Outcome and complications in horses administered sterile or non‐sterile fluids intravenously

Abstract Background Obtaining commercial fluids for intravenous administration (IVF) was challenging during a recent shortage. This necessitated use of custom‐made non‐sterile fluids for intravenous administration (JUGs) in some hospitals. There are no studies comparing outcome of horses treated with JUG versus IVF and limited information is available about adverse effects of JUGs. Hypothesis/Objectives To evaluate death, complications, blood pH, and plasma electrolyte concentrations of horses that received JUG versus IVFs. Animals One hundred eighty‐six horses that received IVFs and 37 that received JUGs. Methods A retrospective review of medical records was performed to identify horses that received IVFs or JUGs during hospitalization. Information including survival to discharge, complications (fever [>38.5°C], jugular vein phlebitis/thrombosis, arrhythmia, or laminitis), blood pH, and plasma electrolyte concentrations were obtained. Results There was no difference (P = .67) in survival to discharge for horses that received JUGs (78%) compared to horses that received IVFs (87%). Horses that received JUGs were more likely to develop a jugular vein complication (3 of 37 versus 1 of 186, odds ratio 17.2 [95% CI 1.9‐389.8], P = .04). Horses that received JUGs were more likely to have electrolyte abnormalities consistent with hyperchloremic metabolic acidosis. Conclusion and Clinical Importance Veterinarians using JUGs should obtain informed client consent because of a potential increased risk of jugular vein complications. Chloride content of JUGs should be considered to limit development of hyperchloremic metabolic acidosis.

disease associated with large volume fluid loss (ie, diarrhea and gastric reflux) or instances when water consumption is precluded and enteral fluid therapy is not tolerated.
There are several limitations of IVF that are available in volumes practical for equine fluid therapy (ie, 3 and 5 L bags). First, most commercially available products are formulated as replacement solutions with electrolyte concentrations similar in composition to ECF. These products are appropriate for short-term replacement of fluid deficits but might not be ideal for long-term (>48 hours) supportive treatment of horses with ongoing fluid losses in gastric reflux or diarrhea, as they might provide an excessive load of sodium and chloride. Long-term, high volume supportive treatment with replacement solutions will result in excess administration of Na + and Cl − , promoting development of edema, hyperchloremic metabolic acidosis, and depletion of body K + content through increased urine output. 1,2 Further, use of chloride-rich IVF (ie, 0.9% NaCl) versus more balanced polyionic solutions (ie, Plasmalyte-148) has been associated with increased morbidity and death in multiple studies in people. [4][5][6][7] Use of 0.9% NaCl in human patients results in the development of a hyperchloremic metabolic acidosis. 4,6,8 The exact cause of increased death in human patients receiving 0.9% NaCl is not known. However, the use of 0.9% NaCl subsequently resulting in a hyperchloremic metabolic acidosis has been associated with decreased renal perfusion 9 and impaired immune function 10 including an increased risk of postoperative infections. 11 Furthermore, in humans the use of replacement fluids in resuscitation of critically ill patients has specifically been associated with degradation of the endothelial glycocalyx further perpetuating poor vascular homeostasis. 12 Moreover, the previous shortage of 3 and 5 L IVF bags for horses 13 made obtaining IVF challenging at best, and impossible for some equine hospitals during that time period. This problem led to increased use of non-sterile custom-made fluids for IV administration formulated by adding bulk electrolyte products to reverse osmosis or distilled water carboys. A high rate of bacterial and low rate of endotoxin contamination in hospital prepared, non-sterile fluids for IV administration was previously reported. 9 However, potential adverse clinical signs associated with the administration of these fluids for IV administration were not reported in horses.
This retrospective study was performed to test the null hypothesis that horses treated with non-sterile custom-made fluids for IV administration (JUG) would not have increased rate of death or complications compared to horses treated with commercial, sterile IVF. Venous blood pH and plasma electrolyte concentrations of hospitalized horses were also evaluated to determine if horses receiving JUG would be more likely to develop acid-base and electrolyte derangements, as compared to horses receiving IVF.

| Data collection
Medical records of horses greater than 6 months of age that were presented to Michigan State University's Veterinary Medical Center and received fluids administered IV between December 2014 and August 2017 were reviewed. Inclusion criteria included administration of IVF or JUG for at least 24 hours. Horses that were administered both IVF and JUG during the same hospitalization period were excluded. Data extracted from medical records included signalment, rectal temperature, HR, RR, PCV, TS, pH, lactate, and electrolyte (Na + , K + , Cl − , and HCO 3− ) concentrations at admission and~24 hours after starting fluid therapy.
The type(s) of fluids for IV administration administered and total volume (mL/kg) of fluids for IV administration administered were recorded. PCV and TS were determined by the microhematocrit method and refractometry, respectively, and the mean values of two samples recorded in medical records were used for analysis. Blood gas parameters, lactate and electrolyte concentrations were determined with a blood gas analyzer (Novaphox Ultra, Novabiomedical, Waltham, Massachusetts).
Outcome (survival to discharge) and development of a fever (rectal temperature >38.5 C), jugular vein phlebitis/thrombosis, arrhythmias, and laminitis were the endpoints of interest.
All IVCs used for horses in this study were 14 gauge × 9 cm poly- To assess for complications, all pages of the medical records were carefully reviewed for documentation of a fever, jugular vein phlebitis/ thrombosis, arrhythmias, and evidence of laminitis (weight shifting, increased digital pulses, sensitivity to hoof testers at the apex of the frog). Jugular vein phlebitis was defined as palpation of vessel wall thickening (with heat, pain, or both) on routine examination and jugular vein thrombosis was defined as complete occlusion of the vessel confirmed by absence of jugular vein distention when the vein was occluded at the base of the neck, below the catheter insertion site.  Table 1.

| Data analysis
Horses were grouped by type of IV fluid administered (IVF or JUG) and missing data was accounted for by imputation of the mean. A propensity score analysis was conducted to estimate the effect of the type of IV fluid administered on survival to discharge and development of fever, jugular vein phlebitis/thrombosis, arrhythmia, and laminitis. The propensity score was used in a regression adjustment in order to account for risk factors that were potential confounders of the outcome variables, while avoiding the degree of overfitting that could result from directly adjusting for all confounders in the model. 14 The score was constructed through a logistic regression analysis that

| Horses and treatments
Two hundred and ninety-nine horses were treated with fluids for IV administration during the 2.5 year study period. Seventy-six horses were excluded because they received both IVF and JUG during the same hospitalization, the owners elected euthanasia in surgery or they had incomplete medical records. One hundred eighty-six horses that received IVF and 37 horses that received JUG were included in the final analyses. Mean age (IVF 12.8 and JUG 13 years) and sex distribution were similar between the two groups and a variety of breeds were represented.

| Admission data
Admission physical exam and clinicopathologic data are detailed in Table 2. Admission parameters were not significantly different between the two groups.

| Changes in clinicopathologic values
T A B L E 1 Composition of commercial fluids and custom-made, non-sterile fluids for IV administration (JUG) used during the study period. All solutes are expressed as mmol/L and osmolarity is expressed as mOsm/L  19 Similarly, jugular vein thrombophlebitis was a short-term complication in 21/252 (7.5%) horses after colic surgery, with the incidence higher in horses manifesting postoperative pain or shock (15 and 20%, respectively). 27 Realistically, reliance on information retrieved from medical records likely resulted in underestimation of jugular vein complications in our study. As an example, when more critical examination was pursued in a study comparing two catheter types, 61% of veins had local perivascular swelling, hematomas or both and ultrasonographic assessment revealed moderate to severe venous pathology in 5.4% of catheterized veins. 16 Although JUG fluids likely had a greater risk of contamination with bacteria and endotoxin, neither of which are found in commercial fluids, 28 introduction of these agents would more likely produce systemic effects (eg, fever, tachycardia, tachypnea, and sweating) 29 rather than localized inflammation or sepsis. Consequently, the mechanism(s) by which custom-made, non-sterile fluids might have contributed to an increased risk of jugular vein complications was not fully elucidated in either our or previous studies. 19 To the best of our knowledge, none of the horses in this study suffered further complications of jugular vein phlebitis/thrombosis after hospital discharge. However, sequela can include ipsilateral swelling of the head and neck, bacteremia, endotoxemia, pulmonary infarction, and vegetative endocarditis. 25,30 Further, thrombosis resulting in permanent occlusion of a jugular vein would make future medical treatment requiring placement of an IVC into the contralateral jugular vein of greater risk for more severe complications, including head swelling and airway occlusion. Unilateral jugular vein thrombosis has also been hypothesized to limit athletic potential; however, a retrospective study found no significant difference in performance of either pleasure or racehorses before and after development of a jugular vein thrombosis, in horses that returned to performance. 31 A guiding principle of fluid therapy is restoration and maintenance of euvolemia and tissue perfusion, while avoiding exacerbation of metabolic disturbances. In humans, use of IVF containing Cl − in concentrations higher than patient serum (or "chloride rich fluids") promotes development of hyperchloremic metabolic acidosis and increased morbidity and risk of death with this metabolic disturbance has gained attention in recent years. 4,11,32,33 In people, development of hyperchloremia has been associated with decreased renal blood flow and renal cortical perfusion, 9 impaired immune function, 10 increased postoperative infection rates 11 and increased risk of death. 11,34,35  with NSAIDs was not assessed as a risk factor in the statistical analysis because of a number of factors including: variable NSAID administration before presentation, differences in clinician preference regarding administration of NSAIDs (timing and dose) and the fact that essentially all horses received 1 or more doses of an NSAID during the initial 24 hours of hospitalization. It should also be noted that although a statistical significance was not reached for fluid type and proportion of horses that developed laminitis or a fever within 24 hours the P-values were <.10, which might be cause for consideration. ORs and 95% CIs were provided so that the reader can determine their level of comfort with these results. Post hoc power calculations indicated that to detect a difference in the proportion of horses affected by laminitis a total of 344 horses per group would have been required and a total of 176 horses per group would have been required to detect a difference in the proportion of horses that had a fever recorded in the first 24 hours after starting fluids.
Finally, given the significant difference in cost between the two types of fluids administered, it is possible that financial constraints of the owner were an additional confounding factor.
In conclusion, we found no evidence that horses receiving JUG fluids were less likely to survive to discharge compared to horses that received IVF. However, there was moderate evidence that horses that received JUG fluids had an increased risk of developing a jugular vein complication. When considering use of JUG fluids, this increased risk should be discussed with owners so that they can make an informed decision. Additionally, use of Cl − -rich JUG formulas resulted in hyperchloremic metabolic acidosis compared to those horses that received IVF. Consequently, hospitals using custom-made, non-sterile fluids for IV administration should consider the amount of Cl − when formulating fluids for IV administration.

ACKNOWLEDGMENT
Results from this study were presented, in part, at the 2018 IVECCS XXIII Symposium in New Orleans LA.