaglandins, blood pressure control, baroreceptor
Gestational proteinuric hypertension, or pregnancyinduced hypertension, may be associated with an imbalance between prostacyclin and thromboxane A 2 • These eicosanoids have potent and opposing effects on vascular reactivity and platelet aggregation. 13 It has been suggested that this imbalance is a contributing factor in the loss of refractoriness to the pressor effects of angiotensin II and platelet aggregation, which characterize this syndrome.4 5 Because thromboxane A 2 is From the Departments of Obstetrics and Gynecology and Pediatrics, The University of Texas Southwestern Medical Center at Dallas. Supported by National Institutes of Health grants HL34150, HD08783, and HD24971, and by American Heart Association Texas Affiliate 86R-071. Received October 13, 1987; revised April 28, 1988; accepted May 22, 1988. Reprint requests: Norman F. Gant, MD, Professor of Obstetrics and Gynecology, The University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd., Dallas, TX 75235.
released by circulating platelets,' it is logical to consider the use of antiplatelet agents to prevent or treat pregnancy-induced hypertension. Two groups of investigators"· 7 have shown a decreased incidence of pregnancy-induced hypertension in women given cyclooxygenase and/or plateletinhibiting drugs. In a controlled trial by Beaufils et al.," aspirin (150 mg/day) combined with dipyridamole (300 mg/day) was tested in women at the end of the first trimester of pregnancy. They were selected to participate because their previous obstetric histories indicated they were likely to develop hypertension. This antiplatelet regimen conferred a clear benefit in terms of the occurrence of gestational protein uric hypertension, intrauterine growth retardation, and perinatal survival. In a subsequent report by Wallenburg et al.," only primigravid women were selected. In these randomized placebo-controlled and double-blind studies, treatment 1035
t\ovember l 988
1036 Spitz et al. Am
J Obstet Gynecol
Table I. Accuracy and precision of plasma prostaglandin assays Concentration (pg I ml)
I
25
Prostaglandin added
5
Thromboxane B2 6-keto-prostaglandin F1. Prostaglandin E2
6.6 ± 4.2 2.2 ± 1.5
24.2 ± 6.1 27.4 ± 2.9
6.4 ± 8.3
23.0 ± 9.8
Values are means ± SD (n
=
I
100
Correlation coefficient
Buffer blank (pg/ml)
50.5 ± 17.4 40.2 ± 18.6
104.8 ± 2.9 114.l ± 60.3
0.999 0.987
7.6 1.6
49.6 ± 10.5
113.0 ±
7.3
0.997
5.3
50
I
6).
with aspirin (60 mg/day) was associated with a significantly lower incidence of pregnancy-induced hypertension. In none of these studies were the concentrations of circulating prostanoids reported. Thus the hypothesis that restoration of angiotensin II refractoriness and improved clinical outcome were consequences of selective inhibition of thromboxane A 2 and sparing of prostacyclin remains unproved. The aim of this study was to determine the effects of low-dose aspirin on changes in vascular reactivity to angiotensin II and whether such changes could be explained by alterations in circulating prostacyclin and thromboxane A2 • Material and methods
In this study we determined the effect of 81 mg of aspirin per day for 1 week beginning at approximately 32 weeks' gestation on (1) the minimum amount of angiotensin II per kilogram per minute required to increase diastolic blood pressure 20 mm Hg (the effective pressor dose) in 22 women sensitive to angiotensin II (effective pressor dose< 10 ng/kg/min), (2) the concentration of thromboxane A 2 in serum, and (3) the ratio of plasma prostacyclin/thromboxane A 2 • We obtained angiotensin II pressor responses in 25 nonsensitive women who served as controls, to provide current data in regard to the effective pressor dose of angiotensin II in nonsensitive women and to permit us to determine whether aspirin returned the angiotensin II pressor response to normal pregnancy values. Sensitivity to angiotensin II is defined as a pressor dose< 10 ng/kg/min, and refractoriness is defined as a pressor dose:;,,, 10 ng/kg/min.4 Nonsensitive women were not treated with aspirin because they were refractory to angiotensin II. The selection of women sensitive to angiotensin II was based on the observations of Gant et al.4 that 50% to 90% of such patients will ultimately have pregnancy-induced hypertension. This study was approved by the Institutional Review Board of The University of Texas Southwestern Medical Center at Dallas, and all study subjects gave informed written consents. To locate women sensitive to angiotensin II, angiotensin II sensitivity tests were performed in a high-risk population recruited from two sources. First, women
admitted to Parkland Memorial Hospital (Dallas, Texas) and the High-Risk Antepartum Care Unit with early onset of pregnancy-induced hypertension (who subsequently became normotensive) were studied. Second, women with positive "roll-over tests" were selected from the antepartum clinics. This test has been used to identify women in whom hypertension is likely to develop later in pregnancy, 10 and it was useful in helping to discover women sensitive to angiotensin II who were candidates for this study. Women were excluded if they were known to have taken drugs such as aspirin 11 or theophylline 12 that could influence either test. Patients with a history of chronic hypertension also were excluded from this study. Diets were unrestricted. The angiotensin II sensitivity test was performed as previously described.4 Women were placed in the left lateral recumbent position in a quiet room. Blood pressure and heart rate were automatically measured from the right arm every 4 minutes by means of an oscillometer (Spacelabs 512D; E. R. Squibb & Sons, Princeton, N.J.). After stabilization of the diastolic blood pressure for at least 15 minutes, an intravenous infusion was started in the left arm. This was connected to an infusion pump with a syringe containing angiotensin II amide (1000 ng/ml in 0.9% sodium chloride, CibaGiegy Pharmaceuticals). Infusion of angiotensin II was started at a rate of approximately 1.5 ng/kg/min and was gradually increased at 4-minute intervals. The test was stopped after the effective pressor dose response was observed. This procedure was repeated after 5 minutes; results in all studies were identical to the first pressor dose response. Data from the two studies were averaged and used as a single observation for statistical analysis. Before the angiotensin II infusion was started, blood was collected for analysis of serum and plasma prostaglandin concentrations. Specific methods for the measurement of platelet-derived serum thromboxane B, (the stable metabolite of thromboxane A 2 ), plasma thromboxane B 2 , 6-keto-prostaglandin F 1• (the stable metabolite of prostacyclin), and prostaglandin E 2 are described below. Prostaglandin assays. Serum concentrations of thromboxane B, reflective of platelet-derived prostanoid were obtained from samples after clotting (45 min-
Low-dose aspirin and angiotensin II presser response
Volume 159 Number 5
20
c ·e
i
20
NULLIPAROUS PAROUS
18
ALL
16
15
c
-
:!
Cl
~
UJ
5
Cl
E
....
0
c
0
0:: 0
UJ
....0:: 0.. ....>
VI VI
0::
a...
UJ
>
i==
10
8
j::
u
u.. u..
12
VI
0
10
Ill Ill
UJ
14
Cl
Ill
0::
- - - NULLIPAROUS - - - - - PAROUS
0. ..le
Cl
1037
....'-'
5
LL.
LL. ....
UJ
6
A 4
"' ';: 0
z
Fig. I. Effective pressor doses in nonsensitive (A) and sensitive women before (B) and after (C) aspirin treatment. (A > B, p < 0.001; B < C, p = 0.037; C
utes at room temperature) and were produced in quantities of 1 to 10 ng/ml; thus a direct assay was performed on the serum. To measure thromboxane B2 , we used standard (up to 500 pg) and unknown amounts of thromboxane B2 mixed in 0.1 moll L phosphatebuffered saline solution (pH = 7.4) with 0.1 % polyvinylpyrrolidone. Duplicate samples of serum (0.025 to 0.lml) were assayed by the addition of 0.1 ml antithromboxane B2 antiserum ( 1 : 16,000 titer) and 0.1 ml 3 H-thromboxane B2 .(New England Nuclear; 12,000 dpm) and were incubated for 12 to 18 hours at 4° C. Antibody-bound thromboxane B2 was separated with dextran-coated charcoal. To measure plasma prostaglandins, the following methods were used. Blood samples (6.0 ml) were drawn into chilled syringes, transferred immediately to chilled tubes containing ethylenediaminetetraacetic acid (4.2 mg) and aspirin (0.3 mg), centrifuged, and stored at - 20° C. Measurement of plasma prostaglandins was by the methods of Campbell and Ojeda. 13 To determine recovery, the appropriate tritiated prostaglandin (3000 dpm) was added to 2 ml acidified plasma (pH = 3.0; glacial acetic acid) and extracted on individual Bond Elut C-18 cartridges (Analytichem International, Harbor City, Calif.). Bond Elut C-18 cartridges, which had
2 BEFORE ASPIRIN
AFTER ASPIRIN
0 Fig. 2. Effective pressor dose in women sensitive to angiotensin II before and after aspirin treatment. (Paired ttest: p < 0.001.)
been washed first with IO ml 95% ethanol and then with IO ml distilled water, were slowly loaded with acidified samples. They were then washed with 5 ml distilled water and 5 ml petroleum ether and were eluted with 5 ml ethyl acetate. The extract was dried and purified by silicic acid chromatography that used solvents of increasing polarity (e.g., toluene : ethyl acetate : methanol). The first fraction, which contained free fatty acids and prostaglandins A2 and B2 , was discarded. The second fraction contained thromboxane B2 , 6-keto-prostaglandin F 1a, and prostaglandin E 2 ; the eluate was dried under vacuum (37° C), reconstituted in phosphate-buffered saline solution, and assayed for thromboxane B2 , 6-keto-prostaglandin F 1a, and prostaglandin E 2 by radioimmunoassay. The assay for thromboxane B2 was the same as that described above for the serum. The 6-keto-prostaglandin F 1a assay procedure used standard (up to 200 pg) and unknown quantities of 6-keto-prostaglandin F 1a (0.1 ml) mixed in 0.1 mol/L phosphate-buffered saline solution (pH = 7.4) with 0.1 % polyvinylpyrrolidone prepared in duplicate aliquots. Antiserum (0.1 ml, 1: 4,000 titer) and 0.1 ml 3 H-6-keto-prostaglandin F 1a (New England Nuclear; 12,000 dpm) were added and the tubes were incubated at 4° C for 12 to 18 hours. Bound ligand and
November 1988
1038 Spitz et al.
J Obstet Gynecol
Am
Table II. Experimental data in nonsensitive, sensitive and aspirin-treated sensitive women Sensitive before aspirin Nonsensitive (nulliparous)
I
Nulliparous
I
Parous
75.6 ± 14.4 (22) 22.5 ± 5.4 (22) 6.3 ± 2.3 (22)
71.4 ± 12.3 (16) 20.8 ± 3.6 (16) 6.1 ± 2.4 (16)
86.9 ± 14.7 (6) 27.3 ± 6.9 (6) 6.9 ± 2.0 (6)
103.8 ± 9.1 (25)
109.9 ± 12.3 (22)
108.8 ± 10.7 (16)
112.8 ± 16.7 (6)
56.0 ± 8.9 (25)
61.1 ± 9.5 (22)
60.0 ± 9.3 (16)
64.3 ± 10.2 (6)
90.6 ± 10.4 (16)
74.2 ± 9.2 (15)
73.2 ± 9.9 (12)
78.3 ± 4.7 (3)
78.0 ± 9.7 (16)
66.4 ± 7.8 (15)
65.5 ± 7.0 (12)
70.3 ± 11.7 (3)
31.3 ± 2.7 (25)
32.7 ± 2.1 (22)
32.7 ± 2.3 (16)
32.8 ± 1.5 (6)
76.7 ± 15.7 (25) 20.3 ± 3.1 (25) 20.1 ± 6.1 (25)
Weight (kg) Age (yr) Effective pressor dose (ng/kg/min) Systolic blood pressure (mm Hg) Diastolic blood pressure (mm Hg) Baseline heart rate (beats/ min) Heart rate during effective pressor dose (beats/min) Gestational age (wk)
All
Values are means ± SD. Numbers in parentheses are numbers of observations.
Table III. Prostaglandin (pg I ml) concentrations in women sensitive to angiotensin II before and after aspirin treatment (81 mg) Sensitive after aspirin
Sensitive before aspirin
Serum Thromboxane B2 Plasma Thromboxane B2 Plasma 6-Ketoprostaglandin F,a Plasma Prostaglandin E2
I
I
I
I
Parous
All
1804 ± 1771 (14)
718 ± 428 (6)
113 ± 180 (16)
132 ± 206 (12)
57 ± 25 (4)
111 ± 98 (16)
130 ± 107 (12)
56 ± 29 (4)
19 ± 10 (16)
19 ± 12 (12)
21 ± 8 (4)
231 ± 83 (16)
244 ± 90 (12)
196 ± 50 (4)
172 ± 80'(16)
163 ± 90 (12)
199 ± 29 (4)
152 ± 58 (16)
155 ± 67 (12)
144 ± 15 (4)
100 ± 52 (16)
95 ± 40 (12)
114 ± 87 (4)
All
1478 ± 1567 (20)
Nulliparous
Nulliparous
Parous
Values are means ± SD. Numbers in parentheses are numbers of observations.
free ligand were separated with dextran-coated charcoal mixed in the assay buffer. For the prostaglandin E2 assay we used standard (up to 200 pg) and unknown quantities of prostaglandin E2 mixed in the assay buffer. Duplicate samples were mixed with 0.1 ml of antisera ( 1: 4,000 titer) and 0. 1 ml 3 H-prostaglandin E2 (New England Nuclear; 12,000 dpm), incubated 12 to 18 hours, and separated with dextran-coated charcoal. The antibodies were the gift of William B. Campbell, PhD, Department of Pharmacology, Southwestern Medical School, Dallas, Tex. Cross-reactivities of the thromboxane B2 antibody with prostaglandins A2 , E., and F2a and with 6-keto-prostaglandin F,a are <0.001 % and that with prostaglandin D2 is 0.003%. Crossreactivities of the 6-keto-prostaglandin F,a antibody with prostaglandins A2 , D2 , E" E 2 , F,a, and F2 a and with 6,15-diketo-prostaglandin F1a, and thromboxane B 2 are 0.20%, 0.88%, 1.14%, 0.59%, 0.88%, 0.62%, 2.8%, and
0.001 %, respectively. Cross-reactivities of the prostaglandin E2 antibody with prostaglandins A 2 , B 2 , D2 , E,, F 1a, and F2a, and with 15-keto-prostaglandin E 2 and 13,14-dihydro-15-keto-prostaglandin E2 are 0.30%, 0.58%, 0.1%, 14.0%, 0.01%, 0.70%, 0.10%, and 0.15%, respectively. Recoveries of thromboxane B2 , 6-ketoprostaglandin Fla• and prostaglandin E2 were consistent for each compound at 86.4% ± 3.0%, 63.4% ± 1.1 %, and 76.1 % ± 9.5%, respectively. Values were adjusted according to individual internal recoveries within each sample for each hormone. The assay coefficients of variation for plasma thromboxane B2 , plasma 6-keto-prostaglandin F 1a, and prostaglandin E2 in our laboratories were 10.2%, 9.8%, and 9.4%, respectively. To reduce variation, all plasma samples were extracted, chromatographed, and assayed in a single assay. When known amounts of these three prostaglandins were added to human plasma, the re-
Low-dose aspirin and angiotensin II pressor response
Volume 159 Number 5
1039
Sensitive after aspirin All
Nulliparous
Parous
10.2 ± 5.5 (17)
10.0 ± 5.8 (13)
10.9 ± 5.2 (4)
112.5 ± 10.0 (18)
111.7 ± 8.9 (13)
114.4 ± 13.6(5)
62.7 ± 8.6 (18)
61.9 ± 9.7 (13)
65.0 ± 4.6 (5)
83.0 ± 13.7 (13)
79.5 ± 13.l (10)
95.0 ± 9.0 (3)
74.3 ± 11.1(13)
71.6 ± 11.2(10)
83.6 ± 2.8 (3)
covered amounts measured after extraction and silicic acid chromatography are as shown in Table I. Statistical analysis. All data are expressed as the mean ± 1 SD. Comparisons between groups were performed by the Student t test for paired and unpaired data as appropriate. In case of unequal variances, Welch's approximation to the t test was used. Results
We studied two groups that consisted of 25 nulliparous women nonsensitive to angiotensin II and 22 women who were sensitive to angiotensin II (Table II). The 22 sensitive participants comprised 16 nulliparous and 6 parous women. The parous women were heavier than the nulliparous women (86.9 ± 14.7 kg versus 71.4 ± 12.3 kg, p = 0.02), but they did not significantly differ in age, gestational age, systolic or diastolic baseline blood pressures, effective pressor dose of angiotensin II, baseline heart rate, or heart rate during the administration of angiotensin II. Nonsensitive and sensitive nulliparous women differed in effective pressor doses (20. l ± 6.1 versus 6.1±2.4 ng/kg/min, p<0.001) and in baseline heart rates (90.6 ± 10.4 versus 73.2 ± 9.9 beats/min, p < 0.001). Heart rates decreased 12.6 ± 9.1 beats/min (13.6% ± 9.6%) during the effective pressor doses of angiotensin II (a reflection of baroreceptor reflex function) in the nonsensitive group and 7.8 ± 6.9 beats/min (10.2% ± 8.0%) in the sensitive women (p = 0.07). When comparisons involved the combination of sensitive nulliparous and parous patients, similar differences between sensitive and nonsensitive women were observed. Repeat studies could not be obtained in five of the aspirin-treated sensitive women (three nulliparous and two parous) because of delivery or withdrawal from the study. As illustrated in Table II and Fig. 1 and 2, 1
week of aspirin treatment (81 mg/day) increased the effective pressor dose of angiotensin II from 6.1 ± 2 .4 to 10.0 ± 5.8 ng/kg/min in the nulliparous group (p = 0.037). In the 17 women studied before and after aspirin treatment, the effective pressor dose of angiotensin II increased from 5.9 ± 2.4 to 10.2 ± 5.5 ng/kg/min (p < 0.001). Although the effective pressor dose of angiotensin II increased after aspirin treatment, the values observed remained less than those recorded in the nonsensitive participants (p < 0.001). One week of aspirin treatment did not significantly alter systolic blood pressure, diastolic blood pressure, baseline heart rate, or heart rate during the effective pressor dose of angiotensin (Table II). As shown in Table III, serum thromboxane B2 in nulliparous sensitive patients fell from 1804 ± 1771 to 132 ± 206 pg/ml (p < 0.001). These values of platelet-derived thromboxane B2 are a measure of therapy compliance. Only in one woman was the thromboxane B2 inhibition doubtful, i.e., 5867 pg/ml before aspirin and 766 pg/ml after aspirin. Her angiotensin II refractoriness improved very little; the effective pressor dose increased from 3.8 ng/kg/min before aspirin to 5.2 ng/kg/min after treatment. As further illustrated in Table III, plasma thromboxane B 2 in nulliparous sensitive women also decreased after 1 week of aspirin treatment from 130 ± 107 to 19 ± 12 pg/ml (p = 0.004), but plasma 6-keto-prostaglandin F 10 also fell from 244 ± 90 to 163 ± 90 pg/ml (p = 0.039) and prostaglandin £ 2 from 155 ± 67 to 95 ± 40 pg/ml (p = 0.014). As illustrated in Fig. 3, the inhibition of plasma thromboxane B 2 (75% ± 19%) was more pronounced (p < 0.001) than for either 6-keto-prostaglandin F 10 (21 % ± 33%) or prostaglandin £ 2 (29% ± 36%). Thus the 6-keto-prostaglandin F 1)thromboxane B2 ratio (Fig. 4) increased from 3.1 ± 2.0 to 12.3 ± 9.9 (p = 0.002).
1040
November 1988 Am .J Obstet Gynecol
Spitz et al.
i
.2s 1. 0
0
-
PLASMA THROHBOXANE 8 2
~
PL ASHA 6-kETO-PGF 1•
~
~
31
NULLIPAROUS PAROUS ALL
-nr
PLASMA PROSU.GLANDIN E2
~
SERUM THROHBOXANE 8 2
~
~
-25%
-50%
-75%
-100%
Fig. 3. Relative inhibition of serum and plasma thromboxane B2 and plasma 6-keto-prostaglandin F 1• and prostaglandin E2 after I week of 81 mg aspirin treatment. Values are illustrated as means ± 1 SD.
When compared with sens1t1ve nulliparous participants (Table III), sensitive parous women before aspirin therapy had lower concentrations of serum thromboxane B2 (1804 ± 1771 versus 718 ± 428 pg/ml, p = 0.047) and plasma thromboxane B2 (130 ± 107 versus 56 ± 29 pg/ml, p;< 0.05). It should be pointed out that not all subjects responded to low-dose aspirin therapy by becoming nonsensitive to infused angiotensin II. Specifically, 9 of 17 (52.9%) women increased their pressor responses to ;;;,10 ng/kg/min; however, only 6 (35.3% of the 17) participants reached doses within 1 SD (> 14 ng/kg/min) of those observed in the nonsensitive patients (20. l ± 6.1 ng/kg/min). Thus, despite becoming more refractory to infused angiotensin II, 8 of 17 (47.1%) women remained sensitive (<10 ng/kg/min) to angiotensin II (Fig. 2). Comment
There is evidence that various hemodynamic changes characteristic of normal pregnancy may depend on a delicate balance between the production of vasodilator and vasoconstrictor prostaglandins and that prostacyclin and thromboxane A 2 could be involved in this phenomenon. 1. 2 · 11 Whereas normal pregnancy may be characterized by increased prostacyclin production relative to thromboxane A 2 , in pregnancy-induced hypertension, the production of these two eicosanoids ap-
pears to be tilted in favor of thromboxane A2 • Enhanced sensitivity to angiotensin II could be an early, if not the first, sign of such an imbalance. 2· 1 · 11 In recent reports the beneficial effect of low-dose aspirin in the prevention of gestational proteinuric hypertension has been attributed to suppression of thromboxane A 2 synthesis with sparing of prostacyclin8 • 9 ; however, this has not been clearly demonstrated. In this study 1 week of aspirin treatment with 81 mg/ day significantly inhibited thromboxane A, production; thus the 6-keto-prostaglandin F 1.lthromboxane B, ratio increased. However, this inhibition was not selective for thromboxane A,, and plasma concentrations of prostacyclin and prostaglandin E, were reduced significantly, 21 % ± 33% and 29% ± 36%, respectively. Although it is not known whether this represents a reduction in prostacyclin production by vessels, comparable doses of aspirin and/ or their cumulative effects have been reported to inhibit vascular prostacyclin production. 15• 16 Moreover, there is considerable in vitro evidence that prostacyclin is the primary prostanoid produced by arteries. 1•· 17 Although prostaglandin E, also can be of vascular origin, 17 its primary source in pregnancy is most likely the uteroplacental unit. 18 In contrast, thromboxane A, is primarily produced by platelets5 · 15- 16 but also may be produced by the placenta.' Thus our data support the view that platelet, vascular, and uteroplacental prostanoid production are significantly altered by low-dose aspirin. Nevertheless, the effect of aspirin treatment on the effective pressor dose of angiotensin II in this study reflects only a partial selective prostaglandin inhibition. While the effect of low-dose aspirin was to cause the women sensitive to angiotensin II to become more refractory to infused angiotensin II, they did not become as refractory to infused angiotensin II as normal pregnant women. Their pressor dose was 10.2 ± 5.5 ng/kg/min as compared with 20.l ± 6.1 ng/kg/min, respectively. Moreover, 4 7% (8 of 17) of these women did not increase their effective pressor dose of angiotensin II> 10 ng/kg/min. On further visual inspection of the pressor responses to angiotensin II depicted in Fig. 2, there appear to be at least two subgroups of women within the sensitive group that received aspirin; one of these is pointed out above. Six others of these women (35%) exhibited an effective pressor dose > 14 ng/kg/min, a value that is within 1 SD of the group of women nonsensitive to angiotensin II. When we further examined the data from this subgroup of women, they were found to have no significant decrease in their plasma levels of prostacyclin after aspirin treatment and only a 65% decrease in plasma thromboxane B,. In contrast, in the seven women whose effective pressor dose of angiotensin II remained <7 ng/kg/min after low-dose aspirin, there was nearly a 25% inhibition of plasma 6-keto-prostaglandin F 1• and an 80% inhibition
Volume 159 Number 5
of thromboxane B2 • Thus there appears to be a substantial pharmacologic difference between these two subgroups of women relative to their prostanoid production, and there are substantial differences in their subsequent vascular responses to infused angiotensin II. It should be pointed out that the two subgroups of women noted previously are relatively small in number, and they were recognized only after examination of Fig. 2. Nevertheless, these observations provide provocative ideas with regard to the aforementioned differences and the heterogeneity of women with pregnancy-induced hypertension. It is possible that the differences seen reflect a difference in the dose of aspirin when corrected for weight. This is not likely, because the mean body weight of the women with less prostaglandin inhibition and more refractoriness to angiotensin II was less, i.e., 69 ± 13 kg as compared with 87 ± 14 kg. It also is not explained by a difference in age or race. However, the differences observed are consistent with the hypothesis that pregnancy-induced hypertension may be a heterogeneous disease with two or three subgroups that may be separated by their pharmacologic and physiologic responses to aspirin and infused angiotensin II, respectively. For example, in the subgroup with increased refractoriness to angiotensin II after aspirin treatment (i.e., >14 ng/kg/min), we observed that prostacyclin production was not inhibited by aspirin, whereas thromboxane A2 was inhibited. Thus not only did they maintain plasma prostacyclin values relative to plasma thromboxane A 2 (which resulted in an increased prostacyclin/thromboxane A 2 ratio), but their vessels also may have retained the capacity to produce additional prostacyclin when stimulated by infused angiotensin II. This concept is supported by our recent observations in vitro, 11 and in vivo (Magness RR, et al., unpublished observations) that angiotensin II causes increased vascular and uteroplacental production of prostacyclin in ovine pregnancy. In contrast, the subgroup of women whose effective pressor dose of angiotensin II increased little or not at all (i.e., remained <7 ng/kg/min) had a nearly 25% decrease in basal plasma 6-keto-prostaglandin F 10 in addition to an 80% decrease in basal plasma thromboxane B2 • Thus their vessels and platelets appear to be very sensitive to prostaglandin inhibition by aspirin treatment. This also suggests that their vasculature might be very different with regard to prostanoid production and that after aspirin treatment their vessels may have lost the capacity to respond to infused angiotensin II by producing additional prostacyclin to offset or antagonize the vasoconstrictor effects of this octapeptide. Nevertheless, there was a greater decrease in thromboxane A 2 than in prostacyclin and.their ratio was increased; however, there was no change in the effective pressor dose of angiotensin II. Although our participant numbers
Low-dose aspirin and angiotensin II presser response
1041
AFTER ASPIRIN
20
N
a:I
x I-
u:-"'
t..:I 0...
0
I-
UJ
10
:lo:: I
'°0
t:= C(
IX
BEFORE ASPIRIN
0
+
Fig. 4. Ratio of 6-keto-prostaglandin F 10 /thromboxane B, before and after 1 week of 81 mg aspirin treatment (p = 0.002). Values are expressed as means ± 1 SD.
are too small to be conclusive, these observations are exciting and subsequently may help to explain why pregnancy-induced hypertension may recur in some women after the first pregnancy; i.e., there may be a primary vascular defect in prostanoid production among these women that could possibly lead to transient forms of hypertension with each pregnancy. These observations also are of interest with regard to the interpretation of the plasma prostacyclin/thromboxane A 2 ratio. While our overall group of women sensitive to angiotensin II had a substantial and significant increase in their prostacyclin/thromboxane A, ratio after low-dose aspirin, there was greater variability in this ratio after treatment. The likelihood that there were two subgroups of women accounts for this. Although both subgroups had decreases in plasma thromboxane B2 levels, one subgroup appeared to have a much greater fall in plasma thromboxane B2 than the other. Thus the ratio was misleading in the prediction of the participants' responses to therapy as reflected by the angiotensin II pressor response. Therefore the ratio should be used and interpreted with a great deal of caution and careful examination of the actual changes in prostanoids. It has been suggested that the prevention of thromboxane A, domination, without alteration of prostacyclin, might decrease the incidence and severity of
1042 Spitz et al.
pregnancy-induced hypertension. 6 • 7 This study was neither randomized nor blinded; therefore it cannot be viewed as a clinical trial that would address this issue. Nevertheless, it is of interest that of the six women sensitive to angiotensin II who became as refractory after aspirin therapy as the nonsensitive participants, only one was found to have pregnancy-induced hypertension at the end of pregnancy and she did not have proteinuria. On the other hand, of the subgroup that remained relatively sensitive to infused angiotensin II (<7 ng/kg/min), all seven had pregnancy-induced hypertension and six of them had proteinuria. These observations serve as further support that there were at least two populations of women who by chance randomly entered our study. They can be separated pharmacologically, physiologically, and clinically. Whether low-dose aspirin will aid in further delineation of these two groups or whether even lower doses of aspirin will yield similar results remains to be ascertained. It also is of interest to determine whether the sensitive subgroup of gravidas were the patients who failed to respond to aspirin treatment in other studies, regardless of dose. 8 · 9 Another observation made in these studies relates to the differences in heart rate observed in the women either not sensitive or sensitive to angiotensin II. Seligman 19 reported that during normotensive pregnancy, baroreceptor function is set at a high level of sensitivity, i.e., it is more responsive to afterload changes than it is in women with pregnancy-induced hypertension. In this study we observed that decreases in heart rate during angiotensin II administration were 12.6 ± 9.1 beats/min in women not sensitive to angiotensin II as compared with 7.6 ± 6.9 beats/min in the group sensitive to angiotensin II (p = 0.07). This observation is consistent with that of Seligman. 19 The increase in pressor dose of angiotensin II after aspirin treatment could not be explained by a change in baroreceptor function because the decrease in heart rate in this group was unchanged by aspirin therapy (Table II). Our studies in sheep, animals that have been shown to be strikingly similar to women in their response to angiotensin II, support the hypothesis that in normal pregnancy the reduced pressor response to angiotensin II is not due in any large part to altered cardiac responses but reflects a real reduction in vascular responsiveness. 20 The same may be said of women. A surprising finding in our study was that the basal heart rate values at equal blood pressures were lower in women sensitive to angiotensin II than in women not sensitive to angiotensin II. If the stroke volume is not increased (and little is known about this at this stage of the disease), it appears that women sensitive to angiotensin II may face increased peripheral vascular re-
November 1988 Am J Obstet Gynecol
sistance and maintain a normal blood pressure at least in part by decreasing their cardiac output. What role, if any, prostaglandins have in this heart rate alteration, and presumably vascular resistance, remains to be determined. We thank Laurie Daniels for her technical assistance in the preparation of this manuscript and the house officers in obstetrics and gynecology at Parkland Memorial Hospital for their assistance in the clinical management of the women described in this study. REFERENCES 1. Fitzgerald DJ, Entman SS, Mulloy K, Fitzgerald GA. Decreased prostacyclin biosynthesis preceding the clinical manifestation of pregnancy-induced hypertension. Circulation l 987;75:956. 2. Goodman RP, Killam AP, Brash AR, Branch RA. Prostacyclin production during pregnancy: comparison of production during normal pregnancy and pregnancy complicated by hypertension. AM J OBSTET GYNECOL 1982;142:817. 3. Walsh SW. Preeclampsia: an imbalance in placental prostacyclin and thromboxane production. AM J 0BSTET GYNECOL 1985; 152:335. 4. Gant NF, Daley GL, Chand S, Whalley PJ, MacDonald PC. A study of angiotensin II pressor response throughout primigravid pregnancy. J Clin Invest 1973;52:2682. 5. Fitzgerald DJ, Mayo G, Catella F, Entman SS, Fitzgerald GA. Increased thromboxane biosynthesis in normal pregnancy is mainly derived from platelets. AM J OBSTET GYNECOL 1987; 157:325. 6. Collins E, Turner G. Maternal effects of regular salicylate ingestion in pregnancy. Lancet 1975;2:335. 7. Crandon AJ, Isherwood DM. Effect of aspirin on incidence of pre-eclampsia. Lancet 1979; 1: 1356. 8. Beaufils M, Uzan S, Donsimoni R, Colau JC. Prevention of pre-eclampsia by early antiplatelet therapy. Lancet 1985; 1:840. 9. Walleriburg HCS, Dekker GA, Makovitz JW, Rotmans P. Low-dose aspirin prevents pregnancy-induced hypertension and preeclampsia in angiotensin-sensitive primigravidae. Lancet 1986; 1: 1. 10. Gant NF, Chand S, Worley RJ, Whalley PJ, Crosby UD, MacDonald PC. A clinical test useful for predicting the development of acute hypertension in pregnancy. AM J 0BSTET GYNECOL 1974;120:1. 11. Everett RB, Worley RJ, MacDonald PC, Gant NF. Effect of prostaglandin synthetase inhibitors on pressor response to angiotensin II in human pregnancy. J Clin Endocrinol Metab 1978;46:1007. 12. Everett RB, Worley RJ, MacDonald PC, Gant NF. Oral administration of theophylline to modify pressor responsiveness to angiotensin II in women with pregnancyinduced hypertension. AM J OBSTET GYNECOL 1978; 132:359. 13. Campbell WB, Ojeda SR. Measurements of prostaglandins by radioimmunoassay. Methods Enzymol. 1987; 141:323. 14. 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 Invest 1985;76:2206. 15. Preston FE, Whipps S,Jackson CA, French AJ, Wyld PJ, Stoddard CJ. Inhibition of prostacyclin and platelet thromboxane A 2 after low-dose aspirin. N Engl J Med 1981;304:76. 16. Hanley SP, Bevan J, Cockbill SR, Heptinstall S. Differ-
Volume 159 Number 5
ential inhibition by low-dose aspirin of human venous prostacyclin synthesis and platelet thromboxane synthesis. Lancet 1981; 1:969. 17. Satoh H, Hosono M, Satoh S. Distinctive effect of angiotensin II on prostaglandin production in dog renal and femoral arteries. Prostaglandins 1984;27:807. 18. Mitchell MD, Flint APF. Prostaglandin production by intra-uterine tissues from periparturient sheep: use of a superfusion technique. J Endocrinol 1978;76: 111.
Low-dose aspirin and angiotensin II presser response
19. Seligman SA. Baroreceptor reflex function in preeclampsia. J Obstet Gynaecol Br Commonw 1971:78:413. 20. Naden RP, Gant NF, Rosenfeld CR. The pressor response to angiotensin II: The roles of peripheral and cardiac responses in pregnant and non pregnant sheep. AM J OBSTET GYNECOL 1984;148:450.
Preeclampsia: A microvesicular fat disease of the liver? Hisanori Minakami, MD," Naoko Oka, MD," Tadashi Sato, MD," Taro Tamada, MD," Yoshikazu Yasuda, MD,b and Norio Hirota, MD< Tochigi-ken, Japan To study the interrelationship between preeclampsia, the syndrome of hemolysis, elevated liver enzymes, and low platelet count, and acute fatty liver of pregnancy, 41 liver specimens from 41 preeclamptic women with and without liver dysfunction were examined for the amount of fat deposited in hepatocytes. All 41 specimens stained with oil red 0 on frozen sections showed a significant amount of microvesicular fat droplets in varying degrees. In contrast, only 11 of the 41 stained conventionally (with hematoxylin-eosin) showed significant fatty infiltration. The density of hepatocellular fat correlated positvely with plasma urate concentration and negatively with the platelet count. These findings suggest that preeclampsia may be one of several microvesicular fatty diseases of the liver and that acute fatty liver of pregnancy may be the most severe form. (AM J OBSTET GYNECOL 1988;159:1043-7.)
Key words: Preeclampsia, fatty liver, hyperuricemia, microvesicular fat disease Acute fatty liver of pregnancy is an uncommon but potentially fatal disorder that may complicate the third trimester.' Acute fatty liver of pregnancy is considered an independent clinical entity whose definitive diagnosis requires histologic examination of the liver.2 A microscopic feature of acute fatty liver of pregnancy is microvesicular fat vacuolation of the hepatocytes.2 Recently, the incidence of acute fatty liver of pregnancy has apparently increased, mainly because of a greater awareness and early recognition of this disorder, particularly of milder cases. 3 The syndrome of hemolysis, elevated liver enzymes, and low platelet count in the third trimester is thought to be a form of preeclampsia, even though it sometimes lacks the traditional signs of this disorder.4· 5 Furthermore, it is reported that traditional preeclampsia complicates liver dysfunction in a considerable number of cases. 6 Since the three clinical conditions-acute fatty liver of preg-
From the Departments of Obstetrics and Gynecology," Surgery,' and Pathology,' ]ichi Medical School. Received for publication December 14, 1987; revised May 13, 1988; accepted May 19, 1988. Reprint requests: Hisanori Minakami, MD, Jichi Medical School, Minamikawachi-machi, I'ochigi,-ken, 329-04 japan.
nancy, the syndrome of hemolysis, elevated liver enzymes, and low platelet count, and preeclampsia with liver dysfunction-share many clinical and laboratory abnormalities, it is possible that they form a disease spectrum. However, there has been no evidence to confirm this hypothesis. To resolve the issue, we performed liver biopsies or autopsies on preeclamptic patients with and without liver dysfunction, in patients who fulfilled the criteria of hemolysis, elevated liver enzymes, and low platelet count syndrome described by Weinstein,' and in patients suspected of having acute fatty liver of pregnancy.
Patients and methods We examined liver specimens from three patients who died of preeclampsia complicated by disseminated intravascular coagulation and hepatic failure in the third trimester and from 38 preeclamptic patients from whom specimens were obtained within 14 days of delivery. Preeclampsia was defined as a systolic or diastolic blood pressure ;;;, 140 or 90 mm Hg on two occasions, recorded 24 hours apart, with the presence of proteinuria (>500 mg/day) in a patient who had been normotensive during the first 20 weeks of pregnancy. Patients ranged in age from 21 to 43 years. All had re-
1043