No. A nonsignificant difference in average GHB concentrations does not mean that every stored specimen remained unchanged. In a ten-case frozen-storage comparison, average heart and peripheral blood results did not change significantly after two years at −20°C, but one heart-blood result more than doubled.
What was actually retested?
Andresen-Streichert and colleagues included a long-term storage assessment within their broader postmortem GHB investigation. Blood from ten autopsy cases was analyzed again 24 months after the first measurement: ten heart-blood specimens and ten peripheral venous-blood specimens. These were paired repeat measurements from ten cases, not twenty independent deaths.
The broader study used specimens analyzed on the day of collection or immediately stored at −20°C. Its principal autopsy series had death-to-autopsy intervals of one to eleven days. Those overall intervals should not be treated as a separately established range for the ten-case storage subgroup.
The comparison tested what happened between two laboratory measurements during frozen storage. It did not compare blood during life with blood at autopsy and did not measure the amount of GHB formed before collection.
The averages and the individual exception
Heart-blood GHB averaged 21.6 ± 16.1 mg/L initially and 20.4 ± 16.8 mg/L at repeat testing. The paired comparison was nonsignificant, p = 0.779. Peripheral blood averaged 17.5 ± 7.2 mg/L initially and 15.2 ± 6.4 mg/L after storage, also nonsignificant, p = 0.312.
The individual trajectories were not identical. Heart-blood GHB increased more than twofold in one case, while the other nine results decreased. Peripheral GHB increased in one case and decreased in nine. All ten peripheral results remained below the study’s proposed 30 mg/L interpretive cutoff at both measurements.
That last finding supports a limited observation: none of these peripheral specimens crossed that threshold. It does not establish the maximum possible storage change in other specimens or show that no case near a threshold could change classification.
What “no significant difference” does and does not show
The authors used paired t tests. A p value above 0.05 means the test did not find a statistically significant difference under its assumptions. It is not a calculated probability that every specimen was stable. The study did not set and demonstrate an equivalence margin for individual specimen changes.
With only ten cases, the data cannot precisely characterize uncommon increases or all combinations of starting concentration, preservation and handling. The authors themselves called for further stability work in urine, vitreous humor and cerebrospinal fluid. A blood-storage result should not be extended to those matrices without additional evidence.
How analytical performance enters the comparison
GHB was quantified by derivatization GC–MS with GHB-d6 internal standard. The method’s detection and quantification limits were 0.5 and 0.6 mg/L; the reported linear calibration range was 1–100 mg/L. Validation at the tested controls gave accuracy of at least 91% and imprecision of 9% or less.
Those validation findings help assess the measurement procedure. They do not prove that every difference on repeat analysis came from storage, nor do they erase the observed specimen-level exception. Analytical variation and specimen change remain separate issues to evaluate.
What to retain in a case review
Keep both concentrations, both analysis dates, specimen site and the documented storage history. An unchanged classification after storage is different from an unchanged numerical concentration. Neither establishes GHB intake, impairment or contribution to death.
For evidence under different storage conditions, see GHB increases in stored postmortem blood. Its refrigeration and preservation experiments should not be conflated with this two-year frozen comparison.
Primary source
Andresen-Streichert H, Jensen P, Kietzerow J, et al. Endogenous gamma-hydroxybutyric acid (GHB) concentrations in post-mortem specimens and further recommendation for interpretative cut-offs. International Journal of Legal Medicine. 2015;129:57–68. doi:10.1007/s00414-014-1051-8.
Leave a Reply