Steroids and heart failure in cats: science, or an old wives’ tale?

1Where it started

In 2004 a group from Minnesota described twelve cats that developed congestive heart failure after a corticosteroid, mostly methylprednisolone acetate, between one and nineteen days after dosing.¹ It is a careful and honest paper, and it did what a case series is supposed to do, which is to flag something worth investigating.

The authors offered two arguments for a real drug effect. The first was that these cats lived longer than would be expected for heart failure caused by hypertrophic cardiomyopathy, so that their disease was probably not ordinary cardiomyopathy. The second was that the temporal association was too strong to ignore. Both arguments are worth taking seriously, and neither of them holds up on examination.

2“These cats lived longer than expected”

Five of the twelve died, four of them within days of presenting, so the survival argument rests on the seven that recovered.

The difficulty is that recovery from heart failure with reversal of wall thickening is not evidence of a drug effect. It is instead the description of transient myocardial thickening, a condition that was characterised properly in 2018.⁷ In that study 21 affected cats were compared with 21 with hypertrophic cardiomyopathy. Wall thickness fell from 6.8 to 4.8 mm over about three months, cardiac treatment was stopped in 20 of 21, and all of them survived. Troponin normalised as the echo improved, which points at a myocardium that was injured and recovered.^{7, 15}

The part of that study which matters most for the steroid question is the exposure history. Corticosteroid exposure was present in 3 of the 21 affected cats and in 2 of the 21 cats with cardiomyopathy. These were drugs given during the antecedent event, listed alongside antibiotics in 12, NSAIDs in 7 and opioids in 5.⁷ If steroids were driving the thickening, that comparison is precisely where a difference should have appeared.

The 2025 series describing a reversible restrictive form found the same pattern, reporting an antecedent event in 11 of 17 affected cats and in none of the 16 controls, with a glucocorticoid injection accounting for exactly one of those eleven events.⁸ Neutering accounted for five of them.

A larger series in 2025 applied explicit criteria to 145 cats with a hypertrophic phenotype and found transient thickening in 16 of the 53 that had serial echocardiograms, with maximum wall thickness falling from 8.3 ± 1.7 to 5.4 ± 0.9 mm over a median of 3.5 months and every affected cat surviving.²⁸ Reversible thickening is therefore not rare, and it is not confined to cats that had received a drug.

The point here is not that the 2004 cats were free of heart disease. The point is that nobody had looked before the steroid was given, so the series cannot separate a drug that caused the thickening from a drug that was given to a cat already carrying it.

3“The temporal association is too strong to ignore”

The objection to this argument is not an objection to the timing. A depot preparation such as methylprednisolone acetate releases drug for weeks, and glucocorticoid effects that require gene transcription do not track plasma concentration in any simple way, so an interval of days between injection and decompensation is entirely compatible with a drug effect. Pharmacokinetic reasoning does not settle this question in either direction.

The objection is instead to what a series of this design can establish. The series is retrospective, the cases were identified precisely because they had both the exposure and the outcome, there are no pre-treatment echocardiograms, and there is no denominator, so nobody knows how many cats in the same practice received steroids over the same period and remained well. A temporal association assembled in that way is a signal worth investigating, but it cannot distinguish a drug effect from the coincidence of a common drug and an undiagnosed cardiomyopathy, which is what a population with a 10–15% prevalence of subclinical disease guarantees will happen at some rate.

4The proposed mechanisms

Four mechanisms have been offered for the claimed effect. Each of them is plausible in isolation, and each of them fails when it is set against the available data.

5What happens when you give cats steroids and watch

Three groups have performed the experiment, and they did not all arrive at the same answer. This literature is routinely summarised as uniformly negative, and that summary is inaccurate.

Rishniw. A Cornell group gave 30 mg of methylprednisolone acetate intramuscularly to eight healthy cats and echoed them at one, two, three and six weeks.³ No ventricular measurement differed at any point, no cat showed evidence of volume overload, and left atrial dimensions in fact fell. Their conclusion was that the results do not support high-dose methylprednisolone causing appreciable plasma volume change. This is a conference abstract of eight animals and has never appeared as a full paper.

Khelik. Ten cats with allergic dermatitis received prednisolone at 1 and 2 mg/kg daily for seven days each in a crossover, against ten matched untreated controls, with blood pressure, glucose, potassium, urinary sodium excretion, troponin, NT-proBNP and a full echocardiogram measured throughout.⁴ Nothing differed between the treated cats and the controls. Only glucose and NT-proBNP moved at the higher dose, and both remained within the reference range. This is the best-controlled study in the set.

Block and Oyama, and this one is positive. Ten client-owned cats starting at least a five-day oral course for a clinical indication were echoed before and after.⁵ There were statistically significant increases in left ventricular internal diameter in diastole and in systole, in left atrial dimension, in the left atrial to aortic root ratio, and in left atrial and auricular area. Total solids rose as well. Absolute NT-proBNP did not change (P = .23), and neither did blood pressure, heart rate, glucose or body weight, although six of the ten cats showed an individual NT-proBNP change greater than 60%, and three of the eight that started below 100 pmol/L finished above that value. One cat went from 30 to 414 pmol/L, developed a new murmur and gallop, and acquired a dilated left ventricle. The authors read the pattern as plasma volume expansion.

Two qualifications belong with that result. The first is that the cohort was not a healthy one, since although eight of the ten cats had no echocardiographic evidence of cardiac disease, one had newly diagnosed complete atrioventricular block with four-chamber dilation and one had hypertrophic obstructive cardiomyopathy. The second is that no cat in the study developed congestive heart failure. What the study demonstrates is a short-term loading effect measurable on echocardiography, which is a smaller claim than the one it is usually cited for.

The remaining study counted outcomes rather than measuring them. Dutch and colleagues compared 732 cats given methylprednisolone acetate with 310 that had never had a steroid.⁶ Heart failure occurred in 0.82% of the treated cats and 1.90% of the untreated ones. Diabetes went the same way, 3.83% against 7.10%.

Steroids do not protect cats from heart failure or from diabetes, so that result indicates that the two groups were not comparable with each other. It is worth being clear about what the study does and does not show. It gives no signal of harm at all, and it remains the only study anyone has published with a control group and a real denominator.

Taken together, the experimental literature is mixed rather than negative. Plasma volume expansion was demonstrated in one study and cardiac dimensional changes in a second, nothing was found in a third or a fourth, and no excess of heart failure appeared in the only cohort that included a comparison group. Nothing in this literature has yet shown a cat entering failure because of a steroid, and none of these studies enrolled the cats in which that is supposed to happen.

6The awkward part: steroids are used to treat heart failure

In human medicine, prednisone is given to heart failure patients to improve diuresis. In a randomised, double-blind, placebo-controlled trial of twenty patients with stable heart failure, prednisone increased maximum daily urine output by about 810 mL and daily sodium excretion by roughly 124 mmol, and creatinine fell.¹¹ The approach has been explored specifically in diuretic resistance.¹²

The same drug class is therefore used deliberately to offload fluid in people with failing hearts, while in cats it is held responsible for causing fluid retention severe enough to precipitate failure. Both positions cannot be straightforwardly true, and the human evidence is the better controlled of the two.

7The same question in people

The claim has a human counterpart, and it is worth setting out in full because the human literature does not point consistently in one direction.

The mechanism has been measured, and it was not there. Whitworth gave six normal subjects five days each of prednisolone 40 mg, methylprednisolone 32 mg, triamcinolone 40 mg and dexamethasone 8 mg, at doses matched for glucocorticoid activity.¹⁷ Their finding, verbatim: “There were no MC effects with any of the steroids. Body weight did not increase and urinary sodium excretion increased rather than decreased. Plasma volume (125I human serum albumin) and haematocrit were unchanged.” Plasma volume was measured directly with radiolabelled albumin, using two of the drug classes given to cats, and it did not move.

The volume effect belongs to cortisol rather than to the synthetics. Cortisol at 50 mg six-hourly for five days “caused a marked antinatriuresis and expansion of extracellular fluid and plasma volume”.¹⁸

The effect is not entirely absent, however. Waller measured urine electrolytes during alternate-day methylprednisolone pulses and found suppression of urinary sodium excretion “for up to 32 hr after pulsing”, with a gradual escape during continued treatment.¹⁹ A synthetic glucocorticoid can therefore retain sodium at pulse doses, although it does so briefly and with escape, and that qualification belongs in this appraisal rather than being omitted from it.

The effect has a ceiling in any case. Fludrocortisone at roughly ten times replacement for ten days moved body weight from 72.2 to 74.0 kg,²⁰ and sodium escape occurs by days four to seven.²¹

The association exists in people too, and it is large. A self-controlled case series of 2,623,327 adults given a single oral burst of fourteen days or less found a heart-failure incidence rate ratio of 2.37 (95% CI 2.13–2.63) at days 5 to 30, attenuating over days 31 to 90, at an absolute rate of 1.3 per 1000 person-years.²² The commonest indications were skin disorders and respiratory tract infections, the same clinical picture as the cats. A Danish nationwide case-crossover across 71,821 incident heart-failure cases independently names corticosteroids among the drug classes associated with cardiovascular outcomes.²³

The association behaves like confounding when it is tested directly. CORT-AHF followed 11,356 patients in acute heart failure, 1,594 of them newly started on a systemic corticosteroid: “There were several differences between treated and untreated patients, essentially because corticosteroid-treated patients were sicker. Although unadjusted outcomes were worse in corticosteroid-treated patients … these differences disappeared after adjustment”. The 90-day mortality hazard ratio was 0.91 (0.76–1.10) without COPD, 0.99 (0.78–1.26) with.²⁴

Where allocation was random, no signal appeared at all. ESCAPe randomised 584 patients with severe pneumonia to methylprednisolone or placebo and counted cardiovascular complications under a definition that explicitly included “new congestive heart failure or acute worsening of long-term congestive heart failure”. By day 28 these occurred in 11 of 251 (4%) against 12 of 234 (5%), odds ratio 0.85 (0.37–1.96).²⁵ The trial was underpowered for rare events, but it addresses the question directly.

The obstetric data are the strongest positive signal and the hardest to read. A self-controlled series of 52,119 women given antenatal corticosteroids found a heart-failure incidence rate ratio of 4.45 (2.63–7.55) at days 5 to 60, but also 3.65 (2.14–6.22) at days 61 to 180.²⁶ A rate ratio that remains near four at six months is difficult to attribute to an acute volume effect.

One human case does look like the feline claim. A 53-year-old man with thyrotoxicosis and Graves’ ophthalmopathy, with “no previous history of cardiac disorder or conditions predisposing to cardiac failure other than thyrotoxicosis”, “suddenly developed onset of acute pulmonary edema and hypertension” after the second day of intravenous methylprednisolone 1 g daily.²⁷ The authors describe their own diagnosis as presumptive. This is a single case, at roughly a hundred times a feline anti-inflammatory dose, in an uncontrolled thyrotoxic patient. It is nonetheless the nearest analogue in the human literature, and it belongs in this appraisal for that reason.

What the human evidence settlesThe proposed mechanism is not a class effect of synthetic glucocorticoids, because plasma volume has been measured directly after methylprednisolone and triamcinolone and it did not change. What the human evidence does not settle is whether a susceptible animal can be tipped over by a depot dose. The same temporal association appears in people at enormous scale, survives naive analysis, and dissolves in the one population where the confounding was addressed directly.

8What might actually be happening

Cats do present in heart failure after a visit to the veterinary practice, and that observation is real enough that every cardiologist has encountered it.

The list of antecedent events reported in these studies is worth reading in full, because it includes anaesthesia and surgery, fluid administration, trauma, upper respiratory infection, vaccination, bite wounds and neutering.^{7, 8} A steroid injection is one entry on a long list, and it is not a prominent one. What those events share is not a drug, but rather a sick and frightened cat being handled.

Stress cardiomyopathy is well described in people. Catecholamine surges produce myocardial stunning, with contraction band necrosis on biopsy, in hearts with no coronary disease.¹³ It presents in the manner of a myocardial infarction and it subsequently resolves. A segmental variant exists that appears as focal or patchy wall thickening, which on a feline echocardiogram would be difficult to distinguish from hypertrophic cardiomyopathy.

Stress-associated myocardial dysfunction is one biologically plausible alternative explanation, and it deserves investigation rather than assent. It would fit the transient thickening data, the rise and fall in troponin, the recovery, and the breadth of the reported triggers. It has not been tested in cats, and it is offered here as a hypothesis rather than as a finding that competes with the steroid explanation on equal evidential footing. Both explanations remain untested.

None of this establishes that the steroid is irrelevant in every case. One Maine Coon has been reported whose obstructive phenotype resolved when steroids were withdrawn and returned on re-exposure.¹⁶ The report concerns a single cat, and a single cat generates a hypothesis rather than a finding.

9Where this leaves it

There is no convincing evidence that glucocorticoids cause congestive heart failure in cats, which is a different statement from there being no evidence at all. There is a case series, there is a demonstrated physiological effect, and there is an echocardiographic study showing measurable cardiac dimensional change. What none of them establishes is causation. The series has no denominator and no pre-treatment echocardiograms, the physiological effect has never been shown to be large enough to decompensate a heart, and the study that found cardiac changes found no heart failure. The largest cohort, which is also the only one with a comparison group, found no excess of heart failure in treated cats.

The defensible position sits between the two that the field currently offers. The claim as it is usually stated, that steroids cause heart failure in cats, is not supported by the evidence. The opposite claim, that a steroid cannot contribute in a cat with an already compromised ventricle, is not supported either, and this appraisal does not advance it. What follows for practice is a matter for the attending clinician and the individual animal, and this page takes no position on it.

The ACVIM consensus statement flags extended-release glucocorticoids as one thing to consider before treating an older cat, and grades that recommendation as low-level evidence.¹⁴ That is an accurate label, and the statement is worth reading as the caution it was intended to be rather than as a finding.

The study that would settle the question is not a complicated one. It would require a prospective study of cats with known heart disease that are given steroids, or failing that a retrospective review of cardiac structure and function in cats with heart disease who received them. The population needed for either design exists in every referral practice, and neither study has been performed.

Until one of them is, the association remains an association, and repeating it more often will not convert it into evidence.

—References

1. Smith SA, Tobias AH, Fine DM, Jacob KA, Ployngam T. Corticosteroid-associated congestive heart failure in 12 cats. Int J Appl Res Vet Med. 2004;2(3):159–170.
2. Ployngam T, Tobias AH, Smith SA, Torres SM, Ross SJ. Hemodynamic effects of methylprednisolone acetate administration in cats. Am J Vet Res. 2006;67(4):583–587. PMID 16579749.
3. Rishniw M, Center SA, Randolph JF, Warner KL, Stokol T. Methylprednisolone acetate fails to alter echocardiographic variables in healthy cats [abstract 33]. In: Oral Research Communications of the 18th ECVIM-CA Congress, Ghent, 4–6 September 2008. J Vet Intern Med. 2008;22(6):1456–1483 (abstract p. 1469). doi:10.1111/j.1939-1676.2008.0204.x.
4. Khelik IA, Berger DJ, Mochel JP, Seo YJ, Palerme JS, Ware WA, Ward JL. Clinicopathologic, hemodynamic, and echocardiographic effects of short-term oral administration of anti-inflammatory doses of prednisolone to systemically normal cats. Am J Vet Res. 2019;80(8):743–755. doi:10.2460/ajvr.80.8.743. PMID 31339769.
5. Block CL, Oyama MA. Echocardiographic and biomarker evidence of plasma volume expansion after short-term steroids administered orally in cats. J Vet Intern Med. 2020;34(1):29–34. doi:10.1111/jvim.15678. PMID 31816124.
6. Dutch WA, Norsworthy GD, Mayfield MR, Mayfield JR, Schmitt CW. Incidences of steroid-induced diabetes mellitus and congestive heart failure in cats given non-immunosuppressive doses of methylprednisolone acetate: 1042 cats. Can Vet J. 2023;64(11):1051–1057. PMID 37915779.
7. Novo Matos J, Pereira N, Glaus T, Wilkie L, Borgeat K, Loureiro J, Silva J, Law V, Kranjc A, Connolly DJ, Luis Fuentes V. Transient myocardial thickening in cats associated with heart failure. J Vet Intern Med. 2018;32(1):48–56. doi:10.1111/jvim.14897. PMID 29243322.
8. Poissonnier C, Guigo P, Foulex P, et al. Transient restrictive cardiomyopathy in cats: first reported case series. Front Vet Sci. 2025;12:1607384. doi:10.3389/fvets.2025.1607384. PMID 40487386.
9. Pereira NJ, Novo Matos J, Baron Toaldo M, Bartoszuk U, Summerfield N, Riederer A, Reusch C, Glaus TM. Cats with diabetes mellitus have diastolic dysfunction in the absence of structural heart disease. Vet J. 2017;225:50–55. doi:10.1016/j.tvjl.2017.04.017. PMID 28720299.
10. Kelly JJ, Mangos G, Williamson PM, Whitworth JA. Cortisol and hypertension. Clin Exp Pharmacol Physiol Suppl. 1998;25:S51–S56. PMID 9809193.
11. Liu C, Chen H, Zhou C, et al. Potent potentiating diuretic effects of prednisone in congestive heart failure. J Cardiovasc Pharmacol. 2006;48(4):173–176. doi:10.1097/01.fjc.0000245242.57088.5b. PMID 17086096.
12. Liu C, Liu K. Effects of glucocorticoids in potentiating diuresis in heart failure patients with diuretic resistance [review]. J Card Fail. 2014;20(9):625–629. doi:10.1016/j.cardfail.2014.06.353. PMID 24969700.
13. Sharkey SW, Lesser JR, Zenovich AG, Maron MS, Lindberg J, Longe TF, Maron BJ. Acute and reversible cardiomyopathy provoked by stress in women from the United States. Circulation. 2005;111(4):472–479. doi:10.1161/01.CIR.0000153801.51470.EB. PMID 15687136.
14. Luis Fuentes V, Abbott J, Chetboul V, Côté E, Fox PR, Häggström J, Kittleson MD, Schober K, Stern JA. ACVIM consensus statement guidelines for the classification, diagnosis, and management of cardiomyopathies in cats. J Vet Intern Med. 2020;34(3):1062–1077. doi:10.1111/jvim.15745. PMID 32243654.
15. Romito G, Elmi A, Guglielmini C, et al. Transient myocardial thickening: a retrospective analysis on etiological, clinical, laboratory, therapeutic, and outcome findings in 27 cats. J Vet Cardiol. 2023;50:51–62. doi:10.1016/j.jvc.2023.09.001. PMID 37924558.
16. Henry P, Culshaw G, Oliveira MI, Glebocka M, Johnsen CT. Recurrent steroid-responsive hypertrophic obstructive cardiomyopathy in a Maine Coon cat. Vet Rec Case Rep. 2025;13(1):e1003. doi:10.1002/vrc2.1003.
17. Whitworth JA, Gordon D, Andrews J, Scoggins BA. The hypertensive effect of synthetic glucocorticoids in man: role of sodium and volume. J Hypertens. 1989;7(7):537–549. doi:10.1097/00004872-198907000-00005. PMID 2760458.
18. Connell JM, Whitworth JA, Davies DL, Lever AF, Richards AM, Fraser R. Effects of ACTH and cortisol administration on blood pressure, electrolyte metabolism, atrial natriuretic peptide and renal function in normal man. J Hypertens. 1987;5(4):425–433. PMID 2822795.
19. Waller DG, Barrett DF, Polak A. Methylprednisolone pulses and urine electrolyte excretion. Arch Int Pharmacodyn Ther. 1988;292:258–265. PMID 3395172.
20. Gaillard CA, Koomans HA, Rabelink TJ, Braam B, Boer P, Dorhout Mees EJ. Enhanced natriuretic effect of atrial natriuretic factor during mineralocorticoid escape in humans. Hypertension. 1988;12(4):450–456. PMID 2971618.
21. Cappuccio FP, Markandu ND, Buckley MG, Sagnella GA, Shore AC, MacGregor GA. Changes in the plasma levels of atrial natriuretic peptides during mineralocorticoid escape in man. Clin Sci. 1987;72(5):531–539. doi:10.1042/cs0720531. PMID 2953526.
22. Yao TC, Huang YW, Chang SM, Tsai SY, Wu AC, Tsai HJ. Association between oral corticosteroid bursts and severe adverse events: a nationwide population-based cohort study. Ann Intern Med. 2020;173(5):325–330. doi:10.7326/M20-0432. PMID 32628532.
23. Abbasi SH, Lund LC, Hallas J, Ernst MT, Pottegård A. Acute cardiovascular effects associated with the use of prescription medications: a Danish nationwide screening study. Br J Clin Pharmacol. 2025;91(7):1947–1956. doi:10.1002/bcp.16406. PMID 39980130.
24. Miró Ò, Takagi K, Gayat E, et al. CORT-AHF study: effect on outcomes of systemic corticosteroid therapy during early management of acute heart failure. JACC Heart Fail. 2019;7(10):834–845. doi:10.1016/j.jchf.2019.04.022. PMID 31521676.
25. Meduri GU, Shih MC, Bridges L, et al. Low-dose methylprednisolone treatment in critically ill patients with severe community-acquired pneumonia. Intensive Care Med. 2022;48(8):1009–1023. doi:10.1007/s00134-022-06684-3. PMID 35723686.
26. Tsai HJ, Wallace BI, Waljee AK, et al. Association between antenatal corticosteroid treatment and severe adverse events in pregnant women. BMC Med. 2023;21:413. doi:10.1186/s12916-023-03125-w. PMID 37907932.
27. Gursoy A, Cesur M, Erdogan MF, Corapcioglu D, Kamel N. New-onset acute heart failure after intravenous glucocorticoid pulse therapy in a patient with Graves ophthalmopathy. Endocrine. 2006;29(3):513–516. PMID 16943591.
28. Park SW, Kim K, Lee YJ, Do YJ, Ro WB, Lee CM. Prevalence and characteristics of transient myocardial thickening in cats with hypertrophic cardiomyopathy phenotypes. Vet Q. 2025;45(1):2593357. doi:10.1080/01652176.2025.2593357. PMID 41358391.

—Declarations