A simple, accurate test for identifying individual animals at increased risk of laminitis would aid prevention. Laminitis-prone ponies have a greater serum insulin response to dexamethasone administration than normal ponies in the summer, but the response during different seasons is unknown.
To test the hypothesis that previously laminitic ponies have a greater insulin response to dexamethasone than normal ponies, which is present during all seasons.
Prospective longitudinal study.
Overnight dexamethasone suppression tests were performed on 7 normal ponies and 5 previously laminitic ponies in spring 2009 and 2010, summer 2008 and 2010, autumn 2009 and winter 2008, while the ponies were at pasture. In spring 2010, a dexamethasone suppression test was performed after the ponies had been fed only hay for 3 weeks. Serum cortisol and insulin concentrations pre- and post dexamethasone were measured. Linear mixed models were used to analyse the data.
Insulin concentrations pre- and post dexamethasone were significantly higher in previously laminitic ponies than in normal ponies during spring 2009 and summer 2008, but there was no difference between groups in spring 2010, summer 2010, autumn 2009 or winter 2008. Insulin concentration varied significantly with season. Diet had no apparent effect on insulin concentration pre- or post dexamethasone in spring 2010. Cortisol concentrations post dexamethasone were significantly higher in previously laminitic ponies than in normal ponies in autumn 2009, with concentrations above the reference range (<25 nmol/l) in both groups in summer 2008 and autumn 2009. Individual ponies had insufficient cortisol suppression in all seasons.
There were significant differences between groups in insulin and cortisol concentrations post dexamethasone during some seasons, but this was not present in all years. Wide interindividual variation in response limits the usefulness of a dexamethasone suppression test for predicting the susceptibility of an individual animal to laminitis.
Abnormal insulin and cortisol responses to dexamethasone must be interpreted in the light of the individual animal, seasonal and annual variation reported here.
In groups of ponies at pasture, certain individuals appear predisposed to laminitis, whereas others seem less susceptible. Identification of ‘at risk’ individuals may help to prevent laminitis in susceptible animals. Endocrine conditions such as pituitary pars intermedia dysfunction (PPID) increase the susceptibility to laminitis. Recently, there has also been interest in the links between insulin resistance, equine metabolic syndrome and pasture-associated laminitis [1-4]. In ponies, peripheral tissue insulin resistance is usually compensated by an increase in pancreatic insulin secretion, resulting in hyperinsulinaemia . However, basal hyperinsulinaemia is an unreliable indicator of insulin resistance owing to individual and diurnal variation in insulin concentrations [6, 7]. Dynamic endocrine testing is thus required to identify insulin resistance.
Dexamethasone suppression tests (DSTs) have been used to document abnormalities in control of insulin secretion in ponies by measuring the change in serum insulin concentration post dexamethasone injection . Laminitis-prone ponies had exaggerated increases in serum insulin concentration compared with normal ponies when tested in summer, and all individuals had normal cortisol suppression . The effect of season on insulin responses to dexamethasone has not been reported in ponies. Other investigators found no seasonal effect on insulin concentration post dexamethasone in normal horses in southern USA . However, basal insulin concentrations in grazing horses in the USA vary seasonally, with concentrations peaking in April  or September , which is positively correlated with pasture carbohydrate content [10, 11]. There was no seasonal variation in insulin concentration in samples obtained from stabled horses fed hay [10, 11].
The DST is commonly used to screen horses and ponies for PPID by documenting failure to suppress serum cortisol concentration post dexamethasone injection . However, it has been shown that the degree of cortisol suppression in clinically normal horses and ponies is affected by season, because 21% of 29 ponies and 40% of 10 horses in north-eastern USA had insufficient cortisol suppression in September, but all suppressed serum cortisol normally in January . When tested monthly for 12 months, up to 53% of 18 clinically normal horses in Michigan had abnormal cortisol suppression from May to October, with the peak number of abnormal tests occurring in August . Season also had a significant effect on cortisol concentration in normal horses in southern USA .
The objective of the present study was to determine whether there were any differences between clinically normal and previously laminitic ponies in serum insulin and cortisol concentrations pre- and post dexamethasone injection during different seasons. A further objective was to document seasonal and annual variation in insulin and cortisol concentrations pre- and post dexamethasone in these ponies.
Materials and methods
Twelve ponies were studied, 7 of which had no recorded history of laminitis (normal ponies; all intact females) and 5 of which suffered recurrent bouts of pasture-associated laminitis, including at least one episode of laminitis in the previous 3 years, diagnosed on clinical signs by experienced equine veterinary surgeons (previously laminitic ponies: 4 intact females and one gelding). None of the ponies had clinical signs of laminitis at the time of testing, was obese or showed any clinical signs of PPID. The mean ± s.d. age and weight of the normal ponies was 19 ± 4.3 years and 314 ± 71 kg and of previously laminitic ponies was 16.3 ± 2.7 years and 289 ± 60 kg, respectively. Body weight did not change significantly with season. The median (interquartile range) body condition score (on a scale of 1–9)  of the normal ponies was 6 (2) and of previously laminitic ponies was 5 (2), which did not vary with season. All were mixed, native British pony breeds (predominantly New Forest and Welsh ponies and cross-breeds) and were unrelated.
Overnight DSTs were performed in spring (April 2009 and April 2010), summer (July 2008 and July 2010), autumn (October 2009) and winter (December 2008) in southern England, while the ponies were maintained on the same pasture. Pasture was supplemented with ad libitum grass hay in December. In spring 2010 (May), a second DST was performed after the ponies had been kept in a dirt paddock without access to grass for 3 weeks, eating only ad libitum hay. In summer 2010, a subset of the ponies (4 normal ponies and 3 previously laminitic ponies) was available to be tested. Food and water were not withheld before any of the tests. Not all DSTs could be performed on the same day within each test month; however, they were performed on all ponies within a few days of each other.
Dexamethasonea 0.04 mg/kg bwt was injected i.m. between 16.00 and 17.00 h, and blood was collected by jugular venepuncture into serum Vacutainers before and 19 h after injection. Blood was allowed to clot at 37°C for 20 min before centrifugation (3000 g) at 4°C for 10 min. Serum was harvested and stored at -80°C. Serum cortisol analysis was by chemiluminescent assay at a commercial laboratoryb, and serum insulin was analysed using a commercially available radioimmunoassay (RIA) kitc, previously validated for use in horses and ponies in the authors' laboratory. When this kit is used to measure high concentrations of insulin in equine samples, dilutional parallelism does not occur if the zero calibrator supplied in the kit is used as a diluent, as recommended by the manufacturers . However, dilutional parallelism is seen when insulin-depleted equine serum (IDS) is used as a diluent . All samples were assayed undiluted, and samples that were expected to contain insulin concentrations above the working range of the RIA (>350 μiu/ml) were diluted using IDS prepared by continuously mixing 50 mg activated charcoald per 1 ml serum for 20 h at 20°C. Serum was then centrifuged twice (3000 g) at 4°C for 20 min, decanting the supernatant after each centrifugation. Serum was then filtered through a 0.2 μm syringe filtere and stored at -80°C. A sample of IDS was analysed in each RIA to confirm that no insulin was detected.
Weather records for 2008–2010 for southern England were obtained from Met Office records online and examined for differences in sunshine hours, rainfall and mean temperature during the spring and summer months of the different years.
Linear mixed modelsf were used to analyse the data, with individual pony as the subject and insulin or cortisol concentrations pre- or post dexamethasone as the dependent variable. Season was included as a repeated measure, with a compound symmetry repeated covariance format. Group (normal ponies or previously laminitic ponies), season and the interaction between group and season were included as fixed effects. Bonferroni post hoc tests were used where significant differences were found in the output of the mixed model. Normality was assessed using visual inspection of histograms and the Kolmogorov–Smirnov test. Results are presented as means ± s.d. if normally distributed or median (interquartile range) if not normally distributed, and significance was set at P≤0.05.
Results from the linear mixed models are shown in Table 1. Insulin concentration predexamethasone was significantly (P<0.001) higher in previously laminitic ponies than in normal ponies during spring 2009 and summer 2008 (Fig 1). In normal ponies, insulin concentration predexamethasone was significantly (P = 0.045) higher in spring 2009 compared with the first test in spring 2010. In previously laminitic ponies, insulin concentrations predexamethasone in spring 2009 and summer 2008 were significantly (P≤0.03) higher than in both tests in spring 2010, summer 2010, autumn and winter (Fig 1). There was no significant difference in insulin concentration predexamethasone in previously laminitic ponies between spring 2009 and summer 2008 (P = 0.072).
Table 1. Linear mixed model results
Insulin concentration predexamethasone
Insulin concentration post dexamethasone
Cortisol concentration predexamethasone
Cortisol concentration post dexamethasone
Each column shows output of a linear mixed model performed using the specified dependent variable. Fixed effects tested were group, season and the interaction between group and season.
P = 0.475
P = 0.091
P = 0.536
Group × season interaction
P = 0.008
P = 0.431
P = 0.045
Insulin concentration post dexamethasone was significantly (P<0.001) higher in previously laminitic ponies compared with normal ponies in spring 2009 and summer 2008 (Fig 2). In normal ponies, insulin concentration post dexamethasone did not vary significantly with season. In previously laminitic ponies, insulin concentrations post dexamethasone in spring 2009 and in summer 2008 were significantly (P<0.001) greater than in both tests in spring 2010, summer 2010, autumn and winter (Fig 2). There was no significant difference in insulin concentration post dexamethasone in previously laminitic ponies between spring 2009 and summer 2008 (P = 0.31).
The difference in insulin concentration post vs. predexamethasone paralleled the insulin concentrations post dexamethasone (data not shown). All ponies increased insulin concentrations post dexamethasone apart from one normal pony in spring 2009, in which insulin concentration decreased by 74 μiu/ml (from 240 to 166 μiu/ml).
In summer 2008, insulin concentration post dexamethasone varied widely. Two previously laminitic ponies had insulin concentrations <200 μiu/ml and the other 3 had insulin concentrations >1200 μiu/ml. Three normal ponies had insulin concentrations post dexamethasone <60 μiu/ml and 3 had insulin concentrations >200 μiu/ml. The final normal pony had an intermediate insulin concentration of 105 μiu/ml.
In spring 2009, one normal pony had an insulin concentration post dexamethasone >500 μiu/ml, and all 5 previously laminitic ponies had insulin concentrations >800 μiu/ml. There was less variation in spring 2010, with 6 of 7 normal ponies having insulin concentrations post dexamethasone <50 μiu/ml and 4 of 5 previously laminitic ponies having insulin concentrations post dexamethasone <110 μiu/ml. There was no significant difference in insulin concentrations pre- or post dexamethasone between the first and 2nd tests in spring 2010 (before and after 3 weeks of grass restriction) in either group.
Serum insulin concentration in winter in both groups was low and responded minimally to dexamethasone injection (mean ± s.d. in normal pony pre, 10 ± 4 μiu/ml and post, 26 ± 15 μiu/ml; previously laminitic ponies pre, 13 ± 11 μiu/ml and post, 28 ± 18 μiu/ml).
Cortisol was not measured in the summer 2010 test. Cortisol concentration post dexamethasone was significantly (P = 0.021) higher in previously laminitic ponies than in normal ponies in autumn (Fig 3). In normal ponies, cortisol concentration post dexamethasone was significantly higher in summer 2008 than in spring 2009 (P = 0.035) and in the second test in spring 2010 (P = 0.025; Fig 3). Cortisol concentration post dexamethasone in normal ponies was also significantly higher in autumn compared with the second test in spring 2010 (P = 0.042). In previously laminitic ponies, cortisol concentration post dexamethasone in autumn was significantly higher than in previously laminitic ponies during all other seasons (P≤0.01; Fig 3). No suppression of serum cortisol occurred in response to dexamethasone in previously laminitic ponies in autumn (median [interquartile range]: pre, 83  nmol/l and post, 81  nmol/l).
Cortisol concentration post dexamethasone was above the laboratory reference range (<25 nmol/l) in both groups in summer 2008 and autumn 2009. Individual ponies did not suppress serum cortisol normally in the first test in spring 2010 (one normal pony), summer 2008 (one normal pony; 2 previously laminitic ponies), autumn 2009 (4 normal ponies; 4 previously laminitic ponies) or winter 2008 (one normal pony; one previously laminitic pony). Different individual ponies had insufficient cortisol suppression in the different seasons. Over all seasons, 4 normal ponies had insufficient cortisol suppression (one normal pony had one abnormal result and 3 normal ponies had 2 abnormal results). Four previously laminitic ponies had insufficient cortisol suppression in at least one season (2 previously laminitic ponies had one abnormal result; one previously laminitic pony had 2 abnormal results; and one previously laminitic pony had 3 abnormal results). Three normal ponies and one previously laminitic pony had normal cortisol suppression during all 6 tests. All individual ponies had normal cortisol suppression post dexamethasone in spring 2009 and in the second test in spring 2010.
Weather records for 2008–2010 showed that January and February 2010 were colder than 2008 or 2009, followed by a drier, sunnier April 2010 (Table 2).
Table 2. Weather records for southern England for January–July 2008–2010
None of the ponies developed clinical signs of laminitis following injection of dexamethasone at any time.
Season significantly affected insulin concentrations in the present study. Basal insulin concentration varied significantly with season and year in both groups in the present study, in agreement with previous studies [6, 7, 11], making this unsuitable for estimation of insulin sensitivity in horses. Significantly increased basal insulin concentrations have been reported in horses in the USA in both April and September, correlating with pasture carbohydrate content [10, 11]. In the present study, basal insulin concentrations differed significantly between normal ponies and previously laminitic ponies in spring 2009 and summer 2008, but this difference was not present in spring or summer 2010. Additionally, there was no difference between groups in autumn or winter.
Insulin concentration post dexamethasone showed a similar pattern to basal insulin, with a significant difference between groups in spring 2009 and summer 2008 but not in spring or summer 2010, autumn or winter. There was no significant seasonal variation in insulin concentration post dexamethasone in normal ponies, but seasonal differences were apparent in previously laminitic ponies. Although the median insulin concentration post dexamethasone was higher in previously laminitic ponies than in normal ponies in summer 2008, there was a wide variation in individual results. Two of 5 previously laminitic ponies had insulin concentrations post dexamethasone that were lower than 3 of 7 normal ponies tested at this time. This restricts the usefulness of the DST for individual ponies. In spring 2009, an insulin concentration post dexamethasone of >550 μiu/ml identified individual ponies based on their history of laminitis. However, in spring 2010, only one previously laminitic pony had an insulin concentration >550 μiu/ml post dexamethasone. All other normal ponies and previously laminitic ponies had insulin concentrations post dexamethasone of <200 μiu/ml. This annual variation in insulin response further limits the usefulness of the test for an individual pony.
The insulin concentrations which should be regarded as indicative of insulin resistance in horses and ponies are still controversial and subject to debate. Recently published guidelines have recommended that fasting insulin concentrations >20 μiu/ml should be considered hyperinsulinaemic in normal horses and ponies, although this ‘normal’ range is dependent on the diagnostic assay used . However, other authors have reported higher basal insulin concentrations (27 ± 8 μiu/ml) in Thoroughbred mares at pasture and not fasted before sampling . This was suggested to indicate lower insulin sensitivity in these grazing horses, but could be due to the effect of grazing continuously before blood sampling . The ponies in the present study were not fasted before blood sampling and were mainly out at pasture, so they would be expected to have high basal insulin concentrations. Additionally, ponies have higher and more variable basal insulin concentrations than other breeds [7, 17, 18].
Previous investigators have designated a cut-off value for plasma insulin of >32 μiu/ml in spring as predictive of the development of clinical laminitis within 3 months . In the present study, 4 of 7 normal ponies had basal insulin concentrations above this cut-off value in spring 2009, yet none have ever been diagnosed with clinical laminitis. All of the previously laminitic ponies in the present study had basal insulin concentrations >32 μiu/ml in spring 2009 but did not develop laminitis in the next 3 months. In the previously reported study , samples were collected before concentrate feeding, but it was not stated whether the ponies were fasted before sampling, which could account for some of the differences between basal concentrations in the 2 studies. The ponies in the previously reported study  were also obese (median body condition score 6.5–7.5), which was not the case for the ponies investigated in the study reported here. Nevertheless, results of the present study suggest that normal ponies may have higher basal insulin concentrations than previously reported.
There were significant differences between insulin concentrations in different years in both groups. The reasons for this variation are unclear but may be due to factors such as environmental and management conditions.
The weather records showed that January and February 2010 were colder than the same months in 2008 or 2009, but April 2010 was drier and sunnier. As a result, the grass in 2010 may have been of poorer quality, with a lower water-soluble carbohydrate (sugars and fructans) content. One of the limitations of the present study was that pasture analysis and comparison of the grazing conditions between seasons was not performed. However, pasture water-soluble carbohydrate content can vary considerably over short periods of time. If differences in pasture water-soluble carbohydrate did occur, this may have caused less stimulation of pancreatic insulin secretion, which could account for the lower basal insulin concentrations and responses to injection of dexamethasone in 2010. There was no effect on insulin concentrations when the ponies did not have access to grass and were fed only hay for 3 weeks (between the first and second tests in spring 2010). This suggests that they were not ingesting high quantities of water-soluble carbohydrate in the grass during spring 2010.
The ponies in the present study are managed in an attempt to prevent laminitis by restricting exposure to grass during perceived high-risk periods, and it is also possible that this management was better in 2010 than in 2008 and 2009, meaning the ponies were less insulin resistant in 2010. Stress is unlikely to have affected the cortisol and insulin results at different seasons, because all tests were performed by the same experienced operator and the ponies were maintained in the same conditions for all testing dates and were well accustomed to handling and blood sampling.
In autumn and winter, DSTs could not distinguish between normal ponies and previously laminitic ponies. In winter, there was minimal change in insulin concentrations in response to dexamethasone and the results were almost identical for normal ponies and previously laminitic ponies. The reasons for the lack of insulin response to dexamethasone at this time of year are unknown but may include variation in diet and nutritional status or photoperiod.
Seasonal changes in the pituitary–adrenal axis have been reported in other seasonal mammals, including sheep and squirrels [20, 21]. It is difficult to separate seasonal changes from concurrent alterations in other factors, such as nutrient availability and reproductive status. Nutritional status may be a key factor affecting responses to dexamethasone in the present study. Previous studies have shown a significant reduction in basal insulin concentration in laminitis-prone, but not normal, ponies 7 days after switching from a grass to a hay diet . Likewise, addition of inulin (a fructan carbohydrate) to a hay-based diet resulted in significantly greater increases in serum insulin concentration in previously laminitic ponies than in normal ponies . These results show that increasing the water-soluble carbohydrate content of the diet using either grass or inulin affected control of insulin secretion in previously laminitic ponies. Other studies have also reported that insulin concentration in grazing horses is positively correlated with pasture carbohydrate content [10, 11]. The water-soluble carbohydrate content of the pasture was likely to have been much lower in the winter compared with the other testing periods, and this may explain the lack of insulin response to dexamethasone at this time of year.
Most of the ponies in the present study were female (11 of 12). No specific oestrous activity was observed during sampling, although further measurements and reproductive ultrasonography were not performed. It is therefore not possible to relate changes in reproductive activity and stage of the oestrous cycle to the different responses to dexamethasone in the present study.
Although there was no significant difference in age between groups, both groups of ponies were aged. It is possible that different results may have been obtained in younger individuals. In female rhesus monkeys, cortisol concentrations 19 h post dexamethasone were significantly higher in aged than in young monkeys . The sensitivity of cortisol secretion to glucocorticoid negative feedback may reduce with ageing .
The results of the present study have also confirmed the findings of previous investigators of abnormal cortisol suppression in summer and autumn [9, 12, 13]. There was no difference in basal cortisol between seasons in the present study, in agreement with previous studies . Plasma adrenocorticotrophic hormone is also used to diagnose PPID but is also subject to seasonal variation, increasing in the late summer and autumn months [11, 18]. The abnormalities detected in the autumn in the present study are thus unlikely to be due to diagnostic inaccuracy of the DST at this time of year but reflect genuine alterations in the pituitary–adrenal axis, with a seasonal upregulation in late summer and autumn .
Although both groups of ponies in the present study suppressed serum cortisol in response to dexamethasone in summer 2008, the median values for both groups post dexamethasone were above the laboratory reference range. All ponies had at least 3 normal results (suppression of serum cortisol) during the course of the study, although 4 individual ponies in each group in the present study had insufficient cortisol suppression in at least one test. Insufficient cortisol suppression occurred during all seasons, indicating that inaccurate results may occur at any time of year. Caution is therefore necessary in interpreting insufficient cortisol suppression in response to dexamethasone in animals without clinical signs of PPID.
In conclusion, seasonal and annual variation in cortisol and insulin responses to dexamethasone occurred in normal ponies and previously laminitic ponies. Although the test was able to distinguish between normal ponies and previously laminitic ponies based on insulin responses on some occasions, the variation in results between different seasons and years means that its usefulness in an individual animal is limited. Further work is necessary to determine the reasons for the seasonal and annual variation in results.
Authors' declaration of interests
There are no conflicts of interest to disclose.
Ethical animal research
The work was carried out under a project licence granted under the Animals (Scientific Procedures) Act 1986 and was approved by the institutional ethics and welfare committee.
Source of funding
K.E.B.-W. was funded by a BBSRC Case Studentship in association with the Waltham Centre for Pet Nutrition.
All authors contributed to study design, data interpretation and editing of the manuscript. K.E.B.-W. designed the study, collected data, executed the study, analysed and interpreted data and wrote the manuscript.