Factors determining reopening of the ductus arteriosus after successful clinical closure with indomethacin Hali Weiss, MD, Bruce Cooper, PhD, Michael Brook, MD, Mureen Schlueter, BS, a n d Ronald Clyman, MD From the Cardiovascular Research Institute, Department of Pediatrics, Universityof California San Francisco, and the California School of Professional Psychology, Alameda
Objective: To examine the role of ductus arteriosus (DA) constriction and Ioss of luminal blood flow in producing permanent closure of the DA in human infants. Methods: We studied 77 newborn infants (gestational age, 23 to 33 weeks) with asymptomatic patent ductus arteriosus (PDA), who had "complete clinical closure" (defined as the disappearance of all PDA signs) after treatment with indomethacin (three doses within 36 hours). All infants had an echocardiogram 24 to 36 hours after the last dose of indomethacin. They were then followed for the development of ductus reopening. Results: Despite the absence of clinical signs, 18 (23%) of 77 infants still had some residual luminal blood flow according to their echocardiograms. The failure to obliterate luminal blood fiow completely was directly related to the infant's postnatal age when treatment was begun and to the amount of fluid administered before treatment. Subsequently the DA reopened in 16 (21%)of 77 infants. As predicted, infants with residual luminal blood flow after indomethacin treatment had a higher rate of subsequent clinical reopening than did those with no luminal flow. In addition, immature infants had a significantly higher reopening rate than did more mature infants. The increased risk of reopening in immature infants was seen even when indomethacin caused complete obliteration of ductus luminal blood flow. Conclusion: The DA of immature infants is resistant to the constriction-induced ischemic damage that is necessary for subsequent permanent closure of the vessel. (J PEDIATR1995;127:466-71)
In most term infants the DA constficts spontaneously during the first 48 hours after birth. This leads to a decrease in ducms luminal blood flow, which in turn causes ischemia of the inner muscle wall. I The ischemic insult inhibits the ability of the ductus to respond to further vasodilator stimulafion. 1
As a result, ductus-related symptoms and signs rarely reappear, 2 and the ductus is irreversibly closed. Previous studies in term lambs have shown that the failure of the ductus to respond to subsequent vasoactive stimulation is directly related to the loss of ductus luminal blood flow. 1
Supported in part by a grant from the U.S. Public Health Service, National Heart, Lung, and Blood Institute (HL 46691), and by a gift from the Perinatal Associates Research Foundation. Submitted for publication Feb. 6, 1995; accepted April 7, 1995. Reprint requests: Ronald Clyman, MD, Box 0544, HSE 1492, University of California School of Medicine, San Francisco, CA 94143. Copyright © 1995 by Mosby-Year Book, Inc. 002-3476/95/$5.00 + 0 9¤23¤65476
I
466
DA PDA
Ductusarteriosus Patentductus arteriosus
[
On the other hand, in preterm infants, who often require treatment with indomethacin to close the ductus, irreversible ductus closure does not necessarily develop. Although indomethacin induces ducms closure in 85% of the premature infants in whom it is used, reappearance of clinical signs,
The Journal of Pediatrics Volume 127, Number 3
Weiss et al.
467
Table I. Antenatal and postnatal determinants of persistent ductus luminal blood flow after treatment with indomethacin Luminal blood flow
Neonatal factors Gestational age (wk) Birth weight (gm) Gender (% male) PIP (cm H20) FIO2 Number of signs Fluid intake (cc/kg per day) Age at treatment (days) RDS (%) Surfactant treatment (%) Maternal factors Betamethasone (%) MgSO4 (%) Preeclampsia (%) Diabetes (%) PROM (%) Chorioamnionitis (%)
Absent (n = 59)
Persistent (n = 18)
Correlation coefficient
p
27.8 _+2.5 1067 -+ 366 52 18 _+5 0.33 _+0.15 1.3 _+0.6 111 _+38 3.7 -+ 3.0 92 87
27.6 _+2.4 1108 _+388 33 18 _+7 0.30 _+0.14 1.5 _+0.8 143 _+43 5.4 _+3.6 95 89
-0.02' 0.07* 0.17~ -0.01" -0.08* -0.01" 0.395 0.40:) 0.05~ 0.03~
0.84 0.56 0.14 0.91 0.47 0.89 0.0005§ 0.0003§ 0.69 0.79
63 53 25 10 15 20
50 44 6 11 27 11
0.044 0.06t 0.22"~ 0.04"~ 0.12"~ 0.11~
0.75 0.59 0.06 0.72 0.29 0.32
Values with plus-minussigns are expressed as mean -- SD. Fto2, Fractionof inspiredoxygen at 24 hours of life; Fluid intake, fluid intake 24 hours before appearanceof symptomatology;MgS04, magnesiumsulfate tocolysis; P1P, peak inspiratorypressure at 24 hours after birth; PROM, prolongedrupmre of membranes;RDS, respiratory distress syndrome. *Pearson correlation,correlaUoncoefficient. "~Chi-squaretest, phi coefficient. :~Partialcorrelationcontrollingfor gestationalage in model, correlationcoefficient. §Staüstically significant(p <0.05). or ductus reopening, occurs in 20% to 100% of the treated infants. 3-12 Whether the increased incidence of ducms reopening in indomethacin-treated premature infants is the result of ineffective ductus closure,1 indomethacin itself, 13 or some other phenomenon attributable to prematurity 1 remains unclear. On the basis of the events that lead tO irreversible closure in the term infant, we hypothesized that ductus reopening after indomethacin- induced clinical closure is caused by persistent ductus luminal blood flow that is beyond the limits of clinical detection. This smdy tested that hypothesis. METHODS Pafients eügible for this study were premature infants admitred to the William H. Tooley Intensive Care Nursery between January 1990 and July 1994, who had a symptomafic p D A within the first 14 days after birth and who were treated with indomethacin. The diagnosis of PDA was made in any infant who had at least one of three cünical signs--murmur, increased precordial acfivity, and bounding pulses--and in whom the diagnosis could be confirmed by echocardiography. Two-dimensional and color Doppler flow-mapping echocardiography was performed with a 5 or 7 MHz h-ansducer interlaced with a model 128 XP echocardiographic system (Acuson, Inc., Mountain View, Calif.). Color Dop-
pier measurements of flow within the DA and two-dimensional measurements of ductal size were made, in addition to calculation of left atriallaortic ratio and evaluation of retrograde flow within the descending aorta. Both continuouswave and pulsed-wave Doppler probes were used to confirm the presence, direction, and timing of ductus flow. All images were recorded on super VHS videotape. All patients were treated with three doses of indomethacin (0.2 mg/kg, 0.1 mg/kg, and 0.1 mg/kg) administered during a 36-hour period (at 0, 12, and 36 hours, respectively). Infants were examined frequently for clinical signs of a PDA. The clinical response to indomethacin was based on the disappearance of ductus-related signs within 48 hours of the start of indomethacin therapy. Infants were considered to have had "complete clinical closure" if all the signs disappeared; to have had a "partial response" if the number of signs was reduced, with some signs still remainingi and to be "nonresponders" if there was no change in the quality of the existing signs. During the study period, 60% of the infants treated with indomethacin had complete clinical closure, 25% were partial responders, and 15% were nortresponders. All partial responders and nonresponders continued to have echocardiographic evidence of persistent ductus luminal blood flow. To test the hypothesis that ductus reopening after in-
468
Weiss et al.
The Journal of Pediatrics September 1995
T a b l e II. Antenatal and postnatal determinants of ductus reopening Permanent ductus status
Neonatal factors GA -<26 wk (%) Birth weight (gm) Gender (% male) Luminal flow (%) PIP (cm H20) F~o2 Fluid intake (ml/kg per day) Age at treatment (days) Number of signs RDS (%) Surfactant treatment (%) Matemal factors Betamethasone (%) MgSO4 Preeclampsia (%) Diabetes (%) PROM (%) Chofioamnionitis (%)
Closed (n = 61)
Reopened (n = 16)
Test statistic
p
31 1098 _+357 46 15 18 +- 7 0.33 +_0.14 107 _+43 3.3 +_2.8 1.7 +_0.8 93 87
69 942 + 374 56 56 19 -+ 3 0.31 + 0.17 131 + 49 4.0 _+2.3 1.2 _+0.7 88 88
4.61' -0.02~ 0.545 11.36§ 0.04t -0.09~ 0.18t 0.09t -0.20~ 0.56:) 0.004:)
0.03* 0.88 0.46 0.0008** 0.76 0.47 0.11 0.42 0.08 0.45 0.95
61 49 23 11 18 16
56 56 19 0 19 25
0.37:~ 0.25:) 0.13:) 3.44:~ 0.004~ 0.595
0.54 0.61 0.71 0.06 0.94 0.44
Values with plus-minus signs are expressed as mean _+SD. FIo2, Fraction of inspired oxygen at 24 hours of life;fluid intake, fluid intake 24 hours before appearance of symptomatology;GA, gestaUonalage; luminalflow, persistent luminal blood flow after indomethacintreatment; MgS04, magnesiumsulfate tocolysis;PIP, peak inspiratory pressure at 24 hours of life; PROM, prolonged rupture of membranes; RDS, respiratory distress syndrome. *Logisfic regression with persistent luminal blood flow in model, likelihoodrafio chi-square. -~Partial correlation with gestational age in mode1, correlation coefficient. ~:Chi-square test, likelihoodratio chi-square. §Logistic regression with gestational age in model, likelihoodraUo chi-sfluare. **Statisücally significant (p <0.05).
domethacin-induced closure is related to persistent luminal blood flow, we analyzed what happened to the group of infants considered to have had complete clinical closure. During the study period, a follow-up echocardiogram was performed 24 to 36 hours after the rinal indomethacin dose to document the presence or absence of persistent luminal blood flow when, according the clinical critefia, the ductus was thought to be closed. Luminal blood flow was considered to be absent if no retrograde color Doppler flow was present either in the region of the ductus or in the pulmonary artery and normal anterograde diastolic flow was present in the descending aorta in the region of the ductus. In no instance was luminal blood flow detected by pulsed Doppler if color Doppler failed to detect its presence. I f any color Doppler flow was seen in the region of the ductus, the ductus was considered to have persistent flow through its lumen. Infants were then followed for the development of clinical findings suggestive of ductus reopening. If any recurred, a third echocardiogram was performed. If this documented the presence of both increased ductus diameter and luminal blood flow, the ductus was considered to have reopened, and the infant either was treated with a second course of
indomethacin or underwent surgical ligation of the ductus. Infants whose ductus-related signs never reappeared did not have any fimher echocardiograms and were considered to have achieved permanent closure of the DA. Statistical analysis. Results are presented as means +_ standard deviations, percentages, and correlation coefficients, depending on levels of measurement. Intergroup differences were evaluated with analyses of variance for continuous variables, Marm-Whitney tests when variables were converted to ranks, and likelihood-ratio chi-square analyses for dichotomous dependent variables. We used logistic regression analysis or partial correlations to evaluate the relationships of specific predictors, controlling for covariates. The significance level for hypothesis tests was p <0.05.
RESULTS During the study period, we identified 77 infants whose ductus had closed clinically as indicated by the complete disappearance of clinical signs after a three-dose course of indomethacin. O f the 77 infants with a clinically closed ductus, 59 (77%) had no residual luminal blood flow on the second echocardiogram obtained within 36 hours of the fi-
The Journal of Pediatrics Volume 127, Number 3
Weiss et al.
00
1
469
4
90 80 70 60 50
5
-o-Partlal elosure - I - Complete closure
2
40 30 20 10 0
i
-
< 24
i
|
25
26
~
27
.
> 28
Gestational Age (weeks)
Figure. Incidence of ductus reopening as a function of gestational age and initial response to indomethacin. Numbers on graph represent the total number of infants with partial or c0mplete closure at each given gestational age. Total number = 77 infants. nal dose of indomethacin. Eighteen infants (23%) continued to have echocardiographic evidence of persistent, diminished luminal blood flow despite the absence of any clinical signs of a PDA. We looked at a number of antenatal and postnatal factors that might determine whether an infant had absent or persistent luminal blood flow after treatment with indomethacin (Table I). We found that the later an infant was treated with indomethacin, the more likely he or she was to have persistent luminal blood flow. Similarly, the more fluid an infant received in the 24 hours preceding the appearance of ductus-related signs, the more likely he or she was to have persistent luminal blood flow on an echocardiogram. Both of these significant predictors of persistent luminal blood flow were independent of gestational age, and each was independent of the other. Despite initial clinical closure with indomethacin, in 16 (21%) of the 77 infants the DA subsequently reopened and required further treatment. This observation was based on the development of clinical signs and evidence of increased ductus diameter and luminal blood flow on the third echocardiogram. Of the 77 infants, 61 (79%) had no recurrence of ductus-related signs during the remainder of their hospital course and were considered to have achieved permanent ductus closure. Consistent with our initial hypothesis, infants with echocardiographic evidence of residual luminal blood flow after treatment with indomethacin had a significantly higher rate of ductus reopening (9/18, 50%) than did infants with no demonstrable luminal blood flow (7/59, 12%; p = 0.001) (Table II). Although infants with persistent luminal blood flow had a higher rate of reopening than did infants with
no flow, there did not appear to be any relationship between the time it took for the ductus to reopen after initial closure and the presence of luminal blood flow. Reopening occurred 17 - 11 days after initial clinical closure in infants with persistent luminal blood flow, whereas it occurred 13 _+ 7 days after initial closure in those with no luminal flow. We also found gestational age to be a strong predictor of ductus reopening. The most immature infants studied, those born at -<26 weeks of gestation, had a significantly higher rate of subsequent ductus reopening (11/30, 37%) than did infants bom at 27 to 33 weeks of gestation (5/47, 11%; p = 0.045). Using a logistic regression model, we found that the significant effect of persistent luminal blood flow on the rate of reopening was independent of the infant's gestational age (Table II). Similarly, we found that the significant effect of immature gestational age on the rate of reopening was independent of the presence of luminal blood flow after treatment with indomethacin (Table 1]; Figure). None of the other potential prenatal factors had any effect on subsequent ductus reopening (Table II). We also were unable to demonstrate any relationship between ductus reopening and indicators of respiratory distress, such as administration of surfactant, peak inspiratory pressure at 24 hours of age, and fraction of inspired oxygen at 24 hours of age. Finally, we looked at the amount of fluid the infant received in the 24 hours before the appearance of ductus-related signs, the infant's age at the time of treatment with indomethacin, and the number of symptoms atthe time of initial recognition (one to three) but were unable to find any relationship with subsequent ductus reopening.
470
Weiss et al.
In this group of infants, ductus reopening was a strong predictor of subsequent chronic lung disease (defined as a requirement for supplemental oxygen for >28 days). In in2 fants whose ductus reopened, chronic lung disease developed at a significantly higher rate (88%) than in those whose ductus remained closed (48%; p : 0.003). Using a logistic regression model, we found that the significant effect of ductus reopening on the incidence of chronic lung disease was independent of the infant's gestational age (p = 0.01). We could not determine whetläer the increased incidence of chronic lung disease was caused by the recurrence of the PDA or whether ductus reopening was just another indicator of the type of infant in whom chronic lung disease was likely to develop. DISCUSSION All infants in our study were considered to have had a successful initial response to indomethacin as indicated by the disappearance of c!inical signs. However, 23% of the infants continued to have echocardiographic evidence of luminal blood flow. This is not surprising, inasmuch as careful, frequent physical examinations have been shown to miss 20% to 28% of PDAs that are clinically silent.2' 14 We found that fluid intake and postnatal age at the time of treatment were the only factors that predicted the persistence of ductus luminal blood flow after indomethacin administration. Fluid intake in the 24 hours preceding treatment was directly related to the persistence of luminal blood flow after treatment. This observation is consistent with earlier studies that found an association between the volume of fluid intake and the development of PDA-related symptoms in premature infants. 15-17In this study, we found that persistent luminal blood flow was seen most commonly in infants who were treated with indomethacin at a later postnatal age. This observation is consistent with earlier studies that found that the clinical closure rate after indomethacin treatment was a function of an infant's postnatal age at the start of treatment. 18 We previously observed that the antenatal use of betamethasone was associated with a decreased incidence of PDA. 19 This appears to be related to the ability of glucocorticoids to alter the sensitivity of the ductus to dilator prostaglandins.20, 21 Our results indicate that if antenatal glucocorticoid administration falls to prevent the development of asymptomatic PDA, so that the infant requires indomethacin treatment, it also falls to protect the infant from ductus reopening after indomethacin-induced closure. Out findings are consistent with out earlier animal studies, which demonstrated that persistent luminal blood flow determines the future responsiveness of the ductus to vasodilator prostaglandins. 1 We hypothesize that the ductus arteriosus of an infant with persistent residual luminal blood
The Journal of Pediatrics September 1995
flow continues to be responsive to prostaglandins and other vasodilators and is therefore more likely to reopen at a later time. This hypothesis is supported by our previous finding that 67% of ductus that reopen after indomethacin-induced closure can be closed successfully by a repeated course of indomethacin. Previously we reported that premature lambs, like premature human infants, have a limited ability to obliterate ducms luminal blood flow after birth. 22 Even in the unusual premature lamb, in which the reduction in luminal blood flow was as great as that in the term lamb, the DA continued to be responsive to prostaglandins, oxygen, and indomethacin.1, 22 In this study, we found a strong relationship between gestational age and the incidence of ductus reopening in human infants. Although we hypothesized that premamre infants were at increased risk of ductus reopening primarily because of their limited abifity to obliterate ducms luminal blood flow completely, this study demonstrates that the significant relationship of ductus reopening to gestational age is independent of the amount of luminal blood flow. Both 0ur human and out animal studies indicate that some other property of the premature infant's ductus prevents it from losing its responsiveness to prostaglandins and predisposes it to subsequent reopening. Whether this is related to factors that make the immature ductus more resistant to ischemia after loss of luminal flow must await fumre experimental studies. We conclude that among infants whose gestational age is 26 weeks or more, the most likely cause of subsequent ducms reopening is failure of the initial three doses of indomethacin to produce complete obliteration of ductus luminal blood flow. Of the 10 infants in this age group in whom the ductus reopened, 7 still had persistent luminal blood flow on echocardiogram and were therefore predisposed to reopening. Whether a longer course of treatment with indomethacin would produce a more complete initial closure and therefore would decrease the risk of reopening is a question that remains unanswered. Earlier use of indomethacin might decrease the number of ductus that subsequently reopen, becuase early treatment is associated with a higher rate of loss of luminal blood flow. On the other hand, when the ductus reopens in infants who are born at or before 25 weeks of gestation, it is not necessarily associated with persistence of luminal blood flow; four of the six infants in this age group whose ductus reopened had no residual luminal blood flow after the initial course of indomethacin. It appears that other factors associated with prematurity may be primarily responsible for the high rate of reopening in this extremely immature age group, We thank Mr. Paul Sagan for his skillful editorial work on this manuscript.
The Journal of Pediatrics Volume 127, Number 3
W e ß s et al.
REFERENCES
1. Clyman RI, Mauray F, Roman C, Heymann MA, Payne B. Factors determining the loss of ductus artefiosus responsiveness to prostaglandin E2. Circulation 1983;68:433-6. 2. Roberson DA, Silverman NH. Color Doppler flow mapping of the patent ductus arteriosus in very low birthweight neonates: echocardiographic and clinical findings. Pediatr Cardiol 1994; 15:219-24. 3. Merritt TA, Harris JP, Roghmann K, et al. Early closure of the patent ductus arteriosus in very low birthweight infants: a controlled trial. J PEDIATR1981;99:281-6. 4. Yeh TF, Wilks A, Betkerur M, Lilien L, Pildes RS. Furosemide prevents the renal side effects of indomethacin therapy in premature infants with patent ductus arteriosus. J PEDIATR1982; 101:433-7. 5. Mahony L, Carnero V, Brett C, Heymann MA, Clyman RI. Prophylactic indomethacin therapy for patent ductus arteriosus in very-low-birth-weight infants. N Engl J Med 1982;306:50610. 6. Gersony WM, Peckham G J, Ellison RC, Miettinen OS, Nadas AS. Effects of indomethacin in premature infants with patent ductus arteriosus: results of a national collaborative study. J PEDbXa'R 1983;102:895-906. 7. Ivey HH, Kattwinkel J, Park TS, Krovetz LJ. Failure of indomethacin to close patent ductus arteriosus in infants weighing under 1000 g. Br Heart J 1979;41:304-7. 8. Brash AR, Hickey DE, Graham TP, Oates JA, Cotton RB. Pharmacokinetics of indomethacin in the neonate. N Engl J Med 1981;305:67-72. 9. Kaapa P, Lanning P, Koivisto M. Early closure of patent ducms arteriosus with indomethacin in preterm infants with idiopathic respiratory distress syndrome. Acta Paediatr Scand 1983;72:179-84. 10. Rudd P, Montanez P, Hallidie-Smith K, Silverman M. Indomethacin treatment for patent ductus arteriosus in very low birthweight infants: double blind trial. Arch Dis Child 1983; 58:267-70. 11. Clyman RI, Manray F, Roman C, Heymann MA, Payne B. El-
12.
13.
14.
15.
16. 17.
18. 19.
20.
21.
22.
471
fect of gestational age on ductus arteriosus response to circulating prostaglandin E» J PEDIATR1983;102:907-11. Mellander M, Leheup B, Lindstrom DP, et al. Recurrence of symptomatic patent ductus arteriosus in extremely premature infants, treated with indomethacin. J l:'zD~r~ 1984;105:13843. Koppel R, Rabinovitch M. Regulation of fetal lamb ductus arteriosus smooth muscle cell migration by indomethacin and dexarnethasone. Pediatr Res 1993;33:352-8. Valdes-Cruz LM, Dudell GG. Specificity and accuracy of echocardiographic and clinical criteria for diagnosis of patent ductus arteriosus in fluid-restricted infants. J PEDIA~ 1981; 98:298-305. Bell EF, Warburton D, Stonestreet B, Oh W. Effect of fluid administration on the development of symptomatic patent ducms arteriosus and congestive heart failure in premature infants. N Engl J Med 1980;302:598-604. Brown E. Increased risk of bronchopulmonary dysplasia in infants with patent ductus arteriosus. J PEDIATR1979;95:865-6. Stevenson JG. Fluid administration in the association of patent ductus arteriosus complicating respiratory distress syndrome. J PZD~ATR1977;90:257-61. Firth J, Pickering D. Timing of indomethacin therapy in persistent ductus. Lancet 1980;I1(8186):144. Clyman RI, Ballard PL, Sniderman S, et al. Prenatal administration of betamethasone for prevention of patent ducms arteriosus. J PEDIATR1981;98:123-6. Clyman RI, Manray F, Roman C, et al. Effects of antenatal glucocorticoid administration on the ductus arteriosus of preterm lambs. Am J Physiol 1981;241:H415-20. Clyman RI, Mauray F, Roman C, Rudolph AM, Heymann MA. Glucocorficoids alter the sensitivity of the lamb ductus arteriosus to prostaglandin F4. J PEDIATR1981;98:126-8. Clyman RI, Campbell D, Heymama MA, Manray F. Persistent responsiveness of the neonatal ductus arteriosus in immamre lambs: a possible cause for reopening of patent ductus arteriosus after indomethacin induced closure. Circulätion 1985; 71:141-5.