October 1978
The Journal o f P E D 1 A T R 1 C S
647
Randomized trial of early closure of symptomatic patent ductus arteriosus in small preterm infants As a result of randomized assignment. 15 preterm infants weighing 1,500 gm or less at birth and who had a symptomatic PDA were treated according to a medical management protocol, and ten according to an earl)' surgical closure protocol. AII infants required mechanical ventilation at the time of stud). entr); which was one week after birth. Birth weight, gestational age, age at onset of congestive failure, age at stud), entr); and the initial morbidit)' of members o f the two groups were similar. The nine surviving infants managed according to the surgical closure protocol were weaned from mechanical ventilation sooner, had a decreased need for digoxin and fttrosemide, achieved gastrointestinal fttnction sooner, and had a smaller hospital bill than the 12 survivors of the medical management group. These results indicate that infants with a s)'mptomatic PDA still requiring mechanical ventilation at one week after birth will benefit from surgical closure o f the ductus at that time.
Robert B. Cotton, M.D., Mildred T. Stahlman, M.D., Harvey W. Bender, M.D.,
Thomas P. Graham, M.D., W. Z. Catterton, M.D.,* and llya Kovar, M.D, Nashville, TemL
CONGESTIVE HEART FAILURE with pulmonary edema due to left-to-right shunting through a patent ductus arteriosus is common among small preterm infants both with and without hyaline membrane disease?" When cardiac and ventilatory failure are progressive and life threatening in spite of aggressive medical management, there is general agreement that immediate intervention to close the ductus is indicated? -s For most preterm infants affected by a PDA, the hemodynamic effects are not immediately life threatening and can be controlled medically without intervention to close the ductus? Even so, small preterm infants weighing less than 1,500 gm at birth frequently have a prolonged course of ductus shunting, and characteristically require days or weeks of mechanical ventilation, In addition, this From the Departments o f Pediatrics and Cardiothoracic Surger), Vanderbilt University School o f Medicine. Supported b), a National tleart and Lung Institute Pulmonar)" SCOR Grant-HL 14214 and the Tennessee Heart Association. ~Recipient of American Lung Association Training Fellowship in Pulmonary Disease.
0022-3476/78/100647+05500.50/0 9 1978 The C. V. Mosby Co.
group is plagued by disorders such as chronic lung disease, necrotizing enterocolitis, intraventricular hemorrhage, retrolental fibroplasia, and problems owing to inadequate nutrition. Whether or not these infants would benefit from early ductus closure has been argued with
See related articles, pp. 639 and 682. Abbreviations used PDA: patent duetus arteriosus HMD: hyaline membrane disease MM: medical management "SC: surgicalclosure conviction, but without the benefit of data collected objectively from'controlled studies. Indeed, much of the enthusiasm for pharmacologically induced ductus constriction 9-'~ is based on the attractive but unproven premise that early ductus closure provides benefits that exceed the risks of potentially adverse drug effects. Prior to the present study, all symptomatic patients with PDA at Vanderbilt University Medical Center Newborn Intensive Care Unit were managed medically, without
Vol. 93. No 4. pp. 647.651
648
Cotton et al.
ductus ligation, unless congestive heart failure and pulmonary edema were thought to be immediately life threatening.' During that time, less than 10% of affected infants underwent ductus ligation because of medical management failure. That experience with a uniform protocol of aggressive medical management provided documentation of the expectations when small preterm infants with a symptomatic PDA are managed without surgical or pharmacologic intervention, and was used as the basis for the design and execution of an evaluation of early ductus closure. This repo/t describes the results of that evaluation by detailing the outcome of infants who were randomly assigned either "surgical closure" or "medical management" treatment protocols. MATERIALS
AND METttODS
Patients selected for the study came from the population ofpreterm infants weighing 1,500 gm or less admitted to the Vanderbilt Newborn Intensive Care Unit. The composition of this population and its management are described in a previous report.' Infants larger than 1,500 gm birth weight were not included because their morbidity and mortality related to a patent ductus are minimal. From the population of small preterm infants, those patients who developed a symptomatic PDA (see below) were treated medically for congestive heart failure and pulmonary edema according to a previously described protocol.' At one week of age, those infants who still had clinical, radiographic, and echocardiographic evidence of a large left-to-right shunt and remained dependent on mechanical ventilation through an endotracheal tube were considered for entry into the study. The decision for study entry was delayed if an infant had not received at least 48 hours of aggressive medical management. If.an infant's ventilatory status appeared tO be improving, study entry wasdelayed until an attempt to remove the endotracheal tube proved unsuccessful. Study entry was also delayed if an exchange transfusion was required because of clinically important clotting abnormalities. Infants with pulmonary infection, severe pulmonary interstitial air dissection, sepsis, meningitis, or suspected intraventricular 9hemorrhage were not enrolled in the study. In the absence of sepsis, necrotizing enterocolitis did not preclude study entry after fluid balance had been restored. If medical management was judged a failure (see below) before consideration for study entry, the ductus was immediately ligated and the patient was not included in the study. When an infant had met all conditions for study entry, an informed consent was obtained from the parents to permit the random assignment of either continued medical management or immediate surgical closure protocols.
The Journal o f Pediatrics October 1978
After permission for randomization had been obtained, the treatment protocol was selected by drawing from a series of identical envelopes. Initially there were 40 envelopes, half containing a designation for MM and half a designation for SC. Diagnosis ofsymptomatle PDA. Symptomatic PDA was diagnosed when clinical evidence of a large left-to-right duetus shunt was accompanied by cardiac failure requiring digitalis therapy.. The criteria for symptomatic PDA included an elevated basal heart rate greater than 150 beats per minute, usually with a gallop rhythm, a hyperdynamic precordium, collapsing pulses, and a characteristic murmur. Most infants had moist rales and all infants with a symptomatic PDA had chest radiographic findings of prominent pulmonary vasculature and evidence of pulmonary edema ranging from hazy lung fields to complete "white-out." These infants also developed cardiomegaly demonstrated by chest radiographs, and left ventricular or left atrial enlargement by echocardiography. Patients with the characteristic murmur of PDA who did not have evidence of congestive heart failure (tachycardia, galloP, pulmonary edema) were not designated as having a symptomatic PDA. As determined by the study entry criteria, the congestive heart failure and pulmonary edema of all study patients were successfully controlled by medical manage: ment before the treatment protocol was randomly assigned. At the time of study entry, the presence of a large left-to-right shunt was documented by the following findings: hyperactive precordium and collapsing pulses, murmur, radiographic evidence of cardiomegaly and increased pulmonary vasculature, and increased left atrial or left ventricular dimensions by echocardiography. Medical management protocol. A program of aggressive medical management as previously described' was continued when an infant was randomly assigned the MM protocol. Following assignment, if congestive heart failure and pulmonary edema subsequently were uncontrolled and the life of the infant was thought to be threatened by the left-to-right shunt, the ductus was closed surgically. Since surgical intervention was a provision of the MM protocol in the case of deterioration after protocol assignment, two MM infants who underwent surgical closure subsequent to randomization were included in the analyses of results. Surgical closure protocol. When an infant was randomly assigned the surgical closure protocol, ductus iigation was scheduled for the following day. Preparation on the day before surgery included insertion of a central venous catheter if one was not already present, and clotting factor analysis including a platelet count. When abnormal clotting function was thought to be clinically important, the
Volume93 Number4
Earl)' closure of PDA in preterm infants
DURATION OF ENDOTRACHEAL 9 ~
I00
INTUBATION
MQdlr 0--0 Surgir
649
Table I. Composition of study groups
MM group
(n=l~ wutvivort t 5 nOn-Survivors) (n= g llurvlvOrl I I floo-lutvlvor']
[
SC group
{ P* value
~o 8 0 x
td
60
)-
~20 0
.
_
z
ib :,~ 2b z'5 sb ~b 4b DAYS AFTER STUDY ENTRY
Fig. I. Illustration of mechanical ventilator dependence of the study groups. The patients who died were considered never to have been successfully extubated. patient underwent an exchange transfusion With fresh blood. The ductus ligation procedure Was carried Out in a heated (28~ operating room approximately 100 meters from the Newborn Intensive Care Unit and on the same floor. The patient . was transported to the operating room in an infant incubatoi'. During the procedure, the infant's body temperature was maintained above 36.2~ with an underlying heating pad and, at times, an overhead radiant heater. Total time away from the Newborn Intensive Care unit ranged from 30 to 60 minutes. Upon return to the unit, mechanical ventilation was resumed with the Babybird respirator, the chest tube was Connected to 8 cm 1t20 constant suction, and the estimated surgical blood loss 9(usually less than 5 ml) was replaced. Postoperatively, many o f the principles Of the MM protocol were observed. Maintenance fluid therapy was limited to _. 0.20
8.4 + 0.4
8.6 • 0.6
" > 0.50
t test-
"[" __. S E M .
around the null hypothesis, stated a priori, that there is no difference in duration until successful weaning from mechanical ventilation and endotracheal intubation whether symptomatic infants.with PDA are managed according to the MM or the SC protocol. Accordingly, dependence on mechanical ventilation via an endotracheal tube was a requirement for study entry, and the number of days between study entry arid final successful extubation was the variable chosen to test the hypothesis. At least once a day, infants in both study groups were evaluated for endotracheal tube removal. If extubation was followed by recurrent episodes of apnea and bradycardia wfiich were unresponsive to tactile stimulation, the infant was intubated again and ventilated mechanically. Statistical method. The Wilcoxan rank" test was selected a priori to decide if there was a difference in duratiori of need for endotracheal intubation following study entry between the MM and SC study grotips. For the SC group, study entry was designated as the day of surgery. For the MM group, study entry was designated as the day after the protocol was randomly assigned, a time corresponding to the day of surgery had the infant been assigned the SC instead of the MM protocol. Patients who died were considered not to have been extub~.ted successfully and were given an equal rank of one place grea!er than the survivor requiring the longest duration of endotracheal intubation. RESULTS Between March, 1975, and January, 1977, 25 infants with birth weights between .794 and 1,361 gm were enrolled in the study. Ten patients were assigned the SC protocol, and 15 the MM protocol. There was no statistically significant difference between the two groups in birt h weight, gcstational age, age at onset of symptomatic PDA, or age at study entry (Table I). Six of the ten SC
650
Cottonet aL
The Journal of Pediatrics October 1978
AVERAGE DALLY H O S P I T A L B I L L [ ] Medical ( n , 5 SurvivOrs) r l Sur~llCOl (n=8 Survivors)
8OO
Too O3 n~ ,r .d o ~
§ I
600 500 400
:
5OO 200 I00
BEFORE STUDY ENTRY
i i{in I-7
7-14
14-21
AFTER STUDY ENTRY
Fig. 2. Average daily hospital bill for each study group before study entry and for three sequential weeks after study entry. Data used for this analysis were available from five b,lb,l survivors and eight SC survivors. Table II. Survivors' complications after study entry
I SC group IMM group Number of survivors lhzlmonary complications Pneumothorax Interstitial air dissection Infection Chronic respirator lung ttemorrhage Necrotizing enterocolitis Retrolental fibroplasia Digitalis intoxication
9 1 2 3 1 1 0 1* 0
12 0 2 5 3 0 3 4"[ 2
*None with permanent cicatricial changes. tTwo with permanent cicatricial changes.
patients and eight of the 15 MM patients v;'ere recovering from IIMD when admitted to the study. Criteria for the diagnosis o f t t M D are given in a previous report.' One SC and three MM patients died, a difference that is not statistically significant. The SC patient died with Klebsiella sepsis. Two of the MM patients died with intraventricular hemorrhage and one died with superior vena caval thrombosis. The 25 study patients were ranked according to the number of days between study entry and final successful removal of the endotracheal tube. The sum of the ranks for the SC patients was 94. According to the Wilcoxan rank test, this value causes rejection (P < 0.05) of the null hypothesis that there is no difference in duration until successful weaning from mechanical ventilation and endotracheal intubation between symptomatic PDA infants managed with and without early ductus closure: Thus, the duration of ventilator dependence following
study entry was less for the SC group than the MM group. Fig. I illustrates this difference. The amount ofdigoxin and furosemide administered to survivors of the two study groups during the prestudy period and for three weeks after study entry was tabulated. Both groups received similar amounts of these drugs before study entry. As expected, the SC group received only a small amount of digoxin the week of study entry and none during the second and third weeks after surgery. The MM group had a continuing need for digoxin and furosemide throughout the three weekly periods following study entry. By the third week, the SC group received far less of the diuretic drug than did the MM group. The age that a sustained intake of at least 80 cal/kg/day via the gastrointestinal tract could be achieved by ihe survivors was regarded as an important index of morbidity. The SC group of patients achieved this level of caloric intake an average of 13 days (• SEM) after study entry, cgmpared to 24 days (___3.5 SEM) for the MM group. As a corollary to the delayed realization of caloric intake via the gastrointestinal tract, seven of the 12 MM survivors had a central venous line in place two weeks or more beyond study entry. This potentially hazardous requirement was necessary in only two of the SC survivor5. The crmplichtions ihat occurred among su~ivors of both study groups after siudy entry are listed in Table II. Important pulmonary complications occurred in both groups.Members of the SC group were spared necrotizing enterocolitis, retrolental fibi'oplasia resulting in permanent cicatrJcial changes, and digitalis intoxication. At the completion of the study, the daily iteriiized hospital bill Was available for five of the MM Survivors an d eight SC survivors (Fig. 2). The average daily hospital bill was similar for both groups before study entry. During the seven-day period beginning with study entry, the SC group had =i slightly higher average daily hospital bill, which was consistent with the additional costs related to surgery. During the second and third weeks after study entry, the SC group's average daily hospital bill declined. During the same periods, the hospital bill of the MM group remained at the level of the first week after study entry. The.average hospital bill of the SC group for the entire period was $1,805 less per patient than the average bill of the MM group. DISCUSSION The results of this prospective study of small preterm infants with symptomatic PDA who were randomly assigned one of two different management protocols
Voho,ze 93 Number 4
indicate that early closure of the ductus is desirable for the population of infants represented by the study groups. Infants who underwent early surgical closure were not ventilator dependent as long as infants assigned the MM protocol. In addition, the SC group of infants required fewer medications, became free of the need of a central venous catheter sooner, were able to meet much of their nutritional needs by the gastrointestinal route at an earlier age, were spared certain serious complications, and incurred smaller hospital bills than the patients managed according to the MM protocol. Since the management protocol was randomly assigned, there was no possibility that the investigators' bias influenced the composition of either study group. That the initial morbidity of the two groups was similar is documented by the similar amounts of digoxin and furosemide received by both groups before study entry and by the similar hospital bills incurred by both groups up until the time of protocol assignment. Also, there was no difference in birth weight, gestational age, age at onset of symptomatic PDA, or age at study entry between the two groups. The decision to terminate the study was made when the null hypothesis was rejected by the Wilcoxan rank test at P < 0.05. Although the grounds for hypothesis rejection are weakened somewhat by stepwise examination of the data, continuation of the study beyond this point could not be justified in view of the other evidence of decreased morbidity in the SC group. The number ofsubjects in this study was limited by the occurrence o f hypothesis rejection after the twenty-fifth patient had completed the protocol. Even though this clinical trial did not involve large numbers of subjects, its design and conduct, which was prospective and included randomized treatment assignment as well as the a priori statement of hypothesis and statistical method, yielded results which indicated that early ductus closure was the superior treatment for this population. The results of this study have provided us a rational solution to an important management dilemma. Even though the congestive heart failure and pulmonary edema are satisfactorily controlled, we believe the infant with a symptomatic PDA who is requiring mechanical ventila-
Earl)' closure of PDA in preterm it~znts
65 1
tion at a week of age will benefit from early surgical closure of the ductus. Any infant with progressive, uncontrolled congestive heart failure or pulmonary edema should undergo immediate ductus closure as a lifesaving measure. At present, aggressive medical management is regarded satisfactory for patients who are medically controlled and do not require mechanical ventilation. Whether indomethacin would be beneficial for these patients or as an alternative to surgery for those who are ventilator dependent remains uncertain. Knowledge that early surgical closure is superior to medical management provides additional justification for the investigational trial of a pharmacologic agent such as indomethacin. REFERENCES 1. Cotton RB, Stahlman MT, Kovar I, and Catterton WZ: Medical management of small preterm infants v,ith symptomatic patent ductus arteriosus, J PEDIA'rR92:467, 1978. 2. Kitterman JA, Edmunds LII, Gregory GA, et al: Patent ductus arteriosus in premature infants, N Engl J Med 287:473, 1972. 3. Zachman RD, Steinmetz GP, Botham RJ, et al: Incidence and treatment of the patent ductus arteriosus in the ill premature neonate, Am Heart J 87:697, 1974. 4. Siassi B, Blanco C, Cabal LA, and Coran AG: Incidence and clinical features of patent ductus arteriosus in lowbirthweight infants, Pediatrics 57:347, 1976. 5. Edmunds LH, Gregory GA,tteymann MA, et al: Surgical closure of the ductus arteriosus in premature infants, Circulation 48:856, 1973. 6. Thibeault DW, Emmanouilides GC, Nelson RJ, Lachman RS, Rosengart RM, and Oh W: Patent ductus arteriosus complicating the respiratory distress syndrome in preterm infants, J PEDI,~rR 86:120, 1975. 7. Neal WA, Bessinger FB, Hunt CE, and Lueas RV: Patent ductus aneriosus complicating respiratory distress syndrome, J PEozxrg 86:127, 1975. 8. Nadas AS: Patent ductus revisited, N Engl J Med 295:563, 1976. 9. Friedman WF, Hirschklau M J, Printz MP, et al: Pharmacologic closure of patent ductus arteriosus in the premature infant, N Engl J Med 295:526, 1976. 10. Heymann MA, Rudolph AM, and Silverman NH: Closure of the ductus arteriosus in premature infants by inhibition of prostaglandin synthesis, N Engl J Med 295:530, 1976. il. Snedecor GW, and Cochran WG: Statistical methods, Ames, Iowa, 1968, The Iowa State University Press, pp 130-131.
The Journal o f P E D 1 A T R 1 C S
647
Randomized trial of early closure of symptomatic patent ductus arteriosus in small preterm infants As a result of randomized assignment. 15 preterm infants weighing 1,500 gm or less at birth and who had a symptomatic PDA were treated according to a medical management protocol, and ten according to an earl)' surgical closure protocol. AII infants required mechanical ventilation at the time of stud). entr); which was one week after birth. Birth weight, gestational age, age at onset of congestive failure, age at stud), entr); and the initial morbidit)' of members o f the two groups were similar. The nine surviving infants managed according to the surgical closure protocol were weaned from mechanical ventilation sooner, had a decreased need for digoxin and fttrosemide, achieved gastrointestinal fttnction sooner, and had a smaller hospital bill than the 12 survivors of the medical management group. These results indicate that infants with a s)'mptomatic PDA still requiring mechanical ventilation at one week after birth will benefit from surgical closure o f the ductus at that time.
Robert B. Cotton, M.D., Mildred T. Stahlman, M.D., Harvey W. Bender, M.D.,
Thomas P. Graham, M.D., W. Z. Catterton, M.D.,* and llya Kovar, M.D, Nashville, TemL
CONGESTIVE HEART FAILURE with pulmonary edema due to left-to-right shunting through a patent ductus arteriosus is common among small preterm infants both with and without hyaline membrane disease?" When cardiac and ventilatory failure are progressive and life threatening in spite of aggressive medical management, there is general agreement that immediate intervention to close the ductus is indicated? -s For most preterm infants affected by a PDA, the hemodynamic effects are not immediately life threatening and can be controlled medically without intervention to close the ductus? Even so, small preterm infants weighing less than 1,500 gm at birth frequently have a prolonged course of ductus shunting, and characteristically require days or weeks of mechanical ventilation, In addition, this From the Departments o f Pediatrics and Cardiothoracic Surger), Vanderbilt University School o f Medicine. Supported b), a National tleart and Lung Institute Pulmonar)" SCOR Grant-HL 14214 and the Tennessee Heart Association. ~Recipient of American Lung Association Training Fellowship in Pulmonary Disease.
0022-3476/78/100647+05500.50/0 9 1978 The C. V. Mosby Co.
group is plagued by disorders such as chronic lung disease, necrotizing enterocolitis, intraventricular hemorrhage, retrolental fibroplasia, and problems owing to inadequate nutrition. Whether or not these infants would benefit from early ductus closure has been argued with
See related articles, pp. 639 and 682. Abbreviations used PDA: patent duetus arteriosus HMD: hyaline membrane disease MM: medical management "SC: surgicalclosure conviction, but without the benefit of data collected objectively from'controlled studies. Indeed, much of the enthusiasm for pharmacologically induced ductus constriction 9-'~ is based on the attractive but unproven premise that early ductus closure provides benefits that exceed the risks of potentially adverse drug effects. Prior to the present study, all symptomatic patients with PDA at Vanderbilt University Medical Center Newborn Intensive Care Unit were managed medically, without
Vol. 93. No 4. pp. 647.651
648
Cotton et al.
ductus ligation, unless congestive heart failure and pulmonary edema were thought to be immediately life threatening.' During that time, less than 10% of affected infants underwent ductus ligation because of medical management failure. That experience with a uniform protocol of aggressive medical management provided documentation of the expectations when small preterm infants with a symptomatic PDA are managed without surgical or pharmacologic intervention, and was used as the basis for the design and execution of an evaluation of early ductus closure. This repo/t describes the results of that evaluation by detailing the outcome of infants who were randomly assigned either "surgical closure" or "medical management" treatment protocols. MATERIALS
AND METttODS
Patients selected for the study came from the population ofpreterm infants weighing 1,500 gm or less admitted to the Vanderbilt Newborn Intensive Care Unit. The composition of this population and its management are described in a previous report.' Infants larger than 1,500 gm birth weight were not included because their morbidity and mortality related to a patent ductus are minimal. From the population of small preterm infants, those patients who developed a symptomatic PDA (see below) were treated medically for congestive heart failure and pulmonary edema according to a previously described protocol.' At one week of age, those infants who still had clinical, radiographic, and echocardiographic evidence of a large left-to-right shunt and remained dependent on mechanical ventilation through an endotracheal tube were considered for entry into the study. The decision for study entry was delayed if an infant had not received at least 48 hours of aggressive medical management. If.an infant's ventilatory status appeared tO be improving, study entry wasdelayed until an attempt to remove the endotracheal tube proved unsuccessful. Study entry was also delayed if an exchange transfusion was required because of clinically important clotting abnormalities. Infants with pulmonary infection, severe pulmonary interstitial air dissection, sepsis, meningitis, or suspected intraventricular 9hemorrhage were not enrolled in the study. In the absence of sepsis, necrotizing enterocolitis did not preclude study entry after fluid balance had been restored. If medical management was judged a failure (see below) before consideration for study entry, the ductus was immediately ligated and the patient was not included in the study. When an infant had met all conditions for study entry, an informed consent was obtained from the parents to permit the random assignment of either continued medical management or immediate surgical closure protocols.
The Journal o f Pediatrics October 1978
After permission for randomization had been obtained, the treatment protocol was selected by drawing from a series of identical envelopes. Initially there were 40 envelopes, half containing a designation for MM and half a designation for SC. Diagnosis ofsymptomatle PDA. Symptomatic PDA was diagnosed when clinical evidence of a large left-to-right duetus shunt was accompanied by cardiac failure requiring digitalis therapy.. The criteria for symptomatic PDA included an elevated basal heart rate greater than 150 beats per minute, usually with a gallop rhythm, a hyperdynamic precordium, collapsing pulses, and a characteristic murmur. Most infants had moist rales and all infants with a symptomatic PDA had chest radiographic findings of prominent pulmonary vasculature and evidence of pulmonary edema ranging from hazy lung fields to complete "white-out." These infants also developed cardiomegaly demonstrated by chest radiographs, and left ventricular or left atrial enlargement by echocardiography. Patients with the characteristic murmur of PDA who did not have evidence of congestive heart failure (tachycardia, galloP, pulmonary edema) were not designated as having a symptomatic PDA. As determined by the study entry criteria, the congestive heart failure and pulmonary edema of all study patients were successfully controlled by medical manage: ment before the treatment protocol was randomly assigned. At the time of study entry, the presence of a large left-to-right shunt was documented by the following findings: hyperactive precordium and collapsing pulses, murmur, radiographic evidence of cardiomegaly and increased pulmonary vasculature, and increased left atrial or left ventricular dimensions by echocardiography. Medical management protocol. A program of aggressive medical management as previously described' was continued when an infant was randomly assigned the MM protocol. Following assignment, if congestive heart failure and pulmonary edema subsequently were uncontrolled and the life of the infant was thought to be threatened by the left-to-right shunt, the ductus was closed surgically. Since surgical intervention was a provision of the MM protocol in the case of deterioration after protocol assignment, two MM infants who underwent surgical closure subsequent to randomization were included in the analyses of results. Surgical closure protocol. When an infant was randomly assigned the surgical closure protocol, ductus iigation was scheduled for the following day. Preparation on the day before surgery included insertion of a central venous catheter if one was not already present, and clotting factor analysis including a platelet count. When abnormal clotting function was thought to be clinically important, the
Volume93 Number4
Earl)' closure of PDA in preterm infants
DURATION OF ENDOTRACHEAL 9 ~
I00
INTUBATION
MQdlr 0--0 Surgir
649
Table I. Composition of study groups
MM group
(n=l~ wutvivort t 5 nOn-Survivors) (n= g llurvlvOrl I I floo-lutvlvor']
[
SC group
{ P* value
~o 8 0 x
td
60
)-
~20 0
.
_
z
ib :,~ 2b z'5 sb ~b 4b DAYS AFTER STUDY ENTRY
Fig. I. Illustration of mechanical ventilator dependence of the study groups. The patients who died were considered never to have been successfully extubated. patient underwent an exchange transfusion With fresh blood. The ductus ligation procedure Was carried Out in a heated (28~ operating room approximately 100 meters from the Newborn Intensive Care Unit and on the same floor. The patient . was transported to the operating room in an infant incubatoi'. During the procedure, the infant's body temperature was maintained above 36.2~ with an underlying heating pad and, at times, an overhead radiant heater. Total time away from the Newborn Intensive Care unit ranged from 30 to 60 minutes. Upon return to the unit, mechanical ventilation was resumed with the Babybird respirator, the chest tube was Connected to 8 cm 1t20 constant suction, and the estimated surgical blood loss 9(usually less than 5 ml) was replaced. Postoperatively, many o f the principles Of the MM protocol were observed. Maintenance fluid therapy was limited to _. 0.20
8.4 + 0.4
8.6 • 0.6
" > 0.50
t test-
"[" __. S E M .
around the null hypothesis, stated a priori, that there is no difference in duration until successful weaning from mechanical ventilation and endotracheal intubation whether symptomatic infants.with PDA are managed according to the MM or the SC protocol. Accordingly, dependence on mechanical ventilation via an endotracheal tube was a requirement for study entry, and the number of days between study entry arid final successful extubation was the variable chosen to test the hypothesis. At least once a day, infants in both study groups were evaluated for endotracheal tube removal. If extubation was followed by recurrent episodes of apnea and bradycardia wfiich were unresponsive to tactile stimulation, the infant was intubated again and ventilated mechanically. Statistical method. The Wilcoxan rank" test was selected a priori to decide if there was a difference in duratiori of need for endotracheal intubation following study entry between the MM and SC study grotips. For the SC group, study entry was designated as the day of surgery. For the MM group, study entry was designated as the day after the protocol was randomly assigned, a time corresponding to the day of surgery had the infant been assigned the SC instead of the MM protocol. Patients who died were considered not to have been extub~.ted successfully and were given an equal rank of one place grea!er than the survivor requiring the longest duration of endotracheal intubation. RESULTS Between March, 1975, and January, 1977, 25 infants with birth weights between .794 and 1,361 gm were enrolled in the study. Ten patients were assigned the SC protocol, and 15 the MM protocol. There was no statistically significant difference between the two groups in birt h weight, gcstational age, age at onset of symptomatic PDA, or age at study entry (Table I). Six of the ten SC
650
Cottonet aL
The Journal of Pediatrics October 1978
AVERAGE DALLY H O S P I T A L B I L L [ ] Medical ( n , 5 SurvivOrs) r l Sur~llCOl (n=8 Survivors)
8OO
Too O3 n~ ,r .d o ~
§ I
600 500 400
:
5OO 200 I00
BEFORE STUDY ENTRY
i i{in I-7
7-14
14-21
AFTER STUDY ENTRY
Fig. 2. Average daily hospital bill for each study group before study entry and for three sequential weeks after study entry. Data used for this analysis were available from five b,lb,l survivors and eight SC survivors. Table II. Survivors' complications after study entry
I SC group IMM group Number of survivors lhzlmonary complications Pneumothorax Interstitial air dissection Infection Chronic respirator lung ttemorrhage Necrotizing enterocolitis Retrolental fibroplasia Digitalis intoxication
9 1 2 3 1 1 0 1* 0
12 0 2 5 3 0 3 4"[ 2
*None with permanent cicatricial changes. tTwo with permanent cicatricial changes.
patients and eight of the 15 MM patients v;'ere recovering from IIMD when admitted to the study. Criteria for the diagnosis o f t t M D are given in a previous report.' One SC and three MM patients died, a difference that is not statistically significant. The SC patient died with Klebsiella sepsis. Two of the MM patients died with intraventricular hemorrhage and one died with superior vena caval thrombosis. The 25 study patients were ranked according to the number of days between study entry and final successful removal of the endotracheal tube. The sum of the ranks for the SC patients was 94. According to the Wilcoxan rank test, this value causes rejection (P < 0.05) of the null hypothesis that there is no difference in duration until successful weaning from mechanical ventilation and endotracheal intubation between symptomatic PDA infants managed with and without early ductus closure: Thus, the duration of ventilator dependence following
study entry was less for the SC group than the MM group. Fig. I illustrates this difference. The amount ofdigoxin and furosemide administered to survivors of the two study groups during the prestudy period and for three weeks after study entry was tabulated. Both groups received similar amounts of these drugs before study entry. As expected, the SC group received only a small amount of digoxin the week of study entry and none during the second and third weeks after surgery. The MM group had a continuing need for digoxin and furosemide throughout the three weekly periods following study entry. By the third week, the SC group received far less of the diuretic drug than did the MM group. The age that a sustained intake of at least 80 cal/kg/day via the gastrointestinal tract could be achieved by ihe survivors was regarded as an important index of morbidity. The SC group of patients achieved this level of caloric intake an average of 13 days (• SEM) after study entry, cgmpared to 24 days (___3.5 SEM) for the MM group. As a corollary to the delayed realization of caloric intake via the gastrointestinal tract, seven of the 12 MM survivors had a central venous line in place two weeks or more beyond study entry. This potentially hazardous requirement was necessary in only two of the SC survivor5. The crmplichtions ihat occurred among su~ivors of both study groups after siudy entry are listed in Table II. Important pulmonary complications occurred in both groups.Members of the SC group were spared necrotizing enterocolitis, retrolental fibi'oplasia resulting in permanent cicatrJcial changes, and digitalis intoxication. At the completion of the study, the daily iteriiized hospital bill Was available for five of the MM Survivors an d eight SC survivors (Fig. 2). The average daily hospital bill was similar for both groups before study entry. During the seven-day period beginning with study entry, the SC group had =i slightly higher average daily hospital bill, which was consistent with the additional costs related to surgery. During the second and third weeks after study entry, the SC group's average daily hospital bill declined. During the same periods, the hospital bill of the MM group remained at the level of the first week after study entry. The.average hospital bill of the SC group for the entire period was $1,805 less per patient than the average bill of the MM group. DISCUSSION The results of this prospective study of small preterm infants with symptomatic PDA who were randomly assigned one of two different management protocols
Voho,ze 93 Number 4
indicate that early closure of the ductus is desirable for the population of infants represented by the study groups. Infants who underwent early surgical closure were not ventilator dependent as long as infants assigned the MM protocol. In addition, the SC group of infants required fewer medications, became free of the need of a central venous catheter sooner, were able to meet much of their nutritional needs by the gastrointestinal route at an earlier age, were spared certain serious complications, and incurred smaller hospital bills than the patients managed according to the MM protocol. Since the management protocol was randomly assigned, there was no possibility that the investigators' bias influenced the composition of either study group. That the initial morbidity of the two groups was similar is documented by the similar amounts of digoxin and furosemide received by both groups before study entry and by the similar hospital bills incurred by both groups up until the time of protocol assignment. Also, there was no difference in birth weight, gestational age, age at onset of symptomatic PDA, or age at study entry between the two groups. The decision to terminate the study was made when the null hypothesis was rejected by the Wilcoxan rank test at P < 0.05. Although the grounds for hypothesis rejection are weakened somewhat by stepwise examination of the data, continuation of the study beyond this point could not be justified in view of the other evidence of decreased morbidity in the SC group. The number ofsubjects in this study was limited by the occurrence o f hypothesis rejection after the twenty-fifth patient had completed the protocol. Even though this clinical trial did not involve large numbers of subjects, its design and conduct, which was prospective and included randomized treatment assignment as well as the a priori statement of hypothesis and statistical method, yielded results which indicated that early ductus closure was the superior treatment for this population. The results of this study have provided us a rational solution to an important management dilemma. Even though the congestive heart failure and pulmonary edema are satisfactorily controlled, we believe the infant with a symptomatic PDA who is requiring mechanical ventila-
Earl)' closure of PDA in preterm it~znts
65 1
tion at a week of age will benefit from early surgical closure of the ductus. Any infant with progressive, uncontrolled congestive heart failure or pulmonary edema should undergo immediate ductus closure as a lifesaving measure. At present, aggressive medical management is regarded satisfactory for patients who are medically controlled and do not require mechanical ventilation. Whether indomethacin would be beneficial for these patients or as an alternative to surgery for those who are ventilator dependent remains uncertain. Knowledge that early surgical closure is superior to medical management provides additional justification for the investigational trial of a pharmacologic agent such as indomethacin. REFERENCES 1. Cotton RB, Stahlman MT, Kovar I, and Catterton WZ: Medical management of small preterm infants v,ith symptomatic patent ductus arteriosus, J PEDIA'rR92:467, 1978. 2. Kitterman JA, Edmunds LII, Gregory GA, et al: Patent ductus arteriosus in premature infants, N Engl J Med 287:473, 1972. 3. Zachman RD, Steinmetz GP, Botham RJ, et al: Incidence and treatment of the patent ductus arteriosus in the ill premature neonate, Am Heart J 87:697, 1974. 4. Siassi B, Blanco C, Cabal LA, and Coran AG: Incidence and clinical features of patent ductus arteriosus in lowbirthweight infants, Pediatrics 57:347, 1976. 5. Edmunds LH, Gregory GA,tteymann MA, et al: Surgical closure of the ductus arteriosus in premature infants, Circulation 48:856, 1973. 6. Thibeault DW, Emmanouilides GC, Nelson RJ, Lachman RS, Rosengart RM, and Oh W: Patent ductus arteriosus complicating the respiratory distress syndrome in preterm infants, J PEDI,~rR 86:120, 1975. 7. Neal WA, Bessinger FB, Hunt CE, and Lueas RV: Patent ductus aneriosus complicating respiratory distress syndrome, J PEozxrg 86:127, 1975. 8. Nadas AS: Patent ductus revisited, N Engl J Med 295:563, 1976. 9. Friedman WF, Hirschklau M J, Printz MP, et al: Pharmacologic closure of patent ductus arteriosus in the premature infant, N Engl J Med 295:526, 1976. 10. Heymann MA, Rudolph AM, and Silverman NH: Closure of the ductus arteriosus in premature infants by inhibition of prostaglandin synthesis, N Engl J Med 295:530, 1976. il. Snedecor GW, and Cochran WG: Statistical methods, Ames, Iowa, 1968, The Iowa State University Press, pp 130-131.
