The most serious peri-operative respiratory event (PRE) in pediatric anesthesia is desaturation or hypoxemia which could lead to cardiovascular collapse or cardiac arrest. Intermittent hypoxic episode especially in infants are also associated with impaired growth, longer-term cardiorespiratory instability and poor neurodevelopmental outcome.12 The mechanism of peri-operative desaturation occurring in normoxia infant brain is quite similar to overabundance of oxygen in the acutely hypoxic infant by using 100% oxygen or hyperoxia for resuscitation of acutely asphyxiated infants which can generate excessive neurotoxic compounds and increase oxidative stress markers.17 Anesthetic agents which involve gamma-aminobutyric acid (GABA)receptors eg; volatile agents, midazolam and N-methyl-D-aspartate receptor (NMDA) receptors eg; nitrous oxide, ketamine can cause neuronal apoptosis, neuronal necrosis, neuronal cell death and memory deficit in rat pups. Moreover, prolonged anesthetic exposure, irrespective of open heart surgery, can influence neurodevelopment of the brain in rodents.17 However, the evidence for anesthetic agents causing apoptosis and neurodegeneration in human neonatal brain is still not clear.

Thus, peri-operative desaturation occurred in young age regardless of severity combined with general anesthesia might possibly affect the long-term impact regarding growth and neurodevelopmental outcome in infant or intelligence outcome in older children. In our study, we are interested in looking at the intelligence outcome, which is a part of neurodevelopment outcome, in preschool children aged ≤ 5 years who developed desaturation peri-operatively. Because we include a wide range of age between newborn to five years old to test neurodevelopment outcome in older children, the intelligence outcome may be more appropriate and can be applied to infants and younger ages. Therefore, the objective of study was to compare intelligence outcome between children who developed peri-operative desaturation and children who did not develop PRE.

Material and methods A historical and concurrent follow up study was conducted at Songklanagarind Hospital, an 853-bed tertiary care hospital in southern Thailand, after approval by the Ethics Committee, Prince of Songkla University at November 15th, 2012. Written informed consent was signed by all parents who participated in the study.

Participants Children aged ≤ 5 years old who received general anesthesia (GA)and developed intraoperative or post-anesthetic care unit (PACU)desaturation were included in the study and was defined as desaturation group. The exclusion criteria were ASA (American Society of Anesthesiologist) classification 4 or 5, preoperative oxygen saturation < 95% at room air, required preoperative endotracheal tube intubation (ETT) or mechanical ventilation, had congenital heart disease or open heart surgery, had neurosurgery, had preoperative delayed development, and had perinatal hypoxia or fetal growth retardation. Perioperative desaturation was defined as oxygen saturation < 95% for more than 10 seconds [9]. Children in desaturation group were matched with children in non-desaturation group one per one by the same sex, same age within the same year of general anesthesia, same type of surgery and same technique of anesthesia. Non-desaturation group was defined as children who did not developed perioperative respiratory events eg; desaturation, laryngospasm, bronchospasm, upper airway obstruction and reintubation. After a child from desaturation group was identified, one child from non-desaturation group was randomly selected from the lists of 3 to 4 children matching with same demographic, same surgery and same anesthesia profile above. The parents of both groups were contacted by the investigator by phone call and invitation postcard. If selected child in the non-desaturation group was declined by the parents or could not be reached, the next child on the list would be randomly selected.

Outcome of interest The outcomes of interest were divided into growth developmental outcome and intelligence outcome. The intelligence outcome and growth developmental outcome were evaluated at least 6 months after exposing to general anesthesia. Children aged between at least 2 years and not more than 9 years were equally compared the outcomes of interest between desaturation and non-desaturation groups.

For growth development outcome, preoperative body weight, percentile weight, height and percentile height were compared with the current body weight, current height, percentile weight and percentile height. Month age of first meaning word and month age of first walk will be asked.

The main outcome will be intelligence score or intelligence quotient (IQ) score. There are two intelligence tests available in Songklanagarind hospital which are suitable for testing intelligence outcome in this study; the Standford Binet form L-M and Wechsler Intelligence Scale for Children, 3rd edition (WISC-III). The same aged of children in desaturation and non-desaturation groups were appointed to have an IQ test during the study period. A child psychologist (T.D.) performed the IQ test in all children. The Standford Binet form L-M was the main intelligence test used in our study. It measures cognitive ability regarding verbal reasoning, quantitative reasoning, abstract and visual reasoning and short-term memory skills and a wide range of IQ score varying from 20 to 140. The suitable age for using the Standford Binet form L-M is between 2 to 7 years old. The WISC-III test was used for supplemental test for children who reached the ceiling aged of 7 years and still could not complete IQ measurement. The WISC-III consists of 2 main subtests eg; the verbal subtests and performance subtests (nonverbal subtest) which the average IQ score was used to compared between 2 groups. The same IQ score of each IQ test was considered to have the similar level of intelligence outcome in the study. The type of IQ test was recorded and compared between 2 groups.

The potential predictor variables The children and family related predictors and anesthesia related predictors were included in the data record. The children and family related predictors were history of prematurity, history of parental smoking, religion, mother age at delivery, mother occupation, father occupation, monthly family income, mother education, father education, children education, training by parents (no, partial or regular) and children capability (full help, partial help or no help).

Definition of training by parents Training by parents was defined as developmental stimulation and training by parents or caregivers including with the following statement eg; (1) teaching them to do thing by themselves compatible to their age, if you do regularly in days or weeks considered to be regular stimulation and training,if you do sometimes or once in a while considered to be partial stimulation and training. (2) asking them a simple question to stimulate their mind, if you do regularly in days or weeks considered to be regular stimulation and training, if you do sometimes or once in a while considered to be partial stimulation and training. (3) Taking them outside for a ride or to the playground and stimulating them by showing things, people, animal or asking questions, if you do regularly in days or weeks by both taking them outside and stimulating their mind considered to be regular stimulation and training, if you do sometimes or once in a while considered to be partial stimulation and training, if you take them outside but not stimulate their mind considered to be partial stimulation and training. (4) Buying them a stimulation toys and play with them, if you buy toys and play with them regularly in days or weeks considered to be regular stimulation and training, if you buy them toys but not play with them considered to be partial stimulation and training or if none of the 4 items is performed considered to be no stimulation and training. Training by parents was classified as 3 categories which were no training, partial training or regular training.

Definition of children capability Children capability was defined as capability of children who are able to do things by themselves compatible to their age. For 2 to 3 years of age, children can climb up stair and down stair, can climb off the cot and can take off their cloth. For 3 to 5 years of age, children can eat by spoon, can take a shower, can brush their teeth and can ride a 3 wheel bike. For 5 to 6 years of age, children can get dressed with no button, can go to the toilet and can wear shoes without shoelaces. For 6 to 8 years of age, children can draw and write on the paper, can get dressed and do the button and can tie a shoelace. Children capability was classified as 3 categories which were full help, partial help and no help.

Anesthesia related predictors were ASA classification, time of repeated GA, type of surgery, choice of GA, technique of GA, induction agents, intubation agents, inhalation agents, gas mixed with oxygen, narcotic and duration of anesthesia.

Statistical analysis Analysis was performed with the R program version 2.14.1. Descriptive statistics was computed for all variables and include frequency, proportion, mean + standard deviation (SD) and median (interquartile range; IQR). Predictor variables were continuous or categorical data either originally or if categorical data, after categorizing by selection of suitable cut points.

The growth developmental outcome eg; the current body weight, current height, percentile weight and percentile height including other continuous data such as age at index GA, the current age, weight and height at index GA, percentile weight and height at index, mother age at delivery, monthly family income, first meaning word, first walk, time of repeated GA, duration of anesthesia were analyzed by Student's t- test or non- parametric rank sum test as appropriate.

Categorical data consisting of children and family profiles eg; gender, history of prematurity, history of parental smoking, religion, mother occupation, father occupation, mother education, father education, children education, training by parents, children capability; and anesthesia profiles eg; ASA (American Society of Anesthesiologist) classification, type of surgery, choice of GA, technique of GA, induction agents, intubation agents, inhalation agents, gas mixed with oxygen and narcotics were compared using the Chi-square test or Fisher 's exact test as appropriate.

The IQ score was compared between desaturation and non-desaturation group using simple linear regression model. The other explanatory variables that may be related to IQ score, percentile weight and percentile height were analyzed using multiple linear regression models if showing some evidence of differing across outcome categories in univariate analysis (p ≤ 0.2). The magnitudes and precision of associations will be indicated by adjusted coefficient and their 95% confidence intervals. In multivariate modeling techniques, association with outcome will be considered significant if the likelihood ratio p-values are ≤ 0.05.

Sample size calculation The sample size calculation is based on comparison of an IQ score between desaturation group and non-desaturation group after receiving general anesthesia. According to National Thai Survey of 72,780 school children over 76 provinces in 2007-2008 using the Standard Progressive Matrices (SPM parallel version; updated 2003)45, the average IQ score of Thai children is 98.59 whereas the average IQ score of southern Thai children is a bit lower (96.85). We assumed that 5 point difference in IQ score would have an impact to the intelligence outcome of school children (aged < 9 year old). Therefore, the sample size for testing hypothesis of a difference between 2 means based on the mean difference of IQ score of 5 and standard deviation of 12 would be 91 children per group under a power of 80% and type I error of 5% and 102 children per group would be required to compensate of 10% drop out under data collection.

Inclusion Criteria:

• Children aged ≤ 5 years old who received general anesthesia and developed intraoperative or post-anesthetic care unit (PACU) desaturation between 2008 and 2011 were included in the study and was defined as desaturation group.

Exclusion Criteria:

• The exclusion criteria were ASA classification 4 or 5, preoperative oxygen saturation < 95% at room air, required preoperative endotracheal tube intubation (ETT) or mechanical ventilation, had congenital heart disease or open heart surgery, had neurosurgery, had preoperative delayed development, and had perinatal hypoxia or fetal growth retardation.

children aged 0-5 years who receive general anesthesia at Songklanajarind Hospital between 2008-2011