- Email: [email protected]
Vasoactive effect of sera from preeclamptic patients
Vasoactive effect of sera from preeclamptic patients HAROLD W. PUFFER SL'/ANN E. CHEEK GARY K. OAKES* ~lA\1
K. KIM
GREGORY R. DANTZLER JOHN KUROHARA NANCY E. WARNER Lw Angeles, California Investigated was a bioassay method for measurement of vasoactivity in the serum of preeclamptic patients. Intravital microscopy was used to measure the diameter of the principal arteriole and venule in the gastrolienal mesentery of mice before (VD,) and after (VD~) the intravenous injection of sera from preeclamptic primigravid patients and from preeclamptic primigravid patients after treatment with magnesium sulfate (MgS0 4). A vasoactive index (VI) was calculated as VD2 /VD,. The VIa (vasoactive index for arteriole) of preeclamptic primigravid patients (0.64 ± 0.39, mean ± SO) was significantly different (p < 0.01) from the VIa of normal primigravid patients
(0.91
±
0.08) and from the Vfa of saline controls (0.99 ± 0.02). Four of the seven preeC!arT'.ptic
primigravid patients had a duration of signs and symptoms dating from the antepartum period (Via = 0.53 ± 0.42), and for three, the diagnosis was made only in the intrapartum period (Via = 0.77 ± 0.38). Sera from the MgS04-treated preeclamptic patients yielded a VIa of 0.91 ± 0.14, which was not significantly different from that for either the normal primigravid patients or the saline controls. This VIa value was significantly different (p < 0.01) from the VIa of the untreated preeclamptic patients. These results suggest that there is a vasoactive substance(s) in the sera of preeclamptic patients which may c-Ontribute to their arteriolar vasospasm. The magnitude of vasoactivity appears to be related to the duration and severity of preeclampsia but is masked or disappears after treatment of the patient with MgSO•. (AM. J. OBSTET. GYNECOL. 142:468, 1982.)
HYPERTENSIVE DISORDERS are generally recognized as being of major concern during pregnancy. They occur in about 5% of pregnancies in the United States, 1 and the most severe form, eclampsia, results in a mortality rate of six per 100,000. with a perinatal mortality rate as high as 20%. Abnormal vasospasm is a basic pathologic feature of
From the Departments of Pathology, Obstetrics and GynPCology, and Nur:sing, Los Angeles County!Universit)' of Southern California Medical Center. Supported by Grant 550 from the American Heart Association of Los Angeles. Presented m part at the Twentyjijth Annual Meeting of the Society for Gynecologic Investigation, Atlanta, Georgia, March 15-18, 1978. Reprint requests: Dr. Harold W. Puffer, Department of Pathology, School of Medicine, University of Southern
the hypertensive syndrome. Although several studies have investigated this phenomenon, 2 the etiology remains unknown. Some workers have attributed the hypertension to increased reactivity of the resistance vessels, whereas others have reported the possible presence of vasoactive substances. 3- 5 Progress in isolation and identification of putative vasoactive substances has been hampered by lack of suitable assay. In a previous study, Puffer and associates 6 reported vasoconstriction in the microvasculature of mice injected with serum from patients with toxemia of pregnancy. The results were reported as qualitative observations. This study was undertaken to explore the use of microcirculatory techniques in the development of a quantitative bioassay procedure for identifying vasoactive factors in hypertensive disorders of pregnancy.
California; 2025 Zoro;,0;[ Ave., Los Angeles, California 90033. *Current address: Department of Obstetrics arui Gynecology, Harbor General Hospital, Torrance, California.
468
Female Swiss albino mice that weighed lO to 20 gm and were fed laboratory chow and water ad libitun1 were anesthetized with chloralose (2 gm/100 ml) and 0002-9378/82/040468+03$00.30/0
©
1982 The C. V. Mosby Co.
Vasoactive effect of sera from preeclamptic patients
Volume 142 Number 4
urethane (10 gm/100 ml) solution, 5 mllkg, injected intraperitoneally. The hair was removed from the left flank, and an incision was made in the abdominal wall over the spleen. The lower portion of the spleen and attached mesentery were gently lifted from the body of the animal. Exposed tissues were bathed with LockeRinger's solution at 3 7° C. The microcirculation was examined by transillumination with the use of borosilicate rod and Leitz stereomicroscope equipped with 2, 4, 8, and l2.5X objective as previously described. 7 The observed area was recorded on videotape and photographed periodically by means of a Leitz monobjective trinocular microscope equipped with a 3.8x objective and 12.5X eyepiece, a Leica M-1 camera, ITP video camera, image processor, electronic micrometer, Sony 3650 video recorder, and a Leitz filar micrometer. Studied were samples of blood from seven preeclamptic, untreated primigravid patients and five preeclamptic primigravid patients who were treated with MgSO~. and samples of cord blood from three primigravid patients who had been treated with MgS0 4 • All samples of serum were obtained in the early intrapartum period. The standard criteria of the American Committee on Maternal Welfare were employed to diagnose preeclampsia.8 Samples from four normal primigravid patients were used as controls. Six control mice were injected with saline solution. Ten-milliliter samples of blood were obtained from each patient. The samples were centrifuged and the decanted serum was used within 1 to 2 hours after collection. The serum, 7.5 ml/kg, was injected intravenously via tail vein in the mouse. Diameters of the most prominent arteriole and venule were measured before and after administration of sera. A vasoactive index, 9 VI, was calculated by the formula VI= VD2/VD1 where VD 1 diameter of vessel prior to administration of serum, and VD 2 = diameter of vessel after administration of serum. Animals were examined post mortem, and selected tissues were prepared for examination by light microscopy. To prevent bias, separate teams collected samples of blood, performed bioassays, and made diagnoses on a double-blind basis. Only after all data had been collected were correlations made between diagnosis and results of the bioassay. At no time prior to this did any team know the results obtained by the other teams. Rigid criteria were applied for diagnosis of patients. Several patients were omitted from the study for lack of accurate data, and some bioassay results were deleted because of improper handling of the samples of blood. That is, all blood was assayed within 1 to 2 hours after colleeiion; aU older samples were discarded and
469
Table I. VIs obtained after administration of sera from preeclamptic primigravid patients to mice Patient
Arteriole (VIa)
Venule (Vlv)
B.O.* M.H.* M. A.t R. G.t M. C.* A. D.* E. R.t
0.46 0.00 0.73 0.42 1.00 0.67 1.17
1.00 0.15 0.82 1.00 0.92 0.52 1.60
Average arteriole VI = 0.64:!: 0.39 (mean:!: SD). Average venule VI= 0.86 ± 1.45 (mean± SD). *Preeclampsia dating from antepartum period. tDiagnosed as preeclamptic in intrapartum period.
Table II. VI obtained after administration of sera from normal primigravid patients to mice Patient
s. s. R. R. M.R. E. P.
Arteriole (VI,)
Venule (Vi,.)
0.93 0.86 1.00 0.83
1.03 0.90 0.92 1.00
Average arteriole VI 0.91 ± 0.08 (mean ± SD). Average venule VI 0.96 ± 0.06 (mean SD).
no tests were performed on samples held longer than 2 hours after collection.
Results and comment Results observed in vivo included changes in vascular diameter, intravascular agglutination, adherence of leukocytes to vessel walls, and changes in flow rate and respiration. Autopsy of animals revealed congested liver and kidneys in four of the mice injected with sera from preeclamptic primigravid patients, and in one mouse injected with sera from a normal primigravid patient. Most striking in vi\'0 effects were vasoconstriction and intravascular agglutination. Changes in vascular diameter were used to calculate a vasoactive index, VI, as described above. Values for VI are shown in Tables 1-V. A VI of l indicated no change in vascular diameter. Vasodilatation resulted in VI > l, whereas vasoconstriction resulted in VI < l. As seen in Table II, the administration of sera from normal primigravid patients resulted in a mean vasoactive index for arterioles (VIa) of 0.91 ± 0.08, and for venules (VI,) of 0.96 ± 0.06. These values did not differ significantly from values obtained by injection of saline solution. VI,.= 0.99 ± 0.02 and VI,= 0.91 ± 0.09 (Table V). In all cases (Tables I-V), the values for VI, were not significantly different. This was to be expected, since venules are known to lack the vasoresponsive capacity inherent in arterioles.
470 Puffer et al.
February 15, J9H2
Am
Table III. VIs obtained after administration of preeclamptic sera from primigravid patients treated with MgS0 4 to mice Patient5 F. M.
B. B. F. G. L.L I. M.
Arteriole (VI,.)
Venule (VI,.)
1.00 0.93 0.78 1.08 0.75
l.l8 0.90 0.83 0.97 0.96
.J. Ohstet. Gynecol
Table V. VI obtained after administration of saline solution at room temperature to mice }vfousr No.
2
3 4
5 6
Arteriole (VI,)
Venule (Vl,.J
0.80 1.00 1.13 1.00 0.88 1.!5
0.83 0.96 0.96 1.02 0.90 0.77
Average arteriole VI= 0.91 ± 0.14 (mean± SD). Average venule VI "" 0.97 ± 1.13 (mean ± SD).
Average arteriole VI 0.99 :t 0.02 (mean ±SD). Average venule VI 0.9I ± 0.09 (mean :t SD).
Table IV. VIs obtained after administration of preeclamptic umbilical cord sera from primigravid patients treated with MgS0 4 to mice
rent from those for either the normal primigravid patients or the saline controls. However, these VC values were significantly different (p < 0.01) from the VI,, of untreated preeclamptic patients. Because vessels of different sizes are known to exhibit different responses, first-order branching arterioles and venules of the splenic artery and vein were selected for observation. Only vessels of the same range of size were used. The mean arteriolal' diameter was 0.059 ± 0.22 mm, and the mean venular diameter was 0.110 ± 0.06 mm. Two samples were heated in a water bath at 42° C for 3 hours. These samples lost significant activity. This inactivation suggests that a heat-labile vasoactive substance may be responsible for the vasoactive response. 4 The above-described results suggest that there is a vasoactive substance(s) in the sera of preeclamptic patients which may contribute to the arteriolar vasospasm observed in this syndrome. The magnitude of vasoactivity appears to be related to the duration and severity of preeclampsia, but is masked or disappears after treatment with MgS04. The methodology described above may be useful as a bioassay for vasoactive substance(s) in the sera of patients suffering from hypertensive disorders of pregnancy.
Patient I. H. M.M. E. N.
I
Arteriole (VI")
1.00 0.50 1.00
\
Venule (VI,)
0.97 0.72 1.08
Average arteriole VI 0.83 0.29 (mean ± SD). Average venule VI 0.92 ± 0.18 (mean± SD).
The VIa of preeclamptic primigravid patients (Table I, 0.64 ± 0.39) was singificantly different (p < 0.0 l) from that of normal primigravid patients (Table II, 0.91 ± O.OR) and from the VIa for saline controls (Table V, 0.99 ± 0.02). Four of the seven preeclamptic primigravid patients had a duration of signs and symptoms dating from the antepartum period (VIa = 0.53 ± 0.42), and for three, the diagnosis was made onlv in the intrapartum period (VIa 0.77 ± 0.38). Sera from preeclamptic patients treated with MgS0 4 (Table Ill) yielded a VIa of 0.91 ± 0.14. Sera from samples of cord blood from preeclamptic patients treated with MgSO.; (Table IV) yielded a VI" of O.R3 ± 0.29. These values were not significantly diffe-
I.
2.
3. 4.
5.
REFERENCES de Alvarez, R. R.: Preeclampsia-eclampsia and hypertensive disorders in pregnancy, Clin. Obstet. Gynecol. 16:47, 1973. Assali, N. S., Longo, L. D., and Holm, L. W.: Toxemia.Jike syndromes in animals, spontaneous and experimental, Obstet. Gynecol. Surv. 15:151, 1960. Hunter, C. A., and Howard, W. F.: A pressor substance (hysterotonin) occurring in toxemia, AM. J 0BSTET. GYNECOL. 79:838, 1960. Vorne, M.S., Jarvinen, P. A., Jarvi,]., Ylostalo, P.R., and Karki, N .: Pressor substances in the posthemodialysis dialysates of patients with toxemia of pregnancy, AM.]. 0BSTET. GYNECOL. 119:610, 1974. Pi rani, B. B. K., and MacGillivray, 1.: The effect of plasma retransfusion on the blood pressure in the puerperium, AM.]. 0BSTET. GYNECOL.l%1:221, 1975.
6. Puffer, H. W., Warner, N. E., and Martin, C. B., Jr.: Preliminary observations of the effect of serum from patients with toxemia of pregnancy, Bib!. Anal. 12:13, 1973. 7. Warner, N. E., .Bunnag, S., Bunnag, S. C., and Bennington,]. L.: Micr01;irculation in the islets of Langethans in experimental endotoxemia in mice, Bibl. Anat; 7:449, 1965. 8. Chesley, L. C.: Hypertensive disorders in pregnancy, in HeUman, L. M., and Pritchard, S. A., editors: WilliaMs Obstetrics, ed. 14, New York, 1971, Appleton-Century· Crofts, pp. 685-746. 9. Puffer, H. W., Warner, N. E., and Barrett, W. P:: Development of an index for evaluating the r~ponse of the microvasculature in C3H mice to various vasoactive drugs, Microcirculation 1:86, 197£.
K. KIM
GREGORY R. DANTZLER JOHN KUROHARA NANCY E. WARNER Lw Angeles, California Investigated was a bioassay method for measurement of vasoactivity in the serum of preeclamptic patients. Intravital microscopy was used to measure the diameter of the principal arteriole and venule in the gastrolienal mesentery of mice before (VD,) and after (VD~) the intravenous injection of sera from preeclamptic primigravid patients and from preeclamptic primigravid patients after treatment with magnesium sulfate (MgS0 4). A vasoactive index (VI) was calculated as VD2 /VD,. The VIa (vasoactive index for arteriole) of preeclamptic primigravid patients (0.64 ± 0.39, mean ± SO) was significantly different (p < 0.01) from the VIa of normal primigravid patients
(0.91
±
0.08) and from the Vfa of saline controls (0.99 ± 0.02). Four of the seven preeC!arT'.ptic
primigravid patients had a duration of signs and symptoms dating from the antepartum period (Via = 0.53 ± 0.42), and for three, the diagnosis was made only in the intrapartum period (Via = 0.77 ± 0.38). Sera from the MgS04-treated preeclamptic patients yielded a VIa of 0.91 ± 0.14, which was not significantly different from that for either the normal primigravid patients or the saline controls. This VIa value was significantly different (p < 0.01) from the VIa of the untreated preeclamptic patients. These results suggest that there is a vasoactive substance(s) in the sera of preeclamptic patients which may c-Ontribute to their arteriolar vasospasm. The magnitude of vasoactivity appears to be related to the duration and severity of preeclampsia but is masked or disappears after treatment of the patient with MgSO•. (AM. J. OBSTET. GYNECOL. 142:468, 1982.)
HYPERTENSIVE DISORDERS are generally recognized as being of major concern during pregnancy. They occur in about 5% of pregnancies in the United States, 1 and the most severe form, eclampsia, results in a mortality rate of six per 100,000. with a perinatal mortality rate as high as 20%. Abnormal vasospasm is a basic pathologic feature of
From the Departments of Pathology, Obstetrics and GynPCology, and Nur:sing, Los Angeles County!Universit)' of Southern California Medical Center. Supported by Grant 550 from the American Heart Association of Los Angeles. Presented m part at the Twentyjijth Annual Meeting of the Society for Gynecologic Investigation, Atlanta, Georgia, March 15-18, 1978. Reprint requests: Dr. Harold W. Puffer, Department of Pathology, School of Medicine, University of Southern
the hypertensive syndrome. Although several studies have investigated this phenomenon, 2 the etiology remains unknown. Some workers have attributed the hypertension to increased reactivity of the resistance vessels, whereas others have reported the possible presence of vasoactive substances. 3- 5 Progress in isolation and identification of putative vasoactive substances has been hampered by lack of suitable assay. In a previous study, Puffer and associates 6 reported vasoconstriction in the microvasculature of mice injected with serum from patients with toxemia of pregnancy. The results were reported as qualitative observations. This study was undertaken to explore the use of microcirculatory techniques in the development of a quantitative bioassay procedure for identifying vasoactive factors in hypertensive disorders of pregnancy.
California; 2025 Zoro;,0;[ Ave., Los Angeles, California 90033. *Current address: Department of Obstetrics arui Gynecology, Harbor General Hospital, Torrance, California.
468
Female Swiss albino mice that weighed lO to 20 gm and were fed laboratory chow and water ad libitun1 were anesthetized with chloralose (2 gm/100 ml) and 0002-9378/82/040468+03$00.30/0
©
1982 The C. V. Mosby Co.
Vasoactive effect of sera from preeclamptic patients
Volume 142 Number 4
urethane (10 gm/100 ml) solution, 5 mllkg, injected intraperitoneally. The hair was removed from the left flank, and an incision was made in the abdominal wall over the spleen. The lower portion of the spleen and attached mesentery were gently lifted from the body of the animal. Exposed tissues were bathed with LockeRinger's solution at 3 7° C. The microcirculation was examined by transillumination with the use of borosilicate rod and Leitz stereomicroscope equipped with 2, 4, 8, and l2.5X objective as previously described. 7 The observed area was recorded on videotape and photographed periodically by means of a Leitz monobjective trinocular microscope equipped with a 3.8x objective and 12.5X eyepiece, a Leica M-1 camera, ITP video camera, image processor, electronic micrometer, Sony 3650 video recorder, and a Leitz filar micrometer. Studied were samples of blood from seven preeclamptic, untreated primigravid patients and five preeclamptic primigravid patients who were treated with MgSO~. and samples of cord blood from three primigravid patients who had been treated with MgS0 4 • All samples of serum were obtained in the early intrapartum period. The standard criteria of the American Committee on Maternal Welfare were employed to diagnose preeclampsia.8 Samples from four normal primigravid patients were used as controls. Six control mice were injected with saline solution. Ten-milliliter samples of blood were obtained from each patient. The samples were centrifuged and the decanted serum was used within 1 to 2 hours after collection. The serum, 7.5 ml/kg, was injected intravenously via tail vein in the mouse. Diameters of the most prominent arteriole and venule were measured before and after administration of sera. A vasoactive index, 9 VI, was calculated by the formula VI= VD2/VD1 where VD 1 diameter of vessel prior to administration of serum, and VD 2 = diameter of vessel after administration of serum. Animals were examined post mortem, and selected tissues were prepared for examination by light microscopy. To prevent bias, separate teams collected samples of blood, performed bioassays, and made diagnoses on a double-blind basis. Only after all data had been collected were correlations made between diagnosis and results of the bioassay. At no time prior to this did any team know the results obtained by the other teams. Rigid criteria were applied for diagnosis of patients. Several patients were omitted from the study for lack of accurate data, and some bioassay results were deleted because of improper handling of the samples of blood. That is, all blood was assayed within 1 to 2 hours after colleeiion; aU older samples were discarded and
469
Table I. VIs obtained after administration of sera from preeclamptic primigravid patients to mice Patient
Arteriole (VIa)
Venule (Vlv)
B.O.* M.H.* M. A.t R. G.t M. C.* A. D.* E. R.t
0.46 0.00 0.73 0.42 1.00 0.67 1.17
1.00 0.15 0.82 1.00 0.92 0.52 1.60
Average arteriole VI = 0.64:!: 0.39 (mean:!: SD). Average venule VI= 0.86 ± 1.45 (mean± SD). *Preeclampsia dating from antepartum period. tDiagnosed as preeclamptic in intrapartum period.
Table II. VI obtained after administration of sera from normal primigravid patients to mice Patient
s. s. R. R. M.R. E. P.
Arteriole (VI,)
Venule (Vi,.)
0.93 0.86 1.00 0.83
1.03 0.90 0.92 1.00
Average arteriole VI 0.91 ± 0.08 (mean ± SD). Average venule VI 0.96 ± 0.06 (mean SD).
no tests were performed on samples held longer than 2 hours after collection.
Results and comment Results observed in vivo included changes in vascular diameter, intravascular agglutination, adherence of leukocytes to vessel walls, and changes in flow rate and respiration. Autopsy of animals revealed congested liver and kidneys in four of the mice injected with sera from preeclamptic primigravid patients, and in one mouse injected with sera from a normal primigravid patient. Most striking in vi\'0 effects were vasoconstriction and intravascular agglutination. Changes in vascular diameter were used to calculate a vasoactive index, VI, as described above. Values for VI are shown in Tables 1-V. A VI of l indicated no change in vascular diameter. Vasodilatation resulted in VI > l, whereas vasoconstriction resulted in VI < l. As seen in Table II, the administration of sera from normal primigravid patients resulted in a mean vasoactive index for arterioles (VIa) of 0.91 ± 0.08, and for venules (VI,) of 0.96 ± 0.06. These values did not differ significantly from values obtained by injection of saline solution. VI,.= 0.99 ± 0.02 and VI,= 0.91 ± 0.09 (Table V). In all cases (Tables I-V), the values for VI, were not significantly different. This was to be expected, since venules are known to lack the vasoresponsive capacity inherent in arterioles.
470 Puffer et al.
February 15, J9H2
Am
Table III. VIs obtained after administration of preeclamptic sera from primigravid patients treated with MgS0 4 to mice Patient5 F. M.
B. B. F. G. L.L I. M.
Arteriole (VI,.)
Venule (VI,.)
1.00 0.93 0.78 1.08 0.75
l.l8 0.90 0.83 0.97 0.96
.J. Ohstet. Gynecol
Table V. VI obtained after administration of saline solution at room temperature to mice }vfousr No.
2
3 4
5 6
Arteriole (VI,)
Venule (Vl,.J
0.80 1.00 1.13 1.00 0.88 1.!5
0.83 0.96 0.96 1.02 0.90 0.77
Average arteriole VI= 0.91 ± 0.14 (mean± SD). Average venule VI "" 0.97 ± 1.13 (mean ± SD).
Average arteriole VI 0.99 :t 0.02 (mean ±SD). Average venule VI 0.9I ± 0.09 (mean :t SD).
Table IV. VIs obtained after administration of preeclamptic umbilical cord sera from primigravid patients treated with MgS0 4 to mice
rent from those for either the normal primigravid patients or the saline controls. However, these VC values were significantly different (p < 0.01) from the VI,, of untreated preeclamptic patients. Because vessels of different sizes are known to exhibit different responses, first-order branching arterioles and venules of the splenic artery and vein were selected for observation. Only vessels of the same range of size were used. The mean arteriolal' diameter was 0.059 ± 0.22 mm, and the mean venular diameter was 0.110 ± 0.06 mm. Two samples were heated in a water bath at 42° C for 3 hours. These samples lost significant activity. This inactivation suggests that a heat-labile vasoactive substance may be responsible for the vasoactive response. 4 The above-described results suggest that there is a vasoactive substance(s) in the sera of preeclamptic patients which may contribute to the arteriolar vasospasm observed in this syndrome. The magnitude of vasoactivity appears to be related to the duration and severity of preeclampsia, but is masked or disappears after treatment with MgS04. The methodology described above may be useful as a bioassay for vasoactive substance(s) in the sera of patients suffering from hypertensive disorders of pregnancy.
Patient I. H. M.M. E. N.
I
Arteriole (VI")
1.00 0.50 1.00
\
Venule (VI,)
0.97 0.72 1.08
Average arteriole VI 0.83 0.29 (mean ± SD). Average venule VI 0.92 ± 0.18 (mean± SD).
The VIa of preeclamptic primigravid patients (Table I, 0.64 ± 0.39) was singificantly different (p < 0.0 l) from that of normal primigravid patients (Table II, 0.91 ± O.OR) and from the VIa for saline controls (Table V, 0.99 ± 0.02). Four of the seven preeclamptic primigravid patients had a duration of signs and symptoms dating from the antepartum period (VIa = 0.53 ± 0.42), and for three, the diagnosis was made onlv in the intrapartum period (VIa 0.77 ± 0.38). Sera from preeclamptic patients treated with MgS0 4 (Table Ill) yielded a VIa of 0.91 ± 0.14. Sera from samples of cord blood from preeclamptic patients treated with MgSO.; (Table IV) yielded a VI" of O.R3 ± 0.29. These values were not significantly diffe-
I.
2.
3. 4.
5.
REFERENCES de Alvarez, R. R.: Preeclampsia-eclampsia and hypertensive disorders in pregnancy, Clin. Obstet. Gynecol. 16:47, 1973. Assali, N. S., Longo, L. D., and Holm, L. W.: Toxemia.Jike syndromes in animals, spontaneous and experimental, Obstet. Gynecol. Surv. 15:151, 1960. Hunter, C. A., and Howard, W. F.: A pressor substance (hysterotonin) occurring in toxemia, AM. J 0BSTET. GYNECOL. 79:838, 1960. Vorne, M.S., Jarvinen, P. A., Jarvi,]., Ylostalo, P.R., and Karki, N .: Pressor substances in the posthemodialysis dialysates of patients with toxemia of pregnancy, AM.]. 0BSTET. GYNECOL. 119:610, 1974. Pi rani, B. B. K., and MacGillivray, 1.: The effect of plasma retransfusion on the blood pressure in the puerperium, AM.]. 0BSTET. GYNECOL.l%1:221, 1975.
6. Puffer, H. W., Warner, N. E., and Martin, C. B., Jr.: Preliminary observations of the effect of serum from patients with toxemia of pregnancy, Bib!. Anal. 12:13, 1973. 7. Warner, N. E., .Bunnag, S., Bunnag, S. C., and Bennington,]. L.: Micr01;irculation in the islets of Langethans in experimental endotoxemia in mice, Bibl. Anat; 7:449, 1965. 8. Chesley, L. C.: Hypertensive disorders in pregnancy, in HeUman, L. M., and Pritchard, S. A., editors: WilliaMs Obstetrics, ed. 14, New York, 1971, Appleton-Century· Crofts, pp. 685-746. 9. Puffer, H. W., Warner, N. E., and Barrett, W. P:: Development of an index for evaluating the r~ponse of the microvasculature in C3H mice to various vasoactive drugs, Microcirculation 1:86, 197£.