Equine Veterinary Journal ISSN 0425-1644 DOI: 10.1111/evj.12391
Increased serum nonesterified fatty acid and low ionised calcium concentrations are associated with post partum colic in mares S. J. HOLCOMBE*, R. M. EMBERTSON†, K. A. KURTZ, H. A. ROESSNER, S. E. WISMER, R. J. GEOR and J. B. KANEENE Department of Large Animal Clinical Sciences, College of Veterinary Medicine, Michigan State University, East Lansing, Michigan, USA † Rood and Riddle Equine Hospital, Lexington, Kentucky, USA. *Correspondence email: [email protected]; Received: 09.04.14; Accepted: 17.11.14
Summary Reasons for performing study: Increased serum nonesterified fatty acids (NEFA) and decreased serum electrolytes are linked to abdomasal displacements in post partum dairy cattle. Post partum colic in mares may be associated with metabolic changes specific to pregnancy and the periparturient period. Objectives: To determine if fluctuations in serum NEFA, ionised calcium (iCa) and magnesium (iMg) occurred in periparturient mares and if these alterations were associated with post partum colic. Study design: Longitudinal observational study. Methods: Mares from 3 farms in central Kentucky were enrolled. Blood samples were collected 14 days prior to the estimated foaling date, within 4 days post parturition, and 14 and 28 days after foaling for batch analysis of serum NEFA, iCa and iMg. Health information was provided by farm managers and veterinarians. Data were analysed using Kruskal–Wallis χ2 statistic for nonparametric data and a matched case/control approach. Repeated measures logistic regression models were developed. Results: Serum NEFAs were higher at 14–1 day before foaling (mean ± s.d., mmol/l), 0.28 ± 0.12, P = 0.04 and from foaling to 4 days after foaling, 0.29 ± 0.20 (P = 0.05) in mares that developed colic compared with those that did not colic, 0.19 ± 0.05 and 0.21 ± 0.14, respectively. Ionised calcium was lower at 15–28 days post foaling in mares that showed colic, 1.50 ± 0.17 compared to mares that did not colic, 1.60 ± 0.12, P = 0.02. Risk of colic in post partum mares increased 38% for each 0.1 mmol/l increase in serum NEFA (odds ratio = 1.38, 95% confidence interval 1.06–1.81, P = 0.02). Conclusions: Mares with post partum colic had significantly higher serum NEFA and lower iCa prior to the colic episode compared with mares that did not develop colic. Monitoring these metabolic alterations may lead to predictive and preventive colic strategies for post partum mares. Keywords: horse; nonesterified fatty acids; parturition; mare; calcium; colic
Introduction Post partum mares are at risk for various forms of colic, including enterocolitis, ileus and large colon volvulus [1–5]. Retrospective reports show that episodes of colic occurred most frequently within 7 days of foaling but up to 90 days post partum [1–5]. Gastric and small intestinal ileus occurred in Thoroughbred mares on average 11 days post partum (range 1–62 days) between February and May [4]. Risk factors for large colon volvulus included being a broodmare, foaling within the last 90 days, increased hours of stabling in the previous 14 days, eating hay or sugar beet (beet pulp), and a change in pasture in the past 28 days [5]. The cause and effect relationships between these factors and the occurrence of colic were not established. The association between parturition and intestinal diseases in horses is well supported but the aetiology remains unclear. The predisposition to post partum gastrointestinal disease is not unique to horses and is well described in dairy cows. After calving, 5–7% of dairy cattle developed abomasum displacements [6,7]. Similar to horses, displacements occurred 11 days post partum (range 3–30 days) [6,7]. The negative energy balance that occurred in late gestation and early lactation, reflected by increased nonesterified fatty acid (NEFA) concentrations in the blood, was associated with abdomasal displacements [6,7]. Elevated concentrations of NEFA were shown to impair various immune and inflammatory functions in dairy cows [8,9]. Similar metabolic responses may occur in periparturient mares and increased serum NEFA concentrations might reflect increased risk for post partum colic. Altered calcium homeostasis also occurs in the periparturient period. Calcium is mobilised in horses during the last trimester to support fetal growth and in the first weeks after foaling to prepare for peak lactation [10–13]. Calcium directly affects gastrointestinal motility because calcium mediates the contractile activity of intestinal smooth muscle [14,15]. Hypocalcaemia is linked to gastrointestinal motility disorders, including development of ileus in horses [15–17]. Furthermore, calcium potentiates the proinflammatory and antimicrobial functions of neutrophils because ionised calcium (iCa) is a second messenger critically involved in activation Equine Veterinary Journal 48 (2016) 39–44 © 2015 EVJ Ltd
of both NADPH oxidase and neutrophil degranulation [18,19]. Ionised calcium depletion in the periparturient period may result in altered gastrointestinal motility and impaired immune function. Hypocalcaemia can be a consequence of hypomagnesaemia [20]. Nineteen of 35 horses admitted with surgical colic had low serum ionised magnesium concentrations (iMg) at the time of hospital admission [16]. Horses that developed post operative ileus had significantly lower serum iMg concentrations compared with the horses that did not experience ileus [16]. Depletion of iMg may alter gastrointestinal motility directly or indirectly by diminishing iCa. The objectives of this study were to determine if fluctuations in serum NEFA, iCa and iMg occurred in periparturient mares and if NEFA, iCa or iMg concentrations predicted colic in post partum mares. The hypothesis of this study was that increased serum NEFA concentrations and decreased serum iCa and iMg concentrations in the periparturient period would be associated with the development of post partum colic in mares.
Materials and methods A longitudinal observational study was performed using broodmares housed on 3 breeding farms in Lexington, Kentucky from January to July 2012 during the foaling season.
Horses All healthy, pregnant mares were eligible for inclusion in the study. Mares were excluded if they aborted the pregnancy during the study, left the farm, or developed colic during the prepartum period. Four nonpregnant mares, selected by the farm managers from each farm, were enrolled to serve as nonpregnant controls. Mares were deemed healthy based on the results of physical examinations and current vaccination and de-worming status. Information regarding feeding, management, foaling dates and episodes of colic were provided by the farm managers and veterinarians. Each of the 3 farms had slightly different feeding and housing management strategies for the mares on the farm. The mares from each specific farm were fed, housed and managed similarly.
39
Changes in serum NEFA and calcium predict colic in post partum
Sampling Twenty millilitres of blood were collected by venipuncture using 18 gauge, 3.175 cm vacutainer needlesa into sterile 10 ml BD serum separator Vacutainer vialsa. Blood samples were collected from the periparturient mares approximately 14 days prior to the estimated foaling date, within 4 days after parturition, and 14 and 28 days after foaling. If a mare was >14 days beyond the anticipated foaling date and had not delivered, the 14 day prepartum blood sample was repeated. Therefore, many mares had samples collected approximately 28 and 14 days prior to foaling. No samples were collected after an episode of colic such that all samples from mares with post partum colic were obtained prior to the colic episode. Blood was drawn from the nongestational mares every 14 days during the course of the study. Samples were collected at various times of the day to coincide with other veterinary procedures that were performed on the mares. None of the mares were sedated or received medications prior to blood collection. All samples were postprandial because mares had 24 h access to hay. Blood samples were sent to the laboratory at Rood and Riddle Equine Hospital, centrifuged for 10 min at 5056 g using an HN-SII general purpose centrifugeb to separate the serum, and frozen at -23°C [21]. Samples were retrieved from the Rood and Riddle laboratory and transported in freezer containers to Michigan State University every 2 weeks and stored at -23°C for batch analysis within 6 months of collection [21,22].
NEFA, iCa and iMg measurements Serum was harvested for measurement of NEFA and iCa and iMg. Nonesterified faty acid concentrations were quantified by the enzymatic colorimetric method, ACS-ACOD (NEFA-HR [2]c), as outlined by Panzani et al., and modified for use in a 96-well microplate format [23]. Briefly, 200 μl of Color Reagent A solution was added to 5.0 μl of standards (serial dilution of NEFA Standard solution 276–76491c), and to samples and a control serum sample and incubated for 10 min at 37°C. Absorbance 1 (Abs1) was read at 550/660 nm using a SpectraMax 340PCd microplate reader. This measurement served as the standard/sample blank. Next, 100 μl of Color Reagent B solution was added to each well and incubated at 37°C for an additional 10 min. Following the incubation, absorbance 2 (Abs2) was read at 550/660 nm. The final absorbance for each well was determined by subtracting Abs1 from Abs2. Nonesterified fatty acid concentration of each sample was then determined using the constructed linear standard curve (absorbance vs. concentration). Standards and samples were assayed in duplicate and the intra- and interassay coefficients of variation were 2.92% and 2.85%, respectively. Ionised calcium and magnesium concentrations were measured using the NOVA8 – Ion-selective analysere at the MSU Diagnostic Center for Populations and Animal Health.
Colic Episodes of colic were recorded for mares on each farm by veterinarians. Data for colic episodes were collected from the time of foaling until 150 days post partum. This time range was chosen to include lactation. Colic was defined as an episode of abdominal pain based on the mare’s behaviour (pawing and rolling) that lasted >2 h and required administration of analgesics. Final diagnoses and treatment (surgical or medical) were obtained from veterinarians on each farm and from veterinarians at Rood and Riddle Equine Hospital if the mares were hospitalised.
Data analysis Descriptive statistics of the entire dataset: Descriptive statistics were generated for all horses in the study, including pregnant mares that experienced colic (cases), pregnant mares that did not colic and nonpregnant mares from the 3 farms. The distributions of serum NEFA, iMg and iCa were assessed using the Shapiro–Wilk test for normality (PROC UNIVARIATE, SAS 9.3)f, which indicated that serum NEFA, iMg and iCa were not normally distributed. Consequently, associations between serum NEFA, iMg and iCa, colic status, and other risk factors that might influence serology were assessed using the Kruskal–Wallis χ2 statistic for the nonparametric data. These risk factors included mare age, the time interval when the sample was collected in relation to the foaling date (28–15 days
40
S. J. Holcombe et al.
prior to foaling, 14–1 day prior to foaling, foaling to 4 days after foaling, 5–14 days after foaling, 15–28 days after foaling), farm and month of sample collection (January, February, March, April, May–June). The case–control study dataset: To describe associations between colic and serum concentrations of NEFA, iCa and iMg from pregnant mares, a matched case–control approach was used with mares from the 3 breeding farms. Colic (case) mares on each farm were identified and matched to 1–2 noncolic herd mates (controls) that foaled close to the same date, to control for the effects of sample month, sampling time interval and farm. Controls were also matched to case mares by age as closely as possible. For mares with colic, only samples that were collected prior to the colic event were included in the analyses. Repeated measures logistic regression models, matching by case and including farm ID to pregnant noncolic herd mates (control) for herd effect (PROC GENMOD, SAS 9.3)f, were developed for individual mare colic status to identify risk factors associated with the occurrence of colic in pregnant mares. Univariable models were developed for each potential risk factor, and factors with P values ≤0.2 were considered for inclusion in the multivariable regression model. Potential interaction was tested by creating logistic regression models with individual risk factors and interaction terms, but no significant interactions were identified. A backward stepwise model building approach was used to develop the final multivariable logistic regression model for colic mares from the matched case–control dataset. All risk factors meeting the inclusion criteria (P≤0.2) were entered into the full model, and the final version of the model was the version with the best combination of individual risk factor P values and model goodness-of–fit (assessed by the Quasi-likelihood information criterion, a modification of the Akaike information criterion for the generalised estimating equations used for the repeated measures regression model). To test for confounding in the final model, comparisons in each risk factor coefficient from the final model were compared with the risk factor coefficient from the univariable model, using change in coefficients >10% as evidence of confounding (there was no evidence of confounding in these models). The effects of risk factors on the occurrence of colic in pregnant mares in the final model were reported as odds ratios with 95% confidence intervals (CI), adjusted to reflect biologically relevant change (0.1 units) in serum levels of NEFA, iMg and iCa.
Results Study population Eleven of 244 mares were excluded from the study because they aborted the pregnancy, developed prepartum colic or left the farm. Data from 233 mares, 221 pregnant and 12 nonpregnant mares were included in the analysis. Three farms in central Kentucky participated in the study with 64, 103 and 54 pregnant mares and 4 nonpregnant mares from each farm. A total of 924 blood samples were collected. Ninety-six samples were from nonpregnant mares, 80 samples from periparturient mares that developed colic and 780 samples from periparturient mares that did not colic.
Colic Twenty-four of the 221 pregnant mares (11%) had at least one episode of post partum colic diagnosed by a veterinarian. The mean ± s.d. age of the mares that developed post partum colic was 9 ± 3 years. Six (25%) mares experienced colic in February, 7 (29%) in March, 5 (21%) in April, 2 (8%) in May, 3 (13%) in June and one (4%) in July. Median time from foaling to development of colic signs was 16 days (range 1–137 days). The interval from foaling to colic for 10 mares was 0–7 days, 8–14 days for one mare, 15–30 days for 4 mares, 31–60 days for 4 mares, 61–90 days for one mare, 91–120 days for 2 mares and 121–137 days for 2 mares. Large colon volvulus occurred 2, 6, 53 and 104 days post partum. Nine mares with colic required surgical correction and 15 mares were treated medically for colic. Diagnoses of surgically treated colic included large colon volvulus (n = 4), nephrosplenic ligament entrapment of the large colon (n = 1), small intestine incarcerated in a mesenteric rent (n = 1), caecal bruise (n = 1), ventral colon impaction (n = 1) and descending colon rupture (n = 1). Diagnosis for medically managed colics included tympany or gas, (n = 8), spasmodic colic, (n = 4), colonic impaction, (n = 2) and enteritis, (n = 1). Equine Veterinary Journal 48 (2016) 39–44 © 2015 EVJ Ltd
Changes in serum NEFA and calcium predict colic in post partum
S. J. Holcombe et al.
Serum NEFA concentrations in 197 pregnant mares without colic and 24 pregnant mares that experienced post partum colic 0.45 *
0.4
NEFA (mmol/l)
Fig 1: Mean serum nonesterified fatty acid (NEFA) ± s.d. concentrations from 14 days prepartum to 4 days post partum for pregnant mares that developed post partum colic were significantly higher compared with concentrations measured in pregnant mares that did not develop post partum colic. The range for serum NEFA concentration for nonpregnant mares from January to June is indicated by the grey band. * indicates significant differences in serum NEFA concentrations between the periparturient mares that developed colic and the periparturient mares that did not colic, P≤0.05. Error bars represents s.d.
Pregnant + colic Pregnant + no colic
*
0.35 0.3 0.25 0.2 0.15 0.1 0.05 0
-28 to 15 days
NEFA descriptive data for the entire study population
-14 to -1 day
+1 to 4 days after foaling
+5 to 14 days
+15 to 28 days
mares compared with nonpregnant mares at 28–15 days before foaling (P = 0.001; Fig 2). Serum iCa was similar for all pregnant mares from 14 days before to 14 days after foaling and were comparable with mean values measured in the nonpregnant mares during the study period (Fig 2). Serum iCa concentrations were significantly lower in pregnant mares that developed post partum colic at 15–28 days post foaling compared with pregnant mares that did not colic, P = 0.02, Fig 2. The iCa concentrations for the nonpregnant mares were 1.57 ± 0.10 mmol/l from January to June 2012. There were no significant differences in serum iMg between pregnant mares that developed colic, pregnant mares that did not colic and the nonpregnant mares.
Sampling time relative to foaling, sampling month, farm, pregnancy and colic significantly affected serum NEFA concentrations, P
Increased serum nonesterified fatty acid and low ionised calcium concentrations are associated with post partum colic in mares S. J. HOLCOMBE*, R. M. EMBERTSON†, K. A. KURTZ, H. A. ROESSNER, S. E. WISMER, R. J. GEOR and J. B. KANEENE Department of Large Animal Clinical Sciences, College of Veterinary Medicine, Michigan State University, East Lansing, Michigan, USA † Rood and Riddle Equine Hospital, Lexington, Kentucky, USA. *Correspondence email: [email protected]; Received: 09.04.14; Accepted: 17.11.14
Summary Reasons for performing study: Increased serum nonesterified fatty acids (NEFA) and decreased serum electrolytes are linked to abdomasal displacements in post partum dairy cattle. Post partum colic in mares may be associated with metabolic changes specific to pregnancy and the periparturient period. Objectives: To determine if fluctuations in serum NEFA, ionised calcium (iCa) and magnesium (iMg) occurred in periparturient mares and if these alterations were associated with post partum colic. Study design: Longitudinal observational study. Methods: Mares from 3 farms in central Kentucky were enrolled. Blood samples were collected 14 days prior to the estimated foaling date, within 4 days post parturition, and 14 and 28 days after foaling for batch analysis of serum NEFA, iCa and iMg. Health information was provided by farm managers and veterinarians. Data were analysed using Kruskal–Wallis χ2 statistic for nonparametric data and a matched case/control approach. Repeated measures logistic regression models were developed. Results: Serum NEFAs were higher at 14–1 day before foaling (mean ± s.d., mmol/l), 0.28 ± 0.12, P = 0.04 and from foaling to 4 days after foaling, 0.29 ± 0.20 (P = 0.05) in mares that developed colic compared with those that did not colic, 0.19 ± 0.05 and 0.21 ± 0.14, respectively. Ionised calcium was lower at 15–28 days post foaling in mares that showed colic, 1.50 ± 0.17 compared to mares that did not colic, 1.60 ± 0.12, P = 0.02. Risk of colic in post partum mares increased 38% for each 0.1 mmol/l increase in serum NEFA (odds ratio = 1.38, 95% confidence interval 1.06–1.81, P = 0.02). Conclusions: Mares with post partum colic had significantly higher serum NEFA and lower iCa prior to the colic episode compared with mares that did not develop colic. Monitoring these metabolic alterations may lead to predictive and preventive colic strategies for post partum mares. Keywords: horse; nonesterified fatty acids; parturition; mare; calcium; colic
Introduction Post partum mares are at risk for various forms of colic, including enterocolitis, ileus and large colon volvulus [1–5]. Retrospective reports show that episodes of colic occurred most frequently within 7 days of foaling but up to 90 days post partum [1–5]. Gastric and small intestinal ileus occurred in Thoroughbred mares on average 11 days post partum (range 1–62 days) between February and May [4]. Risk factors for large colon volvulus included being a broodmare, foaling within the last 90 days, increased hours of stabling in the previous 14 days, eating hay or sugar beet (beet pulp), and a change in pasture in the past 28 days [5]. The cause and effect relationships between these factors and the occurrence of colic were not established. The association between parturition and intestinal diseases in horses is well supported but the aetiology remains unclear. The predisposition to post partum gastrointestinal disease is not unique to horses and is well described in dairy cows. After calving, 5–7% of dairy cattle developed abomasum displacements [6,7]. Similar to horses, displacements occurred 11 days post partum (range 3–30 days) [6,7]. The negative energy balance that occurred in late gestation and early lactation, reflected by increased nonesterified fatty acid (NEFA) concentrations in the blood, was associated with abdomasal displacements [6,7]. Elevated concentrations of NEFA were shown to impair various immune and inflammatory functions in dairy cows [8,9]. Similar metabolic responses may occur in periparturient mares and increased serum NEFA concentrations might reflect increased risk for post partum colic. Altered calcium homeostasis also occurs in the periparturient period. Calcium is mobilised in horses during the last trimester to support fetal growth and in the first weeks after foaling to prepare for peak lactation [10–13]. Calcium directly affects gastrointestinal motility because calcium mediates the contractile activity of intestinal smooth muscle [14,15]. Hypocalcaemia is linked to gastrointestinal motility disorders, including development of ileus in horses [15–17]. Furthermore, calcium potentiates the proinflammatory and antimicrobial functions of neutrophils because ionised calcium (iCa) is a second messenger critically involved in activation Equine Veterinary Journal 48 (2016) 39–44 © 2015 EVJ Ltd
of both NADPH oxidase and neutrophil degranulation [18,19]. Ionised calcium depletion in the periparturient period may result in altered gastrointestinal motility and impaired immune function. Hypocalcaemia can be a consequence of hypomagnesaemia [20]. Nineteen of 35 horses admitted with surgical colic had low serum ionised magnesium concentrations (iMg) at the time of hospital admission [16]. Horses that developed post operative ileus had significantly lower serum iMg concentrations compared with the horses that did not experience ileus [16]. Depletion of iMg may alter gastrointestinal motility directly or indirectly by diminishing iCa. The objectives of this study were to determine if fluctuations in serum NEFA, iCa and iMg occurred in periparturient mares and if NEFA, iCa or iMg concentrations predicted colic in post partum mares. The hypothesis of this study was that increased serum NEFA concentrations and decreased serum iCa and iMg concentrations in the periparturient period would be associated with the development of post partum colic in mares.
Materials and methods A longitudinal observational study was performed using broodmares housed on 3 breeding farms in Lexington, Kentucky from January to July 2012 during the foaling season.
Horses All healthy, pregnant mares were eligible for inclusion in the study. Mares were excluded if they aborted the pregnancy during the study, left the farm, or developed colic during the prepartum period. Four nonpregnant mares, selected by the farm managers from each farm, were enrolled to serve as nonpregnant controls. Mares were deemed healthy based on the results of physical examinations and current vaccination and de-worming status. Information regarding feeding, management, foaling dates and episodes of colic were provided by the farm managers and veterinarians. Each of the 3 farms had slightly different feeding and housing management strategies for the mares on the farm. The mares from each specific farm were fed, housed and managed similarly.
39
Changes in serum NEFA and calcium predict colic in post partum
Sampling Twenty millilitres of blood were collected by venipuncture using 18 gauge, 3.175 cm vacutainer needlesa into sterile 10 ml BD serum separator Vacutainer vialsa. Blood samples were collected from the periparturient mares approximately 14 days prior to the estimated foaling date, within 4 days after parturition, and 14 and 28 days after foaling. If a mare was >14 days beyond the anticipated foaling date and had not delivered, the 14 day prepartum blood sample was repeated. Therefore, many mares had samples collected approximately 28 and 14 days prior to foaling. No samples were collected after an episode of colic such that all samples from mares with post partum colic were obtained prior to the colic episode. Blood was drawn from the nongestational mares every 14 days during the course of the study. Samples were collected at various times of the day to coincide with other veterinary procedures that were performed on the mares. None of the mares were sedated or received medications prior to blood collection. All samples were postprandial because mares had 24 h access to hay. Blood samples were sent to the laboratory at Rood and Riddle Equine Hospital, centrifuged for 10 min at 5056 g using an HN-SII general purpose centrifugeb to separate the serum, and frozen at -23°C [21]. Samples were retrieved from the Rood and Riddle laboratory and transported in freezer containers to Michigan State University every 2 weeks and stored at -23°C for batch analysis within 6 months of collection [21,22].
NEFA, iCa and iMg measurements Serum was harvested for measurement of NEFA and iCa and iMg. Nonesterified faty acid concentrations were quantified by the enzymatic colorimetric method, ACS-ACOD (NEFA-HR [2]c), as outlined by Panzani et al., and modified for use in a 96-well microplate format [23]. Briefly, 200 μl of Color Reagent A solution was added to 5.0 μl of standards (serial dilution of NEFA Standard solution 276–76491c), and to samples and a control serum sample and incubated for 10 min at 37°C. Absorbance 1 (Abs1) was read at 550/660 nm using a SpectraMax 340PCd microplate reader. This measurement served as the standard/sample blank. Next, 100 μl of Color Reagent B solution was added to each well and incubated at 37°C for an additional 10 min. Following the incubation, absorbance 2 (Abs2) was read at 550/660 nm. The final absorbance for each well was determined by subtracting Abs1 from Abs2. Nonesterified fatty acid concentration of each sample was then determined using the constructed linear standard curve (absorbance vs. concentration). Standards and samples were assayed in duplicate and the intra- and interassay coefficients of variation were 2.92% and 2.85%, respectively. Ionised calcium and magnesium concentrations were measured using the NOVA8 – Ion-selective analysere at the MSU Diagnostic Center for Populations and Animal Health.
Colic Episodes of colic were recorded for mares on each farm by veterinarians. Data for colic episodes were collected from the time of foaling until 150 days post partum. This time range was chosen to include lactation. Colic was defined as an episode of abdominal pain based on the mare’s behaviour (pawing and rolling) that lasted >2 h and required administration of analgesics. Final diagnoses and treatment (surgical or medical) were obtained from veterinarians on each farm and from veterinarians at Rood and Riddle Equine Hospital if the mares were hospitalised.
Data analysis Descriptive statistics of the entire dataset: Descriptive statistics were generated for all horses in the study, including pregnant mares that experienced colic (cases), pregnant mares that did not colic and nonpregnant mares from the 3 farms. The distributions of serum NEFA, iMg and iCa were assessed using the Shapiro–Wilk test for normality (PROC UNIVARIATE, SAS 9.3)f, which indicated that serum NEFA, iMg and iCa were not normally distributed. Consequently, associations between serum NEFA, iMg and iCa, colic status, and other risk factors that might influence serology were assessed using the Kruskal–Wallis χ2 statistic for the nonparametric data. These risk factors included mare age, the time interval when the sample was collected in relation to the foaling date (28–15 days
40
S. J. Holcombe et al.
prior to foaling, 14–1 day prior to foaling, foaling to 4 days after foaling, 5–14 days after foaling, 15–28 days after foaling), farm and month of sample collection (January, February, March, April, May–June). The case–control study dataset: To describe associations between colic and serum concentrations of NEFA, iCa and iMg from pregnant mares, a matched case–control approach was used with mares from the 3 breeding farms. Colic (case) mares on each farm were identified and matched to 1–2 noncolic herd mates (controls) that foaled close to the same date, to control for the effects of sample month, sampling time interval and farm. Controls were also matched to case mares by age as closely as possible. For mares with colic, only samples that were collected prior to the colic event were included in the analyses. Repeated measures logistic regression models, matching by case and including farm ID to pregnant noncolic herd mates (control) for herd effect (PROC GENMOD, SAS 9.3)f, were developed for individual mare colic status to identify risk factors associated with the occurrence of colic in pregnant mares. Univariable models were developed for each potential risk factor, and factors with P values ≤0.2 were considered for inclusion in the multivariable regression model. Potential interaction was tested by creating logistic regression models with individual risk factors and interaction terms, but no significant interactions were identified. A backward stepwise model building approach was used to develop the final multivariable logistic regression model for colic mares from the matched case–control dataset. All risk factors meeting the inclusion criteria (P≤0.2) were entered into the full model, and the final version of the model was the version with the best combination of individual risk factor P values and model goodness-of–fit (assessed by the Quasi-likelihood information criterion, a modification of the Akaike information criterion for the generalised estimating equations used for the repeated measures regression model). To test for confounding in the final model, comparisons in each risk factor coefficient from the final model were compared with the risk factor coefficient from the univariable model, using change in coefficients >10% as evidence of confounding (there was no evidence of confounding in these models). The effects of risk factors on the occurrence of colic in pregnant mares in the final model were reported as odds ratios with 95% confidence intervals (CI), adjusted to reflect biologically relevant change (0.1 units) in serum levels of NEFA, iMg and iCa.
Results Study population Eleven of 244 mares were excluded from the study because they aborted the pregnancy, developed prepartum colic or left the farm. Data from 233 mares, 221 pregnant and 12 nonpregnant mares were included in the analysis. Three farms in central Kentucky participated in the study with 64, 103 and 54 pregnant mares and 4 nonpregnant mares from each farm. A total of 924 blood samples were collected. Ninety-six samples were from nonpregnant mares, 80 samples from periparturient mares that developed colic and 780 samples from periparturient mares that did not colic.
Colic Twenty-four of the 221 pregnant mares (11%) had at least one episode of post partum colic diagnosed by a veterinarian. The mean ± s.d. age of the mares that developed post partum colic was 9 ± 3 years. Six (25%) mares experienced colic in February, 7 (29%) in March, 5 (21%) in April, 2 (8%) in May, 3 (13%) in June and one (4%) in July. Median time from foaling to development of colic signs was 16 days (range 1–137 days). The interval from foaling to colic for 10 mares was 0–7 days, 8–14 days for one mare, 15–30 days for 4 mares, 31–60 days for 4 mares, 61–90 days for one mare, 91–120 days for 2 mares and 121–137 days for 2 mares. Large colon volvulus occurred 2, 6, 53 and 104 days post partum. Nine mares with colic required surgical correction and 15 mares were treated medically for colic. Diagnoses of surgically treated colic included large colon volvulus (n = 4), nephrosplenic ligament entrapment of the large colon (n = 1), small intestine incarcerated in a mesenteric rent (n = 1), caecal bruise (n = 1), ventral colon impaction (n = 1) and descending colon rupture (n = 1). Diagnosis for medically managed colics included tympany or gas, (n = 8), spasmodic colic, (n = 4), colonic impaction, (n = 2) and enteritis, (n = 1). Equine Veterinary Journal 48 (2016) 39–44 © 2015 EVJ Ltd
Changes in serum NEFA and calcium predict colic in post partum
S. J. Holcombe et al.
Serum NEFA concentrations in 197 pregnant mares without colic and 24 pregnant mares that experienced post partum colic 0.45 *
0.4
NEFA (mmol/l)
Fig 1: Mean serum nonesterified fatty acid (NEFA) ± s.d. concentrations from 14 days prepartum to 4 days post partum for pregnant mares that developed post partum colic were significantly higher compared with concentrations measured in pregnant mares that did not develop post partum colic. The range for serum NEFA concentration for nonpregnant mares from January to June is indicated by the grey band. * indicates significant differences in serum NEFA concentrations between the periparturient mares that developed colic and the periparturient mares that did not colic, P≤0.05. Error bars represents s.d.
Pregnant + colic Pregnant + no colic
*
0.35 0.3 0.25 0.2 0.15 0.1 0.05 0
-28 to 15 days
NEFA descriptive data for the entire study population
-14 to -1 day
+1 to 4 days after foaling
+5 to 14 days
+15 to 28 days
mares compared with nonpregnant mares at 28–15 days before foaling (P = 0.001; Fig 2). Serum iCa was similar for all pregnant mares from 14 days before to 14 days after foaling and were comparable with mean values measured in the nonpregnant mares during the study period (Fig 2). Serum iCa concentrations were significantly lower in pregnant mares that developed post partum colic at 15–28 days post foaling compared with pregnant mares that did not colic, P = 0.02, Fig 2. The iCa concentrations for the nonpregnant mares were 1.57 ± 0.10 mmol/l from January to June 2012. There were no significant differences in serum iMg between pregnant mares that developed colic, pregnant mares that did not colic and the nonpregnant mares.
Sampling time relative to foaling, sampling month, farm, pregnancy and colic significantly affected serum NEFA concentrations, P
