Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Cytokine aberrations in autism spectrum disorder: a systematic review and meta-analysis



The role of non-diagnostic features in the pathophysiology of autism spectrum disorders (ASDs) is unclear. Increasing evidence suggests immune system alterations in ASD may be implicated in the severity of behavioral impairment and other developmental outcomes. The primary objective of this meta-analysis was to investigate if there is a characteristic abnormal cytokine profile in ASD compared with healthy controls (HCs). We identified relevant studies following a search of MEDLINE, EMBASE, PsycINFO, Web of Knowledge and Scopus. A meta-analysis was performed on studies comparing plasma and serum concentrations of cytokines in unmedicated participants with ASD and HCs. Results were reported according to PRISMA statement. Seventeen studies with a total sample size of 743 participants with ASD and 592 HC were included in the analysis. Nineteen cytokines were assessed. Concentrations of interleukin (IL)-1beta (P<0.001), IL-6 (P=0.03), IL-8 (P=0.04), interferon-gamma (P=0.02), eotaxin (P=0.01) and monocyte chemotactic protein-1 (P<0.05) were significantly higher in the participants with ASD compared with the HC group, while concentrations of transforming growth factor-β1 were significantly lower (P<0.001). There were no significant differences between ASD participants and controls for the other 12 cytokines analyzed. The findings of our meta-analysis identified significantly altered concentrations of cytokines in ASD compared to HCs, strengthening evidence of an abnormal cytokine profile in ASD where inflammatory signals dominate.


Autism spectrum disorder (ASD) is a complex, pervasive neurobiological disorder that is characterized by impairments in the domains of social communication and restricted interests and repetitive behaviors. Despite a variable onset in the manifestation of diagnostic features, the developmental trajectory is usually significantly altered. Heterogeneity in etiology, phenotype, outcome and comorbidities are hallmarks of ASD. The core diagnostic features have long been the focus of research and intervention strategies. However, in recent years, increasing attention has been given to concurrent non-diagnostic features and pathophysiology observed in ASD, including, sleep difficulties,1 immune system abnormalities,2 gastrointestinal disorders3 and mitochondrial dysfunction.4

Narrative reviews of immune system abnormalities and cytokine alterations in ASD have increased in recent years.5,6 Mechanisms leading to alterations in immune system functioning in ASD have been addressed from numerous perspectives, including genetic,7 maternal immune activation,8,9 autoimmunity10 and neuroimmune.11,12 Several studies have reported altered cytokine profiles in ASD compared with healthy controls (HCs).13, 14, 15 Furthermore, associations between the severity of diagnostic features and elevated cytokines in ASD have been identified.16,17

Cytokines are key modulators in the initiation (innate immune system) and maintenance (adaptive immune system) of immune responses, and also facilitate the exchange of information between cells of the immune system and other tissue cells. Cytokines regulate immune responses via pro-inflammatory and anti-inflammatory pathways. The impact of peripheral cytokines most likely extends across the blood–brain barrier (BBB) into the brain parenchyma via active protein transport mechanisms, and by recruitment of activated immune cells across the vascular endothelium.18, 19, 20 Cytokine expression has been identified in the brain tissue of patients with ASD, with microglial and astroglial activation in the cerebellum shown to be associated with an upregulation of cytokines indicating an innate immune response.11 In cerebrospinal fluid, cytokines involved in pro-inflammatory pathways, such as interferon-gamma (IFN-γ) and tumor necrosis factor-α have been reported at increased levels, supporting the hypothesis of a chronic state of cytokine induction in ASD.11,21 Case–control investigations of cytokine concentrations in plasma or serum also suggest several alterations compared with healthy subjects.14,22, 23, 24, 25

Associations between the severity of behavioral impairments identified as core diagnostic features of ASD and ex vivo cytokine levels have been reported.17,26,27 Immune activation in ASD has also been explored, with in vitro techniques using various mitogens, and various allergens as stimuli.13,28, 29, 30, 31, 32 Cytokine profiles have also been investigated in neonatal dried blood samples from a newborn screening biobank, with results suggesting depressed or hypoactive immune cell activity in the neonatal period in ASD.33 The impact of maternal immune activation has been extensively investigated in animal models, with levels of circulating cytokines influencing fetal brain development,9 CNS development34 and changes in the balance of cytokines inducing behavioral abnormalities in offspring.35

Various hypothesis have been proposed addressing the role of cytokines in the pathophysiology of ASD, with the more developed literature in the field of schizophrenia,36,37 bipolar disorder38 and major depression39 providing a framework. The objective of this systematic review is to investigate alterations in cytokine levels in the peripheral blood in patients with ASD compared to HC subjects. An association between cytokine aberrations and ASD has been reported in several studies,17,26,40 however the association is inconsistent between studies and for individual cytokines.14,41,42 We have used meta-analytical techniques to improve the strength of the accumulating evidence implicating immune system abnormalities in the pathophysiology of autism and to address the inconsistency in published data.

Materials and methods

The systematic review and meta-analysis were undertaken and reported in accordance with the PRISMA statement (Preferred Reporting Items for Systematic Reviews and Meta-Analysis).43

Inclusion and exclusion criteria

Studies comparing cytokines in participants with ASD and HC subjects were included. The inclusion criteria were (1) an ASD diagnosis according to criteria established in DSM-III or IV; (2) participants aged over 2 years old; (3) measurement of peripheral blood cytokines in plasma or serum; (4) concentrations of individual cytokines that were available in three or more studies, as per prior meta-analyses;38 and (5) studies which reported cytokine concentrations using a mean and standard deviation, or sample size and P-value. The exclusion criteria were studies which only included (1) medicated participants; (2) measurement of cytokines in cerebrospinal fluid or brain tissue; (3) in vitro analysis of cytokines after mitogen stimulation; (4) measurement of immune markers other than cytokines; and (5) evaluation of cytokine gene expression profiles.

Search strategy

A computerized search of MEDLINE, EMBASE, PsycINFO, Web of Knowledge and Scopus from January 1980 to August 2013 was conducted using the search strategy ‘(autism) and (cytokine or chemokine or interleukin or interferon or inflammation or tumor necrosis factor)’. Duplicates were deleted and the title and abstract of each article were scanned for relevance independently by two researchers. The full text of potentially relevant studies was then retrieved and assessed for eligibility considering established criteria detailed above. The reference lists of relevant studies were also searched for possible studies meeting criteria. Where there was initial disagreement discussion among researchers established universal agreement on studies to be included. A flowchart of information pertaining to identification, screening, eligibility and final studies included was constructed according to PRISMA guidelines43 (see Supplementary Figure 1).

Data extraction

Data for each individual cytokine assessed in eligible studies were extracted into an Excel spreadsheet (sample size, mean, standard deviation and P-value). When a P-value was reported as an inequality rather than an exact value, the value was rounded down to the nearest 5 thousandth to allow compatibility with meta-analysis software. Data for moderator analysis of gender, age, type of control group and blood fraction were also extracted (Supplementary Table 1). One author (AM) extracted all data. The method used to extract data was independently verified by another author (DSQ). Data were then re-checked for accuracy (AM).

Data analysis

Comprehensive Meta-Analysis Version 2 (Biostat Inc., Englewood, NJ, USA) was used for analysis.44 Effect sizes were calculated as Hedges’ g standardized mean difference of cytokine concentrations between ASD and HC groups using means and standard deviations or sample size and P-values. Hedges’ g offers the same interpretation as Cohen’s d, but incorporates a correction factor addressing the potential biases resulting from a small sample size as Cohen’s d45 tends to overestimate the absolute value of the standardized mean difference in small samples. An effect size estimate was calculated for each cytokine analyzed in all eligible studies. Classification of Hedges’ g effect size can be done in the same way as Cohen’s d: 0.2=small; 0.5=medium; and 0.8=large.45

A positive effect size implied that higher levels of the specified cytokine were reported in the ASD group as compared with the HC group. Effect size estimates were calculated using a random-effects model, which estimates a mean of a distribution of effects. Under this model, allocation of study weights is based on the inverse of the total variance, which includes both within-study variance and between-study variance. This method results in a more conservative approach, as the model yields a wider confidence interval when there is significant heterogeneity among the results of eligible studies.46 The alternative, a fixed-effects model, estimates a common effect and allocates weight based on sample size.

Sensitivity analysis was undertaken using forest plots to initially identify outliers. The impact of any outliers was addressed by removing the study reporting the outlier result and comparing the effect size and P-value of the subsequent result. Publication bias was assessed using a P-value obtained for Egger’s test of funnel plot asymmetry.47 Heterogeneity was assessed using Cochrane’s Q and I2 was used to evaluate the impact of heterogeneity.48 The level of significance for the Cochrane Q test was set at 0.10 due to the low power of the test to detect heterogeneity.49 An I2 of 0.25, 0.50 and 0.75 denoted small, moderate and high levels of heterogeneity, respectively.48 When data permitted, moderator analysis was undertaken in an attempt to identify sources of heterogeneity.


A total of 2940 studies were identified from the initial search, after excluding duplicates. Titles and abstracts were scanned resulting in identification of 46 articles to be assessed for eligibility. Several studies were excluded as they analyzed cytokines in cerebrospinal fluid or brain tissue (5 studies); reported stimulated or unstimulated levels of cytokines in vitro (17 studies); included participants who were medicated (1 study); the cytokine analyzed was not assessed in at least two other studies (5 studies) and/or insufficient information was provided on the method used to process blood samples (1 study). Seventeen studies, with a total of 1341 participants (746 with ASD and 595 HC), met inclusion and exclusion criteria. There were a total of 1113 (83%) males and 228 (17%) females. In the ASD group, 634 (85%) participants were male, while in the HC group there were 479 (81%) males. Meta-analysis was performed on data for 19 different cytokines, chemokines and cytokine receptors: interleukin (IL)-4, IL-6, IL-8, IL-10, IL-12p40, IL-17, IL-23, IL-1α, IL-1β, receptor antagonist IL-1RA, IFN-γ, granulocyte colony-stimulating factor, transforming growth factor-beta1 (TGF-β1), tumor necrosis factor-α, eotaxin (also known as CCL11), monocyte chemotactic protein-1 (MCP-1; CCL2), macrophage inflammatory proteins MIP-1α (CCL3) and MIP-1β (CCL4) and RANTES (CCL5). Fifty five other cytokines were unable to be included in the meta-analysis because of either having only been assessed less than three times, or if assessed three or more times, inadequate information was provided in one or more papers (Supplementary Table 2). Table 1 presents characteristics for the groups of studies measuring each cytokine and the effect size estimates for each individual cytokine. Five studies assessed cytokine concentrations in serum,13,14,22,50,51 and twelve studies assessed cytokine concentrations in plasma.15,16,17,24, 25, 26,41,42,52, 53, 54, 55

Table 1 Characteristics and effect size estimates for cytokine concentrations in autism spectrum disorder compared with healthy controls

Meta-analysis on all cytokines resulted in significantly higher concentrations of cytokines in ASD compared to HC (Hedges’ g=0.216, 95% confidence interval 0.086 to 0.346, P=0.001). Heterogeneity was significant (Q=388, P<0.001). For individual cytokine concentrations in peripheral blood, a large and highly significant effect size estimate was associated with decreased concentrations of TGF-β1, while significantly higher concentrations of IL-1β in participants with ASD was associated with a medium effect size (see Table 1). Small-to-medium effect size estimates were identified for concentrations of IL-6, IL-8, eotaxin and MCP-1. Concentrations of IL-1RA were bordering on significance in participants with ASD, a result associated with a medium effect size. There was a large and highly significant effect size for IFN-γ in participants with ASD compared with HC (see Table 1). One outlier study,55 caused significant heterogeneity and resulted in publication bias. Removal of this study from the meta-analysis substantially reduced the effect size, however the results still showed significantly higher concentrations of IFN-γ in participants with ASD compared to HC (Hedges’ g=0.489, 95% confidence interval 0.035–0.943, P=0.035), and this finding was associated with a medium effect size. Heterogeneity was reduced by over 20% (I2 =69). Concentrations of cytokines IL-1α, IL-4, IL-10, IL-12p40, IL-17, IL-23, granulocyte colony-stimulating factor, RANTES, MIP-1α and MIP-1β were all extracted from three studies and yielded nonsignificant effect size estimates. Six studies measuring concentrations of tumor necrosis factor-α also yielded nonsignificant effect size estimates.

Methodological and clinical heterogeneity

Heterogeneity reached significance (P<0.10) for 13 out of the 19 cytokines (Table 1). Data for measurements on the remaining six cytokines was extracted from only three studies. A possible explanation for this apparent homogeneity is the low power the test for heterogeneity has in meta-analyses with a small number of studies.49

Methodological sources of heterogeneity include the timing of blood sampling (reflecting circadian variations) and a fasting or non-fasting state, the time interval between blood collection and separation and storage, the number of freeze-thaw cycles of a sample prior to analysis and the different assay procedures used for analysis. Furthermore, methodological differences in assay procedures and operator reliability are all possible sources of heterogeneity. A substantial lack of information regarding these parameters, and low study numbers for individual cytokines prevented a moderator analysis using any of these variables. A moderator analysis was conducted to assess for potential differences between the level of cytokines in two different blood components, plasma and serum. Subgroups for meta-analysis were formed using data from 73 different cytokine measurements. Eleven of these measurements were performed on sera, while the remainder were performed on plasma samples (85%). Overall results remained significant for both components (plasma: Hedges’ g=0.148, P=0.02; serum: Hedges’ g=0.596, P=0.007).

Possible sources of clinical heterogeneity include age, gender and type of control group (that is, siblings vs non-siblings), intercurrent infection or other inflammatory illness, and previous or unreported medication effects. Extensive overlap between age groupings did not allow for evaluation of the impact of age on effect size and heterogeneity. With the exception of one small study where mean age was approximately 23 years old,22 the age range of participants was between 2 years old and 18 years old (Supplementary Table 1). Some studies excluded children older than 10 years of age to avoid potential effects associated with fluctuations in hormonal changes during puberty on cytokine concentrations. Gender subgroups were either classified as all male (52% of individual cytokines analyzed) or mixed (48% of individual cytokines analyzed), as no study separated out results for female participants. A moderator analysis was performed to investigate the impact of gender on overall results. Results for the male subgroup remained significant (Hedges’ g=0.259, P=0.005), however the effect size for the mixed gender group was reduced when compared with the overall results (Hedges’ g=0.174, P=0.07). Heterogeneity remained significant for both groups (male: Q=147, P<0.001; mixed: Q=236, P<0.001). Despite the differences between the male and mixed subgroups, effect sizes were similar enough to support collapsing the gender groupings for analysis of effect sizes of individual cytokine concentrations.

Two studies included in the meta-analysis used siblings as HCs. One study used related siblings as HCs,41 while the other study used unrelated siblings as the control group.53 In this study all of the HCs had a sibling with an ASD diagnosis, but were unrelated to the children with ASD selected for the study. A meta-analysis was conducted to assess the impact the removal of these studies had on effect sizes for individual cytokines and on overall results (Figure 1). Effect sizes for individual cytokines all increased, as did the overall result (Hedges’ g=0.370, 95% confidence interval 0.206–0.534, P<0.001).

Figure 1

Forest plot of meta-analysis excluding studies using related and unrelated siblings as control groups. The arrow representing the study of El-Ansary and Al-Ayadhi55 shows the hedges' g and confidence interval (CI) spread further than the limit for the effect size range (−4.00 to 4.00).

PowerPoint slide


This meta-analysis was undertaken to investigate alterations in peripheral blood cytokine concentrations in ASD compared to HC subjects, and to explore if there is a characteristic abnormal cytokine profile in ASD. A combined estimate of effect size indicates that there are peripheral alterations in cytokine levels in ASD, with a small effect size. There were elevations in the pro-inflammatory cytokines, IFN-γ, IL-1β and IL-6, and reduced concentrations of the anti-inflammatory cytokine, TGF-β1. Levels of several chemokines associated with recruitment of inflammatory cells, including eotaxin, IL-8 and MCP-1, were elevated. Effect sizes associated with the results for these individual cytokines varied from small to large. These results provide evidence of a heightened inflammatory state and altered cytokine profile, suggesting immune system dysfunction in ASD.

Effect sizes for IFN-γ and TGF-β1 were the largest of the 19 cytokines analyzed, with concentrations in ASD significantly elevated for IFN-γ, and concentrations of TGF-β1 significantly reduced. The roles of IFN-γ are diverse and multi-faceted. IFN-γ amplifies immune system sensitivity and response to pathogens, and coordinates the transition from innate immunity to adaptive immunity.56 IFN-γ is the prototypic T-cell-derived cytokine,57 and through promotion of macrophage activation and cytotoxicity skews the immune response toward a so-called T-helper-1 (Th1) phenotype generally intended to foster immune responses against intracellular pathogens, but also recognized in many auto-immune conditions. IFN-γ also inhibits proliferation of Th2 cells and subsequent production of anti-inflammatory cytokines, including IL-4, IL-5 and IL-10—all of which yielded nonsignificant results in this study. TGF-β1 also performs many cellular functions, including promoting cell growth, proliferation and differentiation, and is a potent immunosuppressant essential for immune homeostasis.22,26,58 IL-1β, a potent pro-inflammatory cytokine, is found in many settings of acute and chronic inflammation, and is known to affect the hypothalamic–pituitary–adrenal axis.59 IL-6, another pro-inflammatory cytokine, also induces production of acute phase proteins and stimulates B-cell differentiation and antibody production.60 Apart from direct actions on immune functions, IL-6 influences many hematologic, hepatic, endocrine and metabolic functions.61 Secretion of IL-6 has previously been reported to be enhanced in those with disruption of the sleep-wake cycle and changes in circulation IL-6 can cause changes in sleep quality.62,63 IL-6 is an important facilitator of central nervous system–immune communication.62 Specifically in ASD, elevated brain levels of IL-6 have been shown to impair neural cell adhesion and migration and is thought to impact synapse formation.12 The three chemokines that were elevated relative to HC, IL-8, eotaxin and MCP-1 are involved in the recruitment of neutrophils, eosinophils, monocytes and T cells from the circulation into settings of tissue inflammation.

This meta-analysis combined results from studies assessing single measures of cytokines in plasma and serum samples. While the meta-analysis technique attempts to adjust for methodological confounders, the procedure is limited by the number of available studies, the number of participants in each study and the information that can be extracted from the included studies, therefore cautious interpretation of the results of this meta-analysis is warranted. There are numerous methodological issues associated with specimen collection, handling and cytokine measurement,64, 65, 66, 67 as well as clinical confounders that should be considered in future research. The former technical concerns include assay reliability, particularly at the low end of the detection range, and variability in results using kits from different assay manufacturers.68 Issues following venipuncture procedure also require consideration, including artifactual ex vivo increases in cytokine levels by release from leukocytes,69 and ex vivo degradation by plasma proteases.70 The effects of anti-coagulants,71 freeze-thaw effects71 and potential for degradation following long-term storage66 all may potentially confound results. Given the clinical heterogeneity inherent in ASD, differences in patient characteristics, including the degree of severity of diagnostic features, should be considered to ensure methodological issues are not confounding results. Biological issues such as circadian rhythm72 and the menstrual cycle, and the effects of high or low body mass index, medication, smoking and alcohol can all influence cytokine concentrations.73 Data relating to these factors should be collected, controlled for and considered when establishing protocols investigating markers of inflammation such as cytokine concentrations.

Future research protocols investigating the relationship between behavioral factors in ASD and inflammation should assess and control for the impact of bio-behavioral factors on peripheral markers of inflammation, in order to facilitate a more robust interpretation of results.73 Longitudinal and concurrent measurement of peripheral blood and cerebrospinal fluid inflammatory markers could also assist with elucidation of the relationship between cytokine levels and pathological effect, and the identification of a characteristic cytokine profile in ASD. Further, larger studies are needed to identify the impact of gender.74 The influence of two studies that used siblings as healthy controls was of particular interest.41,53 Excluding results for cytokines analyzed in these studies lead to an increase in effect size overall and for individual cytokines, skewing an imbalanced immune system even more toward an inflammatory state. This corroborates with previous findings that immune profiles of children with ASD are more similar to siblings without a diagnosis of ASD than to HCs.31,41 These results suggest that siblings should not be employed as HCs, or additional unrelated HC subjects should also be included. Instead, evaluation of immune function in unaffected siblings may provide information that guides investigation into the causal pathway leading to immune system perturbation in ASD.75

To our knowledge, no previous studies have synthesized data on cytokine alterations in ASD. Effect size estimates for seven cytokines were significant (TGF-β1, IFN-γ, IL-1β, IL-8, IL-6, eotaxin and MCP-1). Based on these results we propose there is a pathophysiology in ASD that either manifests as an inflammatory disease process, similar to that of autoimmunity, or as a neuroimmune epiphenomenon that changes susceptibility to factors, such as pathogens and allergens, that trigger inflammatory responses and subsequent abnormalities in the immune system. Evidence of immune system perturbations in ASD, specifically a pro-inflammatory state, has been strengthened by this meta-analysis. Associations between immune system abnormalities and the severity of aberrant behavior in ASD,52 however, could not be investigated in this meta-analysis due to insufficient data. Interestingly, efficacious use of a non-steroidal anti-inflammatory drug as an adjunctive intervention in reducing the severity of diagnostic features76 supports our findings and potential role of the immune system in the manifestation of diagnostic features in ASD. Important considerations to ensure biological and methodological issues do not interfere with the strength of future findings are highlighted by this meta-analysis. A better understanding of the inflammatory biology of ASD and possible associations with behavioral impairments and non-diagnostic features warrants further investigation and may have significant therapeutic implications.


  1. 1

    Goldman SE, Surdyka K, Cuevas R, Adkins K, Wang L, Malow BA . Defining the sleep phenotype in children with autism. Dev Neuropsychol 2009; 34: 560–573.

    Article  Google Scholar 

  2. 2

    Careaga M, Ashwood P . Autism spectrum disorders: from immunity to behavior. Methods Mol Biol 2012; 934: 219–240.

    Article  Google Scholar 

  3. 3

    Valicenti-Mcdermott M, McVICAR K, Rapin I, Wershil BK, Cohen H, Shinnar S . Frequency of gastrointestinal symptoms in children with autistic spectrum disorders and association with family history of autoimmune disease. J Dev Behav Pediatr 2006; 27: S128–S136.

    Article  Google Scholar 

  4. 4

    Rossignol DA, Frye RE . Mitochondrial dysfunction in autism spectrum disorders: a systematic review and meta-analysis. Mol Psychiatry 2011; 17: 290–314.

    Article  Google Scholar 

  5. 5

    Ashwood P, Wills S, Van de Water J . The immune response in autism: a new frontier for autism research. J Leukoc Biol 2006; 80: 1–15.

    CAS  Article  Google Scholar 

  6. 6

    Goines PE, Ashwood P . Cytokine dysregulation in autism spectrum disorders (ASD): possible role of the environment. Neurotoxicol Teratol 2012; 36: 67–81.

    Article  Google Scholar 

  7. 7

    Crespi BJ, Thiselton DL . Comparative immunogenetics of autism and schizophrenia. Genes Brain Behav 2011; 10: 689–701.

    CAS  Article  Google Scholar 

  8. 8

    Garay PA, Hsiao EY, Patterson PH, Kimberley McAllister A . Maternal immune activation causes age-and region-specific changes in brain cytokines in offspring throughout development. Brain Behavior Immunity 2012; 31: 54–68.

    Article  Google Scholar 

  9. 9

    Smith SEP, Li J, Garbett K, Mirnics K, Patterson PH . Maternal immune activation alters fetal brain development through interleukin-6. J Neurosci 2007; 27: 10695–10702.

    CAS  Article  Google Scholar 

  10. 10

    Atladóttir HÓ, Pedersen MG, Thorsen P, Mortensen PB, Deleuran B, Eaton WW et al. Association of family history of autoimmune diseases and autism spectrum disorders. Pediatrics 2009; 124: 687–694.

    Article  Google Scholar 

  11. 11

    Vargas DL, Nascimbene C, Krishnan C, Zimmerman AW, Pardo CA . Neuroglial activation and neuroinflammation in the brain of patients with autism. Ann Neurol 2005; 57: 67–81.

    CAS  Article  Google Scholar 

  12. 12

    Wei H, Zou H, Sheikh AM, Malik M, Dobkin C, Brown WT et al. IL-6 is increased in the cerebellum of autistic brain and alters neural cell adhesion, migration and synaptic formation. J Neuroinflam 2011; 8: 52.

    CAS  Article  Google Scholar 

  13. 13

    Croonenberghs J, Bosmans E, Deboutte D, Kenis G, Maes M . Activation of the inflammatory response system in autism. Neuropsychobiology 2002; 45: 1–6.

    CAS  Article  Google Scholar 

  14. 14

    Ricci S, Businaro R, Ippoliti F, Vasco Lo VR, Massoni F, Onofri E et al. Altered cytokine and BDNF levels in autism spectrum disorder. Neurotox Res 2013; 24: 491–501.

    CAS  Article  Google Scholar 

  15. 15

    Suzuki K, Matsuzaki H, Iwata K, Kameno Y, Shimmura C, Kawai S et al. Plasma cytokine profiles in subjects with high-functioning autism spectrum disorders. PLoS ONE 2011; 6: e20470.

    CAS  Article  Google Scholar 

  16. 16

    Ashwood P, Krakowiak P, Hertz-Picciotto I, Hansen R, Pessah IN, Van de Water J . Associations of impaired behaviors with elevated plasma chemokines in autism spectrum disorders. J Neuroimmunol 2011a; 232: 196–199.

    CAS  Article  Google Scholar 

  17. 17

    Hashim H, Abdelrahman H, Mohammed D, Karam R . Association between plasma levels of transforming growth factor-b1, IL-23 and IL-17 and the severity of autism in Egyptian children. Res Autism Spect Dis 2013; 7: 199–204.

    Article  Google Scholar 

  18. 18

    Banks WA . Blood-brain barrier transport of cytokines: a mechanism for neuropathology. Curr Pharm Design 2005; 11: 973–984.

    CAS  Article  Google Scholar 

  19. 19

    Fung A, Vizcaychipi M, Lloyd D, Wan Y, Ma D . Central nervous system inflammation in disease related conditions: mechanistic prospects. Brain Res 2012; 1446: 144–155.

    CAS  Article  Google Scholar 

  20. 20

    Erickson MA, Dohi K, Banks WA . Neuroinflammation: a common pathway in CNS diseases as mMediated at the blood-brain barrier. Neuroimmunomodulation 2012; 19: 121–130.

    CAS  Article  Google Scholar 

  21. 21

    Chez MG, Dowling T, Patel PB, Khanna P, Kominsky M . Elevation of tumor necrosis factor-alpha in cerebrospinal fluid of autistic children. Pediatr Neurol 2007; 36: 361–365.

    Article  Google Scholar 

  22. 22

    Okada K, Hashimoto K, Iwata Y, Nakamura K, Tsujii M, Tsuchiya KJ et al. Decreased serum levels of transforming growth factor-(beta)1 in patients with autism. Prog Neuro-Psychopharmacol Biol Psychiatry 2007; 31: 187–190.

    CAS  Article  Google Scholar 

  23. 23

    Molloy CA, Morrow AL, Meinzen-Derr J, Schleifer K, Dienger K, Manning-Courtney P et al. Elevated cytokine levels in children with autism spectrum disorder. J Neuroimmunol 2006; 172: 198–205.

    CAS  Article  Google Scholar 

  24. 24

    Enstrom A, Onore C, Hertz-Picciotto I, Hansen R, Croen L, Van de Water J et al. Detection of IL-17 and IL-23 in plasma samples of children with autism. Am J Biochem Biotechnol 2008; 4: 114–120.

    CAS  Article  Google Scholar 

  25. 25

    Singh VK . Plasma increase of interleukin- 12 and interferon-gamma pathological significance in autism. J Neuroimmunol 1996; 66: 143–145.

    CAS  Article  Google Scholar 

  26. 26

    Ashwood P, Enstrom A, Krakowiak P, Hertz-Picciotto I, Hansen R, Croen LA et al. Decreased transforming growth factor beta1 in autism: a potential link between immune dysregulation and impairment in clinical behavioral outcomes. J Neuroimmunol 2008; 204: 149–153.

    CAS  Article  Google Scholar 

  27. 27

    Al-Ayadhi LY, Mostafa GA . Elevated serum levels of interleukin-17 A in children with autism. J Neuroinflamm 2012; 9: 158.

    CAS  Google Scholar 

  28. 28

    Jyonouchi H, Geng L, Ruby A, Reddy C, Zimmerman-Bier B . Evaluation of an association between gastrointestinal symptoms and cytokine production against common dietary proteins in children with autism spectrum disorders. J Pediatr 2005; 146: 605–610.

    CAS  Article  Google Scholar 

  29. 29

    Ashwood P, Krakowiak P, Hertz-Picciotto I, Hansen R, Pessah IN, Van de Water J . Altered T cell responses in children with autism. Brain Behav Immun 2011; 25: 840–849.

    CAS  Article  Google Scholar 

  30. 30

    Malik M, Sheikh AM, Wen G, Spivack W, Brown WT, Li X . Expression of inflammatory cytokines, Bc12 and cathepsin D are altered in lymphoblasts of autistic subjects. Immunobiology 2010; 216: 80–85.

    Article  Google Scholar 

  31. 31

    Saresella M, Marventano I, Guerini FR, Mancuso R, Ceresa L, Zanzottera M et al. An autistic endophenotype results in complex immune dysfunction in healthy siblings of autistic children. Biol Psychiatry 2009; 66: 978–984.

    Article  Google Scholar 

  32. 32

    Vojdani A, Mumper E, Granpeesheh D, Mielke L, Traver D, Bock K et al. Low natural killer cell cytotoxic activity in autism: the role of glutathione, IL-2 and IL-15. J Neuroimmunol 2008; 205: 148–154.

    CAS  Article  Google Scholar 

  33. 33

    Abdallah MW, Larsen N, Mortensen EL, Atladottir HO, Norgaard-Pedersen B, Bonefeld-Jorgensen EC et al. Neonatal levels of cytokines and risk of autism spectrum disorders: An exploratory register-based historic birth cohort study utilizing the Danish Newborn Screening Biobank. J Neuroimmunol 2012; 252: 75–82.

    CAS  Article  Google Scholar 

  34. 34

    Deverman BE, Patterson PH . Cytokines and CNS development. Neuron 2009; 64: 61–78.

    CAS  Article  Google Scholar 

  35. 35

    Hsiao EY, McBride SW, Chow J, Mazmanian SK, Patterson PH . Modeling an autism risk factor in mice leads to permanent immune dysregulation. Proc Natl Acad Sci 2012; 109: 12776–12781.

    CAS  Article  Google Scholar 

  36. 36

    Miller BJ, Buckley P, Seabolt W, Mellor A, Kirkpatrick B . Meta-analysis of cytokine alterations in schizophrenia: clinical status and antipsychotic effects. Bioi Psychiatry 2011; 70: 663–671.

    CAS  Article  Google Scholar 

  37. 37

    Potvin S, Stip E, Sepehry AA, Gendron A, Bah R, Kouassi E . Inflammatorycytokine alterations in schizophrenia: a systematic quantitative review. Biol Psychiatry 2008; 63: 801–808.

    CAS  Article  Google Scholar 

  38. 38

    Modabbernia A, Taslimi S, Brietzke E, Ashrafi M . Cytokine alterations in bipolar disorder: a meta-analysis of 30 studies. Biol Psychiatry 2013; 74: 15–25.

    CAS  Article  Google Scholar 

  39. 39

    Dowlati Y, Herrmann N, Swardfager W, Liu H, Sham L, Reim EK et al. A meta-analysis of cytokines in major depression. Biol Psychiatry 2010; 67: 446–457.

    CAS  Article  Google Scholar 

  40. 40

    Onore C, Enstrom A, Krakowiak P, Hertz-Picciotto I, Hansen R, Van de Water J et al. Decreased cellular IL-23 but not IL-17 production in children with autism spectrum disorders. J Neuroimmunol 2009; 216: 126–129.

    CAS  Article  Google Scholar 

  41. 41

    Napolioni V, Ober-Reynolds B, Szelinger S, Corneveaux JJ, Pawlowski T, Ober-Reynolds S et al. Plasma cytokine profiling in sibling pairs discordant for autism spectrum disorder. J Neuroinflam 2013; 10: 38.

    CAS  Article  Google Scholar 

  42. 42

    Tostes MHFS, Teixeira HC, Gattaz WF, Brandao MAF, Raposo NRB . Altered neurotrophin, neuropeptide, cytokines and nitric oxide levels in autism. Pharmacopsychiatry 2012; 45: 241–243.

    CAS  Article  Google Scholar 

  43. 43

    Moher D, Liberati A, Tetzlaff J, Altman DG . Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Ann Intern Med 2009; 151: 264–269.

    Article  Google Scholar 

  44. 44

    Borenstein M, Hedges L, Higgins J, Rothstein H . Comprehensive meta-analysis version 2. Biostat: Englewood, NJ, 2005.

    Google Scholar 

  45. 45

    Cohen J . Statistical Power Analysis for the Behavioral Sciences. Lawrence Erlbaum Associates: Hillsdale, NJ, 1988.

    Google Scholar 

  46. 46

    Hedges L, Olkin I . Statistical Methods for Meta-Analysis. Academic Press: Orlando, 1985.

    Google Scholar 

  47. 47

    Egger M, Smith GD, Schneider M, Minder C . Bias in meta-analysis detected by a simple, graphical test. Br Med J 1997; 315: 629–634.

    CAS  Article  Google Scholar 

  48. 48

    Higgins JPT, Thompson SG . Quantifying heterogeneity in a meta-analysis. Statist Med 2002; 21: 1539–1558.

    Article  Google Scholar 

  49. 49

    Higgins JPT, Thompson SG, Deeks JJ, Altman DG . Measuring inconsistency in meta-analyses. Br Med J 2003; 327: 557–560.

    Article  Google Scholar 

  50. 50

    Al-Ayadhi LY . Pro-inflammatory cytokines in autistic children in central Saudi Arabia. Neurosciences 2005; 10: 155–158.

    PubMed  Google Scholar 

  51. 51

    Wakkad ASEDE, Saleh MT . The proinflammatory cytokines in children with autism. Pak J Biol Sci 2006; 9: 2593–2599.

    Article  Google Scholar 

  52. 52

    Ashwood P, Krakowiak P, Hertz-Picciotto I, Hansen R, PESSAH I, Van de Water J . Elevated plasma cytokines in autism spectrum disorders provide evidence of immune dysfunction and are associated with impaired behavioral outcome. Brain Behav Immun 2011b; 25: 40–45.

    CAS  Article  Google Scholar 

  53. 53

    Manzardo AM, Henkhaus R, Dhillon S, Butler MG . Plasma cytokine levels in children with autistic disorder and unrelated siblings. Int J Dev Neurosci 2012; 30: 121–127.

    CAS  Article  Google Scholar 

  54. 54

    Sweeten TL, Posey DJ, Shankar S, McDougle CJ . High nitric oxide production in autistic disorder: a possible role for interferon-γ. Biol Psychiatry 2004; 55: 434–437.

    CAS  Article  Google Scholar 

  55. 55

    El-Ansary A, Al-Ayadhi L . Neuroinflammation in autism spectrum disorders. J Neuroinflam 2012; 9: 1–9.

    Google Scholar 

  56. 56

    Schroder K, Hertzog PJ, Ravasi T, Hume DA . Interferon-γ: an overview of signals, mechanisms and functions. J Leukoc Biol 2004; 75: 163–189.

    CAS  Article  Google Scholar 

  57. 57

    Berger A . Science commentary: Th1 and Th2 responses: what are they? Br Med J 2000; 321: 424–424.

    CAS  Article  Google Scholar 

  58. 58

    Gomes FCA, Sousa V de O, Romão L . Emerging roles for TGF-β1 in nervous system development. Int J Dev Neurosci 2005; 23: 413–424.

    CAS  Article  Google Scholar 

  59. 59

    Maes M, Bosmans E, Meltzer HY, Scharpé S, Suy E . Interleukin-1 beta: a putative mediator of HPA axis hyperactivity in major depression? Am J Psychiatry 1993; 150: 1189–1193.

    CAS  Article  Google Scholar 

  60. 60

    Nishimoto N, Kishimoto T . Interleukin 6: from bench to bedside. Nat Clin Pract Rheumatol 2006; 2: 619–626.

    CAS  Article  Google Scholar 

  61. 61

    Papanicolaou DA . The Pathophysiologic roles of interleukin-6 in human disease. Ann Intern Med 1998; 128: 127.

    CAS  Article  Google Scholar 

  62. 62

    Rohleder N, Aringer M, Boentert M . Role of interleukin‐6 in stress, sleep, and fatigue. Ann New York Acad Sci 2012; 1261: 88–96.

    CAS  Article  Google Scholar 

  63. 63

    Vgontzas AN, Papanicolaou DA, Bixler EO, Lotsikas A, Zachman K, Kales A et al. Circadian interleukin-6 secretion and quantity and depth of sleep. J Clin Endo 84: 203–2607.

  64. 64

    Skogstrand K, Thorsen P, Norgaard-Pedersen B, Schendel DE, Sorensen LC, Hougaard DM . Simultaneous measurement of 25 inflammatory markers and neurotrophins in neonatal dried blood spots by immunoassay with xMAP technology. Clin Chem 2005; 51: 1854–1866.

    CAS  Article  Google Scholar 

  65. 65

    Skogstrand K, Ekelund CK, Thorsen P, Vogel I, Jacobsson B, Nørgaard-Pedersen B et al. Effects of blood sample handling procedures on measurable inflammatory markers in plasma, serum and dried blood spot samples. J Immunol Methods 2008; 336: 78–84.

    CAS  Article  Google Scholar 

  66. 66

    Keustermans GCE, Hoeks SBE, Meerding JM, Prakken BJ, de Jager W . Cytokine assays: An assessment of the preparation and treatment of blood and tissue samples. Methods 2013; 61: 10–17.

    CAS  Article  Google Scholar 

  67. 67

    de Jager W, Prakken BJ, Bijlsma JWJ, Kuis W, Rijkers GT . Improved multiplex immunoassay performance in human plasma and synovial fluid following removal of interfering heterophilic antibodies. J Immunol Methods 2005; 300: 124–135.

    CAS  Article  Google Scholar 

  68. 68

    Breen EC, Reynolds SM, Cox C, Jacobson LP, Magpantay L, Mulder CB et al. Multisite comparison of high-sensitvity multiplex cytokine assays. Clin vaccine immunol 2011; 18: 1229–1242.

    CAS  Article  Google Scholar 

  69. 69

    McFaul SJ, Bowman PD, Villa VM . Hemoglobin stimulates the release of proinflammatory cytokines from leukocytes in whole blood. J Lab Clin Med 2000; 135: 263–269.

    CAS  Article  Google Scholar 

  70. 70

    Niwa Y, Akamatsu H, Sumi H, Ozaki Y, Abe A . Evidence for degradation of cytokines in the serum of patients with atopic dermatitis by calcium-dependent protease. Arch Dermatol Res 2000; 292: 391–396.

    CAS  Article  Google Scholar 

  71. 71

    de Jager W, Bourcier K, Rijkers GT, Prakken BJ, Seyfert-Margolis V . Prerequisites for cytokine measurements in clinical trials with multiplex immunoassays. BMC immunol 2009; 10: 52.

    Article  Google Scholar 

  72. 72

    Frey DJ, Fleshner M, Wright KP . The effects of 40 h of total sleep deprivation on inflammatory markers in healthy young adults. Brain Behav Immun 2007; 21: 1050–1057.

    CAS  Article  Google Scholar 

  73. 73

    O’Connor M-F, Bower JE, Cho HJ, Creswell JD, Dimitrov S, Hamby ME et al. To assess, to control, to exclude: effects of biobehavioral factors on circulating inflammatory markers. Brain Behav Immun 2009; 23: 887–897.

    Article  Google Scholar 

  74. 74

    Casimir GJA, Mulier S, Hanssens L, Zylberberg K, Duchateau J . Gender differences in inflammatory markers in children. Shock 2010; 33: 258–262.

    CAS  Article  Google Scholar 

  75. 75

    Goines P, Van de Water J . The immune system's role in the biology of autism. Curr Opin Neurol 2010; 23: 111–117.

    Article  Google Scholar 

  76. 76

    Asadabadi M, Mohammadi MR, Ghanizadeh A, Modabbernia A, Ashrafi M, Hassanzadeh E et al. Celecoxib as adjunctive treatment to risperidone in children with autistic disorder: a randomized, double-blind, placebo-controlled trial. Psychopharmacology 2013; 225: 51–59.

    CAS  Article  Google Scholar 

Download references


We thank Melanie Reuss for assisting with the independent screening of titles, abstracts and full-text articles and subsequent assessment of eligibility. We acknowledge the support of National Health and Medical Research Council project grant (1043664) and Career Development Fellowship (1061922) to Associate Professor Guastella and an Australia Fellowship (494914) to Professor Hickie.

Author information



Corresponding author

Correspondence to A J Guastella.

Ethics declarations

Competing interests

Ian B Hickie is a member of the Medical Advisory Panel for BUPA Health Insurance (Australia) and also a Board Member of Psychosis Australia Trust. From 2012, he is a Commissioner in Australia’s new National Mental Health Commission. He was until January 2012 a director of headspace: the national youth mental health foundation. Professor Hickie was previously the chief executive officer (till 2003) and clinical adviser (till 2006) of beyondblue, an Australian National Depression Initiative. He is supported principally for clinical research in depression and health services and population health initiatives related to anxiety and depression by an NHMRC Australian Medical Research Fellowship (2007–2012). He has led projects for health professionals and the community supported by governmental, community agency and pharmaceutical industry partners (Wyeth, Eli Lily, Servier, Pfizer, AstraZeneca) for the identification and management of depression and anxiety. He has received honoraria for presentations of his own work at educational seminars supported by a number of non-government organizations and the pharmaceutical industry (including Pfizer, Servier and AstraZeneca). He has served on advisory boards convened by the pharmaceutical industry in relation to specific antidepressants, including nefazodone, duloxetine and desvenlafaxine. He leads a new investigator-initiated study of the effects of agomelatine on circadian parameters (supported in part by Servier but also by other NHMRC funding) and has participated in a multicentre clinical trial of the effects of agomelatine on sleep architecture in depression and a Servier-supported study of major depression and sleep disturbance in primary care settings. In addition to national and international government-based grant bodies, investigator-initiated mental health research at the BMRI, he has been supported by various pharmaceutical manufacturers (including Servier and Pfizer) and not-for-profit entities (including the Heart Foundation, beyondblue and the BUPA Foundation).

Additional information

Supplementary Information accompanies the paper on the Molecular Psychiatry website

Supplementary information

PowerPoint slides

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Masi, A., Quintana, D., Glozier, N. et al. Cytokine aberrations in autism spectrum disorder: a systematic review and meta-analysis. Mol Psychiatry 20, 440–446 (2015).

Download citation

Further reading


Quick links