Infantile spasms treatment: ACTH, steroids and vigabatrin

Short answer: infantile spasms are treated with one of three standard medicines: ACTH (an injection), an oral steroid such as prednisolone, or vigabatrin.^1,2,3 For most babies, hormones stop spasms more often than vigabatrin alone, but for spasms caused by tuberous sclerosis complex (TSC), vigabatrin is the first choice.^1,4

Treatment should start fast and be checked fast. The US consensus calls for early recognition, a short course of first-line treatment, a timely EEG to check the response, and a prompt change if it is not working.⁵ This page explains what the evidence shows for each option, so families can follow the neurologist's plan and ask good questions. The condition itself, including its form called West syndrome, is explained separately.

First-line options

• ACTH (adrenocorticotropic hormone). An injection into the muscle. The US label for ACTH gel gives 150 U/m² a day, split into two injections, for 2 weeks, then a 2-week taper.³ UK trials used a synthetic depot form (tetracosactide).⁶
• Oral steroids. Usually prednisolone by mouth. UK trials used at least 40 mg a day for 2 weeks, followed by a 2-week taper.⁶
• Vigabatrin. An oral medicine, approved on its own for infantile spasms from 1 month to 2 years. The dose starts at 50 mg/kg/day and can rise to 150 mg/kg/day, split into two doses.⁷

A 2026 review of 28 guidelines found that earlier guidelines (2004 to 2015) favored ACTH, while recent ones (2016 to 2024) increasingly endorse prednisolone and vigabatrin. Guidelines from low- and middle-income countries uniformly recommend prednisolone.¹

Infantile spasms treatment success rates

Success rates depend on how "success" is defined (spasms stopping, EEG clearing, or a lasting response) and on the cause of the spasms.

• Vigabatrin
• Hormone treatment (ACTH or steroids)

UKISS trial, babies without TSC: spasms stopped by day 14

Hormone treatment

73%

Vigabatrin

54%

US national cohort: lasting response at 3 months

ACTH

55%

Oral steroids

39%

Vigabatrin

36%

Show as a table

Group	Measure	Value
UKISS trial, babies without TSC: spasms stopped by day 14	Hormone treatment	73%
UKISS trial, babies without TSC: spasms stopped by day 14	Vigabatrin	54%
US national cohort: lasting response at 3 months	ACTH	55%
US national cohort: lasting response at 3 months	Oral steroids	39%
US national cohort: lasting response at 3 months	Vigabatrin	36%

• UKISS trial (babies without TSC). By days 13 to 14, spasms had stopped in 73% on hormones and 54% on vigabatrin.⁸ By 12 to 14 months, about three in four were spasm-free in both groups.¹⁰
• US national cohort, 22 centers. A lasting response at 3 months was seen in 55% on ACTH, 39% on oral steroids and 36% on vigabatrin.⁹
• US cohort, 23 centers. Freedom from treatment failure at 60 days was 46% on ACTH, 44% on oral steroids, 37% on vigabatrin and 8% on non-standard medicines.¹¹
• Guideline-based care. At one center using a management guideline, remission at 3 months was 79% with ACTH, 83% with prednisolone and 40% with vigabatrin.¹²
• Reviews. A Cochrane review of 18 trials, most of poor quality, found hormones stop spasms faster and in more babies than vigabatrin.¹³ A 2023 meta-analysis found vigabatrin alone less effective than hormones alone (odds ratio 0.37).⁴

The underlying cause has the strongest effect on the long-term outlook.² Medicines outside the standard three worked much less often as a first treatment (8% in the US cohort above).¹¹

ACTH vs prednisolone for infantile spasms

ACTH has long been the reference treatment, but recent evidence suggests high-dose prednisolone works about as well:

• UKISS and ICISS combined (126 babies on hormones alone). Spasms stopped by days 13 to 14 in 76% on depot ACTH and 67% on prednisolone. The difference was not significant, and development scores and later epilepsy were similar. The authors concluded that either may be recommended.⁶
• Sri Lanka trial (97 babies). Spasms stopped by day 14 in 58% on prednisolone (40 to 60 mg/day) against 37% on synthetic ACTH, and prednisolone worked faster (about 4 days vs 9).¹⁴
• 2025 meta-analysis. No significant difference in spasm control at day 14 or in EEG clearing between ACTH and steroids, but ACTH caused more weight gain. The studies varied a lot.¹⁵
• US cohort. Switching from oral steroids to ACTH was not estimated to change the response rate.¹¹
• The ACTH label trial. In the small trial behind the US label, 13 of 15 babies responded to ACTH against 4 of 14 on prednisone 1 mg/kg twice a day.³ That prednisone dose was lower than in the later trials.

Prednisolone dose matters. In a trial of 63 children, spasms stopped by day 14 in 52% on 4 mg/kg/day against 25% on 2 mg/kg/day, with similar side effects.¹⁶

Vigabatrin first for tuberous sclerosis

TSC is the clear exception to "hormones first." In a small trial, all 11 babies with TSC given vigabatrin stopped having spasms, against 5 of 11 given hydrocortisone.¹⁷ In a meta-analysis, vigabatrin worked much better in TSC than in other causes (odds ratio 5.59).⁴ In a US cohort, 62% of babies with TSC responded to vigabatrin, against 29% of others.¹¹ Every guideline in the 2026 global review recommends vigabatrin first for TSC-related spasms.¹

Hormones and vigabatrin together

In the ICISS trial of 377 babies, adding vigabatrin to hormones stopped spasms between days 14 and 42 in 72%, against 57% on hormones alone.¹⁸ At 18 months, development scores and epilepsy rates were the same in both groups.¹⁹ About a quarter of guidelines since 2017 mention combination treatment.¹

• Hormones plus vigabatrin
• Hormones alone

Hormones plus vigabatrin

72%

Hormones alone

57%

Development at 18 months was the same in both groups.

Show as a table

Group	Measure	Value
ICISS trial: spasms stopped between days 14 and 42	Hormones plus vigabatrin	72%
ICISS trial: spasms stopped between days 14 and 42	Hormones alone	57%

How response is judged

• About two weeks. Most guidelines check response at about 14 days.¹
• Spasms and EEG. A full response usually means no more spasms plus clearing of the EEG pattern called hypsarrhythmia.²⁰ In a large US study, a sustained response meant the last spasm by day 14, no hypsarrhythmia on an EEG at 2 to 4 weeks, and remission lasting to day 30.²¹
• Fast responses. In that study, 43% of 395 babies went into remission within 2 weeks, most of them in the first week. The median time to response was 3 days on vigabatrin, 4 on ACTH and 3 on oral steroids. The authors advise reassessing straight away a baby who has not responded after 1 week.²¹

Keep a simple diary of spasm clusters each day and film any you see. It helps the neurologist judge the response.

If the first treatment does not work

First treatments often fail, but a second one helps many babies. In a US cohort of 118 babies whose first treatment failed, 37% responded to a second.²² Switching to a standard medicine with a different action (from hormones to vigabatrin, or the reverse) worked in 55%, against 25% for other sequences. Babies first treated within 4 weeks of onset also responded better to the second treatment (44% vs 24%).²²

If vigabatrin is not clearly helping, the label advises withdrawing it within 2 to 4 weeks, by tapering rather than stopping suddenly, because of the risk of vision loss.²³

Side effects of each treatment

ACTH and steroids

On the ACTH label, the most common side effects in babies with infantile spasms are infections, convulsions, high blood pressure, irritability and fever. At the recommended dose, 20% had infections and 11% high blood pressure.³ The label also warns about a Cushing-like appearance, salt and water retention, low potassium, and effects on growth. Stopping must be gradual, because the body's own steroid production can be suppressed, and signs of this in babies (poor feeding, tiredness, weakness, weight loss) are easy to miss.³ In pooled studies, ACTH caused more weight gain than oral steroids, with no significant difference in infection, irritability or high blood pressure.¹⁵

Vigabatrin

Vigabatrin can cause permanent loss of side (peripheral) vision. The risk rises with dose and total exposure, and eye checks are required at the start, at least every 3 months, and after stopping.⁷ Brain MRI changes were seen in 22% of babies on vigabatrin against 4% on other treatments. Drowsiness and weight gain also occur.⁷ Symptomatic MRI changes have been linked to giving hormones at the same time in one study.²⁴

In UKISS, side effects were reported about equally often with hormones and with vigabatrin (55% vs 54%).⁸

Why speed and accurate doses matter

Time counts. In UKISS, each step up in the delay to effective treatment was linked to a lower development score at age 4.²⁵ In ICISS, babies whose spasms stopped early had higher development scores at 18 months (79.1 vs 63.2).¹⁹

That is why the dose a baby receives matters from day one. In vigabatrin's main dose trial, 100 to 148 mg/kg/day stopped spasms more often than 18 to 36 mg/kg/day (24 of 67 against 8 of 75 responders).²⁶ Those doses were several times apart, so the link from that trial to dose-measuring accuracy is indirect. Every vigabatrin form is the same medicine.⁷ Vigabatrin powder has to be mixed with water before every dose, while the ready-to-use liquid (Vigafyde) needs no mixing.^7,27 In the one direct comparison, all 30 doses drawn from the liquid were within 5% of the target, while 7 of 30 doses mixed from powder missed by more than 10%, and the authors attribute the powder errors to mixing.²⁸ That evidence has clear limits. It was one small study (30 people), run by employees of Pyros, which holds the Vigafyde application.^28,29 Each person prepared one large lab dose of each form, and the liquid arm used a stand-in solution, with doses calculated from the volume delivered rather than measured.²⁸ No study links the form to seizure outcome.²⁸ Like the powder, the liquid is labeled for infantile spasms from 1 month to 2 years.⁷

What guidelines say

• International League Against Epilepsy (2015): ACTH preferred for short-term spasm control; oral steroids probably effective; a shorter time to treatment may improve outcome.³⁰
• American Academy of Neurology (2012): ACTH more effective than vigabatrin in the short term outside TSC. This guideline has since been retired.³¹
• Global review (2026): vigabatrin first for TSC in every guideline; response usually judged at about 14 days.¹
• American Epilepsy Society (2026): its infant epilepsy guideline leaves infantile spasms out on purpose.³²

Questions to ask the neurologist

• Which first treatment do you recommend for my baby, and why?
• When will we repeat the EEG, and what counts as a response?
• What side effects should we watch for, and who do we call?
• What is the plan if this treatment does not work by 2 weeks?
• If vigabatrin is used, which form, and how do we measure each dose?

Answers to common questions are on the FAQ, and summaries of the key trials are on research.