Secondary Logo

Journal Logo

Pediatric Anesthesiology: Research Reports

Children with Infantile Neuronal Ceroid Lipofuscinosis Have an Increased Risk of Hypothermia and Bradycardia During Anesthesia

Miao, Ning MD*; Levin, Sondra W. MD†‡; Baker, Eva H. MD, PhD§; Caruso, Rafael C. MD; Zhang, Zhongjian MD, PhD; Gropman, Andrea MD¶#; Koziol, Deloris PhD**; Wesley, Robert PhD**; Mukherjee, Anil B. MD, PhD; Quezado, Zenaide M. N. MD*

Author Information
doi: 10.1213/ane.0b013e3181aa6e95
  • Free

Abstract

Neuronal ceroid lipofuscinoses (NCLs), collectively referred to as Batten disease, are a group of predominantly autosomal recessive neurodegenerative storage diseases characterized by lysosomal accumulation of autofluorescent material in neurons and other cell types.1–3 There are four major subtypes of NCLs: infantile neuronal ceroid lipofuscinoses (INCL, also known as CLN 1 or Haltia-Santavuori disease), late-infantile neuronal ceroid lipofuscinoses (LINCL, also known as CLN 2 or Jansky-Bielschowsky disease), juvenile neuronal ceroid lipofuscinoses (JNCL, also known as CLN 3 or Spielmeyer-Sjögren disease), and adult neuronal ceroid lipofuscinoses (also known as CLN 4 or Kufs disease).1–3 These NCL subtypes do differ on age of onset and composition of the storage material. The infantile subtype (INCL) is rare (1 in >100,000 births),4 is the most devastating disease, and has the earliest age of onset. INCL is caused by mutations within the gene CLN1 on chromosome 1p32, which encodes palmitoyl-protein thioesterase-1 (PPT1).5,6 PPT1 cleaves thioester linkages in S-acylated (palmitoylated) proteins facilitating the degradation of fatty acylated proteins by lysosomal hydrolases.5 Thus, deficiency of PPT1 leads to abnormal lysosomal accumulation of palmitoylated proteins, which, in turn, leads to INCL pathogenesis.7

Children with INCL appear normal at birth and seem to acquire developmental skills until the age of 6-11 mo. These children then begin to demonstrate hyperirritability, hypotonia, myoclonic jerks, and seizures.1–3 By the age of 2 yr, severe visual deterioration and slow or absence of pupillary reaction are noted, by 3 yr significant loss of cortical function ensues, and by 4 yr, children with INCL usually manifest an isoelectric electroencephalogram attesting to lack of brain function. These children then live in a vegetative state until they are 8-12 yr, at which time death occurs.1,8–10 INCL is uniformly fatal and, although many manifestations of the disease can be ameliorated with sedatives, centrally acting skeletal muscle relaxants (such as baclofen), and anticonvulsants, effective treatment is nonexistent.

Because of the profound neurologic abnormalities in children with INCL, the anesthetic management of these patients requires an understanding of the natural history of the disease. We report our experience with the anesthetic management of eight children with INCL and have found that hypothermia and bradycardia are frequent occurrences during anesthesia.

METHODS

Study Design and Patients

We conducted a case-control study of children with INCL and control children without the disease. Children in both case (INCL) and control groups were anesthetized for diagnostic procedures between 2001 and 2007 by the same anesthesiologists. Attempts were made to have controls comparable for age, duration of anesthetic, and types of procedures. Children in the INCL group were enrolled in an investigational therapeutic protocol approved by the IRB of the Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health since 2001. Children from the control group did not have INCL and were enrolled in unrelated investigational protocols approved by the IRBs of the National Institute of Child Health and Human Development and National Cancer Institute. Data analyzed included patient demographics, primary diagnoses, anesthetic technique, duration of anesthesia, and changes in temperature and heart rate. IRB review for this report was waived because it analyzed data previously collected for quality assurance, and data were devoid of identifiers.

Procedures

Table 1 outlines locations of procedures and transportation between locations during the anesthetics. All patients with INCL received general anesthesia for electroretinogram (ERG), brain magnetic resonance imaging (MRI), and spectroscopy. For the 2-h ERG, tropicamide 1% and phenylephrine 2.5% were applied to both eyes and a contact lens was placed on one eye as previously described.11 Children in the control group received general anesthesia for MRI, spectroscopy, auditory brain response and/or lumbar punctures, and dental examinations. The ambient temperature in the operating rooms and radiology suite was maintained at 22°C.

Table 1
Table 1:
Procedure Locations and Transportation Routes for Anesthetics in INCL and Control Patients Undergoing Diagnostic Procedures

Anesthetics

Anesthetics were administered by staff anesthesiologists and routine monitoring devices (electrocardiogram, noninvasive arterial blood pressure, pulse oximetry, and capnography) were used during procedures and transportation. Temperature was measured with a tympanic thermometer when patients were in the preoperative holding area (baseline temperature) and postanesthesia care unit (PACU) and with an esophageal probe when patients were in the operating room. For all patients, full body air blankets (Bair Hugger, Arizant Healthcare, Eden Prairie, MN) set at high temperature were used during portions of anesthetics conducted in the operating room. Whereas in the MRI scanner, temperature was not monitored, and patients were fully covered with blankets.

For this study, hypothermia was arbitrarily defined as temperatures below 35.6°C and bradycardia as heart rate below the 5th percentile for age. Therefore, heart rates below 85 bpm for patients aged 13-36 mo, below 75 for 37-72 mo, and below 74 for 73-108 mo were considered bradycardia.

Statistical Methods

Analyses were done with Statistical Analysis System Version 9.1 software (SAS Institute, Cary, NC) and StatXact Version 6 software (Cytel Software Corporation, Cambridge, MA). Because the number of anesthetics per patient differed significantly between the two comparison groups (patients with INCL versus controls), we formulated analyses to ensure valid comparisons. We first compared only the first anesthetic for each patient in INCL and control groups (Table 2, upper section). When analyzing data from the first anesthetic for each patient, interval data are presented as means and standard deviations, t tests were used to compare the two groups, and Fisher’s exact test were used to compare categorical variables.

Table 2
Table 2:
Comparison of Patients with Infantile Neuronal Lipofuscinosis (INCL) Versus Controls Undergoing Anesthesia for Diagnostic Procedures

We then compared all anesthetics for each patient in INCL and control groups. For analyses comparing the two groups with repeating data from all anesthetics for each patient (Table 2, bottom section), we used specialized methods appropriate for the type of outcome (dichotomous or interval): exact tests (using StatXact) for binomial clustered data for dichotomous outcomes (e.g., occurrence/nonoccurrence of hypothermia or bradycardia) or SASs Proc Mixed (for mixed models) for interval data. These methods are necessary because, when patients have repeating outcomes (dependent variables) and associated predictor variables (e.g., comparison group, preprocedure temperature, duration of anesthesia, time to first occurrence of lowest temperature, time in PACU), the repeating outcomes for a given group are no longer statistically independent. In these data, there is only one repeating factor: the possibly multiple anesthesia events a patient had. For dichotomous outcomes, these repeats per patient constitute the cluster (StatXact). In specifying analyses using SASs Proc Mixed, the predictor variables of interest were modeled as fixed effects; patient was a random effect (i.e., a “random intercept” model). The error terms in the mixed effects model were assumed to all be statistically independent, which induced a compound symmetric covariance matrix for the repeating outcomes on a single patient. Because of small sample sizes, we used the Kenward and Roger method for computing the degrees of freedom for tests of fixed effects. Lastly, the variances of the random effects were assumed to be equal in the two comparison groups. The primary outcome of interest was the change in temperature from preprocedure to the lower of the two temperatures recorded for each perianesthetic period. All “time” variables (duration of anesthesia, time to first occurrence of lowest temperature, time in PACU) were not repeated within an anesthetic event but a single measure for each anesthetic event. Results from Proc Mixed repeated measures are presented as the value of the estimator (“least square means” in SAS terminology) ± associated standard error of the estimator and P values. For all analyses, two-sided P < 0.05 was considered statistically significant. For two comparisons of data from all anesthetics, we did not use repeated measures or statistical comparisons. We determined the lowest temperature and slowest heart rate, for each patient from all anesthetic events, and then for each comparison group we obtained an arithmetic mean, standard deviation, and range of these lowest values. Because of the discrepancies between comparison groups in the number of anesthetics per patient, we simply report the descriptive statistics without an associated P value.

RESULTS

Patients

We identified eight children with INCL who cumulatively underwent 31 consecutive anesthetics between 2001 and 2007. Demographic data and clinical findings for INCL and control patients are listed in Tables 2 and 3. All patients with INCL had confirmed lethal mutations of the PPT1 gene. In patients with INCL, the experimental protocol called for ERG and brain MRI at baseline and at approximately every 6 mo after enrollment. With exception of Patient 8, who was admitted at the age of 10 mo and had no overt neurologic deficit, all other patients with INCL showed typical developmental delay and neurologic findings at time of enrollment.

Table 3
Table 3:
Demographics of Patients with Infantile Neuronal Ceroid Lipofuscinosis

The control group comprised 25 children (11 boys and 14 girls) with primary diagnoses of Niemann-Pick disease (n = 12), Smith-Lemli-Optiz syndrome (n = 5), Gaucher disease (n = 3), brain tumors (n = 2), neurofibromatosis Type 1 (n = 1), rheumatoid arthritis (n = 1), and chronic granulomatous disease (n = 1).

Anesthesia Management

Details of the 31 anesthetics for INCL and control groups are shown in Table 4. During ERGs, to avoid interference in the response to optical stimulation, no volatile anesthetic was used and anesthesia was maintained with nitrous oxide and propofol. In addition, in patients with INCL, muscle relaxants were used to avoid eye movement during ERGs as necessary. During other diagnostic procedures, anesthesia was maintained with a volatile anesthetic or propofol at the discretion of anesthesiologists. Considering all procedures, the mean duration of anesthesia was longer in patients with INCL (280 ± 10.3 min [mean ± se]), compared with controls (231 ± 8.2 min [mean ± se]), P = 0.002 (Table 2).

Table 4
Table 4:
Anesthetic Technique Used for Children with Infantile Neuronal Lipofuscinosis (INCL) and Controls Undergoing Imaging and Invasive Diagnostic Procedures

After all anesthetics, patients were routinely admitted to the PACU and subsequently transferred to patient care units, except for two patients with INCL who had hypothermia and after brief PACU stay (Patients 2 and 7) were transferred to the intensive care unit for overnight observation (see later).

Perianesthetic Events: Hypothermia and Bradycardia

To ensure valid comparisons between INCL and control groups, because some patients had repeated studies, we conducted two analyses: one comparing only the first anesthetic from each patient and another comparing all anesthetics for patients in the two groups (Table 2). Although patients with INCL were younger at the time of their first anesthetic (Table 2, P < 0.001), when all anesthetics were considered, there was no significant age difference comparing INCL versus controls (Table 2, P = 0.34). With regard to hypothermia, patients with INCL had a significantly more frequent occurrence of hypothermia than did controls: 18 vs 0, respectively (P < 0.001). In patients with INCL, in these 18 anesthetics when hypothermia ensued (Table 5), the baseline temperature was 36.3 ± 0.1°C (mean ± se), onset of hypothermia (temperature below 35.6°C) was observed relatively early in the course of anesthesia (by 59 ± 12.3 min, mean ± se), and the average temperature decrease was −1.6 ± 0.2°C (mean ± se). Furthermore, in these 18 anesthetics, hypothermia was observed during the first procedure, ERG in 12, and brain MRI in six patients. The presence of hypothermia was not associated with time spent in the PACU (P = 0.98). Figure 1 shows the times during anesthetics when the lowest temperature was recorded in controls and patients with INCL.

Table 5
Table 5:
Significant Events Observed During Anesthetics in Patients with Infantile Neuronal Lipofuscinosis
Figure 1.
Figure 1.:
Time of lowest temperature recorded during anesthesia for diagnostic procedures in infantile neuronal ceroid lipofuscinosis and control patients. Hypothermia was defined as temperature below 35.6°C (indicated by the dashed line).

Patients with INCL had a significantly more frequent occurrence of sinus bradycardia than did controls, 10 vs 1, respectively (Table 2, P < 0.001). As shown in Table 5, in the INCL group, five of eight patients (Patients 3, 4, 5, 6, and 7) displayed 10 episodes, whereas in the control group only 1 of 25 patients displayed 1 episode of sinus bradycardia. In the INCL group, in 8 of 10 episodes, bradycardia was associated with hypothermia, whereas in the remaining two episodes, bradycardia was observed in the absence of hypothermia. All episodes of sinus bradycardia were successfully treated with atropine or glycopyrrolate.

We evaluated possible associations of preanesthetic factors and duration of anesthesia with changes in temperature during anesthetics in INCL and control groups. Disease (INCL versus no INCL [controls]) was significantly associated with decreases in temperature (P < 0.001), whereas age (P = 0.12), baseline temperature (P = 0.09), duration of anesthesia (P = 0.84), and transportation route (P = 0.33) were not. When we evaluated factors associated with the lowest temperature during anesthetics, results were similar except that baseline temperature was significantly associated with lowest temperature observed during anesthetics (P = 0.002). In addition, considering the INCL group alone, duration of anesthesia (P = 0.30), age (P = 0.35), sex (P = 0.62), baseline temperature (P = 0.09), and transportation route (P = 0.33) were not associated with decreases in temperature during anesthesia.

Other Perianesthetic Occurrences

Other perianesthetic events included stridor in INCL Patient 2 who was intubated with an appropriate size endotracheal tube and yet had stridor after extubation with two different anesthetics. Over the 31 anesthetic events, there were no significant changes in arterial blood pressure (data not shown).

Patient 2, because of changes in brain MRI compared with that obtained 6 mo earlier, and Patient 7, because of episodes of transient apnea coupled with decreases in oxygen saturation, were admitted to the intensive care unit for overnight observation. In those two occasions, no deterioration of the baseline neurologic or respiratory examinations was noted and no interventions were necessary. Both patients were discharged without sequelae the following morning.

DISCUSSION

This case-control study showed that children with INCL have lower body temperature at baseline and that, during general anesthesia, they are at increased risk of hypothermia and bradycardia. These findings suggest that children with INCL might have impaired thermoregulation and abnormalities in the cardiac conduction system. Disturbances in temperature regulation are previously unreported phenotype of INCL and further illustrate the critical nature of this disease. Considering these potential thermoregulatory defects in patients with INCL and the known deleterious effects of anesthesia in thermoregulation,12 to avoid the complications associated with profound hypothermia, careful planning for anesthesia in this patient population is imperative.

Our data suggest that, in patients with INCL, hypothermia is a common occurrence during general anesthesia. In our series, seven of eight patients developed hypothermia (defined as temperature <35.6°C) despite the use of active rewarming techniques, careful measures to avoid heat loss, and in the absence of fluid shifts. We found four published reports describing the anesthetic courses of patients with NCLs. One report describes anesthesia for patients with LINCL,13 a second for two patients with JNCL,14 a third for an infant with NCL,15 and a fourth for one patient with adult NCL.16 To our knowledge, this is the first report to describe anesthetic considerations in infants and children with PPT1 mutation-confirmed INCL. Our findings are in concert with two other reports suggesting that mild hypothermia can occur in older patients with LINCL and JNCL.13,14 In our patients with INCL, we recorded temperatures lower than 34°C despite the use of active warming techniques during anesthesia. These episodes of hypothermia occurred in environments (operating rooms and MRI) with controlled ambient temperature (22°C). In some patients, hypothermia was associated with bradycardia and appeared to prolong patients’ stay in the PACU. However, PACU time was not statistically different when patients with and without hypothermia were compared. It is important to note that two patients with INCL who had hypothermia were transferred to an intensive care unit after a short stay in the PACU. Nevertheless, our findings together with those of others strongly suggest that, when anesthetizing a child with INCL, careful attention to core body temperature, as well as cardiac monitoring, and active prevention of heat loss are warranted.

Although general anesthesia alters temperature thresholds for thermoregulation12 and exposure to cold environments (such as MRI suites) can lead to decreases in body temperature, increases in core temperature and even hyperthermia have been reported in children undergoing MRIs.17–19 Our findings of hypothermia in patients with INCL, which in some was observed during MRI, are in contrast to those reports.17–19 Those studies showed that children (2-77 mo) sedated for MRI studies with various drugs had increases, rather than decreases, in core temperature.17,19 Furthermore, hyperthermia, rather than hypothermia, has been reported in children receiving volatile anesthetics during anesthesia for MRI.18 One could postulate that our anesthetic technique or its length could have contributed to hypothermia in patients with INCL. However, we believe these possibilities to be less likely given lack of hypothermia in control patients receiving similar anesthetic techniques during the same time period. Therefore, our findings suggest that children with INCL have increased risk of hypothermia and measures to avoid heat loss are warranted during anesthetics in these patients.

Why might patients with INCL have lower baseline temperature and when subjected to general anesthesia be susceptible to frequent episodes of hypothermia? One possibility is that children with INCL might have impairment of thermoregulation. In support of this hypothesis are studies of circadian temperature recordings in patients with NCL showing significant disturbances in body temperature rhythms. Researchers showed that patients with INCL have maximal temperature in the morning, whereas controls had higher temperatures in the afternoon.20 It is also noteworthy that the only patient with INCL in our series who did not develop hypothermia was the youngest patient (Patient 8), who had not yet developed neurologic deficits at the time of first anesthesia. This may suggest that impairment in thermoregulation follows the development of neurologic deficit in INCL. Although the mechanism for hypothermia is incompletely understood, our findings suggest that during anesthesia, patients with INCL are prone to develop significant hypothermia that can reach potentially harmful levels and warrant interventions. In addition, our findings of possible impairment of thermoregulation in patients with INCL, lend further support to a body of literature suggesting that some neurodegenerative diseases21 may involve and result from abnormalities in metabolic processes associated with energy production and thermoregulation.

Another common occurrence during anesthesia in our patients with INCL was bradycardia. Some of these episodes of bradycardia were associated with hypothermia but others occurred in the absence of hypothermia. Importantly, no episode of bradycardia was coupled with overt hemodynamic compromise and all episodes were successfully treated with anticholinergics. Although we observed no arrhythmia other than sinus bradycardia during anesthesia, others have reported the occurrence of bradycardia, sinus arrest, and severe supraventricular tachycardia during anesthesia in patients with JNCL.22 In that autopsy study, researchers showed that NCLs are associated with degeneration and infiltration of the cardiac conduction system with granular material, left ventricular hypertrophy, ventricular dilation, degenerative myocardial changes, interstitial fibrosis, and fatty replacement.22 Therefore, during the anesthetic of children with NCL, sinus bradycardia can occur with or without hypothermia, and there is the possibility of other arrhythmias given possible involvement of the cardiac conduction system known to occur in this group of neurodegenerative diseases. Therefore, consideration should be given to obtaining preoperative electrocardiograms in some patients with INCL to determine the presence of conduction abnormalities.

Although some may advocate avoiding the use of neuromuscular blocking drugs in patients with NCLs,14 we chose to use a nondepolarizing muscle relaxant in some of our anesthetics and found that its use was not associated with prolonged recovery or prolonged muscle relaxation. With regard to volatile anesthetics, we chose to use sevoflurane for induction of anesthesia in our patients and did not observe worsening of myoclonic activity or episodes of seizures. Although there are theoretical concerns about the use of sevoflurane in young children because of its epileptogenic potential,23,24 we found little evidence to indicate that its use is unsafe in patients with INCL.24

In summary, we examined temperature regulation in patients with INCL undergoing anesthesia and found that in these patients basal body temperature is lower than in controls, and hypothermia and sinus bradycardia are common occurrences during general anesthesia. Further, the degree of hypothermia can be significant and warrants the use of active warming techniques. In our institution, besides using warming techniques, we have taken measures to minimize anesthesia time (decreasing waiting time during anesthetics, rigorously coordinating the schedule of all operators involved in the multiple procedures) and measures to decrease heat loss by turning off air circulation in the bore of the magnet during MRIs.

Our findings suggest a previously unreported phenotype that include increased risk for hypothermia and bradycardia in INCL patients. Although the mechanisms of disturbances in temperature control are incompletely understood and were not explored in this study, this new phenotype has implications for future investigations of the pathophysiology and therapy of INCL.

REFERENCES

1. Goebel HH, Wisniewski KE. Current state of clinical and morphological features in human NCL. Brain Pathol 2004;14:61–9
2. Haltia M. The neuronal ceroid-lipofuscinoses: from past to present. Biochim Biophys Acta 2006;1762:850–6
3. Mole SE, Williams RE, Goebel HH. Correlations between genotype, ultrastructural morphology and clinical phenotype in the neuronal ceroid lipofuscinoses. Neurogenetics 2005;6:107–26
4. Santavuori P, Vanhanen SL, Sainio K, Nieminen M, Wallden T, Launes J, Raininko R. Infantile neuronal ceroid-lipofuscinosis (INCL): diagnostic criteria. J Inherit Metab Dis 1993;16:227–9
5. Camp LA, Verkruyse LA, Afendis SJ, Slaughter CA, Hofmann SL. Molecular cloning and expression of palmitoyl-protein thioesterase. J Biol Chem 1994;269:23212–9
6. Vesa J, Hellsten E, Verkruyse LA, Camp LA, Rapola J, Santavuori P, Hofmann SL, Peltonen L. Mutations in the palmitoyl protein thioesterase gene causing infantile neuronal ceroid lipofuscinosis. Nature 1995;376:584–7
7. Hofmann SL, Peltonen L. The neuronal ceroid lipofuscinosis. In: Scriver CR, Sly WS, Childs B, eds. The metabolic and molecular bases of inherited disease. New York: McGraw-Hill, 2001:3877–94
8. Santavuori P, Lauronen L, Kirveskari E, Aberg L, Sainio K, Autti T. Neuronal ceroid lipofuscinoses in childhood. Neurol Sci 2000;21:S35–41
9. Gardiner RM. Genetic analysis of Batten disease. J Inherit Metab Dis 1993;16:787–90
10. Haltia M. The neuronal ceroid-lipofuscinoses. J Neuropathol Exp Neurol 2003;62:1–13
11. Marmor MF, Zrenner E. Standard for clinical electroretinography (1999 update). International Society for Clinical Electrophysiology of Vision. Doc Ophthalmol 1998;97:143–56
12. Sessler DI. Temperature monitoring and perioperative thermoregulation. Anesthesiology 2008;109:318–38
13. Yamada Y, Doi K, Sakura S, Saito Y. Anesthetic management for a patient with Jansky-Bielschowsky disease. Can J Anaesth 2002;49:81–3
14. Pereira D, Pereira M, Caldas F. Anesthesia management in neuronal ceroid lipofuscinoses. Paediatr Anaesth 2006;16:356–8
15. Gopalakrishnan S, Sidduiqui S, Mayhew JF. Anesthesia in a child with Batten disease. Paediatr Anaesth 2004;14:890–1
16. Defalque RJ. Anesthesia for a patient with Kufs’ disease. Anesthesiology 1990;73:1041–2
17. Bryan YF, Templeton TW, Nick TG, Szafran M, Tung A. Brain magnetic resonance imaging increases core body temperature in sedated children. Anesth Analg 2006;102:1674–9
18. Kussman BD, Mulkern RV, Holzman RS. Iatrogenic hyperthermia during cardiac magnetic resonance imaging. Anesth Analg 2004;99:1053–5
19. Machata AM, Willschke H, Kabon B, Prayer D, Marhofer P. Effect of brain magnetic resonance imaging on body core temperature in sedated infants and children. Br J Anaesth 2009;102:385–9
20. Heikkila E, Hatonen TH, Telakivi T, Laakso ML, Heiskala H, Salmi T, Alila A, Santavuori P. Circadian rhythm studies in neuronal ceroid-lipofuscinosis (NCL). Am J Med Genet 1995;57:229–34
21. Weydt P, Pineda VV, Torrence AE, Libby RT, Satterfield TF, Lazarowski ER, Gilbert ML, Morton GJ, Bammler TK, Strand AD, Cui L, Beyer RP, Easley CN, Smith AC, Krainc D, Luquet S, Sweet IR, Schwartz MW, La Spada AR. Thermoregulatory and metabolic defects in Huntington’s disease transgenic mice implicate PGC-1alpha in Huntington’s disease neurodegeneration. Cell Metab 2006;4:349–62
22. Hofman IL, van der Wal AC, Dingemans KP, Becker AE. Cardiac pathology in neuronal ceroid lipofuscinoses—a clinicopathologic correlation in three patients. Eur J Paediatr Neurol 2001;5(suppl A):213–7
23. Akeson J, Didriksson I. Convulsions on anaesthetic induction with sevoflurane in young children. Acta Anaesthesiol Scand 2004;48:405–7
24. Constant I, Seeman R, Murat I. Sevoflurane and epileptiform EEG changes. Paediatr Anaesth 2005;15:266–74
© 2009 International Anesthesia Research Society