Litman, Ronald S. DO, FAAP*; Soin, Komal BA†; Salam, Abdul MSC‡
Section Editor(s): Davis, Peter J.
One of the most serious side effects of the use of general anesthetics in term and preterm infants is postoperative apnea. This complication is thought to occur from the relative immaturity of the central respiratory centers combined with the effect of residual anesthetics on ventilatory drive. Previous research has revealed that preterm infants less than approximately 48 to 60 weeks' postconceptional age (PCA) who receive general anesthesia have an increased risk of postoperative apnea1–3 and therefore require extended postoperative monitoring. The risk may be similar for term infants younger than about 1 month, although this is based more on case reports.4–7 Similar studies have not been performed in infants who received sedatives for medical procedures; thus, the risk of postsedation apnea and oxygen desaturation relative to chronological and gestational age is unknown.
At our institution, chloral hydrate (CH) sedation is routinely used for infants undergoing magnetic resonance imaging (MRI) examination. Many of these infants were born prematurely and <60 weeks' PCA at the time of the procedure. There are no data to guide postsedation management with regard to length of monitoring and observation after CH administration. Although many different sedatives are used for medical procedures, CH is an important sedative to analyze because its risk is higher in infants than older children8,9; in preterm infants, its active metabolite, trichloroethanol, has a relatively long half-life of up to 40 hours10 and may have an extended duration of action in former preterm infants.11 There have been a number of published studies on the safety of CH for MRI in infants,12,13 but none has specifically focused on the effects of chronological age, gestational age, and PCA.
We performed this retrospective cohort study to examine the clinical course of infants younger than 1 year who underwent MRI with CH, with or without additional sedatives. We aimed to describe the efficacy of CH sedation, the need for additional sedative drugs, and the incidence of oxyhemoglobin desaturation or need for oxygen supplementation. We also sought to determine the relationship between these factors to chronological age in term infants and gestational age and PCA in preterm infants (<37 weeks' gestation). We hypothesized that in term infants, oxyhemoglobin desaturation or need for oxygen supplementation would be directly correlated with younger chronological age, and that these complications in preterm infants would be directly correlated with chronological age, as well as gestational age and PCA.
This study was approved by the Department of Radiology Research Board and the IRB of the Stokes Research Institute of The Children's Hospital of Philadelphia. Using a Department of Radiology clinical database, we identified all infants younger than 1 year who underwent MRI examination with CH sedation (with or without additional sedatives) between July 1, 2001 and September 1, 2002. In all cases, the administration of CH to infants for MRI adhered to a specific protocol: an initial dose of 50 mg/kg for infants younger than 3 months or 75 mg/kg for infants 3 months or older (with a maximal single and total dose of 1 g), and if still awake after 20 minutes, the practitioner had the option of administering an additional CH dose (half the initial dose), intranasal midazolam 0.15 mg/kg, or IV midazolam 0.1 mg/kg. Infants were monitored during the MRI scan by a registered nurse who was supervised by an attending radiologist in the immediate vicinity. Nurses provided oxygen supplementation via nasal cannula at their discretion with the onset of oxyhemoglobin desaturation that did not immediately abate or recurred during the scan.
We excluded infants with congenital heart disease and those receiving mechanical ventilation via endotracheal tube or with presence of a tracheostomy at the time of the MRI. Using names and medical record numbers from the radiology database, each infant's complete medical chart was examined using paper (for outpatients) and electronic (inpatients) records, and, using the pertinent sections of the medical record, a case report form was generated for dependent and independent variables of interest (Table 1) during the preprocedure period, intraprocedure period, and postprocedure period up to 24 hours after the MRI scan. Documentation for these procedural periods is routinely recorded on a time-based sheet by the nurse responsible for continuously monitoring the patient. Heart rate, respiratory rate, oxygen saturation as measured by pulse oximetry (Spo2) value, arterial blood pressure, fractional inspired concentration of oxygen, sedation level, and level of airway patency are recorded every 5 minutes. There is also a field for subjectively recording any other clinical events, such as patient movement and coughing. Temperature was not recorded in the MRI facility. Postprocedural data and events were recorded from nurses' flow sheets, which included all outcome variables as well as temperature on an at least half-hourly basis, and more frequently when the patient was being cared for in an intensive care unit. Separate sections of the MRI sedation record allowed us to determine the length of time of the scan and time to discharge after the scan. Discharge criteria are dictated by hospital policy and include attainment of baseline wakefulness, verbalization, and cardiorespiratory and neurologic status, without evidence of vomiting, pain, or Spo2 <92% while asleep on room air or <95% when awake or at baseline. There was no minimal length of phase 1 recovery stay. Some infants were preselected for overnight admission based on PCA criteria. These infants were included in the analysis as inpatients. All data from case report forms were recorded in a password-protected database (Microsoft Access) for eventual analysis.
For reporting purposes, we defined a term infant as born at or after 37 weeks' gestation, and a preterm infant was defined as born <37 weeks' gestation. Chronological age is defined as the number of weeks after birth, and PCA is defined only in the preterm group of infants as the sum of the gestational age and the chronological age, in weeks.
Statistical analysis was performed on a deidentified data set after all patients were entered into the database. Descriptive statistics are presented as mean ± sd, median with range for quantitative variables, and number (percentages) for qualitative variables. Univariate analysis was performed by using independent sample t test, Mann-Whitney U test, Pearson χ2 test, and Fisher exact test whenever appropriate to associate demographic and clinical factors with infants born preterm versus full term and postprocedure oxyhemoglobin desaturation or need for supplemental oxygen (yes versus no). McNemar test was used to compare the marginal distribution among preprocedure, intraprocedure, and postprocedure oxyhemoglobin desaturation or need for supplemental oxygen. Multiple logistic regression analysis was used to identify independent predictors for postprocedure oxyhemoglobin desaturation or need for supplemental oxygen. A P value of <0.05 was considered statistically significant, and all P values reported are 2-sided. All analyses were performed using SPSS 15.0 for Windows (SPSS, Chicago, IL).
From an initial database consisting of 2031 infants who underwent MRI with CH, 1373 met the inclusion criteria and were eligible for analysis (Table 2). Ineligible subjects consisted of infants whose lungs were mechanically ventilated at the time of the MRI, those who did not receive CH, and those for whom the patient chart could not be located. The frequency distribution of preterm (by PCA) and term (by chronological age) infants is shown in Figure 1.
Preprocedural, intraprocedural, and postprocedural complications for all infants are described in Table 3. Oxyhemoglobin desaturation or need for supplemental oxygen was the most common complication. It was more prevalent during the procedure than in the postprocedure phase (19.6% vs 2.0%) because of the higher frequency of administration of supplemental oxygen by nursing personnel in response to gradually decreasing oxyhemoglobin saturations during the MRI scan. There was a sedation failure rate of 4.8%, whereby the MRI scan was unable to be successfully completed.
A comparison of dependent and independent variables between term and preterm infants was assessed by univariate analysis (Table 4). Preterm infants were more likely to be inpatients, have a lower weight, have a history of apnea, and have lower hemoglobin levels. The incidence of oxyhemoglobin desaturation <90% or the need for supplemental oxygen was approximately 20% in both term and preterm infants. Preterm infants were not more likely to develop postprocedure oxyhemoglobin desaturation than term infants (P = 0.2); however, because there were only 8 preterm infants who developed this complication, a type 2 error may have occurred because of the lack of sufficient statistical power. The clinical characteristics of these infants are detailed in Table 5.
There were 10 instances of postprocedural bradycardia in 8 infants (2 infants had a repeat MRI with CH sedation). Six (2.3%) involved preterm-born infants and there were 4 (0.4%) instances in term infants (P = 0.005). The PCA of the preterm infants ranged from 36 to 40 weeks, and the chronological age of the term infants ranged from 3 to 7 days. Three of the 10 episodes were associated with oxyhemoglobin desaturation. The nadir heart rates in these cases were 45 bpm, 78 bpm, and one unspecified. In the remaining 7 instances of bradycardia that were not associated with oxyhemoglobin desaturation, nadir heart rates ranged from 66 to 99 bpm. All resolved with physical stimulation.
We examined the influence of independent variables based on the occurrence of postprocedure oxyhemoglobin desaturation. Univariate analysis revealed that postprocedure oxyhemoglobin desaturation or need for supplemental oxygen was more likely in inpatients (P < 0.001) and was associated with a lower body weight (3.9 ± 2.1 kg vs 6.6 ± 3.0 kg; P < 0.001), history of apnea (33.3% vs 9.9%; P = 0.001), higher ASA physical status (P = 0.002), and younger chronological age (58.7 ± 82.8 days vs 152 ± 105.9 days; P < 0.0001; Fig. 2). Eighty-six infants (6.4%) in our cohort were classified as ASA physical status I (healthy without systemic disease), and none of these patients demonstrated postprocedure oxyhemoglobin desaturation or need for supplemental oxygen. Infants with postprocedural oxyhemoglobin desaturation tended to have higher hemoglobin levels (14.7 ± 3.8 g/dL vs 12.4 ± 3.0 g/dL; P = 0.035). In term and preterm infants, the presence of preprocedure or intraprocedure oxyhemoglobin desaturation was not associated with postprocedure oxyhemoglobin desaturation or need for supplemental oxygen. When the preterm group was analyzed separately, the risk of postprocedure oxyhemoglobin desaturation or need for supplemental oxygen was directly correlated with younger chronological age (56.0 ± 41.5 days vs 150.6 ± 107.1 days; P = 0.012) and younger PCA (39.5 ± 4.1 days vs 54.4 ± 15.2 days; P = 0.005), but not gestational age. Additional doses of CH or supplementation with midazolam also did not increase the incidence of postprocedural oxyhemoglobin desaturation or need for supplemental oxygen.
Chloral hydrate is a relatively mild sedative that, when administered orally in doses of 50 to 75 mg/kg, induces sleep without untoward respiratory or hemodynamic complications in most infants.14,15 However, in certain patients, CH has been associated with prolonged recovery and oxygen requirement.16 This is likely attributable to the relatively long half-life of its active metabolite, trichloroethanol, which is further delayed in preterm infants.10
Term infants younger than 1 month and preterm infants younger than approximately 48 to 60 weeks' PCA are at increased risk of developing central apnea after general anesthesia.1,3 It has become standard policy in most children's hospitals that these infants are admitted and monitored for at least 12 hours after administration of general anesthesia. Similar studies have not been performed to define the risk of apnea for at-risk infants who receive sedative drugs for medical procedures.
The most important findings in our analysis were that a relatively high proportion (approximately 20%) of both term and preterm infants required oxygen supplementation during the MRI scan, and the risk of postprocedure oxyhemoglobin desaturation or need for oxygen supplementation was directly correlated with younger chronological age in term infants, and directly correlated with lower weight, lower PCA, and younger chronological age in preterm infants. When compared with the term infant group, preterm infants did not have an increased incidence of postprocedure oxyhemoglobin desaturation or need for supplemental oxygen (P = 0.2). However, because this was a retrospective study, no a priori power estimates were performed. A post hoc power estimate revealed that with the current observation of postprocedure oxyhemoglobin desaturation or need for supplemental oxygen between preterm and full-term infants, we would need 5859 cases (1172 preterm and 4687 full term) to achieve 80% power to detect an absolute difference of 1.4% with a 5% level of significance using χ2 test. Thus, we cannot firmly accept the null hypothesis of similar rates of postprocedure oxyhemoglobin desaturation or need for supplemental oxygen between term and preterm groups. Bradycardia, however, was more common in preterm infants after the MRI scan. Allegaert et al.11 reported an increased incidence of bradycardia in former preterm infants who received CH. In their group, the development of bradycardia was associated with a younger gestational age but not a younger PCA.
When the group of infants who developed postsedation oxyhemoglobin desaturation or need for supplemental oxygen was examined more closely, several patterns are evident (Table 5). None of the preterm infants with postsedation oxyhemoglobin desaturation or need for supplemental oxygen was older than 48 weeks' PCA. All preterm infants with postprocedure oxyhemoglobin desaturation or need for supplemental oxygen were inpatients. In our group of term infants, there were 4 outpatients with relatively minor decreases in oxyhemoglobin saturation, and all were discharged home within several hours after the MRI scan completion. Six term infants required prolonged oxygen supplementation. None was older than 3 weeks, and all had significant comorbidities.
Previous studies that defined the risk of apnea after administration of general anesthesia were performed prospectively. Apnea was detected by using pneumography and nasal flow sensors.2 This prospective study design, although scientifically rigorous, is applicable for studying relatively small numbers of patients. A combined analysis was necessary to pool such studies for more meaningful results.3 By using the retrospective methodology in this study, we were not able to detect apnea, per se, but relied on the presence of oxyhemoglobin desaturation or requirement for additional oxygen administration as surrogate outcomes. The addition of supplemental oxygen did not have definite and consistent criteria throughout different locations in our institution, but, in exchange for using these surrogate outcomes, we were able to study a relatively large number of patients to determine relative risk of low chronological age, preterm birth, and other associated factors such as gestational age and PCA.
Our population of infants did not include those with congenital heart disease. At our institution, these patients are cared for and sedated by a separate service with separate sedation protocols.17 In addition, we thought that exclusion of this group was necessary to reduce bias from a population of infants at higher baseline risk for hypoxemia.18
There are additional limitations inherent in the use of a database populated by data obtained from electronic and paper-based patient records. The most important is the absence of documentation of important complications such as oxyhemoglobin desaturation. Thus, our incidence may be an underestimate of the true incidence, which could only be determined by prospective continuous observation. However, we expect that there is no bias toward underreporting based on the chronological age, gestational age, or PCA of our patients. Subjective complications such as upper airway obstruction, bronchospasm, laryngospasm, or flaccidity are also likely to be underestimated or documented to a lesser degree than would an objective number (i.e., Spo2) obtained from a patient monitor. Similarly, because discharge criteria were based on attainment of baseline clinical measures and not time per se, it is possible that some episodes of postdischarge oxyhemoglobin desaturation occurred. These limitations underscore the importance of carefully designed prospective studies to definitively investigate the effects of a variety of types of sedatives on postprocedural complications. Another possible limitation that is particular to this study is the lack of data after discharge from the MRI facility. Vital signs and oxygen requirements are recorded on the nurses' flow sheets up until the time of discharge, but there was no mechanism to identify subsequent sedation-related problems after discharge, such as subclinical oxyhemoglobin desaturation that self-resolved or did not result in an obvious change in the infant's health.
Finally, when considering important complications that guide clinical care, the interpretation of zero numerators must be done with extreme caution.19 For example, if one uses the case of the 86 ASA physical status I patients who did not develop postprocedure oxygen desaturation, the Hanley-Lippman-Hand rule of 3s predicts that we can be 95% confident that the rate of zero complications in this group is not >3 of 86, or approximately 3.5%. It is conceivable that healthy term infants may be safely discharged home without extended monitoring, but because CH causes delayed sedation after discharge home,16 communication with caretakers until the following day is essential. The inpatients in our study group had the advantage of documented vital signs over the postprocedure 24-hour period, but these data were impossible to know for outpatients.
In conclusion, the occurrence of postprocedural oxyhemoglobin desaturation that we were able to detect or need for supplemental oxygen was directly correlated with younger chronological age in term infants and younger PCA in preterm infants. Term infants who developed extended postprocedural hypoxemia were inpatients and had significant comorbidities. Preterm infants had a higher incidence of postprocedure bradycardia.
The authors are grateful to Mary Beth Bartko, MSN, Nurse Coordinator in the Department of Radiology, The Children's Hospital of Philadelphia, for providing the original database. They also acknowledge the invaluable assistance of research assistants in the Department of Anesthesiology and Critical Care at The Children's Hospital of Philadelphia: Sina Shah-Hosseini, MSE, Jared Mendelsohn, BS, Sabaa Dam, BS, and Reshma Pachikara, BS.
1. Kurth CD, Spitzer AR, Broennle AM, Downes JJ. Postoperative apnea in preterm infants. Anesthesiology 1987;66:483–8
2. Kurth CD, Lebard SE. Association of postoperative apnea, airway obstruction, and hypoxemia in former premature infants. Anesthesiology 1991;75:22–6
3. Cote CJ, Zaslavsky A, Downes JJ, Kurth CD, Welborn LG, Warner LO, Malviya SV. Postoperative apnea in former preterm infants after inguinal herniorrhaphy. A combined analysis. Anesthesiology 1995;82:809–22
4. Tetzlaff JE, Annand DW, Pudimat MA, Nicodemus HF. Postoperative apnea in a full-term infant. Anesthesiology 1988;69:426–8
5. Cote CJ, Kelly DH. Postoperative apnea in a full-term infant with a demonstrable respiratory pattern abnormality. Anesthesiology 1990;72:559–61
6. Karayan J, Lacoste L, Fusciardi J. Postoperative apnea in a full-term infant. Anesthesiology 1991;75:375
7. Andropoulos DB, Heard MB, Johnson KL, Clarke JT, Rowe RW. Postanesthetic apnea in full-term infants after pyloromyotomy. Anesthesiology 1994;80:216–9
8. Dalal PG, Murray D, Cox T, McAllister J, Snider R. Sedation and anesthesia protocols used for magnetic resonance imaging studies in infants: provider and pharmacologic considerations. Anesth Analg 2006;103:863–8
9. Malviya S, Voepel-Lewis T, Tait AR. Adverse events and risk factors associated with the sedation of children by nonanesthesiologists. Anesth Analg 1997;85:1207–13
10. Mayers DJ, Hindmarsh KW, Gorecki DKJ, Sankaran K. Sedative/hypnotic effects of chloral hydrate in the neonate: trichloroethanol or parent drug? Dev Pharmacol Ther 1992;19:141–6
11. Allegaert K, Daniels H, Naulaers G, Tibboel D, Devlieger H. Pharmacodynamics of chloral hydrate in former preterm infants. Eur J Pediatr 2005;164:403–7
12. Greenberg SB, Faerber EN, Aspinall CL, Adams RC. High-dose chloral hydrate sedation for children undergoing MR imaging: safety and efficacy in relation to age. AJR Am J Roentgenol 1993;161:639–41
13. Mason KP, Sanborn P, Zurakowski D, Karian VE, Connor L, Fontaine PJ, Burrows PE. Superiority of pentobarbital versus chloral hydrate for sedation in infants during imaging. Radiology 2004;230:537–42
14. Boswinkel JP, Litman RS. Sedating patients for radiologic studies. Pediatr Ann 2005;34:650–4, 656
15. Boswinkel JP, Litman RS. The pharmacology of sedation. Pediatr Ann 2005;34:607–13
16. Malviya S, Voepel-Lewis T, Prochaska G, Tait AR. Prolonged recovery and delayed side effects of sedation for diagnostic imaging studies in children. Pediatrics 2000;105:E42
17. Fogel MA, Weinberg PM, Parave E, Harris C, Montenegro L, Harris MA, Concepcion M. Deep sedation for cardiac magnetic resonance imaging: a comparison with cardiac anesthesia. J Pediatr 2008;152:534–9, 539.e1
18. Heistein LC, Ramaciotti C, Scott WA, Coursey M, Sheeran PW, Lemler MS. Chloral hydrate sedation for pediatric echocardiography: physiologic responses, adverse events, and risk factors. Pediatrics 2006;117:e434–41
19. Hanley JA, Lippman-Hand A. If nothing goes wrong, is everything all right? Interpreting zero numerators. JAMA 1983; 249:1743–5