Do Recommended PAN Investigations Differentiate Cases From Controls?

Key Takeaways

• Abnormal laboratory findings were common in both groups, and recommended blood and throat testing did not distinguish PANS from idiopathic OCD and or tic disorder controls.
• In the PANS subgroup, CSF, MRI, and EEG findings were largely nonspecific; MRI showed no signs of neuroinflammation, and only rare potentially explanatory findings emerged.
• The authors concluded that the full investigation set rarely identified an underlying somatic condition, and incidental findings were often nonactionable.

A Stockholm case-control study found that recommended testing for suspected PANS did not clearly separate cases from psychiatric controls: abnormal findings appeared in 44 of 51 children with PANS and 56 of 58 controls. After correction for multiple comparisons, no laboratory variable remained significantly different between groups. The laboratory workup provided little diagnostic separation between the groups.

This original case-control study drew on prospectively collected data from the longitudinal Karolinska PANS cohort and specialist clinics in Stockholm, Sweden. It included children aged 4 to 18 years: 51 in the PANS group and 58 with idiopathic OCD and or tic disorders as controls. The guideline-recommended PANS laboratory investigations covered 56 variables from blood tests and throat cultures, with CSF analysis, brain MRI, and EEG available in a PANS subsample when clinically indicated. Participants were recruited from January 2020 through September 2023, and analyses were completed in 2024 and 2025. The study was designed to compare diagnostic yield with a psychiatric control group rather than establish causation.

Most children in both groups had several abnormal results, with most PANS participants showing 3 or more and most controls showing 4 or more. Among children with PANS, common findings included low ferritin in 10 of 35 tested, low leukocytes in 12 of 48, and low complement C4 in 9 of 42. Among controls, low leukocytes appeared in 24 of 58 tested, while low orosomucoid and low vitamin D each appeared in 14 of 55. Group A streptococci were found by throat culture in 5 of 30 tested PANS participants and 4 of 47 controls. These abnormalities were broadly nonspecific rather than a distinct PANS profile.

Among the 18 children with PANS who underwent broader neurologic evaluation, CSF pleocytosis appeared in 1 of 14 tested participants, and 2 of 9 had slightly elevated protein fractions, including albumin. None of the 14 tested had oligoclonal bands, and brain MRI showed no signs of neuroinflammation. Incidental benign cysts appeared on 5 of 12 MRI scans, and EEG abnormalities appeared in 3 of 15 participants, with generalized slowing in 1 and focal abnormalities in 2, without epileptiform activity. Blood analysis identified celiac disease in 1 child, and another had laboratory and EEG findings compatible with neuroinflammatory activity without a definite diagnosis before PANS assessment. Overall, the extended workup yielded mostly nonspecific or incidental findings.

The investigators concluded that the comprehensive recommended investigations rarely identified underlying somatic conditions in children with PANS and often produced incidental, nonactionable findings. Interpretation of CSF, MRI, and EEG results was limited because those tests lacked a control comparison, were confined to an 18-participant PANS subgroup, and involved a modest sample. In sensitivity analysis, shorter time from symptom onset to testing was linked to more abnormal findings in PANS, but the model explained less than 3% of variance and was influenced by 4 outliers. Principal component analysis suggested that age, rather than inflammatory markers or group assignment, accounted for much of the observed variance. Overall, the workup detected many abnormalities but offered little diagnostic separation.