Corresponding author: K.T.T. Corley, Anglesey Lodge Equine Hospital, The Curragh, Co Kildare, Ireland; e-mail: firstname.lastname@example.org.
Reasons for Performing Study: Critical illness is associated with hyperglycemia in humans, and a greater degree and duration of hyperglycemia is associated with nonsurvival. Hypoglycemia is also seen in critically ill humans, and is associated with nonsurvival. This might also be true in the critically ill foal.
Objectives: To investigate the association of blood glucose concentrations with survival, sepsis, and the systemic inflammatory response syndrome (SIRS).
Methods: Blood glucose concentrations at admission (515 foals) and 24 hours (159 foals), 36 hours (95), 48 hours (82), and 60 hours (45) after admission were analyzed. Logistic regression analyses were performed to investigate the association of glucose concentrations with survival, sepsis, a positive blood culture, or SIRS.
Results: 29.1% of foals had blood glucose concentrations within the reference range (76–131 mg/dL) at admission, 36.5% were hyperglycemic, and 34.4% were hypoglycaemic. Foals that did not survive to hospital discharge had lower mean blood glucose concentrations at admission, as well as higher maximum and lower minimum blood glucose concentrations in the 1st 24 hours of hospitalization, and higher blood glucose at 24 and 36 hours. Foals with blood glucose concentrations <2.8 mmol/L (50 mg/dL) or >10 mmol/L (180 mg/dL) at admission were less likely to survive. Hypoglycemia at admission was associated with sepsis, a positive blood culture, and SIRS.
Conclusions and Potential Relevance: Derangements of blood glucose concentration are common in critically ill foals. Controlling blood glucose concentrations may therefore be beneficial in the critically ill neonatal foal, and this warrants further investigation.
Critical illness is associated with hyperglycemia in humans.1 This occurs as a result of altered glucose metabolism and insulin resistance, with an increase in hepatic gluconeogenesis despite increased blood glucose and insulin concentrations.1 The degree of hyperglycemia at admission and the duration of hyperglycemia during critical illness are both associated with an adverse outcome in adult and neonatal humans.2–5 Hypoglycemia has also been reported in critically ill humans, and is associated with nonsurvival.6,7
Derangements of blood glucose concentration are also seen in critically ill horses. Hyperglycemia is common in horses with acute abdominal disease, and is associated with nonsurvival, whereas hypoglycemia is uncommon.8 Hypoglycemia in critically ill foals at admission correlates to nonsurvival to hospital discharge in septic foals,9 and with survival to 96 hours but not to hospital discharge in all foals presenting to a neonatal intensive care unit.10 Clinical experience suggests that hyperglycemia is also common in critically ill foals, but the prevalence and association with outcome is not well documented. In light of the interest in blood glucose concentrations in human patients, a retrospective study of blood glucose concentrations in critically ill foals was undertaken.
The purpose of this study was to investigate blood glucose concentrations in critically ill neonatal horses at admission and for the 1st 60 hours of hospitalization, and to investigate whether there is an association of blood glucose concentrations with survival, sepsis and the systemic inflammatory response syndrome (SIRS) in these patients.
Materials and Methods
Foals presenting to Marion DuPont Scott Equine Medical Center (1998–2005); University of California at Davis (1995–2004); Scone Veterinary Hospital (2004); and the Royal Veterinary College (2002–2006) were eligible for this study if they met the inclusion criteria. These criteria were neonatal foals presenting to a neonatal intensive care unit and having blood glucose concentrations analyzed at any point in the 1st 60 hours of hospitalization. Exclusion criteria were foals >7 days old, and where the medical record was unavailable. Data from some of the foals in this study have been published in previous studies.11–13
Analysis of venous blood for blood glucose concentrations is routinely performed in these cases. Different methods of measuring blood glucose were used at the different institutions: blood glucose measurements were obtained by the Critical Care Xpress,a Stat Profile M7,b Hitachi 717 automated chemistry,c and the Accu-check.d
The following data were also recorded on initial examination, where available: signalment, weight, heart rate, respiratory rate, rectal temperature, peripheral white cell count, peripheral neutrophil count and percentage, peripheral immature (band) neutrophil count and percentage, blood pressure measurements, blood culture results and the final diagnosis.
Animals were defined as having the SIRS if 2 or more of the following criteria were met: leucocytosis or leucopenia (peripheral white cell count >12.5 × 109/L or <4 × 109/L), >10% immature (band) neutrophils, hyperthermia or hypothermia (rectal temperature >39.2 °C or <37.2 °C), tachycardia (heart rate > 120 beats/min), tachypnea (respiratory rate > 30 breaths/min) and evidence of sepsis, cerebral ischemia or hypoxia, or trauma.13–16 Animals were defined as being septic if there was a positive blood culture at admission, or microbiological evidence of localized infection, and trauma was defined as clinical or radiological evidence of fractured ribs or head trauma.13 Survival was defined as survival to hospital discharge.
Blood glucose concentrations at admission and during the 1st 72 hours of hospitalization were compared with the reference range of 4.2–7.3 mmol/L (76–131 mg/dL). Blood glucose concentrations at admission, 24, 36, 48, 60 hours, and the 24-hour maximum and minimum glucose concentrations were compared in surviving and nonsurviving animals. The survival of animals with blood glucose concentrations <2.8 mmol/L (50 mg/dL) or >10 mmol/L (180 mg/dL) at admission was compared with those with blood glucose concentrations 2.8–10 mmol/L (50–180 mg/dL) at admission. These “extreme” ranges were set, before data analysis, in order to investigate the clinical observation that foals with particularly high or low glucose concentrations appeared to have the worst outcome. The survival of animals with blood glucose concentrations <4.22 mmol/L (76 mg/dL) and >7.3 mmol/L (131 mg/dL) at admission was compared with those with blood glucose concentrations 4.22–7.3 mmol/L (76–131 mg/dL). Blood glucose concentrations at admission were compared in animals with and without sepsis, a positive blood culture and SIRS. These relationships were examined using logistic regression, and quantified in terms of the odds ratio. A 1-way analysis of variance was used to investigate the association of data origin with blood glucose concentrations. All statistical analyses were performed using commercial statistical software.e,f Results were considered significant if the P value was <.05.
Five hundred and forty-nine foals had records available for evaluation of blood glucose concentrations, 515 of which had blood glucose concentrations recorded during the 1st 60 hours of hospitalization. These constituted the study population.
There were 251 foals for which sex was recorded, comprising 125 colts and 126 fillies. The average age was 34.4 hours (range 0–168 hours; n = 369) and weight 48.1 kg (range 9–130 kg; n = 266). 355 foals were septic, 152 foals were blood culture positive, and 153 had SIRS. No foals with head trauma were identified in the study. Five hundred and fifteen foals had blood glucose concentrations measured at admission, 159 at 24 hours after admission, 95 at 36 hours after admission, 82 at 48 hours after admission, and 45 at 60 hours after admission.
Blood Glucose Concentrations
Of 515 critically ill foals, 188 (36.5%) had glucose above the normal range, 177 (34.4%) below the normal range, and 150 (29.1%) had blood glucose concentrations in the normal range at admission (Fig 1). There was no effect of study institution on blood glucose concentrations.
Hypoglycemia (glucose < 4.2 mmol/L or 75.6 mg/dL) at admission was associated with a worse prognosis for survival to hospital discharge (P < .001), and for each 1 mmol/L (18 mg/dL) increase in blood glucose, the odds of survival to hospital discharge increased by 3.39 (95% CI 2.12–5.41). Hypoglycemia at admission was also associated with sepsis (P < .001), SIRS (P < .001), and a positive blood culture (P= .014). Hyperglycemia (blood glucose > 7.3 mmol/L) was not associated with a worse prognosis for survival to hospital discharge, sepsis, a positive blood culture, or SIRS. Extreme hyperglycemia (blood glucose > 10 mmol/L [180 mg/dL]) was associated with a worse prognosis to hospital discharge (P= .049). Extreme hypoglycemia (blood glucose < 2.8 mmol/L [50.4 mg/dL]) at admission was associated with a worse prognosis to hospital discharge compared with animals with blood glucose concentrations >2.8 mmol/L (>50.4 mg/dL) (P < .001). Extreme hypoglycemia was also associated with sepsis (P < .001) and a positive blood culture (P= .009), but not with SIRS (P= .076).
Animals that did not survive also had a lower blood glucose at the 1st sample after admission (P= .007), and higher maximum (P= .033) and lower minimum (P= .029) blood glucose concentrations in the 1st 24 hours following hospitalization, as well as a higher blood glucose at 24 hours (P < .001) and 36 hours (P= .025).
Blood glucose concentrations at 48 hours or any time point after 48 hours of hospitalization showed no association with survival to hospital discharge.
In this group of critically ill neonatal foals, derangements of blood glucose were very common, with 29.1% of animals being euglycemic at admission. Admission hypoglycemia and hyperglycemia at admission were common in this group of patients. This is in agreement with data from pediatric human intensive care patients, where 18.6% of patients were hypoglycemic, and 61% were hyperglycemic.7 In contrast, 50.2% of adult horses with acute abdominal disease were hyperglycemic at admission, but only 0.4% were hypoglycemic.8 Hypoglycemia could be more common in neonatal foals than in adult horses because foals are born with very low fat and glycogen stores.17 Foals that are not nursing are therefore prone to hypoglycemia. In addition, a high proportion of foals are septicemic at admission. Septicemia may result in hypoglycemia, possibly as a result of the low glycogen reserves and poor nursing behavior in these animals, or secondary to increased catabolism. In addition, asphyxia leads to the rapid metabolism of glucose by the brain and other tissues for energy, so foals that have suffered from an asphyxial or ischemic insult before, during or after parturition may be more prone to hypoglycemia.
Hypoglycemia was associated with nonsurvival to hospital discharge. Extreme hypoglycemia, or blood glucose <2.8 mmol/L (50 mg/dL), was also associated with nonsurvival to hospital discharge when compared with animals with blood glucose >2.8 mmol/L (50 mg/dL). This is in agreement with data from critically ill adult and pediatric humans, in which hypoglycemia is associated with a high case fatality rate.6,7 Hypoxemia, ischemia, and asphyxia increase the vulnerability of the newborn to the detrimental effects of hypoglycemia.18 This means that foals with a degree of perinatal asphyxia syndrome could be at heightened risk of suffering from the ill effects of hypoglycemia, and these patients may benefit the most from treatment aimed at normalizing blood glucose concentrations.
In this group of critically ill neonatal foals, hypoglycemia at admission was associated with sepsis, a positive blood culture, and SIRS. Extreme hypoglycemia (blood glucose concentration <2.8 mmol/L or 50 mg/dL) was associated with sepsis and a positive blood culture, but not with SIRS. The association of hypoglycemia with septicemia, bacteremia, and SIRS may reflect the severity of illness in these patients. More severely compromised foals will be less likely to nurse, and could therefore be more prone to hypoglycemia.
When defined as a blood glucose concentration of >7.2 mmol/L (131 mg/dL), hyperglycemia at admission was not associated with mortality in this group of critically ill equine neonates. However, “extreme hyperglycemia,” defined as blood glucose concentration >10 mmol/L (180 mg/dL), was associated with nonsurvival to hospital discharge. The association of hyperglycemia with nonsurvival agrees with data from critically ill horses8 and humans.1–4,7,19,20 In pediatric human intensive care patients, 68% of deaths occurred in patients with blood glucose levels over 11.11 mmol/L, defined as extreme hyperglycemia in this study.7 It seems that the degree of hyperglycemia is an important factor in deciding the prognosis of these patients. Hyperglycemia itself is thought to be detrimental during critical illness, and may lead to increased production of inflammatory cytokines, hypercoagulation, increased cellular apoptosis, and endothelial dysfunction.21 Hyperglycemia may also impair immune function,22 leaving the patient more vulnerable to infectious complications. It is therefore not surprising that the foals with the highest levels of blood glucose were less likely to survive, as these animals may have suffered most from the detrimental effects of hyperglycemia.
Foals that did not survive to hospital discharge had a lower 1st sample after admission, as well as higher maximum and lower minimum blood glucose concentration during the 1st 24 hours after hospital admission, and a higher blood glucose at 24 and 36 hours. In pediatric human patients, an increase in blood glucose variability is associated with nonsurvival.7 The maximum and minimum concentrations in the present study may reflect blood glucose variability, and thus agree with the data from pediatric human patients. However, this was not specifically measured. Alternatively, this may reflect that hypoglycemia and hyperglycemia at any point up to and including 36 hours after admission is associated with nonsurvival.
In this group of critically ill neonatal foals, blood glucose concentrations at or after 48 hours following admission were not associated with survival. This is in contrast to data from adult horses with acute abdominal disease, where blood glucose concentration up to and including 48 hours after admission was associated with survival.8 Many critically ill foals are placed on total or partial parenteral nutrition shortly after admission to the intensive care unit, with or without intensive glycemic management via the administration of exogenous insulin. In the study of adult horses, none were placed on partial or total parenteral nutrition in the 1st 48 hours of hospitalization.8 Using parenteral nutrition, with or without intensive glycemic management, will undoubtedly affect the blood glucose concentrations at any point in time. The use of IV glucose, partial or total parenteral nutrition, enteral nutrition, and insulin was not investigated in the present study, and may have affected blood glucose concentration in these foals, especially after the admission blood glucose measurement (at which point no animals had received supplemental nutrition in the hospital). Ideally, the management of the foals would be standardized so that all animals received the same type and amount of nutritional support; however, in a clinical situation and with a multicenter study, this is not possible. This may account for the lack of association of blood glucose at or following 48 hours after admission in the present study.
As a retrospective, multicenter study there are a number of limitations. The data in the present study were collected from several institutions, all of which used different methods of glucometry. We could find no difference in the mean glucose concentration in the foals from each participating institution, but this could hide extremes of measurement. A study of glucometry in an equine neonatal intensive care unit showed that a blood gas machine showed good agreement with a laboratory standard of measurement, whereas a glucometer had less than ideal agreement with the laboratory standard.23 However, the glucometer tested was 1 brand, and there might be a difference in the accuracy of various glucometers. A different brand proved to have close agreement with a laboratory standard when used with plasma in adult equine patients, but had less than ideal agreement when used with whole blood.24 However, as the data were collected from geographically varying areas and thus widely differing hospital populations, the results from the present study may be representative of global hospital populations. In addition, as with any retrospective study, the data collection was limited by the availability of the medical record, and also the availability of recorded blood glucose concentrations. The definition of nonsurvival was any foal that was not discharged from the hospital alive. As a retrospective study it was not possible to differentiate between foals that died, were euthanized due to their poor prognosis or intractable disease, or were euthanized due to financial concerns. In addition, it was not possible to investigate whether foals had any treatments before referral that could influence glucose concentrations at admission. Collecting similar data in a prospective manner may provide more information and allow further comparisons to be made. The “extreme” measurement groups were made based on clinical observations that foals with particularly high or low glucose concentrations appeared to have worse outcomes. The negative effect of “extreme” hyperglycemia was seen in a population of horses with acute abdominal disease,8 and the authors wished to investigate the possibility of a similar effect of extremes of glucose measurements in the neonatal foal.
The definition of normoglycemia is somewhat controversial in this age group, and very young foals might normally have a lower glucose concentration than older foals. However, this effect would, if anything, blunt the apparent association of hypoglycemia with nonsurvival in neonatal foals, and was therefore not further considered in the present study.
This study confirms that hypoglycemia and hyperglycemia are common in critically ill foals. At the present time it remains unclear whether these are mediators, or merely markers of mortality in the foal. In order to understand the mechanisms of development of hyperglycemia and hypoglycemia in the neonatal foal, insulin and glucagon concentrations should be investigated in these patients. In addition, it would be interesting to investigate whether controlling hyperglycemia with intensive insulin therapy and controlling hypoglycemia with glucose therapy would reduce mortality in these patients. This would require a large number of patients, and would probably necessitate a multicenter study.
In conclusion, derangements of blood glucose concentrations are very common in critically ill neonatal foals, with only 29.1% of these patients being euglycemic at admission. Hyperglycemia and hypoglycemia occur with approximately equal frequency in these patients. Hypoglycemia and extreme hyperglycemia at admission are associated with nonsurvival to hospital discharge. Foals that do not survive have a lower 1st blood glucose sample after admission, and higher maximum and lower minimum blood glucose concentrations during the 1st 24 hours of hospitalization, as well as higher blood glucose at 24 and 36 hours after admission. Hypoglycemia at admission was associated with sepsis, SIRS, and a positive blood culture. Extreme hypoglycemia was associated with sepsis and a positive blood culture. Hypoglycemia and extreme hyperglycemia at admission may therefore be useful prognostic indicators. In some groups of critically ill humans, intensive glycemic management improves prognosis in the intensive care unit. This may also be true in the critically ill equine neonate, and warrants further investigation.
aNOVA Biochemistry, Waltham, MA
cRoche Diagnostics, Indianapolis, IN
dRoche Diagnostics, NSW, Australia
eSPSS version 14, SPSS Inc, Chicago, IL
fStata 9.0 for Windows, Stata Corp LP, College Station, TX
The authors wish to thank Gemma Pearce, who collected some of the data presented in this study.