Dexmedetomidine versus fentanyl for sedation in extremely preterm infants

Few studies have compared the efficacy and complications of dexmedetomidine (DEX) and fentanyl (FEN) in extremely preterm infants.


I N T RODUC T ION
Proper sedation reduces stress and prevents complications in neonates who undergo uncomfortable procedures such as endotracheal intubation and mechanical ventilation. 1 However, exposure to sedative drugs may have negative effects on the developing brain. 2,3Although clinicians must balance patient comfort with the potentially negative consequences, the clinical evidence and guidelines are limited.
0][11] Respiratory depression lengthens the duration of mechanical ventilation, which may increase the risk of bronchopulmonary dysplasia.Unstable blood pressure is associated with a higher risk of intraventricular hemorrhage (IVH). 11ecreased movements and gastrointestinal function lead to abnormal bowel patterns, meconium plug syndrome, and delayed full enteral feeding. 10In addition, a significant decrease in scores on certain developmental scales has been reported with continuous FEN infusion. 12exmedetomidine (DEX) is a highly selective, centrally acting α-2 agonist that achieves sedation by reducing sympathetic outflow. 6It causes analgesia by inhibiting substance P release, and it shares a potassium channel with opioid receptors.The use of DEX in neonates is appealing because it is associated with reduced opioid and benzodiazepine use, neuroprotective effects in animal models, shorter mechanical ventilation durations, and earlier enteral feeding, which reduces the incidence of necrotizing enterocolitis (NEC). 6,13][15] Due to concerns about opioid use, for example, potentially harmful effects on the central nervous system and the need for their strict management due to legal requirements in Japan, in 2015 our neonatal intensive care unit (NICU) began using DEX instead of FEN as the first-line sedative agent in preterm infants.Against this background, we performed a single-institution, retrospective controlled before-and-after study to assess the relationship between first-line sedative selection and the clinical course of preterm infants.

M ET HOD S
We conducted a single-institution, retrospective controlled before-and-after study of infants admitted to the NICU of Fujita Health University Hospital between April 2010 and December 2018.We enrolled preterm infants whose gestational age was between 22 weeks 0 days and 27 weeks 6 days, and who were administered sedative drugs within 24 h after birth.Exclusion criteria were as follows: born in another hospital, the presence of a major congenital or genetic anomaly, and death or a diagnosis of IVH before the initiation of sedation.Between April 2010 and December 2014, FEN was used as the first-line sedative.It was administered as a continuous infusion at a dosage of 0.5 μg/kg/h, which could be increased up to 2 μg/kg/h.Phenobarbital (PB) was administered for additional sedation at 5 mg/kg if needed.Between January 2015 and December 2018, DEX was used as the first-line sedative agent.It was given as a continuous infusion at a dosage of 0.3 μg/kg/h without loading, which could be increased up to 0.7 μg/kg/h.As with FEN, PB could be administered for additional sedation at 5 mg/ kg.Beginning 72 h after birth, both FEN and DEX were gradually tapered and discontinued.In both periods, the indication for PB administration was when the patient's circulation and respiration became unstable with the state on the neonatal behavioral assessment scale being 5 (alert and active state) or 6 (awake and distressed state), 16 and the nurse's intervention did not decrease the state or improve circulation and respiration, and the final decision to administer PB was made by the physician.
The clinical data of subjects were extracted from electronic medical records.The following patient characteristics were evaluated: gestational age (GA; weeks); birthweight (g); Apgar score at 5 min after birth; sex; multiple births; current or previous maternal chorioamnionitis; small for gestational age (SGA), defined as a birthweight below the 10th percentile for Japanese neonates; 17 respiratory distress syndrome; bronchopulomary dysplasia (BPD), defined as O 2 inhalation or any respiratory support at 36 weeks of postmenstrual age; dry lung syndrome; inhalation of nitric oxide during the first 72 h after birth; administration of hydrocortisone during the first 24 h after birth, which was basically routine unless the physician decided not to administer it due to high blood pressure or other reason; postmenstrual weeks at extubation; use of dopamine (DOA); dobutamine (DOB); nitroglycerine (NTG) and days of age at which full enteral feeding was achieved (100 mL/kg/day).
Information was obtained about the following complications: late-onset circulatory collapse (LCC), defined as refractory hypotension occurring after 7 days of life; 18 patent ductus arteriosus (PDA) requiring administration of a cyclooxygenase inhibitor; PDA requiring surgical closure; any stage of NEC; 19 localized intestinal perforation (LIP); 20 any grade of intraventricular hemorrhage (IVH) or severe IVH (grade 3 or 4); 21 death during hospitalization; and a total developmental quotient (DQ) < 70 at a corrected age of 3 years ± 3 months, as assessed by the Kyoto Scale of Psychological Development (KSPD). 22KSPD was administered by experienced testers who were certified psychologists and blinded to perinatal details during hospitalization.The DQ was derived by developmental age defined for each task dividing by chronological age.A DQ score < 70 was interpreted as significantly delayed performance.
Data were collected on the duration of sedative use and the addition of sedatives.Furthermore, to evaluate the acute cardiovascular effects of sedatives, systolic blood pressure (sBP), diastolic blood pressure (dBP), and mean blood pressure (mBP) were obtained by arterial blood pressure measurement, and heart rate (HR) was measured before sedative use and at 1, 6, 12, 24, and 48 h after sedative initiation.
In this study, we compared clinical courses between the DEX and FEN groups.A composite outcome of death during hospitalization and DQ < 70 at a corrected age of 3 years was the primary outcome.The secondary | 3 of 9 DEX VERSUS FEN FOR PRETERM INFANTS outcomes were death during hospitalization; DQ < 70 at a corrected age of 3 years; duration of mechanical ventilation; days of age at which full enteral feeding was achieved; use of DOA, DOB, and NTG; PDA requiring medication or surgical intervention; and intestinal complications, specifically NEC, LIP, LCC, IVH and additional use of secondary sedatives.Hemodynamically parameters including sBP, dBP, mBP and HR were also compared.
Mann-Whitney U test and Fisher's exact test were used for univariate analysis.For outcomes, which were binary variables, logistic regression analysis was performed for GA which was generally considered to have a great impact 23 and the selection of sedatives, when the number of applicable cases was 20 or more.Because the number of relevant cases for each endpoint was expected to be limited based on the overall number of patients, the variables used for logistic regression analysis were limited.For continuous variables, the postmenstrual weeks at extubation and the days of age at which full enteral feeding was achieved, multiple regression analysis was performed.As explanatory variables, GA and SGA, which have been shown in the past to significantly affect each of the objective variables, 24,25 were used along with the choice of sedative medication.Linear mixed models were used to compare the longitudinal values of sBP, dBP, mBP, and HR in the DEX and FEN groups.To estimate the change in each value, these models included sedative selection, time since sedative initiation, and the interaction term between the covariates of sedative selection and time.Changes in values in the DEX and FEN groups were also evaluated by repeated measures ANOVA.All statistical analyses were performed with EZR statistical software, version 1.61. 26his study was approved by the ethics committee of Fujita Health University (HM20-568).

R E SU LT S
During the study period, 1643 infants were admitted to our NICU, of whom 81 were enrolled in this study (Figure 1).Fifteen infants were excluded, for the following reasons: born in another hospital (n = 2), major congenital anomaly (n = 3), chromosomal anomaly (n = 1), and death or IVH before the initiation of sedation (n = 9).Thirty-three infants received FEN as the first-line sedative drug, while 33 received DEX.The characteristics of the FEN and DEX groups are described in Table 1.The gestational age was significantly lower in the DEX group than in the FEN group, but there were no other significant differences between the two groups.Sixteen subjects did not undergo developmental assessment at a corrected age of 3 years, due to death during hospitalization, being followed up at other hospitals, or not receiving the assessment at the appropriate time.
The clinical information during hospitalization and the results of developmental assessment at a corrected age of 3 years are summarized in Table 2.The median treatment duration was not significantly different between the FEN and DEX groups (94 h and 91 h, respectively).The mean infusion rate of DEX was 0.426/μg/kg in this study.The number of cases administered additional sedative drugs and the doses were significantly fewer in the DEX group than in the FEN group (27.3% vs. 54.5%),and that was also significant after adjusted by GA (p = 0.013).The number of PB doses with additional sedative cases were also significantly fewer in the DEX group (p = 0.006).The DEX group, compared to the FEN group, underwent extubation at a higher number of postmenstrual weeks (30 vs. 29 weeks) and required a longer period to achieve full enteral feeding (14 vs. 11 days).Regarding the use of circulatory agents, the DEX group more frequently required the administration of NTG (51.5% vs. 12.1%) than the FEN group.However, after adjusted by GA, not only NTG use but also DOA use was significantly more frequent by sedative selection (p = 0.003 and p = 0.048, respectively).The incidence of LCC was lower in the DEX group than in the FEN group (45.5% vs. 78.8%),and it was also significant after adjusted by GA (p = 0.023).There were no significant differences between groups in the incidences of BPD, intestinal complications such as NEC and LIP, any grade of IVH, death, DQ < 70 at a corrected age of 3 years, or the composite outcome of death and DQ < 70 at a corrected age of 3 years.
Since the DEX group was associated with a higher number of postmenstrual weeks at extubation and a longer time to achieve full enteral feeding, we performed multiple regression analysis of the number of postmenstrual weeks at extubation and the days of age to achieve full feeding, using normal log transformation with GA, SGA, and sedative selection as the explanatory variables (Tables 3 and 4).This analysis revealed that while SGA was significantly associated with extubation timing, sedative selection was not.On the other hand, sedative selection and GA were significantly associated with days of age to achieve full feeding.
Regarding the acute cardiovascular effects of sedatives, linear mixed models showed no significant differences in sBP, dBP, mBP, or HR according to sedative selection.On the other hand, group analyses with repeated measures ANOVA demonstrated that over the course of 48 h, HR significantly decreased in both groups and sBP significantly increased in the FEN group (Figure 2).

DI SC US SION
This was a single-institution, retrospective, beforeand-after study that compared the continuous infusion of DEX or FEN for sedation in extremely preterm infants whose gestational age was below 28 weeks.Our findings showed that compared to FEN, DEX did not increase the incidences of mortality, DQ < 70 at a corrected age of 3 years, or the composite outcome of both.The groups also showed nonsignificant differences regarding intestinal and central nervous complications.One important aspect of this study concerns the method of DEX administration for preterm infants.There are few reports of DEX administration to preterm infants.O'Mara et al. 13 reported that DEX was initiated at 0.3 μg/kg/h for infants at a mean gestational age of 25 weeks within 48 h of life and administered at a mean infusion rate of 0.6 μg/kg/h for 12 days on average.Chrysostomou et al. 14 reported on DEX administration at 0.05-0.2μg/kg/h with a duration of 24 h, but their target preterm infants were those who were 28 weeks or longer in gestation.The initial and average doses of DEX in this study were not so different from those reported by O'Mara et al. but the duration of administration was shorter.
Regarding important outcomes such as death and developmental delays, there have been few studies of preterm infants treated with continuous DEX. 13,27In this study, DEX did not worsen mortality or developmental outcomes compared to FEN.A previous study in preterm infants showed that continuous infusion of FEN did not increase mortality compared to placebo. 28owever, while one study found no relationship between FEN dose and developmental scores, 29 another reported that a higher dose was associated with worse eye and hand coordination skills at a modified age of 2 years. 12cPherson et al. 30 reported that a higher cumulative FEN dose before term-equivalent age correlated with a lower transverse cerebellar diameter, while there was no correlation between the cumulative FEN dose and development at 2 years.In a neonate rat model, early morphine exposure reduced the levels of brain-derived neurotrophic factor, a marker of synaptic plasticity and a modulator of cognitive function. 31Another study showed that morphine enhanced apoptosis in certain regions of the neonatal rat brain. 32Also, FEN administration to rats in early life promoted changes in behavioral responses. 33In contrast to opioids, DEX had a neuroprotective effect in a murine model of perinatal excitotoxic brain injury.The hemodynamic, respiratory, and gastrointestinal depressant effects of sedatives are of significant concern, especially in preterm infants.First, regarding hemodynamics, in this study neither FEN nor DEX was associated with a significant change in blood pressure or HR, but both were correlated with a significant decrease in HR.On the other hand, the inotropic agents DOA and NTG were used more frequently in the DEX group.Ancora et al. 28 compared FEN and placebo groups in preterm infants and found that the mBP was lower in the placebo group at day 7, but did not differ otherwise, nor did the frequency of vasopressor use.However, that study did not evaluate blood pressure variability in the relatively short time after FEN administration.Chrysostomou et al. 14 reported a 12% decrease in HR and a 14% decrease in sBP in the 24 h after DEX administration to preterm infants, but none of the effects were problematic enough to interrupt treatment.Similarly, Dersch-Mills et al. 34 reported that 80% of preterm infants receiving DEX had bradycardia and 41% had hypotension requiring intervention, but DEX did not need to be discontinued.However, other studies reported that DEX administration to preterm infants had no hemodynamic effects. 13,15Thus, there is no established viewpoint regarding the effect of DEX on hemodynamics in preterm infants.Considering these results, although our report showed more use of cardiovascular agents in the DEX group, this may be because at our institution, acute management is performed while assessing cardiac function with echocardiography, which is widely practiced in Japan. 35,36This therapeutic strategy seems to result in the safe use of sedatives.
In terms of respiratory effects, specifically extubation timing, our study identified no advantages or disadvantages of DEX relative to FEN.The two drugs also showed no significant difference in the incidence of bronchopulmonary dysplasia.O'Mara et al. 13 reported that DEX was associated with a significantly shorter ventilation duration than FEN.They stated that this was unlikely to be due entirely to decreased respiratory drive suppression with DEX and hypothesized that it could have been the result of DEX decreasing lung tissue damage caused by interleukin (IL)-1β and cyclooxygenase-2 (also known as prostaglandin E2), two important mediators of ventilator-induced lung injury whose production was decreased by DEX in a rat model. 37'Mara et al. 13 also mentioned advantages of DEX regarding gastrointestinal function, such as fewer days until achieving full enteral feeding.In our study, however, DEX was associated with significantly more days to achieve full feeding than FEN, even after adjustment for SGA and GA.Our result indicated that the actual difference in reaching full feeding of 1.2 days by of the selection of sedatives.Additionally, no obvious significant difference, odds ratio of NEC, LIP and both of them were high between DEX group and FEN group.In our study, there were not enough cases to evaluate other clinical factors besides sedatives even existed a difference in GA, which might influence intestinal complications.Consideration to our result that differs from the previous study, further study is needed to determine the effects of DEX on intestinal function.
Another clinical benefit of DEX is that it is associated with reduced administration of additional sedation medications, which was confirmed in our study and in that by O'Mara et al. 13 In this study, both the number of cases with additional PB administration and the number of PB doses with additional sedative cases were fewer in DEX group.This may be an advantage in terms of lowering the negative effects on neurodevelopment caused by other sedatives including PB. [30][31][32]38 It should also be noted that the incidence of LCC was low in the DEX group in this study, and even the incidence of LCC was relatively high compared to mean value of Japanese NICUs. 39lthough some interinstitutional disparities in the incidence of LCC have been reported, factors contributing to this disparity have not been clarified, nor have sedative medications been examined.39 Additionally, the factors contributing to the development of LCC in preterm infants are not clearly understood, and it is thought that it involves the inability to cope with stressors due to immature adrenal function, resulting in relative adrenal insufficiency that then causes hypotension and other symptoms.18 The reason for fewer incidences of LCC in the DEX group in our study was not clear.It is entirely hypothetical, DEX has been reported to suppress invasive immune responses, [40][41][42][43] and although the mechanism is uncertain, this may be linked to reduced stress caused by inflammation in preterm infants, thereby reducing the incidence of LCC at our institute, which had a relatively high incidence rate.Of course, further studies are needed to clarify the mechanism.
One limitation of this study is that it used a retrospective, controlled before and after study design.Therefore, gestational age, a very important factor related to prognosis in infants, differed between the DEX and FEN groups.However, other than the switch from FEN to DEX, there were no other changes in treatment strategy.Another limitation of this study is the small sample size, which made it possible to analyze only a limited number of explanatory variables in multivariate analysis.In addition, longerterm developmental prognosis to be evaluated to further assess the benefit from the neuroprotective effects noted in the animal models.Despite these limitations, we believe this study is valuable because it is the first to consider the developmental effects of FEN and DEX in preterm infants born before 28 weeks of gestation, who may require sedation due to their severe condition.Additionally, repeated measures ANOVA analysis of hemodynamic variables for both groups showed significance in systolic blood pressure and heart rate for the fentanyl group and heart rate for the dexmedetomidine group.† Indicates that p-values were less than 0.05 compared to the values at treatment initiation with each sedative, as determined using Bonferroni's multiple comparison test.Abbreviations: bpm, beats per minute; dBP, diastolic blood pressure; DEX, dexmedetomidine; FEN, fentanyl; HR, heart rate; mBP, mean blood pressure; sBP, systolic blood pressure.

CONC
Compared to FEN, DEX did not increase the incidences of short-term complications, death, or developmental delay at a corrected age of 3 years in preterm infants born before 28 weeks of gestation.This study confirmed the advantages of DEX, such as the need for fewer additional sedatives and a lower incidence of LCC, but the benefit regarding gastrointestinal function unclear.Prospective randomized controlled trials that evaluate the long-term effects of DEX on development are needed to further clarify its usefulness.

T HOR CON T R I BU T ION S
C.N. and M.Mi.contributed to the conception and design of this study.C.N., M.Mi., S.K., Y.F., M.Ma., A.K., Y.K., H.U., M.F., and H.B. collected the data.C.N. and M.Mi.analyzed the data and wrote the manuscript.All authors read and approved the final manuscript.

AC K NOW L E DGM E N T S
The main findings of this research were presented at the 125th Annual Meeting of the Japan Pediatric Society, April 15-17, 2022, in Koriyama City, Fukushima Prefecture, Japan.

CON F L IC T OF I N T E R E ST STAT E M E N T
The authors declare no conflict of interest.

F I G U R E 2
The changes in each hemodynamic parameter of the two groups after sedative initiation were not significantly different when analyzed by linear mixed models.Error bars represent standard deviations.p-values indicate results of analysis by linear mixed models.
Characteristics of the fentanyl and dexmedetomidine groups.
T A B L E 1Note: Data are presented as medians (interquartile range) or percentages.p-valuesrepresent results of the Mann-Whitney U test and Fisher's exact test.Abbreviations: DEX, dexmedetomidine; FEN, fentanyl; iNO, inhalation of nitric oxide.
Comparison of clinical information during hospitalization and developmental assessment at a corrected age of 3 years.
Multiple regression analysis of postmenstrual weeks at extubation.
Note: p-values represent results of multiple regression analysis of days of age at full enteral feeding after normal log transformation performed with sedative selection, SGA, and GA as explanatory variables.The adjusted R-squared in this regression equation is 0.152.Abbreviations: DEX, dexmedetomidine; FEN, fentanyl; GA, gestational age; SGA, small for gestational age.