Urinary loss of thiamine is increased by low doses of furosemide in healthy volunteers

Urinary loss of thiamine is increased by low doses of furosemide in healthy volunteers

Urinary loss of thiamine is increased by low doses of furosemide in healthy volunteers JONATHAN RIECK, HILLELHALKIN, SHLOMO ALMOG, HANNA SELIGMAN, AHA...

478KB Sizes 1 Downloads 47 Views

Urinary loss of thiamine is increased by low doses of furosemide in healthy volunteers JONATHAN RIECK, HILLELHALKIN, SHLOMO ALMOG, HANNA SELIGMAN, AHARON LUBETSKY,DAVID OLCHOVSKY, and DAVID EZRA TEL HASHOMER, ISRAEL

Prolonged furosemide treatment is associated with urinary loss of thiamine and thiamine deficiency in some patients with congestive heart failure and low dietary thiamine intake. In the rat, diuretic-induced thiamine urinary loss is solely dependent on increased diuresis and is unrelated to the type of diuretic used. We studied the effects of single intravenous doses of furosemide (I, 3, and 10 mg) and of normal saline infusion (750 mL) on urinary thiamine excretion in 6 volunteers. Over a 6-hour period, furosemide induced dose-dependent increases in urine flow and sodium excretion rates (mean _+SD), from 51 _+ 17 mL/h at baseline to 89 _+29 mL/h, 110 _+38 mL/h, and 183 _+58 mL/h (F = 10.4, P < .002) and from 5.1 _+2.3 mmol/h to 9.4 _+6.8 mmol/h, 12.1 _+2.6 mmol/h, and 20.9 _+ 10.6 mmol/h (F = 6.3, P < .005) for the three doses, respectively (104 _+35 mL/h and 13.0 _+6.2 mmol/h for the saline infusion). During this period the thiamine excretion rate doubled from baseline levels (mean of four 24-hour periods before the diuretic interventions) of 6.4 _+5.1 nmol/h to 11.6 _+8.2 nmol/h (F = 5.03, P < .01, for all four interventions, no difference being found between them), then returning over the following 18 hours to 6.1 _+3.9 nmol/h. The thiamine excretion rate was correlated with the urine flow rate (r = 0.54, P < .001), with no further effect of the type of intervention or sodium excretion rate. These findings complement our previous results in animals and indicate that sustained diuresis of > 100 mL/h induces a nonspecific but significant increase in urinary loss of thiamine in human subjects. Thiamine supplements should be considered in patients undergoing sustained diuresis, when dietary deficiency may be present. (J Lab Clin Med 1999; 134:238-43)

F

u r o s e m i d e - i n d u c e d thiamine loss, m e d i a t e d by increased urinary excretion of the vitamin, was first described by Yui et al.1,2 We have described clinically meaningful thiamine deficiency in a subset of patients with congestive heart failure treated with high doses of the diuretic for over 3 months. 3,4 Other investigators have either not confirmed these findings 5-7 or have ascribed them, at least in part, to concomitant low dietary intake of the vitamin. 8 In a more recent study conducted in rats we were able to demonstrate that the

effect of furosemide on urinary loss of thiamine is not specific to that drug but is a feature c o m m o n to diuresis induced by a thiazide, amiloride, acetazolamide, mannitol, and normal saline infusion. 9 In fact, the sole determinant of the degree of thiamine loss induced by this variety of diuretic interventions was urinary volume or flow rate. The objective of the present study was to assess the effects of low doses of furosemide on the urinary excretion of thiamine in normal volunteers. METHODS

From the Division of Clinical Pharmacology and Toxicology,Department of Medicine, Sheba Medical Center and Tel Aviv University, Sackler School of Medicine. Submitted for publication May 20, 1999; accepted May 26, 1999. Reprint requests: Hillel Halkin, MD, Department of Medicine, Sheba Medical Center, Tel Hashomer 52621, Israel. Copyright © 1999 by Mosby, Inc. 0022-2143/99 $8.00 + 0 5/1/100512

238

Volunteers. The study was approved by the institutional review board. Six consenting healthy adults (ages 21 to 27 years, 5 men, 1 woman) were recruited after screening by medical history, physical examination, routine laboratory tests, and an electrocardiogram. No volunteer was a smoker or drank alcoholic beverages. All were consuming unrestricted diets and none was using any medications or vitamin supplements of any kind.

J L a b Clin M e d V o l u m e 134, N u m b e r 3

R i e c k e t al

T a b l e I. U r i n e v a r i a b l e s

Volume (mL/d) Creatinine (mg/d) Flow (mL/h)

24 hours before

diuretic

interventions

Before I mg

Before 3 mg

Before 10 mg

Before saline

Mean

1224 _+ 454

1218 _ 413

1225 _+ 369

1232 _+ 504

1223 _+ 408

943 + 436

1271 _+ 318

1200 _+ 345

1224 _+ 291

1169 _+ 345

50 _+ 17

51 _+ 19

51 _+ 16

52 _+ 21

51 _+ 17

Sodium (mmol/h)

4.5 _+ 1.6

5.5 _+ 3.5

5.1 _+ 2.3

5.9 +_ 2.1

5,1 _+ 2.3

Potassium (mmol/d)

47 _+ 17

49 _+ 16

53 _+ 21

53 _+ 20

50 _+ 17

Thiamine (nmol/d)

194_+ 91

157_+ 114

125 + 81

141 _+ 118

154_+ 90

Thiamine (nmol/h)*

8.1 _+ 7.6

6.6 _+ 4.7

5.2 _+ 3.4

5.9 _+ 4.8

6.4 _+ 5.1

Thiamine (nmoi/L)l-

143 _+ 91

128 _+ 57

102 _+ 52

119 _+ 681

112 _+ 58

*Rate, tConcentration.

T a b l e II. U r i n e v a r i a b l e s (period

II), a n d

after diuretic

interventions:

Observed

collection

(period

I), u n o b s e r v e d

collection

total for day

Furosemide dose

I mg

3 mg

10 mg

Saline solution

Volume (mL) Period I Period II Days total

485 _+ 154" 603 -+ 191 1088 + 285

551 + 188" 756 _+ 229 1341 _+ 234

877 _+ 216" 676 _+ 250 1553 _+ 430

535 _+ 159 1089 _+ 2 7 4 t 1623 _+ 333

Flow (mL/h) Period I Period II Days total

89 + 29* 33 + 11 46 _+ 14

110 _+ 38* 43 _+ 13 59 -+ 11

183 + 58* 37 + 13 67 _+ 18

104 + 35 59 -+ 1 5 t 68 _+ 14

Sodium excretion Period I (mmol/h) Period II (mmol/h) Days total (mmol/d)

9.4 _ 6,8* 4.1 _+ 1.4 132 _+ 49

12.1 _+ 2,6* 3.7 _+ 1.7 128 _+ 28

20.9 + 10.6" 3.0 --- 1.3 152 _+ 48

13.0 _+ 6.2 5.0 _+ 2.61225 _+ 51

Sodium concentration (mmol/L) Period I Period II

101 _+ 55 131 + 44

109 _+ 22 92 _+ 48

112 _+ 24 95 _+ 29

125 _+ 32 153 _+ 43

Thiamine excretion Period I (nmol/h) Period II (nmol/h) Days total (nmol/d)

10.5 _+ 4.6:~ 8.1 _+ 7.6 156 _+ 67

13.0 _+ 9.9:1: 6.6 _+ 4.7 141 _+ 107

9.7 _+ 2.35 5.2 _+ 3.4 161 _+ 76

13.1 _+ 9.35 5.9 ___4.9 236 ___105

Thiamine concentration (nmol/L) Period I Period II

121 _+ 56 163 _+ 70

60 _+ 29§ 130 _+ 80

55 +_ 13§ 166 + 88

145 _+ 42 156 _+ 53

*Statistically significant dose-related correlations, See Results, period I. 1-Statistically significant effects of saline infusion, See Results, period II, ~Significantly increased thiamine excretion rates as c o m p a r e d with baseline, See Results, period I. §Significantly d e c r e a s e d thiamine concentrations, See Results, period I.

239

240

J Lab Clin Med September 1999

Rieck et al

log Thiamine excretion (nmol/hr)

r=0.54 p<0.001

[20)

/

...... •





• •

:•

12) 120)

,

, •

(1,00)

(,180)

2 log Diuresis (mL/hr) (unlransformed values in brackets)

Fig 1. Correlationof thiamine excretionrate with diuresis (urine flow rate) over all study periods (log-transformedvalues).

Procedures. Volunteers were tested on 4 study days, 1 for each of the four diuretic interventions, at 1-week intervals. After an overnight fast and pre-intervention 24-hour urine collection at home, subjects arrived at the laboratory at 6 aM. At 7 AM the diuretic was administered (either 1 mg, 3 mg, or 10 mg intravenous furosemide or an intravenous infusion of 750 mL normal saline solution administered over 20 minutes). Spontaneously voided urine was collected over an approximate 6-hour observation period. At 10 AM a light meal, consisting of a roll and 200 mL of tea or coffee, was consumed. At 1 PM subjects left the laboratory for home, where they collected all urine voided until 7 AMthe following day. Subjects were instructed to consume the same food items and beverages to which they were accustomed for the remainder of the 24-hour period at home. Analytic methods. Urine concentrations of creatinine, sodium, and potassium were determined by routine methods. Urinary concentrations of thiamine were determined by highpressure liquid chromatography. 10 The completeness of urine collections at home (two periods per treatment) was ascertained by monitoring creatinine excretion. Data analysis. Data are presented throughout as mean _+ SD. Comparisons of differences between the effects of the four diuretic interventions (three doses of furosemide and intravenous saline infusion) on urine volume, flow rate, and excretion of sodium, potassium, and thiamine were done by one-way analysis of variance (unpaired Student t tests or the two-sample median test for non-normally distributed data when only two groups were being compared). The associations between furosemide doses and the same dependent variables were analyzed by linear regression. The associations between thiamine excretion and urine volume, flow rate, and

sodium excretion were explored by multiple linear regression. A level of P < .05 was taken as indicating significance. All analyses were done with GB-STAT statistical programs.

RESULTS There were no significant individual differences in urine variables between the 6 volunteers on the 4 preintervention days (data not shown). Baseline values for the urine variables on the pre-intervention days are presented in Table I. Volume and excretion of creatinine, sodium, potassium, and thiamine did not vary significantly between the 4 days, although the between-individual coefficient of variation was on the order of 50% (75% for thiamine). The excretion rate of thiamine was 6.4 +_5.1 nmol/h, and the range of daily thiamine excretion was 32 to 549 nmol/d (95% confidence limits 102 to 206 nmol/d). T h i a m i n e excretion on the pre-intervention days was significantly correlated with urine flow or volume (r = 0.57, P < .01) and equally with sodium excretion (r = 0.58, P < .01), with neither variable adding to the predictive capacity of the other. Results of the diuretic interventions are given in Table II. The mean duration of the observed collection period was 5.6 _+ 0.5 hours (period I), and that of the unobserved collection at home (period II) was 18.3 _+ 1.1 hours. Total creatinine excretion was not different on the intervention days from the baseline values (1169 _+ 345 mg/d vs 1353 _+ 393 mg/d), indicating the adequacy of the urine collections. The potassium excretion rate was significantly higher than the baseline value after all four interventions (2.5 + 026 mmol/h vs 2.1 _ 0.7 mmol/h; Chi 2 = 5.8, P < .02)(but inasmuch as there

J Lab Clin Med Volume 134, Number 3

were no differences between them, individual values are not presented). During period I, furosemide induced dose-related increases in urine flow, volume, and sodium excretion (F -- 10.4, P < .002; F = 10.1, P < .002; and F = 6.3, P < .005; respectively). (Saline infusion resulted in a diuretic response similar to the 3 mg furosemide dose.) During period II the effects of the three furosemide doses were no longer evident in terms of urine flow, volume, and sodium excretion. However, after the saline infusion there were still significant increments in these three variables (F = 4.45, P < .01; F = 4.83, P < .015; and F = 17.7, P < .001; respectively). Urinary concentrations of sodium did not vary significantly between treatments in both study periods. During period I, the thiamine excretion rate was 2fold higher (for all four interventions) as compared with both the pre-intervention rate and that for period II (mean values 11.6 + 8.2 nmol/h, 6.4 _+ 5.1 nmol/h, and 6.1 _+3.9 nmol/h, respectively; F = 5.03, P < .01). There was no significant difference between the three furosemide doses or the saline infusion in this effect. Thiamine urine concentrations were significantly lower after the 3 mg and 10 mg furosemide doses (F = 3.52, P < .04). In period II, the thiamine excretion rate showed a trend toward higher values only after the saline infusion (F -- 1.6, P < .2) but was significantly correlated with urine flow, volume, and sodium excretion (r = 0.53, P < .012; r = 0.56, P < .01; and r = 0.44, P < .05; respectively). For all observation periods combined, thiamine excretion and urine flow rates were highly correlated (Fig 1). Neither sodium excretion rate nor urine volume was additive in the multiple regression model. DISCUSSION Thiamine is eliminated from the body by renal excretion that is dependent on glomerular filtration rate and on thiamine plasma concentrations. 11 Under normal conditions free thiamine constitutes only a small fraction of total blood concentration, because it is rapidly taken up by blood cells and phosphorylated to thiamine pyrophosphate, which cannot diffuse out of the cell.12 The vitamin undergoes glomerular filtration as well as active tubular reabsorption and secretion,11 with the latter being decreased by concomitant probenecid) 3 At very high plasma thiamine concentrations (>100 nmol/L) achieved only experimentally, or after high therapeutic doses (50 to 200 mg), thiamine renal clearance approaches the magnitude of renal plasma flow. Tubular reabsorption is only half-maximal at plasma concentrations of 20 to 50 nmol/L, which is 2-fold to 5-fold above the levels found in healthy volunteers not taking supplements. 11 In such subjects, under condi-

Rieck et al

241

tions of experimental dietary restriction of thiamine intake, urinary excretion of the vitamin falls within days to levels <50 nmol/d and soon after ceases completely. 13,14 Urinary thiamine excretion below 90 nmol/d parallels the appearance of biochemical evidence of subclinical thiamine deficiency. 15 In healthy volunteers receiving 1.6 mg of the vitamin daily, thiamine excretion was 125 to 1275 nmol/d. 13 The range of daily urinary excretion of thiamine in our study, in which there was no thiamine supplementation, is consistent with these values. In a recent experimental study done in rats, we 9 were able to expand on the observations of Yui et al on furosemide-induced thiamine loss.l, 2 We showed that a variety of other diuretics (including acetazol-amide, chlorothiazide, amiloride, mannitol, and normal saline solution) with differing specific renal mechanisms of action caused similar dose-dependent increases in urinary thiamine excretion. These increased excretion rates were solely dependent on changes induced in urine flow rates and were only partially explained by differences in changes in fractional sodium excretion. The present findings in human volunteers are similar in nature. With no diuretic intervention, thiamine excretion was simply correlated with urine volume (interchangeably, but not additively, with sodium excretion). As expected, furosemide induced dose-dependent increases in urine flow and sodium excretion that lasted no longer than the 6 hours after injection. Urine volume and the amounts of sodium excreted during period I after the low doses of furosemide used were proportional to those found in normal volunteers over 6 hours after an oral dose of 20 mg. 16 Changes in urine flow and sodium excretion rates were proportional to those seen in normal volunteers after 20 mg intravenous furosemide.17 Moreover, they were commensurate with those predicted by pharmacokinetic and dynamic modeling of furosemide effects based on the renal excretion rate of furosemide after intravenous administration.18,19 During this expected period of duration of the diuretic effect, the thiamine excretion rate increased 2-fold over baseline following all three furosemide doses at all flow rates achieved. Thiamine concentrations in urine were significantly reduced after the 3 mg and 10 mg thiamine doses. These changes in thiamine excretion may reflect the transient increase in glomerular filtration rate known to occur after single doses of furosemide, 17 dilution of urinary thiamine by increased urine free water at these doses of furosemide, or a relative increase in tubular reabsorption of the vitamin, given enhanced delivery to putative reabsorption sites, with the process of reabsorption itself being unaffected by furosemide. The diuresis and natriuresis induced by the saline infu-

242

J Lab Clin Med September 1999

Riecket al

sion (similar to those that f o l l o w e d the 3 mg dose of furosemide) also resulted in a doubling of the thiamine excretion rate. A l t h o u g h the precise m e c h a n i s m and sites of thiamine tubular reabsorption are unknown, thiamine u p t a k e b y e r y t h r o c y t e s has b e e n shown to b e i n d e p e n d e n t of sodium transport 20 and in intestinal biopsy s p e c i m e n s has been shown to be unaffected by f u r o s e m i d e 21, 22 This could explain the differential effect on urinary thiamine concentrations, which fell, while those of sodium were unchanged, despite the dose-related increased sodium excretion rates. During period II, after the effect o f the diuretic interventions waned, with urine flow and sodium excretion rates back at baseline levels, thiamine excretion also returned to baseline, maintaining the correlation with urine flow rates. Given our understanding of the renal handling of thiamine as described above, these findings may be taken as indicating that in the absence of thiamine deficiency or of rapidly changing thiamine b l o o d levels, single doses of f u r o s e m i d e will cause enhanced thiamine excretion at urine flow rates in the order of 100 mL/min and greater. We do not know the effects of r e p e a t e d doses, but in the absence of tolerance, the longer the duration of increased diuresis, the greater will be the amount of thiamine lost. Whether or not this results in thiamine deficiency depends on thiamine intake during the p e r i o d of increased diuresis,4, 8 b e c a u s e thiamine body stores are small, 2~ and overt deficiency develops relatively rapidly under conditions of increased demand or a negative balance of intake and loss. 13-15 The main limitations of our study were the uncontrolled periods of urine collection before and after diuretic interventions and the l a c k o f control of thiamine intake. Given the similarity of creatinine excretion over the study days, erroneous collection seems not to have introduced systematic data bias. This is also supported by the similarity of our sodium excretion data to those in earlier studies on the clinical pharmacology of furosemide. 24 With respect to thiamine balance, our results were affected by the considerable between-individual differences in daily excretion. This may explain why we were unable to demonstrate significant increments in total daily excretion of thiamine after the diuretic interventions. Classical work 25 demonstrated a linear relation b e t w e e n thiamine intake and urinary excretion at daily intakes of 0.6 to 2.0 mg, with a 2-fold to 3-fold range of b e t w e e n - i n d i v i d u a l variation in excretion at given levels of intake. Because daily thiamine excretion did not differ significantly over the preintervention days, the element of variability could only have obscured a diuresis-related effect, which was not the case.

In summary, our results show that even low doses of f u r o s e m i d e cause significantly increased urinary loss of thiamine, by a nonspecific mechanism related to the increased diuresis itself. Consequently, under conditions of no supplementation, continued high rates o f d i u r e s i s m a y be a s s o c i a t e d with n e g a t i v e t h i a m i n e balance and clinically significant deficiency. A c c o r d ingly, t h i a m i n e supplements should be c o n s i d e r e d in all patients expected to undergo a period o f sustained diuresis when inadequate dietary intake is possible. REFERENCES

1. Yui Y, Fujiwara H, Mitsui H, Wakabayashi A, Kambara H, Kawai C, et al. Furosemide induced thiamine deficiency [abstract]. Jpn Circ 1978;4:744. 2. Yui Y, Itokawa Y, Kawai C. Furosemide induced thiamine deficiency. Cardiovasc Res 1980;14:537-40. 3. Seligmann H, Halkin H, Rauchfleisch S, Kaufmann N, Tal R, Motro M, et al. Thiamine deficiency in patients with congestive heart failure receiving long term furosemide therapy: a pilot study. Am J Med 1991;91:151-5. 4. Shimon I, Almog S, Vered Z, Seligmann H, Shefi M, Peleg E, et al. Improved left ventricular function after thiamine supplementation in patients with congestive heart failure receiving long term furosemide therapy. Am J Med 1995;98:48590. 5. Levy WC, Soine LA, Huth MM, Fishbein DR Thiamine deficiency in congestive heart failure [letter]. Am J Med 1992; 93:705-6. 6. Kwok T, Falconer-Smith JF, Potter JF, Ives DR. Thiamine status of elderly patients with cardiac failure. Age Ageing 1992;21:67-71. 7. Yue QY, Beermann B, Lindstrom B, Nyquist O. No difference in blood thiamine diphosphate levels between Swedish caucasian patients with congestive heart failure treated with furosemide and patients without heart failure. J Intern Med 1997;242:491-5. 8. Brady JA, Rock CL, Horneffer MR. Thiamine status, diuretic medications and the management of congestive heart failure. J Am Diet Assoc 1995;95:541-4. 9. Lubetsky A, Winaver J, Seligmann H, Olchovsky D, Almog S, Halkin H, et al. Urinary thiamine excretion in the rat: effects of furosemide, other diuretics, and volume load. J Lab Clin Med 1999;134:232-7. 10. Wielders J, Mink C. Quantitative analysis of total thiamine in human blood, milk and cerebrospinal fluid by reversed phase ion pair high performance liquid chromatography. J Chromatogr 1983;277:145-56. 11. Weber W, Nitz M, Looby M. Nonlinear kinetics of the thiamine cation in humans: saturation of nonrenal clearance and tubular reabsorption. J Pharmacokinet Biopharm 1990;18: 501-23. 12. Tallaksen CME, Sande A, Bohmer T, Bell H, Karlsen J. Kinetics of thiamin and thiamin phosphate esters in human blood, plasma and urine after 50 mg intravenously or orally. Eur J Clin Pharmacol 1993;44:73-8. 13. Ziporin Z, Nunes WT, Powell RC, Waring PP, Sauberlich HE. Excretion of thiamine in the urine of young adult males receiving restricted intakes of the vitamin. J Nutr 1965;85: 287-96. 14. Ziporin Z, Nunes WT, Powell RC, Waring PP, Sauberlich HE. Thiamine requirement in the human adult as measured

J Lab Clin Med Volume 134, Number 3

15.

16.

17.

18.

19.

20.

by urinary excretion of thiamine metabolites. J Nutr 1965; 85:297-304. Wood B, Gijsbers A, Goode A, Davis S, Mulholland J, Breen K. A study of partial thiamin restriction in human volunteers. Am J Clin Nutr 1980;33:848-61. Ramsay LE, McInnes GT, Hettiarachchi J, Shelton J, Scott R Bumetanide and frusemide: a comparison of dose-response curves in healthy men. Br J Clin Pharmacol 1978;5:243-7. Passmore AR Copeland S, Johnston GD. A comparison of the effects of ibuprofen and indomethacin upon renal haemodynamics and electrolyte excretion in the presence and absence of frusemide. Br J Clin Pharmacol 1989;27:483-90. Ponto LLB, Schoenwald RD. Furosemide (Frusemide): a pharmacoldnetic/pharmacodyamic review (part I). Clin Pharmacokinet 1990;18:381-408. Ponto LLB, Schoenwald RD. Furosemide (Frusemide): a pharmacokinetic/pharmacodyamic review (part II). Clin Pharmacokinet 1990;18:460-71. Casirola D, Patrini C, Ferrari G, Rindi G. Thiamine trans-

Rieck et al

21.

22.

23.

24.

25.

243

port by human erythrocytes and ghosts. J Membrane Biol 1990;118:1 l-8. Casirola D, Ferrari G, Gastaldi G, Patrini C, Rindi G. Transport of thiamine by brush-border membrane vesicles from rat small intestine. J Physiol 1988;398:329-39. Laforenza U, Patrini C, Alvisi C, Faelli A, Licandro A, Rindi G. Thiamine uptake in human intestinal biopsy specimens, including observations from a patient with acute thiamine deficiency. Am J Clin Nutr 1997;66:320-6. Anonymous. Deaths associated with thiamine deficient total parenteral nutrition. MMWR Morb Mortal Wkly Rep 1989; 38:43-6. Alvan G, Helleday L, Lindholm A, Sanz E, Villen T. Diuretic effect and diuretic efficiency after intravenous dosage of furosemide. Br J Clin Pharmacol 1990;29:215-9. Mickelsen O, Caster WO, Keys A. Statistical evaluation of the thiamine and pyramin excretions of normal young men on controlled intakes of thiamine. J Biol Chem 1947;168: 415-31.