Prolonged Anesthetic Exposure in Children and Factors Associated with Exposure Duration

Abstract

Introduction:

Anesthetic exposure duration has come under scrutiny due to a Food and Drug Administration (FDA) warning against prolonged use of anesthesia in children, defined as exposures longer than 3 hours.

Methods:

Data for 2,613,344 pediatric anesthetic records from the American Society of Anesthesiologist’s National Anesthesia Clinical Outcomes Registry (NACOR) from 2010 to 2015 were analyzed to describe anesthetic duration and the prevalence of prolonged exposures in children. Common pediatric inpatient procedures were independently assessed to determine factors associated with exposure duration.

Results:

The overall mean pediatric anesthetic duration was 83.3 ± 107.4 minutes (median 57 minutes). Prolonged exposures comprised 7% of pediatric anesthetic records, and 15% of anesthetic records in infants. After restricting to common pediatric inpatient procedures (n=96,603) and adjusting for procedure type using a Poisson model, compared to children ≥10 years old, exposures in infants were 20.5% longer (anesthetic duration ratio [ADR], 1.205; [95% confidence interval (CI), 1.202–1.208]). Longer exposures were also seen with a variety of other patient and hospital factors including ASA 4 patients vs. ASA 1 patients, ([ADR], 1.381; [95% CI, 1.376–1.386]), and university hospitals vs. surgery centers ([ADR], 1.241; [95% CI, 1.236–1.246])

Conclusions:

Most pediatric anesthetic exposures last less than one hour with a small percentage lasting over 3 hours. Anesthetic duration for inpatient pediatric procedures however is associated with specific patient and hospital characteristics. These results may help identify children potentially at risk for prolonged anesthetic exposure, and inform procedure time prediction and operating room scheduling.

Introduction:

The evaluation of procedure duration has a variety of applications including the improvement of operating room efficiency and resource allocation.¹ Procedure duration and the associated duration of anesthetic exposure has also gained additional relevance in children due to a recent recommendation by the Food and Drug Administration (FDA) against prolonged anesthetic exposures (defined as exposures longer than 3 hours) in young children given the possible risks to neurodevelopment.² While there is some controversy regarding this recommendation, and the long-term effects of anesthesia on neurodevelopment are still unclear, associations between prolonged or multiple anesthetic exposures and an increased risk of neurodevelopmental deficit have been described.^3-9 However, little is known about the prevalence of prolonged anesthetic exposures in children, or factors that are associated with exposure duration.

Procedure time is commonly reported in individual studies of surgery and anesthesia in children, but these studies are typically performed at single centers with a limited number of patients, and often evaluate just one specific procedure.^10-12 Procedure duration information typically requires abstraction of data from medical records and is therefore difficult to perform on a large-scale multicenter basis. Anesthetic duration for surgical procedures can also vary widely, particularly in children as there is a wide range of patient size, weight, and pathology.¹³ Predictors of duration have been explored in adults, with a multi-center study finding increased anesthesia time to be associated with race, coagulation disorders, and paraplegia.¹⁴ However to our knowledge, only one large scale single center study has evaluated procedure time in children, but was unable to identify any factors associated with anesthesia or procedure duration.¹³

In order to address these gaps in knowledge, the goals of the present study are to use a large multicenter database to: 1) evaluate the incidence of prolonged anesthetic duration as defined by the FDA in children of different ages, and 2) identify patient and hospital factors associated with anesthetic duration.

Methods:

Data Source:

The data were obtained from the American Society of Anesthesiologists (ASA) Anesthesia Quality Institute (AQI) National Anesthesia Clinical Outcomes Registry (NACOR). The NACOR is a voluntary submission registry created with the goal of sharing anesthesia-related data and outcomes. The purpose of the database is to generate a representative view of anesthetic practice in the United States, and to describe patient demographics and quality of anesthetic care on a local and national level.¹⁵ The NACOR data is directly imported from participating institutions through their billing, anesthesia information management, quality assurance, and hospital electronic record systems.¹⁶

In the present study, data from the NACOR Participant User Files from 2010 to 2015 were used. The primary variable of interest was case duration, defined as the reported billing interval between anesthesia start and stop times, and measured in minutes. This duration describes a period of time that includes all time for anesthesia care including patient transfer to a post-operative recovery location. All records with missing values for patient age or case duration were excluded from analysis. Analysis was also restricted to records for patients aged <18 years old.

To evaluate the incidence of prolonged anesthetic duration, patient characteristics and durations for all cases were evaluated. Records were then evaluated after stratifying by patient age in the following categories: age <1 year, ≥1 to <3 years, ≥3 to <5 years, ≥5 to <10 years, and ≥10 to <18 years old. According to the FDA recommendation, the duration of time specified as a prolonged exposure to be avoided in children is 3 hours, and the children potentially at risk were identified as those below 3 years of age.² Therefore, the percentage of cases with durations greater than 3 hours was evaluated in children of each age category. As anesthesia time may also include time for patient transfer after the completion of the intraoperative anesthetic, to calculate a more conservative estimate of anesthetic exposures of greater than 3 hours, the percentage of cases with durations greater than 4 hours was also evaluated.

To assess patient and hospital level factors associated with anesthetic duration while adjusting for the type of procedure, the NACOR was queried for records of children receiving anesthesia for the ten most common inpatient general surgical procedures performed in the United States in 2009.¹⁷ These procedures were identified using Current Procedural Terminology 4^th edition (CPT-4) codes listed in the NACOR records. In order to remove outlier duration records for this analysis, patients in the 0.5 percentile of the shortest and longest durations for each procedure were excluded, leaving 99% of the records for each procedure available for analysis. The variables evaluated for their association with duration of anesthesia were patient sex, age, and ASA Status, as well as the type and US geographical region of the medical facility. The US geographical regions were facilities in the Northeast, Midwest, South, and West regions. The facility types were university hospitals, large community hospitals (consisting of > 500 beds), medium-sized community hospitals (consisting of 100 – 500 beds), small community hospitals (less than 100 beds), specialty hospitals, and surgery centers.

Statistical Analysis:

To evaluate prolonged duration in pediatric procedures, means and standard deviations, and median values of procedure duration were calculated for all available cases. The incidence of prolonged duration in children (>3 hours or >4 hours) of different age groups was then calculated. To assess predictors of duration in the ten most common pediatric general surgical procedures, means and standard deviations, median values, quartiles, and highest and lowest deciles were first calculated. Given that procedure duration is not normally distributed, Poisson regression was used to evaluate the association between the duration of anesthesia and all patient and hospital level covariates. Surgical procedure type was also included as a multilevel categorical variable. Records with missing data in any of the covariates or records with “Other” listed as the facility type were excluded from regression analysis. All statistical analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC).

Results:

The study protocol was approved by the Institutional Review Board at Columbia University (New York, New York) as exempt from written or informed consent. A total of over 30 million records were identified in the NACOR participant user files from 2010 to 2015 with 3,341,655 records in children with age less than 18 years old. After also excluding all records with missing duration of anesthesia, 2,613,344 records remained, with a total of 332 anesthesia practices submitting data.

Prolonged Anesthetic Duration in Children

Of these records, 5.8% (n=152,567) were from children <1 year, 18.6% (n=486,550) from children ≥1 to <3 years, 15.5% (n=406,179) from children ≥3 to <5 years, 26.2% (n=685,294) from children ≥5 to <10 years, and 33.8% (n=882,754) from children ≥10 to <18 years old. The majority of children were ASA 1 or 2. (Table 1) Compared to other age categories, children receiving anesthesia at <1 year of age had higher ASA statuses than children in other age categories. There were similar distributions of facility types based on age, with the exception of more children <1 year of age receiving anesthesia at university and large community hospitals.

Table 1:

Patient Characteristics and Duration of Anesthesia by Age Category

	All ages (n=2,613,344) %	<1 yo (n=152,567) %	≥1 to <3 yo (n=486,550) %	≥3 to <5 yo (n=406,179) %	≥5 to <10 yo (n=685,294) %	≥10 to <18 yo (n=882,754) %
Sex
Male	53.6	59.4	57.8	55.2	53.8	49.5
Female	41.9	33.8	37.4	40.4	41.8	46.6
Missing	4.5	6.8	4.8	4.4	4.4	3.9
ASA Status
1	47	33.7	48.6	49.7	47.4	46.9
2	25.1	20.2	23	23.8	26.1	26.8
3	9.2	19.5	8.9	8	8.3	8.7
≥4	1.6	11	1.6	0.9	0.8	1
Missing	17.1	15.6	17.9	17.6	17.3	16.5
US Geographical Region
Northeast	11.2	10.5	9.2	11	11	12.7
Midwest	27.7	23.3	29.6	26.8	27.8	27.7
South	44.2	49.9	46.4	46.2	44.4	40.9
West	15.4	15.3	13.2	14.4	15.4	17
Missing	1.5	1.1	1.6	1.6	1.4	1.7
Facility Type
University	11.4	15.8	12.1	10.7	10.6	11.3
Large Community	13.9	18.1	13	13	13.5	14.3
Medium Community	23	19.6	21.3	22.1	22.9	25
Small Community	2	1.2	1.8	2.3	2.1	2.1
Specialty	7.2	8.6	7.5	7.3	7.3	6.8
Surgery Center	19.3	8.9	21.8	22	21.4	16.8
Other	0.4	1.1	1.6	1.5	1.4	1.7
Missing	22.7	26.6	20.9	21	20.7	22.1
Duration (minutes)
Mean ± SD	83.3 ± 107.4	118.7 ± 141.3	72.6 ± 123.2	70.9 ± 85.4	71.4 ± 82.6	98.1 ± 114.1
Median (IQR)	57 (59)	81 (87)	44 (57)	53 (51)	53 (49)	67 (67)

The mean duration for all anesthetic records was 83.3 ± 107.4 minutes with a median duration of 57 minutes. The mean durations were similar between age categories with children between 1 and 10 years old, but longer mean durations were seen in children at ages <1 year old and ≥10 years old.

The majority of anesthetic cases were less than 3 hours in duration, although longer procedures were more common in younger children. The percentage of anesthetic cases that lasted for longer than 3 hours for all children was 7.2%, with 4.2% lasting longer than 4 hours. In children <1 year old however, 15.6% of cases lasted longer than 3 hours with 9.4% lasting longer than 4 hours. (Figure 1)

Figure 1: Percentage of Anesthesia Cases with Prolonged Duration.

The percentages of anesthesia cases for the 2,613,344 pediatric anesthetic records reported to the NACOR from 2010 to 2015 with durations >3 hours and >4 hours. Data are presented for all ages and also stratified by patient age.

Factors Associated with Anesthetic Duration

Using the NACOR database, a total of 96,603 records from children receiving one of the ten most common inpatient general surgical procedures were identified. These records originated from 187 anesthesia practices. Compared to the records of children receiving anesthesia for all procedures, these children tended to be younger in age, with a higher percentage of children in the <1 year old category and a lower percentage of children in the ≥10 year old category. (Table 2) These children also had overall higher ASA statuses and were less likely to be performed at surgery centers. The duration of these cases ranged from 22 to 712 minutes with a right skewed distribution. (Figure 2) The overall median duration of these most common pediatric procedures was 75 minutes but ranged from 51 to 180 minutes based on procedure type. (Table 3) The longest duration was found with anti-reflux procedures and bladder ureter reconstructions while gastrostomy tube placements had the shortest median durations. The durations of anesthetic care within the same procedure type also varied, with the highest duration decile ranging from 1.9 to 4.6 times longer than the lowest decile.

Table 2:

Characteristics of Patients Receiving the Ten Most Common Inpatient Procedures in Children

	All Procedures (n=96,603) %
Age
<1 years old	13.5
≥1 to <3 years old	12.1
≥3 to <5 years old	7.4
≥5 to <10 years old	21.4
≥10 to <18 years old	45.5
Sex
Male	60.3
Female	39.7
ASA Status
1	49.7
2	30.4
3	16.1
≥4	3.9
US Geographical Region
Northeast	13.6
Midwest	23.1
South	41.2
West	22.2
Facility Type
University	11.7
Large Community	28.6
Medium Community	39.6
Small Community	3.4
Specialty	11.9
Surgery Center	4.9

Figure 2: Histogram of Anesthetic Durations for All Common Pediatric General Surgical Procedures.

This histogram includes anesthetic records reported to the NACOR from 2010 to 2015. The anesthetic durations of the 96,603 anesthetic records for the ten most common pediatric general surgical inpatient procedures are represented.

Table 3:

Duration in Minutes of Common Pediatric General Surgical Procedures

Procedure Type	n	Duration in Minutes
		Mean ± SD	10th Percentile	25th Percentile	Median	75th Percentile	90th Percentile
Bladder/ureter reconstruction	2,018	191.9 ± 90.2	98	136	180	227	290
Anti-reflux procedure	2,411	168.5 ± 71.2	104	124	152	195	250
Cholecystectomy	6,388	103.1 ± 46.1	63	76	93	119	151
Burn debridement	1,162	98.2 ± 68.9	38	56	81	117	175
PDA ligation	1,648	94.7 ± 69.8	45	57	75	110	155
Appendectomy	40,435	79.3 ± 32.9	53	62	73	89	109
Pyloromyotomy	3,187	80.9 ± 51.2	55	63	73	87	102
Central venous access	16,165	94.1 ± 84.7	39	51	71	100	157
Inguinal hernia	18,590	79.7 ± 50.8	44	55	70	91	118
Gastrostomy tube	4,599	66 ± 53.2	30	37	51	76	117
All Procedures	96,603	87.9 ± 58.6	47	59	75	97	134

All covariates as well as a surgical procedure type variable were included in a multivariable Poisson regression model. After adjusting for procedure type, given the large sample size, nearly all covariates were found to be statistically associated with anesthesia duration. The patient level covariates with potential clinical significance however included ages <1 year old as well as ≥1 to <3 years old, with anesthesia duration ratios (ADR) of 1.205 (95% confidence interval [CI], 1.202–1.208) and 1.125 (95% CI, 1.122–1.128) respectively compared to children ≥10 years old. (Table 4) This can be interpreted as children <1 year old having a 20.5% increased duration of anesthesia compared to children ≥10 years old after accounting for procedure type and other covariates. Male sex was associated with statistically longer duration, but an estimated 1.9% higher duration ([ADR]: 1.019 (95% CI, 1.017–1.020) is unlikely to be clinically significant. Durations in children between 3 and 5, and 5 and 10 years old were 1.2% and 2.9% shorter respectively than children older than 10 years of age, which is also unlikely to be clinically significant. Of all covariates, ASA 4 status was associated with the largest increase in duration with ADR: 1.381 (95% CI, 1.376–1.386), while ASA 3 status also conferred a rate of additional anesthetic duration ADR, 1.161 (95% CI, 1.158–1.163). The hospital level covariates associated with longer anesthetic duration included geographical region, with hospitals located in the Northeast having the longest durations. When evaluating facility type, the longest durations were seen at university hospitals, with ADR, 1.241 (95% CI, 1.236–1.246) compared to the reference of surgical centers, which had the shortest durations overall. This represents a 24.1% higher anesthetic duration in university hospitals compared to surgery centers.

Table 4:

Patient and Hospital Characteristics and Association with Anesthetic Duration

	Anesthesia Duration Ratio (ADR): Estimate (95% CI)
Age
<1 year old	1.205 (1.202 – 1.208)
≥1 to <3 years old	1.125 (1.122 – 1.128)
≥3 to < 5 years old	0.988 (0.985 – 0.991)
≥ 5 to <10 years old	0.971 (0.969 – 0.973)
≥10 to < 18 years old	Reference
Sex
Male	1.019 (1.017 – 1.020)
Female	Reference
ASA Status
1	Reference
2	1.029 (1.028 – 1.031)
3	1.161 (1.158 – 1.163)
≥4	1.381 (1.376 – 1.386)
US Geographical Region
Northeast	1.134 (1.131 – 1.137)
Midwest	1.056 (1.054 – 1.058)
South	1.049 (1.047 – 1.051)
West	Reference
Facility Type
University	1.241 (1.236 – 1.246)
Large Community	1.111 (1.107 – 1.115)
Medium Community	1.068 (1.064 – 1.072)
Small Community	1.001 (0.996 – 1.006)
Specialty	1.112 (1.108 – 1.117)
Surgery Center	Reference

Discussion:

In this study, we found that the median anesthetic duration in children was 57 minutes and that 7% of children had anesthetic durations greater than 3 hours. Longer exposures however were more commonly seen in younger children, with 15% of anesthetic cases lasting longer than 3 hours in children aged <1 year old. While the FDA has recommended avoiding anesthetic exposures of longer than 3 hours in children,² there remains some debate about this particular time threshold. One major reason for the lack of consensus is that the determination of this threshold is based on animal models and not clinical studies.¹⁸ Also while a number of clinical studies have evaluated anesthetic exposure duration, associations with neurodevelopmental deficits have been seen in exposures as short as 35 minutes to as long as 120 minutes^3-6 Interpretation of these studies however is complicated by the fact that a range of many different surgical procedures were included in the studies. As longer anesthetic duration is also related to the complexity of the surgical procedure and underlying comorbid disease, the increased risks of neurodevelopmental deficit cannot be conclusively attributed to the prolonged anesthetic exposure. As a result, while this warning against 3-hour exposures remains, the exact threshold of vulnerability in children, and whether a threshold even exists is still unclear.

When assessing the most commonly performed pediatric general surgery procedures, we found a significant amount of variation in anesthetic duration between procedure types. There was even significant variation within procedures, with the longest anesthetics for a given procedure lasting several times longer than the shortest anesthetic. This may have some important implications for population-based studies of anesthetic neurotoxicity, as there may be significant variability in the dose of anesthesia even for children receiving the same procedure.

In evaluating patient level factors associated with anesthetic duration, after adjusting for procedure type, procedures in children younger than 3 years old and those with ASA status higher than 3 were associated with longer anesthetic durations. These results are in contrast to a prior single center study which was unable to find associations between patient age and ASA status and differences in procedure duration.¹³ As age and ASA status are easily identifiable in patients preoperatively, these results may play a role in helping to predict surgical and anesthetic duration and thereby improve operating room efficiency. With regard to hospital level factors, that university hospitals were associated with the longest anesthetic exposures is not surprising given the participation of surgery and anesthesia trainees in intraoperative patient care. These results however are novel as we were able to compare and quantify the differences between the different facility types.

There are a number of limitations inherent in this study. First, the NACOR data is composed of records from institutions that voluntarily submit data and therefore these records may not be representative of the general population, or of all medical institutions within the United States. The NACOR however is the largest repository of anesthesia billing data in the US and offers data from a mixture of university and community hospitals. Second, there may be inconsistencies and inaccuracies between the participating institutions in the reporting of data. However, since the anesthetic duration is directly related to anesthesia billing, this variable is likely to be accurate. Third, while these durations identify exposure to anesthesia during individual procedures, since individual patients cannot be identified, the total duration of exposure for individual children who may have received multiple exposures cannot be calculated. As a result, if neurodevelopmental risk after anesthetic exposure is cumulative, this analysis may underestimate the total number of children with cumulative exposures of greater than 3 hours. Finally, while our results have found associations between specific patient and hospital level covariates and anesthesia duration, prediction of duration is complicated and a more complex model will likely be needed for anesthesia case time prediction.

Conclusions:

To our knowledge, this study is the first to evaluate anesthetic exposure duration in children and identify specific factors associated with duration using a large, multi-centered database. The results from this study have a number of important implications for the fields of anesthetic neurotoxicity as well as operating room efficiency. If prolonged anesthetic exposure is ultimately found to contribute to neurodevelopmental deficit in children, knowledge of the proportion of children with prolonged exposures will be helpful for determining the number of children potentially at risk for anesthetic induced neurotoxicity. In addition, an understanding of the factors associated with anesthetic exposure duration would not only help to further identify patients potentially at risk for prolonged exposures, but would also allow for the development of systems that may help to more efficiently allocate operating room resources.