Eur J Anaesthesiol 2015; 32:303–310
ORIGINAL ARTICLE
Risk of autistic disorder after exposure to general anaesthesia and surgery A nationwide, retrospective matched cohort study Wen-Ru Ko, Jing-Yang Huang, Yi-Chen Chiang, Oswald Ndi Nfor, Pei-Chieh Ko, Shiou-Rung Jan, Chia-Chi Lung, Hui-Chin Chang, Long-Yau Lin and Yung-Po Liaw BACKGROUND Deficits of learning, memory and cognition have been observed in newborn animals exposed to general anaesthetics. However, conclusions from clinical studies conducted in humans to investigate the relationship between anaesthesia and neurodevelopmental disorders have been inconsistent. Autistic disorder is typically recognised earlier than other neurobehavioural disorders. Although certain genes apparently contribute to autistic disorder susceptibility, other factors such as perinatal insults and exposure to neurotoxic agents may play a crucial role in gene–environmental interaction. OBJECTIVE This study was designed to investigate the association of exposure to general anaesthesia/surgery with autistic disorder. We hypothesised that exposure to general anaesthesia and surgery before 2 years of age is associated with an increased risk of developing autistic disorder. DESIGN A retrospective matched-cohort study. SETTING A medical university. Data from the National Health Insurance Research Database of Taiwan from 2001 to 2010 were analysed. PATIENTS The birth cohort included 114 435 children, among whom 5197 were exposed to general anaesthesia
and surgery before the age of 2 years. The 1 : 4 matched controls comprised 20 788 children. MAIN OUTCOME MEASURES The primary endpoint was the diagnosis of autistic disorder after the first exposure to general anaesthesia and surgery. RESULTS No differences were found in the incidence of autistic disorder between the exposed group (0.96%) and the unexposed controls (0.89%) (P ¼ 0.62). Cox proportional regression showed that the hazard ratio of exposure to general anaesthesia and surgery was 0.93 [95% confidence interval (95% CI) 0.57 to 1.53] after adjusting for potential confounders. Age at first exposure did not influence the risk of autistic disorder. No relationship was found between the total number of exposures and the risk of autistic disorder. CONCLUSION Exposure to general anaesthesia and surgery before the age of 2 years age at first exposure and number of exposures were not associated with the development of autistic disorder.
Published online 13 January 2015
This article is accompanied by the following Invited Commentary: Loepke AW, Hansen TG. Is this your (paediatric patient’s) brain on (anaesthetic) drugs? The search for a potential neurological phenotype of anaesthesia-related neurotoxicity in humans. Eur J Anaesthesiol 2015; 32:298–300.
From the Institute of Medicine, Chung Shan Medical University (W-RK, L-YL), Department of Anaesthesiology, Chung Shan Medical University and Hospital (W-RK), Department of Public Health and Institute of Public Health, Chung Shan Medical University (J-YH, Y-CC, ONN, P-CK, S-RJ, C-CL, Y-PL), Department of Family and Community Medicine, Chung Shan Medical University Hospital (Y-CC, C-CL, Y-PL), Department of Public Health (H-CC), EBM Centre & Library (H-CC), College of Medicine, Chung Shan Medical University (L-YL), Department of Obstetrics and Gynaecology, Chung Shan Medical University Hospital, Taichung, Taiwan (L-YL) Correspondence to Yung-Po Liaw, PhD, 110, Sec.1, Jianguo North Road, Taichung 40201, Taiwan E-mail: [email protected] 0265-0215 ß 2015 Copyright European Society of Anaesthesiology
DOI:10.1097/EJA.0000000000000130
Copyright © European Society of Anaesthesiology. Unauthorized reproduction of this article is prohibited.
304 Ko et al.
Introduction Various animal studies have revealed that general anaesthetics may cause apoptosis and degeneration of the immature nervous system in rodents and primates.1–13 Some authors have also observed long-term learning, memory and cognition deficits in newborn animals exposed to general anaesthetics.3–5,11–14 In humans, synaptogenesis is considered to extend from the embryonic period through 2 years of age and older.15 The conclusions from human clinical studies concerning the association between early general anaesthesia exposure and neurodevelopmental disorders have been discordant. Several studies have revealed that exposure to general anaesthesia in early life significantly increased the risk of developmental cognitive disorders,16–18 whereas other authors have found no evidence of a causal relationship between general anaesthesia and the risks of learning or behaviour disorders.19–22 Some studies that have investigated the total number of general anaesthetics have concluded that young children exposed to multiple, but not single, general anaesthetics were at a higher risk of developing attention-deficit/hyperactivity disorder23 and learning disabilities than those who were not exposed.24,25 Two studies by DiMaggio et al.16,17 included autism as one element of the composite outcomes. They reported an increased risk of developmental or behavioural disorders in children who underwent surgery in early life. Autism is a neurodevelopmental disorder with symptoms that include deficits in social interaction, learning and communication, as well as stereotypical behaviours, and is associated with negative outcomes among affected children and their families. Children with autism are considered to have developmental disorders of brain function.26 The pathogenesis of autistic disorder is not well established. Although certain types of gene have been identified and thought to contribute to susceptibility to autistic disorder,27,28 other factors including trauma,29,30 perinatal insults31–37 and exposure to neurotoxic agents38,39 could play a crucial role in gene– environment interaction. Autistic disorder is typically recognised earlier than are other neurobehavioural disorders and is considered a suitable outcome measurement for studying the neurodevelopmental effect of exposure to anaesthesia in early life. A number of prospective investigations into the association between general anaesthesia and neurodevelopment disorders are in progress, but it will take several years to clarify this relationship. We also observed that all studies published on anaesthesia exposure and neurodevelopmental disorders were conducted in Western countries. No information has been reported regarding an Asian population. We hypothesised that exposure to general anaesthesia and surgery in early life is associated with an increased risk of autistic disorder. A nationwide population-based
analysis was performed using data retrieved from the National Health Insurance Research Database (NHIRD) of Taiwan.
Materials and methods We designed a retrospective matched-cohort study. The study protocol (CSMUH No. CS13180) was reviewed and approved by the Institutional Review Board of Chung Shan Medical University Hospital, Taichung, Taiwan (Chairman C.P. Han) on 26 September 2013. The data were extracted from the NHIRD of Taiwan (http://nhird.nhri.org.tw/en/index.htm ). The National Health Research Institutes of Taiwan constructed the longitudinal database by randomly sampling one million insured people from the year registry (approximately 25 million people) of the National Health Insurance (NHI) programme every 5 years since 2000 (i.e. in 2000, 2005 and 2010). All the registration and claim data of the sampled people constituted the longitudinal database. The healthcare data from each sampled person are updated annually. The original NHIRD covers 99% of the population in Taiwan, and there were no significant differences in age and sex distributions between the people in the longitudinal databases and the people in the original NHIRD (http://nhird.nh ri.org.tw/date_cohort.htm#3; http://nhird.nhri.org.tw/en/ Data_Subsets.html#S3). The randomly sampled populations are thought to be representative of the insured population. The database is maintained and distributed by the Ministry of Health and Welfare and the National Health Research Institutes of Taiwan. Two longitudinal datasets (2005 and 2010), each containing one million samples, were used for analyses. Those that were sampled repeatedly were deleted from the 2010 database. Of the approximately two million insured, children born between 1 January 2001 and 31 December 2007 were all enrolled. The data were checked for general anaesthesia codes and autistic disorder diagnoses (ICD-9-CM 299.00). Children were allocated to the exposed group if the general anaesthesia codes were noted before the age of 2 years, and no diagnosis of autistic disorder had been found before anaesthesia occurred. The Statistical Analysis System (SAS) programme was used to assign a random number (uniform function) to each participant in the birth cohort. Data were sorted by random numbers. Each child in the exposed group was matched to four unexposed children on the basis of sex and birth year/month using PROC SQL in SAS. The index date for the matched controls was defined as the day on which the exposed counterparts had their first exposure. The matched controls had not been diagnosed with an autistic disorder before the index date.
Eur J Anaesthesiol 2015; 32:303–310 Copyright © European Society of Anaesthesiology. Unauthorized reproduction of this article is prohibited.
Autistic disorder after exposure to anaesthesia and surgery 305
Data were checked for the diagnostic codes for perinatal conditions (ICD-9-CM 765, disorders relating to short gestation and low birth weight; 768, intrauterine hypoxia and birth asphyxia; 769, respiratory distress syndrome; 770, other respiratory conditions of fetus and newborn; 771, infections specific to the perinatal period; 775, endocrine and metabolic disturbances specific to the fetus and newborn), congenital anomalies (ICD-9-CM 740), anencephalus and similar anomalies; 741, spina bifida; 742, other congenital anomalies of the nervous system; 743, congenital anomalies of the eye; 744, congenital anomalies of ear, face, and neck; 758, chromosomal anomalies; 759, other and unspecified congenital anomalies), neurological diseases (ICD-9-CM 320 to 326, inflammatory diseases of the central nervous system; 330, cerebral degeneration usually manifest in childhood; 331, other cerebral degeneration; 333, other extrapyramidal diseases and abnormal movement disorders; 334, spinocerebellar disease; 335, anterior horn cell disease; 336, other diseases of the spinal cord; 337, disorders of the autonomic nervous system; 343, infantile cerebral palsy; 344, other paralytic syndromes; 345, epilepsy and recurrent seizures), endocrine diseases (ICD9-CM 243, congenital hypothyroidism; 250, diabetes mellitus; 251, other disorders of pancreatic internal secretion; 252, disorders of the parathyroid gland; 253, disorders of the pituitary gland and its hypothalamic control; 254, diseases of the thymus gland; 255, disorders of the adrenal glands; 256, ovarian dysfunction; 257, testicular dysfunction; 258, polyglandular dysfunction and related disorders; 259, other endocrine disorders) and categories of surgery that had been undergone (skin, limbs, head and neck operations; heart and thoracic operations; abdominal operations; operations on the nervous system). For the exposed participants, observations began on the day of the first exposure to general anaesthesia and surgery. The matched controls were entered on the same day (index date) as their exposed counterparts. All exposed children and the corresponding matched controls were considered to be at risk until a diagnosis of autistic disorder (ICD-9-CM 299.00) was made. We concluded the observation period on 31 December 2010. Two additional analyses were performed on the basis of age at the time of first exposure and cumulative number of exposures. In the first additional analysis, the exposed group was stratified into four categories on the basis of the quartiles of age at the time of the first exposure. In the second analysis, the study participants were stratified into three categories according to the number of exposures (none, single or multiple).
Statistical analysis Multivariate Cox proportional hazards regression was used to assess whether exposure to general anaesthesia
and surgery before the age of 2 years was a risk factor for developing an autistic disorder, both with and without adjustment for potential confounders, including home location, perinatal conditions (ICD-9-CM 765, 768 to 771, 775), congenital anomalies (ICD-9-CM 740 to 744, 758, 759), neurological diseases (ICD-9-CM 320 to 326, 330, 331, 333 to 337, 343 to 345), endocrine diseases (ICD-9CM 243, 250, 251 to 259) and surgical operations (skin, limbs, head and neck; heart and thoracic; abdominal and nervous system). Data were censored for death, leaving the NHI programme or on the last follow-up before 31 December 2010. In the additional analyses, the hazard ratio of autistic disorders was calculated with and without adjusting for the same potential confounders as in the main analysis. The results were summarised as hazard ratio estimates and corresponding 95% confidence intervals (95% CIs). For nominal variables and stratified nominal variables, P values less than 0.05 in the x2 test and Cochran Mantel Haenszel x2 tests, respectively, were considered statistically significant. Analyses were performed using SAS statistical software (version 9.3; SAS Institute, Inc., Cary, North Carolina, USA).
Results A total of 114 435 children were born in Taiwan between 1 January 2001 and 31 December 2007. Among the cohort, 5197 children (3672 boys and 1525 girls) were noted to have undergone general anaesthesia and surgery before the age of 2 years. All were free of autistic disorders before exposure to anaesthesia and surgery. The 1 : 4 matched controls comprised 20 788 children who were neither exposed to general anaesthesia before 2 years of age nor diagnosed with an autistic disorder before the index date (Fig. 1). The mean SD age at diagnosis of autistic disorder was 4.04 1.80 years in this cohort, 3.82 1.68 years in the exposed group and 4.10 1.82 years in the matched controls (P ¼ 0.16). The incidences of autistic disorder before 31 December 2010 were 0.96% (50 of 5197) in the exposed group and 0.89% (185 of 20 788) in the matched controls (P ¼ 0.62). Compared with the unexposed children, those exposed to general anaesthesia and surgery before the age of 2 years were more likely to have been diagnosed with perinatal conditions (P < 0.001), congenital anomalies (P < 0.001), neurological disease (P < 0.001) or endocrine disease (P < 0.001). No significant differences were found between the two groups in home location or parental occupation (Table 1). Multivariate Cox proportional hazards regression showed that congenital anomalies, neurological diseases and endocrine diseases were associated with a risk of autistic disorder without any adjustments. The association of
Eur J Anaesthesiol 2015; 32:303–310 Copyright © European Society of Anaesthesiology. Unauthorized reproduction of this article is prohibited.
306 Ko et al.
Fig. 1
Children born between 1 January 2001 and 31 December 2007 n = 114 435 (M, 59 960; F, 54 475)
GA exposure before the age of two
no GA exposure before the age of two
yearsa
years
n = 5197 (M, 3672; F, 1525)
n = 109 238 (M, 56 288; F, 52 950)
1:4 matched controlsb n = 20 788 (M, 14 688; F, 6100)
Recruitment of study participants. GA, general anaesthesia. aAll the exposed individuals were free of autistic disorder before exposure to anaesthesia and surgery. bThe matched controls were free of autistic disorder before the index dates.
endocrine diseases with autistic disorder became nonsignificant after adjusting for other potential confounders. The association between exposure to general anaesthesia and surgery before the age of 2 years and an autistic disorder was nonsignificant, either before or after adjusting for home location, perinatal conditions, congenital Table 1
anomalies, neurological diseases, endocrine diseases and categories of surgery (Table 2). Age categories at the time of first exposure were defined as 0 to 108 days, 109 to 327 days, 328 to 508 days and 509 days to 2 years according to quartiles of age at the
Characteristics of the individuals with and without general anaesthesia and surgery exposure before the age of 2 years With exposure (n U 5197) n (%)
Sex Female Male Home locationa I II III IV V VI VII Parental occupationb I II III IV Missing data Perinatal conditionsc Congenital anomaliesd Neurological diseasese Endocrine diseasesf Autistic disorder
Without exposure (n U 20 788) n (%)
P 1.00
1525 (29.3) 3672 (70.7)
6100 (29.3) 14 688 (70.7) 0.95
1497 1556 947 719 101 188 189
(28.8) (29.9) (18.2) (13.8) (1.9) (3.6) (3.6)
5838 6429 3846 2750 391 768 766
(28.1) (30.9) (18.5) (13.2) (1.9) (3.7) (3.7)
2697 559 532 619 790 1009 2827 482 189 50
(51.9) (10.8) (10.2) (11.9) (15.2) (19.4) (54.4) (9.3) (3.6) (0.96)
10 939 2059 1913 2587 3290 1984 2846 762 417 185
(52.6) (9.9) (9.2) (12.4) (15.8) (9.5) (13.7) (3.7) (2.0) (0.89)
0.69
ORIGINAL ARTICLE
Risk of autistic disorder after exposure to general anaesthesia and surgery A nationwide, retrospective matched cohort study Wen-Ru Ko, Jing-Yang Huang, Yi-Chen Chiang, Oswald Ndi Nfor, Pei-Chieh Ko, Shiou-Rung Jan, Chia-Chi Lung, Hui-Chin Chang, Long-Yau Lin and Yung-Po Liaw BACKGROUND Deficits of learning, memory and cognition have been observed in newborn animals exposed to general anaesthetics. However, conclusions from clinical studies conducted in humans to investigate the relationship between anaesthesia and neurodevelopmental disorders have been inconsistent. Autistic disorder is typically recognised earlier than other neurobehavioural disorders. Although certain genes apparently contribute to autistic disorder susceptibility, other factors such as perinatal insults and exposure to neurotoxic agents may play a crucial role in gene–environmental interaction. OBJECTIVE This study was designed to investigate the association of exposure to general anaesthesia/surgery with autistic disorder. We hypothesised that exposure to general anaesthesia and surgery before 2 years of age is associated with an increased risk of developing autistic disorder. DESIGN A retrospective matched-cohort study. SETTING A medical university. Data from the National Health Insurance Research Database of Taiwan from 2001 to 2010 were analysed. PATIENTS The birth cohort included 114 435 children, among whom 5197 were exposed to general anaesthesia
and surgery before the age of 2 years. The 1 : 4 matched controls comprised 20 788 children. MAIN OUTCOME MEASURES The primary endpoint was the diagnosis of autistic disorder after the first exposure to general anaesthesia and surgery. RESULTS No differences were found in the incidence of autistic disorder between the exposed group (0.96%) and the unexposed controls (0.89%) (P ¼ 0.62). Cox proportional regression showed that the hazard ratio of exposure to general anaesthesia and surgery was 0.93 [95% confidence interval (95% CI) 0.57 to 1.53] after adjusting for potential confounders. Age at first exposure did not influence the risk of autistic disorder. No relationship was found between the total number of exposures and the risk of autistic disorder. CONCLUSION Exposure to general anaesthesia and surgery before the age of 2 years age at first exposure and number of exposures were not associated with the development of autistic disorder.
Published online 13 January 2015
This article is accompanied by the following Invited Commentary: Loepke AW, Hansen TG. Is this your (paediatric patient’s) brain on (anaesthetic) drugs? The search for a potential neurological phenotype of anaesthesia-related neurotoxicity in humans. Eur J Anaesthesiol 2015; 32:298–300.
From the Institute of Medicine, Chung Shan Medical University (W-RK, L-YL), Department of Anaesthesiology, Chung Shan Medical University and Hospital (W-RK), Department of Public Health and Institute of Public Health, Chung Shan Medical University (J-YH, Y-CC, ONN, P-CK, S-RJ, C-CL, Y-PL), Department of Family and Community Medicine, Chung Shan Medical University Hospital (Y-CC, C-CL, Y-PL), Department of Public Health (H-CC), EBM Centre & Library (H-CC), College of Medicine, Chung Shan Medical University (L-YL), Department of Obstetrics and Gynaecology, Chung Shan Medical University Hospital, Taichung, Taiwan (L-YL) Correspondence to Yung-Po Liaw, PhD, 110, Sec.1, Jianguo North Road, Taichung 40201, Taiwan E-mail: [email protected] 0265-0215 ß 2015 Copyright European Society of Anaesthesiology
DOI:10.1097/EJA.0000000000000130
Copyright © European Society of Anaesthesiology. Unauthorized reproduction of this article is prohibited.
304 Ko et al.
Introduction Various animal studies have revealed that general anaesthetics may cause apoptosis and degeneration of the immature nervous system in rodents and primates.1–13 Some authors have also observed long-term learning, memory and cognition deficits in newborn animals exposed to general anaesthetics.3–5,11–14 In humans, synaptogenesis is considered to extend from the embryonic period through 2 years of age and older.15 The conclusions from human clinical studies concerning the association between early general anaesthesia exposure and neurodevelopmental disorders have been discordant. Several studies have revealed that exposure to general anaesthesia in early life significantly increased the risk of developmental cognitive disorders,16–18 whereas other authors have found no evidence of a causal relationship between general anaesthesia and the risks of learning or behaviour disorders.19–22 Some studies that have investigated the total number of general anaesthetics have concluded that young children exposed to multiple, but not single, general anaesthetics were at a higher risk of developing attention-deficit/hyperactivity disorder23 and learning disabilities than those who were not exposed.24,25 Two studies by DiMaggio et al.16,17 included autism as one element of the composite outcomes. They reported an increased risk of developmental or behavioural disorders in children who underwent surgery in early life. Autism is a neurodevelopmental disorder with symptoms that include deficits in social interaction, learning and communication, as well as stereotypical behaviours, and is associated with negative outcomes among affected children and their families. Children with autism are considered to have developmental disorders of brain function.26 The pathogenesis of autistic disorder is not well established. Although certain types of gene have been identified and thought to contribute to susceptibility to autistic disorder,27,28 other factors including trauma,29,30 perinatal insults31–37 and exposure to neurotoxic agents38,39 could play a crucial role in gene– environment interaction. Autistic disorder is typically recognised earlier than are other neurobehavioural disorders and is considered a suitable outcome measurement for studying the neurodevelopmental effect of exposure to anaesthesia in early life. A number of prospective investigations into the association between general anaesthesia and neurodevelopment disorders are in progress, but it will take several years to clarify this relationship. We also observed that all studies published on anaesthesia exposure and neurodevelopmental disorders were conducted in Western countries. No information has been reported regarding an Asian population. We hypothesised that exposure to general anaesthesia and surgery in early life is associated with an increased risk of autistic disorder. A nationwide population-based
analysis was performed using data retrieved from the National Health Insurance Research Database (NHIRD) of Taiwan.
Materials and methods We designed a retrospective matched-cohort study. The study protocol (CSMUH No. CS13180) was reviewed and approved by the Institutional Review Board of Chung Shan Medical University Hospital, Taichung, Taiwan (Chairman C.P. Han) on 26 September 2013. The data were extracted from the NHIRD of Taiwan (http://nhird.nhri.org.tw/en/index.htm ). The National Health Research Institutes of Taiwan constructed the longitudinal database by randomly sampling one million insured people from the year registry (approximately 25 million people) of the National Health Insurance (NHI) programme every 5 years since 2000 (i.e. in 2000, 2005 and 2010). All the registration and claim data of the sampled people constituted the longitudinal database. The healthcare data from each sampled person are updated annually. The original NHIRD covers 99% of the population in Taiwan, and there were no significant differences in age and sex distributions between the people in the longitudinal databases and the people in the original NHIRD (http://nhird.nh ri.org.tw/date_cohort.htm#3; http://nhird.nhri.org.tw/en/ Data_Subsets.html#S3). The randomly sampled populations are thought to be representative of the insured population. The database is maintained and distributed by the Ministry of Health and Welfare and the National Health Research Institutes of Taiwan. Two longitudinal datasets (2005 and 2010), each containing one million samples, were used for analyses. Those that were sampled repeatedly were deleted from the 2010 database. Of the approximately two million insured, children born between 1 January 2001 and 31 December 2007 were all enrolled. The data were checked for general anaesthesia codes and autistic disorder diagnoses (ICD-9-CM 299.00). Children were allocated to the exposed group if the general anaesthesia codes were noted before the age of 2 years, and no diagnosis of autistic disorder had been found before anaesthesia occurred. The Statistical Analysis System (SAS) programme was used to assign a random number (uniform function) to each participant in the birth cohort. Data were sorted by random numbers. Each child in the exposed group was matched to four unexposed children on the basis of sex and birth year/month using PROC SQL in SAS. The index date for the matched controls was defined as the day on which the exposed counterparts had their first exposure. The matched controls had not been diagnosed with an autistic disorder before the index date.
Eur J Anaesthesiol 2015; 32:303–310 Copyright © European Society of Anaesthesiology. Unauthorized reproduction of this article is prohibited.
Autistic disorder after exposure to anaesthesia and surgery 305
Data were checked for the diagnostic codes for perinatal conditions (ICD-9-CM 765, disorders relating to short gestation and low birth weight; 768, intrauterine hypoxia and birth asphyxia; 769, respiratory distress syndrome; 770, other respiratory conditions of fetus and newborn; 771, infections specific to the perinatal period; 775, endocrine and metabolic disturbances specific to the fetus and newborn), congenital anomalies (ICD-9-CM 740), anencephalus and similar anomalies; 741, spina bifida; 742, other congenital anomalies of the nervous system; 743, congenital anomalies of the eye; 744, congenital anomalies of ear, face, and neck; 758, chromosomal anomalies; 759, other and unspecified congenital anomalies), neurological diseases (ICD-9-CM 320 to 326, inflammatory diseases of the central nervous system; 330, cerebral degeneration usually manifest in childhood; 331, other cerebral degeneration; 333, other extrapyramidal diseases and abnormal movement disorders; 334, spinocerebellar disease; 335, anterior horn cell disease; 336, other diseases of the spinal cord; 337, disorders of the autonomic nervous system; 343, infantile cerebral palsy; 344, other paralytic syndromes; 345, epilepsy and recurrent seizures), endocrine diseases (ICD9-CM 243, congenital hypothyroidism; 250, diabetes mellitus; 251, other disorders of pancreatic internal secretion; 252, disorders of the parathyroid gland; 253, disorders of the pituitary gland and its hypothalamic control; 254, diseases of the thymus gland; 255, disorders of the adrenal glands; 256, ovarian dysfunction; 257, testicular dysfunction; 258, polyglandular dysfunction and related disorders; 259, other endocrine disorders) and categories of surgery that had been undergone (skin, limbs, head and neck operations; heart and thoracic operations; abdominal operations; operations on the nervous system). For the exposed participants, observations began on the day of the first exposure to general anaesthesia and surgery. The matched controls were entered on the same day (index date) as their exposed counterparts. All exposed children and the corresponding matched controls were considered to be at risk until a diagnosis of autistic disorder (ICD-9-CM 299.00) was made. We concluded the observation period on 31 December 2010. Two additional analyses were performed on the basis of age at the time of first exposure and cumulative number of exposures. In the first additional analysis, the exposed group was stratified into four categories on the basis of the quartiles of age at the time of the first exposure. In the second analysis, the study participants were stratified into three categories according to the number of exposures (none, single or multiple).
Statistical analysis Multivariate Cox proportional hazards regression was used to assess whether exposure to general anaesthesia
and surgery before the age of 2 years was a risk factor for developing an autistic disorder, both with and without adjustment for potential confounders, including home location, perinatal conditions (ICD-9-CM 765, 768 to 771, 775), congenital anomalies (ICD-9-CM 740 to 744, 758, 759), neurological diseases (ICD-9-CM 320 to 326, 330, 331, 333 to 337, 343 to 345), endocrine diseases (ICD-9CM 243, 250, 251 to 259) and surgical operations (skin, limbs, head and neck; heart and thoracic; abdominal and nervous system). Data were censored for death, leaving the NHI programme or on the last follow-up before 31 December 2010. In the additional analyses, the hazard ratio of autistic disorders was calculated with and without adjusting for the same potential confounders as in the main analysis. The results were summarised as hazard ratio estimates and corresponding 95% confidence intervals (95% CIs). For nominal variables and stratified nominal variables, P values less than 0.05 in the x2 test and Cochran Mantel Haenszel x2 tests, respectively, were considered statistically significant. Analyses were performed using SAS statistical software (version 9.3; SAS Institute, Inc., Cary, North Carolina, USA).
Results A total of 114 435 children were born in Taiwan between 1 January 2001 and 31 December 2007. Among the cohort, 5197 children (3672 boys and 1525 girls) were noted to have undergone general anaesthesia and surgery before the age of 2 years. All were free of autistic disorders before exposure to anaesthesia and surgery. The 1 : 4 matched controls comprised 20 788 children who were neither exposed to general anaesthesia before 2 years of age nor diagnosed with an autistic disorder before the index date (Fig. 1). The mean SD age at diagnosis of autistic disorder was 4.04 1.80 years in this cohort, 3.82 1.68 years in the exposed group and 4.10 1.82 years in the matched controls (P ¼ 0.16). The incidences of autistic disorder before 31 December 2010 were 0.96% (50 of 5197) in the exposed group and 0.89% (185 of 20 788) in the matched controls (P ¼ 0.62). Compared with the unexposed children, those exposed to general anaesthesia and surgery before the age of 2 years were more likely to have been diagnosed with perinatal conditions (P < 0.001), congenital anomalies (P < 0.001), neurological disease (P < 0.001) or endocrine disease (P < 0.001). No significant differences were found between the two groups in home location or parental occupation (Table 1). Multivariate Cox proportional hazards regression showed that congenital anomalies, neurological diseases and endocrine diseases were associated with a risk of autistic disorder without any adjustments. The association of
Eur J Anaesthesiol 2015; 32:303–310 Copyright © European Society of Anaesthesiology. Unauthorized reproduction of this article is prohibited.
306 Ko et al.
Fig. 1
Children born between 1 January 2001 and 31 December 2007 n = 114 435 (M, 59 960; F, 54 475)
GA exposure before the age of two
no GA exposure before the age of two
yearsa
years
n = 5197 (M, 3672; F, 1525)
n = 109 238 (M, 56 288; F, 52 950)
1:4 matched controlsb n = 20 788 (M, 14 688; F, 6100)
Recruitment of study participants. GA, general anaesthesia. aAll the exposed individuals were free of autistic disorder before exposure to anaesthesia and surgery. bThe matched controls were free of autistic disorder before the index dates.
endocrine diseases with autistic disorder became nonsignificant after adjusting for other potential confounders. The association between exposure to general anaesthesia and surgery before the age of 2 years and an autistic disorder was nonsignificant, either before or after adjusting for home location, perinatal conditions, congenital Table 1
anomalies, neurological diseases, endocrine diseases and categories of surgery (Table 2). Age categories at the time of first exposure were defined as 0 to 108 days, 109 to 327 days, 328 to 508 days and 509 days to 2 years according to quartiles of age at the
Characteristics of the individuals with and without general anaesthesia and surgery exposure before the age of 2 years With exposure (n U 5197) n (%)
Sex Female Male Home locationa I II III IV V VI VII Parental occupationb I II III IV Missing data Perinatal conditionsc Congenital anomaliesd Neurological diseasese Endocrine diseasesf Autistic disorder
Without exposure (n U 20 788) n (%)
P 1.00
1525 (29.3) 3672 (70.7)
6100 (29.3) 14 688 (70.7) 0.95
1497 1556 947 719 101 188 189
(28.8) (29.9) (18.2) (13.8) (1.9) (3.6) (3.6)
5838 6429 3846 2750 391 768 766
(28.1) (30.9) (18.5) (13.2) (1.9) (3.7) (3.7)
2697 559 532 619 790 1009 2827 482 189 50
(51.9) (10.8) (10.2) (11.9) (15.2) (19.4) (54.4) (9.3) (3.6) (0.96)
10 939 2059 1913 2587 3290 1984 2846 762 417 185
(52.6) (9.9) (9.2) (12.4) (15.8) (9.5) (13.7) (3.7) (2.0) (0.89)
0.69
