Providing adequate sedation for critically ill infants with heart disease is a challenging clinical problem. These patients often have labile cardiovascular function and may require several days to weeks of sedation as a component of their intensive care unit management. The challenge remains to find an ideal pharmacologic agent that is not only effective but also safe, without alteration of a patient’s hemodynamic status.

Although it is proposed by some that dexmedetomidine (DEX) is a useful addition to the sedation armamentarium of pediatric cardiac intensive care physicians, the effect of this agent on the hemodynamics of infants is unknown. As an α2-adrenergic agonist of the imidazole subclass, DEX (Precedex, Hospira Worldwide Inc, Lake Forest, IL, USA) is similar in structure to clonidine but with an α2:α1 specificity of nearly 1,600 to 1 [21]. It acts as a sedative, anxiolytic, and analgesic agent through α2-adrenergic agonism in the locus coeruleus [21]. The shorter half-life of DEX (2–3 vs 12–24 h for clonidine) and its availability for administration as an intravenous (IV) agent allows titration by continuous infusion [22].

Because DEX has limited effects on respiratory function, interference with weaning from mechanical ventilation is less likely. In fact, it has been used both as a bridge to tracheal extubation and as sedation for nonintubated patients [7, 8].

Enthusiasm for this newer agent stems from several factors including the lack of an ideal agent for sedation during mechanical ventilation, adverse effects associated with existing agents, and recognition over the past 10 years of the potential deleterious physiologic effects of untreated pain in the acute care setting [1, 20].

Initially, DEX received approval from the Food and Drug Administration (FDA) for the provision of sedation up to 24 h in adults during mechanical ventilation. However, the safety and hemodynamic profile of this drug for neonates and infants with critically ill heart disease still is poorly understood, particularly for infants with a single-ventricle anatomy, those undergoing complex cardiac surgeries, and those receiving an orthotopic heart transplant. In this study, we aimed to address some of these limitations by focusing on the course and hemodynamics of critically ill infants with heart disease who received infusions of DEX during their intensive care course.

Patients and Methods

We retrospectively reviewed the charts of all neonates and infants who received DEX in our pediatric cardiovascular intensive care unit (CVICU) between August 2009 and June 2010. The study population included all critically ill neonates and infants (age, ≤12 months) with congenital or acquired heart disease. Using the records from our pharmacy, we identified all the infants who received DEX during their CVICU stay.

The study excluded those who received any part of their DEX infusion outside the CVICU setting (i.e., operating room, floor, or radiologic suite), those with incomplete or missing medical records, those who did not have an arterial catheter for invasive blood pressure monitoring, those who received DEX infusion for less than 24 h, those who received extracorporeal membrane oxygenation (ECMO) partly or completely during DEX infusion, and those with “do not resuscitate” orders. The Institutional Review Board at Arkansas Children’s Hospital approved the study protocol, and the need for informed consent or assent was waived.

According to routine clinical practice in our CVICU, DEX was initiated as a second or third continuous-infusion sedative after the perceived failure of conventional sedation agents. The decision to initiate a DEX infusion was at the discretion of the medical team caring for the patient. The most common conventional agents used in our CVICU include midazolam infusion, morphine infusion, or both. We initiated DEX as continuous infusion at a dose of 0.1–1.5 mcg/kg/h without a bolus dose.

The daily dose of DEX was determined by averaging the amount of drug each patient received over 24 h and was expressed as mcg/kg/day. Midazolam and morphine doses were determined by averaging the amount of each drug each patient received over 24 h and were expressed as mg/kg/day. The data for patients receiving more than one course of DEX per hospital admission were evaluated as separate events, provided the time between each infusion period was more than 12 h.

We collected demographic data including age, weight, sex, diagnosis, and Risk Adjustment in Congenital Heart Surgery (RACHS-1) score [15]. We also collected detailed information on the dosages of DEX, midazolam, and morphine administered by infusion in the first 24 h, in 24–48 h, and during the last 24 h before discontinuation of the therapy, together with the total duration of infusion for each drug. Given the lack of an established comfort or pain score in our unit, the infusions were adjusted according to the clinical assessment of the physician caring for the patient.

We also collected physiologic data including heart rate, mean arterial pressure (MAP), respiratory rate, systemic oxygen saturation by pulse oximetry, inotrope score [16], near-infrared spectroscopy (NIRS), and central venous pressure (CVP). It is a common practice in our unit to titrate inotropes according to signs and symptoms of cardiac output (e.g., capillary refill, urine output, mixed venous oxygen saturation, serum lactate, and/or arterial blood pH) and not solely on the basis of blood pressures. The baseline data for these variables were collected for 6 h before initiation of DEX (expressed as the mean value over 1 h), for the first hour after initiation of DEX, for 2–6 h after initiation of DEX (expressed as the mean value over 1 h), for 7–24 h after initiation of DEX (expressed as the mean value over 1 h), for the last 6 h before termination of DEX (expressed as the mean value over 1 h), for the first hour after termination of DEX, for the sixth hour after termination of DEX, and for the 24 h after termination of DEX.

The potential side effects evaluated in our study were nausea, vomiting, abdominal distension, dysrhythmias, neurologic abnormalities, seizures, and signs and symptoms of withdrawal. Additional variables evaluated to assess the safety of DEX were blood pH, partial pressure of oxygen (PaO2), partial pressure of carbon dioxide (PaCO2), and serum bicarbonate before initiation of DEX infusion, during the infusion, and after the termination of the infusion.

To assess the efficacy of DEX for sedative and analgesic requirements, we also examined the number of rescue boluses for each category before initiation of the DEX infusion, during the DEX infusion, and after termination of the DEX infusion. Given the lack of an established protocol for sedation management in our unit, the rescue agents were multiple and included midazolam, lorazepam, ketamine, chloralhydrate, and diphenhydramine for sedation and fentanyl and morphine for analgesia. For some patients, narcotics were used for both analgesia (primary effect) and sedation (sedation effect). Rescue agents were labeled as either a “sedation bolus” or an “analgesic bolus” depending on their primary effect.

Statistical Analysis

Continuous variables are presented as medians (Q1, Q3), where Q1 is the 25th percentile, and Q3 is the 75th percentile and/or means ± standard deviations, whereas categorical variables are presented as numbers and percentages. Calculation of p values was performed using the chi-square test or Fisher’s exact test of independence for categorical variables and the Wilcoxon rank-sum test for continuous variables. Quantitative continuous variables were compared between groups using the Mann–Whitney nonparametric test if the variable had a non-normal distribution or using the unpaired Student’s t test if the variable had a normal distribution.

Serial measurements including heart rate, MAP, inotrope score, CVP, NIRS, and respiratory rate were analyzed using random-effects mixed-linear models to account for repeated measures. Separate linear mixed models were suitable for comparing the outcomes over time between neonates and infants, between one- and two-ventricle anatomies, and between heart transplantation and no heart transplantation. Repeated measures of analyses were performed using the MIXED procedure in SAS. A compound symmetry covariance structure was used to account for correlations of measurements of the same subject. A p value of 0.05 was considered significant for study purposes. However, given that there were eight group comparisons (one per period), a Bonferroni corrected alpha level of 0.00625 (0.05/8) was used to control for multiple testing within each model. The p values were not adjusted for multiple testing. Analyses were performed using STATA/MP, version 12 (Stata Corp LP, College Station, TX, USA) or SAS version 9.2 (SAS Institute Inc., Cary, NC, USA).

Results

During the study period, 50 neonates and infants received DEX. Of these 50 patients, 27 were males (54%) and 13 were neonates (26%). The patients had an average age of 3.53 ± 2.64 months and a weight of 4.85 ± 1.67 kg. Whereas 43 patients (86%) received DEX while intubated on a mechanical ventilator, 1 patient received DEX while undergoing noninvasive ventilation. Administration of DEX was after surgery for congenital heart disease (CHD) for 34 patients (68%) and after heart transplantation for 15 patients (30%). Of these 50 infants, 10 (20%) had a single-ventricle anatomy, whereas 13 (26%) had risk adjustment scores (RACHS-1) in the categories of 4 to 6.

Administration of DEX was as a continuous infusion for a median period of 78 h (range, 40–290 h). The patients received an average dose of 9.1 ± 2.8 mcg/kg/day during the first 24 h, 8.7 ± 4.8 mcg/kg/day during the subsequent 24 h, and 7.2 ± 4.8 mcg/kg/day during the last 24 h before termination of DEX usage. The median CVICU stay was 29 days (range, 8–69 days).

Hemodynamic Effects

All the patients remained hemodynamically stable during the DEX infusion. Hemodynamic stability was defined as no need for escalation of inotropic support or IV fluid boluses and no need for holding or decreasing the DEX dose during the infusion. There was a significant decrease in heart rate, MAP, inotrope score, and CVP throughout the DEX infusion.

Heart Rate

The average heart rate increased from a baseline value of 143 to a value of 145 in the first hour after initiation of DEX (p = 0.60), followed by a decrease to 139 in 2–6 h (p = 0.16), to 131 in 7–24 h (p < 0.001), and to 124 in last 6 h (p < 0.001) of DEX infusion. The heart rate started to increase toward baseline values after termination of DEX infusion, although the values still remained significantly lower than the baseline values, with an average heart rate of 134 in 24 h after termination of DEX (p = 0.008) (Fig. 1).

Fig. 1
figure 1

Effects of dexmedetomidine (DEX) infusion on the heart rate of critically ill neonates and infants with heart disease

Full size image

Comparison of the neonates and infants showed no significant difference in heart rate until last 6 h of DEX infusion, when the infants had a significantly lower heart rate (p = 0.001), which continued until 6 h after termination of DEX infusion (p = 0.002) (Table 1). There was no substantial difference in heart rate at any time point between patients with one- and two-ventricle anatomies (Table 1). The patients receiving a heart transplant had a lower heart rate before and during DEX infusion (Table 1). There was no significant difference in heart rate at any time point between patients undergoing complex surgery in RACHS categories 4 to 6 and those undergoing surgery in RACHS categories 1 to 3.

Table 1 Comparisons of hemodynamic variables in three subpopulations among critically ill infants with cardiac disease
Full size table

MAP and Inotrope Score

The average MAP increased from a baseline value of 69 to a value of 70 in the first hour after initiation of DEX (p = 0.34), followed by a decrease to 64 in 2–6 h (p = 0.02) to 61 in 7–24 h (p = 0.0001), and to 60 in last 6 h (p < 0.001) of DEX infusion. After termination of the DEX infusion, MAP started to increase toward baseline values, with the values reaching a baseline MAP of 66 in 24 h (p = 0.12) (Fig. 2).

Fig. 2
figure 2

Effects of dexmedetomidine (DEX) infusion on the mean arterial pressure of critically ill neonates and infants with heart disease

Full size image

Comparison of the neonates and infants showed that the neonates had a lower MAP at baseline (p = 0.01), which continued to be lower throughout the DEX infusion (Table 1). There was no substantial difference in MAP for patients with a one- versus a two-ventricle anatomy or for those undergoing surgery in the RACHS 4–6 category versus the RACHS 1–3 category (Table 1). The patients receiving a heart transplant had a higher MAP at baseline (p < 0.001), which continued to be higher throughout the DEX infusion (Table 1).

The inotrope score also followed the same trends as MAP, with the baseline value of 8.2 decreasing to 7.6 in first hour after initiation of DEX (p = 0.32), followed by a decrease to 7.4 in 2–6 h (p = 0.14), to 7.3 in 7–24 h (p = 0.10), and to 6.5 in the last 6 h (p = 0.002) of DEX infusion. The inotrope score started increasing toward baseline values after termination of the DEX infusion, with the values reaching the baseline value of 7.7 in 24 h after the termination of DEX (p = 0.36) (Fig. 3). The inotrope scores did not differ substantially between the patients in the different subgroups (Table 1).

Fig. 3
figure 3

Effects of dexmedetomidine (DEX) infusion on the inotrope score of critically ill neonates and infants with heart disease

Full size image

CVP

The CVP decreased from the baseline value of 11 to the value of 9 in 2–6 h after initiation of DEX (p < 0.001), to 8 in 7–24 h (p < 0.001), and to 9 in last 6 h (p < 0.001) of DEX infusion. After termination of DEX infusion, CVP started increasing toward baseline values, reaching these values in 24 h after termination of DEX (p = 0.15) (Fig. 4). Comparison of CVP in the different subgroups did not show any substantial differences (Table 1).

Fig. 4
figure 4

Effects of dexmedetomidine (DEX) infusion on the central venous pressure of critically ill neonates and infants with heart disease

Full size image

NIRS

Overall, NIRS did not differ substantially after initiation of the DEX infusion. Subgroup analysis showed no difference in NIRS for neonates versus infants, for patients undergoing surgery in the RACHS 4–6 category versus the RACHS 1–3 category, and for patients undergoing heart transplantation versus no heart transplantation. However, as expected, the patients with a single-ventricle anatomy had a lower baseline NIRS (p = 0.006), which continued to be lower in this subgroup throughout the DEX infusion.

Effect on Sedation and Analgesia

The majority of the patients received continuous IV benzodiazepine and/or opioid infusions before and during their DEX infusion, with 42 patients receiving morphine infusion at an average dose of 2.8 ± 2.6 mg/kg/day before initiation of DEX. After initiation of DEX infusion, 39 patients continued to receive morphine infusion during the 24- to 48-h interval (at an average dose of 2.9 ± 2.8 mg/kg/day), and 34 patients continued to receive morphine infusion in the last 24 h of DEX (at an average dose of 2.4 ± 2.6 mg/kg/day).

Before initiation of DEX, 36 patients received midazolam infusion at an average dose of 3.5 ± 2.4 mg/kg/day. After initiation of DEX infusion, 32 patients continued to receive midazolam infusion in the 24- to 48-h interval (at an average dose of 4.1 ± 2.4 mg/kg/day), and 30 patients continued to receive morphine infusion in the last 24 h of DEX (at an average dose of 3.2 ± 3.0 mg/kg/day). Additional rescue boluses of sedation and analgesia were recorded during the DEX infusion, and then again after termination of the DEX therapy.

Before initiation of DEX, 45 patients received sedation rescue boluses at a median rate of 2 boluses (range, 0–3 boluses). After initiation of DEX, 43 patients needed sedation rescue boluses at a median rate of 1 bolus (range, 0–5 boluses) in the 24- to 48-h interval, and 39 patients received sedation rescue boluses at a median rate of 1 bolus (range, 0–3 boluses) in the last 24 h.

Before the initiation of DEX, 47 patients received analgesic rescue boluses at a median rate of 2 boluses (range, 1–3 boluses). After the initiation of DEX, 43 patients needed analgesic rescue boluses at a median rate of 2 boluses (range, 1–6 boluses) in the 24- to 48-h interval, and 38 patients received analgesic rescue boluses at a median rate of 1 bolus (range, 0–3 boluses) in the last 24 h.

Other Outcomes

No adverse respiratory effects were associated with DEX infusion. The respiratory rates during and after the termination of DEX infusion did not differ in our cohort. Of 43 intubated patients at the time DEX infusion was initiated, 22 were weaned off mechanical ventilation and extubated while receiving DEX infusion. The only patient receiving noninvasive ventilation also was transitioned to high-flow nasal cannula while receiving DEX infusion. No documented episodes of apnea occurred among the patients not intubated at the time DEX infusion was initiated or among the extubated patients receiving DEX infusion.

There was no difference in oxygen saturation, blood pH, PaO2, PaCO2, or serum bicarbonate values before and during DEX infusion. None of the patients had any rhythm abnormalities associated with DEX infusion. No significant abdominal signs or symptoms were noted. None of the patients had increased vomiting or a distended abdomen. No withdrawal symptoms associated with the termination of DEX were reported in the medical records. However, 37 patients (74%) in our cohort received a clonidine patch. No neurologic abnormalities were noted during or after the termination of DEX infusion.

Discussion

The use of DEX in CVICUs has been increasing across the country [2, 8, 9, 18]. We report the safe use of DEX infusion for critically ill neonates and infants with heart disease, including those with a single-ventricle anatomy, those undergoing complex surgeries, and those receiving an orthotopic heart transplant at doses of 0.1–1.5 mcg/kg/h for a median period of 78 h (range, 40–290 h). The major findings of this study were (1) a significant decrease in heart rate, MAP, inotrope score, and CVP among neonates and infants receiving DEX infusion; (2) a decrease in MAP associated with decrease in inotrope score during DEX infusion, suggesting an improved response to existing inotropes and improvement in vascular reactivity; and (3) a lack of substantial difference in major hemodynamic variables between neonates and infants, between single- and two-ventricle repair, between RACHS 4–6 and RACHS 1–3 categories for patients undergoing surgery, and between patients receiving a heart transplant and those receiving no heart transplant.

Chrysostomou et al. [8] were the first to describe the use of DEX as a sedative agent after cardiac surgery for children without significant adverse cardiovascular, respiratory, or gastrointestinal effects. The same group later demonstrated the use of this drug for infants after cardiac surgery [9].

The most frequently reported adverse effects associated with DEX infusion for children include dose-related hypotension and bradycardia [5, 12, 16, 23]. In the current study, DEX infusion was associated with similar findings of hypotension and bradycardia. Because the timing of DEX initiation was not uniform among all patients after surgery, it was difficult to differentiate whether the hemodynamic changes were caused by the potential presence of low cardiac output syndrome or by DEX.

In the current study, a decrease in MAP was not associated with escalation of inotropic support or the need for fluid boluses. In fact, it was associated with a decrease in inotropic support. It has been proposed that α2 agonists such as DEX aid in restoring vascular reactivity to exogenously administered catecholamines for patients with catecholamine-refractory shock [19]. Sympathetic inhibitors such as α2 agonists lower intrasynaptic catecholamine concentrations, leading to reduction in α1 desensitization. This eventually leads to gradual re-sensitization of α1 receptors and a better response to exogenously administered catecholamines [11, 19]. Moreover, this sympatholysis may decrease the myocardial oxygen requirement and myocardial oxygen consumption. These results should be interpreted with caution and investigated in future studies because our findings appear to be counterintuitive regarding sympatho-inhibition in the setting of catecholamine-refractory shock.

We did not observe any bradyarrythmias, clinical respiratory depressions, or significant changes in arterial blood gases or oxygen saturation. These results are consistent with the results seen with clonidine [13] and, in other studies, with DEX [4, 8, 17, 25]. Other investigators have demonstrated that DEX causes mild decreases in PaO2 levels, oxygen saturation, and mild hypercapnia. These changes, however, are often clinically insignificant [4, 9]. We found that DEX administration significantly reduced the concomitant daily dosing of opiates and benzodiazepines via infusion, which is consistent with other pediatric studies performed in CVICU settings [3, 6, 14].

In our study, 74% of the neonates and infants received clonidine after discontinuation of DEX infusion. It is possible that they were manifesting a withdrawal phenomenon with removal of DEX. The withdrawal phenomenon after sedation with DEX is an increasing concern in the literature, with some authors reporting symptoms such as increasing agitation, hypertension, tachycardia, emesis, dilated pupils, diarrhea, increased muscle tone, sneezing, and even seizures with abrupt discontinuation of prolonged DEX infusion (≥96 h) [10, 22, 24].

This study had several limitations. It was a single-center study, so the results may not be generalizable to all centers. The retrospective design of the study rendered it susceptible to design flaws and bias. We also lacked objective data regarding ventricular function and cardiac output during DEX infusion. The calculation for the inotrope score in our study included inodilators such as milrinone that may have falsely affected the score either during or after termination of DEX. Given the limited data, withdrawal symptoms or adverse side effects may have been missed. The decision to initiate DEX and the decision to wean opioid and benzodiazepine infusions after initiation of DEX were done at clinical assessment in the absence of any objective pain or comfort scores, which may have created a selection bias in our study.

Despite these limitations, our study provides a foundation for future prospective trials of DEX in multiple CVICUs. An additional study would be useful for a conclusion that the decrease in conventional sedation is significant. A prospective, observational study with a control group of patients receiving conventional sedation versus an experimental group receiving DEX and conventional sedation, using a comfort score, would provide an interesting opportunity to determine whether the decrease in conventional sedation has an additional value regarding the number of ventilation days and the ICU length of stay after cardiac surgery.

Conclusion

This is one of the largest studies on the use of DEX infusion in infants and children with critical heart disease. Based on our experience, we conclude that DEX infusion in neonates and infants with heart disease is safe from a hemodynamic standpoint and can reduce the concomitant dosing of opioid and benzodiazepine agents. Furthermore, DEX infusion may be useful in reducing vasopressor agent dosing for children with catecholamine-refractory cardiogenic shock. An additional study with a well-defined control group would be useful for concluding that the decrease in inotrope score and conventional sedation is significant.