Cholesterol Levels in Young Children with Autism and Typically Developing Controls

Background: High rates of autism are observed among patients with Smith Lemli Opitz Syndrome (SLOS), a rare genetic disorder with low cholesterol levels resulting from genetic defects in sterol synthesis (Tierney et al., 2001; Sikora et al., 2006).  To investigate whether SLOS might go unrecognized in autism, Tierney et al., (2006) examined a large group of children with ASD for biochemical evidence of SLOS by measuring total cholesterol levels. While no patients with SLOS were detected, a subset of patients was found to have low cholesterol levels, leading to the hypothesis that other sterol deficits might contribute to the ASD phenotype. 

Objectives: To determine if children with autism (AUT) have abnormally low cholesterol levels, in comparison with NHANES norms and a group of typically developing children (TYP). 

Methods: AUT and TYP subjects were participants in a phenomenological study examining subtypes of autism.  Autism is diagnosed by ADI-R, ADOS and clinical judgment.  Children with typical development have normal scores on cognitive, adaptive behavior and ADOS assessments.  Cholesterol levels were taken from a clinical laboratory panel obtained at baseline evaluation.  Normative cholesterol values were drawn from NHANES data according to age and gender. 

Results: Cholesterol measurements were available for 87 AUT (ages 2-8) and 35 TYP (ages 2-8).  Table 1 shows mean cholesterol, and distributions of high, normal and low levels of cholesterol for both groups.  In comparison with the NHANES norms, low levels were over-represented in both groups.  18.4% of AUT and 14% of the TYP sample had levels that were >1.65 SD (5^th centile) below the mean.  11.5 % of AUT and 5.7% of TYP had levels that were >2 SD below the mean (2^nd centile).  The between group differences were not significant by Fisher’s exact testing. 

	N	mean (g/dl)	High (>1.65 SD above the mean)	Normal	Low (<1.65 SD below the mean)	Very Low (<2 SD below the mean)	Lowest (>2.5 SD below the mean)
AUT	87	140.2	4 (4.6%)	67 (77%)	16 (18.4%)	10 (11.5%)	2 (2.3%)
TYP	35	134	0 (0%)	30 (86%)	5  (14%)	2 (5.7%)	0 (0%)

Conclusions: Low cholesterol levels were found more often than expected in both the TYP and AUT groups.  These results demonstrate the importance of having a local control group, as a comparison of the AUT results against the NHANES norms would have suggested a specific deficit in the affected individuals.   Since the TYP group also differed from the larger NHANES sample, the increased rates of low cholesterol levels may be due to other factors, such as lab methodology, family history, or nutritional intake.  Although no significant differences were seen in overall rates of abnormalities, it is interesting to note that 2 AUT children but none of the TYP children were in an “extremely low” group (>2.5 SD below the mean).  Larger samples are needed to further examine the relationship of very low cholesterol levels and autism.