--200 patients with seizure-like activity, 109 ultimately diagnosed with seizure
--31% (of 200 patients) had abnormal prolactin (upper limit of normal ~30mg/dL)
--the numbers:
- sensitivity of this serum prolactin was 42%
- specificity was 82%
- positive predictive value (PPV) of 74%
- negative predictive value (NPV) of 54%
- overall accuracy of 60% in the diagnosis of seizure,
- likelihood ratio of 2.4
REVIEW ARTICLE:
--most studies used 2x baseline serum prolactin level as 'elevated'
--the numbers:
- pooled sensitivity for generalized tonic-clonic seizures (60.0%); for complex partial seizures (46.1%)
- pooled specificity was similar for both (approximately 96%)
- 2 Class II studies were consistent in showing prolactin elevation after tilt-test-induced syncope.
META-ANALYSIS:
--usefulness of raised serum prolactin in diagnosing generalised tonic-clonic seizures (GTSC) in patients presenting to the ED after a single episode of syncope
--13 relevant studies only 3 met the criteria for evaluation
--the numbers: if a serum prolactin concentration is > 3x the baseline when taken within one hour of syncope, then...
- LR (likelihood ratio) of GTSC vs pseudoseizure = 8.92, sensitivity 0.62, specificity 0.89
- LR of GTSC vs. syncope = 4.60, sensitivity 0.71, specificity 0.85
10-SECOND TAKEAWAY:
--serum prolactin in seizure: not too useful in the ED
--generally poor sensitivity, better specificity, but only if tested early (~10-60 min s/p episode)
--so if you can draw it fast, and if its significantly elevated, it might be useful (a lot of if's), but if its low, doesn't mean it's not a seizure
--serum prolactin in seizure: not too useful in the ED
Reference(s): study, review, meta-analysis, picture
No comments:
Post a Comment