- Email: [email protected]
The therapeutic efficacy and cost-effectiveness of aggressive tocolysis for premature labor associated with premature rupture of the membranes
The therapeutic efficacy and cost-effectiveness of aggressive tocolysis for premature labor associated with premature rupture of the membranes Carl P. Weiner, MD," Kate Renk, BSN," and Marie Klugman, PhD• Iowa City, Iowa We conducted a randomized trial comparing bed rest with tocolysis to determine the therapeutic efficacy, safety, and cost-effectiveness of tocolysis for the treatment of preterm labor after membrane rupture. One hundred nine women participated over a 26-month interval. Treatment groups did not differ significantly in terms of gestational age at membrane rupture, gestational age at delivery, birth weight, maternal or fetal infectious morbidity, respiratory distress syndrome, necrotizing enterocolitis, or perinatal mortality. Prolongation of intrauterine time after the onset of uterine contractions was seen in women receiving tocolysis (105.2 ± 157 hours versus 62.1 ± 77 hours, p = 0.06). This prolongation was not associated with a significant reduction in the total cost per surviving infant (tocolysis, $38,593 ± $40,887 versus bed rest, $43, 158 ± $37, 116; p = 0.445). The cost difference was artifactual. The number of very premature infants born (<28 weeks' gestation) was unequal in the two groups (12 in the bed rest group and 5 in the tocolysis group) and skewed the results. Before 28 weeks' gestation tocolysis was associated with a significant increase in intrauterine time after the onset of regular contractions (p = 0.05). However, there was no identifiable perinatal benefit garnered from the additional 5 days. After 28 weeks there were no significant differences between treatment groups in terms of intrauterine time after the onset of regular contractions and total cost per surviving infant. Because tocolysis does not improve perinatal outcome and can itself be associated with major maternal morbidity, it should be avoided after 28 weeks' gestation. Before 28 weeks' gestation tocolysis may greatly increase intrauterine time, but the benefit of this prolongation is not clear. (AM J OBSTET GvNECOL 1988;159:216-22.)
Key words: Preterm delivery, preterm rupture of the membranes, tocolysis, ritodrine, magnesium sulfate
Although preterm premature rupture of the membranes complicates only 1% to 2% of all pregnancies, 1. 2 it is associated with 40% of preterm deliveries.' Bed rest is generally accepted as the preferred management for the uninfected patient with preterm premature rupture of the membranes who is not in labor,4 but other aspects of care remain unsettled. Particularly controversial is the issue of tocolysis in women with premature rupture of the membranes. Whereas tocolysis for premature labor associated with intact membranes is cost-effective,5 many physicians refrain from the use of tocolytic agents in women with pre term premature rupture of the membranes because both their efficacy and their safety in this subgroup are not clear. Almost all published reports of tocolysis for preterm labor preceded by rupture of membranes center From the Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology,' and the Division of Biostatistics, Department of Preventive Medicine,' University of Iowa Medical School. Received for publication October 29, 1987; revised December 21, 1987; accepted January 29, 1988. Reprint requests: Carl P. Weiner, MD, Department of Obstetrics and Gynecology, University of Iowa College of Medicine, Iowa City, IA 52242.
216
on delaying delivery only to administer corticosteroids in the hopes of accelerating fetal lung maturity. Before the conception of the current study there was only one prospective, randomized trial of tocolysis for premature labor after rupture of the membranes with the express goal of prolonging pregnancy >48 hours. 6 That study had three limitations: (l) The number of subjects (n = 30) was probably inadequate to conclude benefit or no benefit from therapy; (2) the dose and dose interval of the oral ~-mimetic agent was not individualized to the patient's response after successful labor inhibition with parenteral drug; and (3) pregnancies at a relatively advanced gestational age (34 to 36 weeks) were included. Recently a second study has been published focusing on gestation <28 weeks. 7 That investigation did not examine medical care costs. Furthermore it was potentially biased by the large number of patients who declined the opportunity to participate-only 79 women were randomized over a 4-year interval. To test the hypothesis that aggressive tocolysis would produce benefits similar to those achieved when used for preterm labor with intact membranes, we con-
Volume 159 Number I
Aggressive tocolysis and PROM
LABOR 3cm or 80%
NO LABOR
I I Amniocenteaia IV Flulda
atop
~ mat..• lrnrneture I
I
----------
Bedreat
I
delivery
Bedreat
~
labor
labor
I
contractions cont inue
1
amnlonltla
I
IV llulda
----------
~
Bedreat
delivery
I
Tocolyaia
~/"--.._ contraction• atop
contraction• continue
I
I
I
cycle repeats 11134 weeks
A
failure
success
amniocenteaia
mature
immature
I
I
cycle repeata t~
34 weeks
~
mature
Immature
exclude
cycle repeata
I tll 34 weeks
I
"
amnionitil
amnionitia
I aucceaa
I
I
deMvery
repeat cycle tll 34 weeka
Detailed
failure
dell very
1
amnkx:enteaia
exclude
Bedreat
~ I I
contracttona sto p
~
delivery
contraction• continue
delivery
continue
I
amnlonltla
>
Conaent, Randomize
Conaent, Rendomlze
~
LABOR 3cm or 80%
~. .A.~/ <
exclude
217
I
delivery
delivery
Skeletal
Fig. I. Generalized flow diagram of the management protocol used. IV, Intravenous.
ducted a randomized trial comparing bed rest with aggressive tocolysis to determine the therapeutic efficacy, safety, and cost-effectiveness of tocolysis for preterm labor after membrane rupture. Methods
From August 1984 through October 1986, all women with documented premature rupture of the membranes at ,;;34 weeks' gestation admitted to the University of Iowa Hospital without either pyrexia (temperature >99.6° F) or uterine tenderness between contractions, and who were not already receiving pharmacologic tocolysis, were solicited to participate in a randomized trial comparing bed rest with aggressive tocolysis for the treatment of premature labor. The protocol was approved by the University of Iowa Committe for Human Investigation. The diagnosis of ruptured membranes required arborization of a smear made with a cotton swab dipped at the cervical os (ferning). Patients were randomized shortly after admission (whether or not in active labor). Sixty percent of randomized women were assigned to the tocolysis group to account for both protocol violations and women who would develop chorioamnionitis before the onset of labor. The sequence of assignment was derived from a random number table and the therapy choice was sealed in sequentially numbered manila envelopes, which were opened after written consent had been given. Tocolysis for lab.or associated with premature rupture of the membranes was unavailable outside of the study.
Protocol. Patients were treated as identically as clinically possible except for tocolysis (Fig. I). To assure a similar degree of hydration in both treatment groups, lactated Ringer's solution was infused at admission in a quantity sufficient to produce urine specific gravity < 1.005. Initial patient evaluation included microbiologic culture of the cervix, a single pelvic examination to rule out umbilical cord prolapse and document cervical effacement and dilation, and, when the quantity of amniotic fluid remaining was adequate, an amniocentesis for Gram stain, culture, and fetal pulmonary maturity studies. Patients were not excluded for a positive Gram stain or culture. When the amniocentesis was unsuccessful on admission (48%), management proceeded along the decision tree. A second attempt later during the hospitalized period was not required for inclusion in the study. Activity for all patients was limited to bathroom privileges. Patients with amniotic fluid studies consistent with fetal lung maturity were excluded. Corticosteroids were not administered. Fetal heart rate testing was performed every other day. The Biophysical Profile was not in use at the University of Iowa in 1984 and was not incorporated into the study design. In the absence of chorioamnionitis, antibiotics were given only for the treatment of bacteriuria (symptomatic or asymptomatic) or group B streptococcal colonization of the cervix or vagina. Induction of labor for women in either group was reserved for clinical amnionitis or fetal distress. Intravenous tocolysis was begun if uterine activity exceeded three contractions per hour. This liberal ap-
218 Weiner, Renk, and Klugman
proach was applied in the hope of treating any uterine activity before it could become well established. Tocolysis was discontinued only for patient intolerance of the drug's side effects, cervical dilation >4 cm, or chorioamnionitis. Tocolytic agents used in our facility included ritodrine, terbutaline, and magnesium sulfate, the functional efficacy of which is similar.8· 9 In the tocolysis group, all but two women received a 13-mimetic initially and all but two of these received ritodrine. Drug infusion was increased rapidly until either uterine activity (as perceived by either the patient or on the external tocodynometer) was all but abolished, or the dose exceeded 450 µg/min for ritodrine or 20 µg/min for terbutalnie. Magnesium sulfate up to 4.5 gm/hr was added liberally if uterine activity could not be controlled by the 13-mimetic agent. After 24 hours of successful tocolysis, dosing with an oral 13-mimetic (usually terbutaline in the interest of cost containment) was begun and both the dose and dose interval were individualized to maintain the maternal heart rate >95 bpm. Recurrent episodes of labor were treated similarly. Adherence to the study protocol was ascertained and the data were collected contemporaneously by one of the authors (K. R.). Gestational age was based on the "best obstetric dates" -a combination of menstrual, historic, and objective criteria. If the postnatal pediatric estimate of gestational age (method of Ballard) differed by >3 weeks, the pediatric estimate was accepted as correct. Morbidity. Maternal morbidity was limited to infection. The patient's temperature was measured at least every 4 hours. The diagnosis of chorioamnionitis either before or during labor required pyrexia and uterine tenderness between contractions and was not made on the basis of laboratory tests alone. When possible, amniotic fluid was obtained on insertion of the intrauterine pressure catheter during labor and cultured. Febrile morbidity required two readings ;;o l 00° F after the first postpartum day. Specimens for anaerobic and aerobic cultures were obtained from the blood, urine, and uterus as part of the standard evaluation of a postpartum fever. The diagnosis of sepsis required positive blood cultures. In the absence of another explanation (e.g., urinary tract or upper respiratory infection), febrile morbidity plus uterine tenderness to palpation was coded as endometritis. Major neonatal complications included sepsis, probable sepsis, respiratory distress syndrome, and necrotizing enterocolitis. The diagnosis of sepsis required either a positive culture from a blood, urine, or cerebral spinal fluid specimen obtained within 6 hours of birth; a chest x-ray film consistent with pneumonia associated with a positive endotracheal aspirate obtained within 6 hours of birth; or a positive antibody test for group B streptococcus in a urine specimen obtained within 6
July 1988 Am J Obstet Gynecol
hours of birth. A neonate was classified as probably septic whenever there was laboratory (e.g., leukopenia, thrombocytopenia) or clinical (e.g., hypotension, tachypnea) evidence of infection that did not fulfill the above criteria. Respiratory distress syndrome was a clinical diagnosis based on the standard criteria. ' 0 Charges. Patient charges (mother and child, hospital, and physician), including all consultant and anesthesia fees, were determined after discharge from the Hospital Business and Accounts Department. These included costs subsequently borne by the State of Iowa Indigent Care Program and federally funded Medicaid. No inflationary adjustments in costs were made because a similar number of women were enrolled each of the 2 years. In the event the infant was transferred from the university hospital to a level II facility before discharge home, outlying physician and hospital costs were obtained by direct correspondence with the care provider. Similar numbers of infants in each treatment group were transferred to local hospitals. The cost per surviving infant (i.e., discharged home alive) by treatment group was defined as the aggregate maternal and neonatal medical cost divided by the number of survivors. For the post hoc analysis that included multiple gestations, the maternal cost for that pregnancy was multiplied by the number of fetuses to yield the equivalent maternal-fetal unit cost. Analysis. Results were assessed by subprograms from the Statistical Analysis System release 5. Assessment involved parametric and nonparametric tests with and without logarithmic conversion where appropriate. Post hoc analyses were performed on the following groups: excluding pregnancies concluded before neonatal viability (25 weeks' gestation), excluding pregnancies complicated by infection, dividing the study population by the gestational age at rupture of the membranes (<28 weeks, ;;.28 weeks), including all randomized patients, and including all randomized patients except those with multiple gestations. A p value ~0.05 was considered significant and p < 0.10 was considered near significant.
Results One hundred seventy-eight women were eligible to participate; 109 (61 %) gave consent and were randomized. Thirty-four patients were excluded from the initial analyses: 8 in whom the protocol was violated, 6 who never received tocolysis either because of chorioamnionitis or advanced cervical dilation, 8 who withdrew from the study before delivery (all in the tocolysis group), and 12 multiple gestations (eight in the bed rest group and four in the tocolysis group). The protocol continues for the latter group, which we consider an entity separate from the singleton gestation. The demographic profile of the remaining participants is listed
Aggressive tocolysis and PROM
Volume 159 Number I
Table I. Demographic profile
No. of patients Maternal age (yr) Gravidity Parity Abortions Gestation at ROM (wk) % Married % White % Tobacco users % High school education or higher % Medicaid or state aid % Prior stillbirth % Medical complications
Table IV. Neonatal results
Bed rest
Tocolysis
42 23.5 ± 4.4 2.1 ± 1.2 0.8 ± 1.0 0.3 ± 0.5 29.4 ± 3.2 57.l 85.7 54.8 69.0
33 25.7 ± 5.7* 2.3 ± 1.2 0.7 ± 0.9 0.5 ± 0.8 30.0 ± 3.2 62.5 90.3 33.3t 68.8
10.0 0.0 2.4
5.3 0.0 6.0
Table II. Cervical status
Cervical dilation (cm) Cervical effacement (%) Station
Bed rest
1.5 ± 1 1.5 ± 1 53 ± 30 49.8 ± 29
-2
Birth weight (gm) Umbilical vein pH Umbilical artery pH Sepsis(%) Probable sepsis (%) RDS(%) NEC(%) Neonatal deaths(%)
Bed rest
Tocolysis
1518 ± 563 7.35 ± 0.09 7.27 ± 0.09 3 (7.1) 17 (40.5) 22 (52.4) IO (23.8) 5 (11.9)
1648 ± 536 7.34 ± 0.08 7.25 ± 0.10 3 (9.))
9 15 6 3
(27.3) (45.4) (18.2) (9.1)
Data are mean ± SD. All differences nonsignificant. RDS, Respiratory distress syndrome; NEC, necrotizing enterocolitis.
Table V. Comparisons by selected intervals
Data are mean ± SD. ROM, Rupture of membranes. *p = 0.06. tp < 0.053.
Total
219
-3
Tocolysis
1.6 ± 1 57 ± 31 -1.8
From membrane rupture to delivery <24 hr <48 hr <72 hr <24 hr <48 hr <72 hr
Bed rest(%)
Tocolysis (%)
7.1 23.8 38.l 42.1 63.2
6.1 12.1 30.3 24.1 55.1 62.l
71.1
All differences nonsignificant.
Data are mean ± SD. All differences nonsignificant.
Table III. Maternal results
Gestation at delivery (wk) % Cesarean section % Maternal infection Hospital days before delivery
Bed rest
Tocolysis
30.l ± 3.3 16.7 26.2 5.2 ± 6.0
3.10 ± 2.9 12.1 27.3 6.7 ± 7.2
Data are mean ± SD. All differences nonsignificant.
in Table I. All subsequent analyses were based on the 75 remaining women. Women randomized to the tocolysis group were slightly older (p = 0.06) and less likely to smoke in excess of I 0 cigarettes per day (p = 0.053; Table I). There were no differences between women in the two groups in terms of either cervical status (Table II) or percentage of women in active labor on admission to the university hospital. Although 71 of the l 09 women randomized had regular uterine activity documented on admission, prodromal signs of membrane rupture were only infrequently noted by the patient. Regular uterine activity preceded rupture in 7.6%, cramping in 13.2%, and spotting in 9.4% of the women. More than 80% of women were unaware of any change suggesting rupture of the membranes was imminent. This .intimates that patient education on the warning signs of premature rupture of the membranes will not end the problem.
Seventy-nine women (72%) had a urine specific gravity > 1.005 at admission and received a fluid bolus. Of these 59 women were having regular uterine contractions when the bolus was begun; 42% of these stopped contracting with hydration alone. There was no difference between treatment groups in response to hydration. Most fetuses had a reactive heart rate pattern at admission (95%). Variable decelerations (consistent with intermittent umbilical cord compression) were common. Seventy-four percent had at least mild variable decelerations and 3.9% had at least one severe variable (<90 beats/ min for ;;,;45 seconds) deceleration. Approximately 50% ( 17) of women in the tocolysis group received two agents. The mean maximum dose of ritodrine was 225 µg/min and that of magnesium sulfate was 2.9 gm. Despite the aggressiveness of drug therapy, there were no major complications attributable to the tocolytic agents during the study. No significant differences between the treatment groups were observed for the interval from membrane rupture to delivery, proportion of cesarean delivery, or maternal infectious morbidity (Table Ill). There were also no significant or near significant differences in neonatal outcome for birth weight, umbilical cord blood gases, sepsis or probable sepsis, respiratory distress syndrome, necrotizing enterocolitis, or neonatal mortality (Table IV). However, there was a near significant increase in
220
Weiner, Renk, and Klugman
July 1988 Am J Obstet Gynecol
Table VI. Interval from onset of regular contractions to delivery and cost per survivor by subgroup Bed rest
Tocolysis
I
Cost per suroivor ($)
Interoal (hr)
58 ± 75* 62 ± 77
51,145 ± 52,196 43,158 ± 37,116
117 ± 174* 96 ± 141
48,215 ± 40,618 48,232 ± 44,524
69 ± 79
39,156 ± 34,364
108 ± 159
33,759 ± 32,286
58 ± 75*
51,145 ± 52,196
129 ± 194*
49,091 ± 36,975
Group
Interoal (hr)
All patients All patients except multiple births All patients except delivery <25 wk All patients except protocol violation
J
Cost per suroivor ($)
Data are mean ± SD.
*p < 0.05.
Table VII. Interval from onset of regular uterine contractions gestational age at membrane rupture
to delivery and cost per survivor by
<28 wk Bed rest
No. of patients Interval (hr) Neonatal deaths Cost per survivor ($)
*p
= 0.05.
tp
=
12 53.4 ± 87* 4
82,871 ± 30,650t
>28 wk
I
Tocolysis
Bed rest
5 232.8 ± 312 2 118,206 ± 42,172
24 68.5 ± 70
I
Tocolysis
25 78.7 ± 93
1
I
23,302 ± 22,779
22,670 ± 15, 195
O.D7.
the duration of intrauterine time after the onset of regular contractions in the tocolysis group (bed rest, 62.l ± 77 hours; tocolysis, 105.2 ± 157 hours; p = 0.06). The percent of women delivered 24, 48, and 72 hours after rupture of membranes and the onset of regular uterine contractions is shown in Table V. There were no significant differences between the two groups. A cost analysis was performed in case small benefits otherwise missed by the above analysis might be reflected in a reduction of the overall medical cost. Women who received tocolysis had a near-significant increase in hospital costs (bed rest, $4399 ± $3266; tocolysis, $4883 ± $2491; p = 0.076) and a significant increase in physician fees (bed rest, $939 ± $863; tocolysis, $1419 ± $1003;p = 0.005). There were no significant cost differences between treatment groups for the neonates for either the hospital (bed rest, $26,839 ± $30,651; tocolysis, $23,900 ± $33,728; p = 0.602) or physician (bed rest $5674 ± $6009; tocolysis, $4882 ± $6779; p = 0.418). The overall cost per survivor for the tocolysis group was suggestive of benefit but not significantly reduced (bed rest, $43, 158 ± $37,116; tocolysis, $38,593 ± $40,887; p = 0.4453). Should this percent reduction in costs hold, we calculate that a study population in excess of 3000 women with premature rupture of the membranes <34 weeks' gestation would be needed to reach a significant difference.
In an attempt to understand further the absolute reduction in costs per surviving neonate and the nearsignificant increase in intrauterine time achieved by tocolysis after the onset of regular uterine contractions, the above analyses were repeated with these modifications: first, excluding pregnancies complicated by either maternal or neonatal infection; second, excluding all pregnancies delivered before viability; third, including all randomized patients; fourth, excluding only multiple gestations (Table VI); and fifth, by dividing the 75 women according to gestational age at rupture of the membranes <28 or ~28 weeks' gestation. The results were similar to those already reported in Tables I to II I for each of the first four post hoc analyses. Only the last analysis yielded findings of interest (Table VII). The distribution of patients into the treatment arms was skewed in the group with delivery <28 weeks' gestation (bed rest, 12; tocolysis, 5). The added intrauterine time and the reduction in costs per survivor associated with tocolysis was wholly explainable by this unequal distribution. However, there was an absolute reduction in the cost per survivor (2.7%) and an increase in intrauterine time after the onset of regular contractions (15%) associated with tocolysis when delivery occurred after 28 weeks' gestation. However, achievement of an 80% power for these values in this subset would require 6200 patients with rupture of the membranes <34 weeks.
Volume 159 Number I
We considered the possibility that a clinically silent chorioamnionitis before birth might render the uterus unresponsive to tocolysis and mask a significant effect of tocolysis in uninfected patients. After excluding pregnancies of women with evidence of antepartum chorioamnionitis or postpartum endometritis or neonates with confirmed septicemia (16 in the bed rest group and 11 in the tocolysis group), there was a nearsignificant increase in the interval from onset of regular contractions to delivery in the tocolysis group (bed rest, 60.4 ± 79.2 hours; tocolysis, 116.5 ± 176.5 hours; p = 0.06). Although there was no significant difference in the total medical costs per survivor (bed rest, $45,302 ± $40,025; tocolysis, $37,413 ± $44,732;p = 0.390), the difference does amount to a 17% reduction in costs. Six hundred patients would be required for the detection of this difference at the 0.05 level.
Comment Preterm premature rupture of the membranes is a major cause of perinatal morbidity and mortality. Numerous investigators have concluded that the risk of neonatal respiratory distress outweighs the risk of perinatal or maternal sepsis in the absence of proved fetal lung maturity.'·'· 11 · " Presently, bed rest under close medical supervision rather than immediate induction of labor is the standard management of the patient with preterm premature rupture of the membranes who is not in labor. Management of the associated premature labor, which occurs in 80% of women within 7 days of preterm premature rupture of the membranes, 12 has not been adequately studied. The use of tocolysis for preterm labor associated with intact membranes does improve neonatal outcome.5 A similar analysis has not been used with the patient with preterm premature rupture of the membranes. Because labor may be a sign of incipient chorioamnionitis, and chorioamnionitis is associated with an increase in perinatal morbidity and mortality, many physicians refrain from tocolysis, concerned that any prolongation of intrauterine time would increase infectious complications. However, if tocolysis were otherwise safe and effective, this practice would deny those benefits to the majority of patients who are free of infection. Some investigators have advocated tocolysis only for an interval long enough to administer corticosteroids for the enhancement of fetal lung maturity (48 hours) and then delivery of the patient. 1• Unfortunately, the efficacy of maternal corticosteroid administration after rupture of the membranes is in doubt." Much of the confusion undoubtedly stems from the paucity of information concerning either the efficacy or safety of tocolysis for labor after preterm premature rupture of the membranes. At the study's inception, we could identify only one prospective, randomized
Aggressive tocolysis and PROM 221
trial of tocolysis for labor after preterm premature ru pture of the membranes. Although the number of pregnancies studied was small (n = 30), those investigators 6 could not demonstrate an increase in infection with tocolysis but they did find an increase in the percentage of women delivered ~24 hours after rupture of the membranes. We hypothesized that the benefits of tocolysis might have been clear if their treatment protocol had been more aggressive and their analysis had involved multiple neonatal parameters. The present investigation addresses not only the therapeutic efficacy and safety of tocolysis after preterm premature rupture of the membranes, but also, for the first time, its cost efficacy. Aggressive tocolysis initiated because of regular uterine contractions (in contrast to awaiting cervical dilation) failed to significantly increase overall intrauterine time above that achieved by bed rest. However, the intrauterine time achieved after the onset of regular uterine contractions was increased by almost 48 hours when tocolysis was used. There was no evidence that tocolysis increased either maternal or perinatal infectious complications. Furthermore, the total medical cost per surviving infant was reduced in association with tocolysis. Unfortunately, these salutary effects were likely an artifact produced by an unequal distribution between treatment groups of infants delivered at <28 weeks' gestation. These infants had significantly higher costs per survivor (all pregnancies delivered <28 weeks, $93,264, versus $23,284 if delivered after 28 weeks). Despite a significant increase in intrauterine time achieved with tocolysis after the onset of regular contractions in the group delivering at <28 weeks' gestation, there was a near-significant increase in cost per survivor (p = 0.07). Whether this trend is real or a skewing brought about by the small population and unequal distribution is unknown. In contrast to the infants delivered at <28 weeks' gestation, there were no significant differences between treatment groups with delivery after 28 weeks. Although there was an absolute reduction in the cost per survivor associated with tocolysis (2. 7% ), power analysis revealed that the difference would require a study population in excess of 6200 women with rupture of the membranes between 28 and 34 weeks' gestation. It seems unlikely tocolysis offers any clinical benefit should rupture of the membranes occur after 28 weeks. Deleting from analysis those pregnancies complicated by either maternal or neonatal infectious morbidity failed to alter the aforenoted pattern. The cost per neonatal survivor was reduced by 17% in association with tocolysis, but confirmation of this difference would require at least a trebling of the sample size. Thus the lack of a benefit from tocolysis for the treatment of labor associated with premature rupture of the
222
Weiner, Renk, and Klugman
membranes cannot be attributed solely to a few patients in whom tocolysis failed because of infection. We conclude that treatment of labor after preterm premature rupture of the membranes does not improve perinatal outcome after 28 weeks' gestation. Because therapy itself entails major maternal risks such as pulmonary edema, tocolysis should not be used for these patients. At <28 weeks' gestation tocolysis may increase intrauterine time, but no benefit could be demonstrated. The use of tocolysis in these very preterm pregnancies would not in our population be interdicted out of concern for increasing infectious morbidity, but its benefit remains unclear. REFERENCES I. Druzin ML, Toth M, Ledger WJ. Nonintervention in pre-
2. 3.
4.
5.
mature rupture of the amniotic membranes. Surg Gynecol Obstet 1986; 163:5-8. Fayez JA, Hasan AA, Jonas HS, Miller GL. Management of premature rupture of the membranes. Obstet Gynecol 1978;52:17-21. Daikoku NH, Kaltreider DJ.Johnson TRB,JohnsonJWC, Simmons MA. Premature rupture of membranes and preterm labor: Neonatal infection and perinatal mortality risks. Obstet Gynecol 1981;58:417-25. Gibbs RS, Sweet RL. Maternal and fetal infections. In: Creasy RK, Resnik R, eds. Maternal fetal medicine: Principles and practice. Philadelphia: W. B. Saunders, 1984: 607-13. Korenbrot CC, Aalta LH, Laros RK. The cost effectiveness of stopping preterm labor with beta-adrenergic treatment. N Engl J Med 1984;310:691-6.
July 1988 Am J Obstet Gynecol
6. Christensen KK, Ingemarsson I, Leideman T, Solum H, Svenningsen N. Effect of ritodrine on labor after premature rupture of the membranes. Obstet Gynecol 1980;55: 187-90. 7. Garite TJ, Keegan KA, Freeman RK, Nageotte MP. A randomized trial of ritodrine tocolysis versus expectant management in patients with premature rupture of membranes at 25 to 30 weeks' gestation. AMJ OBSTET GYNECOL 1987; 157:388-93. 8. Petrie RH. Tocolysis using magnesium sulfate. Semin Perinatol 1981 ;5:266-9. 9. Caritis SN, Toig G, Heddinger LA, Ashmead G. A doubleblind study comparing ritodrine and terbutaline in the treatment of preterm labor. AM J OBSTET GYNECOL 1984; 150:7-14. JO. Stahlman MT. Acute respiratory disorders of the newborn. In: Avery GB, ed. Neonatology-pathophysiology and management of the newborn. Philadelphia: JB Lippincott, 1981 :507-13. 11. Varner MW, Galask RP. Conservative management of premature rupture of the membranes. AM J 0BSTET GYNECOL 1981; 140:39-45. 12. Kappy KA, Cetrulo CL, Knuppel RA, et al. Premature rupture of the membranes: A conservative approach. AM J 0BSTET GYNECOL 1979; 134:655-61. 13. Garite TJ. Premature rupture of the membranes: The enigma of the obstetrician. AM J OBSTET GYNECOL 1985; 151: 1001-5. 14. Mead PB. Management of the patient with premature rupture of the membranes. Clin Perinatol 1980;7:243-55. 15. Garite TJ, Freeman RK, Linzey EM, Braly PS, Dorchester WL. Prospective randomized study of corticosteroids in the management of premature rupture of the membranes and the premature gestation. AM J 0BSTET GYNECOL 1981; 141 :508-15.
Bound volumes available to subscribers Bound volumes of the AMERICAN JOURNAL OF OBSTETRICS AND GYNECOLOGY are available to subscribers (only) for the 1988 issues from the Publisher, at a cost of $54.00 ($77.00 international) for Vol. 158 (January-] une) and Vol. 159 (j uly-December). Shipping charges are included. Each bound volume contains a subject and author index and all advertising is removed. Copies are shipped within 60 days after publication of the last issue in the volume. The binding is durable buckram with the JOURNAL name, volume number, and year stamped in gold on the spine. Payment must accompany all orders. Contact The C. V. Mosby Company, Circulation Department, 11830 Westline Industrial Drive, St. Louis, Missouri 63146, USA; phone (800) 325-41 77, ext. 351. Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular JOURNAL subscription.
Key words: Preterm delivery, preterm rupture of the membranes, tocolysis, ritodrine, magnesium sulfate
Although preterm premature rupture of the membranes complicates only 1% to 2% of all pregnancies, 1. 2 it is associated with 40% of preterm deliveries.' Bed rest is generally accepted as the preferred management for the uninfected patient with preterm premature rupture of the membranes who is not in labor,4 but other aspects of care remain unsettled. Particularly controversial is the issue of tocolysis in women with premature rupture of the membranes. Whereas tocolysis for premature labor associated with intact membranes is cost-effective,5 many physicians refrain from the use of tocolytic agents in women with pre term premature rupture of the membranes because both their efficacy and their safety in this subgroup are not clear. Almost all published reports of tocolysis for preterm labor preceded by rupture of membranes center From the Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology,' and the Division of Biostatistics, Department of Preventive Medicine,' University of Iowa Medical School. Received for publication October 29, 1987; revised December 21, 1987; accepted January 29, 1988. Reprint requests: Carl P. Weiner, MD, Department of Obstetrics and Gynecology, University of Iowa College of Medicine, Iowa City, IA 52242.
216
on delaying delivery only to administer corticosteroids in the hopes of accelerating fetal lung maturity. Before the conception of the current study there was only one prospective, randomized trial of tocolysis for premature labor after rupture of the membranes with the express goal of prolonging pregnancy >48 hours. 6 That study had three limitations: (l) The number of subjects (n = 30) was probably inadequate to conclude benefit or no benefit from therapy; (2) the dose and dose interval of the oral ~-mimetic agent was not individualized to the patient's response after successful labor inhibition with parenteral drug; and (3) pregnancies at a relatively advanced gestational age (34 to 36 weeks) were included. Recently a second study has been published focusing on gestation <28 weeks. 7 That investigation did not examine medical care costs. Furthermore it was potentially biased by the large number of patients who declined the opportunity to participate-only 79 women were randomized over a 4-year interval. To test the hypothesis that aggressive tocolysis would produce benefits similar to those achieved when used for preterm labor with intact membranes, we con-
Volume 159 Number I
Aggressive tocolysis and PROM
LABOR 3cm or 80%
NO LABOR
I I Amniocenteaia IV Flulda
atop
~ mat..• lrnrneture I
I
----------
Bedreat
I
delivery
Bedreat
~
labor
labor
I
contractions cont inue
1
amnlonltla
I
IV llulda
----------
~
Bedreat
delivery
I
Tocolyaia
~/"--.._ contraction• atop
contraction• continue
I
I
I
cycle repeats 11134 weeks
A
failure
success
amniocenteaia
mature
immature
I
I
cycle repeata t~
34 weeks
~
mature
Immature
exclude
cycle repeata
I tll 34 weeks
I
"
amnionitil
amnionitia
I aucceaa
I
I
deMvery
repeat cycle tll 34 weeka
Detailed
failure
dell very
1
amnkx:enteaia
exclude
Bedreat
~ I I
contracttona sto p
~
delivery
contraction• continue
delivery
continue
I
amnlonltla
>
Conaent, Randomize
Conaent, Rendomlze
~
LABOR 3cm or 80%
~. .A.~/ <
exclude
217
I
delivery
delivery
Skeletal
Fig. I. Generalized flow diagram of the management protocol used. IV, Intravenous.
ducted a randomized trial comparing bed rest with aggressive tocolysis to determine the therapeutic efficacy, safety, and cost-effectiveness of tocolysis for preterm labor after membrane rupture. Methods
From August 1984 through October 1986, all women with documented premature rupture of the membranes at ,;;34 weeks' gestation admitted to the University of Iowa Hospital without either pyrexia (temperature >99.6° F) or uterine tenderness between contractions, and who were not already receiving pharmacologic tocolysis, were solicited to participate in a randomized trial comparing bed rest with aggressive tocolysis for the treatment of premature labor. The protocol was approved by the University of Iowa Committe for Human Investigation. The diagnosis of ruptured membranes required arborization of a smear made with a cotton swab dipped at the cervical os (ferning). Patients were randomized shortly after admission (whether or not in active labor). Sixty percent of randomized women were assigned to the tocolysis group to account for both protocol violations and women who would develop chorioamnionitis before the onset of labor. The sequence of assignment was derived from a random number table and the therapy choice was sealed in sequentially numbered manila envelopes, which were opened after written consent had been given. Tocolysis for lab.or associated with premature rupture of the membranes was unavailable outside of the study.
Protocol. Patients were treated as identically as clinically possible except for tocolysis (Fig. I). To assure a similar degree of hydration in both treatment groups, lactated Ringer's solution was infused at admission in a quantity sufficient to produce urine specific gravity < 1.005. Initial patient evaluation included microbiologic culture of the cervix, a single pelvic examination to rule out umbilical cord prolapse and document cervical effacement and dilation, and, when the quantity of amniotic fluid remaining was adequate, an amniocentesis for Gram stain, culture, and fetal pulmonary maturity studies. Patients were not excluded for a positive Gram stain or culture. When the amniocentesis was unsuccessful on admission (48%), management proceeded along the decision tree. A second attempt later during the hospitalized period was not required for inclusion in the study. Activity for all patients was limited to bathroom privileges. Patients with amniotic fluid studies consistent with fetal lung maturity were excluded. Corticosteroids were not administered. Fetal heart rate testing was performed every other day. The Biophysical Profile was not in use at the University of Iowa in 1984 and was not incorporated into the study design. In the absence of chorioamnionitis, antibiotics were given only for the treatment of bacteriuria (symptomatic or asymptomatic) or group B streptococcal colonization of the cervix or vagina. Induction of labor for women in either group was reserved for clinical amnionitis or fetal distress. Intravenous tocolysis was begun if uterine activity exceeded three contractions per hour. This liberal ap-
218 Weiner, Renk, and Klugman
proach was applied in the hope of treating any uterine activity before it could become well established. Tocolysis was discontinued only for patient intolerance of the drug's side effects, cervical dilation >4 cm, or chorioamnionitis. Tocolytic agents used in our facility included ritodrine, terbutaline, and magnesium sulfate, the functional efficacy of which is similar.8· 9 In the tocolysis group, all but two women received a 13-mimetic initially and all but two of these received ritodrine. Drug infusion was increased rapidly until either uterine activity (as perceived by either the patient or on the external tocodynometer) was all but abolished, or the dose exceeded 450 µg/min for ritodrine or 20 µg/min for terbutalnie. Magnesium sulfate up to 4.5 gm/hr was added liberally if uterine activity could not be controlled by the 13-mimetic agent. After 24 hours of successful tocolysis, dosing with an oral 13-mimetic (usually terbutaline in the interest of cost containment) was begun and both the dose and dose interval were individualized to maintain the maternal heart rate >95 bpm. Recurrent episodes of labor were treated similarly. Adherence to the study protocol was ascertained and the data were collected contemporaneously by one of the authors (K. R.). Gestational age was based on the "best obstetric dates" -a combination of menstrual, historic, and objective criteria. If the postnatal pediatric estimate of gestational age (method of Ballard) differed by >3 weeks, the pediatric estimate was accepted as correct. Morbidity. Maternal morbidity was limited to infection. The patient's temperature was measured at least every 4 hours. The diagnosis of chorioamnionitis either before or during labor required pyrexia and uterine tenderness between contractions and was not made on the basis of laboratory tests alone. When possible, amniotic fluid was obtained on insertion of the intrauterine pressure catheter during labor and cultured. Febrile morbidity required two readings ;;o l 00° F after the first postpartum day. Specimens for anaerobic and aerobic cultures were obtained from the blood, urine, and uterus as part of the standard evaluation of a postpartum fever. The diagnosis of sepsis required positive blood cultures. In the absence of another explanation (e.g., urinary tract or upper respiratory infection), febrile morbidity plus uterine tenderness to palpation was coded as endometritis. Major neonatal complications included sepsis, probable sepsis, respiratory distress syndrome, and necrotizing enterocolitis. The diagnosis of sepsis required either a positive culture from a blood, urine, or cerebral spinal fluid specimen obtained within 6 hours of birth; a chest x-ray film consistent with pneumonia associated with a positive endotracheal aspirate obtained within 6 hours of birth; or a positive antibody test for group B streptococcus in a urine specimen obtained within 6
July 1988 Am J Obstet Gynecol
hours of birth. A neonate was classified as probably septic whenever there was laboratory (e.g., leukopenia, thrombocytopenia) or clinical (e.g., hypotension, tachypnea) evidence of infection that did not fulfill the above criteria. Respiratory distress syndrome was a clinical diagnosis based on the standard criteria. ' 0 Charges. Patient charges (mother and child, hospital, and physician), including all consultant and anesthesia fees, were determined after discharge from the Hospital Business and Accounts Department. These included costs subsequently borne by the State of Iowa Indigent Care Program and federally funded Medicaid. No inflationary adjustments in costs were made because a similar number of women were enrolled each of the 2 years. In the event the infant was transferred from the university hospital to a level II facility before discharge home, outlying physician and hospital costs were obtained by direct correspondence with the care provider. Similar numbers of infants in each treatment group were transferred to local hospitals. The cost per surviving infant (i.e., discharged home alive) by treatment group was defined as the aggregate maternal and neonatal medical cost divided by the number of survivors. For the post hoc analysis that included multiple gestations, the maternal cost for that pregnancy was multiplied by the number of fetuses to yield the equivalent maternal-fetal unit cost. Analysis. Results were assessed by subprograms from the Statistical Analysis System release 5. Assessment involved parametric and nonparametric tests with and without logarithmic conversion where appropriate. Post hoc analyses were performed on the following groups: excluding pregnancies concluded before neonatal viability (25 weeks' gestation), excluding pregnancies complicated by infection, dividing the study population by the gestational age at rupture of the membranes (<28 weeks, ;;.28 weeks), including all randomized patients, and including all randomized patients except those with multiple gestations. A p value ~0.05 was considered significant and p < 0.10 was considered near significant.
Results One hundred seventy-eight women were eligible to participate; 109 (61 %) gave consent and were randomized. Thirty-four patients were excluded from the initial analyses: 8 in whom the protocol was violated, 6 who never received tocolysis either because of chorioamnionitis or advanced cervical dilation, 8 who withdrew from the study before delivery (all in the tocolysis group), and 12 multiple gestations (eight in the bed rest group and four in the tocolysis group). The protocol continues for the latter group, which we consider an entity separate from the singleton gestation. The demographic profile of the remaining participants is listed
Aggressive tocolysis and PROM
Volume 159 Number I
Table I. Demographic profile
No. of patients Maternal age (yr) Gravidity Parity Abortions Gestation at ROM (wk) % Married % White % Tobacco users % High school education or higher % Medicaid or state aid % Prior stillbirth % Medical complications
Table IV. Neonatal results
Bed rest
Tocolysis
42 23.5 ± 4.4 2.1 ± 1.2 0.8 ± 1.0 0.3 ± 0.5 29.4 ± 3.2 57.l 85.7 54.8 69.0
33 25.7 ± 5.7* 2.3 ± 1.2 0.7 ± 0.9 0.5 ± 0.8 30.0 ± 3.2 62.5 90.3 33.3t 68.8
10.0 0.0 2.4
5.3 0.0 6.0
Table II. Cervical status
Cervical dilation (cm) Cervical effacement (%) Station
Bed rest
1.5 ± 1 1.5 ± 1 53 ± 30 49.8 ± 29
-2
Birth weight (gm) Umbilical vein pH Umbilical artery pH Sepsis(%) Probable sepsis (%) RDS(%) NEC(%) Neonatal deaths(%)
Bed rest
Tocolysis
1518 ± 563 7.35 ± 0.09 7.27 ± 0.09 3 (7.1) 17 (40.5) 22 (52.4) IO (23.8) 5 (11.9)
1648 ± 536 7.34 ± 0.08 7.25 ± 0.10 3 (9.))
9 15 6 3
(27.3) (45.4) (18.2) (9.1)
Data are mean ± SD. All differences nonsignificant. RDS, Respiratory distress syndrome; NEC, necrotizing enterocolitis.
Table V. Comparisons by selected intervals
Data are mean ± SD. ROM, Rupture of membranes. *p = 0.06. tp < 0.053.
Total
219
-3
Tocolysis
1.6 ± 1 57 ± 31 -1.8
From membrane rupture to delivery <24 hr <48 hr <72 hr <24 hr <48 hr <72 hr
Bed rest(%)
Tocolysis (%)
7.1 23.8 38.l 42.1 63.2
6.1 12.1 30.3 24.1 55.1 62.l
71.1
All differences nonsignificant.
Data are mean ± SD. All differences nonsignificant.
Table III. Maternal results
Gestation at delivery (wk) % Cesarean section % Maternal infection Hospital days before delivery
Bed rest
Tocolysis
30.l ± 3.3 16.7 26.2 5.2 ± 6.0
3.10 ± 2.9 12.1 27.3 6.7 ± 7.2
Data are mean ± SD. All differences nonsignificant.
in Table I. All subsequent analyses were based on the 75 remaining women. Women randomized to the tocolysis group were slightly older (p = 0.06) and less likely to smoke in excess of I 0 cigarettes per day (p = 0.053; Table I). There were no differences between women in the two groups in terms of either cervical status (Table II) or percentage of women in active labor on admission to the university hospital. Although 71 of the l 09 women randomized had regular uterine activity documented on admission, prodromal signs of membrane rupture were only infrequently noted by the patient. Regular uterine activity preceded rupture in 7.6%, cramping in 13.2%, and spotting in 9.4% of the women. More than 80% of women were unaware of any change suggesting rupture of the membranes was imminent. This .intimates that patient education on the warning signs of premature rupture of the membranes will not end the problem.
Seventy-nine women (72%) had a urine specific gravity > 1.005 at admission and received a fluid bolus. Of these 59 women were having regular uterine contractions when the bolus was begun; 42% of these stopped contracting with hydration alone. There was no difference between treatment groups in response to hydration. Most fetuses had a reactive heart rate pattern at admission (95%). Variable decelerations (consistent with intermittent umbilical cord compression) were common. Seventy-four percent had at least mild variable decelerations and 3.9% had at least one severe variable (<90 beats/ min for ;;,;45 seconds) deceleration. Approximately 50% ( 17) of women in the tocolysis group received two agents. The mean maximum dose of ritodrine was 225 µg/min and that of magnesium sulfate was 2.9 gm. Despite the aggressiveness of drug therapy, there were no major complications attributable to the tocolytic agents during the study. No significant differences between the treatment groups were observed for the interval from membrane rupture to delivery, proportion of cesarean delivery, or maternal infectious morbidity (Table Ill). There were also no significant or near significant differences in neonatal outcome for birth weight, umbilical cord blood gases, sepsis or probable sepsis, respiratory distress syndrome, necrotizing enterocolitis, or neonatal mortality (Table IV). However, there was a near significant increase in
220
Weiner, Renk, and Klugman
July 1988 Am J Obstet Gynecol
Table VI. Interval from onset of regular contractions to delivery and cost per survivor by subgroup Bed rest
Tocolysis
I
Cost per suroivor ($)
Interoal (hr)
58 ± 75* 62 ± 77
51,145 ± 52,196 43,158 ± 37,116
117 ± 174* 96 ± 141
48,215 ± 40,618 48,232 ± 44,524
69 ± 79
39,156 ± 34,364
108 ± 159
33,759 ± 32,286
58 ± 75*
51,145 ± 52,196
129 ± 194*
49,091 ± 36,975
Group
Interoal (hr)
All patients All patients except multiple births All patients except delivery <25 wk All patients except protocol violation
J
Cost per suroivor ($)
Data are mean ± SD.
*p < 0.05.
Table VII. Interval from onset of regular uterine contractions gestational age at membrane rupture
to delivery and cost per survivor by
<28 wk Bed rest
No. of patients Interval (hr) Neonatal deaths Cost per survivor ($)
*p
= 0.05.
tp
=
12 53.4 ± 87* 4
82,871 ± 30,650t
>28 wk
I
Tocolysis
Bed rest
5 232.8 ± 312 2 118,206 ± 42,172
24 68.5 ± 70
I
Tocolysis
25 78.7 ± 93
1
I
23,302 ± 22,779
22,670 ± 15, 195
O.D7.
the duration of intrauterine time after the onset of regular contractions in the tocolysis group (bed rest, 62.l ± 77 hours; tocolysis, 105.2 ± 157 hours; p = 0.06). The percent of women delivered 24, 48, and 72 hours after rupture of membranes and the onset of regular uterine contractions is shown in Table V. There were no significant differences between the two groups. A cost analysis was performed in case small benefits otherwise missed by the above analysis might be reflected in a reduction of the overall medical cost. Women who received tocolysis had a near-significant increase in hospital costs (bed rest, $4399 ± $3266; tocolysis, $4883 ± $2491; p = 0.076) and a significant increase in physician fees (bed rest, $939 ± $863; tocolysis, $1419 ± $1003;p = 0.005). There were no significant cost differences between treatment groups for the neonates for either the hospital (bed rest, $26,839 ± $30,651; tocolysis, $23,900 ± $33,728; p = 0.602) or physician (bed rest $5674 ± $6009; tocolysis, $4882 ± $6779; p = 0.418). The overall cost per survivor for the tocolysis group was suggestive of benefit but not significantly reduced (bed rest, $43, 158 ± $37,116; tocolysis, $38,593 ± $40,887; p = 0.4453). Should this percent reduction in costs hold, we calculate that a study population in excess of 3000 women with premature rupture of the membranes <34 weeks' gestation would be needed to reach a significant difference.
In an attempt to understand further the absolute reduction in costs per surviving neonate and the nearsignificant increase in intrauterine time achieved by tocolysis after the onset of regular uterine contractions, the above analyses were repeated with these modifications: first, excluding pregnancies complicated by either maternal or neonatal infection; second, excluding all pregnancies delivered before viability; third, including all randomized patients; fourth, excluding only multiple gestations (Table VI); and fifth, by dividing the 75 women according to gestational age at rupture of the membranes <28 or ~28 weeks' gestation. The results were similar to those already reported in Tables I to II I for each of the first four post hoc analyses. Only the last analysis yielded findings of interest (Table VII). The distribution of patients into the treatment arms was skewed in the group with delivery <28 weeks' gestation (bed rest, 12; tocolysis, 5). The added intrauterine time and the reduction in costs per survivor associated with tocolysis was wholly explainable by this unequal distribution. However, there was an absolute reduction in the cost per survivor (2.7%) and an increase in intrauterine time after the onset of regular contractions (15%) associated with tocolysis when delivery occurred after 28 weeks' gestation. However, achievement of an 80% power for these values in this subset would require 6200 patients with rupture of the membranes <34 weeks.
Volume 159 Number I
We considered the possibility that a clinically silent chorioamnionitis before birth might render the uterus unresponsive to tocolysis and mask a significant effect of tocolysis in uninfected patients. After excluding pregnancies of women with evidence of antepartum chorioamnionitis or postpartum endometritis or neonates with confirmed septicemia (16 in the bed rest group and 11 in the tocolysis group), there was a nearsignificant increase in the interval from onset of regular contractions to delivery in the tocolysis group (bed rest, 60.4 ± 79.2 hours; tocolysis, 116.5 ± 176.5 hours; p = 0.06). Although there was no significant difference in the total medical costs per survivor (bed rest, $45,302 ± $40,025; tocolysis, $37,413 ± $44,732;p = 0.390), the difference does amount to a 17% reduction in costs. Six hundred patients would be required for the detection of this difference at the 0.05 level.
Comment Preterm premature rupture of the membranes is a major cause of perinatal morbidity and mortality. Numerous investigators have concluded that the risk of neonatal respiratory distress outweighs the risk of perinatal or maternal sepsis in the absence of proved fetal lung maturity.'·'· 11 · " Presently, bed rest under close medical supervision rather than immediate induction of labor is the standard management of the patient with preterm premature rupture of the membranes who is not in labor. Management of the associated premature labor, which occurs in 80% of women within 7 days of preterm premature rupture of the membranes, 12 has not been adequately studied. The use of tocolysis for preterm labor associated with intact membranes does improve neonatal outcome.5 A similar analysis has not been used with the patient with preterm premature rupture of the membranes. Because labor may be a sign of incipient chorioamnionitis, and chorioamnionitis is associated with an increase in perinatal morbidity and mortality, many physicians refrain from tocolysis, concerned that any prolongation of intrauterine time would increase infectious complications. However, if tocolysis were otherwise safe and effective, this practice would deny those benefits to the majority of patients who are free of infection. Some investigators have advocated tocolysis only for an interval long enough to administer corticosteroids for the enhancement of fetal lung maturity (48 hours) and then delivery of the patient. 1• Unfortunately, the efficacy of maternal corticosteroid administration after rupture of the membranes is in doubt." Much of the confusion undoubtedly stems from the paucity of information concerning either the efficacy or safety of tocolysis for labor after preterm premature rupture of the membranes. At the study's inception, we could identify only one prospective, randomized
Aggressive tocolysis and PROM 221
trial of tocolysis for labor after preterm premature ru pture of the membranes. Although the number of pregnancies studied was small (n = 30), those investigators 6 could not demonstrate an increase in infection with tocolysis but they did find an increase in the percentage of women delivered ~24 hours after rupture of the membranes. We hypothesized that the benefits of tocolysis might have been clear if their treatment protocol had been more aggressive and their analysis had involved multiple neonatal parameters. The present investigation addresses not only the therapeutic efficacy and safety of tocolysis after preterm premature rupture of the membranes, but also, for the first time, its cost efficacy. Aggressive tocolysis initiated because of regular uterine contractions (in contrast to awaiting cervical dilation) failed to significantly increase overall intrauterine time above that achieved by bed rest. However, the intrauterine time achieved after the onset of regular uterine contractions was increased by almost 48 hours when tocolysis was used. There was no evidence that tocolysis increased either maternal or perinatal infectious complications. Furthermore, the total medical cost per surviving infant was reduced in association with tocolysis. Unfortunately, these salutary effects were likely an artifact produced by an unequal distribution between treatment groups of infants delivered at <28 weeks' gestation. These infants had significantly higher costs per survivor (all pregnancies delivered <28 weeks, $93,264, versus $23,284 if delivered after 28 weeks). Despite a significant increase in intrauterine time achieved with tocolysis after the onset of regular contractions in the group delivering at <28 weeks' gestation, there was a near-significant increase in cost per survivor (p = 0.07). Whether this trend is real or a skewing brought about by the small population and unequal distribution is unknown. In contrast to the infants delivered at <28 weeks' gestation, there were no significant differences between treatment groups with delivery after 28 weeks. Although there was an absolute reduction in the cost per survivor associated with tocolysis (2. 7% ), power analysis revealed that the difference would require a study population in excess of 6200 women with rupture of the membranes between 28 and 34 weeks' gestation. It seems unlikely tocolysis offers any clinical benefit should rupture of the membranes occur after 28 weeks. Deleting from analysis those pregnancies complicated by either maternal or neonatal infectious morbidity failed to alter the aforenoted pattern. The cost per neonatal survivor was reduced by 17% in association with tocolysis, but confirmation of this difference would require at least a trebling of the sample size. Thus the lack of a benefit from tocolysis for the treatment of labor associated with premature rupture of the
222
Weiner, Renk, and Klugman
membranes cannot be attributed solely to a few patients in whom tocolysis failed because of infection. We conclude that treatment of labor after preterm premature rupture of the membranes does not improve perinatal outcome after 28 weeks' gestation. Because therapy itself entails major maternal risks such as pulmonary edema, tocolysis should not be used for these patients. At <28 weeks' gestation tocolysis may increase intrauterine time, but no benefit could be demonstrated. The use of tocolysis in these very preterm pregnancies would not in our population be interdicted out of concern for increasing infectious morbidity, but its benefit remains unclear. REFERENCES I. Druzin ML, Toth M, Ledger WJ. Nonintervention in pre-
2. 3.
4.
5.
mature rupture of the amniotic membranes. Surg Gynecol Obstet 1986; 163:5-8. Fayez JA, Hasan AA, Jonas HS, Miller GL. Management of premature rupture of the membranes. Obstet Gynecol 1978;52:17-21. Daikoku NH, Kaltreider DJ.Johnson TRB,JohnsonJWC, Simmons MA. Premature rupture of membranes and preterm labor: Neonatal infection and perinatal mortality risks. Obstet Gynecol 1981;58:417-25. Gibbs RS, Sweet RL. Maternal and fetal infections. In: Creasy RK, Resnik R, eds. Maternal fetal medicine: Principles and practice. Philadelphia: W. B. Saunders, 1984: 607-13. Korenbrot CC, Aalta LH, Laros RK. The cost effectiveness of stopping preterm labor with beta-adrenergic treatment. N Engl J Med 1984;310:691-6.
July 1988 Am J Obstet Gynecol
6. Christensen KK, Ingemarsson I, Leideman T, Solum H, Svenningsen N. Effect of ritodrine on labor after premature rupture of the membranes. Obstet Gynecol 1980;55: 187-90. 7. Garite TJ, Keegan KA, Freeman RK, Nageotte MP. A randomized trial of ritodrine tocolysis versus expectant management in patients with premature rupture of membranes at 25 to 30 weeks' gestation. AMJ OBSTET GYNECOL 1987; 157:388-93. 8. Petrie RH. Tocolysis using magnesium sulfate. Semin Perinatol 1981 ;5:266-9. 9. Caritis SN, Toig G, Heddinger LA, Ashmead G. A doubleblind study comparing ritodrine and terbutaline in the treatment of preterm labor. AM J OBSTET GYNECOL 1984; 150:7-14. JO. Stahlman MT. Acute respiratory disorders of the newborn. In: Avery GB, ed. Neonatology-pathophysiology and management of the newborn. Philadelphia: JB Lippincott, 1981 :507-13. 11. Varner MW, Galask RP. Conservative management of premature rupture of the membranes. AM J 0BSTET GYNECOL 1981; 140:39-45. 12. Kappy KA, Cetrulo CL, Knuppel RA, et al. Premature rupture of the membranes: A conservative approach. AM J 0BSTET GYNECOL 1979; 134:655-61. 13. Garite TJ. Premature rupture of the membranes: The enigma of the obstetrician. AM J OBSTET GYNECOL 1985; 151: 1001-5. 14. Mead PB. Management of the patient with premature rupture of the membranes. Clin Perinatol 1980;7:243-55. 15. Garite TJ, Freeman RK, Linzey EM, Braly PS, Dorchester WL. Prospective randomized study of corticosteroids in the management of premature rupture of the membranes and the premature gestation. AM J 0BSTET GYNECOL 1981; 141 :508-15.
Bound volumes available to subscribers Bound volumes of the AMERICAN JOURNAL OF OBSTETRICS AND GYNECOLOGY are available to subscribers (only) for the 1988 issues from the Publisher, at a cost of $54.00 ($77.00 international) for Vol. 158 (January-] une) and Vol. 159 (j uly-December). Shipping charges are included. Each bound volume contains a subject and author index and all advertising is removed. Copies are shipped within 60 days after publication of the last issue in the volume. The binding is durable buckram with the JOURNAL name, volume number, and year stamped in gold on the spine. Payment must accompany all orders. Contact The C. V. Mosby Company, Circulation Department, 11830 Westline Industrial Drive, St. Louis, Missouri 63146, USA; phone (800) 325-41 77, ext. 351. Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular JOURNAL subscription.