Exploring the stability of whole blood samples for remote health care applications towards reliable home-based blood testing

Abstract

Objectives

To evaluate the stability of commonly used biomarkers in whole blood under various storage conditions, in order to assess their suitability for home-based blood collection in remote care settings.

Methods

Whole blood samples were stored at 4–8°C, 20–25°C and 37°C for up to 72 h. Six pooled samples and two healthy volunteer samples were aliquoted and analysed at four time points (T0, T24, T48 and T72 h). A panel of 47 routine chemistry and immunochemistry biomarkers was measured in all samples. Recoveries relative to T0 were compared to within subject coefficients of variation of the individual biomarkers.

Results

Most biomarkers remained stable at refrigerated and room temperature conditions up to 72 h with a subset of biomarkers, mostly proteins and tumour markers also showing good to acceptable stability at 37°C. Several biomarkers, including potassium, inorganic phosphate and iron, consistently fell outside acceptable limits under multiple conditions. Others, such as sodium, calcium, ferritin, aspartate aminotransferase and cytokeratin fragment 21-1, were unstable only at 37°C. A few transient deviations were observed, but these were not consistent over time.

Conclusions

Home-based blood sampling is feasible for a broad range of biomarkers, provided that heat exposure during transport is minimised. These findings are a first step in the further validation of selected analytes under real-world, capillary blood sampling conditions.

Keywords

Remote health care biomarker stability whole blood capillary blood collection diagnostic testing

Introduction

Health care systems worldwide face increasing pressure due to population ageing and the growing burden of chronic diseases. By 2050, over 2 billion people will be aged 60 or older, paralleled by rising rates of non-communicable diseases (NCDs) such as cardiovascular disease, diabetes and cancer, together responsible for more than 70% of global deaths.¹ These trends are expected to drive demand for health care services, while many systems continue to struggle with workforce shortages and limited resources.^2,3

To address these challenges, care delivery must shift toward sustainable, scalable and patient-centred models. Accelerated by the COVID-19 pandemic, remote care such as telemedicine, home monitoring and remote follow-up has emerged as a viable alternative to in-person care,^4–6 improving both accessibility and patient empowerment.⁷

One promising approach is self-collection of capillary blood, either via fingerprick or automated sampling devices.^8–10 In the first issue of CCLM this year which was accompanied by an editorial from Plebani, Poland and Cobbaert highlighted blood self-sampling devices as a key innovation in this transition, describing them as ‘the future of laboratory diagnostics’ due to their potential for cost-effectiveness, logistical efficiency, patient convenience and seamless integration into fully automated laboratory workflows.¹¹

After collection, patients can mail the samples to a central laboratory for analysis, eliminating the need for a health care facility visit. While home-based sampling offers clear logistical and clinical benefits, pre-analytical challenges remain. Samples collected at home are exposed to variability in transport time and ambient temperature, which may affect analyte stability and thus clinical interpretation.^12,13

To explore this, we conducted an exploratory study assessing the stability of a broad range of laboratory parameters in whole blood stored under different temperature conditions. The goal was to identify which biomarkers remain robust under delayed or non-standardised conditions and which may require stricter protocols. These insights aim to inform future implementation strategies for remote blood collection.

Methods

Study design

We conducted an exploratory laboratory study at the Department of Clinical Chemistry of Erasmus University Medical Centre to assess the stability of routine chemistry and immunochemistry biomarkers in whole blood stored under various temperature conditions over time. Residual venous blood collected in lithium heparin tubes (BD Vacutainer®, Becton & Dickinson, UK) for blood gas analysis was used to create six pooled samples. Additionally, two individual samples were obtained from a healthy male and female volunteer. The healthy volunteer samples were obtained by venipuncture and collected in lithium heparin tubes, identical to the pooled samples.

Samples were stored under three conditions: refrigerated (4–8°C), room temperature (20–25°C) and elevated temperature (37°C), using a laboratory refrigerator or incubator. The selected temperatures reflect plausible transport conditions for home-based blood sampling: refrigerated (4–8°C), ambient postal conditions (20–25°C) and a worst-case heat exposure scenario (37°C). All samples were protected from light. Analyses were performed at four time points: baseline (T0), 24 h (T24), 48 h (T48) and 72 h (T72). For pooled samples, T0 analyses occurred within 6 h of the initial blood gas analysis and for the volunteer samples, within 1 h of phlebotomy.

Each sample (six pooled + two volunteers) was divided into twelve 800 μL aliquots in 13 × 75 mm false-bottom tubes (Sarstedt, Germany). An aliquot volume of 800 µL was required to allow analysis of the complete biomarker panel at all time points. At each time point and temperature, a separate aliquot was analysed according to standard laboratory procedures.

Laboratory analyses

All whole blood aliquots were centrifuged for 5 min at 3000 g. A comprehensive panel of 47 routine chemistry and immunochemistry biomarkers was measured in all samples. All parameters were analysed on a cobas 8000 system (chemistry: C502/C702 module; immunochemistry: E801 module) according to the manufacturer’s instructions (Roche Diagnostics, Switzerland). An overview of the full panel is provided in Table 1.

Table 1.

Overview of the 47 Analytes in This Study. The Numbers Between the Brackets Represent the T0 Average, Minimum and Maximum Biomarker Concentrations of the Samples (6 Pooled and 2 Volunteers), Respectively.

Biomarker category	Biomarker	Abbreviation	T0 average	Min–Max concentration
Electrolytes	Sodium	Na	139 mmol/L	138–144 mmol/L
	Potassium	K	4.2 mmol/L	3.8–4.5 mmol/L
	Chloride	Cl	103 mmol/L	100–105 mmol/L
	Calcium	Ca	2.39 mmol/L	2.21–2.52 mmol/L
	Phosphate	PO₄	1.27 mmol/L	1.10–1.37 mmol/L
	Magnesium	Mg	0.81 mmol/L	0.67–0.87 mmol/L
	Iron	Fe	13 µmol/L	9.6–18 µmol/L
Metabolites	Urea	Ure	7.9 mmol/L	4.4–14 mmol/L
	Creatinine	Cre	145 µmol/L	75–210 µmol/L
	Cystatin C	CysC	1.66 mg/L	0.91–2.24 mg/L
	Uric acid	UA	0.35 mmol/L	0.28–0.41 mmol/L
	Total bilirubin	TBil^a	60 µmol/L	—
	Conjugated bilirubin	DBil^a	51 µmol/L	—
Enzymes	Alanine aminotransferase	ALT	34 U/L	17–57 U/L
	Aspartate aminotransferase	AST	35 U/L	23–55 U/L
	Gamma-glutamyl transferase	GGT	97 U/L	16–240 U/L
	Alkaline phosphatase	AP	158 U/L	110–221 U/L
	Creatine kinase	CK	140 U/L	59–345 U/L
Proteins	Total protein	TP	70 g/L	63–77 g/L
	Albumin	Alb	35 g/L	30–42 g/L
	Immunoglobulin G	IgG	11 g/L	8.9–13 g/L
	Immunoglobulin A	IgA	1.71 g/L	1.06–2.69 g/L
	Immunoglobulin M	IgM	1.08 g/L	0.76–1.54 g/L
	Transferrin	Trf	2.4 g/L	2.1–2.7 g/L
	Ferritin	Fer	424 µg/L	105–1081 µg/L
	Complement C3	C3	1.34 g/L	1.06–1.86 g/L
	Complement C4	C4	0.25 g/L	0.19–0.42 g/L
	Apolipoprotein B	ApoB	0.78 g/L	0.58–1.11 g/L
Lipids	Triglycerides	TG	1.74 mmol/L	0.80–3.56 mmol/L
	Low-density lipoprotein-cholesterol	LDL-chol	2.38 mmol/L	1.89–3.65 mmol/L
	High-density lipoprotein-cholesterol	HDL-chol	1.15 mmol/L	0.95–1.26 mmol/L
	Total cholesterol	Chol	4.21 mmol/L	3.40–5.50 mmol/L
Vitamins	Vitamin B12	B12	532 pmol/L	325–764 pmol/L
	Holo-transcobalamin	aB12	164 pmol/L	83–216 pmol/L
	Folate	Fol	20 nmol/L	8.9–35 nmol/L
Cardiac marker	N-terminal pro-B-type natriuretic peptide	NT-proBNP	135 pmol/L	4.0–700 pmol/L
Tumour markers	Cancer antigen 125	CA125	26 kU/L	5.0–74 kU/L
	Cancer antigen 19-9	CA19-9	16 kU/L	4.0–33 kU/L
	Cancer antigen 15-3	CA15-3	51 kU/L	6.0–292 kU/L
	Cytokeratin fragment 21-1	CYFRA 21.1	2.93 µg/L	1.12–6.17 µg/L
	Protein induced by vitamin K absence II	PIVKA-II	87 ng/mL	16–293 ng/mL
	Alpha-fetoprotein	AFP	40 µg/L	—
	Human chorionic gonadotropin	hCG^a	3899 IU/L	—
	Total prostate-specific antigen	tPSA^a	0.51 µg/L	0.10–1.3 µg/L
Sample integrity markers	Haemolytic index	—	6.3 µmol/L	2.0–13 µmol/L
	Icteric index	—	23	10–79
	Lipemic index	—	9.6	8.0–13

^aFor total- and conjugated bilirubin, alpha foetal protein and human chorion gonadotropin, 7/8 samples showed very low T0 concentrations. Accordingly, the absolute concentration of one pool was used.

Data analysis

To assess relative stability, the measured aliquot was normalised to the baseline (T0) value (Recovery=(Tx/T0-1)×100%) and averaged, resulting in a mean recovery. For each biomarker, mean recovery percentages were calculated per time point and storage condition and visualised in a heatmap. A recovery below 0% indicated a decrease and above 0% indicated an increase in recovery. We used the within-subject coefficients of variation percentages (%CVi) as published on the EFLM biological variation database (https://biologicalvariation.eu) to assess whether the recovery values were within an acceptable range. For biomarkers with a %CVi>10% and biomarkers with no meta-data leading to a %CVi, a maximum recovery of 10% was used. Recovery percentages were classified into three categories: ‘good stability’ (recovery percentage lower than the %CVi), ‘acceptable stability’ (recovery percentage between the %CVi and 10%) and ‘unacceptable stability’ (recovery percentage >10%).

To prevent misinterpretation of recovery values over time and at different temperatures, we excluded results from biomarkers with very low analyte concentrations in which the percentile change was more than 10% but in which the absolute change in concentration had no clinical relevance. For example, a conjugated bilirubin concentration with an initial concentration of 3 umol/L decreasing to 2 umol/L gives a recovery of −33% but this decrease is of no clinical relevance.

Results

In curating the data, we investigated whether filtering out outliers had an impact on the stability results. Using a cut-off z-score of 2 to detect outliers, we did see some improvement in variation coefficients of certain biomarkers at certain time points and temperatures; however, this did not impact the stability classification. We therefore decided not to exclude outliers in the analysis.

The heatmap in Figure 1 shows the normalised mean scores for the three time points and the three different temperature settings.

Figure 1.

Heatmap for the mean normalised values relative to analyte specific CVi, at three time points under three temperature conditions.

Electrolytes

While chloride showed acceptable recoveries for all three temperatures and time points, sodium, calcium and magnesium showed acceptable to good recovery percentages for both refrigerated and room temperature for up to 72 h. For inorganic phosphate, stability was good at 4–8°C for three time points and only for 24 h at room temperature. Iron and potassium however showed limited whole blood stability with good recoveries for iron up to 48 h at 4–8°C and an acceptable recovery of 4.9% for potassium at room temperature for 24 h.

Metabolites

All metabolites, including total bilirubin and conjugated bilirubin, were highly stable with good recovery percentages at both refrigerated and room temperature for up to 72 h. For total bilirubin and conjugated bilirubin, all but one pool (pool 4) was excluded from analysis due to very low T0 concentrations. Pool 4 had a T0 concentration of 60 and 51 umol/L for total and conjugated bilirubin, respectively. While total bilirubin was stable for up to 48 h at 37°C, conjugated bilirubin was slightly less stable showing good recoveries up to 72 h at room temperature. Creatinine and cystatin C remained stable in whole blood with acceptable to good recoveries for up to 72 h at 37°C.

Enzymes

With the exception of AST and alkaline phosphatase, both showing unacceptable recoveries at 37°C, all enzymes tested remained stable in whole blood at all three temperatures and time points.

Lipids

All lipids demonstrated excellent stability across nearly all time points and conditions. Two exceptions were triglycerides (−11% after 3 days at 37°C) and HDL-cholesterol (+11% after 1 day at 37°C).

Vitamins

Vitamin B12 and active vitamin B12 were generally stable. Folate, however, showed a consistent decrease over time at room temperature with mean recoveries of −17% and −24% after 2 and 3 days, respectively.

Proteins

Like the enzymes, proteins showed overall good to acceptable recoveries at all temperatures and time points. Exceptions were ferritin (not stable at 37°C at all time points), complement C3 (+12% recovery at room temperature after 3 days), complement C4 (+14% recovery at 37°C after 1 day) and apo-lipoprotein B (+12% recovery at 37°C after 1 day).

Cardiac marker

NT-proBNP showed good recoveries up to 3 days at 4–8°C and room temperature. At 37°C, NT-proBNP remained stable up to 24 h after which recoveries decreased to −11% and −17% after 2 and 3 days, respectively.

Tumour markers

All tumour markers showed good stability under all conditions. Total PSA and CYFRA 21.1 were the exception. While tPSA showed an unacceptable stability after 3 days at 37°C, CYFRA 21.1 declined dramatically at 37°C for all time points. For alpha foetal protein and human chorionic gonadotropin, all but one pool (both pool 1) was excluded from analysis due to very low T0 concentrations. Pool 1 had a T0 concentration of 40 ug/L and 3899 U/L for alpha foetal protein and human chorionic gonadotropin, respectively.

Sample integrity markers

In Figure 1, the integrity indices are shown as averaged absolute values. The icteric index remained stable for all three temperatures and time points. The haemolytic and lipemic index increased substantially at 37°C.

Discussion

This exploratory study evaluated the stability of 47 commonly requested routine chemistry and immunochemistry biomarkers in whole blood stored under various temperature conditions. The thing that sets our exploratory study apart from previously published whole blood stability studies (i.e. Refs. 14 and 15) is that we included a higher storage temperature of 37°C and prolonged storage times, extending the stability analysis for 48 and 72 h. This enabled us to assess whether biomarkers were suitable candidates for home-based (self)sampling with patients sending in the samples to a central lab via postal mail.

Most biomarkers remained stable in whole blood when stored at refrigerated (4–8°C) or room temperature (20–25°C). Elevated temperature (37°C), however, substantially reduced stability of several biomarkers, underscoring the importance of proper sample conditioning in remote care. The selected temperatures reflect plausible transport conditions for home-based blood sampling, including a worst-case heat exposure scenario. In contrast to stability criteria derived from controlled laboratory workflows, this exploratory study focused on prolonged and less controlled transport conditions relevant to remote sampling, in line with CRESS recommendations.¹⁶

At elevated temperature, instability was observed mainly for electrolytes (except chloride), uric acid, conjugated bilirubin, AST, alkaline phosphatase, ferritin NT-proBNP and CYFRA 21.1. For these biomarkers, remote sampling is still viable albeit under strict temperature control. Interestingly, even at 37°C a fair set of biomarkers, mainly proteins and tumour markers, showed acceptable to good stability for up to 3 days.

A subset of biomarkers, mostly electrolytes like potassium, inorganic phosphate and iron, consistently fell outside acceptable limits under multiple conditions, indicating that these are less suitable for remote workflows.

HDL-cholesterol, complement C4 and apolipoprotein B appeared unstable after 24 h at 37°C; however, these observations were transient and not consistent over time with recoveries at subsequent time point at this temperature remained within acceptable limits. This transient deviation may warrant further investigation.

In the context of remote sampling, transport conditions, particularly temperature exposure and transit time, are harder to control than in clinical settings, where the pre-analytical phase is tightly managed. Home-based sampling relies on patients to collect and mail samples, introducing variability that may affect integrity. The WHO guideline Use of Anticoagulants in Diagnostic Laboratory Investigations similarly stresses that both anticoagulant choice and transport conditions are critical for reliable results.¹⁷ While strategies like insulated mailers and phase-change materials can reduce risk,^18,19 they add complexity and cost, and transport methods must balance clinical value with economic feasibility.

In light of other strategies in remote sampling, there are numerous publications on the successful use of dried blood spots (DBSs) in remote care.²⁰ While the DBS applications focus mainly on neonatal screening and therapeutic drug monitoring, the use cases in routine clinical chemistry and immunochemistry are limited.²¹ One feasible explanation for this is that there are no solutions to integrate the processing of DBS in a total laboratory automation concept, making it labour intensive.

As health care systems face growing pressure from ageing populations, chronic disease and workforce shortages, more sustainable care models are needed. Remote and home-based solutions, including home blood collection, could help relieve pressure on health care infrastructure.^3,22 Monitoring transit time and temperature during transport are essential to assess whether a biomarker concentration is reliable. Modalities like mobile temperature dataloggers, for example, the TAPP datalogger (https://tapp.online), offer possibilities for this.

This study provides an initial assessment of remote blood testing feasibility by evaluating the stability of commonly used analytes under different storage conditions. Rather than drawing firm conclusions, we aimed to identify which analytes are robust enough to justify further validation in real-world settings. These findings can help inform future efforts to develop safe, reliable remote testing protocols.

Several limitations should be noted. First, we used pooled venous samples from standard phlebotomy, which may not fully reflect capillary blood obtained via home-based devices. Although high concordance between capillary and venous samples has been reported for selected analytes in remote settings, this study was intended as an exploratory first step, and further validation in genuine capillary blood is required.⁸ Second, the sample size was limited, and for most biomarkers clinically abnormal cases were not included. For total and conjugated bilirubin, alpha-fetoprotein and human chorionic gonadotropin, measurable baseline concentrations were available in only one sample pool, which precludes drawing firm conclusions regarding stability and warrants cautious interpretation of these findings. Third, all experiments were conducted under controlled lab conditions, which may not represent real-world variability in home collection and transport. Finally, our stability threshold (±10% deviation from baseline) was arbitrary and may not align with clinical cut-offs for all analytes.

Nonetheless, these findings offer a valuable starting point for prioritising analytes for further investigation. Future studies should validate results using capillary blood under real-world conditions and include clinically relevant populations to assess diagnostic accuracy and practical utility.

Conclusions

This study demonstrates that many commonly used analytes in whole blood remain stable under refrigerated and room temperature conditions for up to 72 h, supporting the feasibility of home-based blood sampling in remote care. However, elevated temperatures significantly affect the stability of several parameters, highlighting the need for temperature-controlled transport. These findings inform future efforts to develop reliable, patient-friendly remote diagnostic pathways.

Footnotes

ORCID iDs

M. H. Elise van Driel

Christian Ramakers

Ethical approval

This study was conducted in accordance with the Declaration of Helsinki. Residual lithium-heparin whole blood samples from routine clinical care were used after completion of diagnostic procedures and were fully anonymised before analysis. According to Erasmus MC policy on the use of left-over material for medical research, formal ethics approval was not required for this type of study (Erasmus MC, Left-over material used for medical research, available at: ).

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Guarantor

Christian Ramakers is responsible for the manuscript’s integrity.

Contributorship

The study was conceived and designed by Christian Ramakers and Dirk J. Grünhagen. Christian Ramakers and Alan Han carried out the experimental work and performed part of the analyses. M.H. Elise van Driel conducted the remaining analyses and drafted the manuscript. All authors contributed to reviewing and editing the manuscript and approved the final version.

Data Availability Statement

The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request. Requests may be sent to initiating author Dr. C. Ramakers (c.ramakers@erasmusmc.nl).

References

World Health Organization . Noncommunicable diseases. Available from: https://www.who.int/news-room/fact-sheets/detail/noncommunicable-diseases (2023, Accessed 15 April 2025).

World Health Organization . Global strategy on human resources for health: workforce 2030. Geneva: WHO, 2016.

Jones

Dolsten

. Healthcare on the brink: navigating the challenges of an aging society in the United States. NPJ Aging 2024; 10(1): 22.

Sirina

Andrea

Kevin

. Covid-19 and health care’s digital revolution. N Engl J Med 2020; 382(23): e82.

Peyroteo

Ferreira

Elvas

, et al. Remote monitoring systems for patients with chronic diseases in primary health care: systematic review. JMIR mHealth uHealth 2021; 9(12): e28285.

Greenhalgh

Vijayaraghavan

Wherton

, et al. Virtual online consultations: advantages and limitations (VOCAL) study. BMJ Open 2016; 6(1): e009388.

van Driel

MHE

Wullaert

Voigt

, et al. Patient-led follow-up after curative colorectal cancer surgery: the primary outcome analysis of the prospective, multicentre FUTURE-primary implementation study. Eur J Cancer 2025; 226: 115599.

Voigt

Wullaert

Gobardhan

, et al. Feasibility, reliability and satisfaction of (automated) capillary carcinoembryonic antigen measurements for future home-based blood sampling: the prospective CASA-I study. Colorectal Dis 2024; 26(8): 1560–1568.

Tasso Inc . Tasso+ device for blood collection. Available from: https://www.tassoinc.com/ (2024, Accessed 15 April 2025).

10.

Health

. TAP II – Touch-Activated Phlebotomy device. Available from: https://www.yourbiohealth.com/ (2024, Accessed 15 April 2025).

11.

Poland

DCW

Cobbaert

. Blood self-sampling devices: innovation, interpretation and implementation in total lab automation. Clin Chem Lab Med 2025; 63(1): 3–13.

12.

Lippi

Guidi

Mattiuzzi

, et al. Preanalytical variability: the dark side of the moon in laboratory testing. Clin Chem Lab Med 2006; 44(4): 358–365.

13.

Wang

. How long can we store blood samples: a systematic review and meta-analysis. EBioMedicine 2017; 24: 277–285.

14.

van Balveren

Huijskens

Gemen

, et al. Effects of time and temperature on 48 routine chemistry, haematology and coagulation analytes in whole blood samples. Ann Clin Biochem 2017; 54(4): 448–462.

15.

Abraham

Agrawal

Acharya

, et al. Effect of temperature and time delay in centrifugation on stability of select biomarkers of nutrition and non-communicable diseases in blood samples. Biochem Med 2019; 29(2): 020708.

16.

Cornes

Simundic

Cadamuro

, et al. The CRESS checklist for reporting stability studies: on behalf of the European Federation of Clinical Chemistry and Laboratory Medicine (EFLM) Working Group for the Preanalytical Phase (WG-PRE). Clin Chem Lab Med 2020; 59(1): 59–69.

17.

World Health Organization . Use of anticoagulants in diagnostic laboratory investigations. Geneva, 2002: Report No.: WHO/DIL/LAB/99.1 Rev.2.

18.

Schaff

Collier

Iacovetti

, et al. Validation of a reusable, thermally protective specimen shipper for decentralized sample collections. The Journal of Applied Laboratory Medicine 2025; 10(4): 874–886.

19.

Thibault

Beauséjour

Jacques

, et al. Overnight storage of whole blood: cooling and transporting blood at room temperature under extreme temperature conditions. Vox Sang 2014; 106(2): 127–136.

20.

Lehmann

Delaby

Vialaret

, et al. Current and future use of “dried blood spot” analyses in clinical chemistry. Clin Chem Lab Med 2013; 51(10): 1897–1909.

21.

Jacobson

Rahbari

Schwartz

, et al. Dried blood spots to assess cardiovascular-kidney-metabolic health. J Am Heart Assoc 2025; 14(6): e037454.

22.

Kurashvili

Reinhold

Järvis

. Managing an ageing healthcare workforce: a systematic literature review. J Health Organisat Manag 2022, ahead-of-print(ahead-of-print).