Present address: Swedish-Norwegian Foundation for Equine Research, SE-161 89 Stockholm, Sweden. Email: firstname.lastname@example.org
Reasons for performing study: Since nonsteroidal anti-inflammatory drugs, such as flunixin, on account of their anti-inflammatory and analgesic properties, are used in both racing and equestrian sport horses, the question has been raised as to whether these drugs affect the physiological responses to exercise and thus performance potential.
Objectives: The aims of this investigation were to study the effects of flunixin on cardiorespiratory, metabolic and locomotor parameters in horses during intense treadmill exercise.
Methods: Six Standardbred trotters underwent an incremental treadmill exercise test to fatigue, without drug and then after administration of flunixin meglumine (1.1 mg/kg bwt i.m.). Heart rate (HR), oxygen uptake and stride length were measured and venous blood samples drawn repeatedly during the test.
Results: Heart rates were found to be significantly higher at submaximal speeds, while the velocity causing a HR of 200 beats/min was significantly decreased after treatment with flunixin. Maximal HR and plasma lactate concentration 5 min after exercise were unchanged after medication. Flunixin caused higher plasma lactate concentrations at all speeds and the lactate threshold was decreased, compared with baseline values. Oxygen uptake levelled off at the highest velocities and did not change after flunixin treatment. Stride length was increased after treatment, although not at the highest velocities.
Conclusion: The increased HR and lactate responses to exercise after flunixin treatment indicate that it does influence physiological responses, but does not improve the performance potential of clinically healthy horses. However, the lengthened stride during submaximal exercise after medication could imply undetected subclinical lameness, masked in some of the horses, i.e. they have performed with a longer stride at the cost of a higher heart rate and an increased lactate concentration.
The nonsteroidal anti-inflammatory drug (NSAID) flunixin meglumine is commonly used in equine practice for musculoskeletal inflammatory conditions and treatment of colic and abdominal pain (Tobin 1981; Lees and Higgins 1985; Kallings 1993). Flunixin is a nonspecific blocker of cyclo-oxygenase enzymes (COX), inhibiting the production of prostaglandins (PG) and thromboxanes (TX) (Lees and Higgins 1984; Soma et al. 1992). The major effects of NSAIDs are anti-inflammatory, analgesic and antipyretic, all related to their prostanoid-inhibiting effect and depending on the different influence on the constitutive COX-1/inflammation-induced COX-2 enzymes (Rang et al. 1995).
Even if blocking of cyclo-oxygenase is the main mechanism explaining the anti-inflammatory and analgesic effects, some of these drugs also have COX-independent anti-inflammatory and analgesic effects (Tegeder et al. 2001).
Since NSAIDs, on account of their anti-inflammatory and analgesic properties, are so commonly used in both racehorses and pleasure horses and in some jurisdictions even permitted in competitions, the question has been raised as to whether these drugs affect the physiological responses to exercise and thus the performance potential.
The aim of the current investigation was to study the effects of flunixin on cardiorespiratory, metabolic and locomotor parameters in Standardbred trotters during incremental treadmill exercise to fatigue.
Materials and methods
Six Standardbred trotters, one mare and 5 geldings (age 5–12 years, weighing 425–525 kg), were used in the study. On pre-examination they were clinically healthy and well accustomed to the treadmill before undergoing the tests. These experimental horses had been regularly trained and used in different research projects with sufficient washout periods in between. In order to keep the horses fit but avoid training effects, they performed submaximal/maximal exercise workouts once a week either on a treadmill or track. They were housed in conventional stables and fed a normal diet consisting of grain and hay and had water ad libitum.
The study was performed after approval of the local ethical committee for animal experiments.
The horses underwent 2 standardised exercise tests on a high-speed treadmill (Säto)1. At least one week (1–3 weeks) after the baseline test (drug-free) they performed the flunixin test (blinded), starting 4 h after administration of the drug (Finadyne; 1.1 mg/kg bwt, given i.m. in the neck)2, i.e. each horse served as its own control.
The exercise test comprised of an incremental test to fatigue on an inclined (6.25%) treadmill. The horses were fitted with a face mask and trotted, after a warm-up, at 1 min stepwise speed increments of 6, 7, 8, 9, 10 and 11 m/s or until they could no longer keep pace with the treadmill at a trot. Time to fatigue was defined as the point at which the horses were no longer able to keep pace at the last speed on the treadmill, despite intense but humane encouragement.
Measurements of the exercise parameters were made before and during each speed step and after exercise. If a horse was unable to complete the 60 s of the last speed step, samples were taken and measurements made just before the treadmill was stopped and the time to fatigue noted.
Heart rate (HR), monitored electrocardiographically with a Mingograph 8043, was recorded before the test and during the last 15 s of each exercise step. The HR response to exercise, expressed as the velocity (V) causing a HR of 200 beats/min, V200 (m/s), was calculated from the linear HR/V relationship (Persson 1983, 1997).
The oxygen uptake (VO2; l/min) was determined in an open-flow system using a gas analyser (Servomex integrated into Oximeter 3200)4,5. Recordings were made at 15 s intervals during the test. VO2max was defined as the levelling off of VO2 at increasing speed, a respiratory exchange ratio (R = VCO2/VO2) exceeding 1 and a high lactate concentration.
Venous blood samples were collected from an indwelling catheter in a jugular vein before each speed, during the last 15 s at each speed and 5 min after exercise. Plasma lactate concentrations (mmol/l) were determined with a lactate analyser (GM7)6. The lactate threshold, VLA4 (m/s), was estimated from the exponential function of blood lactate and treadmill velocity (Persson 1983, 1997). Blood lactate (BLA) calculations were made by using the regression between the plasma and blood lactate concentrations (Persson 1997).
Stride frequency and stride length
Stride frequency was calculated by timing 50 paces at each speed, and stride length (SL; m) was calculated from stride frequency and speed.
The results are expressed as mean ± s.d. The Wilcoxon signed rank test was used to compare differences at each observation, with and without treatment. The level of significance was set at P<0.05.
The heart rate showed a linear increase with exercise up to 9 m/s and then levelled off. After flunixin treatment, HR increased significantly during submaximal exercise to above 8 m/s compared with the baseline value and consequently V200 decreased significantly, from 8.52 ± 0.52 m/s at baseline to 7.95 (± 0.57) m/s after treatment (Fig 1a). VO2 increased linearly with exercise up to 10 m/s, when it levelled off. Flunixin did not affect this uptake, although a significant increased value was noted at one particular speed (8 m/s) (Fig 1b). The respiratory exchange ratio was not affected by flunixin.
Plasma lactate concentration increased exponentially and was, after administration of flunixin, significantly increased at all steps during exercise, compared with baseline (Fig 1c). The lactate threshold, VLA4, was significantly decreased, from 8.67 ± 0.57 to 8.26 ± 0.42 m/s after drug treatment. The mean plasma lactate value 5 min post exercise was not significantly changed between the test without medication (20.1 ± 5.1 mmol/l) and the test with medication of flunixin (25.5 ± 5.4 mmol/l), although 5 out of 6 horses had higher values for lactate after medication.
Stride and fatigue
Stride length increased significantly with speed after flunixin treatment, except at the higher intensities (Fig 1d). The mean time to fatigue at the highest speed the horse could sustain did not significantly differ from baseline after flunixin treatment, although it diminished in 3 of the horses.
The plasma concentrations of flunixin at the times of the tests (4–4.5 h after administration) ranged from 1.2–2.4 µg/ml (mean 1.9 µg/ml).
It was found in the present study that flunixin treatment affected physiological responses during treadmill exercise to fatigue. The higher heart rates during submaximal exercise and the decrease seen in V200, were consistent with previous preliminary findings (Kallings and Persson 1983).
Effects of another NSAID, phenylbutazone, have been investigated in a similiar way during submaximal exercise in treadmill trials. After repeated oral administration of phenylbutazone (2.5 mg/kg bwt, b.i.d.), V200 values were significantly decreased, by 2%, compared with those in a drug-free test. Oxygen uptake, respiration and other exercise tolerance related variables did not change significantly during the exercise test, starting 4 h after the last drug administration (Kallings et al. 1987). Therefore, flunixin and phenylbutazone increased the heart rate in horses performing submaximal treadmill exercise. However, during sustained submaximal exertion (1 h on an inclined treadmill) no effects of phenylbutazone (4.4 mg/kg bwt i.v.) on HR, cardiac output, right atrial pressure or other haemodynamic variables were found (Hinchcliff et al. 1994).
Although effects of prostanoid inhibition on the cardiovascular responses to submaximal exercise have been documented in man (Cowley et al. 1984, 1985; Stassen et al. 1984; Wilson and Kapoor 1993) and in horses performing treadmill exercise (Kallings and Persson 1983; Olsen et al. 1992; Hinchcliff et al. 1994), there has been only limited information of the effects of NSAIDs on the systemic haemodynamic responses to intense exercise in horses. In agreement with the results in the present study, an incremental exercise stress test 1 h after treatment with flunixin (1.1 mg/kg bwt i.v.), showed no significant effect on mean peak oxygen consumption or carbon dioxide production. Unfortunately, HR and lactate responses were not reported in that study, but it was shown that flunixin prevented the exercise induced elevations in serum TXB2 (significantly reduced for 8 h and returned to normal by 12 h) and PGF1α (Colahan et al. 2002). A similar effect on TXB2 was observed (significantly decreased for 12 h with a return by 24–48 h) in a submaximal study on phenylbutazone (5 mg/kg bwt first day, then 2.5 mg/kg bwt for 2.5 days per os) (Kallings et al. 1987). Increased heart rate and right atrial pressure have been observed in Standardbred horses during exertion in treadmill tests following repeated administration of phenylbutazone (8.8 mg/kg bwt per os for 2 days, then 4.4 mg/kg i.v. 1 h pre-exercise). It was also found that phenylbutazone nullified the exertion-induced increases in plasma 6-keto-PGF1α and TXB2 and it was concluded that prostanoids probably mediate or modulate some of the systemic haemodynamic responses to exertion in horses (Mitten et al. 1996). In our submaximal test, increased heart rate and no effect on oxygen uptake were found (Kallings et al. 1987). The accentuation of the exercise-induced increase in heart rate by flunixin in the present study, and phenylbutazone in previous studies (Kallings et al. 1987; Mitten et al. 1996), may have been due to inhibition of prostanoid production and an increased sensitivity to sympathetic simulation of the SA node in the heart, as suggested by Mitten et al. (1996). As mentioned earlier, heart rates during sustained submaximal exercise showed no effect of phenylbutazone in the study by Hinchcliff et al. (1994). This may then be related to the fact that no changes were seen in 6-keto-PGF1α or PGE in that study and the change seen in TXB2 was not marked. Phenylbutazone (4.4 mg/kg bwt i.v., for 2 days with final dose 4 or 24 h before test) had no effect on right atrial and pulmonary vascular pressures or on the heart rate in Thoroughbreds undergoing strenuous treadmill exercise in a study by Manohar et al. (1996). They suggested that these contradictory findings vis-à-vis previous studies might be related to breed differences and/or different standards of fitness of the horses in the 2 studies. However, submaximal exercise responses were not reported in that study. The unchanged heart rate at maximal exertion due to NSAID-treatment is in agreement with findings in the present study, as well as in our study on the track where flunixin did not show any significant effect on the heart rate responses during a simulated race (Kallings et al. 1999).
In conclusion, the results indicate in the current as well as in previous studies, that flunixin increases HR during submaximal but not at maximal exercise.
Effects on lactate response
The lactate concentrations during exercise were increased after medication and consequently, VLA4 was decreased. At the end of exercise all horses had high lactate levels in plasma and their respiratory exchange ratio was >1, implying maximal exertion. Five out of 6 horses showed higher plasma lactate accumulation 5 min post exercise after flunixin treatment. The reason why one flunixin-treated horse had lower lactate values after work may be that this horse performed less intense exercise as it became fatigued after only 10 s in the last speed step. In our study on the track, flunixin had no statistically significant effects on plasma lactate concentrations post exercise (Kallings et al. 1999).
Previous studies have shown that VLA4 was not significantly changed after flunixin treatment and the lactate accumulation after exercise was lower than or similar to the corresponding baseline value (Kallings and Persson 1983). Similarly, blood lactate accumulation did not change significantly during submaximal exercise (with 2 min steps) following administration of phenylbutazone (Kallings et al. 1987; Mitten et al. 1996).
These results could be attributed to the fact that in the previous submaximal exercise test model the horses ran longer during each step (2 min each step) than in the present study, allowing the blood lactate to be dispersed to other tissues and organs.
In the present intense study, the speed increase was faster, with 1 min at each step up to 11 m/s. This indicates an earlier recruitment of fast-twitch glycolytic muscle fibres. It has been shown that fibre recruitment pattern occur from type I (slow twitch) to type IIA and type IIB (fast twitch) fibres with incremental speed (Lindholm et al. 1974; Valberg 1986). There will therefore be a quicker increase in the accumulation of lactate, which was relatively higher after flunixin treatment. Lactate concentration in blood is dependent not only on the release of lactate from muscle, but also on its rate of removal from blood. The increased level of plasma lactate after flunixin administration in the current study could therefore be due to factors such as a reduced uptake in surrounding tissues (e.g. muscle fibres) or organs (heart, liver, kidneys, etc.). It may also have been a result of an increased production (glycolysis), an increased efflux from muscle to blood or reduced transport into the erythrocytes.
In the present study lactate was measured in plasma, and in the earlier studies in whole blood, which might suggest that the uptake of lactate by red blood cells is reduced by flunixin treatment. It is reported that NSAIDs may exert nonprostanoid effects such as inhibition of transmembrane anion transport in erythrocytes (Abramson and Weissmann 1989).
Inhibition of prostanoids by the cyclo-oxygenase pathway as the only mechanism of action has been discussed. Nonprostanoid effects have also been described (Boothe 1995). At high concentrations, NSAIDs seems to uncouple protein-protein interactions within the plasma membrane and therefore interfere with a variety of cell membrane processes such as oxidative phosphorylation and cellular adhesion (Abramson and Weissmann 1989; Weissmann 1991; Rang et al. 1995).
Beside their inflammatory effects, prostanoids have numerous other effects in the body, including action on the haemostatic system. Prostacyclin (PGI2) and PGE2 act, in principle, as vasodilators and PGF2α and TXA2 as vasoconstrictors (Adams 1995). It has been discussed whether prostanoids are involved in the regulation of regional blood flow during exercise - increased blood flow to the active skeletal muscle and reduced blood flow to organs not requiring it would be beneficial - but varying results have been reported in different studies on prostanoids in different systems (Birks et al. 1991). In man, release of vasodilatory PGs contributes to exercise induced arteriolar vasodilatation and hyperaemia in skeletal muscle (Wilson and Kapoor 1993). Increases in plasma PGs have been reported in marathon runners (Demers et al. 1981; Ronni-Sivula et al. 1993) and during treadmill and cycle exercise (Mehta et al. 1983).
The NSAIDs acetylsalicylic acid and indomethacin attenuate the exercise-induced increase in calf blood flow and attenuate the increase in blood flow to a nonexercising limb in human subjects performing cycle exercise (Cowley et al. 1984, 1985). Indomethacin has been shown to reduce forearm blood flow both at rest and during exercise (Wilson and Kapoor 1993).
The increased plasma lactate levels recorded in our study could therefore be a result of interference of flunixin with the blood flow, and/or the oxidative phosphorylation, in skeletal muscle. They could also reflect a change in muscle fibre recruitment, with an increased contribution from the glycolytic type II fibres.
Effects on fatigue, stride and movement
The mean time to fatigue after medication did not differ, which was in agreement with the study of Colahan et al. (2002).
Stride length (SL) in the present study increased at submaximal velocities after treatment with flunixin compared with baseline values. Since the treadmill velocities were the same on both occasions, the mean stride frequencies were subsequently reduced, although not significantly so (except at 7 m/s).
When the horses in the present study were studied on the track, no significant changes in maximum speed, stride length or stride duration could be proven after administration of flunixin, although some effects on the locomotion pattern were observed. In forelimb, stance time decreased, swing time increased and the range of limb angle during the stance phase decreased after flunixin was administered (Kallings et al. 1999). The change in stance time with respect to stride time was decreased, consistent with a study of induced lameness (Buchner et al. 1995). Studies of effects of phenylbutazone on lame horses have given similar results (Drevemo et al. 1995). A possible explanation, that these changes might be due to a mitigation of pain in the musculoskeletal system and with a suspicion of subclinical lameness, has been discussed (Kallings et al. 1999).
In the current treadmill trial, the increased SL after flunixin administration might have been caused, if not by changes in muscle fibre recruitment, then by subclinical lameness and a consequently modified locomotory pattern in these horses.
In a study on the effects of drugs on the performance of horses it was found that phenylbutazone improved their performance in time trials. Phenylbutazone was administered (6.6 mg/kg bwt i.m.) to 4 horses 23 h before the trial. The rather surprising improvement was explained by assuming that the horses, previously regarded as sound, were actually lame (Sanford 1983). This conclusion, that phenylbutazone relieved subclinical lameness rather than actually simulating and improving performance, was supported by a study on the effects on locomotor response in laboratory experiments. Phenylbutazone neither stimulated nor depressed the horses at clinically used doses (Tobin 1981; Tobin et al. 1986).
Stride length is considered to be sensitive to the action of NSAIDs and to quantify the effect on lameness (Tobin 1981). In a pharmacokinetic/pharmacodynamic model with induced carpal arthritis it was reported that NSAIDs have a potential effect of increasing SL by 10%. Flunixin increased the stride length of horses with induced joint inflammation by some 6–16% compared with control (saline), with a peak effect at 4–6 h and an EC50 of 0.9 µg/ml plasma (Toutain et al. 1994).
Flunixin concentration and efficacy
To ensure effective levels of flunixin during the current trial, the plasma concentrations of the drug were determined and found to range from 1.2–2.8 µg/ml (4–4.5 h after administration). Drug concentration after a recommended dose is reported to peak at 1.6 µg/ml and the onset of action to occur within 2 h, peaking at 2–16 h. (Boothe 1995). The median maximal concentration (Cmax) after i.m. administration has been shown to be 2.3 µg/ml plasma (range 1.8–3.3 µg/ml), the median time to reach Cmax (Tmax) to be 76.6 min (31.7–97 min) (Toutain et al. 1994; Dyke et al. 1997). The mean plasma flunixin concentration of 1.9 µg/ml in our study was therefore within the suggested range of efficacy.
Since the present study did not demonstrate any increase in oxygen uptake, while heart rate and plasma lactate levels were elevated, flunixin does not appear to have had a positive effect on the physiological response to exercise in these horses. On the other hand, the longer stride after flunixin medication rather implies a positive effect on the locomotory pattern. However, the possibility that this might have been related to the analgesic/antiphlogistic effect of the drug on subclinical lameness or pain cannot be ruled out. The results of this study indicate that this could lead to longer stride length, but at the cost of higher heart rate and lactate responses during submaximal exercise.
It may therefore be considered that NSAIDs do not seem to enhance the performance potential but they may affect the actual capacity by masking pain and (subclinical) lameness. In racing and competitive sports, this kind of therapeutic use could jeopardise the welfare of the equine athlete.
The authors are grateful to Professor Lars-Erik Appelgren for kind interest and valuable criticism of the manuscript. We also wish to thank Professor Ulf Bondesson and the Equine Drug Research Laboratory for carrying out the plasma assays of flunixin.
The study was supported financially by the Swedish Trotting Association (STC).