Metabolic and lipidomic stability in EDTA plasma from hospitalized patients: impact of refrigerated storage time and additional freeze-thaw cycle.
Author(s): Guttorm SJT, Gukasyan L, Prebensen C, Karlsson NG, Holten AR, Kvale D, Rootwelt H, Elgstøen KBP, Amundsen EK
Publication: Metabolomics, 2026, Vol. 22, Page
PubMed ID: 42024200 PubMed Review Paper? No
Purpose of Paper
This paper compared the liquid chromatography-mass spectrometry (LC-MS) metabolome and lipidome of plasma stored in the collection tube for 0, 24 and 72 h at 4°C before freezing and after 1 or 2 freeze-thaw cycles.
Conclusion of Paper
Principal Component Analysis (PCA) of the plasma metabolome or lipidome was unaffected by storage at 4°C for up to 72 h or an additional freeze-thaw cycle, with specimens instead clustering by patient source. Over 87% of the metabolites and 90% of the lipids remained stable in plasma for up to 72 h of pre-centrifugation storage. However, significant changes occurred in levels of individual metabolites, including lactate, hypoxanthine, oxoproline, glutamate, and aspartate and several polyunsaturated fatty acids (PUFAs).
Studies
-
Study Purpose
This study compared the liquid chromatography-mass spectrometry (LC-MS) metabolome and lipidome of plasma stored in the collection tube for 0, 24 and 72 h before freezing and after 1 or 2 freeze-thaw cycles. EDTA blood was collected from 20 hospitalized patients (diagnosis included non-specified infection, cancer, chronic lung disease, heart failure and chronic kidney disease) and plasma was separated by centrifugation at 2200 g for 10 min. Plasma from one tube per patient was immediately aliquoted and frozen, while the remaining specimens were stored in the original tube for 24 or 72 h at 4°C before aliquoting and freezing. Plasma aliquots were stored at -80°C for approximately 1 month. Specimens were thawed for 1 h, analyzed, and then refrozen for 3 months before re-analysis. The metabolome and lipidome were analyzed by LC-MS. The metabolome was annotated using in-house libraries while the lipidome was annotated using LipidSearch (v5.1), LipidBlast, and LipidMaps.
Summary of Findings:
A total of 3,516 metabolite features were detected in plasma specimens, 2,196 and 1,320 metabolites were detected in positive and negative ionization modes, respectively. Principal Component Analysis (PCA) clustered specimens from the same patient together, regardless of storage duration. However, when plasma was stored in the original tube for 24 h at 4°C after centrifugation ANOVA revealed significant differences in levels of 145 (6.6%) metabolites in positive mode and 45 (3.4%) in negative mode , both of which this increased to 277 (12.6%) metabolites in positive mode and 86 (6.5%) in negative mode when specimens were stored for 72 h. Of the metabolite features that were affected by storage duration, 53 were annotated with a confidence level of 1-3. For 48 of these, including lactate, hypoxanthine, oxoproline, glutamate, and aspartate, storage had an effect greater than Cohen’s d±0.25. An additional freeze-thaw cycle did not have a noticeable effect on PCA clustering. When plasma that was freeze-thawed once versus twice 8.7% (positive ionization) and 7.1% (negative ionization) of metabolites were differentially expressed, but only 5 of these metabolites were annotated with a confidence level of 1-3. An additional freeze-thaw cycle had an effect greater than Cohen’s d±0.25 on the level of 4 of these metabolites: glycerol, fatty acid (20:4), 2-amino-4-cyanobutanoic acid, 1-Hydroxy-2-napthoic acid.
A total of 7,973 lipid features were detected, 5,281 in positive mode and 2,692 in negative mode. PCA of the lipidome clustered plasma specimens predominantly by patient source, with no clear effect of storage. Plasma storage (4°C) for 24 h resulted in differential expression of 386 (7.3%) and 155 (5.8%) of lipids in positive and negative ion mode, respectively, but only 367 (6.9%) and 132 (4.9%) of lipids, respectively, after 72 h of storage. Of the differentially expressed lipid features, 79 were annotated with a confidence level of 1-3, and for 68 of these lipid features, the effect was greater than Cohen’s d±0.25. The lipids affected included polyunsaturated fatty acids (PUFAs), including fatty acid (20:4), phosphatidylcholine (20:4_22:6), phosphatidylethanolamine (18:1_20:4), phosphatidylserine (18:0_20:4) and lysophosphatidylcholines. An additional freeze-thaw cycle did not have a noticeable effect on PCA clustering of the lipidome, but 9.4% (positive ionization) and 10.8% (negative ionization) of lipids were differentially expressed between specimens freeze-thawed once and twice. Of these lipids, 85 were sufficiently annotated, and an effect greater than Cohen’s d±0.25 was observed for 27 of these lipids (including acylcarnitines; sphingomyelins; and several Polyunsaturated fatty acid containing lipids, including the polyunsaturated fatty acids lysophosphatidylcholine (20:4), triacylglycerol (16:0_18:2_22:6), and phosphatidylcholine (18:1_22:6)).
Biospecimens
Preservative Types
- Frozen
Diagnoses:
- Cardiovascular Disease
- Neoplastic - Not specified
- Other diagnoses
Platform:
Analyte Technology Platform Small molecule LC-MS or LC-MS/MS Lipid LC-MS or LC-MS/MS Pre-analytical Factors:
Classification Pre-analytical Factor Value(s) Storage Freeze/thaw cycling 1 cycle
2 cycles
Storage Storage duration 0 h
24 h
72 h