1. Introduction
Idiopathic or trigeminal-mediated headshaking in horses is a debilitating and painful condition that can seriously compromise athletic performance, daily life activities, and quality of life, representing a major welfare problem that often results in euthanasia [
1]. Although the condition has been described in the literature for over one hundred years, its pathogenesis remains poorly understood [
2]. The clinical signs consist of sudden violent head shakes, nose itching and rubbing, snorting, striking their face with their front limbs, burning, tingling, or electric-like sensations that fit the description for neuropathic pain [
3,
4,
5]. The behavior and anxious facial expression of the horse appear to be that of stress. Indeed, studies on somatosensory nerve conduction of the trigeminal nerve in horses with idiopathic headshaking demonstrated lower threshold to trigger nerve conduction (10 times lower) of the maxillary branch of the trigeminal nerve in affected horses compared to control horses [
6,
7]. These findings suggested a functional abnormality and led to renaming the condition as trigeminal-mediated headshaking, a term that differentiates it from other causes of headshaking [
1,
6,
7]. Furthermore, the low threshold for firing of the nerve explains why sometimes apparently innocuous stimuli such as light, sound, wind, or touch might trigger apparent intense pain [
7]. Other times, the signs appear spontaneously [
7]. The disorder is seen primarily in geldings with an age onset of about 8 to 9 years, and it has a seasonal component, happening more often in the spring and summer months [
3]. A number of treatment options have been reported with inconsistent results [
1,
8,
9,
10,
11,
12].
The cause of trigeminal-mediated headshaking is unknown, however, a multifactorial etiology is suspected because 60% of affected horses present exacerbation of clinical signs in a seasonal fashion (mainly spring and summer months) and geldings are over-represented (72%) [
1]. These factors may include environmental, dietary, and hormonal contributions. Dietary component might be associated with seasonality, as the switch of forage from a winter crop to a spring crop coincides with onset of exacerbation of signs. Dietary components, including electrolytes and mineral content, in forage are known to change with season, soil, and harvest management [
13,
14,
15]. Moreover, it has been shown that changes in blood pH and electrolytes, especially calcium and magnesium, affect nerve conduction [
16,
17,
18]. The current study was implemented to evaluate the potential effects of experimental changes in blood pH and electrolytes on headshaking [
16]. This preliminary study sought to utilize horses diagnosed with trigeminal-mediated headshaking syndrome and assess the effects of intravenous infusion of 5% dextrose, hypertonic saline, and hypertonic sodium bicarbonate solutions on the frequency of headshaking. We hypothesized that alterations in blood pH and electrolytes will affect headshaking behavior.
2. Materials and Methods
2.1. Subjects and Facilities
This study included a total of six male castrated horses with naturally-occurring headshaking that were donated to the Center for Equine Health at University of California, Davis, CA, USA. The inclusion criteria consisted of fulfillment of the diagnosis of trigeminal-mediated headshaking by exclusion of all other causes of headshaking. At the time of donation, horses received all recommended vaccinations and deworming per Center for Equine Health protocols. Prior to entering the study, all horses underwent a thorough physical and neurologic examination performed by a board-certified, large-animal internist and neurologist (Monica Aleman) followed by a detailed diagnostic workup. Diagnostic workup included oral, ophthalmic and otoscopic examinations, complete cell blood count, serum biochemical profile, skull radiographs, and upper airway endoscopy. Breeds of enrolled horses included Quarter Horse breeds (n = 4) and Thoroughbreds (n = 2), ages 5 to 13 years, weighing 472–580 kg. The horses were housed in covered box stalls bedded with wood shavings, having free access to fresh water (automatic waterer), and fed twice daily a hay diet consisting of grass-alfalfa mix with a dietary cation–anion balance (DCAB = [Na+ + K+] − [Cl− + SO4−]) of 31 mEq/100g.
2.2. Experimental Design
The study was a randomized controlled crossover (3 by 3) experimental design, where each horse served as its own control. The horses were randomized to one of the three treatment groups with a washout period greater than a week between treatments. All horses received all three treatments. Intravenous catheters were placed in the left jugular vein under aseptic technique. Horses received sterile fluids intravenously: Dextrose Solution (DS) group received 5% dextrose solution at 2 mL/kg bwt, Hypertonic Saline (HS) group received NaCl 7.5% solution at 4 mL/kg bwt, and Hypertonic Sodium Bicarbonate (HB) group received NaHCO3 8.4% solution at 2 mmol/kg bwt. The infusion time was 30 min, and after infusion of treatment, the catheters were removed.
2.3. Sample Collection
Heparinized blood samples were collected by venipuncture from the right jugular vein at times 0 (baseline, before infusion) and 5, 15, 30, 60, 120 min post infusion and placed on ice immediately.
2.4. Blood Analysis
Venous blood pH, standard base excess (SBE), HCO
3−, Na
+, Cl
−, K
+, Ca
2+, glucose, and lactate in heparinized blood samples were determined using an ABL815 FLEX (Radiometer America Inc., Brea, CA, USA). Total magnesium (tMg) and ionized magnesium (Mg
2+) in heparinized blood samples were determined using NOVA 8 (NOVA Biomedical, Waltham, MA, USA). The strong ion difference (SID) was calculated using the Stewart equation as follows: [(Na
+ + K
+ + Ca
2+ + Mg
2+) − (Cl
−)] [
19]. Anion Gap (AG) was calculated using the equation as follows: [(Na
+ + K
+) − (Cl
− + HCO
3−)] [
19].
2.5. Behavioral Analysis
Horses were placed in individual round pens, without tack or halters, and evaluated by three independent, trained evaluators for headshaking behavior while at a walk (1 min), trot (1–3 min), canter (1 min), and walk again (1 min) at 5 time points (i.e., T0 (baseline, before infusion) and T15, 30, 60, 120 min post infusion). After each time point, the horses were returned to their immediately adjacent stalls. Evaluators were unaware of which treatment each horse received on any particular day. The headshaking behavior (including headshakes/minute, head tossing/minute, nose rubbing/minute, dropped head/minute, and snorting/minute) was recorded for each horse during each level of exercise at each time point by the three evaluators. A median of each headshaking behavior per minute for each horse at each time point during each level of exercise was obtained. Variability between evaluators was determined.
2.6. Statistical Analysis
Data were analyzed using Stata Statistical Software, Release 14, StataCorp LP 2015, College Station, TX. A multilevel mixed-effects Poisson regression model examined the main and interactive fixed effects of treatment groups (DS, HS, or HB), time period 0 (baseline), and 15, 30, 60, 120 min post infusion, period, and breed, with individual horses as the random effect, on headshaking. Each period (walk, trot, canter, walk) was evaluated separately. Within each period, a model was created to evaluate the possible interaction between treatment and time; if the interaction was significant, two further sets of analyses (effect of treatment at individual times, effect of time within individual treatments) were performed. If the interaction was not significant, then a main effects-only model was fit. Results are presented as incidence rate ratios (IRR), p values, and 95% confidence intervals (CI). A color-coded dot plot with the regression line was used to display longitudinal data for individual horses.
A multilevel mixed-effects analysis of variance model was also used to look at the main and interactive effects of treatment groups (DS, HS, or HB) and time period (baseline or 0, and 5, 15, 30, 60, 120 min post infusion), with individual horse as random effect, on blood parameters (pH, K+, Na+, Cl−, SBE, HCO3−, Ca2+, glucose, lactate, tMg, Mg2+) and calculated values of SID and AG. Each period (walk, trot, canter, walk) was evaluated separately. Within each period, a model was created to evaluate the possible interaction between treatment and time; if the interaction was significant, two further sets of analyses (effect of treatment at individual times, effect of time within individual treatments) were performed. If the interaction was not significant, then a main effects-only model was fit.
2.7. Ethical Approval
Ethical approval was provided by UC Davis IACUC on 24 March 2015. This institution is accredited by the Association for Assessment and Accreditation of Laboratory Animal Care, International (AAALAC). This institution has an Animal Welfare Assurance on file with the Office of Laboratory Animal Welfare (OLAW). The Assurance Number is A3433-01.
4. Discussion
This study evaluated changes in blood pH and electrolytes following administration of 5% dextrose solution, hypertonic saline, and hypertonic sodium bicarbonate solution with concomitant effect on headshaking behavior in the short term. The minimal effects observed were not expected to last. There was an effect of treatment with HB treatment having a 58% incident rate reduction in median headshakes/minute when compared to DS across all gaits (walk, trot, canter, walk). The reduction was compared to DS treatment which is a solution that is converted to water and would have minimal effect on pH. Despite the fact that there was great variability in the severity of the headshaking behavior between horses and variability of the same horses on different days, the IRR analysis takes into account these variables and provides a way to look at the effect of solution in headshaking behavior. The greatest reduction in headshakes/minute was 67% IRR for the trot with HB treatment. The effect of time at T15 was the first exercise period after receiving intravenous fluid treatment and the cause for that is speculated as the first time to exercise after being restrained for fluid therapy. The increase in rate of headshakes/minute at T60 after HB treatment is speculated as being the effect of intravenous fluids not lasting long, and an hour after administration the effects of the solution were going away. In this study, Thoroughbreds were less severely affected overall, however, they were more sensitive to changes associated with the solutions given. It would be beneficial to increase animal numbers to verify the effect of breed.
Overall, administration of 5% dextrose solution at 2 mL/Kg bwt caused minimal (not significant) increase in pH and a small drop in Na+ and Mg2+. The mild drop in Na+ is likely a result of the dilution effect of the infused solution. The mild drop in Mg2+ could be explained by a glucose-mediated increase in transcellular transport of magnesium into the cells leading to a slight decrease in plasma Mg2+; additionally, mild diuresis associated with the transient hyperglycemia and presumed glucosuria could have occurred.
Administration of hypertonic saline (8 times the physiologic concentrations of Na+ and Cl− in blood) at 4 mL/Kg bwt caused mild decrease in blood pH (remaining within normal ranges), mild decrease in SBE, AG, and SID from baseline, mild hypernatremia and marked hyperchloremia, mild drop in Ca2+ (although remaining within reference range), and mild hypomagnesemia (characterized by decrease in Mg2+ below normal range and mild decrease in tMg). These changes are likely associated with Na+ and Cl− overload, leading to potent volume expansion and natriuresis. This explains the persistent and slightly worsening concentrations of the Mg2+ up to the last sampling (120 min).
Finally, administration of hypertonic bicarbonate solution (6 times the physiologic concentrations of Na+ in blood) at 2 mL/Kg bwt caused moderate increase in blood pH, marked increase in HCO3− and SBE (the latter reaching 3 times the baseline concentration), mild decrease in SID, mild decrease in Cl− (although remaining within normal ranges), and mild decrease in AG. Similar to the effect seen after HS, HB led to mild decrease in Ca2+ and tMg (although remaining within normal ranges), and hypomagnesemia (characterized by decreased Mg2+) that persisted, worsening slightly over time, up to the last sampling (120 min).
The blood results from this study showed transient changes in blood variables, with significant short-lived changes in acid–base status (blood pH, SBE, HCO
3−, SID, and AG) when the horses were infused with HB, and mild changes in acid–base status (blood pH, SBE, SID, and AG) following HS. When horses were infused with DS, there were virtually no changes in acid–base status from baseline. It is expected that the changes in acid–base status are short-lived and quickly corrected [
20]. Indeed, by 30 min post infusion of HB, blood pH was back to the reference range. It is surprising that the acidifying effects of hypertonic saline were so minimal. Treatment with either hypertonic solution led to a decrease in Mg
2+, likely due to volume expansion and natriuresis. Although the changes in pH and electrolytes were modest, these subtle changes could lead to alterations in nerve firing, thus affecting headshaking.
Baseline values for SBE for all horses were above the reference range, and the calculated AG was below reference range at baseline. This finding was most likely due to diet (DCAB of 31 mEq/100 kg). This could have accounted for the mild acidifying effect of HS administration. The baseline measurements for most of the horses showed marginal levels of Mg
2+ (reference range from UC Davis VMTH Clinical Diagnostic Laboratory Services: 0.47–0.70 mmol/L). This is surprising considering that the horses should have been receiving their magnesium requirement of 13 mg/kg bwt/day, based on hay analysis of their feed [
21,
22]. Average absorption of magnesium content in feed by horses is 49.5% with a range of 30–60% [
21]. Magnesium absorption can also be impaired by fiber and phytate levels; however, phytate levels were not measured in the hay [
22]. It is expected that alkalemia would reduce Mg
2+ and Ca
2+ [
23,
24], but the decrease in Mg
2+ and Ca
2+ following HS and HB were comparable. This is likely a result of the overpowering effect of volume expansion and natriuresis of both solutions. One could speculate that decrease in Ca
2+ would result in less activation of the trigeminal nerve into firing. Greater concentrations of ionized calcium in the blood mean that they can activate release of neurotransmitters to signal the nerve to fire [
25]. It has been shown that calcium does affect neuropathic pain in the trigeminal nerve [
18]. The effects of calcium and magnesium on neuropathic pain should be further researched.
The main limitations of the current study included natural disease having a broad range of clinical signs and severity, the small group of affected horses, and lack of control groups (affected but untreated group and unaffected group). The observed changes in headshaking behavior following the infusion of the different solutions to this group of horses demonstrated a relative decrease in headshaking following HB treatment. A larger group of affected horses might be needed to fully investigate these effects, changes in pH and electrolytes, in such an impairing disorder. Further studies are warranted to establish if dietary changes, especially with respect to calcium and magnesium, affect headshaking behavior and trigeminal nerve firing.