- Email: [email protected]
Early Caffeine and Weaning from Mechanical Ventilation in Preterm Infants: A Randomized, Placebo-Controlled Trial
ARTICLE IN PRESS THE JOURNAL OF PEDIATRICS • www.jpeds.com
ORIGINAL ARTICLES
Early Caffeine and Weaning from Mechanical Ventilation in Preterm Infants: A Randomized, Placebo-Controlled Trial Cynthia M. Amaro, MD*, Jose A. Bello, MD*, Deepak Jain, MD, Alexandra Ramnath, MD, Carmen D'Ugard, RRT, Silvia Vanbuskirk, RN, Eduardo Bancalari, MD, and Nelson Claure, MSc, PhD Objective To evaluate in a randomized, double-blind, placebo-controlled trial the effect of early caffeine on the age of first successful extubation in preterm infants. Study design Preterm infants born at 23-30 weeks of gestation requiring mechanical ventilation in the first 5 postnatal days were randomized to receive a 20 mg/kg loading dose followed by 5 mg/kg/day of caffeine or placebo until considered ready for extubation. The placebo group received a blinded loading dose of caffeine before extubation. Results Infants were randomized to receive caffeine (n = 41) or placebo (n = 42). Age at first successful extubation did not differ between early caffeine (median, 24 days; IQR, 10-41 days) and control groups (median, 20 days; IQR, 9-43 days; P = .7). An interim analysis at 75% enrollment showed a trend toward higher mortality in 1 of the groups and the data safety and monitoring board recommended stopping the trial. Unblinded analysis revealed mortality did not differ significantly between the early caffeine (9 [22%]) and control groups (5 [12%]; P = .22). Conclusions Early initiation of caffeine in this group of premature infants did not reduce the age of first successful extubation. A nonsignificant trend toward higher mortality in the early caffeine group led to a cautious decision to stop the trial. These findings suggest caution with early use of caffeine in mechanically ventilated preterm infants until more efficacy and safety data become available. (J Pediatr 2018;■■:■■-■■). Trial Registration ClinicalTrials.gov NCT01751724.
A
large proportion of extreme premature infants require prolonged mechanical ventilation. This is associated with increased risk for bronchopulmonary dysplasia (BPD) and poor neurodevelopmental outcome.1 Caffeine, a methylxanthine and adenosine receptor antagonist, is a potent stimulant of central respiratory activity and is an effective treatment of apnea of prematurity and to avoid extubation failure.2,3 However, adenosine has been shown to preserve brain cell survival and prevent brain energy failure by preserving adenosine triphosphate levels in experimental hypoxia and ischemia models.4-6 Although caffeine has been used for these indications for the last 3 decades, until recently there were no long-term safety data to support its use.7 The large, randomized, controlled trial, Caffeine for Apnea of Prematurity, showed that caffeine started for these indications within the first 10 days after birth did not have long-term negative effects on neurologic outcome or survival.8,9 Analysis of predischarge outcomes showed a shorter duration of mechanical ventilation and a lower incidence of BPD in the caffeine compared with the placebo group. A secondary analysis showed that infants started on caffeine in the first 3 days while receiving mechanical ventilation derived the most benefit in terms of BPD and neurodevelopment.10 Since then, the use of caffeine has increased and often is started earlier in mechanically ventilated preterm infants.11 Strategies to provide mechanical respiratory support have changed from controlled ventilation to a more gentle support, where the ventilator is used to assist the infant’s spontaneous breathing. This step has been achieved in large part through the use of patient-triggered modes of mechanical ventilation. The effective use of these modes to reduce the need for mechanical ventilation must rely on the spontaneous respiratory drive, but this drive is frequently inconsistent in extremely premature infants. Thus, the use of respiratory stimulants such as caffeine seems to be a good alternative to decrease the duration of mechanical ventilation in this population. The efficacy and safety of this approach have not been evaluated by prospective, randomized, controlled trials. We postulated that early initiation of caffeine in mechanically ventilated preterm From the Division of Neonatology, Department of Pediatrics, University of Miami Miller School of Medicine, infants will hasten the weaning process during the first course of mechanical venMiami, FL tilation. The primary objective of this trial was to evaluate the effect of early caf*Contributed equally. Supported by the University of Miami Project NewBorn, a feine administration on the age of first successful extubation in mechanically philanthropic organization that did not participate in any aspect of the research. The authors declare no conflicts of interest.
BPD DSMB FiO2 NICU
Portions of this study were presented as an abstract at the Pediatric Academic Societies annual meeting, April 30-May 3, 2016, Baltimore, Maryland.
Bronchopulmonary dysplasia Data safety and monitoring board Fraction of inspired oxygen Neonatal intensive care unit
0022-3476/$ - see front matter. © 2018 Elsevier Inc. All rights reserved. https://doi.org10.1016/j.jpeds.2018.01.010
1 FLA 5.5.0 DTD ■ YMPD9716_proof ■ March 6, 2018
THE JOURNAL OF PEDIATRICS • www.jpeds.com ventilated preterm infants. The secondary objective was to evaluate the effect of early caffeine on total duration of mechanical ventilation and oxygen supplementation and on the incidence of BPD.
Methods This single-center, double-blind, placebo-controlled, randomized clinical trial (ClinicalTrials.gov: NCT01751724) was conducted in the neonatal intensive care unit (NICU) at Holtz Children’s Hospital of the Jackson Health System—University of Miami Medical Center. Premature infants born between 23 and 30 weeks of gestation who required mechanical ventilation in the first 5 postnatal days were eligible for the study. Infants with major congenital anomalies and infants that were small for gestational age (birth weight <3rd percentile) were excluded. Endpoints The primary endpoint of the trial was the age of first successful extubation, defined as postnatal age of first extubation after which the infant remained extubated for >24 hours. Secondary respiratory endpoints included total duration of mechanical ventilation, duration of oxygen supplementa-
Volume ■■
•
■■ 2018
tion, and the incidence of BPD. An oxygen or ventilator day was defined as the need for oxygen or mechanical ventilation for ≥12 hours over a 24-hour period. BPD was defined as need for supplemental oxygen for ≥28 days and at 36 weeks postmenstrual age. Severe BPD was defined as need for supplemental oxygen for ≥28 days and fraction of inspired oxygen (FiO2) of ≥0.30 and/or positive pressure respiratory support at 36 weeks postmenstrual age. Neonatal morbidities including pulmonary hemorrhage, echocardiography-confirmed patent ductus arteriosus, necrotizing enterocolitis defined as Bell’s stage II or III, spontaneous intestinal perforation, grade 3 or 4 intraventricular hemorrhage, periventricular leukomalacia, and retinopathy of prematurity stage 3 or 4 were documented. Interventions After screening and enrollment with parental consent during the first 5 days after birth, mechanically ventilated infants were randomized to the early caffeine or control group (Figure 1). Randomization was stratified according to gestational age, either 23-26 or 27-30 weeks, using sealed, opaque envelopes. Stratification was aimed at balancing the number of infants within each gestational age bracket. Randomization and study drug preparation were done by the NICU pharmacy staff. Knowl-
Figure 1. Trial flow diagram. Study protocol from screening to the end of the intervention. 2
Amaro et al FLA 5.5.0 DTD ■ YMPD9716_proof ■ March 6, 2018
■■ 2018
ORIGINAL ARTICLES
edge of treatment assignment was limited to the pharmacy staff. Investigators and clinical teams were unaware of the assigned treatment. Infants in the early caffeine group received a loading dose of caffeine citrate 20 mg/kg followed by a maintenance dose of 5 mg/kg/day. Infants in the control group received an equivalent volume of normal saline bolus and maintenance. The absolute dose of study drug was adjusted weekly to changes in weight. Routine ventilator management continued until the clinical team made the decision to extubate the infant. Infants in the control group received a blinded loading dose of caffeine citrate 20 mg/kg before extubation, as is routinely done in this NICU. Infants in the early caffeine group received a blinded loading dose of placebo. This was done to prevent excessive caffeine levels in this group. At this point, the investigators and clinical team remained blinded to the assigned intervention. If the infant remained extubated for >24 hours (definition of successful extubation), the study intervention ended and open-label caffeine at 5 mg/kg/day was given to all infants as per standard of care. Infants requiring reintubation in ≤24 hours remained in the assigned group and continued to receive blinded maintenance doses of caffeine or placebo. These steps continued until the infant was successfully extubated (as defined previously). Caffeine levels were not measured in the participating infants during the trial intervention period to avoid unmasking of the groups. Respiratory management followed unit guidelines. These measures include intubation and the administration of surfactant to infants with respiratory distress syndrome who require an FiO2 of >0.3 to keep arterial oxygen saturation between 90% and 95%. Subsequently, the ventilator mandatory rate and peak pressure were adjusted to maintain tidal volume between 4 and 7 mL/kg, pH between 7.25 and 7.4, and PaCO2 between 45 and 55 mm Hg during the first week after birth and between 45 and 65 mm Hg thereafter. Clinical teams were reminded to consider extubation when infants met the following extubation criteria: FiO2 of <0.4, positive endexpiratory pressure of ≤5 cm H2O, a mandatory rate of <20 cycles per minute, a peak pressure of <18 cm H2O, and pressure support of <10 cm H2O to keep arterial blood gases within these ranges. After extubation, clinical teams considered reintubation only if the infant met ≥1 of the following criteria: ≥2 episodes of severe apnea or bradycardia requiring bag-mask ventilation in a 4-hour period, an increase in FiO2 of >0.2 above the pre-extubation baseline and of >0.5, or an increase in PaCO2 of >15 mm Hg above the pre-extubation baseline and of >70 mm Hg. Enrolled infants were followed until 36 weeks postmenstrual age, discharge, or death, whichever came first. Statistical Considerations and Analyses The primary endpoint of the trial was the age at first successful extubation. We estimated that a total enrollment of 110 infants was needed to detect a reduction in this endpoint by >7 days, from a historical baseline of 15 ± 21 days, with 80% power and 5% significance.
Categorical variables were compared by the Pearson c2 or Fisher exact test. Continuous variables were compared by the Student t test or Mann-Whitney U test, depending on the distribution. The Kaplan-Meier log-rank test was used to compare time-to-event variables, such as age at first successful extubation. P < .05 was considered statistically significant. All statistical analyses were performed using IBM SPSS statistics package (IBM Corporation, Armonk, New York). Safety Monitoring Monitoring for side effects of caffeine was done as a part of routine clinical care. A dose of study drug could be withheld, decreased, or completely stopped in infants with persistent tachycardia (defined as a heart rate of >190 beats per minute for 4 hours after excluding other causes) or seizure activity (documented by electroencephalogram and not explained by other causes). The occurrence of these events was documented. An independent data safety and monitoring board (DSMB) was appointed. Interim analyses were planned at 50% and 75% enrollment to monitor safety. The DSMB board was presented with blinded group data on mortality and rates of adverse events. Ethics The trial was approved by the Institutional Review Board of the University of Miami Human Subjects Research Office and the Jackson Health System Office of Research. Written informed parental consent was obtained before enrollment.
Results The trial was conducted from January 2013 to December 2015. The CONSORT diagram in Figure 2 shows the flow of participants from screening to randomization; 208 infants admitted to the NICU were screened for eligibility and 143 were deemed eligible; 65 infants were excluded. Of these, 52 were not intubated during the first 5 days, 5 had congenital anomalies, and 1 was small for gestational age. Eighty-seven infants were enrolled after obtaining parental consent. Of these, 1 infant died before randomization and 86 infants were randomized. Three of the infants who were randomized were withdrawn from the trial, 2 of them by the parents, which prevented collection of any further data for inclusion in the intention to treat analyses. One randomized infant was withdrawn by the investigators because of extubation and the administration of open-label caffeine by the clinical team before starting the administration of the assigned study drug. Eighty-three randomized infants completed the trial intervention and follow-up, and were included in the analysis, 41 in the early caffeine and 42 in the control group. Table I shows the baseline characteristics of the population. The early caffeine and control groups did not differ with regard to gestational age, birth weight, ethnicity, or exposure to antenatal steroids. There was a greater proportion of male infants in the early caffeine compared with the control group and a trend toward a higher proportion of infants with a 5-minute Apgar score of <5 in the early caffeine group compared with the
Early Caffeine and Weaning from Mechanical Ventilation in Preterm Infants: A Randomized, Placebo-Controlled Trial FLA 5.5.0 DTD ■ YMPD9716_proof ■ March 6, 2018
3
THE JOURNAL OF PEDIATRICS • www.jpeds.com
Volume ■■
Figure 2. CONSORT diagram. Flow of trial participants from screening to randomization.
control group. Infants enrolled in the trial were intubated at a mean age of 3 hours and were randomized at a mean age of 48 hours. The respiratory support parameters at the time of randomization did not differ between groups. There were only 2 protocol deviations. One infant did not receive a blinded study drug bolus before extubation, but was extubated successfully. Another infant who failed extubation within 24 hours was mistakenly randomized for a second time when reintubated. This infant was analyzed in the group assigned in the original randomization.
Table I. Population demographics and characteristics at randomization Characteristic Gestational age (wk) Birth weight (g) Male Black Antenatal steroids 5-Minute Apgar score of <5 Age at intubation (h) Received surfactant Age at randomization (h) FiO2 at randomization Mean airway pressure at randomization
Early caffeine (n = 41)
Control (n = 42)
P
25.7 (24.3-27.0) 670 (605-915) 30 (73) 19 (46) 37 (90) 10 (24) 3.0 ± 8.2 39 (95) 48 ± 25 0.25 (0.21-0.37) 9 (8-11)
26.1 (24.2-28.4) 720 (643-894) 17 (41) 21 (50) 40 (95) 4 (9.5) 3.1 ± 7.0 40 (95) 49 ± 31 0.23 (0.21-0.30) 8 (8-10)
.425 .450 .003 .739 .433 .071 .449 1.000 .874 .195 .112
Data are mean ± SD, median (IQR), or n (%).
The first interim analysis (blinded to group assignment) conducted at 50% enrollment showed a trend toward a higher mortality rate in one of the groups. At the time, the mortality rate in both groups was below the center’s historic mortality rate in the 2 years preceding the trial for a similar population and the DSMB recommended that enrollment should continue. At the second interim analysis at 75% enrollment (83 infants), the DSMB recommended terminating the trial because of a persistent trend of higher mortality, although not statistically significant, in the same group. At this time, the mortality in this group slightly exceeded the historic mortality rate for a similar population in this center. In agreement with the DSMB, the investigators terminated enrollment in the trial. Unblinded analysis revealed a nonsignificant difference in mortality between the early caffeine and control groups (Table II). The proportion of infants who died before extubation (while on trial intervention) and the age of death did not differ significantly between groups. There were no instances where the study drug needed to be stopped or decreased owing to side effects. Analysis of the primary endpoint revealed no difference in the age at first successful extubation between the early caffeine (median, 24 days; IQR, 10-41 days) and control groups (median, 20 days; IQR, 9-43 days; P = .703 by Kaplan-Meier log-rank test; see Table III and Figure 3). Analysis of the secondary respiratory endpoints showed the total duration of mechanical ventilation or oxygen supplementation did not differ between the groups. The incidence
4
Amaro et al FLA 5.5.0 DTD ■ YMPD9716_proof ■ March 6, 2018
■■ 2018
ORIGINAL ARTICLES
Table II. Primary and secondary endpoints and mortality Outcome
Early caffeine (n = 41)
Age at first successful extubation (days) Total duration of mechanical ventilation, days‡ Total duration of oxygen supplementation, days‡ BPD‡,§ Severe BPD‡,¶ BPD or death Severe BPD or death Death before discharge Death while on trial intervention Age at death (d)
24 32 55 15 10 23 18 9 6 30
(10 to 41) (11 to 43) (31 to 86) (46) (30) (56) (44) (22) (15) (7 to 44)
Control (n = 42) 20 26 59 20 12 24 16 5 4 10
(9 to 43) (10 to 44) (36 to 106) (53) (32) (57) (38) (12) (9.5) (8 to 13)
P .703 .723 .531 .546 .908 .923 .591 .222 .520 .898
Mean difference or relative risk −2 7 3 0.86 0.96 0.98 1.15 1.84 1.54 23
(−11 to 6)* (−10 to 23)* (−28 to 34)* (0.53 to 1.40)† (0.48 to 1.93)† (0.67 to 1.43)† (0.69 to 1.94)† (0.68 to 5.04)† (0.47 to 5.05)† (−59 to 105)*
Data are median (IQR), n (%), *mean difference (95% CI), or †relative risk (95% CI). ‡Among infants alive at 36 weeks of age. §BPD defined as receiving oxygen for ≥28 days and at 36 weeks postmenstrual age. ¶Severe BPD defined as receiving oxygen for ≥28 days and ≥30% oxygen and/or positive pressure respiratory support at 36 weeks postmenstrual age.
of BPD, severe BPD, or the composites BPD or death and severe BPD or death did not differ between groups (Table II). The groups did not differ with regard to other morbidities and complications (Table III). Before successful extubation, 8 infants in each group had a failed extubation attempt that required reintubation within 24 hours (20% and 19% in the early caffeine and control group, respectively; P = .15). After the first successful extubation (end of the trial intervention), 13 infants in the early caffeine and 8 infants in the control group required reintubation at a later time while receiving open-label caffeine (20% and 19% in the early caffeine and control group, respectively; P = .24). Post hoc analyses to examine the effect of early caffeine on the primary endpoint, age at first successful extubation, and on death adjusting for gender, gestational age, and an Apgar score of <5 at 5 minutes were conducted using generalized linear models and multivariate binary logistic regression analysis. These analyses showed no association between the group assignment and any of the outcomes. Only gestational age was associated significantly with the age at first successful extubation and with an increased risk of death.
of mechanical ventilation in extremely premature infants with severe respiratory failure. Contrary to the hypothesis, early initiation of caffeine in these infants did not reduce the age of first successful extubation. Similarly, there were no differences between the early caffeine and control groups in secondary respiratory outcomes of duration of mechanical ventilation and oxygen supplementation or incidence of BPD. Although there is a strong rationale for the use of respiratory stimulants in infants with inconsistent respiratory drive, early caffeine administration did not hasten weaning from
Discussion This randomized trial prospectively evaluated the impact of early caffeine administration on the duration of the first course
Table III. Neonatal morbidities and other complications Outcome Pulmonary hemorrhage Patent ductus arteriosus Necrotizing enterocolitis* Spontaneous intestinal perforation Septicemia Intraventricular hemorrhage grade ≥ 3 Periventricular leukomalacia Retinopathy of prematurity ≥ stage 3†
Early caffeine (n = 41) 7 36 7 6 12 12 4 3
Data are n (%). *Necrotizing enterocolitis defined as Bell's stage II or III. †Among infants alive at 36 weeks of age.
(17) (88) (17) (15) (30) (30) (9.8) (9.1)
Control (n = 42) 4 38 2 5 10 6 1 5
(9.5) (91) (4.8) (12) (24) (14) (2.4) (13)
P .311 .738 .088 .714 .573 .098 .202 .716
Figure 3. Age at first successful extubation. Kaplan-Meier survival curves shows the proportion of infants who are still in their first course of mechanical ventilation in the 41 infants of the early caffeine group (solid line) and the 42 infants in the placebo control (dashed line) group. Vertical lines mark the median age at first successful extubation (caffeine: median, 24 [IQR, 9.541.0] vs control: median, 20 [IQR, 9.3-43.0]; P = .703 by KaplanMeier log-rank test). Circles indicate the time of each death in each group.
Early Caffeine and Weaning from Mechanical Ventilation in Preterm Infants: A Randomized, Placebo-Controlled Trial FLA 5.5.0 DTD ■ YMPD9716_proof ■ March 6, 2018
5
THE JOURNAL OF PEDIATRICS • www.jpeds.com mechanical ventilation. This finding is likely due to the greater influence of the underlying lung disease than the inconsistent respiratory drive in prolonging the need for ventilatory support in this particular population. An important factor to consider is that the trial intervention ended after the first successful extubation and, thereafter, all infants received open-label caffeine as standard of care. This point is important because one-third of the deaths in the caffeine group occurred after the first successful extubation when both groups were receiving caffeine. A large retrospective cohort study also reported higher mortality with early compared with late caffeine administration.11 However, as noted by the authors of that study, this difference may be in part explained by survival bias in the late caffeine group, that is, although all infants received caffeine in the early administration group, only those surviving the first days after birth received caffeine in the late administration group. The administration of a bolus of caffeine before extubation is standard practice in most centers. In this trial, the control group received a pre-extubation bolus of caffeine, whereas infants in the early caffeine group received a pre-extubation bolus of placebo. This practice was aimed at maintaining an equal chance for successful extubation in both groups. However, the pre-extubation caffeine bolus could potentially have conferred an advantage to the control group, because these infants may have had a higher caffeine level at the time of extubation. It is reassuring that the proportion of infants who had a failed extubation attempt before their first successful extubation was similar between the 2 groups. The total duration of mechanical ventilation was not influenced by the early administration of caffeine. This finding may be in large part explained by the lack of effect of caffeine on the first respiratory course. Because the trial intervention ended with the first successful extubation and all infants received caffeine before extubation and remained on caffeine thereafter, the group assignment should not influence the duration of secondary courses of mechanical ventilation. A different trial design comparing caffeine and placebo not only during the first course but throughout all courses of mechanical ventilation may have yielded different findings. However, such a comparison could be confounded by the prolonged halflife of the open-label caffeine administration between courses of mechanical ventilation. The findings from this trial need to be interpreted carefully owing to the smaller than planned statistical power that resulted from the early termination. This limitation is important, especially as it relates to mortality; any group imbalance occurring by chance may have a greater impact when evaluated with a smaller than planned sample size. Also, it is possible that the effect of caffeine on mortality and neonatal
Volume ■■ morbidities may be overestimated owing to the early termination of the trial.12 In summary, the early initiation of caffeine in this group of mechanically ventilated, extremely premature infants did not reduce the age of first successful extubation, the duration of need for supplemental oxygen, or BPD. An unexpected nonsignificant trend toward higher mortality in the caffeine group led to a cautionary decision to stop the trial. These findings suggest caution with early initiation of caffeine until larger, multicenter, randomized, controlled trials determine the efficacy and safety of this approach in this population. ■ We gratefully thank the infants and families that participated in the study, attending clinical teams, NICU pharmacy, and the DSMB members. Submitted for publication Aug 31, 2017; last revision received Nov 27, 2017; accepted Jan 3, 2018 Reprint requests: Nelson Claure, MSc, PhD, Division of Neonatology, Department of Pediatrics, University of Miami Miller School of Medicine, PO Box 016960 R-131, Miami, FL 33101. E-mail: [email protected]
References 1. Walsh MC, Morris BH, Wrage LA, Vohr BR, Poole WK, Tyson JE, et al. Extremely low birthweight neonates with protracted ventilation: mortality and 18-month neurodevelopmental outcomes. J Pediatr 2005;146:798804. 2. Henderson-Smart DJ, Davis PG. Prophylactic methylxanthines for endotracheal extubation in preterm infants. Cochrane Database Syst Rev 2010;(12):CD000139. 3. Henderson-Smart DJ, De Paoli AG. Methylxanthine treatment for apnoea in preterm infants. Cochrane Database Syst Rev 2010;(12):CD000140. 4. Thurston JH, Hauhard RE, Dirgo JA. Aminophylline increases cerebral metabolic rate and decreases anoxic survival in young mice. Science 1978;201:649-51. 5. Dunwiddie TV, Masino SA. The role and regulation of adenosine in the central nervous system. Annu Rev Neurosci 2001;24:31-55. 6. Boutilier RG. Mechanisms of cell survival in hypoxia and hypothermia. J Exp Biol 2001;204:3171-81. 7. Clark RH, Bloom BT, Spitzer AR, Gerstmann DR. Reported medication use in the neonatal intensive care unit: data from a large national data set. Pediatrics 2006;117:1979-87. 8. Schmidt B, Roberts RS, Davis P, Doyle LW, Barrington KJ, Ohlsson A, et al. Caffeine therapy for apnea of prematurity. N Engl J Med 2006;354:2112-21. 9. Schmidt B, Anderson PJ, Doyle LW, Dewey D, Grunau RE, Asztalos EV, et al. Survival without disability to age 5 years after neonatal caffeine therapy for apnea of prematurity. JAMA 2012;307:275-82. 10. Davis PG, Schmidt B, Roberts RS, Doyle LW, Asztalos E, Haslam R, et al. Caffeine for Apnea of Prematurity trial: benefits may vary in subgroups. J Pediatr 2010;156:382-7. 11. Dobson NR, Patel RM, Smith PB, Kuehn DR, Clark J, Vyas-Read S, et al. Trends in caffeine use and association between clinical outcomes and timing of therapy in very low birth weight infants. J Pediatr 2014;164:992-8, e993. 12. Viele K, McGlothlin A, Broglio K. Interpretation of clinical trials that stopped early. JAMA 2016;315:1646-7.
6
Amaro et al FLA 5.5.0 DTD ■ YMPD9716_proof ■ March 6, 2018
ORIGINAL ARTICLES
Early Caffeine and Weaning from Mechanical Ventilation in Preterm Infants: A Randomized, Placebo-Controlled Trial Cynthia M. Amaro, MD*, Jose A. Bello, MD*, Deepak Jain, MD, Alexandra Ramnath, MD, Carmen D'Ugard, RRT, Silvia Vanbuskirk, RN, Eduardo Bancalari, MD, and Nelson Claure, MSc, PhD Objective To evaluate in a randomized, double-blind, placebo-controlled trial the effect of early caffeine on the age of first successful extubation in preterm infants. Study design Preterm infants born at 23-30 weeks of gestation requiring mechanical ventilation in the first 5 postnatal days were randomized to receive a 20 mg/kg loading dose followed by 5 mg/kg/day of caffeine or placebo until considered ready for extubation. The placebo group received a blinded loading dose of caffeine before extubation. Results Infants were randomized to receive caffeine (n = 41) or placebo (n = 42). Age at first successful extubation did not differ between early caffeine (median, 24 days; IQR, 10-41 days) and control groups (median, 20 days; IQR, 9-43 days; P = .7). An interim analysis at 75% enrollment showed a trend toward higher mortality in 1 of the groups and the data safety and monitoring board recommended stopping the trial. Unblinded analysis revealed mortality did not differ significantly between the early caffeine (9 [22%]) and control groups (5 [12%]; P = .22). Conclusions Early initiation of caffeine in this group of premature infants did not reduce the age of first successful extubation. A nonsignificant trend toward higher mortality in the early caffeine group led to a cautious decision to stop the trial. These findings suggest caution with early use of caffeine in mechanically ventilated preterm infants until more efficacy and safety data become available. (J Pediatr 2018;■■:■■-■■). Trial Registration ClinicalTrials.gov NCT01751724.
A
large proportion of extreme premature infants require prolonged mechanical ventilation. This is associated with increased risk for bronchopulmonary dysplasia (BPD) and poor neurodevelopmental outcome.1 Caffeine, a methylxanthine and adenosine receptor antagonist, is a potent stimulant of central respiratory activity and is an effective treatment of apnea of prematurity and to avoid extubation failure.2,3 However, adenosine has been shown to preserve brain cell survival and prevent brain energy failure by preserving adenosine triphosphate levels in experimental hypoxia and ischemia models.4-6 Although caffeine has been used for these indications for the last 3 decades, until recently there were no long-term safety data to support its use.7 The large, randomized, controlled trial, Caffeine for Apnea of Prematurity, showed that caffeine started for these indications within the first 10 days after birth did not have long-term negative effects on neurologic outcome or survival.8,9 Analysis of predischarge outcomes showed a shorter duration of mechanical ventilation and a lower incidence of BPD in the caffeine compared with the placebo group. A secondary analysis showed that infants started on caffeine in the first 3 days while receiving mechanical ventilation derived the most benefit in terms of BPD and neurodevelopment.10 Since then, the use of caffeine has increased and often is started earlier in mechanically ventilated preterm infants.11 Strategies to provide mechanical respiratory support have changed from controlled ventilation to a more gentle support, where the ventilator is used to assist the infant’s spontaneous breathing. This step has been achieved in large part through the use of patient-triggered modes of mechanical ventilation. The effective use of these modes to reduce the need for mechanical ventilation must rely on the spontaneous respiratory drive, but this drive is frequently inconsistent in extremely premature infants. Thus, the use of respiratory stimulants such as caffeine seems to be a good alternative to decrease the duration of mechanical ventilation in this population. The efficacy and safety of this approach have not been evaluated by prospective, randomized, controlled trials. We postulated that early initiation of caffeine in mechanically ventilated preterm From the Division of Neonatology, Department of Pediatrics, University of Miami Miller School of Medicine, infants will hasten the weaning process during the first course of mechanical venMiami, FL tilation. The primary objective of this trial was to evaluate the effect of early caf*Contributed equally. Supported by the University of Miami Project NewBorn, a feine administration on the age of first successful extubation in mechanically philanthropic organization that did not participate in any aspect of the research. The authors declare no conflicts of interest.
BPD DSMB FiO2 NICU
Portions of this study were presented as an abstract at the Pediatric Academic Societies annual meeting, April 30-May 3, 2016, Baltimore, Maryland.
Bronchopulmonary dysplasia Data safety and monitoring board Fraction of inspired oxygen Neonatal intensive care unit
0022-3476/$ - see front matter. © 2018 Elsevier Inc. All rights reserved. https://doi.org10.1016/j.jpeds.2018.01.010
1 FLA 5.5.0 DTD ■ YMPD9716_proof ■ March 6, 2018
THE JOURNAL OF PEDIATRICS • www.jpeds.com ventilated preterm infants. The secondary objective was to evaluate the effect of early caffeine on total duration of mechanical ventilation and oxygen supplementation and on the incidence of BPD.
Methods This single-center, double-blind, placebo-controlled, randomized clinical trial (ClinicalTrials.gov: NCT01751724) was conducted in the neonatal intensive care unit (NICU) at Holtz Children’s Hospital of the Jackson Health System—University of Miami Medical Center. Premature infants born between 23 and 30 weeks of gestation who required mechanical ventilation in the first 5 postnatal days were eligible for the study. Infants with major congenital anomalies and infants that were small for gestational age (birth weight <3rd percentile) were excluded. Endpoints The primary endpoint of the trial was the age of first successful extubation, defined as postnatal age of first extubation after which the infant remained extubated for >24 hours. Secondary respiratory endpoints included total duration of mechanical ventilation, duration of oxygen supplementa-
Volume ■■
•
■■ 2018
tion, and the incidence of BPD. An oxygen or ventilator day was defined as the need for oxygen or mechanical ventilation for ≥12 hours over a 24-hour period. BPD was defined as need for supplemental oxygen for ≥28 days and at 36 weeks postmenstrual age. Severe BPD was defined as need for supplemental oxygen for ≥28 days and fraction of inspired oxygen (FiO2) of ≥0.30 and/or positive pressure respiratory support at 36 weeks postmenstrual age. Neonatal morbidities including pulmonary hemorrhage, echocardiography-confirmed patent ductus arteriosus, necrotizing enterocolitis defined as Bell’s stage II or III, spontaneous intestinal perforation, grade 3 or 4 intraventricular hemorrhage, periventricular leukomalacia, and retinopathy of prematurity stage 3 or 4 were documented. Interventions After screening and enrollment with parental consent during the first 5 days after birth, mechanically ventilated infants were randomized to the early caffeine or control group (Figure 1). Randomization was stratified according to gestational age, either 23-26 or 27-30 weeks, using sealed, opaque envelopes. Stratification was aimed at balancing the number of infants within each gestational age bracket. Randomization and study drug preparation were done by the NICU pharmacy staff. Knowl-
Figure 1. Trial flow diagram. Study protocol from screening to the end of the intervention. 2
Amaro et al FLA 5.5.0 DTD ■ YMPD9716_proof ■ March 6, 2018
■■ 2018
ORIGINAL ARTICLES
edge of treatment assignment was limited to the pharmacy staff. Investigators and clinical teams were unaware of the assigned treatment. Infants in the early caffeine group received a loading dose of caffeine citrate 20 mg/kg followed by a maintenance dose of 5 mg/kg/day. Infants in the control group received an equivalent volume of normal saline bolus and maintenance. The absolute dose of study drug was adjusted weekly to changes in weight. Routine ventilator management continued until the clinical team made the decision to extubate the infant. Infants in the control group received a blinded loading dose of caffeine citrate 20 mg/kg before extubation, as is routinely done in this NICU. Infants in the early caffeine group received a blinded loading dose of placebo. This was done to prevent excessive caffeine levels in this group. At this point, the investigators and clinical team remained blinded to the assigned intervention. If the infant remained extubated for >24 hours (definition of successful extubation), the study intervention ended and open-label caffeine at 5 mg/kg/day was given to all infants as per standard of care. Infants requiring reintubation in ≤24 hours remained in the assigned group and continued to receive blinded maintenance doses of caffeine or placebo. These steps continued until the infant was successfully extubated (as defined previously). Caffeine levels were not measured in the participating infants during the trial intervention period to avoid unmasking of the groups. Respiratory management followed unit guidelines. These measures include intubation and the administration of surfactant to infants with respiratory distress syndrome who require an FiO2 of >0.3 to keep arterial oxygen saturation between 90% and 95%. Subsequently, the ventilator mandatory rate and peak pressure were adjusted to maintain tidal volume between 4 and 7 mL/kg, pH between 7.25 and 7.4, and PaCO2 between 45 and 55 mm Hg during the first week after birth and between 45 and 65 mm Hg thereafter. Clinical teams were reminded to consider extubation when infants met the following extubation criteria: FiO2 of <0.4, positive endexpiratory pressure of ≤5 cm H2O, a mandatory rate of <20 cycles per minute, a peak pressure of <18 cm H2O, and pressure support of <10 cm H2O to keep arterial blood gases within these ranges. After extubation, clinical teams considered reintubation only if the infant met ≥1 of the following criteria: ≥2 episodes of severe apnea or bradycardia requiring bag-mask ventilation in a 4-hour period, an increase in FiO2 of >0.2 above the pre-extubation baseline and of >0.5, or an increase in PaCO2 of >15 mm Hg above the pre-extubation baseline and of >70 mm Hg. Enrolled infants were followed until 36 weeks postmenstrual age, discharge, or death, whichever came first. Statistical Considerations and Analyses The primary endpoint of the trial was the age at first successful extubation. We estimated that a total enrollment of 110 infants was needed to detect a reduction in this endpoint by >7 days, from a historical baseline of 15 ± 21 days, with 80% power and 5% significance.
Categorical variables were compared by the Pearson c2 or Fisher exact test. Continuous variables were compared by the Student t test or Mann-Whitney U test, depending on the distribution. The Kaplan-Meier log-rank test was used to compare time-to-event variables, such as age at first successful extubation. P < .05 was considered statistically significant. All statistical analyses were performed using IBM SPSS statistics package (IBM Corporation, Armonk, New York). Safety Monitoring Monitoring for side effects of caffeine was done as a part of routine clinical care. A dose of study drug could be withheld, decreased, or completely stopped in infants with persistent tachycardia (defined as a heart rate of >190 beats per minute for 4 hours after excluding other causes) or seizure activity (documented by electroencephalogram and not explained by other causes). The occurrence of these events was documented. An independent data safety and monitoring board (DSMB) was appointed. Interim analyses were planned at 50% and 75% enrollment to monitor safety. The DSMB board was presented with blinded group data on mortality and rates of adverse events. Ethics The trial was approved by the Institutional Review Board of the University of Miami Human Subjects Research Office and the Jackson Health System Office of Research. Written informed parental consent was obtained before enrollment.
Results The trial was conducted from January 2013 to December 2015. The CONSORT diagram in Figure 2 shows the flow of participants from screening to randomization; 208 infants admitted to the NICU were screened for eligibility and 143 were deemed eligible; 65 infants were excluded. Of these, 52 were not intubated during the first 5 days, 5 had congenital anomalies, and 1 was small for gestational age. Eighty-seven infants were enrolled after obtaining parental consent. Of these, 1 infant died before randomization and 86 infants were randomized. Three of the infants who were randomized were withdrawn from the trial, 2 of them by the parents, which prevented collection of any further data for inclusion in the intention to treat analyses. One randomized infant was withdrawn by the investigators because of extubation and the administration of open-label caffeine by the clinical team before starting the administration of the assigned study drug. Eighty-three randomized infants completed the trial intervention and follow-up, and were included in the analysis, 41 in the early caffeine and 42 in the control group. Table I shows the baseline characteristics of the population. The early caffeine and control groups did not differ with regard to gestational age, birth weight, ethnicity, or exposure to antenatal steroids. There was a greater proportion of male infants in the early caffeine compared with the control group and a trend toward a higher proportion of infants with a 5-minute Apgar score of <5 in the early caffeine group compared with the
Early Caffeine and Weaning from Mechanical Ventilation in Preterm Infants: A Randomized, Placebo-Controlled Trial FLA 5.5.0 DTD ■ YMPD9716_proof ■ March 6, 2018
3
THE JOURNAL OF PEDIATRICS • www.jpeds.com
Volume ■■
Figure 2. CONSORT diagram. Flow of trial participants from screening to randomization.
control group. Infants enrolled in the trial were intubated at a mean age of 3 hours and were randomized at a mean age of 48 hours. The respiratory support parameters at the time of randomization did not differ between groups. There were only 2 protocol deviations. One infant did not receive a blinded study drug bolus before extubation, but was extubated successfully. Another infant who failed extubation within 24 hours was mistakenly randomized for a second time when reintubated. This infant was analyzed in the group assigned in the original randomization.
Table I. Population demographics and characteristics at randomization Characteristic Gestational age (wk) Birth weight (g) Male Black Antenatal steroids 5-Minute Apgar score of <5 Age at intubation (h) Received surfactant Age at randomization (h) FiO2 at randomization Mean airway pressure at randomization
Early caffeine (n = 41)
Control (n = 42)
P
25.7 (24.3-27.0) 670 (605-915) 30 (73) 19 (46) 37 (90) 10 (24) 3.0 ± 8.2 39 (95) 48 ± 25 0.25 (0.21-0.37) 9 (8-11)
26.1 (24.2-28.4) 720 (643-894) 17 (41) 21 (50) 40 (95) 4 (9.5) 3.1 ± 7.0 40 (95) 49 ± 31 0.23 (0.21-0.30) 8 (8-10)
.425 .450 .003 .739 .433 .071 .449 1.000 .874 .195 .112
Data are mean ± SD, median (IQR), or n (%).
The first interim analysis (blinded to group assignment) conducted at 50% enrollment showed a trend toward a higher mortality rate in one of the groups. At the time, the mortality rate in both groups was below the center’s historic mortality rate in the 2 years preceding the trial for a similar population and the DSMB recommended that enrollment should continue. At the second interim analysis at 75% enrollment (83 infants), the DSMB recommended terminating the trial because of a persistent trend of higher mortality, although not statistically significant, in the same group. At this time, the mortality in this group slightly exceeded the historic mortality rate for a similar population in this center. In agreement with the DSMB, the investigators terminated enrollment in the trial. Unblinded analysis revealed a nonsignificant difference in mortality between the early caffeine and control groups (Table II). The proportion of infants who died before extubation (while on trial intervention) and the age of death did not differ significantly between groups. There were no instances where the study drug needed to be stopped or decreased owing to side effects. Analysis of the primary endpoint revealed no difference in the age at first successful extubation between the early caffeine (median, 24 days; IQR, 10-41 days) and control groups (median, 20 days; IQR, 9-43 days; P = .703 by Kaplan-Meier log-rank test; see Table III and Figure 3). Analysis of the secondary respiratory endpoints showed the total duration of mechanical ventilation or oxygen supplementation did not differ between the groups. The incidence
4
Amaro et al FLA 5.5.0 DTD ■ YMPD9716_proof ■ March 6, 2018
■■ 2018
ORIGINAL ARTICLES
Table II. Primary and secondary endpoints and mortality Outcome
Early caffeine (n = 41)
Age at first successful extubation (days) Total duration of mechanical ventilation, days‡ Total duration of oxygen supplementation, days‡ BPD‡,§ Severe BPD‡,¶ BPD or death Severe BPD or death Death before discharge Death while on trial intervention Age at death (d)
24 32 55 15 10 23 18 9 6 30
(10 to 41) (11 to 43) (31 to 86) (46) (30) (56) (44) (22) (15) (7 to 44)
Control (n = 42) 20 26 59 20 12 24 16 5 4 10
(9 to 43) (10 to 44) (36 to 106) (53) (32) (57) (38) (12) (9.5) (8 to 13)
P .703 .723 .531 .546 .908 .923 .591 .222 .520 .898
Mean difference or relative risk −2 7 3 0.86 0.96 0.98 1.15 1.84 1.54 23
(−11 to 6)* (−10 to 23)* (−28 to 34)* (0.53 to 1.40)† (0.48 to 1.93)† (0.67 to 1.43)† (0.69 to 1.94)† (0.68 to 5.04)† (0.47 to 5.05)† (−59 to 105)*
Data are median (IQR), n (%), *mean difference (95% CI), or †relative risk (95% CI). ‡Among infants alive at 36 weeks of age. §BPD defined as receiving oxygen for ≥28 days and at 36 weeks postmenstrual age. ¶Severe BPD defined as receiving oxygen for ≥28 days and ≥30% oxygen and/or positive pressure respiratory support at 36 weeks postmenstrual age.
of BPD, severe BPD, or the composites BPD or death and severe BPD or death did not differ between groups (Table II). The groups did not differ with regard to other morbidities and complications (Table III). Before successful extubation, 8 infants in each group had a failed extubation attempt that required reintubation within 24 hours (20% and 19% in the early caffeine and control group, respectively; P = .15). After the first successful extubation (end of the trial intervention), 13 infants in the early caffeine and 8 infants in the control group required reintubation at a later time while receiving open-label caffeine (20% and 19% in the early caffeine and control group, respectively; P = .24). Post hoc analyses to examine the effect of early caffeine on the primary endpoint, age at first successful extubation, and on death adjusting for gender, gestational age, and an Apgar score of <5 at 5 minutes were conducted using generalized linear models and multivariate binary logistic regression analysis. These analyses showed no association between the group assignment and any of the outcomes. Only gestational age was associated significantly with the age at first successful extubation and with an increased risk of death.
of mechanical ventilation in extremely premature infants with severe respiratory failure. Contrary to the hypothesis, early initiation of caffeine in these infants did not reduce the age of first successful extubation. Similarly, there were no differences between the early caffeine and control groups in secondary respiratory outcomes of duration of mechanical ventilation and oxygen supplementation or incidence of BPD. Although there is a strong rationale for the use of respiratory stimulants in infants with inconsistent respiratory drive, early caffeine administration did not hasten weaning from
Discussion This randomized trial prospectively evaluated the impact of early caffeine administration on the duration of the first course
Table III. Neonatal morbidities and other complications Outcome Pulmonary hemorrhage Patent ductus arteriosus Necrotizing enterocolitis* Spontaneous intestinal perforation Septicemia Intraventricular hemorrhage grade ≥ 3 Periventricular leukomalacia Retinopathy of prematurity ≥ stage 3†
Early caffeine (n = 41) 7 36 7 6 12 12 4 3
Data are n (%). *Necrotizing enterocolitis defined as Bell's stage II or III. †Among infants alive at 36 weeks of age.
(17) (88) (17) (15) (30) (30) (9.8) (9.1)
Control (n = 42) 4 38 2 5 10 6 1 5
(9.5) (91) (4.8) (12) (24) (14) (2.4) (13)
P .311 .738 .088 .714 .573 .098 .202 .716
Figure 3. Age at first successful extubation. Kaplan-Meier survival curves shows the proportion of infants who are still in their first course of mechanical ventilation in the 41 infants of the early caffeine group (solid line) and the 42 infants in the placebo control (dashed line) group. Vertical lines mark the median age at first successful extubation (caffeine: median, 24 [IQR, 9.541.0] vs control: median, 20 [IQR, 9.3-43.0]; P = .703 by KaplanMeier log-rank test). Circles indicate the time of each death in each group.
Early Caffeine and Weaning from Mechanical Ventilation in Preterm Infants: A Randomized, Placebo-Controlled Trial FLA 5.5.0 DTD ■ YMPD9716_proof ■ March 6, 2018
5
THE JOURNAL OF PEDIATRICS • www.jpeds.com mechanical ventilation. This finding is likely due to the greater influence of the underlying lung disease than the inconsistent respiratory drive in prolonging the need for ventilatory support in this particular population. An important factor to consider is that the trial intervention ended after the first successful extubation and, thereafter, all infants received open-label caffeine as standard of care. This point is important because one-third of the deaths in the caffeine group occurred after the first successful extubation when both groups were receiving caffeine. A large retrospective cohort study also reported higher mortality with early compared with late caffeine administration.11 However, as noted by the authors of that study, this difference may be in part explained by survival bias in the late caffeine group, that is, although all infants received caffeine in the early administration group, only those surviving the first days after birth received caffeine in the late administration group. The administration of a bolus of caffeine before extubation is standard practice in most centers. In this trial, the control group received a pre-extubation bolus of caffeine, whereas infants in the early caffeine group received a pre-extubation bolus of placebo. This practice was aimed at maintaining an equal chance for successful extubation in both groups. However, the pre-extubation caffeine bolus could potentially have conferred an advantage to the control group, because these infants may have had a higher caffeine level at the time of extubation. It is reassuring that the proportion of infants who had a failed extubation attempt before their first successful extubation was similar between the 2 groups. The total duration of mechanical ventilation was not influenced by the early administration of caffeine. This finding may be in large part explained by the lack of effect of caffeine on the first respiratory course. Because the trial intervention ended with the first successful extubation and all infants received caffeine before extubation and remained on caffeine thereafter, the group assignment should not influence the duration of secondary courses of mechanical ventilation. A different trial design comparing caffeine and placebo not only during the first course but throughout all courses of mechanical ventilation may have yielded different findings. However, such a comparison could be confounded by the prolonged halflife of the open-label caffeine administration between courses of mechanical ventilation. The findings from this trial need to be interpreted carefully owing to the smaller than planned statistical power that resulted from the early termination. This limitation is important, especially as it relates to mortality; any group imbalance occurring by chance may have a greater impact when evaluated with a smaller than planned sample size. Also, it is possible that the effect of caffeine on mortality and neonatal
Volume ■■ morbidities may be overestimated owing to the early termination of the trial.12 In summary, the early initiation of caffeine in this group of mechanically ventilated, extremely premature infants did not reduce the age of first successful extubation, the duration of need for supplemental oxygen, or BPD. An unexpected nonsignificant trend toward higher mortality in the caffeine group led to a cautionary decision to stop the trial. These findings suggest caution with early initiation of caffeine until larger, multicenter, randomized, controlled trials determine the efficacy and safety of this approach in this population. ■ We gratefully thank the infants and families that participated in the study, attending clinical teams, NICU pharmacy, and the DSMB members. Submitted for publication Aug 31, 2017; last revision received Nov 27, 2017; accepted Jan 3, 2018 Reprint requests: Nelson Claure, MSc, PhD, Division of Neonatology, Department of Pediatrics, University of Miami Miller School of Medicine, PO Box 016960 R-131, Miami, FL 33101. E-mail: [email protected]
References 1. Walsh MC, Morris BH, Wrage LA, Vohr BR, Poole WK, Tyson JE, et al. Extremely low birthweight neonates with protracted ventilation: mortality and 18-month neurodevelopmental outcomes. J Pediatr 2005;146:798804. 2. Henderson-Smart DJ, Davis PG. Prophylactic methylxanthines for endotracheal extubation in preterm infants. Cochrane Database Syst Rev 2010;(12):CD000139. 3. Henderson-Smart DJ, De Paoli AG. Methylxanthine treatment for apnoea in preterm infants. Cochrane Database Syst Rev 2010;(12):CD000140. 4. Thurston JH, Hauhard RE, Dirgo JA. Aminophylline increases cerebral metabolic rate and decreases anoxic survival in young mice. Science 1978;201:649-51. 5. Dunwiddie TV, Masino SA. The role and regulation of adenosine in the central nervous system. Annu Rev Neurosci 2001;24:31-55. 6. Boutilier RG. Mechanisms of cell survival in hypoxia and hypothermia. J Exp Biol 2001;204:3171-81. 7. Clark RH, Bloom BT, Spitzer AR, Gerstmann DR. Reported medication use in the neonatal intensive care unit: data from a large national data set. Pediatrics 2006;117:1979-87. 8. Schmidt B, Roberts RS, Davis P, Doyle LW, Barrington KJ, Ohlsson A, et al. Caffeine therapy for apnea of prematurity. N Engl J Med 2006;354:2112-21. 9. Schmidt B, Anderson PJ, Doyle LW, Dewey D, Grunau RE, Asztalos EV, et al. Survival without disability to age 5 years after neonatal caffeine therapy for apnea of prematurity. JAMA 2012;307:275-82. 10. Davis PG, Schmidt B, Roberts RS, Doyle LW, Asztalos E, Haslam R, et al. Caffeine for Apnea of Prematurity trial: benefits may vary in subgroups. J Pediatr 2010;156:382-7. 11. Dobson NR, Patel RM, Smith PB, Kuehn DR, Clark J, Vyas-Read S, et al. Trends in caffeine use and association between clinical outcomes and timing of therapy in very low birth weight infants. J Pediatr 2014;164:992-8, e993. 12. Viele K, McGlothlin A, Broglio K. Interpretation of clinical trials that stopped early. JAMA 2016;315:1646-7.
6
Amaro et al FLA 5.5.0 DTD ■ YMPD9716_proof ■ March 6, 2018