Is Neonatal Hyperbilirubinemia Associated with Autism Spectrum Disorders - Research Paper Example

Epidemiology Abstract Merrill (2010) defines epidemiology as the investigation of the distribution and determining factors of health related states or events in human populations and the application of this study to the prevention and control of health problems. Epidemiology involves sound methods of scientific investigation. Epidemiologic investigations involve descriptive and analytical methods which draw on statistical techniques fro describing data and evaluating hypotheses, biological principles and the causal theory. This papers endeavors to give a comparison of two peer reviewed journal articles which examine the same hypothesis but employ different study designs. The two articles have a similar hypothesis. Both articles are trying to investigate whether neonatal jaundice also known as neonatal/infant hyperbilirubinemia might be related to Autism Related Disorders (ARDs). The first article, Pediatrics, retrieved from the official journal of the American Academy of Pediatrics employs the case-control cohort study while the second article, retrieved from online publications uses retrospective case control design. Therefore, this paper will examine how each investigation in each article was conducted. It will address the strengths and weaknesses the two approaches. The paper will conclude with suggestions on how I would design an ideal study to investigate the hypothesis. Key words: neonatal hyperbilirubinemia/jaundice (elevated levels of bilirubin in infants) Epidemiology The research objective with regard to both journal articles was to investigate the association between neonatal hyperbilirubinemia/jaundice and autism spectrum disorders. As stated earlier, the investigations carried out in the journal article of pediatrics employed case control cohort study. A large, population based case control study, was carried out within an integrated health plan with extensive computerized data resources involving prospectively laboratory collected results and diagnoses (Pediatrics Digest as cited in Lisa et al., 2005). Case and control subjects were identified from the cohort of infants who were born at a northern California Kaiser Permanente (KP) facility between January 1995 and December 1998 and remained KP members for more than two years after birth (n=73 291) (Pediatrics Digest as cited in Lisa et al., 2005). Case subjects (n=393) were defined as children for whom an ASD diagnosis, i.e., Asperger’s syndrome or pervasive development disorder were recorded in KP outpatients clinical data bases at any time between January 1995 and December 1998…All the children were between the 4 and 7 years old at the time the data base was scanned. 5 control subjects per case subject, were randomly selected from the cohort of children who did not have the diagnosis of ASD recorded in the clinical database (n=1965) (Pediatrics Digest as cited in Lisa et al., 2005). Control subjects were frequency matched with case subjects according to gender, birth, year and hospital of birth… The final study sample was restricted to singleton infants at gestational (Pediatrics Digest as cited in Lisa et al., 2005) of greater than 35 weeks for whom information on neonatal bilirubin levels was available. Information on the neonatal hyperbilirubinemia was derived from bilirubin tests recorded in the KP region wide Integrated Laboratory Information System which contains the date, time and results of all laboratory tests preformed for KP patients (Pediatrics Digest as cited in Lisa et al., 2005). The maximum bilirubin level measured in mg/dl was defined as the highest recorded bilirubin level within the first 30 days of life. Data on phototherapy was derived from the computerized patient databases, which contained detailed information on diagnoses and procedures as well as dates and locations of the patient visits. This data was found to be 94% sensitive and 100% specific when compared to chart review (Pediatrics Digest as cited in Lisa et al., 2005). Maternal age at delivery, child gender, gestational age and birth weight were determined from the information recorded in the KP inpatient data base (Pediatrics Digest as cited in Lisa et al., 2005). Data on maternal age and educational attainment ant the time of delivery were obtained from the state of California birth certificate data bases (Pediatrics Digest as cited in Lisa et al., 2005). Differences in categorical variables between case subjects and control subjects were compared with x2 statistics…Differences in continuous variables were compared with t tests…The risks of autism associated with maximum bilirubin levels above cutoff points were estimated as odd ratios and 95% confidence intervals with multivariate logistic regression analyses (Pediatrics Digest as cited in Lisa et al., 2005). Infants for whom no bilirubin measurements were made were assumed to have levels below the cut off points (Pediatrics Digest as cited Lisa et al., 2005). Maternal and infant characteristics associated with maximal bilirubin levels and infant case statuses were included as covariates in multivariate analyses…All study procedures were approved by KP Northern California Institutional Review Board and California State Committee for the Protection of Human Subjects (Pediatrics Digest as cited in Lisa et al., 2005). Results the first article Several studies reported no association although, some showed that jaundice occurred unusually among children later diagnosed with autism. The results were based on 338 case subjects an 1817 control subjects in the final study population (Pediatrics Digest as cited in Lisa et al., 2005). The gestational and birth weight characteristics distribution were similar between case and control subjects. Male subjects outnumbered the female subjects by a ratio of 4:1 in both cases. The mean age for delivery mothers was greater than that of control mothers. However, case mothers averaged more years of education. Bilirubin measurements (greater than 1) in the first 30 days of life showed similarities (case subjects=27.8% and control subjects= 27.5%) (Pediatrics Digest as cited in Lisa et al., 2005). Among those tested, there was no significant disparity in the mean number of total serum bilirubin tests per infant; the maximum bilirubin measured and the proportion of infants who received phototherapy. All 35 children were observed to check ASD cases. The first scenario showed 19 children (54%) had it (Diagnostics and statistical manual, had some detailed information abstracted) while the other scenario conducted by an expert review (included the abstracted information) showed 46% had ASD (Pediatrics Digest as cited in Lisa et al., 2005). Most previous studies on the association between jaundice and ASD included hyperbilirubinemia only as obstetric sub optimality, did not adequately define cut-off values of hyperbilirubinemia or relied on parental reports of jaundice (Pediatrics Digest as cited in Lisa et al., 2005). Conclusion for the first article There is no association between neonatal hyperbilirubinemia and ASD The second article employed systematic review using published guidelines of Meta-analyses of Observation Studies in Epidemiology (MOOSE) and the Assessment of Multiple Systematic Reviews (AMSTAR) (Pediatrics Digest as cited in Lisa et al., 2005). Only publications in English and those with an evaluation on neonatal jaundice and ASD were included. On the other hand, case series studies, case reports and multiple publications of a given material from the same author, were excluded The search strategy used was supposed to identify all published completed studies up to 2009 in Medline and Pub Med and references of published manuscripts. Index terms included premature infants, neonates, jaundice, hyperbilirubinemia, prenatal factor, neonatal factors, autism and autism spectrum disorder (Pediatrics Digest as cited in Lisa et al., 2005). Data was then collected from each study independently using predesigned form including author, publication year, study design, study matching criteria, gestational age subjects, number of subjects and data raw using 2 x 2 tables for neonatal jaundice and ASD. Meta-analysis was performed for the association between jaundice and ASD using Stata 10. The analysis produces forest plots as a schematic description of the Meta analysis results (Pediatrics Digest as cited in Lisa et al., 2005). Summery random effects were reported using pooled odds ratios QR and 95% confidence intervals (CI), and were calculated around each summer effect. The random effects model assumes that analyzed studies are a random sample of a hypothetical population under study. Random effects models are considered most appropriate when there is heterogeneity (p less than 0.050) (Pediatrics Digest as cited in Lisa et al., 2005). Each of these studies was examined for sources of clinical heterogeneity including study design, year of the study, population characteristics, procedures’ for ascertaining autism and jaundice and analysis of co-founders. Begs funnel plot and eggers test were performed to determine publication bias. A p value less than 0.005 was considered significant for publication bias (Pediatrics Digest as cited in Lisa et al., 2005). Results for the second article Meta analysis of two studies involving premature infants showed no association between neonatal jaundice and ASD in infants (OR 0.7, 95% CI 0.308-1.02) (Sangiv, Smith & Honhyue, 2011). However, some results appeared different especially that did not meet all selection parameters. Thirteen observational studies met the inclusion criteria. However, three studies that did not include basic information on the number of subjects with jaundice and/or autism; but included jaundice in the overall optimal risk scores were excluded (Sangiv, Smith & Honhyue, 2011). Some research results were published in both English and non English Manuscripts. In this case, the non English manuscripts were excluded. Among 13 included studies, there was a statistical association between jaundice and ASD. There was no evidence of population bias. There was significant heterogeneity (QR=1.43, 95% CI 1.22-1.67 and p=0.000) (Sangiv, Smith & Honhyue, 2011). Most studies were designed as retrospective case control studies. Earlier studies were small and used siblings as matched controls. Studies also differed in concentration and evaluated association between ASD and moderate to severe jaundice (Sangiv, Smith & Honhyue, 2011). Most studies that showed no association between jaundice and ASD included infants with a degree of jaundice. However, the studies that demonstrated an association between jaundice and ASD were larger and more recent (Sangiv, Smith & Honhyue, 2011). Most studies included both premature and term infants with jaundice. Both studies reported a positive association between jaundice and ASD only in term patients. However, some studies showed no association between the two conditions in premature infants (Sangiv, Smith & Honhyue, 2011). Conclusion for the second article There was no significant association between neonatal jaundice and ASD in premature infants but, the two conditions have an association in term infants. The difference can be explained by bilirubin metabolism and pathogenesis of bilirubin induced neurotoxicity (Sangiv, Smith & Honhyue, 2011). Bilirubin may be conjugated or not, and observational bias may explain the difference. Only the free unconjugated bilirubin is neuro- toxic. Therefore, a causal relationship between jaundice and ASD cannot be determined from this literature. Secondly, the majorities of these studies were explanatory and not hypothesis driven (Sangiv, Smith & Honhyue, 2011). The studies involved term and preterm infants but only two separate studies for premature infants. Fourth, studies that used siblings, for matching controls, used siblings of either gender, heterogeneous ages and unrelated to birth order in the family (Sangiv, Smith & Honhyue, 2011). Lastly different diagnostic criteria were used for the diagnosis of ASD and exceedingly few used standardized instruments for the measurement of ASD (Sangiv, Smith & Honhyue, 2011). The first article had more detailed information from the specimens collected that included gender, weight, gestational age, educational level, maternal race est. Thus, the findings can be generalized to the general population. The studies also produced accurate data. On the contrary, the findings in the second article cannot be generalized to the general population. Most contained observational biases. The first article has superior methodology mainly because the data collected was 95% correct with minimal biases, components that are key to obtaining precise results. Therefore, the first article is believable. The approach used in the first article seems to be ideal for examining this hypothesis. References Merrill, M. (2010). Introduction to epidemiology. Boston: Jones and Bartlett Publishers Lisa, A et al., 2011.Neonatal Hyperbilirubinemia and Risk of Autism Spectrum Disorders. Pediatrics Digest, 115, 2004-1870. Sanjiv, B, Smith, T and Hongyue, W. (2011). Is Neonatal Jaundice Associate with Autism Spectrum Disoders? A systematic Review. Journal of Autism Development Disorder, 41, 1555- 1463. Read More