Stability of steroid hormones in dried blood spots (DBS)

Introduction

The use of dried blood spots (DBS) for research and diagnostic purposes has gained significant attention in recent years due to its convenience and efficiency. DBS sampling allows patients or study subjects, with proper training, to sample themselves at home [1], [2], [3], [4].

A prerequisite for at-home sampling is stability of the analytes in DBS during the transport to and storage in the laboratory [5, 6]. Guidelines for shipping and handling DBS samples in general are available to help maintain sample quality [5, 7]. Sample quality, and then mainly stability of the DBS can be influenced by the storage temperature [8], [9], [10]. Pivotal factors, such as humidity and the specific method and material employed for sampling, have been identified as important factors influencing the quality of DBS samples [11], [12], [13].

DBS-sampling is used in some patient groups, like children with phenylketonuria [14]. Other patient groups that may also benefit from sampling DBS at-home are individuals that need regular control of their steroid hormone levels, such as pediatric and adult patients with CAH, and hypogonadal men and transmen undergoing testosterone therapy with regular controls of their testosterone concentrations. In addition, as steroid hormones have circadian rhythms and vary during the menstrual cycle, also timed at-home DBS sampling are desired for diagnostic and study purposes.

Although steroids in general are known to be rather stable, DBS samples sent by regular mail to the laboratory can encounter very low to very high temperatures in the mailbox [15]. The same implies for studies or trials, as analysis of the DBS samples may occur in batches, implicating considerable storage time. Therefore, the primary objective of this study is to investigate the stability of steroid hormone concentrations (cortisol, cortisone, corticosterone, testosterone, androstenedione, and 17-OHP) in DBS by storing DBS samples under various temperature conditions for prolonged periods of time in order to establish the possible shipping and storage conditions.

Materials and methods

Data collection

Experiment 1 (between-spot variation)

Samples of 10 healthy volunteers (5 male, 5 female), consisting of 5 DBS per volunteer, were collected and analyzed to determine the between-spot variation (coefficient of variation; CV). In all 5 spots per volunteer steroid concentrations were measured. The between-spot variation was determined to set the acceptable limits for variation in the stability experiments [2], [3], [4].

Experiment 2 (storage at 37 °C)

DBS samples of 5 male and 5 female healthy volunteers were used to test the stability of steroid hormones in DBS when stored at 37 °C. Directly after the blood spots had dried, hormone concentrations were measured and set as 100 % (day 0). The other spots were stored together with desiccants at 37 °C. Next, in these spots, concentrations of steroid hormones were measured after 2, 7 and 14 days. These results were plotted against concentrations at day 0 to determine the changes in concentration and thus the stability of the steroid hormones.

Experiment 3 (short term storage)

Groups of 5 males and 5 females sampled 5 DBS spots each which were used to test the stability for the three temperature conditions (−20 °C, 4 °C and room temperature (RT)). Directly after the blood spots had dried, hormone concentrations were measured and set as 100 % (day 0). The other four spots were stored together with desiccants at −20 °C (group 1), 4 °C (group 2) or RT (group 3). Next, in these spots, concentrations of steroid hormones were measured after 7 and 14 days. These results were plotted against concentrations at day 0 to determine the changes in concentration and thus the stability of the steroid hormones.

Experiment 4 (long term storage)

During experiment 3 two more DBS were taken by the healthy volunteers for long term storage. Steroid hormone concentrations were measured after 3 and 6 months of storage at −20 °C (group 1), 4 °C (group 2) or RT (group 3).

The timeline of the experiments 2–4 per storage temperature is shown in Figure 1 .

Figure 1:

Timeline of the experiments per storage temperature. The upper panel shows the timeline of measurements for steroid hormones in DBS stored at 37 °C. Experiments 3 and 4 are explained in the lower panel and show the timeline for steroid hormone measurements for DBS stored at −20 °C, 4 °C and RT. Data from experiment 4 is compared to baseline (day 0, Exp. 3). RT, room temperature.

Results

Experiment 1. Between-spot variation

The results of the between-spot variation per volunteer and per steroid are shown in Figure 2. Median CVs for all steroid hormones were <10 % (testosterone 9 % (6–18 %), androstenedione 7 % (5–10 %), cortisol 7 % (4–14 %), cortisone 7 % (5–12 %) corticosterone 10 % (6–13 %)) respectively, except for 17-OHP, which showed a CV of 13 % (8–17 %). Detailed information can be found in Table 1. Based on this experiment, we considered steroid concentrations to be stable in the next experiments if the average shift remained within ±15 % from baseline.

Figure 2:

Coefficient of variation (CV) of the 5 spots per volunteer and per steroid hormone (n=10). Individual CVs, median and 95% CI are shown. 17-OHP concentrations of females were below lower limit of quantitation (LLOQ), thus only male data is plotted (n=5). 17-OHP, 17-hydroxyprogesterone.

Table 1:

Between spot variations. Variations in steroid concentrations between five spots, per volunteer (n=10).

	n	Testosterone	17-OHP	Androstenedione	Cortisol	Cortisone	Corticosterone
Median, nmol/L 95 % CI, nmol/L CV, %	1	0.64 0.6–0.65 3	<LLOQ	2.4 2.3–2.4 2	685 662–699 2	52.6 51.1–54.3 2	29.1 28.3–29.6 2
	2	9.6 9.5–11.6 9	2.52 2.48–2.98 8	2.1 1.9–2.3 6	286 267–334 9	45.8 43.1–53.5 9	14.5 13.7–18.5 13
	3	0.15 0.14–0.22 21	<LLOQ	0.81 0.80–0.94 7	129 113–133 7	22.5 20.2–23.5 6	2.8 2.4–2.9 7
	4	7.2 6.8–8.0 7	0.95 0.89–1.22 13	1.9 1.7–2.0 8	201 187–217 6	30.8 27.7–32.8 6	7.3 6.6–7.7 6
	5	0.32 0.27–0.35 10	<LLOQ	1.8 (1.7–2.0) 7	109 102–115 5	31.8 29.1–33.7 6	1.7 1.4–1.9 10
	6	9.8 9.6–14.3 18	1.4 1.3–1.9 17	2.2 2.0–2.9 15	126 115–173 18	17.3 (16.7–23.8) 16	3.3 3.0–4.3 15
	7	0.53 0.50–0.62 9	<LLOQ	3.0 2.9–3.5 8	80 78–92 7	12.3 12.0–15.7 12	1.4 1.3–1.7 10
	8	0.22 0.22–0.25 6	<LLOQ	2.2 2.0–2.2 5	166 150–167 4	20.3 18.4–20.8 5	4.8 4.5–5.4 7
	9	5.9 5.6–7.6 14	1.9 1.4–2.2 16	3.1 3.0–3.8 10	264 251–338 14	29.7 29.2–36.4 10	9.5 9.1–12.1 13
	10	8.7 7.4–9.3 9	1.6 1.5–1.8 9	2.1 1.9–2.3 7	168 146–184 9	23.7 21.2–26.1 9	9.6 8.2–10.6 9
Median CV of the 10 listed samples, %		9	13	7	7	7	10

1. Median concentration and 95 % CI are shown, with coefficient of variation of each volunteer and per steroid. When concentrations were below LLOQ, samples were excluded. LLOQ was 0.14 nmol/L for testosterone, 0.52 nmol/L for androstenedione, 0.7 nmol/L for 17-OHP, cortisone and corticosterone, and 1.4 nmol/L for cortisol. CV, Coefficient of variation (CV=standard deviation/mean*100); 17-OHP, 17-hydroxyprogesterone; LLOQ, lower limit of quantitation.

Experiment 2. Storage at 37 °C

All steroid hormone concentrations remained within ±15 % change from baseline after 2 days of storage at 37 °C (0.5 %, −3%, 3 %, −5%, −11 %, −6%, respectively for testosterone, 17-OHP, androstenedione, cortisol, cortisone, and corticosterone) (Figure 3). Supplemental Table 1 shows the median shifts in nmol/L and percentage shifts compared to baseline per time point.

Figure 3:

Relative levels of testosterone, 17-OHP, androstenedione, cortisol, cortisone, and corticosterone concentrations during short term storage at 37 °C. Percentages are compared to baseline (day 0). Median is shown, error bars indicate IQR. 17-OHP; 17-hydroxyprogesterone.

Median changes of all steroid hormone concentrations measured in DBS at both day 7 and day 14 stayed within ±15 % change, compared to baseline, except for cortisone (day 7: −24 % and day 14: −33 %).

Experiment 3. Short term storage

The % shift of steroid hormone concentrations up to 14 days compared to the baseline for different temperatures are presented in Figure 4.

Figure 4:

Relative levels of testosterone, 17-OHP, androstenedione, cortisol, cortisone, and corticosterone during short-term storage, compared to baseline (day 0). Median is shown, error bars indicate IQR. 17-OHP, 17-hydroxyprogesterone; RT, room temperature.

Testosterone. Median changes of testosterone concentrations in DBS compared to baseline stored at −20 °C and RT stayed within ±15 % change up to 14 days of storage; only at day 7 the samples stored at −20 °C showed a change in concentration of + 22 %. Concentrations in DBS stored at 4 °C exceeded the ±15 % limit after 7 days of storage (−16 %).

17-OHP. Median changes of 17-OHP concentrations in DBS compared to baseline stored at −20 °C, 4 °C and RT stayed within ±15 % change up to 14 days of storage.

Androstenedione. Median changes of androstenedione concentrations in DBS compared to baseline stored at −20 °C and 4 °C stayed within ±15 % change up to 14 days of storage. When stored at RT concentrations remained within ±15 % change up to 7 days of storage but androstenedione concentrations exceeded this after 14 days of storage compared to baseline (19 %).

Cortisol. Median changes of cortisol concentrations in DBS compared to baseline stored at −20 °C, 4 °C and RT stayed within ±15 % change up to 14 days of storage.

Cortisone. Median changes of cortisone concentrations in DBS compared to baseline stored at −20 °C, 4 °C, and RT stayed within ±15 % change up to 14 days of storage.

Corticosterone. Median changes of corticosterone concentrations in DBS compared to baseline stored at −20 °C, 4 °C and RT stayed within ±15 % change up to 14 days of storage.

Detailed results are depicted in Supplemental Table 2, which shows the median shifts in nmol/L and percentage shifts compared to baseline for all storage conditions per time point.

Experiment 4. Long term storage

The % shift of steroid hormone concentrations up to 6 months compared to the baseline steroid concentration for different temperatures are presented in Figure 5.

Figure 5:

Relative levels of testosterone, 17-OHP, androstenedione, cortisol, cortisone, and corticosterone during 6 months storage at −20 °C, 4 °C and room temperature, compared to baseline (day 0). Median is shown, error bars indicate IQR. 17-OHP, 17-hydroxyprogesterone; RT, room temperature.

Testosterone. Median changes of testosterone concentrations in DBS compared to baseline, when stored at −20 °C, 4 °C and RT, stayed within ±15 % change up to 6 months of storage.

17-OHP. Median changes of 17-OHP concentrations in DBS compared to baseline, when stored at −20 °C, stayed within ±15 % change up to 3 months of storage. Stored at 4 °C and RT, concentrations in DBS stayed within ±15 % change up to 6 months.

Androstenedione. Median changes of androstenedione concentrations in DBS compared to baseline, when stored at RT, remained within ±15 % up to 3 months of storage. When stored at −20 °C and 4 °C concentrations in DBS stayed within ±15 % change up to 6 months.

Cortisol. Median changes of cortisol concentrations in DBS compared to baseline, when stored at −20 °C and 4 °C, stayed within ± 15 % change up to 6 months, and when stored at RT only up to 3 months.

Cortisone. Median changes of cortisone concentrations in DBS compared to baseline were more than 15 % after 3 and 6 months of storage at both 4 °C and RT (−15.4 % and −35 %). When stored at −20 °C, concentrations in DBS stayed within ± 15 % change up to 6 months.

Corticosterone. Median changes of corticosterone concentrations in DBS compared to baseline, when stored at −20 °C, 4 °C and RT, stayed within ± 15 % change up to 6 months.

Detailed results are depicted in Supplemental Table 3, which shows the median shifts in nmol/L and percentage shifts compared to baseline for all storage conditions per time point.

Discussion

The present study investigated the stability of steroid hormone concentrations in DBS samples under various temperature conditions for prolonged periods of time. The aim was to simulate possible shipping and storage conditions before enabling DBS sampling for steroid hormones at-home for diagnostic and research purposes. Stability experiments were performed for testosterone, 17OHP, androstenedione, cortisol, cortisone, and corticosterone in DBS. We also aimed to measure 11-deoxycortisol and 21-deoxycortisol, but concentrations were below LLOQ (<0.2 and <1 nmol/L, respectively) in all healthy volunteers. Therefore, no further experiments on these two hormones were conducted, nor reported.

All steroid hormones in DBS samples were stable for short term (up to 14 days) during temperatures between −20 °C and 37 °C, except for cortisone which decreased significantly after 7 days at 37 °C (−24 %). Compared to the other steroid hormones tested, cortisone was also less stable at room temperature (±20 °C) and at 4 °C for the longer term. Cortisol did not remain stable at higher temperatures on the long term either, and seemed to decrease at RT when stored for more than 3 months. Literature describes the decomposition of cortisone and cortisol into 11-ketoandrostenedione and 11-hydroxyandrostenedione, respectively, which occurs faster at higher temperatures and in dry extracts compared to in solution [17]. For testosterone, 17-OHP and androstenedione we observed some variation in the data at several time points of analysis probably due to a combination of analytical variation (which was up to 7.5 %) and between-spot variation (which was up to 13 %), as it was seen randomly at different temperature conditions. Grecso et al. has corroborated the stability of 17-OHP, androstenedione and cortisol concentrations in DBS for up to 1 year at −70 °C, −20 °C, and 4 °C, along with a 4 week stability window at RT [18], using the same criteria as our study: analyte levels within ±15 % change from baseline were considered stable.

The between-spot variation we observed arises most likely from inhomogeneous DBS samples [19], due to the self-sampling of DBS by individuals [20, 21], and the manual punching of the spots during laboratory processing of the DBS samples (not always exactly on the same place in each spot) [20]. Other factors such as humidity at higher temperatures could also affect sample quality [13, 22], for which we mitigated by storing samples in air-tight plastic bags with desiccant. So, probably inhomogeneity of DBS leads to more variation than we are used to observe in serum samples which are, when mixed well, highly homogenous. The (pre)analytical variation and subsequent impact during clinical monitoring of patients could be reduced by measuring duplicate punches from two blood spots. Besides, biological variation of steroid hormone concentrations in DBS could be dealt with by giving patients clear instructions when to sample their blood (for instance between 8 and 9 AM, or before taking their maintenance dosage).

The cut-off of ±15 % change from baseline was based on the between-spot variation but might be too tight for the low concentration range. For instance, relative percentage changes tend to be larger for low testosterone concentrations (in general in samples taken from women), compared to higher concentrations (in samples taken from men). In DBS samples stored at −20 °C, women had a median absolute testosterone concentration of 0.37 nmol/L, while men had a median absolute concentration of 10.7 nmol/L at day 0. An absolute shift of 0.1 nmol/L would result in a 27 % relative shift for women, while for men this would be only a 1 % relative shift. The absolute concentrations in the DBS samples can thereby affect the results of the stability experiments and thus suggests that our cut-off of ±15 % change from baseline could possibly be too tight for the low concentration range. In half of our samples, 17-OHP concentrations were below the LLOQ, as our volunteers were all healthy subjects. This limited our sample size for this group. Despite the influence of these factors and our limited sample size of (maximal) 10 volunteers per group we have achieved reliable results which are attributed to our precise in-house developed and validated LC-MS/MS method [16].

It would be interesting for future research to measure 17-OHP in patients with CAH as they will have higher concentrations, to increase the knowledge regarding 17-OHP and additionally study the stability of 11-deoxycortisol and 21-deoxycortisol in DBS.

To conclude, steroid hormones remain stable at temperatures between −20 °C and up to 37 °C. Thus, self-sampling DBS at home could potentially be used for monitoring steroid hormone levels of patient groups with hormonal dysfunctions, such as pediatric and adult patients with CAH, hypogonadal men and transmen receiving testosterone therapy. Storage up to 6 months is possible at −20 °C for all steroids, and for most steroids at 4 °C or even RT. Storage by −20 °C has the preference if interested in androstenedione, cortisone, and cortisol measurements. Overall, we demonstrated stability of steroid hormone concentrations in DBS under various conditions which may be encountered during shipping to the diagnostic laboratory and during long-term storage before analysis.