Effects of phenylbutazone, firocoxib, and dipyrone on the diuretic response to furosemide in horses

1 INTRODUCTION

Nonsteroidal anti-inflammatory drugs (NSAIDs) are administered to horses to alleviate pain, fever, and inflammation; however, excessive or prolonged administration can have adverse effects on the gastrointestinal tract (GIT) and kidneys.¹ NSAIDs inhibit cyclooxygenase (COX) enzymes that catalyze oxidation of arachidonic acid to prostaglandins (PGs). There are several COX isoforms and COX-1 derived PGs are traditionally viewed as homeostatic, while COX-2 is considered an “inducible” isoform upregulated with inflammation.²

In the kidney, however, both COX-1 and COX-2 have “constitutive” expression and activity.³ COX-1 is the predominant isoform in glomerular mesangial cells, arteriolar endothelial cells, and collecting duct epithelium in most species, while COX-2 is found in glomerular cells, medullary interstitial cells, macula densa (MD) cells, and tubular epithelium in the thick ascending limb (TAL).³ COX-1 derived PGE₂ attenuates afferent arteriole vasoconstriction in response to decreased renal perfusion and antagonizes the antidiuretic effect of arginine vasopressin (AVP) in collecting ducts, and medullary COX-2 PGE₂ production is increased to maintain perfusion of the normally hypoxic medulla during periods of renal hypoperfusion.^{3, 4} The overall effect of renal COX activity is to assist with maintenance of renal blood flow (RBF), glomerular filtration rate (GFR), and urine output during normal physiologic conditions and to preserve RBF, especially to the inner medulla, and GFR during renal hypoperfusion.^{2, 3}

Furosemide is a loop diuretic that causes saluresis, primarily by blocking the apical Na⁺/K⁺/2Cl⁻ cotransporter (NKCC-2) in the TAL of the loop of Henle.^{5, 6} In addition, furosemide activates COX isoforms in different nephron segments. For example, furosemide-induced COX-1 PGE₂ production antagonizes the effect of AVP to increase water reabsorption in collecting ducts⁴ while furosemide-induced COX-2 PGE₂ production in MD cells can lead to further inhibition of tubular reabsorption of Na⁺ and Cl⁻ in the TAL.⁷

Nonselective NSAIDs phenylbutazone and meloxicam inhibit both COX-1 and COX-2, and pretreatment with phenylbutazone attenuates the diuretic response to furosemide by ~30%.^8-10 Treatment of horses with meloxicam, a COX-2 preferential NSAID that becomes nonselective at higher plasma concentrations, produces a similar inhibition of furosemide-induced diuresis.¹⁰ Firocoxib, a second generation NSAID highly selective for COX-2 (a COXIB), was approved for use in horses about a decade ago with a goal of providing a similar analgesic effect as phenylbutazone with less risk of adverse GIT effects.¹¹ However, use of COXIBs does not appear to confer “renoprotection” in humans.^{12, 13} COXIBs demonstrated attenuation of furosemide diuresis, comparable with nonselective COX inhibitors in cats and rats.^{14, 15} Dipyrone, termed an atypical NSAID because of both PG dependent (through COX-3 activity) and independent effects, is approved as an antipyretic agent for use in horses.¹⁶ However, use of dipyrone in the intensive care unit has been associated with a dose-dependent increased risk of acute kidney injury (AKI) in people.¹⁷ The effects of firocoxib and dipyrone on the renal responses to furosemide in horses are unknown. The purpose of this study was to evaluate furosemide-induced diuresis after pretreatment with firocoxib or dipyrone, as compared to pretreatment with phenylbutazone. We hypothesized that there would be less inhibition of furosemide-induced diuresis and natriuresis after pretreatment with firocoxib or dipyrone, as compared to phenylbutazone.

2 MATERIALS AND METHODS

3 RESULTS

All mares tolerated restraint in stocks and instrumentation without sedation during the experiments. Three of the 32 experiments had to be repeated because of dislodging of a ureteral catheter (n = 2) and leakage of urine around a ureteral catheter into the bladder (n = 1). No changes in HR, either in response to furosemide administration over time (P = .24) or between treatments (P = .73), were observed. After furosemide administration PCV was increased (P < .001) from the −15 minutes value after 45, 60, and 180 minutes and TS was increased (P < .001) from the −15 minutes value through 240 minutes. However, there were no differences between treatments (Figure 1). Serum Na⁺ concentration was increased (P < .001) from the −15 minutes value at 45 minutes after furosemide administration while serum K⁺ was decreased (P < .001) from the −15 minutes value at 45 and 90 minutes after furosemide administration. After furosemide administration serum Cl⁻ was decreased (P < .001) from the −15 minutes from 45 through 240 minutes (Figure 2).

The total volume of urine produced after furosemide administration was ~25% greater (P < .001) with FU, as compared to PB, FX, and DP, and there was no difference in urine volume produced between the 3 NSAID treatments (Table 1). There was no difference (P = .27) in total water intake between the treatments: FU 6.4 ± 3.5 L; PB 6.4 ± 4.1 L; FX 7.0 ± 5.5 L; and DP 4.5 ± 4.5 L. With all treatments, USG was decreased (P < .001) from the −30 to 0 minute collection period value during all collection periods after furosemide administration and there were no differences in USG between treatment groups (Figure 3).

Urine Na⁺ concentration was increased (P < .001) from the −30 to 0 minute collection period value during the 0 to 60 and 60 to 120 minutes collection periods while urine K⁺ and Cl⁻ concentrations were decreased (P < .001) from the −30 to 0 minute collection period values during all subsequent collection periods after furosemide administration. There were no differences in urine electrolyte concentrations between treatment groups during any collection period (Figure 4). Total excretion of Na⁺ and Cl⁻, not standardized for bodyweight, during the 4 hours after furosemide administration was ~25% and ~13% greater (P < .001 for both) with FU, as compared to PB, FX, and DP, and there were no differences between the 3 NSAID treatments (Table 1). Although total excretion of K⁺, again not standardized for bodyweight, during the 4 hours after furosemide administration was ~10% greater for FU, as compared to PB, FX, and DP, this was not a significant finding (P = .63) and there were no differences in total K⁺ excretion between the 3 NSAID treatments.

Total Cr excretion during the 4 hours after furosemide administration was not different (P = .23) between treatment groups (Table 1). When serum and urine Cr and electrolyte concentrations were used to calculate fractional clearances of Na⁺, K⁺, and Cl⁻ (FCl_Na, FCl_K, and FCl_Cl), all electrolyte clearances were increased (P < .01 for all) from −30 to 0 minute collection period values during the 0 to 60 and 60 to 120 minutes collection periods after furosemide administration. Further, increases in FCl_Na and FCl_Cl during the 0 to 60 minutes collection period were greater (P < .01) for FU than the 3 NSAID treatments (Figure 5).

Although mean urine production during the 4 hours after furosemide administration was decreased by a similar degree with all NSAID treatments, there was considerable variation among horses. Specifically, the magnitude of diuresis during the 4 hours after furosemide administration ranged from 16.2 to 30.0 mL/kg for the FU treatment. In comparison to FU, the range of urine volume difference was 72% to 102%, 60% to 91%, and 68% to 111% for PB, FX, and DP, respectively (Figure S1). The differences in urine output between FU and the 3 NSAID treatments were less apparent in horses that had lower values for urine production (<20 mL/kg) with FU during the 4 hours after furosemide administration.

4 DISCUSSION

This study compared the effects of a COXIB (FX) and an atypical COX inhibitor (DP) to a nonselective COX inhibitor (PB) on renal function in horses by evaluating the diuretic and natriuretic response to furosemide after NSAID administration. Pretreatment with all NSAIDs resulted in ~25% lower total urine volume after furosemide administration as compared to furosemide alone. Furosemide-induced natriuresis and chloruresis were similarly attenuated after pretreatment with all NSAIDs. The results refute our hypothesis that pretreatment with FX or DP would have a lesser inhibitory effect on furosemide-induced diuresis and natriuresis than pretreatment with PB and are consistent with previous studies investigating the renal effects of phenylbutazone and meloxicam in horses.^8-10 We also found considerable, but inconsistent, interindividual variation in both the magnitude and attenuation of furosemide diuresis after treatment with the different NSAIDs.

After furosemide administration, PCV and TS increased in all groups, similar to previous reports.^{8, 19, 20} The increase in TS was consistent with a 10% to 15% decrease in plasma volume in the hour after furosemide administration.^{20, 21} A lesser increase in PCV, as compared to TS, has been attributed to sequestration of erythrocytes within the spleen.^{5, 19} Initially, serum Na⁺ concentration increased and serum K⁺ concentration decreased after furosemide administration in all treatment groups; however, these serum electrolyte concentrations returned to prefurosemide values by 180 minutes. We could not find a previous report documenting an increase in serum Na⁺ concentration after furosemide administration to horses. Our horses did not have access to water during movement to stocks and instrumentation (~1.5 hours) and were not offered water until 90 minutes after furosemide administration, despite having free access to hay. Thus, a mild increase in serum Na⁺ concentration could be an expected response to a proportionally greater loss of water, supported by maximal urine Na⁺ concentrations of ~100 mmol/L from 0 to 60 minutes after furosemide administration for all treatments. A decrease in serum K⁺ concentration has been reported after furosemide administration and has been attributed to redistribution of K⁺ to the intracellular space, rather than increased urinary K⁺ excretion.^{5, 19, 21} However, although urine K⁺ concentration decreased nearly 10-fold from 0 to 60 minutes after furosemide administration, we also observed a doubling of FCl_K during this same collection period suggesting that increased K⁺ excretion could also play a role in development of transient hypokalemia. Finally, serum Cl⁻ concentration decreased 45 minutes after furosemide administration and remained decreased through the remainder of the 4 hours urine collection period. Hypochloremia develops consequent to increased urinary Cl⁻ excretion as evidenced by the minimal decrease in urine Cl⁻ concentration and substantial increase in FCl_Cl after furosemide administration.^{5, 9, 19}

In comparison to prefurosemide values, urine Na⁺ concentration and FCl_Na increased from 0 to 120 minutes after furosemide administration. In contrast, urine K⁺ and Cl⁻ concentrations decreased and remained lower than baseline values through the remainder of the urine collection period, although FCl_K and FCl_Cl returned to prefurosemide values from 180 to 240 minutes. Because urine Na⁺ concentration is typically low in horses not receiving dietary Na⁺ supplementation (24 ± 7 mmol/L before furosemide administration in the mares in this study), the nearly 5- and 10- to 15-fold increases in urine Na⁺ concentration and FCl_Na, respectively, were expected responses to furosemide administration with a decrease in Na⁺ reabsorption because of blockade of NKCC-2 cotransporters in the TAL. In contrast, although urine K⁺ and Cl⁻ concentrations decreased after furosemide administration, baseline values were greater than baseline urine Na⁺ concentration. Consequently, with furosemide-induced diuresis absolute urine K⁺ and Cl⁻ concentrations decreased (substantially more for K⁺ than Cl⁻) while total excretion and FCl_K and FCl_Cl increased. All NSAID treatments attenuated the increases in FCl_Na and FCl_Cl during the initial hour after furosemide administration, consistent with a previous study that investigated the pharmacologic interaction of furosemide and phenylbutazone in horses.⁸ In all, urine electrolyte losses in the 4 hours after furosemide administration (Table 1) were comparable to previous reports^{9, 19} and approached ~7 and ~14% of extracellular fluid (ECF) Na⁺ and Cl⁻ contents, respectively, assuming ~14 000 mmol of Na⁺ and ~11 000 mmol of Cl⁻ in the ECF of a 500 kg horse. Despite an increase in FCl_K during the 2 hours after furosemide administration, total loss of K⁺ was <3% of intracellular fluid (ICF) K⁺ content (assuming ~28 000 mmol of K⁺ in the ICF of a 500 kg horse).

The primary pharmacologic action of furosemide is attachment to the luminal Cl⁻ binding site of the NKCC-2 cotransporter in the TAL, leading to decreased solute reabsorption. Because 20% and 25% of Na⁺ and Cl⁻ are reabsorbed via the NKCC-2 cotransporter in this nephron segment, furosemide administration results in a large increase in solute delivery to the distal nephron and increased output of urine.^{5, 6, 22} However, furosemide administration also leads to changes in systemic and renal hemodynamics, with a net effect of increasing RBF, and potentially GFR.^{5, 6} Further, PGE₂ production in TAL cells increases in response to furosemide administration through combined effects of increased release of arachidonic acid from phospholipids, increased COX-2 expression, and decreased degradation of PGE₂ by renal PGE₂ 9-ketoreducatse and 15-hydroxyprostaglandin dehydrogenase.^{7, 23, 24} Local action of PGE₂ in TAL and MD cells contributes to a further decrease in solute reabsorption. Finally, furosemide-induced PGE₂ production in cortical collecting duct epithelial cells can antagonize the antidiuretic effect of AVP.^{4, 25}

There are several mechanisms by which administration of NSAIDS could attenuate the diuretic and saliuretic responses to furosemide. First, pretreatment with phenylbutazone abolished furosemide-induced changes in systemic hemodynamics, including decreases in stroke volume and cardiac output and right atrial and right ventricular pressures, in resting horses.⁸ However, inhibition of these hemodynamic responses to furosemide administration was considered unlikely to contribute substantially to attenuation of furosemide-induced diuresis.^{5, 8} Next, because furosemide administration directly stimulates PGE₂ production in TAL and MD cells, further decreasing solute absorption, inhibition of COX-2 derived PGE₂ in this nephron segment could attenuate inhibition of solute reabsorption and diuresis.²³ Finally, production of either a concentrated or dilute final urine is fine-tuned by AVP action to regulate insertion of aquaporins (water channels) in cortical and medullary collecting duct epithelial cells. Attenuation of AVP-induced water reabsorption across isolated perfused rabbit collecting ducts by PGE₁ was first documented in 1968 (aquaporins were not discovered until 1992).⁴ It was proposed that PGs dampened the activity of AVP to prevent overshoots in AVP-induced collecting duct permeability.²⁶ Fifty years later the actions of PGE₁ and PGE₂ are now recognized to be mediated through binding to 1 of 4 subtype receptors (EP1-EP4) with incongruous effects on collecting duct water permeabilty.⁴ PGE₂ action on EP4 and EP2 receptors increase intracellular cAMP, promoting phosphorylation, translocation, and expression of aquaporin 2 (AQP2), thereby increasing collecting duct water permeability and urine concentration. In contrast, PGE₂ action on EP1 and EP3 receptors appears to downregulate AQP2 function by increasing intracellular Ca⁺⁺ and attenuating cAMP production, respectively, leading to diuresis.^{4, 27} Currently, this complex interaction of PGE₂ with its multiple receptors is incompletely understood but in studies examining the effects of NSAIDs on furosemide-induced diuresis, NSAIDs attenuate diuresis by up to 50%.^{3, 4, 6}

Our study was not designed to determine the mechanism(s) by which NSAIDs attenuate furosemide-induced diuresis; thus, the proportion of saliuresis attributable to decreased solute reabsorption versus potential interference with AVP action on collecting ducts remains uncertain. Our interest was in comparing the effects of a nonspecific COX inhibitor (phenylbutazone), a COXIB (firocoxib), and an atypical NSAID (dipyrone) on the diuretic responses of horses to furosemide administration. Again, we found no difference in attenuation of furosemide-induced diuresis and saliuresis after pretreatment with these 3 NSAIDs. A limitation of our investigation was that we studied healthy, euhydrated mares and it is possible that these NSAIDs could have different effects in dehydrated or diseased horses with altered RBF. Acknowledging this limitation, our data do suggest that use of COXIBs might not be “renoprotective” in diseased equids, as compared to nonselective COX inhibitors. Additionally, the variability in furosemide-induced diuresis observed between individual mares in this study, with and without pretreatment with NSAIDs, was considerable and emphasizes the need to monitor renal function when using all NSAIDs in diseased equids.

CONFLICT OF INTEREST DECLARATION

Authors declare no conflict of interest.

OFF-LABEL ANTIMICROBIAL DECLARATION

Trimethoprim-sulfamethoxazole used in the horses included in this manuscript is not labeled for prophylaxis or treatment of urinary tract infection or similar indications in the United States.

INSTITUTIONAL ANIMAL CARE AND USE COMMITTEE (IACUC) OR OTHER APPROVAL DECLARATION

Approved by the Michigan State University IACUC (PROTO202000077).

HUMAN ETHICS APPROVAL DECLARATION

Authors declare human ethics approval was not needed for this study.