Sulindac is not renal sparing in man

258

We investigated the claimed renal-sparing effect of the cyclooxygenase inhibitor sulindac. Fifteen normal women following a diet of 50 mEq salt a day were randomly assigned to 5 days of either placebo, sulindac, 200 mg b.i.d., or indomethacin, 25 mg q.i.d., after first serving as their own controls. Renal effects were assessed by the excretion rate of prostaglandin (PG) E2 (an index of renal PG synthesis), sodium balance, plasma renin activity (PRA), and the response to furosemide. Systemic effects were assessed by collagen- induced platelet aggregation and thromboxane B, formation and by the urinary excretion of a systemically formed metabolite of PGF,,, (PGF-M). Both sulindac and indomethacin resulted in a positive sodium balance and a reduction in 24-hour urinary PGE, excretion (range 49% to 86%). Basal PRA was decreased by indomethacin only, but the increases in PRA and in urinary PGE, excretion in response to furosemide were inhibited by both sulindac and indomethacin. Sulindac reduced the natriuresis induced by furosemide, and indomethacin reduced the rise in inulin clearance after furosemide. Thus the two nonsteroidal anti-inflammatory drugs had similar effects on the kidney. Indomethacin had a greater effect than sulindac on the inhibition of collagen-induced platelet aggregation and thromboxane synthesis and the two drugs had equivalent effects on the reduction of PGF-M excretion. Peak plasma drug concentrations of indomethacin (1.9 ± 0.4 jig/ml) and sulindac sulfide (7.7 ± 1.9 "Tim') were those associated with clinical efficacy. We conclude that sulindac has renal effects qualitatively comparable to those of indomethacin. Therefore, caution should continue to be exercised with the use of sulindac, as with other nonsteroicial anti-inflammatory drugs, in clinical situations in which cydooxygenase inhibitors have been associated with deterioration of renal function. (CLIN PHARMACOL THER 38:258-265, 1985.)

Douglas G. Roberts, M.D.,* John G. Gerber, M.D.,** John S. Barnes, B.S., Gary 0. Zerbe, Ph.D., and Alan S. Nies, M.D. Denver, Colo.

Sulindac (Clinoril; Merck, Sharp & Dohme) is a

newer nonsteroidal anti-inflammatory drug (NSAID) reported to differ from other NSAIDs in that it spares renal prostaglandin (PG) synthetase.578' This would have a potential advantage in some disease states such as severe hepatic dysfunction, in which renal PGs are important in the maintenance of renal blood flow (RBF) and glomerular filtration rate (GFR). However, before clear recommendations can be made about the use of

From the Division of Clinical Pharmacology, Departments of Med- icine and Pharmacology, and Department of Preventive Medicine and Biometrics, University of Colorado School of Medicine.

Supported in part by Merck, Sharp & Dohme Research Laboratories, West Point, Pa., General Clinical Research Center Program Grant No. M01 RR 00051-22 from the Division of Research Resources, National Institutes of Health (NIH), and Grant No. HL21308 from the Heart, Lung and Blood Institute, NIH.

Received for publication Feb. 20, 1985; accepted June 8, 1985.

Reprint requests to: John G. Gerber, M.D., University of Colorado Health Sciences Center, Box C-237, 4200 E. Ninth Ave., Denver, CO 80262.

*Supported by NIH Training Grant in Clinical Pharmacology No. GM07063-09.

**Established Investigator of the American Heart Association.

sulindac in disease states, the evidence about its renal sparing properties should be unequivocal. In this report we show that sulindac is not specifically renal sparing; therefore, caution should continue to be exercised with its use in situations in which impairment of renal func- tion with NSAIDs may be expected.

METHODS Subjects. This study was performed in the Clinical

Research Center of the University of Colorado Health Sciences Center after approval by the Human Subjects Committee. Informed consent was obtained from all subjects.

Subjects were 15 normal, nonsmoking women be- tween the ages of 18 and 55 years, within 15% of ideal weight for height,' with no active illness or any history of renal disease, and with normal physical examination results. All had normal laboratory screening results in- cluding a complete blood count, differential and platelet count, levels of electrolytes, fasting blood sugar, BUN, creatinine, SGOT, alkaline phosphatase, total bilirubin, and total protein, urinalysis, urine culture, and a neg- ative test for pregnancy. They had to be drug free for 2 weeks before the study.

VOLUME 38 NUMBER 3

Table I. Subject characteristics

The study lasted 10 days and was timed to begin within 10 days after menstrual onset. Throughout the study, all subjects took individually labeled, identically appearing packets of drug at 7 AM and 1, 7, and 11

PM. The study was divided into two sequential 5-day periods. During the first 5 days (control) all subjects received placebo (single blind), while during the second 5 days (trial) the subjects were randomly assigned (dou- ble blind) to placebo, sulindac, 200 mg b.i.d. (7 AM

and 7 Pm), or indomethacin, 25 mg q.i.d. There were five subjects in each treatment group.

During the study, a 50 mEq sodium and 60 mEq potassium per day diet was provided and vital signs and weight were determined daily. Twenty-fourhour urine samples were collected daily for determination of electrolytes, creatinine, and osmolality. On days 3

and 8, an aliquot was screened for drugs other than sulindac or indomethacin. On days 4 and 9 the ex- cretion rates of PGE2 and the tetranor metabolite of PGF, (5a ,7a-dihydroxy-11-ketotetranorprostane- 1 ,16-dioic acid; PGF-M) were determined.

The control and trial periods each consisted of 3 days as an outpatient and 2 days as an inpatient. As inpa- tients, the subjects were required to be flat in bed and fasted from midnight until 1 PM the following day. At 6:30 AM, the subjects were given their morning med- ication with 200 ml water. A bladder catheter was placed and loading and maintenance intravenous doses of inulin (50 mg/kg; 30 mg/min) and p-amino hippurate (PAH; 120 mg; 8 mg/min) were given. After allowing 45 minutes for equilibration, a baseline (control) 45- minute urine collection with a midpoint blood collection was obtained. Furosemide, 40 mg, was then intrave-

Sulindac is not renal sparing in man 259

nously injected over 5 seconds, and serial urine and blood samples were collected until 1 PM (at 0 to 15, 15 to 30, 30 to 60, 60 to 120, and 120 to 300 minutes after furosemide). Measured water, sodium, and po- tassium losses were replaced intravenously in the third hour. The urine collection on days 5 and 10 began at 1 PM. Subjects were observed overnight and discharged the next morning.

Baseline measurements included platelet aggregation and thromboxane B2 generation, plasma drug concen- trations, inulin clearance (GFR), PAH clearance (to es- timate renal plasma flow), plasma renin activity (PRA), and the excretion rates of PGE, and electrolytes. All de- terminations except drug levels and the platelet studies were repeated at various times after furosemide dosing.

Assays. Sodium and potassium were measured by ion exchange on a Beckman Electrolyte 2 Analyzer (Beckman Instruments Inc.); bilirubin was measured on a Beckman 24 Spectrophotometer; glucose was ana- lyzed with a Beckman Glucose 2 Analyzer; osmolality was determined by freezing-point depression with an Advanced Instruments, Inc. model 3D II; and inulin and PAH were analyzed with an Autoanalyzer 1 (Techni- con Instruments Corp.).' All other chemistry param- eters were measured with a System 203S Analyzer (Gil- ford Instrument Laboratories Inc.). PRA was deter- mined by standard methods.' Urinary PGE2 was ex- tracted into dichloromethane at pH 3 and measured by RIA with an antibody from Institut Pasteur, Paris.' Urinary PGF-M was measured by RIA with an anti- body.* Standard and tritiated PGF-M were produced

*Obtained from Dr. E. Granstrom.

Placebo Sulindac Indomethacin

Age (yr) 33 -± 5 35 ± 2 30 -± 3

Height (cm) 165 ± 1 161 ± 1 165 -± 3

Systolic blood pressure (mm Hg) Entry 112 -± 2 106 ± 7 108 ± 5

Day 4 96 ± 4 105 ± 4 103 -± 4 Day 9 99 -± 6 102 ± 3 98 -± 3

Diastolic blood pressure (mm Hg) Entry 76 ± 5 72 ± 5 77 ± 3

Day 4 68 ± 1 75 ± 3 75 ± 5

Day 9 71 ± 5 72 -± 4 67 ± 1

Weight (kg) Entry 61.6 ± 2.7 54.9 -± 2.3 60.2 ± 3.1 Day 4 59.9 ± 2.5 54.0 ± 2.2 58.8 ±- 2.9 Day 9 59.2 ± 2.4 54.3 -± 2.1* 59.1 ± 3.0*

Data are TC SE. *P < 0.05 compared with placebo.

260 Roberts et al.

150

E 50

100

Placebo ± S.E.M. a Sulindac o Indomethacin

- ---- -p 2 _ce

0 4 (5) 6 7 8 9 ( 10)

Day

Fig. 1. Daily urinary sodium excretions for groups. All groups

started on placebo; drug groups were crossed over on day 6

to either sulindac or indomethacin. Days 5 and 10 are cor-

rected to a 24-hour equivalent (see text).

by a variation of published methods'''. from 13,14-di- hydro-15-keto-PGF2r, (tritiated from Amersham Corp.; cold from Cayman Chemical). Modifications included addition of adenosine diphosphate (ADP), 0.5 mmol/ L, and D,L-carnitine, 0.24 mmol/L, to the reaction mix- ture, extraction of PGF-M at pH 3 into ethyl acetate, and purification of the evaporated residue by reversed- phase HPLC. The presence of PGF-M was confirmed by gas chromatography/mass spectrometry. Platelet ag- gregation studies were performed with platelet-rich plasma in a Payton aggregometer with collagen and ADP as the proaggregatory agents.' The percent ag- gregation was determined 4 minutes after addition of collagen or ADP. Thromboxane B, generated in the platelet-rich plasma was measured by RIA with an an- tibody from Seragen.

Plasma drug concentrations were measured in du- plicate by a modification of current methods:5'35 Briefly, 200 pA plasma was mixed with 10 ill ethanolic internal standard (indomethacin [Sigma Chemical Co.] for the sulindacs; sulindac sulfoxide [Merck, Sharpe & Dohme] for indomethacin) and 200 p.1 750 mmol/L lactic acid buffer at pH 3.0. The mixture was extracted twice with 4 ml CH2C12, dried under nitrogen at 40° C, and taken up in 100 p,1 mobile phase. Reversed-phase HPLC was performed with a mobile phase of 40% to 60% acetonitrile in 45 mmol/L aqeuous phosphate buffer at pH 3.0, and a nonlinear gradient (no. 7 setting Waters Model 660 Solvent Programmer, slow upstroke; Waters Associates, Inc.) at a flow rate of 1.5 ml/min over 20 minutes with ultraviolet detection at 327 nm

CAN PHARMACOI. THER SEPTEMBER 1985

(Waters Model 440 Variable Wavelength UV Spectro- photometer).

Statistics. Data were analyzed by multivariate one- way ANOVA (three levels) appropriate for analysis of repeated measurements over time.1132 Intragroup com- parisons between trial and control periods and com- parisons between groups were determined. The latter allows direct drug-to-drug comparison. Comparisons of drug to placebo are one tailed. Comparisons between drugs or between placebo periods are two tailed. All data are expressed as the X -± SE. A P value <0.05 was considered significant.

RESULTS The entry profile and the effect of any therapy on

weight and blood pressure are listed in Table I. The groups were comparable in age, weight, and blood pres- sure. Systolic and diastolic sitting blood pressures gen- erally fell over the course of the study, but there were no significant differences between day 9 and day 4

blood pressures in any group. Subjects receiving pla- cebo showed statistically significant weight loss during both control and trial periods of the low-salt diet. Drug groups showed weight loss only for the first (control) half, with no weight loss during drug dosing. Intergroup analysis showed that the absence of continued weight loss while taking sulindac and indomethacin was sig- nificantly different from placebo, but the two drugs did not differ from each other.

Fig. 1 shows daily outpatient urinary sodium excre- tion for each group. The values for inpatient days 5 and 10 do not include the acute study period and are cor- rected to a 24-hour equivalent. At the start of the control period, sodium excretion differed between groups and probably reflected group differences in the subject's diets. The subjects achieved sodium balance, as sodium excretion equilibrated at about 40 mEq/day by day 4

following the standard diet. For statistical analysis, the nonequilibrium period (days 1 through 3) was excluded. During the trial period, sodium excretion decreased compared with control values for both indomethacin (P < 0.01) and sulindac (P < 0.05) but not for pla- cebo. There was no significant difference between in- domethacin and sulindac. Chloride excretion (data not shown) was similar to sodium excretion. Potassium ex- cretion (data not shown) did not show as marked an initial equilibrium phase, and only the sulindac group showed a significant (P = 0.01) decrease, which was relatively small (from 39 -± 4 to 31 2 mEq).

The 24-hour urinary excretion of PGE, was compared for day 4 (control) and day 9 (trial period; Table II). The control values between the groups did not differ

VOLUME 38 NUMBER 3

Table II. Effects of sulindac and indomethacin on excretion of PGE, and PGF-M, urine volume, and on platelet function

Data are X -± SE. *P < 0.05 compared with placebo. tP < 0.05 compared with sulindac.

significantly. During the trial period the placebo group showed a decrease in PGE2 excretion of 24% (not significant), whereas the indomethacin and sulindac groups showed significantly decreased PGE, excretion, by 78% and 49%, respectively. By intergroup com- parison, the effect of indomethacin was greater than that of sulindac (P < 0.05).

Urinary excretion of PGF-M was examined on days 4 and 9 (Table II). Both indomethacin and sulindac decreased PGF-M excretion as compared with the sub- jects' own control value (P < 0.05), while placebo did not. By intergroup comparison the drugs did not differ from each other.

ADP-stimulated platelet aggregation was minimally affected by sulindac and indomethacin. Collagen-stim- ulated aggregation was inhibited to a small (11%) but significant extent by sulindac (P < 0.05), and to a

greater extent (54%) by indomethacin (P < 0.01). In- tergroup comparison showed that the effect of sulindac on collagen-induced platelet aggregation was signifi- cantly less than that of indomethacin (P < 0.01). The small amount of thromboxane released from platelets with ADP was suppressed by both sulindac and indo- methacin. With collagen-induced platelet aggregation, which causes much greater synthesis of thromboxane

Sulindac is not renal sparing in man 261

B2, sulindac decreased thromboxane B, release by 31% (P < 0.05), while indomethacin decreased thrombox- ane B, release by 97% (P < 0.01). Intergroup drug comparison revealed that the effect of indomethacin was significantly greater than that of sulindac (Table 11).

The response to furosemide is shown in Figs. 2 and 3. Neither sulindac nor indomethacin had great effects on the natriuretic response to furosemide. Compared with the placebo group, sulindac significantly blunted urinary sodium excretion (P < 0.05). Indomethacin had a similar effect (P = 0.07; Fig. 2). Furosemide induced a large increase in the urinary excretion of PGE2, which was significantly inhibited by both su- lindac and indomethacin (Fig. 2), and the drug effects did not differ. Only indomethacin decreased basal PRA, but 20 minutes after furosemide both sulindac and in- domethacin inhibited the rise in PRA, an effect signif- icantly different from placebo (Fig. 3). Although at 60 minutes the rise in PRA continued to be inhibited by both sulindac and indomethacin, the difference from the placebo group was of borderline significance (P = 0.07). Neither sulindac nor indomethacin af- fected PAH clearance during baseline or after furose- mide. The GFR before furosemide was not affected by sulindac or indomethacin. However, compared with the

Placebo Sulindac lndomethacin

PGE excretion (ng/gm creatinine) Control (day 4) 96 ± 8 91 ± 10 127 -± 28 Drug (day 9; % control) 73 ± 7 (76%) 46 ± 6 (51%)* 28 ± 3 (22%)*t

PGF-M excretion (rig/gm creatinine) Control 11.6 ± 1.0 10.5 ± 0.9 16.2 ± 2.8 Drug (% control) 11.1 -± 1.1 (96%) 6.8 ± 0.6 (65%)* 9.2 -± 0.4 (57%)*

24-hr urine volume (L) Control 1.89 ± 0.39 2.30 ± 0.64 1.76 -± 0.22 Drug 1.82 ± 0.46 2.45 -± 0.71 1.48 ± 0.19

Platelet aggregation (% maximum) Collagen (5 ig/m1)

Control (day 5) 73 ±- 3 76 ± 2 74 -± 3 Drug (day 10; % control) 73 ± 5 (99%) 68 -± 5 (89%)* 34 ± 8 (46%)*t

ADP (10' mol/L) Control 65 ± 4 66 ± 3 60 ± 8 Drug (% control) 62 ± 5 (96%) 56 ± 2 (85%)* 55 ± 2 (90%)

Platelet thromboxane B, release (ng/ml) Collagen

Control (day 5) 42 ± 7 41 -± 8 53 ± 10 Drug (day 10; % control) 45 ± 8 (108%) 28 -± 7 (69%)* 2 -± 2 (3%)*t

ADP Control 17 ± 10 6 ± 2 11 ± 6 Drug (% control) 13 ± 8 (76%) 0.3 -± 0.3 (5%)* Below detection (0%)

262 Roberts et al.

er2.5 Ui 0

g 2.5

2.0

Id 1.5

g 1.0 t 0.5

:13

.E 5 2.5

2.0

1.5

INDOMETHACIN GROUP 100

SULINDAC GROUP

0.5

-0' 80 60 120 180 240 Time, minutes after Furosemide

Fig. 2. Left, Timed urinary sodium excretions after furosemide on control day 5 (D) and trial day 10 (0). Right, Urinary PGE, excretion after furosemide on control day 5 () and trial day 10 (0). Bars represent standard error of the means. Solid lines represent placebo, dashed lines represent sulindac, and dotted lines represent indomethacin.

placebo group, the furosemide-induced changes in GFR were reduced by indomethacin for the first three col- lection periods (0 to 15, 15 to 30, and 30 to 60 minutes) and by sulindac at the 15-to-30--minute collection period.

Plasma drug concentrations were as follows: sulindac sulfide (active metabolite), 7.7 ± 1.9 1.1g/m1; sulindac (inactive parent compound), 6.0 -± 2.4 p.g/m1; sulin- dac sulfone (inactive metabolite), 3.9 ± 0.7 Rg/m1; and indomethacin, 1.9 ± 0.4 pig/ml. These peak levels were in the expected range, and all subjects were taking their assigned drugs. Urine drug screening results did not show the presence of salicylates or other drugs.

DISCUSSION NSAIDs are useful for the relief of fever, pain, and

inflammation. Recent evidence has implicated these drugs as causing acute renal failure in clinical condi- tions that are associated with glomerular disease or cir- culatory stresses in which both the sympathetic nervous system and the renin-angiotensin system are acti- vated. '''''°-"'" PGs have been demonstrated to modulate the renal effects of these endogenous vasoconstrictors. It is thought that inhibition of PGs by NSAIDs in these settings results in unopposed renal vasoconstriction by angiotensin II and norepinephrine, resulting in acute renal failure.'

The development of an NSAID that would spare renal cyclooxygenase would be an important advance in ther- apeutics. Sulindac has been reported to be renal spar-

300

200

400

300 et.

0.1

200 CO 0. >1 100

0 300

200

1NDOMETHACIN GROUP

SULINDAC GROUP

PLACEBO GROUP

30 60 120 180 240 Time, minutes after Furosemide

GUN PHARMACOI, THEE SEPTEMBER 1985

2.0

1.5

1.0

0.5

ing.578'" Sulindac is a structural analog of indometh- acin, but is inactive as dosed and must be reduced to the sulfide form to be active.' Experiments in rat and rabbit kidneys have shown that the kidney has the en- zymatic machinery to oxidize sulindac sulfide to the inactive forms (the sulfoxide and sulfone).' Thus the kidney may be exposed to less of the active form than other organ systems. Consistent with this hypothesis is the finding that little active sulfide is excreted in the urine.' On the other hand, the sulfide is present in the circulation and in the renal tissue itself,' and thus should be able to affect PG synthesis in the renal vas- culature. In addition, infusion of sulindac sulfide into the isolated kidney results in PG inhibition even though no sulfide appears in the venous effluent.' Nonetheless, recent clinical reports have suggested that sulindac may indeed be safe in situations in which other NSAIDs have caused a reduction in renal function.'''

Because of the importance of this question of selec- tive cyclooxygenase inhibition, we have compared the effect of sulindac with those of indomethacin and pla- cebo given for 5 days to normal women.' The doses of the drugs were those used clinically, and for sulindac the dose used was that reported to have no effect on renal PG synthesis in normal subjects or patients with renal disease. We used collagen-stimulated platelet thromboxane B2 release as well as the urinary excretion of PGF-M as indices of systemic cyclooxygenase in- hibition, and urinary PGE, excretion as a measure of renal cyclooxygenase inhibition. In addition, we looked

1.0 PLACEBO GROUP 100

VOLUME 38 NUMBER 3

at the effects of these drugs on furosemide-stimulated PGE, production and renin release as additional tools to evaluate effects of the drug on renal cyclooxygenase.

Our findings indicate that sulindac does not appear to be renal sparing. The data for this conclusion are several. First, both sulindac and indomethacin induced sodium retention on a moderately sodium-restricted diet and prevented the continued weight loss that occurred in the placebo group. This clinically important effect may reflect either vascular or tubular consequences of inhibiting renal cyclooxygenase. Second, both drugs reduced urinary PGE2 excretion. Urinary PGE, excre- tion is considered to reflect de novo renal PG synthesis, whereas the urinary excretion of the prostacyclin me- tabolite, 6-keto-PGF,, may derive from systemic as well as renal sources. '72831 Third, the response to fu- rosemide was similarly altered by the two drugs. Fu- rosemide resulted in a rapid increase in PGE2 excretion that was inhibited by sulindac and indomethacin. The increase in renin activity after furosemide, which is

dependent on PG synthesis, also was inhibited by both sulindac and indomethacin. Finally, sodium excretion after furosemide was decreased to a small extent by sulindac and probably by indomethacin as well.

We have evidence that indomethacin may have had greater effects than sulindac on some indices of renal PG effect (basal PRA, PGE, excretion, and GFR). However, the small group size of our study precludes determination of small differences between the drugs. The size of the groups was chosen as the minimum number to confirm the large differences in the drugs that had been reported in the literature. Clearly, such large differences did not exist in our study.

Were the systemic effects of sulindac and indometh- acin comparable? The doses used and the blood levels achieved were those associated with clinical anti-in- flammatory effects. Compliance was good as judged by these drug concentrations. Effects of the drugs on the renal excretion of PGF-M were comparable. The PGF- M arises from hepatic metabolism of systemically pro- duced PGF2. A reduction in the urinary excretion of PGF-M, therefore, is indicative of a reduction of sys- temic PGF2 production. Both indomethacin and sul- indac had moderate effects on this index of systemic PG synthesis. As another index of systemic effect, col- lagen- and ADP-induced platelet aggregation and thromboxane production were investigated. ADP has less effect than collagen on the stimulation of platelet aggregation through thromboxane synthesis, and thus to discern differences between cyclooxygenase inhibi- tors, collagen-induced platelet aggregation may be pre- ferred.''' Indomethacin had a greater effect than sul-

Sulindac is not renal sparing in man 263

Time, minutes after Furosemide

Fig. 3. PRA on control day 5 () and trial day 10 (0). Bars represent standard error of the means. Solid lines represent placebo, dashed line represents sulindac, and dotted line rep- resents indomethacin.

indac on collagen-induced platelet aggregation and thromboxane production. This implies that either the platelets are spared by sulindac, or that indomethacin had a greater systemic effect than sulindac.

The literature on the renal effects of sulindac is note- worthy for discrepant results .4,7,8,16,23,29.30.37 Although study protocols vary, some of the different results re- main unexplained. A major question is whether sulindac has systemic effects comparable with those of the other NSAIDs in the various studies. Different methods have been used to evaluate the systemic effects of NSAIDs, but there is no concensus as to their relative meaning or which may be best. We have attempted to assess systemic effect by measuring the excretion of a systemic PG metabolite and by measuring collagen-induced platelet aggregation and thromboxane B2 production. Others have used ADP-induced platelet aggregation, which is not dependent on cyclooxygenase, or ADP- induced thromboxane B2 production, and thus is an insensitive measure of cyclooxygenase, since only a small amount of thromboxane is formed. Still others have used thromboxane production by clotted blood, an indirect measure, the meaning of which relative to platelet thromboxane production induced by a stan- dardized stimulus such as collagen is not known.

A final source of differences between studies are

c_ _c

5.0

INDOMETHACIN GROUP

.c c7) o -- 2

0 0 --6) 5.0 SULINDAC GROUP

-------------- -- - 0

-- >

0

5.0 PLACEBO GROUP

CC

co

2

0.. 0 0 20 60

264 Roberts et al.

variations in the PG assays because of different anti- bodies and different extraction techniques. Our assay for urinary PGE2 has been confirmed by gas chroma- tography-mass spectrometry over a wide range of urine flow and PGE2 excretion rates.' We have also subjected some of the urine samples from subjects receiving pla- cebo, indomethacin, and sulindac to TLC' and consis- tently found 90% of the PGE2 immunoreactivity to co- migrate with authentic PGE2. It has been reported that urine flow rate can affect PGE2 excretion acutely.' However, the change in urine flow after furosemide was not affected by the NSAIDs, and so the decreases of 70% to 90% in PGE2 excretion cannot be accounted for by this mechanism. Similarly, the 24-hour urine volumes were not changed by indomethacin or sulindac (Table II), although there was a significant decrease in PGE2 excretion by both drugs.

Recently, Brater et al.4 and Sedor et al.' have pub- lished studies similar to ours comparing the renal and extrarenal effects of NSAIDs to those of sulindac. In the first study, ibuprofen and naproxen were compared with sulindac after administration of the drugs over a

12-hour period. Differences in terms of the renal and platelet effects were very minor among the three drugs. Brater et al. concluded that sulindac was no more renal sparing than the other two NSAIDs. In the study by Sedor et al., the drug protocol was similar to ours in

that sulindac, 200 mg b.i.d., was compared with in- domethacin, 25 mg q . i.d. Quantitatively, indomethacin was more potent in inhibiting renal excretion of PGF2, although clearly sulindac had an effect as well. In terms of renal function, neither drug had a major effect in this healthy female population. In terms of platelet ef- fect, sulindac and indomethacin were equivalent in al- tering ADP-stimulated thromboxane B2 release in plate- let-rich plasma. However, as discussed above, ADP is not a potent stimulator of platelet thromboxane release, and consequently this model may not be the most sen- sitive method to assess systemic PG inhibition. Another difference between our data and those of Sedor et al. is that in our subjects the peak sulindac sulfide con- centration was 7.7 ± 1.9 Rg/ml, while Sedor et al. reported a value one third this concentration.

Our data from normal subjects may not apply to subjects with renal disease or those with very high levels of circulating angiotensin II or norepinephrine. We used a modest degree of sodium restriction and furosemide dosing to stimulate renal PG production to simulate the conditions that might exist in disease states. However, the fact that we saw significant effects of sulindac on renal function and cyclooxygenase under these mild conditions cannot be ignored. One would expect pa-

CI,IN PHARMACOL THER SEPTEMBER 1985

tients with renal disease, severe cirrhosis with ascites, dehydration, congestive heart failure, or nephrosis to be at least as sensitive to the effects of sulindac as normal subjects. Caution should therefore be applied with the use of sulindac, as with the use of other NSAIDs, in these conditions.

We thank Dr. Elisabeth Granstrom, Stockholm, for PGF- M antibody; Drs. Bob Strife and Paul Fennessey for gas chromatography/mass spectrometry confirmation of PGF-M; and Dr. Athanassios Chremos and Ms. Mary Ann Tupy-Visich of Merck, Sharp and Dohme Research for administrative as- sistance.

References Ariz L, Donker AJM, Brentjens JRH, van der Hem GK: The effect of indomethacin on proteinuria and kidney function in the nephrotic syndrome. Acta Med Scand 179:121-125, 1976. Barnes JS, Dubocovich ML, Nies AS, Gerber JG: Lack of interaction between tricyclic antidepressants and clon- idine at the alpha2-adrenoceptor on human platelets. CLIN

PHARMACOL THER 32:744-748, 1982. Boyer TD, Zia P, Reynolds TB: Effect of indomethacin and prostaglandin A, on renal function and plasma renin activity in alcoholic liver disease. Gastroenterology 77:215-222, 1979. Brater DC, Anderson S, Baird B, Campbell WB: Effects of ibuprofen, naproxen and sulindac on prostaglandins in men. Kidney Int 27:66-73, 1985. Bunning RD, Barth WF: Sulindac, a potentially renal- sparing non-steroidal anti-inflammatory drug. JAMA 248:2864-2867, 1982. Charo IF, Feinman RD, Detwiler TC: Interrelations of platelet aggregation and secretion. J Clin Invest 60:866- 873, 1977. Ciabattoni G, Cinotti G, Pierucci A, et al: Effects of sulindac and ibuprofen in patients with chronic glomer- ular disease. N Engl J Med 310:279-283, 1984. Ciabattoni G, Pugliese F, Cinotti GA, Patrono C: Renal effects of anti-inflammatory drugs. Eur J Rheum Inflam 3:210-221, 1980. Ciabattoni G, Pugliese F, Spaldi M, Cinotti GA, Patrono C: Radioimmunoassay measurement of prostaglandin E, and F2 human urine. J Endocrinol Invest 2:173-182, 1979. Clive DM, Stoff JS: Renal syndromes associated with non-steroidal antiinflammatory drugs. N Engl J Med 310:563-572, 1984. Cole JWL, Grizzle JE: Applications of multivariate anal- ysis of variance to repeated measurements experiments. Biometrics 22:810-828, 1966. Dray F, Charbonnel B, Maclouf J: Radioimmunoassay of prostaglandins F2, E, and E, in human plasma. Eur J

Clin Invest 5:311-318, 1975. Duggan DE, Hare LE, Ditzler CA, Lei BW, Kwan KC:

VOLUME 38 NUMBER 3

Disposition of sulindac. CLIN PHARMACOL THER 21:326- 335, 1977. Duggan DE, Hooke KF, Risley EA, Shen TY, VanArman CG: Identification of the biologically active form of sul- indac. J Pharmacol Exp Ther 201:8-13, 1977. Dusci LJ, Hackett LP: Determination of sulindac and its metabolites in serum by high performance liquid chro- matography. J Chromatogr 171:49-93, 1979. Ferguson RJ, Vlasses PH, Riley LJ, et al: Sulindac and ibuprofen inhibit IV furosemide-stimulated renin release but not natriuresis in sodium-replete men. CLIN PHAR-

MACOL THER 35:239, 1984. (Abst.) Frolich JC, Wilson TW, Sweetman BJ, et al: Urinary prostaglandins: Identification and origin. J Clin Invest 55:763-770, 1975. Fuhr F, Kaczmarczyk J, Kruttgen C-D: Eine einfache colorimetrische Methode zur Inulinbestimmung fur Ni- eren-clearance-untersuchungen bei stoffwechselgesun- den und Diabetikem. Klin Wochen 33:729-730, 1955. Gerber JG, Olson RD, Nies AS: Interrelationship be- tween prostaglandins and renin release. Kidney Int

19:816-821, 1981. Gerrard JM: Platelet aggregation and the influence of prostaglandins. In Lands WEM, Smith WL, editors: Methods in enzymology, vol 86. New York, 1982, Ac-

ademic Press, pp 642-654. Granstrom E. Kindahl H: Radioimmunoassay of 504,7a- dihydroxy- I 1- ketotetranorprostane- I ,I6-dioic acid, a

major prostaglandin F metabolite in blood and urine. In Lands WEM, Smith WL, editors: Methods in enzymol- ogy, vol 86. New York, 1982, Academic Press, pp 339-356. Hamberg M: Metabolism of prostaglandins in rat liver mitochondria. Eur J Biochem 6:135-146, 1968.

Harvey RB, Brothers AJ: Renal extraction of para-ami- nohippurate and creatinine measured by continuous in vivo sampling of arterial and renal-vein blood. Ann NY Acad Sci 102:46-54, 1962. Kronborg I, Daskalopoulos D, Katkov W, Zipser RD: Influence of sulindac and indomethacin on renal function and furosemide-induced decreases in patients with cir- rhosis and ascites. Clin Res 32:14A, 1984. (Abst.) Miller MJS, Bednar MM, McGiff JC: Renal metabolism of sulindac: Functional implications. J Pharmacol Exp Ther 231:449-456, 1984. Patrono C, Wennmalm A, Ciabattoni G, Norvak J. Pug- hese F, Cinotti GA: Evidence for extra-renal origin of urinary prostaglandin E., in healthy men. Prostaglandins 18:623-629, 1979.

Sulindac is not renal sparing in man 265

Roberts DG, Strife RJ, Gerber JG, Murphy RC, Nies AS: Effect of urine flow on prostaglandin E2 excretion in man. Clin Res 31:440A, 1983. (Abst.) Rosenkranz B, Kitajima W, Frolich JC: Relevance of urinary 6-keto-prostaglandin Fro, determination. Kidney Int 19:755-759, 1979. Ryan JR, Chremos AN, Vargas R, Johnson CL, Mc- Mahon FG: The influence of nonsteroidal antiinflam- matory drugs (NSAID) [Sulindac (S) and naproxen (N)] in natriuretic, diuretic and renin (PRA) response to fu- rosemide (F). CLIN PHARMACOL THER 34:271, 1984. (Abst.) Sedor JR, Williams SL, Chremos AN, Johnson CL, Dunn MJ: Effects of sulindac and indomethacin on renal prostaglandin synthesis. CLIN PHARMACOL THER 36:85- 91, 1984. Seyberth HW, Muller H, Solding K, Wille L, Hackenthal E: Urinary excretion rate of 6-keto-prostaglandin F, as an index of circulating prostacyclin. In Samuelsson B, Paoletti R, Ramwell P, editors: Advances in prostaglan- din, thromboxane, and leukotriene research, vol 11. New York, 1983, Raven Press, pp 533-538. Starmer CF, Grizzle JE: A computer program for analysis of data by general linear models. Institute of Statistics Mimeograph Series No. 560. Chapel Hill, NC, 1968,

University of North Carolina. Statistical Bulletin No. 40, November-December, 1959, Metropolitan Life Insurance Company, New York, NY.

Stockigt J, Collins R, Biglieri E: Determination of plasma renin concentration by angiotensin I immunoassay. Di- agnostic impact of precise measurement of subnormal renin in hyperaldosteronism. Circ Res 28/29(suppl II):

175-189, 1971. Swanson BN, Boppana VK: Measurement of sulindac and its metabolites in human plasma and urine by high performance liquid chromatography. J Chromatogr 225:123-140, 1981.

Walshe JJ, Venuto RC: Acute oliguric renal failure in-

duced by indomethacin: Possible mechanism. Ann Intern Med 91:47-49, 1979. Wong DG, Spence JD, McDonald JWD: Nonsteroidal antiinflammatory drugs (NSAID) vs. placebo in hyper- tension treated with diuretics and p-blocker. CLIN PHAR-

MACOL THER 35:284, 1984. (Abst.) Zipser RD, Hoefs JC, Speckart PF, Zia PK, Horton R:

Prostaglandins: Modulators of renal function and pressor resistance in chronic liver disease. J Clin Endocrinol Metab 48:895-900, 1979.