Oscar R. Leeuwenkamp, Willem J. F. van der VIJgh,Brigitta C. P. IlUsken, Paul LIps, and J. ..... Milsom S, Ibbertson HK, Hannon S, Shaw D, Pybus J. Simple.
CLIN. CHEM. 35/9, 1911-1914 (1989)
Quantification of Strontium in Plasma and Urine with Flameless Atomic Absorption Spectrometry Oscar R. Leeuwenkamp, Willem J. F. van der VIJgh,Brigitta C. P. IlUsken, Paul LIps, and J. Coen Netelenbos This analytical method for determination of Sr in plasma and urine involves flameless atomic absorption spectrophotometry (FAAS). Drying, charring, and atomization were optimized with respect to temperature, temperature ramp, and duration for Sr in dilute HNO3 and Sr in plasma diluted 20-fold with dilute HNO3. Calibration curves (r >0.995) were linear in the concentration range 5-250 g/L for Sr in various media, with intercepts negligibly small except for the calibration curves in 1:1-diluted plasma and undiluted urine. The estimated detection limits for Sr in 20-fold-diluted plasma and 50-fold-diluted urine were 2 and 3 zgIL, respectively. Endo9enous Sr in plasma and urine was estimated at 16 (SD 8) g/L and 158 (SD 26) zg/L (n = 6), respectively. Intra- and interassay CVs were 9.1% and 5.3% for 20-fold-diluted plasma at a Sr concentration of 25 g/L, and 6.9% and 4.8% at a concentration of 250 g/L. The respective CV5 were 8.2% and 1.2% for 50-fold-diluted urine at the low concentration, and 4.0% and 4.6% at the high concentration. In a pharmacokinetic pilot study of 2.5 mmol of Sr orally administered to a healthy volunteer, the peak plasma concentration of Sr, 4.4 mg/L, decayed bi-exponentially [ti/2,a = 24 h, t1/2.p = 77 h]; the estimated first-order absorption rate constant was 0.005 min; and the observed decay (day 0-6) of the urinary Sr/creatinine ratio closely paralleled the plasma decay [11,2 = 70 h]. AddItional Keyphrases: calcium
bone
urine
pharma-
cokinetics after oral administration
Excessive age-related bone loss (osteoporosis) and associated fractures present a serious public health care problem in the developed countries, affecting about 25% of the female population and 5% to 10% of the male population (1). Osteoporosis is a multifactorial bone disease, but inadequate intestinal absorption of calcium has been identified as a major factor in its pathogenesis (1). For this reason, methods for assessment of the intestinal calcium absorption will be powerful tools in the clinical investigation of osteoporosis. Currently available methods are based upon (a) oral administration of 5-10 Ci of 45Ca or 47Ca (singleisotope method) (2) or (b) simultaneous administration of one of the isotopes orally and the other intravenously (double-isotope method) (2). The radiation exposure makes both of these methods unsuitable for (longitudinal) clinical studies in patients and certainly not in volunteers. Use of 45Ca is limited to individuals older than 40 y because of its long radiation decay half-life of 165 days. Thus, an alternative method involving a nonradioactive and nontoxic substitute for
Department ofInternal Medicine,Clinical Research Laboratory, Free University Hospital, De Boelelaan 1117, 1081 HV Amsterdam, The Netherlands. Received January 12, 1989;acceptedMay 31, 1989.
is desirable. In many biological processes, including in vitro intestinal absorption, Sr behaves like Ca (3-5), suggesting that Sr is a potential substitute for calcium in assessment of intestinal calcium absorption. Recently, Reid et al. (6) and Milsom et al. (7) developed a Sr absorption test as an alternative for the Ca absorption tests currently used. Use of Sr as a nonradioactive substitute for radioactive calcium meant that a procedure for its determination in plasma had to be developed. Currently, Sr can be quantified by conventional flame atomic absorption spectrophotometry (AAS) (8), flameless atomic absorption spectrophotometry (FAAS) (9), inductively-coupled plasma emission spectrometry with a tantalum filament vaporizer (10), direct-current plasma echelle spectrometry (11), and x-ray spectroscopy (12). In contrast to the other techniques, AAS and FAAS are available to the general clinical laboratory. The advantages of FAAS over AAS are its higher sensitivity and its feature of temperature programming. Here we describe a sensitive, selective analytical procedure based upon FAAS for determination of Sr in either plasma or urine. The analytical method developed has been applied to a pilot pharmacokinetic study, and some prelimcalcium
inary
pharmacokinetic
data are reported.
Materials and Methods Materials The analyses were performed with a Perkin-Elmer 5000 graphite-furnace atomic absorption spectrophotometer equipped with an HGA-500 furnace, an AS-40 autosampler, and pyrolytically coated graphite-furnace tubes. Throughout, nitrogen was used as purge gas (300 mL/min, during atomization 50 mL/min) and 20-p.L sample aliquots were injected. SrCl2 61120 (Sigma) and HNO3 (Merck) were of analytical grade. We used doubly distilled water to prepare all solutions. .
Optimization Furnace temperature programming was optimized in the order atomization, charring, atomization, drying. Initially, we studied the effect of the atomization temperature on the measured absorbance by decreasing the atomization temperature step-wise, starting at an atomization temperature of 2700 #{176}C (duration 10 s). During this procedure, drying took place at 120 #{176}C for 25 s and charring at 200 #{176}C for 30 s. The lowest acceptable atomization temperature was determined from the plot of absorbance vs atomization temperature. At this temperature, we then determined the maximum charring temperature at which no analyte was lost, using the absorbance vs charring temperature plot. At this temperature, we examined the effect of the temperature ramp and the duration of the charring step on the absorbance. Thereafter, the atomization temperature was optimized, followed by studying the effect of temperature ramp and duration of atomization on the measured absorCLINICAL CHEMISTRY, Vol. 35, No. 9, 1989
1911
bance. Using the determined optimum conditions for the charring step and the atomization step, we optimized the drying-step variables: temperature, temperature ramp, and duration. This optimization procedure was used to assay 150 zg of Sr per liter of dilute HNO3 (2 mLIL). An identical optimization procedure was used to assay 100 pg of Sr per liter of plasma diluted 20-fold with the dilute HNO3. Pilot Pharmacokinetic Study
Statistical Analysis The interassay SD (between-day variation) and the pooled intra-assay SD (within-day variation) were calculated according to: /k
interassay
SD
(2,
= m=
-
1
k denotes the number of series measured, the overall mean for the k series measured, and m the mean of each series m. where
A male volunteer (age 32 y, body weight 80 kg, creatinine clearance 120 mljmin), who had fasted since midnight, received 2.5 mmol of SrCl2 6H20 (667 mg) in 200 pooled intra-assay SD = mL of distilled water at 0900 hours. He was allowed to eat 4 h after receiving this dose. Blood was sampled into evacuated heparinized tubes 1k i (Terunio, Leuven, Belgium) at preselected times during m)2/(N - k) m=1 n=1 eight days. Plasma was separated by centrifugation (2000 x g, 10 mm) at ambient temperature and stored at -20 #{176}C where N denotes the total number of reference samples until analysis. measured, k the total number of series measured, Xmn the In addition, 24-h urine portions were collected during value measured for reference sample n in series m, m the days one to six, their volumes determined, and aliquots mean for each series, and i the number of reference samples stored at -20 #{176}C. in each series. From the SDs the respective CVs were calculated (CV = (SD/mean value) 100%). Analysis for Sr in Plasma and Urine
‘sJ
Before analysis, slowly
thawed
the plasma and urine
at ambient
temperature.
samples The thawed
were
Pharmacokinetic Analysis
sam-
The curve for concentration (in plasma) vs time was fit to the equation for an orally administered drug with twocompartmental pharmacokinetics and a zero lag-time (13). The pharmacokinetic parameters were calculated by the curve-stripping method in the order terminal half-life (t1,2), t1,, and first-order absorption rate constant (ka). The area under this plasma curve was estimated according to the linear trapezoidal rule (13).
ples were vortex-mixed, and urinary sediments were removed by centrifugation (2000 x g, 10 mm). The plasma and urine samples were diluted with the dilute HNO3. Sr was measured in plasma diluted 20-fold with the dilute HNO3 and 20-giL aliquots of the dilutions were injected for analysis. All samples were measured by the optimized temperature programming (Table 1) in the automatic background correction mode (tungsten/halogen source). All the plasma samples from the pharmacokinetic pilot study were measured concurrently. Sr was measured in urine after 50-fold dilution with dilute HNO3. Apart from the samples, each series contained two sets of calibration standards made in the diluted biological matrix, one at the beginning of the series and the other at the end. All measurements were done in duplicate. To correct for a gradual change in sensitivity caused by the aging of the graphite tube, we calculated the Sr concentrations by interpolation between the slopes and intercepts of both calibration curves, assuming linear changes with time. Urinary creatinine was determined to calculate the urinary Sr/creatinine
ratio.
plasma to monitor renal the time of the study.
Creatinine
function
was also determined
(creatinine
Results and Discussion Optimization In the first step of the optimization procedure we found the minimal atomization temperature to be 2400#{176}C. At this atomization temperature, the background was negligibly
small
in the
charring
temperature
range
studied
(200-2200 #{176}C). The maximum temperature at which no analyte was atomized was 1400 #{176}C (Figure 1). Ramp to and duration of the charring temperature did not clearly affect the absorbance and the reproducibility. The optimum atomization temperature appeared to be 2600 #{176}C (Figure 2).
in
clearance) at
0.5
0.4
Table 1. Temperature Programming of Furnace for
Analysis of Sr
0.2
S
Time, $ step Drying Charring Atomization ‘Bum-out”
Temp., #{176}CRamp 100 30 1400 1 2600 0 2500
1
Hold
Recycle
Base
30 10 5 10
#{149}8 0.1
0.
-5
-6
Conditions: A = 460.7 nm, low slit 0.14 nm, read 5 8 in peak-heightmode, purgegas N2, internalflow300 mL/min and 50 mL/min during atomization.
1912
0.3
C
CLINICAL CHEMISTRY, Vol.35, No. 9, 1989
800 charring
1600 tamp
2400
(#{176}c) Fig. 1. Absorbance vs charring temperature for assaysof 150 ig of Sr per literof diluteHNO3 () and 100 zg of Sr per literof plasma diluted20-foldwithdilute HNO3 (A) Values are the meanofduplicate measurements
0.5
a
A
. U
C 0
0.2 4 0.1 0 2200
2400 Atomization
2600 2800 temp #{176}CJ
Fig.2. Absorbancevs atomizationtemperaturefor assays of 150 pg of Sr per liter of dilute HNO3 () and 100 pg of Sr per liter of plasma diluted 20-fold with the diluteHNO3 (A) Values are the mean ofduplicate measurements To ensure optimal atomization, we set the atomization temperature 200#{176}C above the determined minimal atomization temperature. Using a charring temperature of 1400 #{176}C and an atomization temperature of 2600#{176}C, we obtained maximum absorbance with the ramp in the maximum power mode. A temperature ramp of 1 s decreased the measured absorbance by about twofold. Use of the maximum power mode did not affect the reproducibility. A temperature ramp smaller than 25 s and a drying temperature of 120#{176}C or higher incidentally caused sputtering of the injected solution. Therefore, we chose a ramp of 30 s and a drying temperature of 110 #{176}C. Maximum absorption was obtained with a drying step of 30 s. Essentially the same results were obtained with optimization of the temperature programming parameters for the analysis of 100 pg of Sr per liter of plasma diluted 20-fold with the dilute HNO3. Given this similarity of results, we used identical temperature programming settings for the analysis of Sr in urine. In the furnace temperature program a “bum-out” step was included to remove residual Sr. Because a temperature of 2500#{176}C appeared as effective as 2700 #{176}C, we used the lower temperature to extend tube life. Table 1 summarizes the optimized furnace temperature settings, which were used throughout the measurements. Calibration Curves, Sensitivity, and Detection Umit In dilute HNO3 and in plasma diluted 20-fold with the dilute HNO3, identical linear calibration curves were obtained over a concentration range of 5-250 pg/L, with absorbance (A) = 2.62[Sr] + 0.011 (r >0.995). The sensitivity of the analysis for Sr in 20-fold-diluted plasma, defined as the concentration required to obtain a signal of 0.0044 A, was 1.6 pg/L. The detection limit (3 SD of the blank/slope ratio) for Sr in the dilute HNO3 and 20fold-diluted plasma was 2 pg/L (0.016 A). In urine diluted 50-fold with the dilute HNO3 a linear calibration curve was obtained in the concentration range 5-250 pgfL, A = 2.66[Sr] + 0.022 (r >0.995). The sensitivity and detection limit were 1.6 pgfL and 3 pg/L (0.030 A), respectively. The calibration curves for the diluted plasma and urine specimens were used to determine the Sr in plasma and urine samples from the pharmacokinetic pilot study. We prepared calibration curves for Sr in 1:1-diluted plasma and undiluted urine, so we could estimate the endogenous concentrations of Sr in plasma and urine from the intercepts of these calibration curves. These curves were also linear up to 250 p.git.
Precision and Accuracy The within-day variation (pooled intra-assay variation) and the between-day variation (interassay variation) for the determinations of Sr in the diluted plasma and diluted urine specimens were determined at Sr concentrations of 25 and 250 pg/L. The highest concentration corresponded to the anticipated maximum concentration of about 5 mg of Sr per liter of plasma, subsequent to oral administration of 2.5 mmol of Sr in the phaz-macokinetic pilot study. The low concentration for the diluted plasma samples was simply an order of magnitude lower. The highest concentration in the diluted urine corresponded to the urinary Sr concentration measured in the urine during the first 0-2 h of the pharmacokinetic pilot study (12.5 mg/L). Our results are given in Table 2. The inter- and intraassay variations are satisfactory at the concentrations studied. From the high accuracy and the small intercepts for the calibration lines, as mentioned in the previous paragraph, we conclude that (a) the analysis in the diluted plasma and the diluted urine specimens is free from interference by endogenous compounds, (b) no acid wash procedure is needed to remove contaminants from the labware to be used, and (c) urinary sediments neither adsorb nor entrap significant amounts of Sr. This selective analysis of Sr results from the minimization of atomization interferences through a combination of both matrix modification with dilute HNO3 and application of (optimized) temperature programming, which allows the removal of interfering matrix components before atomization. Endogenous Concentrations of Sr In Plasma and Urine Because endogenous Sr was not detectable in the diluted plasma and urine specimens, we prepared calibration lines from 1:1-diluted plasma and undiluted urine. In these media, linear calibration graphs were obtained with substantial intercepts, most probably attributable to endogenous Sr in these body fluids. The endogenous concentrations of Sr measured in plasma and urine of six healthy individuals were 16 (SD 8) pg of Sr per liter of plasma (range 10-28 pg/L) and 158 (SD 26) pg of Sr per liter of urine (range 113-200 pg/L). These values are consistent with values reported in the literature (14-17) (Table 3). Therefore, it is plausible that the intercepts of the calibration curves in 1:1-diluted plasma and undiluted urine are due to endogenous Sr.
Table 2. PrecIsion and Accuracy for Determinations Sr In Plasma (Diluted 20-Fold) and Urine (Diluted 50-Fold) CV, %
Sr concn, pg/L Prepared
Plasma 25 250 Urine
25 250
of
Mean (SD) maaured
Accuracy,
24.7 (1 .8) 239.5 (11.9)
-1.2
9.1
-4.2
6.9
5.3 4.8
25.1 (2.1)
+0.4
8.2 4.0
4.6
239.7(8.0)
%b
-4.1
lntra-aaaay#{176}interaesaya
1.2
= 16. b Deviationofthemeanconcentration from the theoretical value. #{176}Pooled wIthin-dayvariation (n = 4). d Between-dayvariation (n = 4, k = 4).
CLINICAL CHEMISTRY, Vol. 35, No. 9, 1989
1913
order absorption rate constant (ka) was 0.005 min1. The constant ka might be a valuable pharmacokinetic parame-
Table 3. Concentrations of Endogenous Sr In Plasma and Urine of Healthy Volunteers Concn, pg/L
T.chnlquea
Plasma 16 45 57
This study(n = 6)
FAAS
AES
14 15 16 17
AAS
XRF XRF
10-50
30 Urine 158 150
ter for assessment of intestinal absorption. The area under the plasma concentration-time curve was calculated to be 16800 mg min/L. The decay (day 0-6) of the urinary Sr/creatinine ratio closely paralleled the decrease of the plasma concentration of Sr (t112 = 70 h vs t112 = 77 h). At day six of the pilot study, 46.9 mg Sr was cleared into the urine. Thus the cumulative urinary excretion was 21% of the administered dose. Assuming a bioavailability of 30% (19), we calculated that 71% of the absorbed Sr dose was cleared into the urine in the first six days.
Reference
This study(n = 6)
FAAS AAS
15
AES, atomic emission spectrometry; XRF, photon-inducedx-ray fluorescence.
We conclude that the developed analytical method based on FAAS is a sensitive, precise, and accurate procedure for the assay of Sr in plasma and urine. ical tool can be used in the further
Pharmacokinetics
analyt-
absorption test.
of Oral Sr
References 1. Mundy GR. Bone resorption and turnover in health and disease.
We applied this analytical procedure to plasma and urine samples from the pharmacokinetic pilot study. In the 20-fold and 50-fold dilutions of the plasma and urine samples collected before the oral administration of 2.5 mmol of SrCl2 the concentration of Sr was close to the respective detection limits. Thus we conclude that the analysis is not complicated by contaminants originating from the urine-collection containers and the blood-collection device and that only absorbed Sr was measured in the diluted plasma and urine specimens. Typically, the difference between duplicate measurements did not exceed 2%. Figure 3 shows the plasma concentration-time curve and a plot of the 24-h urinary Sr/urinary creatinine ratio vs time. The maximum plasma concentration of 4.4 mgfL was reached about 4 h after administration. These findings are in agreement with the observations of Reid et al. (6) and Milsom et al. (7). At the time of the peak plasma concentration (t = 4 h), the fraction of the administered dose in plasma and extracellular fluid, assuming that the volume of these body fluids is equal to 15% of the body weight (18), was 0.24. This value lies within the normal range of 0.09-0.30 reported by Milsom et al. (7). Two distinct elimination phases were observed in the plasma concentrationtime curve. Half-lives, calculated with the curve-stripping procedure, were 24 and 77 h, respectively, and the first-
Bone 1987;8:S9-16. 2. Nordin BEC, Horsman A, Aaron J. Diagnostic procedures. In: Nordin BEC, ed. Calcium, phosphateand magnesium metabolism. Edinburgh: Churchill, Livingstone, 1976:489-90. 3. Eisenberg E, Gordan GB. Skeletal dynamics in man measured by nonradioactive strontium. J Clin Invest 1961;40:1809-25. 4. Papworth DG, Patrick G. The kinetics of influx of calcium and strontium into the rat intestine in vitro. J Physiol 1970;210:999-
1020. 5. Spencer H, Li M, Samachson J, Laszlo D. Metabolism of strontium-85 and calcium-45 in man. Metabolism 1960;9:916-25. 6. Reid IR, Pybus J, Lim TMT, Hannon S. Thbertson HK. The assessment of intestinal calcium absorption using stable strontium. Calcif Tissue mt 1986;38:303-5. 7. Milsom S, Ibbertson HK, Hannon S, Shaw D, Pybus J. Simple test of intestinal calcium absorption measured by stable stron-
Br Med J 1987;295:231-4. 8. Delves HT, Sepherd G, Vinter P. Determination of eleven metals in small samples of blood by sequential solvent extraction and atomic absorption spectrophotometry. Analyst 197 1;96:260-2. 9. Powell LA, Tease R. Determination of calcium, magnesium, strontium and siliconin brines by graphite furnace atomic absorp-
tium.
100
10.
N E
This powerful
development of an Sr
1.0.
10
: Dl
1.0
0.1.
#{149}
tion spectrometry. Anal Chem 1982;54:2154-8. 10. Grime JK, Vickers TJ. Determination of lithium in microliter samples of blood serum using flame atomic emission spectrometry with a filament vaporizer. Anal Chem 1975;47:432-3. 11. Berndt H, Mes8erschmidt J. Simultaneous determination of Na, K, Ca, Li, Fe, Cu and Zn in human serum using direct current plasma. Fresenius Z Anal Chem 1981;301:104-.5. 12. Valkovic V. Analysis of biological material for trace elements using x-ray spectroscopy.Boca Raton, FL: CRC Press, 1980:242 p. 13. Gibaldi M, Perrier D. In: Drug and pharmaceutical sciences, Vol. 15, Pharmacokinetics,2nd ed. Marcel Dekker: New York, 1982:45-53, 445-9. 14. MatusiewiczH. Determination ofnatural levelsof lithium and strontium in human blood serum by discrete injection and atomic emission spectroscopy with a nitrous oxide-acetylene flame. Anal Chim Acta 1983;136:215-23. 15. Schroeder HA, Nason AP. Trace-element analysis in clinical chemistry [Review]. Clin Chem 1971;17:461-74.
16. Iyengar GV, Kolliner WE, Bowen HJH, eds. The elemental composition of human tissues and body fluids. Weinheim: Verlag
Chemie, 1978:151 p. 00.1
________________________________________ 2
3
4 tim.
5
6
7
6
0.1 10
(days)
Fig. 3. Semilogarithmic plot of plasma concentration vs time curve (I) and the urinary Sr/creatinine ratio vs time (0) after oral administrationof 2.5 mmolof SrCI2
1914 CLINICAL CHEMISTRY, Vol. 35, No. 9, 1989
17. Diem K, Lentner C. In: Tables scientifique, 7th ed. Basel: Dokumenta Geigy, 1972:823. 18. Nordin BEC, Young MM, Oxby C, Bulusu L. Calculation of calcium absorption rate from plasma radioactivity. Clin Sci
1986;35:177-82. 19. Hart H, Spencer H. Rate of initial entry of 47Caand sa Sr from the intestine into the vascular space. Proc Soc Exp Biol Med
1967;126:365-71.
