Skip to main content

Effect of long-term frozen storage and thawing of stool samples on faecal haemoglobin concentration and diagnostic performance of faecal immunochemical tests.

Gies A, Niedermaier T, Weigl K, Schrotz-King P, Hoffmeister M, Brenner H

Publication: Clin Chem Lab Med (2019)PMID: 31655793View on PubMed

Purpose

This paper investigated the effects of long-term storage on fecal hemoglobin (Hb) levels using two different assays in specimens from patients with colorectal carcinoma (CRC), advanced adenoma (AA), advanced neoplasms, or none of the aforementioned conditions. Importantly, specimens were freeze-thaw cycled before the first measurement and freeze-thaw cycled 1-2 additional times before reanalysis.

Conclusion

When analyzed a second time after storage using the original method, mean Hb levels were significantly higher using the RIDAscreen assay and lower using the OC-sensor but, when further broken down by diagnosis, the differences using RIDAscreen were only significant for patients without CRC or AA (P<0.0001) and the differences using OC-sensor were only significant in patients with AA or advanced neoplasms (CRC or AA). The RIDAscreen sensitivity was non-significantly higher and specificity non-significantly lower in the second analysis, but the OC-Sensor sensitivity and specificity were both non-significantly lower in the second analysis. The test results differed between the first and second analysis for 10% of specimens when analyzed with RIDAscreen and for 5% of specimens when analyzed using the OC-sensor. Area under the curve (AUC) analysis found similar curves between the first and second analysis for both assays. Overall, Hb levels were modestly to strongly correlated between the two measurements using both methods, but Hb levels as determined by RIDAscreen were not correlated between the measurements in patients with CRC.

Studies (1)

Purpose

This study investigated the effects of long-term storage on fecal hemoglobin levels using two different assays in specimens from patients with CRC, AA, advanced neoplasms, or none of the aforementioned conditions. Fecal specimens were collected from a single bowel movement before screening colonoscopy from 516 patients of which 16 were diagnosed with CRC and 200 with advanced adenoma. An additional 50 specimens from patients with colorectal cancer were included. Patients stored feces frozen and brought them to the laboratory (3% thawed in transport). Specimens were stored at -20°C, shipped on dry ice to a central laboratory, and 504 were thawed for the first time and Hb levels were measured using the ELISA based RIDASCREEN 2-85 days after collection. Specimens were immediately refrozen at -80°C for 1 day to 44 months. A total of 551 specimens were analyzed (for the first or second time) and Hb levels determined using immunoturbidity based OC-Sensor. The remaining specimen was frozen again and stored 7-8 years before the second evaluation using both RIDASCREEN and OC-Sensor.

Summary of Findings

The mean Hb as determined by the RIDAscreen assay was significantly higher when measured the second time than when measured after the first thaw (P=0.0005). When further broken down by diagnosis, the means were higher in each group but only significant in patients without CRC or advanced adenoma (P<0.0001), although this was possibly due to low sample numbers. In contrast, mean levels of Hb using OC-sensor were lower when reanalyzed than in the first analysis (P=0.0326) but the differences were only significant in patients with AA (P=0.0179) or advanced neoplasms (P=0.0164). The RIDAscreen sensitivity was higher and specificity lower in the second analysis than the first and OC-Sensor sensitivity and specificity were lower in the second analysis than the first but the differences were not significant. Importantly, the test results differed between the first and second analysis for 10% of specimens when analyzed with RIDAscreen and for 5% of specimens when analyzed using the OC-sensor. AUC analysis found similar curves between the first and second analysis using both methods and only differed by 2.8% using RIDAscreen (not significant) and 2.6% using OC-sensor (P= 0.0147) for advanced neoplasms. When analyzed by RIDAscreen, Hb levels were modestly to strongly correlated between the two measurements for most patients (R=0.647-0.796, P<0.0001, all), but levels were not correlated in patients with CRC. Hb levels as determined by OC-sensor were strongly to very strongly correlated between the two measurements (R=0.755-0.912, P<0.0001, all).

  • Biospecimens

    FluidFeces
  • Preservation Types

    Frozen
  • Diagnoses

    Not specified
    NeoplasticCarcinoma
    NeoplasticBenign
  • Platforms

    ProteinELISA
    ProteinImmunoturbidometric assay
  • Pre-analytical Factors

    ClassificationPre-Analytical FactorValue(s)
    PreaquisitionDiagnosis/ patient condition
    AA
    Advanced neoplasms (either AA or CRC)
    CRC
    Neither AA or CRC
    StorageFreeze/thaw cycling
    1 cycle
    2 cycles
    3 cycles
    StorageStorage duration
    2-85 days
    6-10 years
    6-8 years
    ELISA SpecificTechnology platform
    Immunoturbidity