JOURNAL OF BONE A N D MINERAL RESEARCH Volume 5, Number 8, 1990 Mary Ann Liebert, Inc., Publishers
Fluoride Bioavailability from Slow-Release Sodium Fluoride Given with Calcium Citrate CHARLES Y.C. PAK, KHASHAYAR SAKHAEE, CAROL PARCEL, JOHN POINDEXTER, BEVERLEY ADAMS, ABDILLAHI BAHAR, and ROBERT BECKLEY
ABSTRACT Clinical pharmacology of slow-release sodium fluoride given with calcium citrate was examined in acute and long-term studies. Following a single oral administration of 50 mg slow-release sodium fluoride, a peak serum fluoride concentration (Cmax)of 184 ng/ml was reached in 2 h; thereafter, serum fluoride concentration declined with a T,,, of 5.9 h. The concurrent administration of calcium citrate (400 mg calcium) gave an equivalent Tmax(time required to attain C,,,) and T,,,, but a lower C,,, of 135 ng/ml. The coadministration of a meal with fluoride also reduced C,,, but increased T,,,. The area under the serum concentration curve of slow-release sodium fluoride was reduced 17-27% by a meal or calcium citrate. Thus, calcium citrate reduced fluoride absorption and peak fluoride concentration in serum of slow-release sodium fluoride but did not affect the time required to reach peak concentration or the rate of subsequent decline. The effect of a meal was similar, except for a longer period required to reach peak serum concentration. During longterm administration of 25 mg slow-release sodium fluoride coadministered with 400 mg calcium as calcium citrate on a twice daily schedule, the trough level of serum fluoride could be kept between 95 and 190 ng/ml, believed to be the therapeutic window.
INTRODUCTION
T
of fluoride to stimulate bone formation"," is the basis for the use of sodium fluoride in the treatment of osteoporosis.'1 4 L A major drawback to this treatment, however, is the potential for the formation of abnormal bone with defective mechanical properties."' Although exact mechanisms for this disturbance have not been established, an abnormally high circulating concentration of fluoride'6)and an inadequate calcium absorption have been implicated. ('I Fluorosis of bone and teeth has been described when serum concentration of fluoride exceeds 190 ng/ml, believed to be the toxic threshold.(R-'O) A defective mineralization of newly formed bone matrix, or osteomalacia, has been reported during fluoride therapy when calcium supplementation is inadequate. ('I1 H E W E L L -KN OW N ABILITY
To overcome these problems, we have advocated the use of sodium fluoride in a slow-release formulation and calcium supplementation with calcium itr rate.'^' In a previous report,'"' the slow-release sodium fluoride given with a standard breakfast was shown to produce a slower rise as well as a lower rate of subsequent decline in serum fluoride concentration when compared to the action of a rapid-release preparation. During long-term treatment with slowrelease sodium fluoride, the trough concentration of serum fluoride could be kept between 95 and 190 ng/ml,I4 ' I ) believed to be the therapeutic window.[61 In several reports,(13,141 calcium citrate was found to be more absorbable than calcium carbonate when taken on an empty stomach. The treatment format, adopted for simplicity and assurance of optimum compliance, was one in which slow-release sodium fluoride was given together with calcium
Center for Mineral Metabolism and Clinical Research, Southwestern Medical School, University of Texas Southwestern Medical Center at Dallas, 5323 Harry Hines Blvd., Dallas, TX 7 5 2 3 5 4 8 8 5 ,
857
PAK ET AL.
858
citrate orally on an empty stomach on a twice daily schedule.‘“) The clinical pharmacology of this format was never described. In this communication, we report pharmacokinetics of fluoride appearance in serum following a single oral administration of slow-release sodium fluoride and changes induced by calcium citrate and a meal.
MATERIALS AND METHODS
Study I . A single-dose bioavailability study with slow-release sodium fluoride: Effect of calcium citrate or meal given concurrently A group of four postmenopausal women (48-62 years, mean age 5 5 years) and eight premenopausal women (2542 years, mean age 32 years) participated in the study. They were free of peptic ulcer disease or significant renal impairment (endogenous creatinine clearance < 40 mi/ minute). They participated in four phases of study chosen in random order (assignment made by an independent statistician). Each phase lasted 4 days. During phase I (sodium flouride or NaF alone), 50 mg slow-release sodium fluoride (two tablets of Slow Fluoride, Mission Pharmacal Co.) was given orally on an empty stomach at 8 a.m. of the test day, with breakfast withheld. During phase 11 (NaF with calcium), 50 mg slow-release sodium fluoride was given with 400 mg calcium as calcium citrate (Citracal, Mission Pharmacal Co.) and breakfast was withheld. During phase I11 (NaF with meal), 50 mg slow-release sodium fluoride was given with a standard breakfast. The breakfast comprised 200 ml orange juice, one poached egg, two slices of toast with apple jelly, 1/2 cup cream of wheat, 1 cup decaffeinated coffee, and 12 g cheese; it contained 200 mg calcium and 260 mg phosphorus. During phase IV (placebo), two placebo tablets of identical appearance and size as slow-release sodium fluoride containing excipient only (also prepared by Mission) were given on an empty stomach without breakfast. Although a typical single dose of slow-release sodium fluoride is 25 mg,(41a higher dose of 50 mg was utilized here in a single-dose bioavailability study to maximize changes between phases. On days 1-3 of each phase, subjects were stabilized at home on a limited fluoride intake by avoidance of fluoride-containing toothpaste and tea and by drinking distilled water instead of fluorinated water. They were admitted to the general clinical research center (GCRC) on the afternoon of day 3, to be begun and kept on a constant metabolic diet (800 mg P, 800 mg Ca, and 100 mEq sodium per day). O n day 4 at 8:OO a.m., subjects received respective medications outlined before for the four phases. Breakfast was withheld during phases 1, 11, and IV. Lunch wa provided at noon and dinner at 5:00 p.m. Venous blood samples ( 5 ml each) were obtained at 8:00 a.m. (before administration of the test dose), 9:OO a.m., 10:00 a.m., 11:00 a.m., 12 noon, 2:OO p.m., 4:OO p.m., 6:00 p.m., and 8:OO p.m. Each sample was assayed for fluoride. Distilled water was drunk at the following schedule: 300 ml each at 8:OO p.m. of day 3, midnight, 8:00 a m . , 1O:OO a.m., noon, and 5:00
p.m. Other fluids (made with distilled water) except tea (in regulated constant volumes) were allowed. In each subject, the identical meals and fluids were consumed during the four GCRC admissions. There was no rest period between phases, because 3 days of stabilization in each phase was considered sufficient to remove the effect of the previous fluoride dose.
Study 2. Effect of calcium citrate given before or after slow-release sodium fluoride on fluoride bioavailability A total of nine subjects from the previous study participated in this investigation: three were postmenopausal women (48-62 years, mean age 54 years) and six were premenopausal women (25-42 years, mean age 32 years). They underwent two additional phases of study, chosen in random order. In phase V (NaF with Ca before), 50 mg slow-release sodium fluoride was given at 8 a.m. and 400 mg calcium as calcium citrate taken at 6 a.m. In phase VI (NaF with Ca after), 50 mg slow-release sodium fluoride was provided at 8 a.m. and 400 mg calcium as calcium citrate was given at 9 a.m. Breakfast was withheld. Dietary preparation, fluid intake, and tests were the same as in study 1.
Study 3. Effect of long-term treatment with slow-release sodium fluoride on serum fluoride A group of 52 patients with primary osteoporosis participated in the study. There were 11 men and 41 women (mean age 60.2 years, 27-84 years). They participated in an intermittent treatment of slow-release sodium fluoride, comprised of repeated cycles of 13 months each. In each cycle, they took 25 mg slow-release sodium fluoride together with 400 mg calcium as calcium citrate on an empty stomach before breakfast and again at bedtime during the first 12 months. During the last month of each cycle, they took calcium citrate only and sodium fluoride was withheld. A venous blood sample was obtained in the fasting state before the morning dose of slow-release sodium fluoride, before and every 3 months of treatment in each cycle, and analyzed for fluoride.
Analytic methods Serum fluoride was determined by the method of Ekstrand using a fluoride To prevent fluoride contamination, acid-washed glassware or fluoride-free plastic containers were used during sample collection and analysis.
Statistical methods Classic pharmacokinetic analysis‘”’ was employed for each study of fluoride bioavailability, broadly defined here as the rate and extent of fluoride entry into the systemic circulation following oral administration. For each bioavailability study, the peak serum fluoride concentration
859
FLUORIDE BlOAV AILABILITY (C,,,), peak time ( Tmax),serum fluoride half-life ( T , , , ) , and area under the curve (AUC) were calculated. The T,,, gave the time in hours (after fluoride dose) when the peak serum fluoride concentration was obtained. The T , / 2represented the half-life of the decline in serum fluoride concentration from its peak to the value at 12 h. The rate o f decline was calculated by nonlinear regression analysis. The AUC represented the area under the serum concentration curve from time zero to 12 h later in units of ng h/ml. It was calculated using the trapezoid rule. The AUC provided a reflection of the overall fluoride absorption from the gastrointestinal tract. However, it is not identical to total fluoride absorption because some of absorbed fluoride is taken up by bone and excreted in urine. A more direct measure of fluoride absorption from a test substance may be obtained by comparing its AUC with the AUC of a substance believed to be completely absorbed (for example, plain sodium fluoride given alone). For variables Cmax, T,,,, T , , , , and AUC, a single-factor repeated measures analysis of variance was performed to assess differences among the phases. The Student-NewmanKeuls multiple comparisons were used for pairwise comparisons if the analysis of variance was significant at (Y = 0.05. For C,,,, Tmax,TI/*,and AUC, the placebo phase was omitted from the analysis. For assessing the significance of difference in serum fluoride concentration from a single-dose bioavailability study, a preliminary two-factor repeated measures analysis of variance was performed, the two repeated factors being phase and time of sample. Since the phase x time interaction was significant O-, < O.OOOl), a single-factor analysis of variance (with repeated factor phase) was performed at each time. The Student-Newman-Keuls multiple comparisons were used for pairwise comparisons if the analysis of variance was significant at CY = 0.05. For serum fluoride concentrations during long-term treatment, the significant difference in serum fluoride during treatment from pretreatment was assessed by the paired /-test. The Bonferroni inequality was used to adjust for multiplicity of testing when appropriate.
RESULTS Serum fluoride concentration curves following a single oral dose of slow-release sodium fluoride were indistinguishable between premenopausal women and postmenopausal women. When 50 mg slow-release sodium fluoride was given with 400 mg calcium as calcium citrate, serum fluoride concentrations during subsequent 12 h were not significantly different between four postmenopausal women and eight premenopausal women (Fig. I ) . Thus, results from all subjects were combined. ESJect of culciutn and meal ,fluoride hiouvuiluhili~~v
on
Following an oral administration of the placebo medication, \erum fluoride concenlration remained low and unchanged during t h e wbwquent 12 h (big. 2). When FIOW-
-
E 150
W (I)
i
d-Premenopausal
- ~-Postmenopausal
A
joli l 0
0
2
1
I
1
1
I
4
6
8
10
12
Time, hours
t
FIG. 1. Serum fluoride concentration after a single dose of 50 mg slow-release sodium fluoride given with 400 mg calcium as calcium citrate in four postmenopausal and eight premenopausal women. The arrow indicates the time of fluoride administration.
-
2oo
E
-
1
-o-a-
2 150
NaF Alone NaF with Calcium NaF with Meal Placebo N = 12
m
0
g
lL
100
E
2
m
50
(I)
0
1
L
I
0 f
2
I
I
I
6 8 Time, hours 4
I
I
10
12
FIG. 2. Serum fluoride concentration after a single dose of 50 mg slow-release sodium fluoride, given alone, with 400 mg calcium as calcium citrate or with a meal in 12 subjects. The arrow indicates the time of fluoride administration. (a) I-, < 0.05 between NaF alone and NaF with calcium or NaF with meal; (b) y < 0.05 between NaF with calcium and NaF with meal. release sodium fluoride was administered, alone or with calcium citrate or a meal, serum fluoride concentration was significantly higher than the corresponding placebo value during all 12 h of the study (p < 0.05). However, the serum fluoride concentration in phase I1 (NaF with calcium) was significantly lower than in phase I (NaF alone), except at 12 h. The serum fluoride concentration during the first 4 h of NaF with a mean (phase 111) was significantly lower than during the corresponding time period of NaF alone. Compared to NaF with calcium, NaF with a meal gave a significantly lower serum fluoride concentration at 1 and 2 h and higher values at 6, 8, 10, and 12 h.
860
PAK ET AL.
During NaF with calcium, the C,,, and AUC were significantly lower than during NaF alone (Table 1). However, Tmaxand T , , , were not different. For NaF with a meal, the C,,, and AUC were also significantly reduced, compared to NaF alone. However, T,,, was significantly higher. Between phase I1 (NaF with calcium) and phase 111 (NaF with meal), T,, was significantly higher for phase 111.
Effect of calcium given before and after slo w-release sodium fluoride Compared to NaF alone, no significant difference in serum fluoride concentration, Cmax,Tmax,T L , , ,or AUC was noted when calcium citrate was given 2 h before or 1 h after slow-release sodium fluoride (Fig. 3). The AUC for NaF alone was 1302 f 257 ng h/ml, which was not significantly different from 1187 + 378 ng h/ml for NaF with Ca before or 1144 f 253 ng- h/ml for NaF with calcium after.
Serum fluoride concentration during long-term treatmen t Before and after 1 month of withdrawal of slow-release sodium fluoride therapy (0, 13, 26, and 39 months), serum fluoride concentration (fasting value, before the morning dose of fluoride) was below 95 ng/ml (Fig. 4). During treatment with slow-release sodium fluoride (3, 6, 9, 12, 16, 19, 22, 25, 29, 32, 35, and 38 months), serum fluoride concentration was mostly above 95 ng/ml and below 190 ng/ml.
DISCUSSION This study examined the effect of calcium and a meal on fluoride bioavailability. The concurrent administration of calcium citrate with slow-release sodium fluoride reduced the peak fluoride concentration in serum and fluoride absorption (reflected by area under the curve), each by 27%,
without affecting the time required to reach peak concentration or the subsequent rate of decline in fluoride concentration. This effect of calcium was negligible when calcium citrate was given 2 h before or 1 h after fluoride administration. The ingestion of a meal delayed the time required to reach the peak fluoride concentration in serum, possibly because of prolonged gastric emptying caused by food. However, a similar reduction in peak fluoride concentration and in fluoride absorption was observed, as with calcium. The conditions of this study may be different from the usual clinical setting when fluoride is given long term to postmenopausal women with osteoporosis. However, the following considerations suggest that the two situations are comparable. First, the fluoride bioavailability was indistinguishable between postmenopausal and premenopausal women. Second, the acute effect of a single dose of slowrelease sodium fluoride (50 mg) in untreated patients described in this paper is qualitatively similar to the effect of 25 mg slow-release sodium fluoride in patients maintained on long-term treatment with fluoride. In this acute study, the increment in serum fluoride concentration from basal to peak level was 101 ng/ml when 50 mg slow-release fluoride was given with 400 mg C a as calcium citrate. In a prea single dose of 25 mg slow-release sodium vious fluoride with a meal in untreated patients raised serum fluoride concentration by 40 ng/ml. In the same study,"" the difference between the trough and peak levels of fluoride in serum was 39 ng/ml in patients maintained on long-term treatment with slow-release sodium fluoride (25 mg) with a meal twice daily. The fluoride absorption from a single dose of 50 mg slow-release sodium fluoride (given with a meal), estimated from the difference in the area under the curve from the placebo phase, was approximately twice that of 25 mg slow-release sodium fluoride given (with a meal) as a single dose or as a part of long-term treatment.llzlThus, fluoride bioavailability did not seem to be affected substantially by previous treatment with slow-release sodium fluoride. During long-term treatment with 25 mg slow-release SOdium fluoride given together with 400 mg calcium as cal-
TABLE l . EFFECTOF CALCIUM CITRATE OR MEALGIVEN CONCURRENTLY ON FLUORIDE BIOAVAILABILITY~ Phase
I NaF alone Cm,, ng/ml AUC, ng h/ml Tmax, h Tit,,
h
184 + 33 1314 f 231 2.0 f 1.0 5.9
f
1.0
II NaF with calcium
135 i 24b 961 i 157b 1.6 f 0.7c 7.1 f 1.3
aAll values represent mean + SD. N = 12. bp < 0.05 between phases I and I1 and between I and 111. cp < 0.05 between phases 11 and 111.
III
IV
NaF with meal
Placebo
136 1096 3.6 13.2
f f f f
33b 241b 2.0b 20.4
-
-
FLUORIDE BIOAVAILABILITY
2oo
861
r
-+-
NaF Alone NaF t Ca Before + NaF t Ca Alter
E cn 150 c
i E & 3 L
N = 9
100
E2 l 50
Q)
(1
v)
oL
I
1
I
I
6 0 Ti me, hours
2
0
4
t
I
I
10
12
FIG. 3. Serum fluoride concentration following a single dose of 50 mg slow-release sodium fluoride, given alone, with 400 mg calcium as calcium citrate given 2 h before (NaF + Ca before) or 1 h after (NaF + C a after) fluoride administration in nine subjects. The arrow indicates the time of fluoride administration.
-
. E
300 250
The fluctuation in serum fluoride concentration from basal to peak levels was about 40 ng/ml."zl This discussion provides justification of our general therapeutic guidelines with slow-release sodium fluoride. The simplest format recommended initially is the combined use of slow-release sodium fluoride with calcium citrate on an empty stomach. If this format is unacceptable or causes gastrointestinal side effects, slow-release sodium fluoride can be given with a light meal without sacrificing fluoride absorption. It is unlikely that treatment with an immediate-release sodium fluoride (25 mg twice or three times a day) could maintain the serum fluoride concentration between 95 and 190 ng/ml without a wide circadian fluctuation, as has been shown here for slow-release sodium fluoride (25 mg twice a day). A pharmacokinetic analysis of the previous acute fluoride bioavailability study'"' revealed a much higher peak fluoride concentration in serum following administration of a rapid-release sodium fluoride (25 mg) than after taking the same dose as slow-release sodium fluoride (220 versus 84 ng/ml, p < 0.05). Fluoride absorption, estimated from the difference in area under the curve from that of the placebo phase, was more than twofold higher for the rapid-release sodium fluoride preparation."*' In our unpublished study of patients maintained on long-term treatment with slow-release sodium fluoride (25 mg) with calcium citrate (400 mg Ca) twice a day, serum fluoride concentration could be kept between 95 and 190 ng/ml during 12 h following 25 mg slow-release sodium fluoride with 400 mg Ca as calcium citrate. However, serum fluoride concentration reached a peak level of 454 ng/ml 1 h after oral administration of plain sodium fluoride (25 mg) with 500 mg Ca as calcium carbonate. In summary, the oral administration of slow-release sodium fluoride with calcium citrate averted the attainment of sharp high peaks of fluoride in serum and produced a slow gradual decline in serum fluoride. Thus, the convenient treatment format of 25 mg slow-release sodium fluoride and 400 mg calcium as calcium citrate given together on an empty stomach on a twice daily schedule assured maintenance of serum fluoride within 95-190 ng/ml, believed to be the therapeutic window, with a modest fluctuation between trough and peak levels.
1
m c 0
0 0
a LL
E
2
Q u)
I
0
6
12
I
19
25
I
32
38
Month
FIG. 4. Fasting (trough) concentration of serum fluoride during long-term administration of slow-release sodium fluoride (25 mg twice a day) and calcium citrate (400 mg calcium twice a day). Sodium fluoride treatment was interrupted for 1 month after each 12 months of treatment, as indicated by vertical dotted lines. Significant difference from the pretreatment value is indicated by (a) p < O.ooO1; (b) p = 0.0001-0.003, and (c) p = 0.05-0.003. Numbers on graph depicts sample size. X = insufficient data for statistical analysis. Bars indicate mean f SEM.
ACKNOWLEDGMENTS Supported by USPHS grants R01-AR16061, M01RR00633, and POI-DK20543.
REFERENCES cium citrate on an empty stomach on a twice daily schedule, the trough levels of serum fluoride could be kept within 95-190 ng/ml. In a previous report,'") serum fluoride was also shown to be maintained within 95-190 ng/ml during long-term treatment with 25 mg slow-release sodium fluoride taken with breakfast or a snack twice a day.
1. Farley JR, Wergedal JE, Baylink DJ 1983 Fluoride directly stimulates proliferation and alkaline phosphatase activity of bone-forming cells. Science 222:330-332. 2. Zerwekh JE, Morris AC, Padalino PK, Gottschalk F. Pak CYC Fluoride and parathyroid hormone raises intracellular calcium in human osteoblasts. J Bone Min Res 5:S131-S135. 3. Riggs LB, Seeman E, Hodgson SF, Taves DR, O'Fallon MW
862
4.
5.
6. 7.
8. 9.
10.
11. 12.
PAK ET AL. 1982 Effect of the fluoride/calcium regimen on vertebral fracture occurrence in postmenopausal osteoporosis. N Engl J Med 306:446-450. Pak CYC, Sakhaee K , Zerwekh JE, Parcel C , Peterson R, Johnson K 1989 A safe and effective treatment of primary osteoporosis by intermittent application of slow-release sodium fluoride: Augmentation of vertebral bone mass and inhibition of fractures. J Clin Endocrinol Metab 68:150-159. Wolinsky I, Simkin A, Guggenheim K 1972 Effects of fluoride on metabolism and mechanical properties of rat bone. Am J Physiol 223(1):46-50. Taves DR 1970 New approach to the treatment of bone disease with fluoride. Fed Proc 29(3):1185-1187. Guggenheim K , Simkin A, Wolkinsky I 1976 The effect of fluoride on bone of rats fed diets deficient in calcium or phosphorus. Calcif Tissue Res 22:9-17. Angmar-Mansson 8, Whitford GM 1984 Enamel fluorosis related to plasma F levels in the rat. Caries Res 18:25-32. Suttie JW, Faltin EC 1971 Effect of a short period of fluoride ingestion on dental fluorosis in cattle. Am J Vet Res 32: 217-222. Harrison JE, Hitchman AJW, Hasany SA, Hitchman A, Tam CS 1984 The effect of diet calcium on fluoride toxicity in growing rats. Can J Physiol Pharmacol 62:259-265. Riggs LB, Jowsey J 1972 Treatment of osteoporosis with fluoride. Semin Drug Treatment 2:27-33. Pak CYC, Sakhaee K , Gallagher C, Parcel C, Peterson R, Zerwekh JE, Lemke M, Britton F, Hsu MC, Adams B 1986 Attainment of therapeutic fluoride levels in serum without
13.
14.
15.
16.
17.
major side effects using a slow release preparation of sodium fluoride in postmenopausal osteoporosis. J Bone Min Res 1: 563-571. Nicar M J , Pak CYC 1985 Calcium bioavailability from calcium carbonate and calcium citrate. J Clin Endocrinol Metab 61:391-393. Harvey JA, Zobitz MM, Pak CYC 1988 Dose dependency of calcium absorption: A comparison of calcium carbonate and calcium citrate. J Bone Min Res 3:253-285. Pak CYC, Sakhaee K, Zerwekh J E Effect of intermittent therapy with a slow release fluoride preparation. J Bone Min Res 5:s 149-S 155. Ekstrand J 1977 A micromethod for the determination of fluoride in blood plasma and saliva. Calcif Tissue Res 23: 225-228. Ekstrand J , Alvan G , Boreus LO, Norlin A 1977 Pharmacokinetics of fluoride in man after single and multiple doses. Eur J Clin Pharmacol 12:311-317.
Address reprint requests to: Dr. Charles Y.C. Pak Center for Mineral Metabolism and Clinical Research Southwestern Medical School University of Texas Southwestern Medical Center at Dallas 5323 Harry Hines Blvd. Dallas, TX 75235-8885 Received for publication October 19, 1989; in revised form March 26, 1990; accepted March 26, 1990.
Fluoride Bioavailability from Slow-Release Sodium Fluoride Given with Calcium Citrate CHARLES Y.C. PAK, KHASHAYAR SAKHAEE, CAROL PARCEL, JOHN POINDEXTER, BEVERLEY ADAMS, ABDILLAHI BAHAR, and ROBERT BECKLEY
ABSTRACT Clinical pharmacology of slow-release sodium fluoride given with calcium citrate was examined in acute and long-term studies. Following a single oral administration of 50 mg slow-release sodium fluoride, a peak serum fluoride concentration (Cmax)of 184 ng/ml was reached in 2 h; thereafter, serum fluoride concentration declined with a T,,, of 5.9 h. The concurrent administration of calcium citrate (400 mg calcium) gave an equivalent Tmax(time required to attain C,,,) and T,,,, but a lower C,,, of 135 ng/ml. The coadministration of a meal with fluoride also reduced C,,, but increased T,,,. The area under the serum concentration curve of slow-release sodium fluoride was reduced 17-27% by a meal or calcium citrate. Thus, calcium citrate reduced fluoride absorption and peak fluoride concentration in serum of slow-release sodium fluoride but did not affect the time required to reach peak concentration or the rate of subsequent decline. The effect of a meal was similar, except for a longer period required to reach peak serum concentration. During longterm administration of 25 mg slow-release sodium fluoride coadministered with 400 mg calcium as calcium citrate on a twice daily schedule, the trough level of serum fluoride could be kept between 95 and 190 ng/ml, believed to be the therapeutic window.
INTRODUCTION
T
of fluoride to stimulate bone formation"," is the basis for the use of sodium fluoride in the treatment of osteoporosis.'1 4 L A major drawback to this treatment, however, is the potential for the formation of abnormal bone with defective mechanical properties."' Although exact mechanisms for this disturbance have not been established, an abnormally high circulating concentration of fluoride'6)and an inadequate calcium absorption have been implicated. ('I Fluorosis of bone and teeth has been described when serum concentration of fluoride exceeds 190 ng/ml, believed to be the toxic threshold.(R-'O) A defective mineralization of newly formed bone matrix, or osteomalacia, has been reported during fluoride therapy when calcium supplementation is inadequate. ('I1 H E W E L L -KN OW N ABILITY
To overcome these problems, we have advocated the use of sodium fluoride in a slow-release formulation and calcium supplementation with calcium itr rate.'^' In a previous report,'"' the slow-release sodium fluoride given with a standard breakfast was shown to produce a slower rise as well as a lower rate of subsequent decline in serum fluoride concentration when compared to the action of a rapid-release preparation. During long-term treatment with slowrelease sodium fluoride, the trough concentration of serum fluoride could be kept between 95 and 190 ng/ml,I4 ' I ) believed to be the therapeutic window.[61 In several reports,(13,141 calcium citrate was found to be more absorbable than calcium carbonate when taken on an empty stomach. The treatment format, adopted for simplicity and assurance of optimum compliance, was one in which slow-release sodium fluoride was given together with calcium
Center for Mineral Metabolism and Clinical Research, Southwestern Medical School, University of Texas Southwestern Medical Center at Dallas, 5323 Harry Hines Blvd., Dallas, TX 7 5 2 3 5 4 8 8 5 ,
857
PAK ET AL.
858
citrate orally on an empty stomach on a twice daily schedule.‘“) The clinical pharmacology of this format was never described. In this communication, we report pharmacokinetics of fluoride appearance in serum following a single oral administration of slow-release sodium fluoride and changes induced by calcium citrate and a meal.
MATERIALS AND METHODS
Study I . A single-dose bioavailability study with slow-release sodium fluoride: Effect of calcium citrate or meal given concurrently A group of four postmenopausal women (48-62 years, mean age 5 5 years) and eight premenopausal women (2542 years, mean age 32 years) participated in the study. They were free of peptic ulcer disease or significant renal impairment (endogenous creatinine clearance < 40 mi/ minute). They participated in four phases of study chosen in random order (assignment made by an independent statistician). Each phase lasted 4 days. During phase I (sodium flouride or NaF alone), 50 mg slow-release sodium fluoride (two tablets of Slow Fluoride, Mission Pharmacal Co.) was given orally on an empty stomach at 8 a.m. of the test day, with breakfast withheld. During phase 11 (NaF with calcium), 50 mg slow-release sodium fluoride was given with 400 mg calcium as calcium citrate (Citracal, Mission Pharmacal Co.) and breakfast was withheld. During phase I11 (NaF with meal), 50 mg slow-release sodium fluoride was given with a standard breakfast. The breakfast comprised 200 ml orange juice, one poached egg, two slices of toast with apple jelly, 1/2 cup cream of wheat, 1 cup decaffeinated coffee, and 12 g cheese; it contained 200 mg calcium and 260 mg phosphorus. During phase IV (placebo), two placebo tablets of identical appearance and size as slow-release sodium fluoride containing excipient only (also prepared by Mission) were given on an empty stomach without breakfast. Although a typical single dose of slow-release sodium fluoride is 25 mg,(41a higher dose of 50 mg was utilized here in a single-dose bioavailability study to maximize changes between phases. On days 1-3 of each phase, subjects were stabilized at home on a limited fluoride intake by avoidance of fluoride-containing toothpaste and tea and by drinking distilled water instead of fluorinated water. They were admitted to the general clinical research center (GCRC) on the afternoon of day 3, to be begun and kept on a constant metabolic diet (800 mg P, 800 mg Ca, and 100 mEq sodium per day). O n day 4 at 8:OO a.m., subjects received respective medications outlined before for the four phases. Breakfast was withheld during phases 1, 11, and IV. Lunch wa provided at noon and dinner at 5:00 p.m. Venous blood samples ( 5 ml each) were obtained at 8:00 a.m. (before administration of the test dose), 9:OO a.m., 10:00 a.m., 11:00 a.m., 12 noon, 2:OO p.m., 4:OO p.m., 6:00 p.m., and 8:OO p.m. Each sample was assayed for fluoride. Distilled water was drunk at the following schedule: 300 ml each at 8:OO p.m. of day 3, midnight, 8:00 a m . , 1O:OO a.m., noon, and 5:00
p.m. Other fluids (made with distilled water) except tea (in regulated constant volumes) were allowed. In each subject, the identical meals and fluids were consumed during the four GCRC admissions. There was no rest period between phases, because 3 days of stabilization in each phase was considered sufficient to remove the effect of the previous fluoride dose.
Study 2. Effect of calcium citrate given before or after slow-release sodium fluoride on fluoride bioavailability A total of nine subjects from the previous study participated in this investigation: three were postmenopausal women (48-62 years, mean age 54 years) and six were premenopausal women (25-42 years, mean age 32 years). They underwent two additional phases of study, chosen in random order. In phase V (NaF with Ca before), 50 mg slow-release sodium fluoride was given at 8 a.m. and 400 mg calcium as calcium citrate taken at 6 a.m. In phase VI (NaF with Ca after), 50 mg slow-release sodium fluoride was provided at 8 a.m. and 400 mg calcium as calcium citrate was given at 9 a.m. Breakfast was withheld. Dietary preparation, fluid intake, and tests were the same as in study 1.
Study 3. Effect of long-term treatment with slow-release sodium fluoride on serum fluoride A group of 52 patients with primary osteoporosis participated in the study. There were 11 men and 41 women (mean age 60.2 years, 27-84 years). They participated in an intermittent treatment of slow-release sodium fluoride, comprised of repeated cycles of 13 months each. In each cycle, they took 25 mg slow-release sodium fluoride together with 400 mg calcium as calcium citrate on an empty stomach before breakfast and again at bedtime during the first 12 months. During the last month of each cycle, they took calcium citrate only and sodium fluoride was withheld. A venous blood sample was obtained in the fasting state before the morning dose of slow-release sodium fluoride, before and every 3 months of treatment in each cycle, and analyzed for fluoride.
Analytic methods Serum fluoride was determined by the method of Ekstrand using a fluoride To prevent fluoride contamination, acid-washed glassware or fluoride-free plastic containers were used during sample collection and analysis.
Statistical methods Classic pharmacokinetic analysis‘”’ was employed for each study of fluoride bioavailability, broadly defined here as the rate and extent of fluoride entry into the systemic circulation following oral administration. For each bioavailability study, the peak serum fluoride concentration
859
FLUORIDE BlOAV AILABILITY (C,,,), peak time ( Tmax),serum fluoride half-life ( T , , , ) , and area under the curve (AUC) were calculated. The T,,, gave the time in hours (after fluoride dose) when the peak serum fluoride concentration was obtained. The T , / 2represented the half-life of the decline in serum fluoride concentration from its peak to the value at 12 h. The rate o f decline was calculated by nonlinear regression analysis. The AUC represented the area under the serum concentration curve from time zero to 12 h later in units of ng h/ml. It was calculated using the trapezoid rule. The AUC provided a reflection of the overall fluoride absorption from the gastrointestinal tract. However, it is not identical to total fluoride absorption because some of absorbed fluoride is taken up by bone and excreted in urine. A more direct measure of fluoride absorption from a test substance may be obtained by comparing its AUC with the AUC of a substance believed to be completely absorbed (for example, plain sodium fluoride given alone). For variables Cmax, T,,,, T , , , , and AUC, a single-factor repeated measures analysis of variance was performed to assess differences among the phases. The Student-NewmanKeuls multiple comparisons were used for pairwise comparisons if the analysis of variance was significant at (Y = 0.05. For C,,,, Tmax,TI/*,and AUC, the placebo phase was omitted from the analysis. For assessing the significance of difference in serum fluoride concentration from a single-dose bioavailability study, a preliminary two-factor repeated measures analysis of variance was performed, the two repeated factors being phase and time of sample. Since the phase x time interaction was significant O-, < O.OOOl), a single-factor analysis of variance (with repeated factor phase) was performed at each time. The Student-Newman-Keuls multiple comparisons were used for pairwise comparisons if the analysis of variance was significant at CY = 0.05. For serum fluoride concentrations during long-term treatment, the significant difference in serum fluoride during treatment from pretreatment was assessed by the paired /-test. The Bonferroni inequality was used to adjust for multiplicity of testing when appropriate.
RESULTS Serum fluoride concentration curves following a single oral dose of slow-release sodium fluoride were indistinguishable between premenopausal women and postmenopausal women. When 50 mg slow-release sodium fluoride was given with 400 mg calcium as calcium citrate, serum fluoride concentrations during subsequent 12 h were not significantly different between four postmenopausal women and eight premenopausal women (Fig. I ) . Thus, results from all subjects were combined. ESJect of culciutn and meal ,fluoride hiouvuiluhili~~v
on
Following an oral administration of the placebo medication, \erum fluoride concenlration remained low and unchanged during t h e wbwquent 12 h (big. 2). When FIOW-
-
E 150
W (I)
i
d-Premenopausal
- ~-Postmenopausal
A
joli l 0
0
2
1
I
1
1
I
4
6
8
10
12
Time, hours
t
FIG. 1. Serum fluoride concentration after a single dose of 50 mg slow-release sodium fluoride given with 400 mg calcium as calcium citrate in four postmenopausal and eight premenopausal women. The arrow indicates the time of fluoride administration.
-
2oo
E
-
1
-o-a-
2 150
NaF Alone NaF with Calcium NaF with Meal Placebo N = 12
m
0
g
lL
100
E
2
m
50
(I)
0
1
L
I
0 f
2
I
I
I
6 8 Time, hours 4
I
I
10
12
FIG. 2. Serum fluoride concentration after a single dose of 50 mg slow-release sodium fluoride, given alone, with 400 mg calcium as calcium citrate or with a meal in 12 subjects. The arrow indicates the time of fluoride administration. (a) I-, < 0.05 between NaF alone and NaF with calcium or NaF with meal; (b) y < 0.05 between NaF with calcium and NaF with meal. release sodium fluoride was administered, alone or with calcium citrate or a meal, serum fluoride concentration was significantly higher than the corresponding placebo value during all 12 h of the study (p < 0.05). However, the serum fluoride concentration in phase I1 (NaF with calcium) was significantly lower than in phase I (NaF alone), except at 12 h. The serum fluoride concentration during the first 4 h of NaF with a mean (phase 111) was significantly lower than during the corresponding time period of NaF alone. Compared to NaF with calcium, NaF with a meal gave a significantly lower serum fluoride concentration at 1 and 2 h and higher values at 6, 8, 10, and 12 h.
860
PAK ET AL.
During NaF with calcium, the C,,, and AUC were significantly lower than during NaF alone (Table 1). However, Tmaxand T , , , were not different. For NaF with a meal, the C,,, and AUC were also significantly reduced, compared to NaF alone. However, T,,, was significantly higher. Between phase I1 (NaF with calcium) and phase 111 (NaF with meal), T,, was significantly higher for phase 111.
Effect of calcium given before and after slo w-release sodium fluoride Compared to NaF alone, no significant difference in serum fluoride concentration, Cmax,Tmax,T L , , ,or AUC was noted when calcium citrate was given 2 h before or 1 h after slow-release sodium fluoride (Fig. 3). The AUC for NaF alone was 1302 f 257 ng h/ml, which was not significantly different from 1187 + 378 ng h/ml for NaF with Ca before or 1144 f 253 ng- h/ml for NaF with calcium after.
Serum fluoride concentration during long-term treatmen t Before and after 1 month of withdrawal of slow-release sodium fluoride therapy (0, 13, 26, and 39 months), serum fluoride concentration (fasting value, before the morning dose of fluoride) was below 95 ng/ml (Fig. 4). During treatment with slow-release sodium fluoride (3, 6, 9, 12, 16, 19, 22, 25, 29, 32, 35, and 38 months), serum fluoride concentration was mostly above 95 ng/ml and below 190 ng/ml.
DISCUSSION This study examined the effect of calcium and a meal on fluoride bioavailability. The concurrent administration of calcium citrate with slow-release sodium fluoride reduced the peak fluoride concentration in serum and fluoride absorption (reflected by area under the curve), each by 27%,
without affecting the time required to reach peak concentration or the subsequent rate of decline in fluoride concentration. This effect of calcium was negligible when calcium citrate was given 2 h before or 1 h after fluoride administration. The ingestion of a meal delayed the time required to reach the peak fluoride concentration in serum, possibly because of prolonged gastric emptying caused by food. However, a similar reduction in peak fluoride concentration and in fluoride absorption was observed, as with calcium. The conditions of this study may be different from the usual clinical setting when fluoride is given long term to postmenopausal women with osteoporosis. However, the following considerations suggest that the two situations are comparable. First, the fluoride bioavailability was indistinguishable between postmenopausal and premenopausal women. Second, the acute effect of a single dose of slowrelease sodium fluoride (50 mg) in untreated patients described in this paper is qualitatively similar to the effect of 25 mg slow-release sodium fluoride in patients maintained on long-term treatment with fluoride. In this acute study, the increment in serum fluoride concentration from basal to peak level was 101 ng/ml when 50 mg slow-release fluoride was given with 400 mg C a as calcium citrate. In a prea single dose of 25 mg slow-release sodium vious fluoride with a meal in untreated patients raised serum fluoride concentration by 40 ng/ml. In the same study,"" the difference between the trough and peak levels of fluoride in serum was 39 ng/ml in patients maintained on long-term treatment with slow-release sodium fluoride (25 mg) with a meal twice daily. The fluoride absorption from a single dose of 50 mg slow-release sodium fluoride (given with a meal), estimated from the difference in the area under the curve from the placebo phase, was approximately twice that of 25 mg slow-release sodium fluoride given (with a meal) as a single dose or as a part of long-term treatment.llzlThus, fluoride bioavailability did not seem to be affected substantially by previous treatment with slow-release sodium fluoride. During long-term treatment with 25 mg slow-release SOdium fluoride given together with 400 mg calcium as cal-
TABLE l . EFFECTOF CALCIUM CITRATE OR MEALGIVEN CONCURRENTLY ON FLUORIDE BIOAVAILABILITY~ Phase
I NaF alone Cm,, ng/ml AUC, ng h/ml Tmax, h Tit,,
h
184 + 33 1314 f 231 2.0 f 1.0 5.9
f
1.0
II NaF with calcium
135 i 24b 961 i 157b 1.6 f 0.7c 7.1 f 1.3
aAll values represent mean + SD. N = 12. bp < 0.05 between phases I and I1 and between I and 111. cp < 0.05 between phases 11 and 111.
III
IV
NaF with meal
Placebo
136 1096 3.6 13.2
f f f f
33b 241b 2.0b 20.4
-
-
FLUORIDE BIOAVAILABILITY
2oo
861
r
-+-
NaF Alone NaF t Ca Before + NaF t Ca Alter
E cn 150 c
i E & 3 L
N = 9
100
E2 l 50
Q)
(1
v)
oL
I
1
I
I
6 0 Ti me, hours
2
0
4
t
I
I
10
12
FIG. 3. Serum fluoride concentration following a single dose of 50 mg slow-release sodium fluoride, given alone, with 400 mg calcium as calcium citrate given 2 h before (NaF + Ca before) or 1 h after (NaF + C a after) fluoride administration in nine subjects. The arrow indicates the time of fluoride administration.
-
. E
300 250
The fluctuation in serum fluoride concentration from basal to peak levels was about 40 ng/ml."zl This discussion provides justification of our general therapeutic guidelines with slow-release sodium fluoride. The simplest format recommended initially is the combined use of slow-release sodium fluoride with calcium citrate on an empty stomach. If this format is unacceptable or causes gastrointestinal side effects, slow-release sodium fluoride can be given with a light meal without sacrificing fluoride absorption. It is unlikely that treatment with an immediate-release sodium fluoride (25 mg twice or three times a day) could maintain the serum fluoride concentration between 95 and 190 ng/ml without a wide circadian fluctuation, as has been shown here for slow-release sodium fluoride (25 mg twice a day). A pharmacokinetic analysis of the previous acute fluoride bioavailability study'"' revealed a much higher peak fluoride concentration in serum following administration of a rapid-release sodium fluoride (25 mg) than after taking the same dose as slow-release sodium fluoride (220 versus 84 ng/ml, p < 0.05). Fluoride absorption, estimated from the difference in area under the curve from that of the placebo phase, was more than twofold higher for the rapid-release sodium fluoride preparation."*' In our unpublished study of patients maintained on long-term treatment with slow-release sodium fluoride (25 mg) with calcium citrate (400 mg Ca) twice a day, serum fluoride concentration could be kept between 95 and 190 ng/ml during 12 h following 25 mg slow-release sodium fluoride with 400 mg Ca as calcium citrate. However, serum fluoride concentration reached a peak level of 454 ng/ml 1 h after oral administration of plain sodium fluoride (25 mg) with 500 mg Ca as calcium carbonate. In summary, the oral administration of slow-release sodium fluoride with calcium citrate averted the attainment of sharp high peaks of fluoride in serum and produced a slow gradual decline in serum fluoride. Thus, the convenient treatment format of 25 mg slow-release sodium fluoride and 400 mg calcium as calcium citrate given together on an empty stomach on a twice daily schedule assured maintenance of serum fluoride within 95-190 ng/ml, believed to be the therapeutic window, with a modest fluctuation between trough and peak levels.
1
m c 0
0 0
a LL
E
2
Q u)
I
0
6
12
I
19
25
I
32
38
Month
FIG. 4. Fasting (trough) concentration of serum fluoride during long-term administration of slow-release sodium fluoride (25 mg twice a day) and calcium citrate (400 mg calcium twice a day). Sodium fluoride treatment was interrupted for 1 month after each 12 months of treatment, as indicated by vertical dotted lines. Significant difference from the pretreatment value is indicated by (a) p < O.ooO1; (b) p = 0.0001-0.003, and (c) p = 0.05-0.003. Numbers on graph depicts sample size. X = insufficient data for statistical analysis. Bars indicate mean f SEM.
ACKNOWLEDGMENTS Supported by USPHS grants R01-AR16061, M01RR00633, and POI-DK20543.
REFERENCES cium citrate on an empty stomach on a twice daily schedule, the trough levels of serum fluoride could be kept within 95-190 ng/ml. In a previous report,'") serum fluoride was also shown to be maintained within 95-190 ng/ml during long-term treatment with 25 mg slow-release sodium fluoride taken with breakfast or a snack twice a day.
1. Farley JR, Wergedal JE, Baylink DJ 1983 Fluoride directly stimulates proliferation and alkaline phosphatase activity of bone-forming cells. Science 222:330-332. 2. Zerwekh JE, Morris AC, Padalino PK, Gottschalk F. Pak CYC Fluoride and parathyroid hormone raises intracellular calcium in human osteoblasts. J Bone Min Res 5:S131-S135. 3. Riggs LB, Seeman E, Hodgson SF, Taves DR, O'Fallon MW
862
4.
5.
6. 7.
8. 9.
10.
11. 12.
PAK ET AL. 1982 Effect of the fluoride/calcium regimen on vertebral fracture occurrence in postmenopausal osteoporosis. N Engl J Med 306:446-450. Pak CYC, Sakhaee K , Zerwekh JE, Parcel C , Peterson R, Johnson K 1989 A safe and effective treatment of primary osteoporosis by intermittent application of slow-release sodium fluoride: Augmentation of vertebral bone mass and inhibition of fractures. J Clin Endocrinol Metab 68:150-159. Wolinsky I, Simkin A, Guggenheim K 1972 Effects of fluoride on metabolism and mechanical properties of rat bone. Am J Physiol 223(1):46-50. Taves DR 1970 New approach to the treatment of bone disease with fluoride. Fed Proc 29(3):1185-1187. Guggenheim K , Simkin A, Wolkinsky I 1976 The effect of fluoride on bone of rats fed diets deficient in calcium or phosphorus. Calcif Tissue Res 22:9-17. Angmar-Mansson 8, Whitford GM 1984 Enamel fluorosis related to plasma F levels in the rat. Caries Res 18:25-32. Suttie JW, Faltin EC 1971 Effect of a short period of fluoride ingestion on dental fluorosis in cattle. Am J Vet Res 32: 217-222. Harrison JE, Hitchman AJW, Hasany SA, Hitchman A, Tam CS 1984 The effect of diet calcium on fluoride toxicity in growing rats. Can J Physiol Pharmacol 62:259-265. Riggs LB, Jowsey J 1972 Treatment of osteoporosis with fluoride. Semin Drug Treatment 2:27-33. Pak CYC, Sakhaee K , Gallagher C, Parcel C, Peterson R, Zerwekh JE, Lemke M, Britton F, Hsu MC, Adams B 1986 Attainment of therapeutic fluoride levels in serum without
13.
14.
15.
16.
17.
major side effects using a slow release preparation of sodium fluoride in postmenopausal osteoporosis. J Bone Min Res 1: 563-571. Nicar M J , Pak CYC 1985 Calcium bioavailability from calcium carbonate and calcium citrate. J Clin Endocrinol Metab 61:391-393. Harvey JA, Zobitz MM, Pak CYC 1988 Dose dependency of calcium absorption: A comparison of calcium carbonate and calcium citrate. J Bone Min Res 3:253-285. Pak CYC, Sakhaee K, Zerwekh J E Effect of intermittent therapy with a slow release fluoride preparation. J Bone Min Res 5:s 149-S 155. Ekstrand J 1977 A micromethod for the determination of fluoride in blood plasma and saliva. Calcif Tissue Res 23: 225-228. Ekstrand J , Alvan G , Boreus LO, Norlin A 1977 Pharmacokinetics of fluoride in man after single and multiple doses. Eur J Clin Pharmacol 12:311-317.
Address reprint requests to: Dr. Charles Y.C. Pak Center for Mineral Metabolism and Clinical Research Southwestern Medical School University of Texas Southwestern Medical Center at Dallas 5323 Harry Hines Blvd. Dallas, TX 75235-8885 Received for publication October 19, 1989; in revised form March 26, 1990; accepted March 26, 1990.
