Materials and Methods

Forty-four horses and 4 ponies were used. All procedures were approved by an institutional animal care and use committee, and owners of privately owned animals gave consent to the tests. Eleven horses and 4 ponies had clinical signs of PPID, 4 horses had equivocal signs of PPID, and 29 horses had no clinical signs of PPID. Horses and ponies in the clinical PPID group obviously had at least 2 of the classical clinical signs such as laminitis, hirsutism, abnormal hair shedding, abnormal fat deposition (near the tail head, over the hindquarters, along the dorsum of the neck, or in the supraorbital fossae), lethargy, loss of epaxial muscle mass and pendulous abdomen, or recurrent infections. Polydipsia and polyuria, 2 additional classical signs, were not used as diagnostic criteria because many horses were housed outdoors, which prevented accurate assessment. Horses were allocated to the equivocal category if their clinical appearance was suspicious but not classical or if they had only 1 sign that suggested PPID; none had the typical long or wavy hair coat. Clinically normal horses had none of the clinical signs of PPID. Within this group, 1 horse developed clinical signs of PPID after the first TRH test and further tests were performed when the horse was clinically abnormal. Sex, breed, age, body weight as determined by use of a weight tape, and BCS were recorded. The latter was determined by use of established criteria.^19,20 In 15 horses, the TRH test was repeated; in 13, the test was repeated during a different season than the original test. One horse had 5 tests, 3 had 4 tests, 3 had 3 tests, and 8 had 2 tests. Except for 4 horses located at the Atlantic Veterinary College, Prince Edward Island, Canada, all horses and ponies were located in the mid-Atlantic region of the United States.

Twenty-four horses without signs of PPID, 4 horses and 1 pony with signs of PPID, and 4 horses with equivocal signs of PPID were euthanized and necropsied within 22 days of a TRH test. One pony with PPID was necropsied 150 days after the test. The clinically normal and equivocally affected horses were euthanized for reasons unrelated to endocrine or metabolic disease. A gross postmortem examination was performed, and midsagittal sections of the pituitary gland were examined histologically after fixation in neutral-buffered 10% formalin. Horses were classified as abnormal if there was microscopic evidence of hyperplasia of the pars intermedia. A diagnosis of pituitary pars intermedia hyperplasia was made when a sagittal section within 1 mm of the true midline that included a portion of the third ventricle had cellular proliferations projecting into the pars nervosa. Fourteen sections were also reviewed later by a different pathologist. Horses were grouped into the following categories: PPID-0PM (n = 9), PPID-PH (6), CN-0PM (5), CN-0PH (10), CN-PH (14), and EQ-PH (4). All PPID and EQ horses that were necropsied had pituitary pars intermedia hyperplasia.

The TRH test and DST were performed by the same technicians (except for the 4 horses at the Atlantic Veterinary College), with horses in their familiar surroundings. An 18-ga IV catheter^c was aseptically placed 30 minutes before the TRH test. Two baseline jugular blood samples were obtained 5 minutes apart prior to injecting 1 mg of synthetic TRH^d and flushing the catheter with 10 mL of heparinized saline (0.9% NaCl) solution. The TRH was reconstituted via sterile technique under a biohazard hood with sterile Dulbecco PBS^e to a concentration of 1 mg of TRH/mL. One-milliliter aliquots were filtered with a 0.22-Mm syringe filter^f and stored in sterile Eppendorf tubes at –70°C until use. Subsequent blood samples were obtained at 2, 4, 6, 8, 10, 12, 14, 30, 60, and 180 minutes after injection of TRH, and the catheter was flushed with heparinized saline solution between each sample. A volume of 6 mL was aspirated and discarded prior to obtaining each blood sample. Samples were collected in plastic 6-mL potassium EDTA tubes that were centrifuged for separation of plasma within 45 minutes of collection. Plasma was stored at −20°C in plastic tubes, placed on ice packs, and sent by overnight mail to the Animal Health Diagnostic Center at Cornell University. A sequential immunometric assay^g that used chemiluminescence for signal generation was used to measure ACTH concentration.² The test is generally specific for ACTH fragment 1-39 but has approximately 12% to 14% cross-reactivity with fragment 18-39.

For evaluation of possible seasonal effect, the year 2004 was examined for duration of daylight hours in Chester County by use of a Web site²¹ and a curve was constructed depicting daylight length changes. On the basis of dates that horses were tested, daylight intervals evaluated were an increase in daylight from 10 to 12.5 hours, duration of daylight from 13 to 15 hours, decrease in daylight from 12.5 to 10 hours, and < 10 hours of daylight.

Variables evaluated as predictors of histologic pituitary pars intermedia hyperplasia were AUC for ACTH concentration, baseline ACTH concentration, maximum ACTH concentration, and percentage increase in ACTH concentration from baseline. To evaluate the association of these variables with season, clustered regression analysis was used. Clustering enabled use of results of repeated tests on horses. The kinetic invariants (1 slope and 2 intercepts) were estimated by use of software^h by fitting biexponential models directly to the ACTH concentrations for each horse.^22,23 The Fisher exact test was used to verify that groups were equally represented across seasons. The curve was analyzed to determine whether values at each sampling time differed among groups. Mean ACTH concentrations were compared among groups at 5 and 10 minutes prior to and 10, 30, and 60 minutes after injection of TRH. To establish the relationship between the variables and outcome (PH at postmortem), clustered logistic regression analysis was used. Final selection among the variables as best predictors of PH at postmortem was based on sensitivity-specificity versus cutoff probability plots, receiver-operator curve plots, and results of the Fisher exact test. All analyses were performed with software. Except as indicated, P < 0.05 was considered significant.

Because of the wide variation in ACTH values, data were converted to a log scale for ease of analysis. Regression analysis with the Bonferroni correction for multiple comparisons was used to evaluate ACTH values at baseline and 10 and 30 minutes after TRH administration. When the Bonferroni correction was applied, P < 0.003 was used. The log of the percentage increase from baseline ACTH concentrations to that at various time points; log of the AUC; and log of the maximum ACTH value and that at baseline and 2, 4, 6, 8, 10, 14, 30, and 60 minutes were compared among age groups (≤ 10 years, > 10 to < 13 years, 13 to < 15 years, 15 to b�20 years, and > 20 years), clinical and postmortem categories, sex, and seasons. Because the published reference range for plasma ACTH concentration in the laboratory is 9 to 35 pg/mL, the sensitivity and specificity of ACTH > 35 pg/mL for diagnosing PPID were determined at baseline and 30 and 60 minutes after TRH administration. No attempt was made to statistically evaluate potential influence of sex or breed on results of the TRH test because of lack of even distribution among the groups of horses.

Seventeen horses were ≤ 10 years; only 1 horse had clinical signs of PPID. Thirteen horses were > 10 to < 15 years of age; 6 were CN-PH, 3 were CN-0PH, 1 was EQ-PH, and 3 were PPID-0PM. Eight horses were from 15 to 20 years of age; 3 had clinical signs, 1 had equivocal signs, and 4 had no clinical signs. Ten horses were > 20 years old; 8 had clinical signs of PPID, and 2 had equivocal signs. The PPID group was composed of 4 castrated males and 11 females; the clinically normal group contained 20 castrated males, 8 females, and 1 male. The EQ group was composed of 3 castrated males and 1 female. There were 14 Thoroughbreds (11 clinically normal and 3 with signs of PPID) and 11 Quarter Horses (8 clinically normal and 3 with equivocal signs of PPID). Only the group with PPID had Morgan Horses (n = 3) and ponies (4), and only the clinically normal group had warmbloods (3). A mixture of other breeds comprised 5 members of the PPID group, 7 of the clinically normal group, and 1 in the group with equivocal signs. Body weight (mean ± SD) was 241 ± 41 kg (530 ± 90 lb) for the ponies, 487 ± 57 kg (1,071 ± 125 lb) for horses with PPID and EQ-PH horses, 530 ± 82 kg (1,166 ± 180 lb) for CN-0PM and CN-0PH horses, and 494 ± 63 kg (1,087 ± 139 lb) for CN-PH horses. Body condition score was 5.3 ± 1 for ponies, 5.2 ± 1.6 for horses with PPID and EQ-PH horses, 5 ± 1.3 for CN-0PM and CN-0PH horses, and 4.8 ± 1.1 for CN-PH horses. Except for the ponies, there were no significant difference in weights among these groups, and the BCS did not differ significantly among groups.

Dexamethasone suppression tests were performed on 2 CN-0PM horses (n = 3 tests), 7 CN-0PH horses, 10 CN-PH horses (12), 5 PPID-0PM horses (11), 3 PPID-PH horses, and 3 EQ-PH horses (6). The test was repeated in different seasons in 5 horses. The DST was performed from 5 to 8 hours after administration of TRH. This time interval allowed sufficient time between the tests because cortisol and ACTH concentrations returned to baseline from 1 to 3 hours after TRH administration. After obtaining a baseline sample, dexamethasone (40 Mg/kg [18 Mg/lb])ⁱ was administered IM in the left side of the neck between 2:00 and 3:00 PM and a second serum sample was collected between 8:00 and 9:00 AM the following day. Plasma was stored at −20°C in plastic tubes and handled as described. Cortisol concentrations were measured by use of a chemiluminescent immunoassay validated for horses.²⁴ Results of the DST were considered abnormal if cortisol concentration exceeded 1.0 μg/dL 18 to 20 hours after administration of dexamethasone.

Results

Baseline plasma ACTH concentrations were within reference range (≤ 35 pg/mL) in all CN-0PH horses, in 6 of 7 tests in 5 CN-0PM horses, in 9 of 14 tests in CN-PH horses, in 6 of 16 tests in 9 PPID-0PM horses, in 2 tests in PPID-PH horses, and in 3 of 8 tests in 4 EQ-PH horses. Concentrations of ACTH at 5 and 10 minutes prior to and 10, 30, and 60 minutes after TRH administration were examined (Table 1). In the EQ-PH group in 2 horses' tests, 1 baseline concentration exceeded 36 pg/mL and the other baseline sample was < 35 pg/mL. One PPID-0PM horse had baseline ACTH concentrations within reference range in 2 of 4 tests. Two other PPID-0PM horses that were tested multiple times never had baseline values within reference range. Variations in baseline ACTH values (r 10 pg/mL) were detected during an individual TRH test in 11 of the 30 tests in horses with PPID and EQ-PH horses; in 4 of these individual tests, there was a difference > 50 pg/ mL. The CN-0PH horses had the lowest and least variable baseline ACTH concentrations (mean ± SD, 18.1 ± 2.6 pg/mL and 19.7 ± 4.4 pg/mL). Baseline samples varied by < 6 pg/mL in all clinically normal horses except 1 with PH, in which values were 79.5 and 146 pg/mL. Five of 14 CN-PH horses had baseline concentrations r 35 and b 150 pg/mL. In the PPID-0PM and PPID-PH groups, baseline ACTH concentrations were extremely variable.

After administration of TRH, the ACTH concentrations increased in all horses, but the increase was significantly greater in horses with PPID. Peak ACTH concentrations occurred in the 2-, 4-, or 6-minute samples in all groups; in 9 tests, concentration peaked at 2 minutes; in 37 tests, concentration peaked at 4 minutes; and in 15 tests, concentration peaked at 6 minutes. Six of 14 CN-PH horses had a peak ACTH concentration from 200 to 500 pg/mL. None of the CN-0PM horses (n = 5; 7 tests) had ACTH concentration > 150 pg/mL at any time. The ACTH concentration exceeded 500 pg/mL in 10 of 16 tests in 10 PPID-0PM horses and in 5 of 6 PPID-PH horses. At 10 minutes, none of the CN-0PH horses had ACTH concentrations > 81 pg/mL, whereas values exceeded this in 7 of 14 tests in CN-PH horses, 14 of 16 tests in PPID-0PM horses, and all PPID-PH horses. In the 10 CN-0PH horses, only 1 had an ACTH concentration > 35 pg/mL at 30 minutes (43.5 pg/mL). The ACTH concentration exceeded 35 pg/mL at 30 minutes in 7 of 14 CN-PH horses and in all PPID-PH horses. The ACTH concentrations for each of the groups were compared at 10 minutes prior to and 10, 30, and 60 minutes after TRH administration. At 10 minutes after TRH administration, values in the PPID-PH group differed significantly from values in the CN-0PM, CN-0PH, and CN-PH groups, but comparisons of ACTH concentrations among all groups for each of the times by use of the Bonferroni correction revealed no other significant differences. Horses older than 12 years of age were significantly more likely to have an abnormal response to TRH with significantly higher log baseline ACTH concentration, log AUC, log percentage increase ACTH concentration, and log maximum ACTH concentration; however, all 15 horses > 12 years of age that were necropsied had PH, confounding any effect of age. The log AUC, log baseline ACTH concentration, log percentage increase in ACTH concentration, log maximum ACTH concentration, and log ACTH concentration at 10 and 30 minutes were significantly greater for horses with PPID, compared with clinically normal horses. Values in the CN-PH group did not differ from those in the CN-0PH group. Values in the PPID-0PM, PPID-PH, and EQ-PH groups differed significantly from those in the CN-0PH and CN-PH groups (Table 2; Figure 1).

Box-and-whisker plots of log-transformed values of variables associated with ACTH concentrations in various groups of horses administered TRH. Horizontal line indicates median, box indicates interquartile range, bars indicate the range, and circles indicate outliers. lbase = Log of values obtained before administration of TRH. lauc = Log of area under the ACTH concentration curve. lpost 10 min = Log of ACTH values obtained 10 minutes after administration of TRH. lmax = Log of maximum ACTH concentration. lpct increase = Log of percentage increase in ACTH concentration after administration of TRH. lpost 30 min = Log of ACTH values obtained 30 minutes after administration of TRH.

Citation: Journal of the American Veterinary Medical Association 231, 3; 10.2460/javma.231.3.417

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Box-and-whisker plots of log-transformed values of variables associated with ACTH concentrations in various groups of horses administered TRH. Horizontal line indicates median, box indicates interquartile range, bars indicate the range, and circles indicate outliers. lbase = Log of values obtained before administration of TRH. lauc = Log of area under the ACTH concentration curve. lpost 10 min = Log of ACTH values obtained 10 minutes after administration of TRH. lmax = Log of maximum ACTH concentration. lpct increase = Log of percentage increase in ACTH concentration after administration of TRH. lpost 30 min = Log of ACTH values obtained 30 minutes after administration of TRH.

Citation: Journal of the American Veterinary Medical Association 231, 3; 10.2460/javma.231.3.417

• Download Figure
• Download figure as PowerPoint slide

Box-and-whisker plots of log-transformed values of variables associated with ACTH concentrations in various groups of horses administered TRH. Horizontal line indicates median, box indicates interquartile range, bars indicate the range, and circles indicate outliers. lbase = Log of values obtained before administration of TRH. lauc = Log of area under the ACTH concentration curve. lpost 10 min = Log of ACTH values obtained 10 minutes after administration of TRH. lmax = Log of maximum ACTH concentration. lpct increase = Log of percentage increase in ACTH concentration after administration of TRH. lpost 30 min = Log of ACTH values obtained 30 minutes after administration of TRH.

Citation: Journal of the American Veterinary Medical Association 231, 3; 10.2460/javma.231.3.417

• Download Figure
• Download figure as PowerPoint slide

Three measurements of ACTH taken at baseline and 30 and 60 minutes after TRH administration were examined for specificity and sensitivity to determine whether a horse had PPID by use of > 35 pg/mL as the cutoff value. In horses that were necropsied and had multiple tests, analyses were performed on the test result obtained closest to the time of necropsy. When CN-0PH horses (n = 10) were compared with PPID-PH horses (6), the sensitivities for baseline, 30-minute, and 60-minute samples were 71%, 100%, and 86%, respectively, and specificities were 100%, 89%, and 89%, respectively. When CN-0PH horses were compared with CN-PH horses (n = 14), sensitivities for baseline, 30-minute, and 60-minute samples were 36%, 50%, and 43%, respectively, and specificities were 100%, 89%, and 89%, respectively, for PH. To determine whether the model predicted the histologic diagnosis of pars intermedia hyperplasia (PPID-PH and CN-PH groups), logistic regression with a predicted cut point probability > 0.5 revealed high sensitivity for baseline (5- and 10-minute samples) and 30-minute ACTH concentrations and generally lower specificity, except for the 30-minute ACTH value for which specificity was slightly greater than sensitivity. Sensitivity, positive predictive value, and correct diagnosis rates were 100%, 72%, and 72%, respectively, by use of the baseline value of > 35 pg of ACTH/mL. Because no horses with histologically normal pituitary glands had baseline values > 35 pg/mL, it was not possible to determine specificity or negative predictive value. By use of the log ACTH values obtained 5 minutes before TRH administration, the sensitivity, positive predictive value, specificity, negative predictive value, and correct diagnosis rates for PH were 86%, 75%, 24%, 40%, and 69%, respectively. By use of the log ACTH values obtained 10 minutes before TRH administration, these values were 89%, 75%, 24%, 44%, and 70%. At 30 minutes after administration of TRH, by use of the value > 35 pg/mL to predict histologic PH, sensitivity was 77%, positive predictive value was 92%, specificity was 82%, and negative predictive value was 58%, yielding a correct diagnosis rate of 72%. The 30-minute log ACTH value permitted differentiation between the presence and absence of PH with a sensitivity of 91%, specificity of 44%, positive predictive value of 82%, and negative predictive value of 64%, resulting in a 79% correct diagnosis rate.

Although a seasonal effect was evident when data from all the horses were analyzed, numbers of horses tested were not evenly distributed among the different seasons. There was no statistical evidence of an actual association of groups with seasons (Table 3). During the time of daylight shortening from 12.5 to 10 hours (autumn), the log baseline ACTH concentration, log maximum ACTH concentration, log AUC, log 10-minute ACTH concentration, and 30-minute ACTH concentration were significantly higher, compared with the periods of daylight < 10 hours (winter) and the time of increase from 10 to 12.5 hours duration (spring); the log percentage increase in ACTH concentration was significantly greater, compared with daylight < 10 hours. During the time from 13 to 15 hours of daylight (summer), the log maximum ACTH concentration, log AUC, log percentage increase in ACTH concentration, and log 10-minute ACTH concentration were significantly higher than when daylight was < 10 hours. The log maximum ACTH concentration and log percentage increase in ACTH concentration were also significantly greater for light duration of 13 to 15 hours, compared with light increasing from 10 to 12.5 hours. The difference in log 30-minute ACTH concentration between the 13- to 15-hour period and < 10-hour period almost reached significance (P = 0.053). There were no differences in any variables between the period of light increasing from 10 to 12.5 hours and the period of light duration from 13 to 15 hours. Among 13 horses with tests repeated in different seasons, 5 horses' baseline ACTH values were higher in September, compared with samples obtained in other months. Baseline ACTH concentrations were higher in 4 of the 9 tests conducted between September 10 and November 17, when daylight was decreasing from 12.5 to 10 hours. The seasonal effect did not change the individual horse's group classification.

Following administration of TRH, 6 horses had muscle trembling that varied from minor and transient in 3 horses to moderate and generalized, but of less than several minutes' duration in the other 3 horses. When the test was repeated in 2 of the latter horses, no adverse effects were seen. Many horses licked their lips immediately after administration, a few horses yawned, several had flehmen, and 2 horses transiently coughed.

Results of the DST were within reference range in 7 CN-0PH horses, and in 7 of 12 times, the test was performed in 10 CN-PH horses. One CN-PH horse had both a normal and an abnormal test result, and another had 2 abnormal test results. In 3 CN-PH horses, cortisol concentration did not decrease to < 1.0 μg/dL from a high baseline value (11.9, 15.0, 12.0, and 9.0 μg/dL in the horse with 2 tests). In another horse, cortisol concentration did not decrease to < 1.0 μg/dL from a baseline value of 2.9 μg/dL. One of 3 PPID-PH horses had results of the DST that were within reference range. Results of 9 of the 11 DSTs performed in 5 PPID-0PM horses were within reference range, and 1 horse had normal and abnormal results. Results of 6 DSTs in 3 EQ-PH horses were within reference range. There was no seasonal effect on results of the DST. Results of the DST were abnormal only in PPID-PH, PPID-0PM, and CN-PH horses (9/26 tests), but results were more frequently within reference range (17/26) in those groups. In addition, 3 of the horses with an abnormal DST result also had a result within reference range when tested on another occasion.

Sensitivity and specificity of the DST were determined for the 3 PPID-PH and 6 CN-0PH horses. Specificity was 100%, and sensitivity was 66%, but the sensitivity decreased to 23% when the PPID group included the 10 PPID-0PM horses. The test was 100% specific in differentiating between CN-PH and CN-0PH horses, but sensitivity was only 33%. When both the DST and TRH tests were performed in the same horse within a 24-hour period, results of the DST were within reference range and results of the TRH test were abnormal 19 times, both were abnormal 4 times, both were within reference range 11 times, and results of the DST were abnormal and results of the TRH test were within reference range in 2 CN-PH horses.

Discussion

Because we considered it likely that administration of TRH would have a maximal effect far earlier than 30 minutes, as in a previous report,¹⁵ we wished to evaluate the response curve for ACTH concentration; a marked early response was seen, and the time for return to baseline concentrations was determined. Unlike the previous report, the increase in ACTH in CN-0PH horses in the present study was not significant at 30 minutes after administration.

As expected, administration of TRH elicited an increase in ACTH concentrations in clinically normal and abnormal horses, but the increase, maximum ACTH concentration, and persistence of high concentrations (ie, the AUC) were greater in abnormal horses. The TRH test also revealed abnormal responses in older clinically normal horses that had pars intermedia hyperplasia. Because none of the older horses that were necropsied had histologically normal pituitary glands, it was not possible to accurately determine specificity of the test in older horses. The relevance of mild pars intermedia hyperplasia with regard to clinical PPID is still unclear. Results of a recent study²⁵ indicated lack of consensus among 7 pathologists when assessing pituitary glands from aged horses with mild clinical signs of PPID. To date, there is a lack of a concise definition of histologic findings for diagnosis of PPID, no data on pathologic changes that accompany disease progression, and no reports of potential seasonal changes in the appearance of the pituitary gland in horses. An age-related increase in pars intermedia lesions has been reported in nonpregnant mares and mares overall as well as in geldings, but not in lactating or pregnant mares.²⁶ Because histologic changes in clinically normal aged horses may overlap with those seen in the early stages of PPID, a definitive statement about the relevance of pars intermedia hyperplasia may be difficult to support. Interpretation of histologic lesions is confounded by the histologic sections being 2-dimensional and the lack of accurate assessment of the quantity of hyperplastic tissue. In the present study, 1 pathologist evaluated sections from all horses and another pathologist also reviewed histologic sections from 14 horses. There was disagreement on the presence of PH in 2 CN-PH and 2 EQ-PH horses, which was not unexpected.²⁵ The ACTH concentration at 30 minutes after administration of TRH was > 35 pg/mL in one of those CN-PH horses (compatible with PPID), and baseline ACTH concentrations were within reference range in both CN-PH horses. The ACTH concentrations were high in 1 EQ-PH horse at baseline and in both horses at 30 minutes after administration of TRH, which was compatible with PPID.

Histologic evidence of PH did not appear to be associated with season because there was no association between season and postmortem classification. The only horse with abnormal results of the TRH test that had no evidence of pars intermedia hyperplasia was tested at the end of July and was located in Prince Edward Island where daylight duration is longer than in Chester County. That horse's anomalous response to administration of TRH occurred despite having a baseline ACTH concentration within reference range, and the reason for the unusual response was unclear. Unfortunately, the test could not be repeated to verify whether it was a true response or a spurious result. The ACTH concentrations in Soay rams are highest in June and July,²⁷ but the seasonal pattern for horses is unknown; the only report⁸ on seasonal effect was limited to comparing baseline ACTH concentrations in September with those in May and January. Although all but 1 of those horses were clinically normal, 9 of 10 horses and 6 of 28 ponies were 13 years of age or older, suggesting potential underlying PH. Following publication of that study, one of the horses subsequently developed clinical signs of PPID and another was confirmed as having pars intermedia hyperplasia. Response of ACTH secretion to different hormones or dexamethasone was reported to be affected by season in rams, with an increased pituitary response in summer and autumn.²⁷ Whether a seasonal effect on endocrine responses would be enhanced in areas with extreme seasonal changes in duration of daylight is speculative. Although daylight was significantly associated with results of the TRH test in the present study, a definitive statement about seasonal variation in ACTH secretion or response to dexamethasone or TRH administration cannot be made. The study was not originally designed to investigate a seasonal effect; not every month was represented, and groups of horses were not evenly distributed among seasons. Because duration of daylight or its waxing and waning, and not month per se, is most likely to be the important factor that affects the hypothalamic-pituitary axis, the authors suggest use of daylight hours and not season in future studies because changes in daylight hours within a season can be quite large.

The TRH test appeared to have high sensitivity and specificity, but its specificity was less than that of the DST. Dexamethasone suppression test responses were not reliably associated with PPID or PH, although an abnormal response was detected only in horses with PPID or histologic evidence of PH. The horse with an abnormal test result in September had a normal response in July, but normal DST responses were also seen in September in PPID-0PM and EQ-PH horses. The influence of season still requires further study. Similar to an earlier report^b of inconsistent responses when DSTs were repeated in a small number of clinically abnormal horses, in the present study, results were inconsistent and some abnormal horses repeatedly had test results within reference range. In previous reports,^10,13 values for the sensitivity and specificity have ranged from above to below those of the present study.

There was no correlation among resting ACTH concentrations, response to TRH, and the DST response. The reason for the discrepancy between the DST and TRH test results was unclear. Exposure of pituitary explants from clinically normal horses to TRH¹⁵ elicits ACTH release only from the pars distalis, not the pars intermedia, but to date, no studies have been published on the response of explants from horses with PPID. Although TRH receptor mRNA has been identified in equine pars intermedia, protein expression of the receptor has not yet been detected and information is lacking about its function in the pituitary gland of horses with PPID.¹⁵ A normal response to the DST with a decrease in circulating cortisol concentration suggests that ACTH is being secreted from the pars distalis, where it is under glucocorticoid feedback regulation, unlike secretion from the pars intermedia. An abnormal result of the DST in horses with PPID suggests that the pars intermedia is the major source of ACTH. However, in rats, glucocorticoid receptors can be induced in the pars intermedia following removal of neural influences.²⁸ Although it is unknown whether this occurs in horses, there is evidence supporting the loss of hypothalamic dopaminergic neurons and nerve terminals in horses with PPID.^1,29

Immunohistochemical studies have detected staining for ACTH in pars intermedia of clinically normal horses and in pituitary adenomas, with more intense staining in the latter.^30-33 The explanation for a normal plasma cortisol concentrations in horses with simultaneously increased ACTH concentrations remains unclear. One possibility for the existence of a normal baseline cortisol concentration and also normal response to the DST in horses with high baseline ACTH concentrations is that the circulating ACTH is not bioactive. This could also explain why some horses with PH or adenoma do not have adrenal hyperplasia. Current immunoassays for hormones depend on recognition of antigenic structure, not biological function. Experiments in humans, dogs, and sheep reveal that the basal secretion of immunoreactive ACTH is associated with little ACTH bioactivity but that various stressors increase immunoreactive ACTH and the proportion of bioactive ACTH.^34-37 Detection of biological function requires specific bioassays and is not practical for clinical diagnostic use. However, there is evidence that the ratio of immunoreactive ACTH to bioactive ACTH is higher in pars intermedia adenomas in horses with severe PPID, compared with pars intermedia in clinically normal horses.³¹ Another study^j that used the same assay as the present study revealed that the ratio of biological to immunoreactive ACTH of plasma in horses with PPID was much less than in clinically normal horses and concluded that the tumorous pars intermedia of the pituitary gland secreted substantially more biologically inactive ACTH than the normal pituitary gland. However, this does not explain the basis for adrenal hyperplasia in horses with cortisol concentrations within reference range.

Because all horses older than 13 years that were necropsied had pars intermedia hyperplasia in this study, it was not possible to determine the effect of age on the ACTH response to administration of TRH. A study³⁸ comparing healthy horses 8 to 12 years of age with those > 20 years of age detected no difference in ACTH concentrations between the 2 groups; however, no horses were necropsied, and some could have had subclinical PPID.

The present study was not designed to determine whether abnormal results of the TRH test are correlated with future onset of clinical signs, although it is of interest that one of the horses initially chosen as a clinically normal subject had abnormal results of the TRH test and developed signs of PPID during the next 6 months. When performed shortly before horses were necropsied, the TRH test did not appear highly sensitive in predicting the presence of PH in clinically normal horses. Whether this is because of the uncertainty regarding histologic diagnosis, because the histologic changes are biologically unimportant, or because activity or numbers of TRH receptors on melanotropes or pro-opiomelanocortin processing are variable in hyperplastic pars intermedia is unknown.

A 1-mg dose of TRH administered IV appeared safe, although 6 horses had bouts of trembling after administration. When the test was repeated in several of these horses, they did not have muscle trembling. It was common for horses to transiently yawn, smack their lips, or display flehmen immediately after administration of TRH. These signs have not been reported previously, but it is unclear whether such transient signs could have occurred but were not witnessed. The present study used the same chemical grade of TRH that has been used in other published studies, although other sources are available. In all horses, as in previous studies, a 1-mg dose of TRH was administered regardless of body weight. However, the authors do not believe that the different dose affected results. By inspection of individual responses to TRH, there was no evidence to suggest an influence of weight or BCS or a difference between horses and ponies.

The reason for the apparent high prevalence of PH in this population was unknown. Such hyperplasia may not be pathologic. Until horses are evaluated and retested over a number of years, the clinical importance of abnormal endocrine tests in clinically normal horses will not be known. Oxidative stress has been suggested to have a potential role in loss of dopaminergic neurons, leading to decreased amounts of dopamine in the pars intermedia and altered metabolism of pro-opiomelanocortin,¹ but a more recent study³⁹ revealed no evidence of oxidative stress in affected horses, although manganese superoxide dismutase activity in the pars intermedia decreased with age. Pesticides and chemicals, lack of antioxidants, or underlying disease leading to increased concentrations of free radicals can lead to oxidative stress, and ergot alkaloids can affect pituitary hormone secretion. None of these was an apparent contributing factor in this study.

The sensitivity and specificity of the baseline ACTH concentration indicated that a consistently high endogenous ACTH concentration is sufficient for diagnosis of PPID. All horses with histologically normal pituitary glands had low baseline ACTH concentrations (≤ 35 pg/mL), but concentrations within reference range were also seen in some horses with PPID, PH, or both. In clinically normal horses in this study, there was little variation in ACTH concentrations in blood samples taken just 5 minutes apart, but variation can occur when concentrations are high, making single samples inaccurate for monitoring the clinical status of a horse with suspected or probable PPID or for evaluating seasonal changes. At least 2 baseline samples should be assayed. If high ACTH concentrations are detected in autumn, several additional samples should be evaluated to determine whether the high values are consistent or attributable solely to seasonal variation. On the basis of current data, ACTH concentrations are unlikely to exceed 250 pg/mL in a clinically normal pony or 150 pg/mL in a clinically normal horse in September.⁸ Concentrations of ACTH consistently exceeding these values in September suggest PPID, but horses and ponies with high concentrations that are less than the mentioned values should be tested in another season, unless clinical signs suffice for diagnosis. Normal baseline concentrations of ACTH may be detected in horses with pars intermedia hyperplasia, and in these horses, evaluation of the ACTH response at 10 and 30 minutes after TRH administration is clinically useful for diagnosis. Values exceeding 100 and 35 pg/mL at 10 and 30 minutes after TRH administration, respectively, indicate that PPID is highly likely. No CN-0PH horse's values exceeded those values except 1 horse with 44 pg/mL at 30 minutes. Only 3 of 30 tests in PPID-0PM, PPID-PH, and EQ-PH horses resulted in values less than those values, and repeated tests in the 3 horses revealed abnormal results (2 in June and 1 in late autumn). Limiting numbers of samples to 2 at baseline and 1 each at 10 and 30 minutes after TRH administration makes this test clinically useful in an ambulatory practice.