Autumn ACTH in the horse: what a seasonal reference range really tells you

Why the value rises in autumn in every horse

Adrenocorticotropic hormone, or ACTH, comes from the pituitary gland. In the horse its concentration in blood follows an annual rhythm. A British analysis compared, in 2012, 156 clinic horses with no sign of pituitary disease against 941 animals presumed affected. Both groups showed the same pattern, with a maximum from August to October. Autumn therefore does not lift only the values of affected animals, it lifts them all.

That governs how a laboratory report is read. If the rise affects healthy and affected horses alike, a high autumn value is not on its own a sign of disease. What changes is the gap between the two groups, widest from August to October. Autumn is therefore the season in which the distinction succeeds best, and the season in which an unchanged limit would misclassify most healthy horses.

What drives this rhythm?

Day length. An Australian study based on monthly sampling over a year stresses that the rise is not confined to the autumn months, but follows the photoperiod. A laboratory analysis across ten breeds places the trough in April and the peak in September.

The same analysis sets out the timing. The rise appears to begin in late April or early May, as the days lengthen, and falls back quickly after the equinox in late September, once the day drops below twelve hours. Measuring in late September or in October means measuring on a very steep part of the curve. The word autumn is shorthand: the real quantity is day length, not the name of the month.

Where the limits of 29 and 47 come from

The two figures that appear most often in reports and guides have a single source: this British study of 2012. It set the upper reference limit at 29 pg/ml from November to July and at 47 pg/ml from August to October. Its authors concluded that ACTH remains usable all year, provided the interval belonging to the period is used.

The origin of these figures deserves a look. The comparison group was made up of clinic patients admitted for other reasons, not a random sample of healthy horses. And the classification of cases rested on clinical suspicion together with a value above the limit for that period. The limit was therefore justified partly by itself. This does not devalue the study, which remains the primary source of both figures, but it explains the different limits obtained since.

Other countries have calculated their own intervals, and they differ markedly. A South African study drew blood every month for a year from 80 clinically healthy horses under twelve years old. The upper limit was 21.4 pg/ml in the local early summer and 60.6 pg/ml in the local autumn, almost three times as high in the same population and the same laboratory. Their conclusion deserves attention: the annual rhythm has the shape seen elsewhere, only the height of the limits differs.

Why two laboratories return different figures

A measured value is not a fixed property of the horse. In 2013 a German study sent four laboratories, for fourteen horses, two samples drawn at the same moment and labelled differently. Within one laboratory the two values from one horse differed by about 19 pg/ml on average; in one case 121 pg/ml against under 5. Between laboratories the mean difference reached about 26 pg/ml, and a twenty-year-old gelding was reported at 60.9 or at 108 pg/ml depending on the laboratory. Centrifuged at once, the sample gave 11.6 pg/ml more on average.

The study dates from 2013, and today's analysers are no longer the same. Its message survives all the same, because it compared laboratories, not horses. A methodological study from 2026 follows on: the published thresholds hold almost exclusively for a single analyser. Between two analysers from the same manufacturer a threshold transferred reliably; between manufacturers it did not, because the deviation depended on the concentration.

Can I compare two laboratory reports directly?

Only with reservations. As long as both name the same method and the same reference interval, the comparison is meaningful. If they come from different laboratories or analysers, a difference may arise from the measurement alone. The interval supplied with the value belongs to it.

The autumn value does not measure ACTH alone

A British study had the same pony samples assayed by two widespread methods. In autumn the two differed by 38.6 pg/ml on average, the following spring by only 5.1. A gap about seven times larger in autumn does not argue for chance, but for the idea that something else is measured alongside in that season.

The trail leads to a protein fragment abbreviated CLIP, which arises in the same part of the pituitary gland as ACTH and corresponds to its terminal portion. A laboratory study from 2023 showed that CLIP binds to the same capture antibody as ACTH, unlike a second related fragment. Endogenous CLIP was detected in the plasma of a pony sampled in autumn. The study establishes the binding; its share of the autumn rise is expressly left open.

What this means in the yard is shown by the same series in ponies without laminitis. Of 88 animals sampled in two consecutive seasons, 56 were above the 47 pg/ml limit in autumn. Of those, 39 were back below the 29 pg/ml limit the following spring, with no treatment at all. Assay interference, a genuine seasonal swing or both: the question is unsettled.

How well the value separates affected from healthy horses

On diagnostic accuracy, a summary analysis brings together eleven studies. Across all thresholds the basal value picked up about seven affected horses in ten and correctly classified nine unaffected in ten. At the commonly used threshold of 35 pg/ml the detection of affected animals came out lower still. The authors expressly advise against the test as screening and in a horse without clinical signs.

Their own worked example shows why: the same test makes different errors in different horses, and the probability before testing decides which error prevails.

An independent systematic review arrives at a statement that is easy to remember: the value rules out better than it rules in. Mean detection of affected animals there was about 76 per cent, and of healthy animals correctly classified about 95 per cent. The authors declined to summarise in a single figure, given the strong differences between studies and the bias they found.

The British professional body published a guideline in 2024, drawn up under a formal appraisal procedure. It puts the overall accuracy of the basal value at 88 to 92 per cent in autumn and at 70 to 86 per cent outside autumn, depending on the probability before testing. And it draws an unusual conclusion: with no workable thresholds that allow for every influence, it argues for grey zones rather than a single dividing line.

How likely is the disease to begin with?

That depends above all on age. A systematic review puts the most robust frequency estimate at 21.2 per cent in horses and ponies aged fifteen and over, and at 2.9 per cent across the whole equine population. Under ten years the probability before testing is regarded as low. Advancing age is also the only established risk factor in that review.

Breed, age and feeding move the limit with them

No reference interval holds for every horse. An Australian study of 399 animals without signs of disease measured around both equinoxes. In autumn ponies of other breeds sat about 1.55 times and Shetland ponies about 2.67 times higher than Thoroughbreds. In spring these groups did not differ. The breed difference is therefore seasonal, not permanent.

A laboratory analysis across ten breeds adds that the window differs as well. Arabians and donkeys were higher mainly from May to November, Shetland and Welsh breeds only from July to November. Breed and month cannot therefore be judged separately, only together.

Age works in two directions at once. An American study at nine sites, with monthly sampling, found higher values all year in horses over fifteen and a more pronounced autumn rise as well. Age is therefore not only a risk factor for the disease: it already shifts the value in apparently healthy animals.

Feeding can move the value too. In one trial sixteen healthy horses received four concentrate feeds of comparable energy value in changing order, with measurements in March, May, August and October. In October the value reached 57.7 pg/ml against 13.2 in March. Older horses on the starch-rich feed reached 60.0 pg/ml, similarly fed younger ones 15.7. The authors see here a risk of wrongly classifying a horse fed on plenty of starch. Sixteen animals at one site: a pointer, not a rule.

Testing in autumn or not: the contradiction resolves

One side recommends autumn, because the gap between affected and healthy animals is then widest and the accuracy highest. The other advises against autumn, because the values swing most. Both rest on real observations and describe the same curve from two ends.

The swing has been measured directly. An Australian study repeated a stimulation test in the same horses two weeks apart. In winter it found no difference between the two appointments and very good repeatability. In autumn the values came out markedly lower at the second appointment, the variation within one horse was greater, and misclassifications doubled. The season that separates best is therefore the least stable. Twenty horses at one site: a signal, not a basis for decisions.

The resolution lies not in the date, but in the limit the value is read against. A prospective study of 106 horses determined two kinds of limit for each month. A diagnostic threshold raised detection of affected animals from about six to nine in ten, and so catches mild cases. A reference interval raised the share of healthy animals correctly classified from 85 to 98 per cent, avoiding needless treatment. Which limit is right depends on which error weighs more. That decision is clinical, not arithmetical.

A laboratory database of more than 75,000 horses reaches the same result. Differences from week to week stay small for most of the year and become material only from mid-June to early December, with the largest swing in late September and early October. Measuring then calls for a narrow seasonal reference, not a coarse one.

What a high value says about laminitis risk

Many owners read ACTH above all as a laminitis warning. The data support that link only in part. The British guideline notes that equids with pituitary disease and raised insulin at the same time appear more at risk, but that ACTH is not an independent predictor of laminitis risk.

A cohort of 374 ponies with no previous laminitis, examined twice a year for up to four years, confirms this. What remained in the joint model were fasting insulin, insulin one hour after a sugar dose, the fat tissue hormone adiponectin, and diverging hoof growth. ACTH was not independently associated with laminitis. In the group with the highest insulin values, about one pony in ten, an estimated 69 per cent developed the disease over four years, against 6 per cent in the lowest group.

From Switzerland comes a small study of 51 clinically healthy elderly horses, which looked at what accompanied radiographic changes of chronic laminitis. Only age did: horses aged 26 to 32 showed them more often than those aged 15 to 25. A value above the seasonally adjusted interval was not associated, nor were body condition or the cresty neck score. Fifty-one horses, a single sampling point: the study shows an absent link, it does not prove an absence.

Published upper reference limits for basal ACTH by region, comparison group and season

Original analysis

Limitations and uncertainty

• The editorial requirement to base at least half the sources at the level of systematic review, meta-analysis or randomised trial cannot be met in this field: only four of the studies used here reach that level, the rest being guidelines, cohorts and case-control, cross-sectional and laboratory studies.
• The two most quoted limits, 29 and 47 pg/ml, come from a single study of 2012. Its comparison group consisted of clinic patients, and the disease cases were defined in part by an ACTH value above that same limit, which anchors the estimate in itself.
• For this disease there is no independent reference standard. Several of the studies analysed rest on composite case definitions that contain the test under evaluation, and the summary analysis names exactly this point, along with the case-control design, as reasons for artificially good accuracy.
• The published thresholds hold almost exclusively for a single analyser. Between analysers from different manufacturers the transfer is not linear, so a limit read in the literature does not simply apply to any given laboratory report.
• No seasonal reference intervals of its own exist for Switzerland. The one Swiss study found here used an existing, seasonally adjusted interval; it did not calculate one. The limits quoted come from the United Kingdom, South Africa and Australia.
• The study on scatter between laboratories dates from 2013 and describes a generation of analysers that is now superseded. It shows that differences between laboratories can exist, not how large they are in a Swiss laboratory today.

Open questions

• How much of the measured autumn rise is due to the CLIP fragment binding to the capture antibody, and how much to genuinely higher ACTH release?
• Which seasonal reference intervals apply at the latitude of Switzerland and to the breeds common here, in particular native ponies and donkeys?
• Do breed-specific and month-specific limits change the course of affected horses, or do they only shift the classification on the laboratory report?
• Why is the test least repeatable in exactly the season in which it separates best?

Frequently asked questions

My horse has an autumn ACTH value of 60. Is that bad?

No diagnosis follows from that figure alone. Whether 60 pg/ml sits above or below the limit depends on the period, the laboratory and the analyser: in British data the autumn limit is 47 pg/ml, in a South African population 60.6, at one Australian site 94. Breed, age and feeding come on top, and they shift the value in healthy animals. The value only becomes informative together with the reference interval of the measuring laboratory and the clinical signs recorded by the veterinary practice.

Should I have it measured again in spring?

That is a question for the practice, not one this text answers. What the data say: in a British series 56 of 88 ponies were above the 47 pg/ml limit in autumn, and 39 of those 56 were back below the spring limit of 29 the following spring, with no treatment at all. A single high autumn value therefore remains provisional. Conversely, values swing markedly less in winter, which makes a repeat in that season easier to compare.

Why does my report show a different reference interval from the one I found online?

Because reference intervals are tied to a laboratory. They are calculated from a particular comparison group, with a particular assay method and for a particular period. A methodological study from 2026 notes that the published thresholds hold almost exclusively for a single analyser and do not simply transfer between manufacturers. The interval on your report belongs to your value. A figure from a guide belongs to another laboratory and is no yardstick for your result.

My pony is only twelve. Can the value still be raised?

Yes, and on its own that says little. In equids under ten the probability of this disease is regarded as low, and the lower it is, the more of the abnormal values are falsely positive: in the worked example of the summary analysis, with two affected animals in a hundred horses, about 127 falsely positive results fall on 1,000 animals tested. At the same time, Shetland ponies sat about 2.7 times higher than Thoroughbreds in autumn, while in spring they were level. Breed and season therefore explain a considerable part of such findings.

Does a high ACTH value mean my horse will get laminitis?

No, at least not on its own. The British guideline notes that ACTH is not an independent predictor of laminitis risk. In a cohort of 374 ponies with no previous laminitis, what remained in the joint model were fasting insulin, insulin after a sugar dose, the fat tissue hormone adiponectin and diverging hoof growth, but not ACTH. A Swiss study of 51 elderly horses likewise found no link between a value above the seasonally adjusted interval and radiographic laminitis changes; only age was associated there.

Is the stimulation test more reliable than a simple blood sample?

It performs better in several studies, but not throughout. In a prospective study of 106 horses the overall accuracy of the stimulation test exceeded that of the basal value. A twelve-month study of 63 horses qualifies this: detection of affected animals was not meaningfully better all year, and the gain concerned mainly animals without clear clinical signs. This test also follows the annual rhythm, with the lowest values from February to May and the highest from August to October.

Sources

1. Menzies-Gow NJ, Banse HE, Duff A, Hart N. BEVA primary care clinical guidelines: Diagnosis and management of equine pituitary pars intermedia dysfunction. Equine Veterinary Journal, 2024 (Professional guideline | Horse)DOI 10.1111/evj.14009
   Overall accuracy of the basal ACTH value lies, depending on the probability before testing, between 88 and 92 per cent in autumn and between 70 and 86 per cent outside autumn; because thresholds allowing for every influence cannot be set workably, the guideline recommends grey zones, and it notes that ACTH is not an independent predictor of laminitis.
2. Meyer JC, Hunyadi LM, Ordóñez-Mena JM. The accuracy of ACTH as a biomarker for pituitary pars intermedia dysfunction in horses: A systematic review and meta-analysis. Equine Veterinary Journal, 2022 (Meta-analysis | Horse)DOI 10.1111/evj.13500
   Eleven primary studies yielded a pooled detection rate for affected animals of 0.72 and a rate of correctly classified healthy animals of 0.88; the authors recommend the basal value as a confirmatory test where the probability before testing is high, expressly not as screening and not in horses without clinical signs.
3. Tatum RC, McGowan CM, Ireland JL. Evaluation of the sensitivity and specificity of basal plasma adrenocorticotrophic hormone concentration for diagnosing pituitary pars intermedia dysfunction in horses: A systematic review. The Veterinary Journal, 2021 (Systematic review | Horse)DOI 10.1016/j.tvjl.2021.105695
   With mean detection of affected animals at 75.5 per cent and a mean rate of correctly classified healthy animals of 95.2 per cent, the basal value rules the disease out better than it rules it in; the authors declined a meta-analysis because of the bias found and the strong differences between studies.
4. Ireland JL, McGowan CM. Epidemiology of pituitary pars intermedia dysfunction: A systematic literature review of clinical presentation, disease prevalence and risk factors. The Veterinary Journal, 2018 (Systematic review | Horse)DOI 10.1016/j.tvjl.2018.03.002
   Across 29 included publications the most robust frequency estimates are 21.2 per cent in equids aged fifteen and over and 2.9 per cent in the overall population, and the only established risk factor remains advancing age, while statements on breed and sex remain contradictory.
5. Jacob SI, Geor RJ, Weber PSD, Harris PA. Effect of dietary carbohydrates and time of year on ACTH and cortisol concentrations in adult and aged horses. Domestic Animal Endocrinology, 2018 (Randomised trial | Horse)DOI 10.1016/j.domaniend.2017.10.005
   In sixteen healthy horses the basal value reached 57.7 in October against 13.2 pg/ml in March, and older animals on starch-rich feed reached 60.0 against 15.7 pg/ml in similarly fed younger ones: the authors treat feeding as a possible confounder able to lead to wrongly classifying a horse as affected.
6. Durham AE, Clarke BR, Potier JFN, Hammarstrand R. Clinically and temporally specific diagnostic thresholds for plasma ACTH in the horse. Equine Veterinary Journal, 2021 (Cross-sectional study | Horse)DOI 10.1111/evj.13292
   In a laboratory database of 75,892 horses, differences from week to week stay small for most of the year and become material from mid-June to early December, with the maximum in late September and early October; the authors call for period-specific thresholds and a choice of limit governed by the less desirable error in the individual case.
7. Durham AE, Potier JF, Huber L. The effect of month and breed on plasma adrenocorticotropic hormone concentrations in equids. The Veterinary Journal, 2022 (Cross-sectional study | Horse)DOI 10.1016/j.tvjl.2022.105857
   Across ten breeds an annual rhythm emerges with a trough in April and a peak in September, Arabians and donkeys being raised mainly from May to November and Shetland and Welsh breeds only from July to November: breed and month must be judged together.
8. Bamford NJ, Stewart AJ, El-Hage CM, Bertin FR. Investigation of breed differences in plasma adrenocorticotropic hormone concentrations among healthy horses and ponies. The Veterinary Journal, 2023 (Cross-sectional study | Horse)DOI 10.1016/j.tvjl.2023.105995
   In 399 animals without signs of disease, ponies of other breeds sat 1.55 times and Shetland ponies 2.67 times higher than Thoroughbreds in autumn, while in spring there was no difference: the breed effect is seasonal, and autumn reference intervals must take breed into account.
9. Copas VEN, Durham AE. Circannual variation in plasma adrenocorticotropic hormone concentrations in the UK in normal horses and ponies, and those with pituitary pars intermedia dysfunction. Equine Veterinary Journal, 2012 (Cross-sectional study | Horse)DOI 10.1111/j.2042-3306.2011.00444.x
   Primary source of the limits of 29 pg/ml from November to July and 47 pg/ml from August to October: the annual variation exists in both groups, with the highest values from August to October, and ACTH remains usable all year provided the reference interval matching the period is used.
10. Fisher D, Schliewert EC, Hooijberg EH. Temporally specific adrenocorticotropic hormone reference intervals for horses in South Africa. Journal of the South African Veterinary Association, 2022 (Cohort study | Horse)DOI 10.36303/JSAVA.001
    In 80 clinically healthy horses under twelve years old with monthly sampling over a year, the upper reference limits were 21.4 pg/ml in early summer, 60.6 in autumn and 22.3 in winter: the shape of the annual rhythm matches that seen elsewhere, the height of the limits is specific to place and laboratory.
11. Secombe CJ, Tan RHH, Perara DI, Byrne DP. The Effect of Geographic Location on Circannual Adrenocorticotropic Hormone Plasma Concentrations in Horses in Australia. Journal of Veterinary Internal Medicine, 2017 (Cohort study | Horse)DOI 10.1111/jvim.14782
    In 81 clinically unremarkable horses at two sites, with monthly sampling over twelve months, the rise is not confined to the autumn months but is linked to the change in photoperiod; the upper limits differed markedly between Perth and Townsville and must be determined site by site.
12. Pinn-Woodcock TL, Llanos-Soto SG, Ivanek R, Goodrich E. Seasonal elevation in equine adrenocorticotropic hormone occurs throughout the contiguous United States and is influenced by latitude and age. Journal of the American Veterinary Medical Association, 2025 (Cohort study | Horse)DOI 10.2460/javma.25.04.0223
    In 201 apparently healthy horses at nine sites, animals over fifteen sit higher all year and additionally show a more pronounced autumn rise, which is why season and age must enter together into the reading of a basal value.
13. Knowles EJ, Moreton-Clack MC, Shaw S, Harris PA. Plasma adrenocorticotropic hormone (ACTH) concentrations in ponies measured by two different assays suggests seasonal cross-reactivity or interference. Equine Veterinary Journal, 2018 (Cohort study | Horse)DOI 10.1111/evj.12797
    Two assay methods differed by 38.6 pg/ml on average in autumn and by only 5.1 in spring; of 88 ponies sampled in both seasons, 56 were above 47 pg/ml in autumn, of which 39 were back below 29 pg/ml the following spring without treatment, which the authors attribute to an altered pituitary peptide pattern in autumn.
14. Knowles EJ, Hyde C, Harris PA, Elliott J. Short Communication: Identification of equine corticotropin-like intermediate lobe peptide (CLIP) binding to an adrenocortipcotrophic hormone (ACTH) assay capture antibody. Domestic Animal Endocrinology, 2023 (Laboratory study | Horse)DOI 10.1016/j.domaniend.2023.106785
    The CLIP fragment binds to the same capture antibody as ACTH, while a second related fragment does not, and endogenous CLIP was detected in the plasma of a pony sampled in autumn: the interference is established, its share of the measured rise is not quantified.
15. Gehlen H, Bradaric Z. [Study on the reproducibility of ACTH concentrations in plasma of horses with and without equine Cushing syndrome]. Berliner und Münchener Tierärztliche Wochenschrift, 2013 (Laboratory study | Horse)PMID 23901591
    In fourteen horses and 132 samples spread across four laboratories, duplicate samples drawn at the same moment differed by about 19 pg/ml on average within one laboratory and by about 26 pg/ml between laboratories, and immediate centrifugation raised the value by 11.6 pg/ml on average.
16. Durham AE, Lopes A. Transference of diagnostic thresholds for equine plasma ACTH from the Immulite 1000 to Immulite 2000XPi and Tosoh AIA systems. Domestic Animal Endocrinology, 2026 (Laboratory study | Horse)DOI 10.1016/j.domaniend.2026.107015
    The published thresholds hold almost exclusively for a single analyser; between two analysers from the same manufacturer the transfer was linear, whereas between manufacturers it was concentration-dependent and therefore possible only through a conversion of its own for each threshold.
17. Horn R, Stewart AJ, Jackson KV, Dryburgh EL. Clinical implications of using adrenocorticotropic hormone diagnostic cutoffs or reference intervals to diagnose pituitary pars intermedia dysfunction in mature horses. Journal of Veterinary Internal Medicine, 2021 (Case-control study | Horse)DOI 10.1111/jvim.16017
    In 106 horses tested monthly, a diagnostic threshold raised detection of affected animals from 0.61 to 0.87, whereas a reference interval raised the share of correctly classified healthy animals from 0.85 to 0.98: the choice of limit is a clinical decision about which error weighs more.
18. Adams AA, Siard-Altman MH, Reedy SE, Barker D. Evaluation of seasonal influences on adrenocorticotropic hormone response to the thyrotropin-releasing hormone stimulation test and its accuracy for diagnosis of pituitary pars intermedia dysfunction. The Veterinary Journal, 2023 (Cohort study | Horse)DOI 10.1016/j.tvjl.2023.106035
    Over twelve months in 63 horses, values after stimulation were lowest from February to May and highest from August to October; accuracy exceeded that of the basal value, but detection of affected animals was not meaningfully better all year, and the gain concerned mainly animals without clear clinical signs.
19. Kam YN, McKenzie K, Coyle M, Bertin FR. Repeatability of a thyrotropin-releasing hormone stimulation test for diagnosis of pituitary pars intermedia dysfunction in mature horses. Journal of Veterinary Internal Medicine, 2021 (Case-control study | Horse)DOI 10.1111/jvim.16281
    In twenty horses tested twice two weeks apart, the stimulation test was very well repeatable in winter, whereas in autumn the values were lower at the second appointment, the scatter within one horse was greater, and twice as many misclassifications occurred.
20. Christen G, Precht C, van der Kolk J, Fouché N. Age over 25 years, but not plasma adrenocorticotropic hormone concentration above the seasonally adjusted reference range is predictive for radiographically assessed changes of chronic laminitis in elderly horses. Schweizer Archiv für Tierheilkunde, 2020 (Cross-sectional study | Horse)DOI 10.17236/sat00283
    In 51 clinically healthy elderly horses only age was associated with radiographic changes of chronic laminitis, whereas an ACTH value above the seasonally adjusted reference range, body condition and the cresty neck score showed no association.
21. Knowles EJ, Elliott J, Harris PA, Chang YM. Predictors of laminitis development in a cohort of nonlaminitic ponies. Equine Veterinary Journal, 2023 (Cohort study | Horse)DOI 10.1111/evj.13572
    In a cohort of 374 ponies with no previous laminitis followed over 891 pony-years, ACTH was not independently associated with laminitis; what remained in the joint model were fasting insulin, insulin after a sugar dose, adiponectin and diverging hoof growth, with estimated four-year frequencies of 6, 22 and 69 per cent across the three insulin groups.