| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Briefs |
1 Child and Adolescent Psychiatry, University Medical Center Utrecht, HP A01.468, Postbox 85500, 3508 GA Utrecht, The Netherlands
2 Endocrinology Laboratory, University Medical Center Utrecht, HP KC.03.063.0, Postbox 85090, 3508 AB Utrecht, The Netherlands
3 Department of Psychiatry, University Medical Center Nijmegen, HP 333, Postbox 9101, 6500 HB Nijmegen, The Netherlands
aauthor for correspondence: fax 31-30-2505487, e-mail C.deWeerth{at}psych.azu.nl
The determination of cortisol in saliva has become popular for human research on stress reactions (1)(2)(3)(4)(5). Depending on the sensitivity and reliability of the assays used, the required sample volume varies between 0.025 and 2 mL of saliva (6)(7)(8). Infants and toddlers, however, often produce only small amounts of saliva and are usually sampled by swabbing the mouth with cotton dental rolls (5) or commercial cotton swabs (Salivette; Sarstedt Inc.) (9), or by pipettes or alternative devices that aspirate saliva from the floor of the mouth (10)(11)(12)(13). Cotton rolls must either be centrifuged to obtain saliva (9) or be placed in the barrel of a syringe (needleless), from which the saliva is expressed into a vial by compression of the plunger (5). With these procedures, saliva remaining in the swabs is thus lost for analysis. When we tested seven different types of cotton rolls, we found that, depending on the individual type, 135–450 µL of saliva could not be centrifuged from the rolls.
Oral stimulants (such as presweetened Kool-Aid crystals) can increase saliva production, but they affect the concentration of cortisol (14). Finally, in the case of Salivettes, the material covering the cotton swab is hard and makes sampling unpleasant.
In this report, we present a new method that uses soft cotton swabs without hard covering material and solvent extraction of cortisol from saliva in the cotton.
Saliva was collected from volunteers in the laboratory and from infants and toddlers participating in studies on cortisol and behavior. Volunteers and the parents of the infants gave informed consent. These studies had been approved by the Medical Ethical Committee of the University Medical Center Utrecht. After collection, either direct or with use of cotton rolls, the samples were stored in closed containers at -20 °C for periods of up to several weeks. We placed 4-cm cotton rolls with a diameter of 8 mm (article no. 900-2005; Henri Schein) individually in disposable 5-mL syringes (PE + PP; Becton Dickinson), closed the syringes with a small plastic cap, and weighed them. For the saliva collection, the cotton roll was taken out of the syringe and the child’s mouth was swabbed by introducing one end of the cotton roll into the buccal cavity. The experimenter moved the roll in the child’s mouth, trying to induce sucking. To obtain as much saliva as possible, after 1–2 min, the experimenter took the roll out of the child’s mouth, turned it around, and introduced the dry end into the child’s mouth. After an additional 1–2 min, the cotton roll was put back in the syringe. The syringe was stored in the dark at -18 to -20 °C and later transported to the laboratory where it was once again weighed. The increase in weight was caused by the amount of saliva on the cotton, 1 mg being equivalent to 1 µL of saliva.
When the volume of saliva was 50–250 µL, cortisol was extracted from the cotton by opening the syringe at both sides and rinsing the cotton roll in the syringe with 1 mL of 960 mL/L ethanol, followed by centrifugation of the syringe at 1500g for 5 min. The resulting liquid was evaporated, and when the volume of saliva was <0.1 mL or the volume was equivalent to the volume of saliva collected, the residue was dissolved in 100 µL of 0.01 mol/L phosphate-buffered saline (pH 7.0) containing 2 g/L bovine serum albumin. After the solution had stood for at least 15 min with repeated mixing with a vortex-mixer, 25 µL was used for the measurement of cortisol by RIA (15).
Direct measurements of cortisol in saliva required 25 µL of saliva. The detection limit of the direct assay was 0.5 nmol/L; the within-assay imprecision (CV) was 4% at 10 nmol/L (n = 10), and the between-assay CV was 9% at 4 nmol/L (n = 69) and 5% at 10 nmol/L (n = 69).
Over an experimental range of 2–20 nmol/L cortisol, concentrations measured after extraction were highly comparable to those measured directly. At the lowest volumes, the imprecision of the measurements increased, but the concentrations with and without extraction did not differ significantly: with solvent extraction they were 117% (22%), 98% (12%), and 107% (11%) of the values measured directly for volumes of 50, 100, and 200 µL, respectively (n = 24 at each volume).
Methanol, isopropanol, and acetone were also tested as possible alternative solvents, but the highest recoveries (>95% of added cortisol) were consistently obtained with ethanol.
The smallest amount of saliva that could be used for the extraction technique with reliable results was evaluated by calculating the mean cortisol concentrations obtained for a series of 459 samples collected in 1- to 3-year-old children screened for possible developmental problems. The absolute amounts of saliva per sample ranged from 10 µL to 1057 µL. The cortisol concentrations were then sorted by the volumes of collected saliva, from smallest to largest, and the running means for 25 consecutive cortisol values were calculated (Fig. 1
). The means (SD) were significantly higher for volumes <50 µL [13.7 (7.0) nmol/L; n = 95] than for samples >50 µL [8.6 (3.6) nmol/L; n = 364; P <0.0001, unpaired t-test with Welch correction, not assuming equal variances].
|
In addition, we calculated the imprecision for 100- and 50-µL samples. For three 100-µL samples from healthy adults, the interassay CV for the individual samples was 3.4–8.5% (n = 5–16). For 50-µL samples, the CV was 19% (n = 24), close to the limits of precision for the RIA. Therefore, only samples >50 µL were used.
To test the stability of cortisol in the cotton at room temperature and during storage in a refrigerator or freezer for periods of up to 3 days, we absorbed known amounts of saliva on the cotton rolls in syringes, closed the syringes, and stored them in the dark. At a concentration range of 7–11 nmol/L, we detected no significant changes in cortisol concentrations after 3 days at room temperature (
20 °C) with saliva volumes of 50–400 µL. Taking all volumes together, the mean concentration after storage was 100% (9.9%) of that in samples stored directly in the freezer (n = 12 at each of the three temperatures). In practice, all samples collected at home were either stored in the freezer (-20 °C) or in the refrigerator for a maximum of 3 days. The stability of cortisol in saliva at -20 °C was excellent: aliquots of pooled samples of saliva used for internal quality assessment for >1 year yielded CVs between 5.1% at 3.8 nmol/L and 3.4% at 19.1 nmol/L (n = 24).
For some assays (16), contamination of saliva with milk affects the measured cortisol. We tested samples of human milk (n = 8) and eight different formula milk preparations at a 1:10 (100 µL of milk or formula and 900 µL of saliva) dilution. At these concentrations, formula milk had no effect on the concentration of cortisol measured, the mean values being 99% (2.4%) of the original saliva concentrations, whereas high amounts of human milk caused a slight decrease in cortisol values, to 91% (6%) of the original concentrations.
As an illustration, the technique as described has been used to evaluate the effects of a known stressor for infants of 11 days, i.e., a standardized physical examination (17) that takes 20 min. Saliva samples from 114 healthy, normally developing infants were collected before (representing basal cortisol) and 40 and 60 min after the start of the examination (representing stress and recovery cortisol, respectively). Salivary sample sizes were 23–1025 µL of saliva; 2% of the samples were <50 µL (and were therefore not included), and 45% were 50–200 µL. During stress, cortisol increased significantly, from 10.5 (5.0) to 17.4 (9.2) nmol/L (Wilcoxon signed-ranks test, z = -6.2; P <0.001), decreasing significantly after the examination, to 14.6 (7.4) nmol/L (z = -5.2; P <0.001).
Although we elected to use our in-house RIA, it is conceivable that the quantitative extraction technique may be readily adapted for cortisol determinations with other, more sophisticated, commercially available methods.
In practice, volumes collected on the cotton rolls varied between 10 and 1057 µL of saliva with the percentage of samples with a volume <200 µL varying between 20% and 80%, depending on the experience of the sampler, the sampling population, and other factors. For example, in the data set used to estimate the smallest amount of saliva that can be used, 45% of the 459 samples contained 50–100 µL of saliva and 22% of the samples contained <50 µL. The extraction method enables analysis of low-volume samples that cannot be assayed by established methods, substantially increasing the number of samples for which results can be obtained; for example, the 58% of samples in our example data set that had volumes of 50–250 µL.
In conclusion, the new method enables reliable measurements of cortisol in small amounts of saliva (as are often obtained from young infants and toddlers) and thus avoids the use of saliva production stimulants and loss of data because of insufficient sampling volume. The method can be used for any clinical or research purpose in which cortisol assessments in saliva are required.
Acknowledgments
This research was supported by Grant 575-25-009 from the Netherlands Organization for Scientific Research (NWO).
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
