- Email: [email protected]
Transplacental transfer of zidovudine in the near-term pregnant baboon
Transplacental transfer of zidovudine in the near-term pregnant baboon Gary D. V. Hankins, MD,. Curtis L. Lowery, Jr., MD,c Richard T. Scott, MD,. W. Robert Morrow, MD: K. D. Carey, DVM, PhD," M. Michelle Leland, DVM,b and Edward V. Colvin, MDC Lackland Air Force Base and San Antonio, Texas, and Birmingham, Alabama Approximately one third of infants born to human immunodeficiency virus type 1 seropositive mothers have evidence of infection or of acquired immunodeficiency syndrome by the age of 18 months. One fifth of infected infants also have died by age 18 months. This prevalence, combined with the demonstration that zidovudine (formerly azidothymidine) can decrease mortality and the frequency of opportunistic infections in patients with acquired immunodeficiency syndrome or acquired immunodeficiency syndrome-related complex, may lead to increasing use of azidothymidine in pregnancy despite a paucity of information regarding its pharmacokinetics. To further investigate the distribution of azidothymidine and its inactive metabolite S'-glucuronide azidothymidine in the mother, fetus, and amniotic fluid, 12 near-term pregnant baboons were given oral azidothymidine (21 mg/kg/day in four divided doses every 6 hours, equivalent to the usual nonpregnant human dose of 1500 mg/day). Specimens of maternal blood, fetal arterial blood obtained by percutaneous umbilical cord blood sampling, and amniotic fluid were obtained after from one to 17 doses of azidothymidine. Azidothymidine levels were measured by radioimmunoassay with the INCSTAR commercial radioimmunoassay kit and using Escherichia coli [3-glucuronidase for determination of S'-glucuronide azidothymidine levels. Paired analyses revealed Significant concentration gradients between amniotiC fluid, fetal serum, and maternal serum for both azidothymidine (p < 0.019) and S'-glucuronide azidothymidine (p < 0.002). The amniotic fluid S'-glucuronide azidothymidine level increased with increasing doses of azidothymidine despite the fact that the maternal azidothymidine and S'-glucuronide azidothymidine concentrations were unchanged. This accumulation of amniotic fluid S'-glucuronide azidothymidine may provide a functional drug reservoir and contribute to the higher fetal concentrations of the medication and its metabolite. Alternatively, the higher fetal levels may represent slower clearance in the fetus than in the mother. Further studies appear warranted with respect to possible adverse fetal effects, especially bone marrow suppression with prolonged and chronic exposure to azidothymidine. (AM J OBSTET GVNECOL 1990;163:728-32.)
Key words: Azidothymidine, acquired immunodeficiency syndrome, placental transfer Few disease states have garnered national and international attention and action as quickly as the acquired immunodeficiency syndrome (AIDS). Demand for responsiveness by the scientific and medical community has been intense and action has been prompt. The Food and Drug Administration approved clinical testing of zidovudine (azidothymidine) within 5 days of receipt of its sponsor's application. The review and approval of the zidovudine New Drug Application required only 107 days and clinical studies of only 22 months. 1 Optimism over the demonstration that azidothymidine adFrom the Wilford Hall United States Air Force Medical Center, Lackland Air Force Base: Southwest Foundation for Biomedical Research, San Antonio,' and the University of Alabama at Birmingham.' Presented at the Tenth Annual Meeting of the Society of Perinatal Obstetricians, Houston, Texas, January 23-27, 1990. The opinions expressed in this manuscript are those of the authors and not necessarily those of the United States Air Force or the Department of Defense. Reprint requests: Cary D. V. Hankins, MD, Wilford Hall USAF Medical CenteriSCHO, Lackland AFB, TX 78236-5300. 616122358
728
ministration could decrease mortality and the frequency of opportunistic infections in a selected group of subjects with AIDS or AIDS-related complex 2 has resulted in ongoing phase I clinical trials of zidovudine in pregnancy. Notwithstanding, published data on azidothymidine in pregnancy, especially in the fetus, are limited to a single study involving one pregnant macaque monkey.3 In an acute preparation involving intravenous administration of zidovudine (44.1 J.Lg/kg/min) for 3Y2 hours before cesarean birth, azidothymidine was demonstrated to cross the placenta with a ratio of the steady-state concentration of the drug in fetus/ mother of 0.74. To more closely approximate the usual therapy used in actual clinical practice, we administered the drug per os to 12 pregnant baboons, in the third trimester of pregnancy. Because the drug is metabolized in the liver by attachment of glucuronic acid in the 5' position with subsequent renal excretion of both the free and conjugated drug, the ability of the fetus to metabolize the drug might be impaired by delayed maturation of the glucuronide enzyme system.
Volume 163 Number 3
Transplacental transfer of zidovudine and its metabolite
Indeed, in nonpregnant adults with concurrent acetaminophen and azidothymidine treatment, a higher frequency of hematologic complications have been reported and attributed to increased azidothymidine toxicity as a result of competitive inhibition of hepatic metabolism by the glucuronide system" Accordingly, information on the degree to which azidothymidine and its metabolite 5'-glucuronide azidothymidine cross the placenta and the levels achieved in the fetus and the amniotic fluid with prolonged long-term therapy is of great importance in assessment of its toxicity and in evaluation of the relative risk/benefit ratio of treatment during pregnancy.
Table I. Number of doses of oral azidothymidine and weight-adjusted daily dosage Animal No.
1 2 3 4 5 6 7 8 9 10
11
Material and methods
Twelve near-term pregnant baboons (Papio cyanocephalus) with a mean gestational age of 169.6 ± 9.0 days were identified for this study after approval by the Animal Use and Care Committee and Investigational Review Committee of the respective institutions. Each animal was studied only once to reduce the effects of anesthesia on the dam and her fetus. The animals had received from one to 17 doses of azidothymidine when studied (Table I). Oral zidovudine was given to the animals in a dose of 100 mg every 6 hours; the powder was mixed with peanut butter or jelly and ingestion by the animals was observed. The dose was selected to approximate that reported by Fischl et aJ.2 of 250 mg every 4 hours for a total daily dose of 1500 mg. If the average weight of a pregnant woman is assumed to be 70 kg, this would be equivalent to 21.4 mg/kg/day. Because of variance in the weight of the animals, the actual mean dose was 22.9 ± 2.3 mg/kg/day of azidothymidine (range, 19.9 to 26.5 mg/kg/day) (Table I). After initial sedation with ketamine hydrochloride all animals were intubated and a surgical level of anesthesia was maintained with a combination of nitrous oxide and halothane. Maternal stability was ensured by continuous monitoring of heart rate and rhythm and of arterial blood pressure via a radial arterial catheter. Assessment of the fetal heart rate was by ultrasonographic visualization. The maternal abdomen was sterilely prepped and draped; amniocentesis and percutaneous umbilical cord sampling were performed under ultrasonographic guidance as previously reported. 5 In each instance the fetal umbilical artery was sampled; visual identification of the umbilical artery was verified by performing a blood gas analysis on the specimen obtained. A 3 ml aliquot of fetal blood and 10 ml of maternal blood were obtained simultaneously. In each instance the amniotic fluid (7 ml) was obtained before the percutaneous umbilical cord blood sampling to avoid contamination with either maternal or fetal blood. Both maternal and fetal blood were centrifuged and the serum was separated and labeled. The samples were
12
Weight (kg)
1 1 4 5 8 9 9 12 13 16 17 17
19.6 17.9 15.1 19.4 15.9 19.4 16.5 16.9 15.9 20.1 19.3 16.0 17.67 ± 1.7
729
Dosage (mgl kgl day)
20.4 22.3 26.5 20.6 25.2 20.6 24.2 23.7 25.2 19.9 20.7 25.0 22.9 ± 2.3
stored at - 70° C pending analysis for azidothymidine and 5' -glucuronide azidothymidine levels. Amniotic fluid was frozen without centrifugation; in each instance the fluid was clear and free of contamination with either blood or meconium. All assays for azidothymidine were performed by the Wellcome Research Laboratories by radioimmunoassay with the INCSTAR commercial radioimmunoassay kit. Each specimen was analyzed both before and after incubation with Escherichia coli [3-glucuronidase; the initial value represented the unconjugated azidothymidine level and the value after incubation represented the total (conjugated and unconjugated azidothymidine) level. The 5' -glucuronide azidothymidine concentration was obtained by subtracting the initial azidothymidine concentration from that after incubation with the [3-glucuronidase. 6 Statistical methods. Evaluation of the relative concentrations of azidothymidine in maternal serum, fetal serum, and amniotic fluid were compared with paired analyses. Similar comparisons were also done for 5'glucuronide azidothymidine concentrations. The effect of increasing number of doses of azidothymidine on azidothymidine and 5' -glucuronide azidothymidine concentrations was studied by linear regression analysis of azidothymidine and 5' -glucuronide azidothymidine levels in maternal serum, fetal serum, and amniotic fluid. An alpha error <0.05 was considered statistically significant and corrections for multiple comparisons were used as indicated. Results
The mean azidothymidine concentration in maternal serum (58.6 ± 62.5 ng/ml), fetal serum (78.4 ± 79 ng/ml), and amniotic fluid (254.4 ± 107.8 ng/ml) is shown in Fig. 1; the concentration progressively increased from mother to fetus to amniotic fluid. Similar gradients were observed with 5' -glucuronide azido-
730
Hankins et al.
September 1990 Am J Obstet Gynecol
2500
-
WW'4a
MATERNAL
e
2000
to-
1500
.s
CJI
N
'II(
I
FETAL
I
J
AMNIOTIC
Co:)
c: 0
to-
1000
N
'II(
500 0
AZT
GAZT
Fig. 1. Azidothymidine and 5'-glucuronide azidothymidine concentrations in maternal serum, fetal serum, and amniotic fluid in near-term pregnant baboons after oral administration of azidothymidine. A clear concentration gradient was present with progressively higher levels being attained in maternal serum, fetal serum, and amniotic fluid (p < 0.0001).
thymidine; the maternal serum value was 203.9 ± 189.9 ng/ml, the fetal serum value was 511.8 ± 399.3 ng/ml, and the amniotic fluid value was 2194.1 ± 1208.2 ng/m!. In the evaluation of each individual animal, in only one instance did the maternal azidothymidine level exceed that of the fetus (p = 0.019). That particular animal received eight doses of azidothymidine. Amniotic fluid azidothymidine levels exceeded those in maternal serum in all 12 animals (p < 0.0001) and exceeded those in the fetus in 11 of 12 animals (p < 0.0001). The single exception was an animal that had received only one dose of azidothymidine approximately 3 hours before sampling. In this particular animal the maternal concentration was 144 ng/ml, fetal concentration 163 ng/ml, and maniotic fluid concentration 147 ng/m!. The concentration of azidothymidine in maternal serum, fetal serum, or amniotic fluid showed no apparent relationship to the number of doses received (p = NS), especially within the range of four to 17 doses (Fig. 2). Similar to the findings with azidothymidine, 5'glucuronide azidothymidine levels in the fetus exceeded those in the mother in 11 of 12 animals (p = 0.002), and amniotic fluid concentrations greatly exceeded those in the maternal serum in all cases (p < 0.0001). Amniotic fluid 5'-glucuronide azidothymidine levels exceeded those of the fetus in 11 of 12 animals (p = 0.0004); the one exception was the animal that had received only one dose of azidothymidine approximately 3 hours before sampling. In this animal the maternal 5'-glucuronide azidothymidine level was 410 ng/ml, the fetal level was 852 nglml, and the amniotic fluid level was 651 ng/ml. There was no detectable change in maternal or fetal
circulating concentrations with increasing numbers of doses (p = NS). Significantly, the concentration of 5'glucuronide azidothymidine in amniotic fluid did show a positive correlation with the number of doses of medication the animals had received before sampling (p < 0.01) (Fig. 3). Comment The full impact of AIDS on women and children in the United States remains a future, but certain, event. Currently the reported cases of AIDS in children in the United States are <2% of the number in adults,7 and serologic prevalence of human immunodeficiency virus among childbearing women remains low (2.111000).8 However, this is not universally the case; Ryder et al. 9 reported that 5.8% of pregnant women screened in Zaire were seropositive for human immunodeficiency virus type 1. By 1 year after delivery 21 % of infants of seropositive women were dead compared with 3.8% of control infants whose mothers tested negative. Furthermore, 7.9% of the surviving infants had AIDS by the age of 1 year. Similar results have been reported by the French Collaborative Study GrouplO who found that approximately one third of the infants born to seropositive women had evidence of human immunodeficiency virus type 1 infection or of AIDS by age 18 months and that 20% of this group had died. Treatment of pregnant women with medications potentially toxic to both the woman and her fetus raises several questions: Who is being treated and what are the indications for treatment? Does the fetus possess certain characteristics that may either augment or protect it from drug toxicity? To what extent does the drug
Transplacental transfer of zidovudine and its metabolite
Volume 163 Number 3
731
500 MATERNAL
400
E ....
01
c
FETAL AMNIOTIC
300
"
l-
N
oc{
, I
, :.\ 1 \
200
I : '.1
Ito
I/'~\
I:
100
t·
J!
,. ..
0
, \
I:
',\
1
1
1 I: '.\1: _¥..
.\
'.\
,:
\
'.1
I.:
'.1
r ......\.,.J \
'.1
:1
..... ;,. \;..
~
l
.. •
4
/,
,...,,...........
,I' •••••' ,
• ..............~...
8
5
9
9
12
13
16
..
17
17
Number of Doses of AZT
Fig. 2. Azidothymidine concentrations in maternal serum, fetal serum, and amniotic fluid versus the number of doses of azidothymidine. The concentration of azidothymidine was not altered significantly in any of the three compartments with increasing numbers of doses of azidothymidine.
5000 MATERNAL
4000 E ....01 C
FETAL AMNIOTIC
3000
l-
N
oc{
C)
2000
1000
,/'\ ,.,
1'... /
••••••
.•. ••• 4
I
,"\
,
,
',I ....... ~
v/
0
I
• •••••
" Io • •
5
8
a
•• a
'-t..~".:''''''....,. ""
• • • • :--. . . . . . . . . . . . . . .
9
9
12
13
16
17
17
Number of Doses of AZT
Fig. 3. 5'-Glucuronide azidothymidine concentrations in maternal serum, fetal serum, and amniotic fluid versus the number of doses of azidothymidine. 5'-Glucuronide azidothymidine amniotic fluid concentrations rose with increasing dosages of azidothymidine (p < 0.01), whereas maternal and fetal serum levels remained unchanged.
cross the placenta and can maternal drug levels be used to extrapolate to fetal drug levels? In the case of azidothymidine many, if not all, of these questions remain unanswered. Indeed, to date it has not been established whether fetal infection occurs principally in utero by transplacental routes, as occurs with intrauterine rubella, or intrapartum as with the hepatitis B virus. 7 By use of an animal model with placentation similar
to that of humans, we found that the serum concentrations of both azidothymidine and 5' -glucuronide azidothymidine were higher in the fetus than in the mother, irrespective of the number of doses of oral medication the mothers had ingested. After intravenous dosing with azidothymidine in nonpregnant adults, the compound has glucuronic acid attached in the 5' position in the liver, thereby inactivating a large
732
Hankins et al.
percentage of the drug. Further metabolism involves renal clearance; 60% of the drug is excreted in the urine as 5' -glucuronide azidothymidine and 25% as azidothymidine. 11 The increased fetal! maternal ratio of the drug may result from an impaired ability to conjugate the drug by the fetal liver because it is well established that this particular enzyme system does not mature until well after birth. In addition, renal clearance of azidothymidine in the newborn, and by extrapolation of the fetus still in utero, is markedly depressed in relation to that of adults and does not approach the same rate until several days of life. 3 Alternatively, and perhaps synergistically, the high concentrations of both azidothymidine and its metabolite in the amniotic fluid may serve as a source of oral therapy for the fetus via in utero swallowing and absorption. A role for fetal swallowing in the further metabolism of amniotic fluid azidothymidine is supported by the fact that the amniotic fluid azidothymidine concentration did not increase in proportion to the number of doses, whereas concentration of its metabolite 5'glucuronide azidothymidine did increase (Fig. 3). With prolonged courses of azidothymidine treatment Richman et al. 4 observed that in 24 % of patients anemia developed, defined as a hemoglobin level <7.5 gm/dl, compared with only 4% of patients treated with placebo. Twenty-one percent of azidothymidine recipients and 4% of placebo recipients required multiple red cell transfusions with hematologic toxicity apparently augmented by concurrent use of acetaminophen. Such toxicities may be even more common and severe in the setting of very active bone marrow, such as characterized by both the pregnant woman and her fetus. A parallel between the effects of acetaminophen in the adult and the immature liver of the fetus should be anticipated with respect to more intense hematologic toxicity. Continued usage of these drugs in pregnancy before further detailed animal studies would seem at a minimum to merit intense ultrasonographic surveillance for the potential development of severe fetal anemia and hydrops. In conclusion, on the basis of data obtained from the pregnant baboon, the fetal levels of azidothymidine will equal or exceed those in the mother. It is anticipated
September 1990 Am J Obstet Gynecol
that azidothymidine tOXICIty In pregnancy, for both mother and fetus, will equal (or exceed) that of nonpregnant patients. We concur that "the priority given studies of possible intrauterine, or neonatal interventions, whether pharmacologic or immunologic, to combat human immunodeficiency virus infection should remain exceedingly high. "7 REFERENCES 1. FDA Drug Bull 1987;17:14-24. 2. Fischl MA, Richman DD, Grieco MH, et al. The efficacy of azidothymidine (AZT) in the treatment of patients with AIDS and AIDS-related complex. N Engl J Med 1987; 317: 185-91. 3. Unad kat JD, Lopez AA, Schuman L. Transplacental transfer and the pharmacokinetics of zidovudine (ZDV) in the near term pregnant macaque. In: Program and abstracts of the Twenty-eighth Interscience Conference on Antimicrobial Agents and Chemotherapy, Los Angeles, October 1988. Los Angeles: American Society of Microbiology, 1988:372. 4. Richman DE, Fischl MA, Grieco MH, et al. The toxicity of azidothymidine (AZT) in the treatment of patients with AIDS and AIDS-related complex. N Engl J Med 1987; 317:192-7. 5. Berkowitz RL, Chitkara U, GoldbergJD, Wilkins I, Chervenak FA, Lynch L. Intrauterine intravascular transfusions for severe red blood cell isoimmunization: ultrasound-guided percutaneous approach. AM J OBSTET GyNECOL 1986;155:574-81. 6. TadepaJIi S, Quinn R, Puckett LD, Orf JW, Goldensoph CR, Donner JE. Adaptation of ZDV-TRAC for simultaneous monitoring of AZT and GAZT. In: Proceedings of the annual meeting of the American Association of Clinical Chemistry, Atlanta, July 1989. Atlanta: American Association of Clinical Chemistry, 1989:40. 7. Katz SL, Wilfert CM. Human immunodeficiency virus infection of newborns. N Engl J Med 1989;320:1687-8. 8. Hoff R, Berardi VP, Weiblen BJ, Mahoney-Trout L, Mitchell ML, Grady GF. Seroprevalence of human immunodeficiency virus among childbearing women. N Engl J Med 1988;318:525-30. 9. Ryder RR, Nsa W, Hassig SE, et al. perinatal transmission of the human immunodeficiency virus type I to infants of seropositive women in Zaire. N Engl J Med 1989; 320: 1637-42. 10. Blanche S, Rouzioux C, Moscato MLG, et al. A prospective study of infants born to women seropositive for human immunodeficiency virus type I. N Engl J Med 1989; 320:1643-8. 11. Blum MR, Liao SHT, Good SS, DeMiranda P. Pharmacokinetics and bioavailability of zidovudine in humans. Am J Med 1988;85(suppl 2A): 189-94.
Key words: Azidothymidine, acquired immunodeficiency syndrome, placental transfer Few disease states have garnered national and international attention and action as quickly as the acquired immunodeficiency syndrome (AIDS). Demand for responsiveness by the scientific and medical community has been intense and action has been prompt. The Food and Drug Administration approved clinical testing of zidovudine (azidothymidine) within 5 days of receipt of its sponsor's application. The review and approval of the zidovudine New Drug Application required only 107 days and clinical studies of only 22 months. 1 Optimism over the demonstration that azidothymidine adFrom the Wilford Hall United States Air Force Medical Center, Lackland Air Force Base: Southwest Foundation for Biomedical Research, San Antonio,' and the University of Alabama at Birmingham.' Presented at the Tenth Annual Meeting of the Society of Perinatal Obstetricians, Houston, Texas, January 23-27, 1990. The opinions expressed in this manuscript are those of the authors and not necessarily those of the United States Air Force or the Department of Defense. Reprint requests: Cary D. V. Hankins, MD, Wilford Hall USAF Medical CenteriSCHO, Lackland AFB, TX 78236-5300. 616122358
728
ministration could decrease mortality and the frequency of opportunistic infections in a selected group of subjects with AIDS or AIDS-related complex 2 has resulted in ongoing phase I clinical trials of zidovudine in pregnancy. Notwithstanding, published data on azidothymidine in pregnancy, especially in the fetus, are limited to a single study involving one pregnant macaque monkey.3 In an acute preparation involving intravenous administration of zidovudine (44.1 J.Lg/kg/min) for 3Y2 hours before cesarean birth, azidothymidine was demonstrated to cross the placenta with a ratio of the steady-state concentration of the drug in fetus/ mother of 0.74. To more closely approximate the usual therapy used in actual clinical practice, we administered the drug per os to 12 pregnant baboons, in the third trimester of pregnancy. Because the drug is metabolized in the liver by attachment of glucuronic acid in the 5' position with subsequent renal excretion of both the free and conjugated drug, the ability of the fetus to metabolize the drug might be impaired by delayed maturation of the glucuronide enzyme system.
Volume 163 Number 3
Transplacental transfer of zidovudine and its metabolite
Indeed, in nonpregnant adults with concurrent acetaminophen and azidothymidine treatment, a higher frequency of hematologic complications have been reported and attributed to increased azidothymidine toxicity as a result of competitive inhibition of hepatic metabolism by the glucuronide system" Accordingly, information on the degree to which azidothymidine and its metabolite 5'-glucuronide azidothymidine cross the placenta and the levels achieved in the fetus and the amniotic fluid with prolonged long-term therapy is of great importance in assessment of its toxicity and in evaluation of the relative risk/benefit ratio of treatment during pregnancy.
Table I. Number of doses of oral azidothymidine and weight-adjusted daily dosage Animal No.
1 2 3 4 5 6 7 8 9 10
11
Material and methods
Twelve near-term pregnant baboons (Papio cyanocephalus) with a mean gestational age of 169.6 ± 9.0 days were identified for this study after approval by the Animal Use and Care Committee and Investigational Review Committee of the respective institutions. Each animal was studied only once to reduce the effects of anesthesia on the dam and her fetus. The animals had received from one to 17 doses of azidothymidine when studied (Table I). Oral zidovudine was given to the animals in a dose of 100 mg every 6 hours; the powder was mixed with peanut butter or jelly and ingestion by the animals was observed. The dose was selected to approximate that reported by Fischl et aJ.2 of 250 mg every 4 hours for a total daily dose of 1500 mg. If the average weight of a pregnant woman is assumed to be 70 kg, this would be equivalent to 21.4 mg/kg/day. Because of variance in the weight of the animals, the actual mean dose was 22.9 ± 2.3 mg/kg/day of azidothymidine (range, 19.9 to 26.5 mg/kg/day) (Table I). After initial sedation with ketamine hydrochloride all animals were intubated and a surgical level of anesthesia was maintained with a combination of nitrous oxide and halothane. Maternal stability was ensured by continuous monitoring of heart rate and rhythm and of arterial blood pressure via a radial arterial catheter. Assessment of the fetal heart rate was by ultrasonographic visualization. The maternal abdomen was sterilely prepped and draped; amniocentesis and percutaneous umbilical cord sampling were performed under ultrasonographic guidance as previously reported. 5 In each instance the fetal umbilical artery was sampled; visual identification of the umbilical artery was verified by performing a blood gas analysis on the specimen obtained. A 3 ml aliquot of fetal blood and 10 ml of maternal blood were obtained simultaneously. In each instance the amniotic fluid (7 ml) was obtained before the percutaneous umbilical cord blood sampling to avoid contamination with either maternal or fetal blood. Both maternal and fetal blood were centrifuged and the serum was separated and labeled. The samples were
12
Weight (kg)
1 1 4 5 8 9 9 12 13 16 17 17
19.6 17.9 15.1 19.4 15.9 19.4 16.5 16.9 15.9 20.1 19.3 16.0 17.67 ± 1.7
729
Dosage (mgl kgl day)
20.4 22.3 26.5 20.6 25.2 20.6 24.2 23.7 25.2 19.9 20.7 25.0 22.9 ± 2.3
stored at - 70° C pending analysis for azidothymidine and 5' -glucuronide azidothymidine levels. Amniotic fluid was frozen without centrifugation; in each instance the fluid was clear and free of contamination with either blood or meconium. All assays for azidothymidine were performed by the Wellcome Research Laboratories by radioimmunoassay with the INCSTAR commercial radioimmunoassay kit. Each specimen was analyzed both before and after incubation with Escherichia coli [3-glucuronidase; the initial value represented the unconjugated azidothymidine level and the value after incubation represented the total (conjugated and unconjugated azidothymidine) level. The 5' -glucuronide azidothymidine concentration was obtained by subtracting the initial azidothymidine concentration from that after incubation with the [3-glucuronidase. 6 Statistical methods. Evaluation of the relative concentrations of azidothymidine in maternal serum, fetal serum, and amniotic fluid were compared with paired analyses. Similar comparisons were also done for 5'glucuronide azidothymidine concentrations. The effect of increasing number of doses of azidothymidine on azidothymidine and 5' -glucuronide azidothymidine concentrations was studied by linear regression analysis of azidothymidine and 5' -glucuronide azidothymidine levels in maternal serum, fetal serum, and amniotic fluid. An alpha error <0.05 was considered statistically significant and corrections for multiple comparisons were used as indicated. Results
The mean azidothymidine concentration in maternal serum (58.6 ± 62.5 ng/ml), fetal serum (78.4 ± 79 ng/ml), and amniotic fluid (254.4 ± 107.8 ng/ml) is shown in Fig. 1; the concentration progressively increased from mother to fetus to amniotic fluid. Similar gradients were observed with 5' -glucuronide azido-
730
Hankins et al.
September 1990 Am J Obstet Gynecol
2500
-
WW'4a
MATERNAL
e
2000
to-
1500
.s
CJI
N
'II(
I
FETAL
I
J
AMNIOTIC
Co:)
c: 0
to-
1000
N
'II(
500 0
AZT
GAZT
Fig. 1. Azidothymidine and 5'-glucuronide azidothymidine concentrations in maternal serum, fetal serum, and amniotic fluid in near-term pregnant baboons after oral administration of azidothymidine. A clear concentration gradient was present with progressively higher levels being attained in maternal serum, fetal serum, and amniotic fluid (p < 0.0001).
thymidine; the maternal serum value was 203.9 ± 189.9 ng/ml, the fetal serum value was 511.8 ± 399.3 ng/ml, and the amniotic fluid value was 2194.1 ± 1208.2 ng/m!. In the evaluation of each individual animal, in only one instance did the maternal azidothymidine level exceed that of the fetus (p = 0.019). That particular animal received eight doses of azidothymidine. Amniotic fluid azidothymidine levels exceeded those in maternal serum in all 12 animals (p < 0.0001) and exceeded those in the fetus in 11 of 12 animals (p < 0.0001). The single exception was an animal that had received only one dose of azidothymidine approximately 3 hours before sampling. In this particular animal the maternal concentration was 144 ng/ml, fetal concentration 163 ng/ml, and maniotic fluid concentration 147 ng/m!. The concentration of azidothymidine in maternal serum, fetal serum, or amniotic fluid showed no apparent relationship to the number of doses received (p = NS), especially within the range of four to 17 doses (Fig. 2). Similar to the findings with azidothymidine, 5'glucuronide azidothymidine levels in the fetus exceeded those in the mother in 11 of 12 animals (p = 0.002), and amniotic fluid concentrations greatly exceeded those in the maternal serum in all cases (p < 0.0001). Amniotic fluid 5'-glucuronide azidothymidine levels exceeded those of the fetus in 11 of 12 animals (p = 0.0004); the one exception was the animal that had received only one dose of azidothymidine approximately 3 hours before sampling. In this animal the maternal 5'-glucuronide azidothymidine level was 410 ng/ml, the fetal level was 852 nglml, and the amniotic fluid level was 651 ng/ml. There was no detectable change in maternal or fetal
circulating concentrations with increasing numbers of doses (p = NS). Significantly, the concentration of 5'glucuronide azidothymidine in amniotic fluid did show a positive correlation with the number of doses of medication the animals had received before sampling (p < 0.01) (Fig. 3). Comment The full impact of AIDS on women and children in the United States remains a future, but certain, event. Currently the reported cases of AIDS in children in the United States are <2% of the number in adults,7 and serologic prevalence of human immunodeficiency virus among childbearing women remains low (2.111000).8 However, this is not universally the case; Ryder et al. 9 reported that 5.8% of pregnant women screened in Zaire were seropositive for human immunodeficiency virus type 1. By 1 year after delivery 21 % of infants of seropositive women were dead compared with 3.8% of control infants whose mothers tested negative. Furthermore, 7.9% of the surviving infants had AIDS by the age of 1 year. Similar results have been reported by the French Collaborative Study GrouplO who found that approximately one third of the infants born to seropositive women had evidence of human immunodeficiency virus type 1 infection or of AIDS by age 18 months and that 20% of this group had died. Treatment of pregnant women with medications potentially toxic to both the woman and her fetus raises several questions: Who is being treated and what are the indications for treatment? Does the fetus possess certain characteristics that may either augment or protect it from drug toxicity? To what extent does the drug
Transplacental transfer of zidovudine and its metabolite
Volume 163 Number 3
731
500 MATERNAL
400
E ....
01
c
FETAL AMNIOTIC
300
"
l-
N
oc{
, I
, :.\ 1 \
200
I : '.1
Ito
I/'~\
I:
100
t·
J!
,. ..
0
, \
I:
',\
1
1
1 I: '.\1: _¥..
.\
'.\
,:
\
'.1
I.:
'.1
r ......\.,.J \
'.1
:1
..... ;,. \;..
~
l
.. •
4
/,
,...,,...........
,I' •••••' ,
• ..............~...
8
5
9
9
12
13
16
..
17
17
Number of Doses of AZT
Fig. 2. Azidothymidine concentrations in maternal serum, fetal serum, and amniotic fluid versus the number of doses of azidothymidine. The concentration of azidothymidine was not altered significantly in any of the three compartments with increasing numbers of doses of azidothymidine.
5000 MATERNAL
4000 E ....01 C
FETAL AMNIOTIC
3000
l-
N
oc{
C)
2000
1000
,/'\ ,.,
1'... /
••••••
.•. ••• 4
I
,"\
,
,
',I ....... ~
v/
0
I
• •••••
" Io • •
5
8
a
•• a
'-t..~".:''''''....,. ""
• • • • :--. . . . . . . . . . . . . . .
9
9
12
13
16
17
17
Number of Doses of AZT
Fig. 3. 5'-Glucuronide azidothymidine concentrations in maternal serum, fetal serum, and amniotic fluid versus the number of doses of azidothymidine. 5'-Glucuronide azidothymidine amniotic fluid concentrations rose with increasing dosages of azidothymidine (p < 0.01), whereas maternal and fetal serum levels remained unchanged.
cross the placenta and can maternal drug levels be used to extrapolate to fetal drug levels? In the case of azidothymidine many, if not all, of these questions remain unanswered. Indeed, to date it has not been established whether fetal infection occurs principally in utero by transplacental routes, as occurs with intrauterine rubella, or intrapartum as with the hepatitis B virus. 7 By use of an animal model with placentation similar
to that of humans, we found that the serum concentrations of both azidothymidine and 5' -glucuronide azidothymidine were higher in the fetus than in the mother, irrespective of the number of doses of oral medication the mothers had ingested. After intravenous dosing with azidothymidine in nonpregnant adults, the compound has glucuronic acid attached in the 5' position in the liver, thereby inactivating a large
732
Hankins et al.
percentage of the drug. Further metabolism involves renal clearance; 60% of the drug is excreted in the urine as 5' -glucuronide azidothymidine and 25% as azidothymidine. 11 The increased fetal! maternal ratio of the drug may result from an impaired ability to conjugate the drug by the fetal liver because it is well established that this particular enzyme system does not mature until well after birth. In addition, renal clearance of azidothymidine in the newborn, and by extrapolation of the fetus still in utero, is markedly depressed in relation to that of adults and does not approach the same rate until several days of life. 3 Alternatively, and perhaps synergistically, the high concentrations of both azidothymidine and its metabolite in the amniotic fluid may serve as a source of oral therapy for the fetus via in utero swallowing and absorption. A role for fetal swallowing in the further metabolism of amniotic fluid azidothymidine is supported by the fact that the amniotic fluid azidothymidine concentration did not increase in proportion to the number of doses, whereas concentration of its metabolite 5'glucuronide azidothymidine did increase (Fig. 3). With prolonged courses of azidothymidine treatment Richman et al. 4 observed that in 24 % of patients anemia developed, defined as a hemoglobin level <7.5 gm/dl, compared with only 4% of patients treated with placebo. Twenty-one percent of azidothymidine recipients and 4% of placebo recipients required multiple red cell transfusions with hematologic toxicity apparently augmented by concurrent use of acetaminophen. Such toxicities may be even more common and severe in the setting of very active bone marrow, such as characterized by both the pregnant woman and her fetus. A parallel between the effects of acetaminophen in the adult and the immature liver of the fetus should be anticipated with respect to more intense hematologic toxicity. Continued usage of these drugs in pregnancy before further detailed animal studies would seem at a minimum to merit intense ultrasonographic surveillance for the potential development of severe fetal anemia and hydrops. In conclusion, on the basis of data obtained from the pregnant baboon, the fetal levels of azidothymidine will equal or exceed those in the mother. It is anticipated
September 1990 Am J Obstet Gynecol
that azidothymidine tOXICIty In pregnancy, for both mother and fetus, will equal (or exceed) that of nonpregnant patients. We concur that "the priority given studies of possible intrauterine, or neonatal interventions, whether pharmacologic or immunologic, to combat human immunodeficiency virus infection should remain exceedingly high. "7 REFERENCES 1. FDA Drug Bull 1987;17:14-24. 2. Fischl MA, Richman DD, Grieco MH, et al. The efficacy of azidothymidine (AZT) in the treatment of patients with AIDS and AIDS-related complex. N Engl J Med 1987; 317: 185-91. 3. Unad kat JD, Lopez AA, Schuman L. Transplacental transfer and the pharmacokinetics of zidovudine (ZDV) in the near term pregnant macaque. In: Program and abstracts of the Twenty-eighth Interscience Conference on Antimicrobial Agents and Chemotherapy, Los Angeles, October 1988. Los Angeles: American Society of Microbiology, 1988:372. 4. Richman DE, Fischl MA, Grieco MH, et al. The toxicity of azidothymidine (AZT) in the treatment of patients with AIDS and AIDS-related complex. N Engl J Med 1987; 317:192-7. 5. Berkowitz RL, Chitkara U, GoldbergJD, Wilkins I, Chervenak FA, Lynch L. Intrauterine intravascular transfusions for severe red blood cell isoimmunization: ultrasound-guided percutaneous approach. AM J OBSTET GyNECOL 1986;155:574-81. 6. TadepaJIi S, Quinn R, Puckett LD, Orf JW, Goldensoph CR, Donner JE. Adaptation of ZDV-TRAC for simultaneous monitoring of AZT and GAZT. In: Proceedings of the annual meeting of the American Association of Clinical Chemistry, Atlanta, July 1989. Atlanta: American Association of Clinical Chemistry, 1989:40. 7. Katz SL, Wilfert CM. Human immunodeficiency virus infection of newborns. N Engl J Med 1989;320:1687-8. 8. Hoff R, Berardi VP, Weiblen BJ, Mahoney-Trout L, Mitchell ML, Grady GF. Seroprevalence of human immunodeficiency virus among childbearing women. N Engl J Med 1988;318:525-30. 9. Ryder RR, Nsa W, Hassig SE, et al. perinatal transmission of the human immunodeficiency virus type I to infants of seropositive women in Zaire. N Engl J Med 1989; 320: 1637-42. 10. Blanche S, Rouzioux C, Moscato MLG, et al. A prospective study of infants born to women seropositive for human immunodeficiency virus type I. N Engl J Med 1989; 320:1643-8. 11. Blum MR, Liao SHT, Good SS, DeMiranda P. Pharmacokinetics and bioavailability of zidovudine in humans. Am J Med 1988;85(suppl 2A): 189-94.