- Email: [email protected]
Recent indomethacin tocolysis is not associated with neonatal complications in preterm infants
Recent indomethacin tocolysis is not associated with neonatal complications in preterm infants Stephen T. Vermillion, MD, and Roger B. Newman, MD Charleston, South Carolina OBJECTIVE: We sought to determine whether indomethacin tocolysis immediately before delivery is associated with any increased complications in neonates delivered between 24 and 32 weeks’ gestation. STUDY DESIGN: We performed a case-control analysis of neonates delivered between 24 and 32 weeks’ gestation after maternal indomethacin treatment for preterm labor. All infants were delivered within 48 hours of indomethacin exposure. Seventy-five patients were matched with 150 control subjects in a patient/control ratio of 1:2. Matching variables in order of priority included gestational age at delivery, fetal number, betamethasone exposure >24 hours before delivery, magnesium sulfate use, mode of delivery, infant sex, and race. Data were analyzed by using the Student t test, χ2 analysis, and the Fisher exact test and Yates’ correction. RESULTS: The gestational age at delivery (mean ± SD) was 28.7 ± 2.3 weeks in the indomethacin group and 28.3 ± 2.1 weeks for the control subjects. Birth weights (mean ± SD) were 1121 ± 243 and 1141 ± 287 g, respectively. All mothers received both magnesium sulfate and betamethasone before delivery. The median cumulative dose of indomethacin was 225 mg. The median interval from last dose of indomethacin until delivery was 12 hours. There were no significant differences between the groups in the incidence of necrotizing enterocolitis (odds ratio, 1.12; 95% confidence interval, 0.31-3.84), grade III/IV intraventricular hemorrhage (odds ratio, 0.37; 95% confidence interval, 0.10-1.19), patent ductus arteriosus (odds ratio, 0.85; 95% confidence interval, 0.44-1.64), bronchopulmonary dysplasia (odds ratio, 0.97; 95% confidence interval, 0.49-1.91), pulmonary hypertension (odds ratio, 0.49; 95% confidence interval, 0.02-4.80), anuria (odds ration, 1.21; 95% confidence interval, 0.22-6.01), thrombocytopenia (odds ratio, 1.14; 95% confidence interval, 0.53-2.42), sepsis (odds ratio, 1.21; 95% confidence interval, 0.22-6.01), or neonatal death (odds ratio, 1.34; 95% confidence interval, 0.55-3.25). CONCLUSION: Maternal indomethacin exposure immediately before delivery was not associated with increased neonatal complications for infants delivered between 24 and 32 weeks’ gestation. (Am J Obstet Gynecol 1999;181:1083-6.)
Key words: Indomethacin, tocolysis, neonatal complications
Indomethacin, a potent prostaglandin synthetase inhibitor, has been demonstrated to be a successful tocolytic agent in several prospective trials.1-4 Recent studies have raised concern over the possibility of increased neonatal complications after antenatal indomethacin exposure,5-8 whereas other investigations have been unable to demonstrate such risks.9-12 Subanalyses from 2 of these studies demonstrated an increased risk of neonatal complications only if delivery occurred within 48 hours of indomethacin exposure.7, 8 These studies, however, in-
From the Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology, Medical University of South Carolina. Presented at the Nineteenth Annual Meeting of the Society for MaternalFetal Medicine, San Francisco, California, January 18-23, 1999. Reprint requests: Stephen T. Vermillion, MD, Department of Obstetrics and Gynecology, Medical University of South Carolina, 96 Jonathan Lucas St, Suite 634, Charleston, SC 29425. Copyright © 1999 by Mosby, Inc. 0002-9378/99 $8.00 + 0 6/6/100993
cluded patients receiving indomethacin for maintenance tocolysis (>72 hours). Prolonged therapy with indomethacin has been demonstrated to increase the risk of premature closure of the ductus arteriosus.13, 14 Shortterm tocolysis does not appear to engender this same risk because of the demonstrated ability of the fetal ductus arteriosus to promptly reverse restrictive flow patterns after short-term exposure.13, 14 The purpose of this study was to determine whether maternal exposure to a short course (≤72 hours) of indomethacin tocolysis within 48 hours of delivery is associated with any increased neonatal complications for infants delivered between 24 and 32 weeks’ gestation. Methods We performed a case-control analysis of neonates delivered between 24 and 32 weeks’ gestation after maternal tocolysis with indomethacin. Data were collected and entered into the Medical University of South Carolina 1083
1084 Vermillion and Newman
November 1999 Am J Obstet Gynecol
Table I. Demographic characteristics of indomethacin-treated and control groups Variable
Indomethacin group (n = 75)
Control group (n = 150)
Statistical significance
Weight (g) Gestational age (wk) Multiple gestations Cesarean delivery Female infant Race Black White Other
1121 ± 243 28.7 ± 2.3 16 (21.3%) 27 (36.0%) 40 (53.3%)
1141 ± 287 28.3 ± 2.1 30 (20.6%) 58 (38.6%) 91 (60.6%)
P = .71 P = .63 P = .81 P = .69 P = .36
53 (70.6%) 19 (25.3%) 3 (4.1%)
118 (78.6%) 29 (19.3%) 3 (2.1%)
P = .24 P = .38 P = .66
Table II. Neonatal outcomes of indomethacin-exposed and control groups Indomethacin group (n = 75) Outcome Necrotizing enterocolitis Patent ductus arteriosus Ligated Intraventricular hemorrhage, grade I/II Intraventricular hemorrhage, grade III/IV Bronchopulmonary dysplasia Anuria Pulmonary hypertension Thrombocytopenia Sepsis Death
Control group (n = 150)
No.
%
No.
%
Odds ratio and 95% confidence interval
Statistical significance
5 21 8 5
6.7 28.0 1.1 6.7
9 47 20 6
6.0 31.3 1.3 4.0
1.12 (0.31-3.84) 0.85 (0.44-1.64) 0.78 (0.30-1.98) 1.71 (0.44-6.63)
P = .92 P = .72 P = .57 P = .51
4
5.3
20
13.3
0.37 (0.10-1.19)
P = .07
19 3 1 15 3 11
25.3 4.0 1.3 20.0 4.0 14.6
39 5 4 27 5 17
26 3.3 2.7 18.0 3.3 11.3
0.97 (0.49-1.91) 1.21 (0.22-6.01) 0.49 (0.02-4.80) 1.14 (0.53-2.42) 1.21 (0.22-6.01) 1.34 (0.55-3.25)
P = .96 P = .90 P = .67 P = .86 P = .90 P = .62
perinatal database between January 1992 and July 1997. Only those neonates exposed to 24 to 72 hours of antenatal indomethacin for tocolysis and delivered within 48 hours of the last dose were included in the analysis. Adjunctive tocolysis with indomethacin at the study institution is limited to pregnancies at gestational ages of <32 weeks with persistent uterine activity despite aggressive parenteral therapy with magnesium sulfate. A total of 75 cases were identified, and each case was matched with 2 control subjects for a total of 150 control patients. Matching variables in order of priority included gestational age at delivery, fetal number, exposure to betamethasone for >24 hours before delivery, magnesium sulfate exposure, mode of delivery, infant sex, and race. The estimated gestational age on admission was confirmed by antepartum ultrasonographic examination. All patients received antibiotic prophylaxis for group B Streptococcus and a single course of betamethasone consisting of 12 mg on admission, which was repeated 24 hours later. Patients with ruptured membranes on admission, preeclampsia, insulin-requiring diabetes, or an anomalous fetus were excluded from the study. Neonatal outcomes analyzed included the following: birth weight, confirmed necrotizing enterocolitis, patent
ductus arteriosus, patent ductus arteriosus requiring surgical ligation, grade I/II and grade III/IV intraventricular hemorrhage, bronchopulmonary dysplasia, anuria, thrombocytopenia, pulmonary hypertension, sepsis, and death. Indomethacin exposure characteristics, including length of exposure, cumulative dose, and interval from last dose until delivery, were also determined. Only cases of confirmed necrotizing enterocolitis were included. Confirmed necrotizing enterocolitis required the presence of either bowel perforation or intramural intestinal air on abdominal radiograph. Patent ductus arteriosus was diagnosed after confirmatory echocardiography. Neonates with a patent ductus arteriosus requiring surgical closure had all failed medical treatment with postnatal indomethacin. All infants underwent a cranial ultrasonographic examination within 7 days of delivery and weekly thereafter as necessary to detect intraventricular hemorrhage. Intraventricular hemorrhage was graded by using the criteria of Papile et al.15 Bronchopulmonary dysplasia was based on clinical suspicion in the context of defined radiographic features. Those infants without urine production within the first 12 hours of life were considered to be anuric. Thrombocytopenia was defined as a platelet count of less than
Volume 181, Number 5, Part 1 Am J Obstet Gynecol
100 × 109/L. Persistent pulmonary hypertension of the newborn was defined by previously described echocardiographic criteria.16 Sepsis was diagnosed only when the neonate had positive blood or cerebrospinal fluid cultures. Data were analyzed by using the Student t test for continuous variables and χ2 analysis and the Fisher exact test and Yates’ correction for differences among proportions. All values of P < .05 were considered significant. Results As designed, the groups were not statistically different with respect to mean gestational age at delivery, birth weight, percentage of multiple gestations, rate of cesarean delivery, infant sex, or race (Table I). All neonates received at least 24 hours of magnesium sulfate and 2 doses of betamethasone before delivery. Patients exposed to indomethacin for tocolysis received a median cumulative dose of 225 mg (range, 150350 mg). The median interval from last dose until delivery was 12 hours (range, 2-48 hours). All treated patients received a 50-mg loading dose of indomethacin administered rectally followed by 25 mg every 6 to 8 hours administered orally. No patient received >72 hours of indomethacin therapy. There were no significant differences in the incidences of necrotizing enterocolitis, patent ductus arteriosus, patent ductus arteriosus requiring surgical ligation, grade I/II intraventricular hemorrhage, bronchopulmonary dysplasia, anuria, thrombocytopenia, pulmonary hypertension, sepsis, or death between the 2 groups (Table II). Although not statistically significant (P = .07), the incidence of grade III/IV intraventricular hemorrhage was increased in the control group (13.3%) compared with the group exposed to antenatal indomethacin (5.3%). Comment In our study maternal indomethacin exposure within 48 hours of delivery was not associated with any detectable increase in neonatal complications among infants delivered between 24 and 32 weeks’ gestation. The decision to use indomethacin adjunctively with magnesium sulfate was not randomly assigned but rather was dependent on the attending physician, and thus case selection may be subject to potential biases. To minimize the potential biases associated with nonrandomization, the 2 groups were carefully matched by using a 2:1 ratio of control subjects to patients for multiple factors known to affect neonatal outcome. Limitations of the previously cited nonrandomized studies include smaller sample sizes, incorporation of fewer matching characteristics, and less specific inclusion criteria than used in our analysis.5-7, 12 A post hoc analysis demonstrated that the sample size in this study had a
Vermillion and Newman 1085
power of >0.95 to detect a difference in the incidences of confirmed necrotizing enterocolitis (19% vs 2%) or grade III/IV intraventricular hemorrhage (28% vs 3%) of the degree previously reported by Norton et al5 and Souter et al,7 respectively. We recognize that there are limitations in estimating the sufficiency of a sample size from power calculations performed by using published outcomes with large proportional differences. Therefore a revised power analysis was performed that reduced the treatment effect by two thirds in both the Norton study (7% vs 2% for necrotizing enterocolitis) and the Souter study (9% vs 3% for grade III/IV intraventricular hemorrhage), respectively. By using these effect sizes, our sample was still capable of generating a power of >0.80 to detect a significant difference. Another strength of this study is that we included only those patients exposed to a short course of indomethacin (≤72 hours) immediately preceding delivery between 24 and 32 weeks’ gestation. This design selects the highestrisk infants for neonatal complications and acutely exposes them to indomethacin immediately before delivery. Recently, Souter et al7 demonstrated that infants exposed to indomethacin within 48 hours of delivery had an increase in the incidence of all grades of intraventricular hemorrhage, patent ductus arteriosus, and failure of the patent ductus arteriosus to close with postnatal indomethacin therapy. Infants that were delivered >48 hours from the last maternal exposure did not demonstrate any significant differences in outcomes. Their study differed from ours in that it had a smaller sample size and less stringent inclusion and exclusion criteria. Specifically, Souter et al included patients with ruptured membranes and those receiving longer-term (>72 hours) tocolysis with indomethacin. We specifically addressed those issues by limiting our inclusion criteria to those patients delivered of their neonates within 48 hours of the last indomethacin exposure and by matching the groups for a larger number of potentially significant demographic and treatment variables. The study by Norton et al5 suggested an increased risk of intraventricular hemorrhage, necrotizing enterocolitis, and persistent patent ductus arteriosus. Again, similar methodologic concerns exist, primarily because of the smaller sample size, fewer matching variables, and a variable relationship between indomethacin exposure and delivery. In addition, the increased risk of intraventricular hemorrhage demonstrated in that study was the result of the inclusion of grade II intraventricular hemorrhage along with the more clinically significant grade III/IV hemorrhage. Only grade II intraventricular hemorrhage was significantly associated with indomethacin exposure in the study of Norton et al,5 whereas the incidences of grade III/IV were similar between exposed and control groups. Although not statistically significant, the incidence of
1086 Vermillion and Newman
advanced grades of intraventricular hemorrhage in our study were decreased in the indomethacin-exposed group. The protective effects of low-dose indomethacin administered during the first 72 hours of life to very-lowbirth-weight neonates for the prevention of intraventricular hemorrhage have been demonstrated.17, 18 One of the proposed mechanisms of this protective effect is the ability of indomethacin to increase cerebral vascular resistance and stabilize cerebral hemodynamics in the premature infant.19 It is possible that the limited doses of indomethacin administered shortly before delivery may be providing similar prophylactic effects to those demonstrated with neonatal administration. Indomethacin has achieved increasing popularity as a tocolytic agent with efficacy that is comparable to that of both magnesium sulfate and β-adrenergic receptor agents. In these clinical trials the neonatal outcomes were excellent, but those samples were composed of larger, near-term infants for whom tocolytic therapy had been successful. Concerns have been raised as to the potential neonatal effects of acute exposure to indomethacin immediately before delivery, especially when the neonate is of <32 weeks’ gestation. We conclude that the short-term use of indomethacin for tocolysis at <32 weeks’ gestation was not associated with any increased risk of neonatal morbidity for those patients who were subsequently delivered within 48 hours of the last indomethacin dose. Additionally, antenatal indomethacin exposure in our patients was associated with a reduced risk of advanced grades of neonatal intraventricular hemorrhage, although further studies are needed to explore this relationship. REFERENCES
1. Niebyl JR, Blake DA, White RD, Kumor KM, Dubin NH, Robinson JC, et al. The inhibition of premature labor with indomethacin. Am J Obstet Gynecol 1980;136:1014-9. 2. Besinger RE, Niebyl JR, Keyes WG, Johnson TRB. Randomized comparative trial of indomethacin and ritodrine for the longterm treatment of preterm labor. Am J Obstet Gynecol 1991;164:981-8. 3. Morales WJ, Madhav H. Efficacy and safety of indomethacin compared with magnesium sulfate in the management of preterm labor: a randomized study. Am J Obstet Gynecol 1993;169:97-102.
November 1999 Am J Obstet Gynecol
4. Bivens HA, Newman RB, Fyfe DA, Campbell BA, Stramm SL. Randomized trial of oral indomethacin and terbutaline sulfate for long-term suppression of preterm labor. Am J Obstet Gynecol 1993;169:1065-70. 5. Norton ME, Merrill J, Cooper BAB, Kuller JA, Clyman R. Neonatal complications after the administration of indomethacin for preterm labor. N Engl J Med 1993;329:1602-7. 6. Major CA, Lewis DF, Harding JA, Porto MA, Garite TJ. Tocolysis with indomethacin increases the incidence of necrotizing enterocolitis in the low-birth-weight neonate. Am J Obstet Gynecol 1994;170 (1 Pt 1):102-6. 7. Souter D, Harding J, McCowan L, O’Donnell C, McLeay E, Baxendale H. Antenatal indomethacin—adverse fetal effects confirmed. Aust N Z J Obstet Gynaecol 1998;38:11-6. 8. Marpeau L, Bouillie J, Barrat J, Milliez J. Obstetrical advantages and perinatal risks of indomethacin: a report of 818 cases. Fetal Diagn Ther 1994;9:110-5. 9. Dudley DKL, Hardie MJ. Fetal and neonatal effects of indomethacin used as tocolytic agent. Am J Obstet Gynecol 1985;151:181-4. 10. Niebyl JR, Witter FR. Neonatal outcome after indomethacin treatment of preterm labor. Am J Obstet Gynecol 1986;155:747-9. 11. Wurtzel D. Prenatal administration of indomethacin as a tocolytic agent: effect on neonatal renal function. Obstet Gynecol 1990;76:689-92. 12. Gardner MO, Owen J, Skelly S, Hauth JC. Preterm delivery after indomethacin—a risk factor for neonatal complications? J Reprod Med 1996;41:903-6. 13. Vermillion ST, Scardo JA, Lashus AG, Wiles HB. The effect of indomethacin tocolysis on fetal ductus arteriosus constriction with advancing gestational age. Am J Obstet Gynecol 1997;177:25661. 14. Moise KJ. Effect of advancing gestational age on the frequency of fetal ductal constriction in the association with maternal indomethacin use. Am J Obstet Gynecol 1993;168:1350-3. 15. Papile LA, Burstein J, Burstein R, Koffler H. Incidence and evolution of subependymal and intraventricular hemorrhage: a study of infants with birth weights less than 1,500 gm. J Pediatr 1978;92:529-34. 16. Morin SC. Persistent pulmonary hypertension of the newborn. In: Enmanouilides GC, Allen HD, Riemenschneider TA, Gupgesell HP, editors. Moss and Adams heart disease of the infant, child, and adolescent. 5th ed. Baltimore: Williams & Wilkins; 1995. p. 599-614. 17. Ment LR, Duncan CC, Ehrenkranz RA, Kleinman CS, Taylor KJ, Scott DT, et al. Randomized low-dose indomethacin trial for prevention of intraventricular hemorrhage in very low birth weight neonates. J Pediatr 1988;112:948-55. 18. Bandstra ES, Montalvo BM, Goldberg RN, Pacheco I, Ferrer PL, Flynn J, et al. Prophylactic indomethacin for prevention of intraventricular hemorrhage in premature infants. Pediatrics 1988;82:533-42. 19. Yanowitz TD, Yao AC, Werner JC, Pettigrew KD, Oh W, Stonestreet BS. Effects of prophylactic low-dose indomethacin on hemodynamics of very low birthweight infants. J Pediatr 1998;132:28-34.
Key words: Indomethacin, tocolysis, neonatal complications
Indomethacin, a potent prostaglandin synthetase inhibitor, has been demonstrated to be a successful tocolytic agent in several prospective trials.1-4 Recent studies have raised concern over the possibility of increased neonatal complications after antenatal indomethacin exposure,5-8 whereas other investigations have been unable to demonstrate such risks.9-12 Subanalyses from 2 of these studies demonstrated an increased risk of neonatal complications only if delivery occurred within 48 hours of indomethacin exposure.7, 8 These studies, however, in-
From the Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology, Medical University of South Carolina. Presented at the Nineteenth Annual Meeting of the Society for MaternalFetal Medicine, San Francisco, California, January 18-23, 1999. Reprint requests: Stephen T. Vermillion, MD, Department of Obstetrics and Gynecology, Medical University of South Carolina, 96 Jonathan Lucas St, Suite 634, Charleston, SC 29425. Copyright © 1999 by Mosby, Inc. 0002-9378/99 $8.00 + 0 6/6/100993
cluded patients receiving indomethacin for maintenance tocolysis (>72 hours). Prolonged therapy with indomethacin has been demonstrated to increase the risk of premature closure of the ductus arteriosus.13, 14 Shortterm tocolysis does not appear to engender this same risk because of the demonstrated ability of the fetal ductus arteriosus to promptly reverse restrictive flow patterns after short-term exposure.13, 14 The purpose of this study was to determine whether maternal exposure to a short course (≤72 hours) of indomethacin tocolysis within 48 hours of delivery is associated with any increased neonatal complications for infants delivered between 24 and 32 weeks’ gestation. Methods We performed a case-control analysis of neonates delivered between 24 and 32 weeks’ gestation after maternal tocolysis with indomethacin. Data were collected and entered into the Medical University of South Carolina 1083
1084 Vermillion and Newman
November 1999 Am J Obstet Gynecol
Table I. Demographic characteristics of indomethacin-treated and control groups Variable
Indomethacin group (n = 75)
Control group (n = 150)
Statistical significance
Weight (g) Gestational age (wk) Multiple gestations Cesarean delivery Female infant Race Black White Other
1121 ± 243 28.7 ± 2.3 16 (21.3%) 27 (36.0%) 40 (53.3%)
1141 ± 287 28.3 ± 2.1 30 (20.6%) 58 (38.6%) 91 (60.6%)
P = .71 P = .63 P = .81 P = .69 P = .36
53 (70.6%) 19 (25.3%) 3 (4.1%)
118 (78.6%) 29 (19.3%) 3 (2.1%)
P = .24 P = .38 P = .66
Table II. Neonatal outcomes of indomethacin-exposed and control groups Indomethacin group (n = 75) Outcome Necrotizing enterocolitis Patent ductus arteriosus Ligated Intraventricular hemorrhage, grade I/II Intraventricular hemorrhage, grade III/IV Bronchopulmonary dysplasia Anuria Pulmonary hypertension Thrombocytopenia Sepsis Death
Control group (n = 150)
No.
%
No.
%
Odds ratio and 95% confidence interval
Statistical significance
5 21 8 5
6.7 28.0 1.1 6.7
9 47 20 6
6.0 31.3 1.3 4.0
1.12 (0.31-3.84) 0.85 (0.44-1.64) 0.78 (0.30-1.98) 1.71 (0.44-6.63)
P = .92 P = .72 P = .57 P = .51
4
5.3
20
13.3
0.37 (0.10-1.19)
P = .07
19 3 1 15 3 11
25.3 4.0 1.3 20.0 4.0 14.6
39 5 4 27 5 17
26 3.3 2.7 18.0 3.3 11.3
0.97 (0.49-1.91) 1.21 (0.22-6.01) 0.49 (0.02-4.80) 1.14 (0.53-2.42) 1.21 (0.22-6.01) 1.34 (0.55-3.25)
P = .96 P = .90 P = .67 P = .86 P = .90 P = .62
perinatal database between January 1992 and July 1997. Only those neonates exposed to 24 to 72 hours of antenatal indomethacin for tocolysis and delivered within 48 hours of the last dose were included in the analysis. Adjunctive tocolysis with indomethacin at the study institution is limited to pregnancies at gestational ages of <32 weeks with persistent uterine activity despite aggressive parenteral therapy with magnesium sulfate. A total of 75 cases were identified, and each case was matched with 2 control subjects for a total of 150 control patients. Matching variables in order of priority included gestational age at delivery, fetal number, exposure to betamethasone for >24 hours before delivery, magnesium sulfate exposure, mode of delivery, infant sex, and race. The estimated gestational age on admission was confirmed by antepartum ultrasonographic examination. All patients received antibiotic prophylaxis for group B Streptococcus and a single course of betamethasone consisting of 12 mg on admission, which was repeated 24 hours later. Patients with ruptured membranes on admission, preeclampsia, insulin-requiring diabetes, or an anomalous fetus were excluded from the study. Neonatal outcomes analyzed included the following: birth weight, confirmed necrotizing enterocolitis, patent
ductus arteriosus, patent ductus arteriosus requiring surgical ligation, grade I/II and grade III/IV intraventricular hemorrhage, bronchopulmonary dysplasia, anuria, thrombocytopenia, pulmonary hypertension, sepsis, and death. Indomethacin exposure characteristics, including length of exposure, cumulative dose, and interval from last dose until delivery, were also determined. Only cases of confirmed necrotizing enterocolitis were included. Confirmed necrotizing enterocolitis required the presence of either bowel perforation or intramural intestinal air on abdominal radiograph. Patent ductus arteriosus was diagnosed after confirmatory echocardiography. Neonates with a patent ductus arteriosus requiring surgical closure had all failed medical treatment with postnatal indomethacin. All infants underwent a cranial ultrasonographic examination within 7 days of delivery and weekly thereafter as necessary to detect intraventricular hemorrhage. Intraventricular hemorrhage was graded by using the criteria of Papile et al.15 Bronchopulmonary dysplasia was based on clinical suspicion in the context of defined radiographic features. Those infants without urine production within the first 12 hours of life were considered to be anuric. Thrombocytopenia was defined as a platelet count of less than
Volume 181, Number 5, Part 1 Am J Obstet Gynecol
100 × 109/L. Persistent pulmonary hypertension of the newborn was defined by previously described echocardiographic criteria.16 Sepsis was diagnosed only when the neonate had positive blood or cerebrospinal fluid cultures. Data were analyzed by using the Student t test for continuous variables and χ2 analysis and the Fisher exact test and Yates’ correction for differences among proportions. All values of P < .05 were considered significant. Results As designed, the groups were not statistically different with respect to mean gestational age at delivery, birth weight, percentage of multiple gestations, rate of cesarean delivery, infant sex, or race (Table I). All neonates received at least 24 hours of magnesium sulfate and 2 doses of betamethasone before delivery. Patients exposed to indomethacin for tocolysis received a median cumulative dose of 225 mg (range, 150350 mg). The median interval from last dose until delivery was 12 hours (range, 2-48 hours). All treated patients received a 50-mg loading dose of indomethacin administered rectally followed by 25 mg every 6 to 8 hours administered orally. No patient received >72 hours of indomethacin therapy. There were no significant differences in the incidences of necrotizing enterocolitis, patent ductus arteriosus, patent ductus arteriosus requiring surgical ligation, grade I/II intraventricular hemorrhage, bronchopulmonary dysplasia, anuria, thrombocytopenia, pulmonary hypertension, sepsis, or death between the 2 groups (Table II). Although not statistically significant (P = .07), the incidence of grade III/IV intraventricular hemorrhage was increased in the control group (13.3%) compared with the group exposed to antenatal indomethacin (5.3%). Comment In our study maternal indomethacin exposure within 48 hours of delivery was not associated with any detectable increase in neonatal complications among infants delivered between 24 and 32 weeks’ gestation. The decision to use indomethacin adjunctively with magnesium sulfate was not randomly assigned but rather was dependent on the attending physician, and thus case selection may be subject to potential biases. To minimize the potential biases associated with nonrandomization, the 2 groups were carefully matched by using a 2:1 ratio of control subjects to patients for multiple factors known to affect neonatal outcome. Limitations of the previously cited nonrandomized studies include smaller sample sizes, incorporation of fewer matching characteristics, and less specific inclusion criteria than used in our analysis.5-7, 12 A post hoc analysis demonstrated that the sample size in this study had a
Vermillion and Newman 1085
power of >0.95 to detect a difference in the incidences of confirmed necrotizing enterocolitis (19% vs 2%) or grade III/IV intraventricular hemorrhage (28% vs 3%) of the degree previously reported by Norton et al5 and Souter et al,7 respectively. We recognize that there are limitations in estimating the sufficiency of a sample size from power calculations performed by using published outcomes with large proportional differences. Therefore a revised power analysis was performed that reduced the treatment effect by two thirds in both the Norton study (7% vs 2% for necrotizing enterocolitis) and the Souter study (9% vs 3% for grade III/IV intraventricular hemorrhage), respectively. By using these effect sizes, our sample was still capable of generating a power of >0.80 to detect a significant difference. Another strength of this study is that we included only those patients exposed to a short course of indomethacin (≤72 hours) immediately preceding delivery between 24 and 32 weeks’ gestation. This design selects the highestrisk infants for neonatal complications and acutely exposes them to indomethacin immediately before delivery. Recently, Souter et al7 demonstrated that infants exposed to indomethacin within 48 hours of delivery had an increase in the incidence of all grades of intraventricular hemorrhage, patent ductus arteriosus, and failure of the patent ductus arteriosus to close with postnatal indomethacin therapy. Infants that were delivered >48 hours from the last maternal exposure did not demonstrate any significant differences in outcomes. Their study differed from ours in that it had a smaller sample size and less stringent inclusion and exclusion criteria. Specifically, Souter et al included patients with ruptured membranes and those receiving longer-term (>72 hours) tocolysis with indomethacin. We specifically addressed those issues by limiting our inclusion criteria to those patients delivered of their neonates within 48 hours of the last indomethacin exposure and by matching the groups for a larger number of potentially significant demographic and treatment variables. The study by Norton et al5 suggested an increased risk of intraventricular hemorrhage, necrotizing enterocolitis, and persistent patent ductus arteriosus. Again, similar methodologic concerns exist, primarily because of the smaller sample size, fewer matching variables, and a variable relationship between indomethacin exposure and delivery. In addition, the increased risk of intraventricular hemorrhage demonstrated in that study was the result of the inclusion of grade II intraventricular hemorrhage along with the more clinically significant grade III/IV hemorrhage. Only grade II intraventricular hemorrhage was significantly associated with indomethacin exposure in the study of Norton et al,5 whereas the incidences of grade III/IV were similar between exposed and control groups. Although not statistically significant, the incidence of
1086 Vermillion and Newman
advanced grades of intraventricular hemorrhage in our study were decreased in the indomethacin-exposed group. The protective effects of low-dose indomethacin administered during the first 72 hours of life to very-lowbirth-weight neonates for the prevention of intraventricular hemorrhage have been demonstrated.17, 18 One of the proposed mechanisms of this protective effect is the ability of indomethacin to increase cerebral vascular resistance and stabilize cerebral hemodynamics in the premature infant.19 It is possible that the limited doses of indomethacin administered shortly before delivery may be providing similar prophylactic effects to those demonstrated with neonatal administration. Indomethacin has achieved increasing popularity as a tocolytic agent with efficacy that is comparable to that of both magnesium sulfate and β-adrenergic receptor agents. In these clinical trials the neonatal outcomes were excellent, but those samples were composed of larger, near-term infants for whom tocolytic therapy had been successful. Concerns have been raised as to the potential neonatal effects of acute exposure to indomethacin immediately before delivery, especially when the neonate is of <32 weeks’ gestation. We conclude that the short-term use of indomethacin for tocolysis at <32 weeks’ gestation was not associated with any increased risk of neonatal morbidity for those patients who were subsequently delivered within 48 hours of the last indomethacin dose. Additionally, antenatal indomethacin exposure in our patients was associated with a reduced risk of advanced grades of neonatal intraventricular hemorrhage, although further studies are needed to explore this relationship. REFERENCES
1. Niebyl JR, Blake DA, White RD, Kumor KM, Dubin NH, Robinson JC, et al. The inhibition of premature labor with indomethacin. Am J Obstet Gynecol 1980;136:1014-9. 2. Besinger RE, Niebyl JR, Keyes WG, Johnson TRB. Randomized comparative trial of indomethacin and ritodrine for the longterm treatment of preterm labor. Am J Obstet Gynecol 1991;164:981-8. 3. Morales WJ, Madhav H. Efficacy and safety of indomethacin compared with magnesium sulfate in the management of preterm labor: a randomized study. Am J Obstet Gynecol 1993;169:97-102.
November 1999 Am J Obstet Gynecol
4. Bivens HA, Newman RB, Fyfe DA, Campbell BA, Stramm SL. Randomized trial of oral indomethacin and terbutaline sulfate for long-term suppression of preterm labor. Am J Obstet Gynecol 1993;169:1065-70. 5. Norton ME, Merrill J, Cooper BAB, Kuller JA, Clyman R. Neonatal complications after the administration of indomethacin for preterm labor. N Engl J Med 1993;329:1602-7. 6. Major CA, Lewis DF, Harding JA, Porto MA, Garite TJ. Tocolysis with indomethacin increases the incidence of necrotizing enterocolitis in the low-birth-weight neonate. Am J Obstet Gynecol 1994;170 (1 Pt 1):102-6. 7. Souter D, Harding J, McCowan L, O’Donnell C, McLeay E, Baxendale H. Antenatal indomethacin—adverse fetal effects confirmed. Aust N Z J Obstet Gynaecol 1998;38:11-6. 8. Marpeau L, Bouillie J, Barrat J, Milliez J. Obstetrical advantages and perinatal risks of indomethacin: a report of 818 cases. Fetal Diagn Ther 1994;9:110-5. 9. Dudley DKL, Hardie MJ. Fetal and neonatal effects of indomethacin used as tocolytic agent. Am J Obstet Gynecol 1985;151:181-4. 10. Niebyl JR, Witter FR. Neonatal outcome after indomethacin treatment of preterm labor. Am J Obstet Gynecol 1986;155:747-9. 11. Wurtzel D. Prenatal administration of indomethacin as a tocolytic agent: effect on neonatal renal function. Obstet Gynecol 1990;76:689-92. 12. Gardner MO, Owen J, Skelly S, Hauth JC. Preterm delivery after indomethacin—a risk factor for neonatal complications? J Reprod Med 1996;41:903-6. 13. Vermillion ST, Scardo JA, Lashus AG, Wiles HB. The effect of indomethacin tocolysis on fetal ductus arteriosus constriction with advancing gestational age. Am J Obstet Gynecol 1997;177:25661. 14. Moise KJ. Effect of advancing gestational age on the frequency of fetal ductal constriction in the association with maternal indomethacin use. Am J Obstet Gynecol 1993;168:1350-3. 15. Papile LA, Burstein J, Burstein R, Koffler H. Incidence and evolution of subependymal and intraventricular hemorrhage: a study of infants with birth weights less than 1,500 gm. J Pediatr 1978;92:529-34. 16. Morin SC. Persistent pulmonary hypertension of the newborn. In: Enmanouilides GC, Allen HD, Riemenschneider TA, Gupgesell HP, editors. Moss and Adams heart disease of the infant, child, and adolescent. 5th ed. Baltimore: Williams & Wilkins; 1995. p. 599-614. 17. Ment LR, Duncan CC, Ehrenkranz RA, Kleinman CS, Taylor KJ, Scott DT, et al. Randomized low-dose indomethacin trial for prevention of intraventricular hemorrhage in very low birth weight neonates. J Pediatr 1988;112:948-55. 18. Bandstra ES, Montalvo BM, Goldberg RN, Pacheco I, Ferrer PL, Flynn J, et al. Prophylactic indomethacin for prevention of intraventricular hemorrhage in premature infants. Pediatrics 1988;82:533-42. 19. Yanowitz TD, Yao AC, Werner JC, Pettigrew KD, Oh W, Stonestreet BS. Effects of prophylactic low-dose indomethacin on hemodynamics of very low birthweight infants. J Pediatr 1998;132:28-34.