ACTH, phenobarbital and infantile spasms — evidence for Yonah

The plan, as of 31 October 2026 (fact-checked)

This plan was re-checked claim by claim against the original papers, labels and official pages by a separate review on 31 October; the corrections it found are built in. It is a weighing of the evidence to take to his doctors, not a prescription.

1. Start ACTH now: on 1 November if at all possible. Today (31 October) is day 6 after the first spasm (counting 25 October as day 0). In the largest trial, babies treated within 7 days had the best development scores at 18 months:

Sources: response, PMID 27838190 (Jan 2017) [FT]; Vineland, PMID 30236380 (Oct 2018) [FT]. This is an association within a trial, not proof that delay causes the drop, but no study found that waiting helps.

2. PET and MEG: treatment first. - His MRI is already done, so the one imaging step that the US consensus wants before ACTH is met: "Ideally, MRI should be obtained prior to the initiation of therapy" (PMID 20608959 (Oct 2010) [FT]). - Guidelines say treatment should not wait for imaging (Indian AOCN–IES guideline: "treatment should be started without waiting for the imaging", PMID 33452776 (Jan 2021) [FT]; US consensus, PMID 20608959 (Oct 2010) [FT]). They mainly suggest PET for spasms that don't respond to treatment. - ACTH may change the PET picture. In 12 babies with West syndrome scanned twice, the under-active areas often went back to normal once spasms stopped (Maeda 1994, PMID 7911927 (Jun 1994) [ABS]). Steroids lowered brain glucose readings in adults (Fulham 1995, PMID 7674016 (Oct 1995) [ABS]). Nothing was found on MEG. - So ask: can the PET be done before the first ACTH dose, in the next day or two? If not, start ACTH anyway. Waiting for 16 November moves him into the 15–28-day group above.

3. Dose: ask what they plan and why. - Japan's usual starting dose is 0.0125 mg/kg/day or less (over 80% of Japanese paediatric neurologists, PMID 29731400 (Sep 2018) [ABS]). At about 5–5.5 kg that is about a quarter to under a third of the UK trial starting dose (0.5 mg every other day). It is a fifth only at about 4 kg. - Within Japan's range, one small Japanese randomised trial (25 babies) compared 0.005 with 0.025 mg/kg/day: no difference in how many responded, and less sleepiness and brain shrinkage on the low dose (PMID 10522523 (Oct 1999) [ABS]). Japan's dose has never been compared with the UK or US trial doses. - Worldwide, a 2026 review of reviews rated "high-dose hormonal therapy outperformed low-dose" as moderate certainty (PMID 41762758 (Apr 2026) [ABS]). Its summary does not define high or low dose, and does not say whether ACTH and prednisolone were pooled, so it cannot settle the choice within Japan's range. - The Japanese asphyxia study does not cover babies like him. In 48 babies with spasms after oxygen-lack injury, a higher total ACTH dose went with longer control (weak: r=0.36, p=0.047). But that was tested only in the 35 whose epilepsy began at 3 months or later. Babies whose spasms began at 0–2 months, as Yonah's did, were left out because their control was "extremely short"; they did significantly worse. Infected babies had received higher doses (infection in 38%), which is an association, not proof of cause (doi:10.11251/ojjscn.55.344 (2023) [FT]). - Our weighing of this: ask for the upper part of the Japanese range (up to 0.025 mg/kg/day, the top of the range Saitama Children's Medical Center uses: 0.010–0.025, PMID 39705749 (Jan 2025) [FT]) rather than the lowest doses. The costs are more sleepiness, temporary brain shrinkage on scans and more infections. Timing is the strongest evidence here; dose is not. - Step-up at day 7 if spasms continue, as the UK trial did (0.5 → 0.75 mg; PMID 27838190 (Jan 2017) [FT]). - EEG at about day 14 of treatment, with the next step agreed in advance.

4. Course: "a month" matches standard practice. That is 2 weeks of daily injections, then a taper. The UK trial stopped all hormones after day 29 (PMID 27838190 (Jan 2017) [FT]). Saitama gives 2 weeks daily, then every other day for 1–2 weeks, then twice weekly for 1 week (PMID 39705749 (Jan 2025) [FT]). Japanese practice varies, and some centres stop at 14 days with no taper (PMID 29731400 (Sep 2018) [ABS]).

5. Vigabatrin: the backup, not the first choice. - Hormones first: hormones stopped spasms in more babies than vigabatrin: 73% vs 54% at day 14 (UKISS, PMID 15541450 (Nov 2004) [ABS]). In a small Italian trial, ACTH did better "in perinatal hypoxic/ischemic injury", which is Yonah's background (PMID 9578521 (Dec 1997) [ABS]). - Vision risk: vigabatrin can cause permanent loss of side vision (Japanese label: about 1 in 3 patients) [L]. - Where it comes in: if ACTH fails at the day-14 EEG (switch or add); or added from the start (ICISS: 72% vs 57% spasm-free, but no development difference at 18 months, PMID 27838190 (Jan 2017), PMID 30236380 (Oct 2018) [FT]); or if the PET or MEG finds a malformation, where vigabatrin did better (PMID 9578521 (Dec 1997) [ABS]). - Ask: is Kyoto University Hospital registered to prescribe Sabril? In Japan only registered hospitals and doctors can (Sabril label, SRSP registration system) [L].

6. Before the first dose: - check chickenpox history and immunity (Cortrosyn Z label: chickenpox during treatment can be fatal) [L]; - check for congenital infection such as CMV (cytomegalovirus). The US label says ACTH must not be used in infants under 2 "who have suspected congenital infections" [L]; - agree how often blood pressure, heart echo/ECG, head scans and blood tests will be checked (Cortrosyn Z label requires head CT and heart checks in children) [L].

7. Vaccines. - Don't delay ACTH for vaccines. The US CDC's default is to start immune-suppressing treatment 2 weeks after inactivated and 4 weeks after live vaccines (CDC, Altered Immunocompetence, updated Jun 2024). But no guideline asks a baby with spasms to wait that long. The German spasms guideline says inactivated vaccines are possible during hormone treatment, with a possibly weaker response (PMC9643068 [G]). - Inactivated (hepatitis B, pneumococcal, 5-in-1): ask the team whether to give them now or during the course, and whether to repeat any doses afterwards. For the 5-in-1, ask about the US precaution: whooping-cough vaccine in "infantile spasms" is deferred "until neurologic status clarified and stabilized", decided case by case (CDC, Contraindications and Precautions, updated Jul 2024). - Rotavirus (live): don't give it right before ACTH. US, Australian and German rules all say about 4 weeks between a live vaccine and high-dose steroids or hormones. If ACTH runs about 1–29 November, 4 weeks later is about 27 December. - The recommended last first dose is 12 December. - Japan's rules allow a later first dose with explanation and consent (MHLW notice, Mar 2026 [G]). The last first dose that still completes the course is about 17–18 January 2027 for Rotarix and about mid-February 2027 for RotaTeq. - US, UK and Japan Pediatric Society guidance do not recommend a first dose after 15 weeks. This is a choice to discuss with the team. - Catch-up: Kyoto City's free catch-up lists ウェスト症候群（点頭てんかん） (West shōkōgun / tentō tenkan) as a qualifying illness. It needs an application with a doctor's opinion. Rotavirus is excluded (Kyoto City). - BCG later: the Japanese BCG label says not during, or within 6 months after, long-term or high-dose corticosteroids (prednisolone etc.). ACTH is not named, so applying this to ACTH is an inference. Ask when BCG can be given [L].

What to say to the doctors:

"His spasms started on 25 October, so today is day 6. In the UK trial, babies treated within 7 days had the best development scores, so we'd like ACTH started tomorrow. We'd prefer the upper part of the Japanese dose range, with a step-up at day 7 if spasms continue, an EEG at about day 14, and a plan for vigabatrin if it hasn't worked. Is this hospital registered for Sabril? If the PET can be done before the first dose, please do it; otherwise please don't hold ACTH for it. Please give the inactivated vaccines as you think best, and plan rotavirus for after the course."

Background to the plan

Facts as reported: born 30 August 2026 after neonatal asphyxia; spasms (eye rolling, then body spasms) first seen 25 October; high-resolution video-EEG on 30–31 October shows spasms plus left occipital focal seizures; an earlier EEG elsewhere was normal; his first MRI (Japan Baptist Hospital, Kyoto) showed nothing; phenobarbital was suggested (until 18 November, review 9 November) but he is not on it; his doctors propose a month of ACTH; PET and MEG are planned for 16–17 November; he is in hospital now on video-EEG; 2-month vaccines (appointment was 4 November) not yet given.

1. Phenobarbital does not treat infantile spasms. In Japan's largest national study (501 children), phenobarbital as the first treatment kept spasms away for a mean of about 2 weeks (0.5 months), and none of the 23 children became seizure-free; ACTH as the first treatment kept them away for a mean of 16.2 months (Takahashi, PMID 34773678 (Mar 2022) [FT]; the 11 hospitals in this study treat many severe cases, and most data predate vigabatrin in Japan). In children who, like Yonah, had spasms and focal seizures, phenobarbital controlled the spasms for 0.2 months and the focal seizures for 0.0 months — but only 2 children got it, against ACTH 2.0 and 13.3 months in 6 children, and the authors found the difference not statistically significant. In the large US cohort, "non-standard" drugs stopped spasms in only 3 of 32 (9%) infants in 2016 — a group that included phenobarbital — against 36–55% for ACTH, prednisolone or vigabatrin (PMID 26704170 (Mar 2016) [FT]). (A 2021 analysis found 4 of 51 (8%), but none of those 51 received phenobarbital — mostly topiramate and levetiracetam — so it speaks to non-standard drugs in general, not phenobarbital, PMID 34266919 (Sep 2021) [FT].) Children given non-standard drugs were more likely to have severe developmental problems, which partly explains the gap. No guideline found anywhere — 43 checked — recommends phenobarbital for spasms, and the US consensus says the evidence "does not support" it (PMID 20608959 (Oct 2010) [FT]).
2. Even for his focal seizures, phenobarbital would not be the best-supported choice. In infants 1–12 months with new epilepsy, babies were seizure-free within 3 months and stayed so to 6 months on that one drug in 6 of 38 (16%) on phenobarbital vs 47 of 117 (40%) on levetiracetam (PMID 29435578 (Apr 2018); cohort — the phenobarbital babies were younger, median 3.0 vs 5.2 months). The 2026 American Epilepsy Society guideline "suggests the use of levetiracetam rather than phenobarbital", citing "potential neurotoxicity and adverse cognitive effects" with prolonged phenobarbital (PMID 41969626 (May 2026) [FT]).
3. Delay matters, and the clock started on 25 October. In the largest trial (377 infants), development at 18 months was 78.2 if treatment started within 7 days of the first spasm, 77.3 at 8–14 days (no real difference), but 72.3 at 15–28 days (about 6 points lower, a statistically significant difference), 68.8 at 29 days–2 months and 65.5 after 2 months (PMID 30236380 (Oct 2018) [FT]). Counting 25 October as day 0: today, 31 October, is day 6; starting by 1 November keeps him in the best (≤7-day) group; 16 November (PET) = day 22 and 18 November = day 24, both in the 15–28-day group. This is an association within a trial (babies treated later also differed in other ways), but every study that measured delay points the same way, and none found that waiting helps.
4. Japan's own guideline sets a 2-week limit. The Japan Epilepsy Society West syndrome guideline: "If another therapy is given before ACTH, its effect should be judged within 2 weeks, and if ineffective, ACTH should be given"; and "ACTH should be used as soon as possible after onset" (JES guideline PDF [G], draft PDF dated May 2009; the "within 1 month" target is worded for spasms with no identified cause). The US consortium says children not responding to a standard treatment within 1 week "should be reassessed immediately" (PMID 36038267 (Nov 2022)).
5. Spasms plus focal seizures point to a focal brain injury, and such spasms are harder to control and relapse more whatever the drug (PMID 34773678 (Mar 2022) [FT]; PMID 22889572 (May 2013)). Authors in Japan, the US and India recommend identifying children who might benefit from epilepsy surgery early: video-EEG and a careful MRI review first; PET when MRI is unclear or surgery is being considered (PMID 20608959 (Oct 2010) [FT]; PMID 26803282 (Feb 2016); PMID 33452776 (Jan 2021) [FT]). After asphyxia, injury at the back of the brain is typical (posterior "watershed" injury in 80.6% of children with this scar pattern, PMID 35882664 (Jan 2023)); low blood sugar after birth also causes occipital injury (96.5% of 170 children, PMID 32446209 (Jul 2020)), so it is worth knowing whether he had it.
6. The 2-month vaccines need a decision now, because they interact with hormone treatment. Inactivated vaccines (hepatitis B, pneumococcal, 5-in-1) work best given 1–2 weeks before hormones start. Rotavirus (live, by mouth) has hard age limits: first dose recommended by 12 December 2026; UK/US rules never allow starting from 13 December, while Japan allows a later first dose with consent (about mid-January for Rotarix, mid-February for RotaTeq); Japan's free catch-up rule excludes rotavirus (§13).

What the evidence supports asking for: see the plan at the top of this section. In short: spasm-specific treatment (ACTH first; prednisolone is not insurance-covered in Japan) started now, not after the PET/MEG or on 9 or 18 November; ask what treatment is planned for the focal seizures, and why phenobarbital rather than levetiracetam (if phenobarbital is ever started, it should not be stopped suddenly — its label warns that this can cause status epilepticus); a plan for the 2-month vaccines before hormones start; the MRI findings; and an early discussion of whether he could be a surgery candidate if medicines fail.

How well it works - ACTH is one of three standard first treatments for infantile spasms, with high-dose prednisolone and vigabatrin (2026 review of 28 guidelines worldwide, PMID 42274306 (Sep 2026) [FT]). - Hormones (ACTH or prednisolone) stop spasms in more babies than vigabatrin in the short term: 73% vs 54% (UK trial, ages 2–12 months, PMID 15541450 (Nov 2004) [ABS]); 74% vs 48% (Italian trial, PMID 9578521 (Dec 1997) [ABS]). The evidence for this is rated moderate-to-high certainty (2026 review of 20 reviews, PMID 41762758 (Apr 2026) [ABS]). - ACTH and high-dose prednisolone look roughly equal: the pooled UK/international trial arms gave 76% vs 67% (not a significant difference, PMID 36621063 (Jan 2023) [FT]); one trial favoured prednisolone (58% vs 37%, PMID 26216500 (Sep 2015) [FT]); one small trial favoured very-high-dose natural ACTH (87% vs 29%, against low-dose prednisone, PMID 8604274 (Mar 1996) [FT]). Certainty is low. - Higher hormone doses work better (moderate certainty, PMID 41762758 (Apr 2026) [ABS]) — and Japan's usual ACTH dose is far lower than the doses tested in the main trials (§8). - In the best-studied countries, spasms stop in roughly 55–75% of babies with the first standard treatment, and responders mostly respond within the first week (median 3–4 days, PMID 36038267 (Nov 2022)).

The positives - Stops spasms in a majority of babies, usually within days. - Long track record (first reported 1958). - Japan's ACTH is inexpensive (¥6,737 per vial) and insurance-covered. - In Japan's national study it gave the longest spasm-free periods of any first treatment (mean 16.2 months vs 0.5 months for phenobarbital, PMID 34773678 (Mar 2022) [FT]). - No study has shown that ACTH itself causes lifelong intellectual or physical disability (§9).

The dangers - Infection is the most important: 13% in the UK-style trial arm (PMID 27838190 (Jan 2017) [FT]); 10.6% with Japanese low dose (doi:10.11251/ojjscn.54.34 (2022) [FT]); 38% of 48 Japanese babies with spasms after asphyxia (doi:10.11251/ojjscn.55.344 (2023) [FT]). ACTH can hide fever, and chickenpox during treatment can be fatal (Japanese label [L]). - High blood pressure (11–44% across studies; 34% in the Japanese asphyxia cohort), irritability (35–69%), weight gain, salt and potassium changes. - Thickening of the heart muscle (seen on echo from week 1, usually without symptoms, PMID 35620145 (2022)), reversible "brain shrinkage" on scans and occasionally bleeding under the skull membrane (subdural haematoma; 3 of 116 in a Japanese series, PMID 39705749 (Jan 2025) [FT]). - Adrenal suppression after stopping (rates from 0 of 11 to 76% depending on regimen and test, §10). - Deaths are rare in modern trials (none attributed to treatment among 377 infants, PMID 27838190 (Jan 2017) [FT]); historically, with much higher doses, 8 of 162 (4.9%) died, mostly of pneumonia (Finland 1960–76, PMID 6254450 (Sep 1980) [FT]).

Bottom line across countries: what best predicts long-term development is the cause of the spasms, then whether spasms (and the EEG) are brought under control quickly — not which hormone is used. The best-documented programmes (the UK/European trial network and the US consortium) use a standard drug first (never vitamin B6, valproate or phenobarbital), start within days, check the EEG at about 2 weeks, and switch quickly if the first drug fails (§8).

Bottom line (searched from the 1950s to October 2026, in English, Japanese and other languages): no study of any design shows that phenobarbital controls infantile spasms as well as the standard treatments; every comparative dataset points the other way. Confidence: moderately high. There is no randomised trial, because no one has thought it worth testing; but independent large cohorts in Japan and the US, and every guideline, agree, and nothing credible points the other way.

• Japan, national study, 501 children (2022): phenobarbital first: mean 0.5 months spasm-free (n=23), none seizure-free; ACTH first: 16.2 months (n=30) (PMID 34773678 (Mar 2022) [FT]).
• Japan, 2023, babies whose spasms began with a "suppression-burst" EEG: phenobarbital "could not suppress at all" (全く抑制できなかった) — 0 months in 3 uses (doi:10.11251/ojjscn.55.427 (2023) [FT]).
• US consortium: non-standard drugs: 9% response in 2016 (a group including phenobarbital) and 8% in 2021 (a group without phenobarbital) (above). When a hormone had failed, switching to another non-standard drug worked in 1 of 30 (3.3%); switching to vigabatrin in 22 of 65 (34%) (PMID 40528701 (Oct 2025) [ABS]).
• Guidelines: none of the 28 in the 2026 worldwide review, nor 15 others checked individually, recommends phenobarbital for spasms.
• Evidence on the other side is only single cases in rare genetic or Ohtahara-type epilepsies, one child out of 21 in "spasm status", and a 2010 Japanese expert workshop that found it useful only as an add-on after ACTH failed.

Where phenobarbital genuinely is first choice: seizures in newborns (the first 28 days of life) — ILAE 2023 guideline (PMID 37655702 (Oct 2023)), Cochrane 2023 (PMID 37873971 (Oct 2023)), based mainly on one trial where 80% of newborns became seizure-free on phenobarbital vs 28% on levetiracetam (PMID 32385134 (Jun 2020)). In Japan its labelled uses are newborn seizures and status epilepticus (Nobelbal label, Jun 2026 [L]); for status epilepticus the label says a fast-acting drug should be used first.

After neonatal seizures, keeping a baby on phenobarbital did not lower the later epilepsy rate (US registry, about 280 babies, PMID 34028496 (Jul 2021) and later follow-up).

Phenobarbital and development: in a randomised trial in toddlers (8–36 months, treated for febrile seizures), IQ was about 7 points lower after 2 years, with a reading deficit at school age (PMID 2242106 (Feb 1990); PMID 10416094 (Jul 1999)). In newborn animals including monkeys it causes brain-cell death (report 11). These were not babies with spasms, but they are why guidelines now prefer other drugs for infant epilepsy.

Why might doctors still choose it? Documented reasons are its role for newborn seizures and status epilepticus, its long Japanese approval for those uses, and its use for co-existing focal seizures. None of these is evidence that it treats spasms. Published trials and cohorts routinely start standard spasm treatment while another anti-seizure drug continues for other seizure types (14% of the infants in the largest trial were already on one, PMID 27838190 (Jan 2017) [FT]); no guideline requires the focal seizures to be controlled first.

Short answer: yes, probably — mainly for development; speed matters less for whether the spasms stop.

How it was measured. "Lead time" is the time from the first spasm to starting treatment. The big UK trials grouped babies by lead time and measured (a) whether spasms stopped and (b) development later on the Vineland scale (average 100).

Whether spasms stopped (ICISS trial, 374 infants, PMID 27838190 (Jan 2017) [FT]):

In plain words: babies treated within 4 weeks did about equally well at stopping spasms; the clear drop came after 2 months.

Development at 18 months (same trial, 359 infants, PMID 30236380 (Oct 2018) [FT]):

In plain words: after the first two weeks, each longer delay went with a lower average score — about 3.3 points per step (before adjustment; after adjusting for other factors, lead time remained a significant independent predictor). A 6-point gap (78 vs 72) is a noticeable but modest difference in everyday skills; the gap between 78 and 65 is larger. Caveat: babies treated later were also older and at higher risk to begin with, so part of this gap is not caused by the delay itself.

The "93.3 vs 66.2" figure, explained. In the earlier UK trial, among children with no identified cause for their spasms, the Vineland score at age 4 was 93.3 (close to the population average of 100) if treated within 7 days, but 66.2 (well below average — more than two standard deviations below 100, by our arithmetic) if treated more than 2 months after the spasms began (PMID 21668442 (Jul 2011) [ABS]). Yonah's spasms do have a likely cause (asphyxia), so this subgroup is not directly his — but it shows how much delay can matter.

What "early spasm control" means in these studies: spasms stopped within about 2 weeks of starting treatment and stayed stopped (the trial measured days 14 to 42), ideally with the EEG cleared. In ICISS, the 229 infants whose spasms stopped (no spasms from day 14 to day 42) scored 79.1 on the Vineland scale at 18 months vs 63.2 for the 129 whose spasms continued — about 16 points higher (PMID 30236380 (Oct 2018) [FT]). This is an association within the trial, not a randomised comparison.

Other studies: in Israel, all 22 children with no identified cause treated within 1 month had normal long-term cognition, vs 40% of those treated after 1–6.5 months (PMID 15009227 (Mar 2004)); in Slovenia, mental outcome worsened once hypsarrhythmia lasted more than 3 weeks (PMID 17201726 (Dec 2006)). Two 2026 cohorts (Korea, India) found the link weakened once the cause was accounted for (PMID 42730593 (Sep 2026) [FT]; PMID 42917491 (Oct 2026) [ABS]). No study specific to spasms after asphyxia showed that faster treatment improves long-term development (report 10) — but none showed that waiting is safe either.

What guidelines say: treat "as soon as possible" (German guideline, PMID 35882373 (Dec 2022) [G]); see a specialist within 24 hours (UK NICE [G]); judge any non-ACTH first treatment within 2 weeks then switch (Japan Epilepsy Society [G]); reassess non-responders after 1 week (US consortium, PMID 36038267 (Nov 2022)).

• How common: HIE is one of the most common single known causes of spasms — 10% in the UK trial cohort (PMID 20726878 (Oct 2010)), 14% in a Japanese national ACTH registry (Nagai 2023 [FT]). After HIE, about 4% of babies develop spasms overall, more after severe HIE (12.5%), none after mild HIE in one cohort (PMID 31518001 (Jan 2020) [FT]). Spasms usually start at 4–8 months in cooled-HIE cohorts — Yonah's started earlier, at about 8 weeks.
• How the treatments do in this group (no trial has been done only in HIE):
• Hormones have the most support for injury-type causes: ACTH did better than vigabatrin "in perinatal hypoxic/ischemic injury" in a small trial (PMID 9578521 (Dec 1997) [ABS]); ACTH stopped spasms in 65% of "structural-acquired" cases vs 31% of genetic ones in a Japanese centre (PMID 33360040 (Feb 2021)).
• Japanese national series of 48 babies with spasms after oxygen-lack injury, given low-dose synthetic ACTH: spasms controlled for at least 2 months in 71%, but only 33% still free of spasms at last follow-up and 21% free of all seizures; a higher total ACTH dose went with longer control (a weak correlation, r=0.36, p=0.047, tested only in the 35 babies whose epilepsy began at 3 months or later; babies whose spasms began at 0–2 months, like Yonah, were left out because their control was "extremely short", and they did significantly worse), and infected babies had received higher doses (infections 38% overall); treatment started late, at a mean of 10 months of age (doi:10.11251/ojjscn.55.344 (2023) [FT]).
• Vigabatrin: about half respond in small Japanese series of acquired injury (5 of 10; 4 of 8; 3 of 7) (PMID 39406059 (Nov 2024); PMID 42476001 (Aug 2026)).
• A 2026 Indian cohort of mostly acquired causes found high-dose oral steroids better than vigabatrin at 1 month (PMID 42917491 (Oct 2026) [ABS]).
• Spasms plus focal seizures after asphyxia: associated with a focal or one-sided injury and with harder-to-control, relapsing spasms. In a Japanese centre, ACTH stopped spasms at first in 100% of "focal" cases but 74% relapsed (vs 38% in diffuse cases); a second ACTH course gave remission in 6 of 8 (PMID 22889572 (May 2013) [ABS]). Parietal or occipital EEG discharges within 1 month after ACTH went with later relapse (12 babies with a preterm-type white-matter injury, PMID 37207538 (Jul 2023) [ABS]).
• Long-term outlook after HIE-related spasms is dominated by the original injury (Japanese series: 98% intellectual disability, 80% motor disability at a mean of 7.8 years — older, later-treated children). Outcomes vary greatly with how much of the brain was injured: small one-sided lesions at the back of the brain had favourable epilepsy outcomes in one series (PMID 8330586 (Jul 1993)).
• PET and MEG: PET shows areas of low brain activity that MRI can miss and is used mainly to plan surgery; MEG can show where "generalised-looking" spasms start. No study shows that doing them early improves outcomes; their value is in choosing and planning surgery when medicines fail (report 10). Guidelines say treatment should not wait for imaging (PMID 33452776 (Jan 2021) [FT]; PMID 20608959 (Oct 2010) [FT]). Treatment may change PET findings: under-active areas often normalised once spasms stopped (PMID 7911927 (Jun 1994) [ABS]), and steroids lower brain glucose readings in adults (PMID 7674016 (Oct 1995) [ABS]). Surgery for posterior lesions after perinatal injury made 75% seizure-free in an Indian series (PMID 34214835 (Sep 2021)) and 84.6% in a Chinese one (PMID 38070409 (Jan 2024)); a 2024 meta-analysis found 67% spasm-free after resection (PMID 39807246 (2024) [FT]).

• Yes — one of three, with vigabatrin first for TSC in all 26 guidelines that make recommendations (PMID 42274306 (Sep 2026) [FT]).
• The world has moved toward prednisolone since about 2016: guidelines 2004–2015 recommended ACTH in 6 of 6; 2016–2024: prednisolone in 18 of 20 (90%), ACTH in 11 of 20 (55%).
• US: ACTH fell from 78% of first treatments (2006) to 18% (2020) (PMID 36600453 (Mar 2023) [ABS]); in 2026, 22 of 33 tertiary hospitals start with high-dose prednisolone and only 4 with ACTH (PMID 42569898 (Oct 2026) [FT]).
• UK: prednisolone + vigabatrin; ACTH not mentioned (NICE NG217, updated Aug 2026 [G]).
• Australia / NZ: prednisolone first; ACTH rarely, for refractory cases (RCH Melbourne Oct 2025; NZ national guideline Jul 2026 [G]).
• WHO: added prednisolone, not ACTH, to the children's essential-medicines list in 2025 (WHO).
• Japan: ACTH first in guidance; prednisolone not insurance-covered.

Prednisolone vs vigabatrin — not the same. Prednisolone is a steroid hormone given by mouth (same family as ACTH). Vigabatrin is a non-hormonal anti-seizure medicine. Hormones stop spasms in more babies short-term; vigabatrin is preferred in TSC and when steroid risk is high; vigabatrin carries a risk of permanent visual-field loss (about 1 in 3 patients per the Japanese label's boxed warning [L]; 5 of 28 testable infant-exposed children in a Finnish population study, PMID 34716587 (May 2022) [ABS]) and MRI changes.

Why most countries prefer prednisolone: 1. Similar results to ACTH in pooled trials (low certainty) (PMID 39900086 (Apr 2025); PMID 40886708 (Aug 2025); PMID 36621063 (Jan 2023) [FT]). 2. Cost: a 14-day US ACTH course cost $100,457 in 2020 vs $89 for prednisolone (PMID 36600453 (Mar 2023) [ABS]). 3. By mouth, at home rather than daily injections usually in hospital. 4. Availability: ACTH is unavailable in Korea, Thailand, Brazil and effectively Canada.

Why Japan prefers ACTH — judged on the evidence: 1. Insurance: high-dose oral steroids for spasms are not covered ("国外では高用量ステロイド療法もあるが，わが国では…保険適用外", Japanese TSC guideline 2025 [G]). 2. Guideline history: the Japan Epilepsy Society guideline rejected oral steroids as "40% or less" effective — a figure from low-dose prednisone studies; later trials showed high-dose prednisolone works much better than low-dose (e.g. 51.6% vs 25%, PMID 25048310 (Oct 2014)). 3. Japanese ACTH is cheap, so cost does not push Japan away from it. 4. Verdict (report 13): Japan's preference rests on system factors and history, not on evidence that ACTH beats high-dose prednisolone.

No country "wins" on long-term development — that depends mostly on the cause and on stopping the spasms. The strongest evidence for short-term control comes from the UK/European ICISS trial regimen and the US consortium centres. Their shared features: a standard drug first, started within days, EEG check at about 2 weeks, fast switch if it fails.

The best-documented regimen (ICISS, 377 infants aged 2–14 months, 103 hospitals in the UK, Germany, Switzerland, Australia, New Zealand; PMID 27838190 (Jan 2017) [FT]): - Prednisolone by mouth 10 mg four times a day (40 mg/day), raised to 20 mg three times a day (60 mg/day) if spasms continue on day 7; or tetracosactide depot (Synacthen Depot) 0.5 mg by injection every other day, raised to 0.75 mg if spasms continue on day 7. - With or without vigabatrin 100–150 mg/kg/day. - 2 weeks of hormone, then a prednisolone taper; EEG at day 14 read by doctors who didn't know the treatment. - Result: spasms stopped in 72% with hormone + vigabatrin vs 57% with hormone alone; no treatment-related deaths; no difference in development at 18 months (PMID 30236380 (Oct 2018) [FT]).

Is Japan's low dose as good? The evidence is weak. - Japan's usual dose (≤0.0125 mg/kg/day) is about one-quarter to one-fifth of the UK trial dose for a 4–5 kg baby (the researchers' own arithmetic: UK 0.5 mg every other day ≈ 0.05 mg/kg/day at 5 kg; the UK dose is fixed, so the ratio changes with weight — about 5× at 4 kg, 3.3× at 6 kg — and is about 6× if the trial's step-up to 0.75 mg is used). - For it: small Japanese studies (n=25 randomised; n=138, n=135 cohorts) found no clear loss of effect at lower doses and less brain shrinkage (PMID 10522523 (Oct 1999); PMID 11781414 (Jan 2002); PMID 16647410 (Apr 2006) [ABS]). - At Yonah's likely weight (about 5–5.5 kg) the Japanese dose is about a quarter to under a third of the UK starting dose. - Against it: those studies are small and old (1999–2006); in the largest, only 38% had an "excellent" EEG result (PMID 11781414 (Jan 2002) [ABS]); worldwide pooled evidence says higher hormone doses work better (moderate certainty, PMID 41762758 (Apr 2026) [ABS]; its summary does not define high or low dose or say whether ACTH and prednisolone were pooled); in China, a low natural-ACTH dose was clearly worse (spasms and EEG cleared at day 14 in 14.7% at 1 unit/kg/day vs 36.4% at 1.1–1.9 and 41.4% at 2–4 units/kg/day; retrospective, single centre, PMID 28421469 (Aug 2017) [ABS]); and in Japanese babies with spasms after asphyxia, a higher total dose went with longer control (a weak correlation in a subset of 35 whose epilepsy began at 3 months or later; babies with onset at 0–2 months, like Yonah, were excluded; doi:10.11251/ojjscn.55.344 (2023) [FT]). - Within Japan's range it has been tested once: 0.005 vs 0.025 mg/kg/day in 25 babies, with no difference in response (PMID 10522523 (Oct 1999) [ABS]). It has never been tested against the UK or US trial doses, the regimens with the best trial evidence.

Bottom line (report 12): no study shows that ACTH itself causes lifelong intellectual or physical disability. Confidence that it has not been shown to: moderate. Confidence that it has been shown not to: low — long-term studies are few and small, no study followed heart, bone, kidney, adrenal or hearing function into adulthood, and there are no long-term randomised comparisons at all for Japan's low-dose regimen.

• What decides lifelong outcome is mainly the cause: about 1 in 8 children with a known cause have a good developmental outcome vs about 1 in 2 with no known cause (meta-analysis, PMID 25524855 (Jan 2015)). Second comes how quickly and completely spasms are controlled.
• Large trials found no lasting difference between treatments: UK trial — no difference between hormones and vigabatrin at 14 months or 4 years (better with hormones in children with no known cause) (PMID 16239177 (Nov 2005); PMID 20457702 (May 2010)); pooled trials — no difference between synthetic ACTH and prednisolone at 14–18 months (PMID 36621063 (Jan 2023) [FT]); a US database of 986 matched children — "very similar" diagnoses over up to 5 years after ACTH or prednisolone (PMID 35493808 (2022)).
• One signal to take seriously: in a Sri Lankan trial, children given synthetic ACTH had more delay in 2 of 5 developmental areas two years after treatment than those given prednisolone (they were 1.5–34 months old at entry, so many were older than Yonah) (PMID 33894665 (May 2021) [ABS]) — but in that trial ACTH also stopped spasms less often (37% vs 58%), so the delay may reflect worse spasm control rather than harm from ACTH.
• Physical effects of ACTH are mostly temporary: "brain shrinkage" on scans usually returns to baseline within 1–6 months (not in every child — 6 of 24 in one study); heart-wall thickening reverses (mild thickening was still present 2 months after stopping in one study); adrenal function recovers in about 2–5 months; weight and blood-fat changes faded by school age; helper immune cells were still below pre-treatment levels at 6 and 12 months, though within normal limits. No lasting eye damage in a 2–33-year follow-up.
• Very long term: in Finland, 214 children followed 20–35 years — about a quarter of survivors had normal or near-normal intelligence as adults; a third had died, mostly those with a known cause. In Korea, where ACTH is unavailable, about 90% still had intellectual disability after 7.9 years — showing severe outcomes happen without ACTH too.

The Japanese label has no contraindications section but warns chickenpox can be fatal and requires head CT and ECG/echo in children. The US label contraindicates use in infants under 2 with suspected congenital infection (e.g. CMV, cytomegalovirus, a common virus that can infect a baby in the womb), with live vaccines at immunosuppressive doses, uncontrolled high blood pressure and heart failure, and (from October 2026) hypersensitivity, after an open FDA safety signal. Labels outside Japan say not to combine tetracosactide with valproate (severe jaundice in children); the Japanese label does not mention this, yet valproate is often given before ACTH in Japan (PMID 29731400 (Sep 2018)).

Due at 2 months in Japan (they can be given on the same day when the doctor considers it necessary; Japan Pediatric Society schedule, Apr 2026): - Rotavirus ロタウイルス — live, by mouth (Rotarix or RotaTeq) - Hepatitis B B型肝炎 — inactivated - Pneumococcal 小児用肺炎球菌 — inactivated - 5-in-1 五種混合 (diphtheria, whooping cough, tetanus, polio, Hib) — inactivated

Inactivated vaccines are safe during steroid/ACTH treatment in every guideline found; the problem is weaker protection. Best given before hormones start: up to 1 week before (German infantile-spasms guideline), 2 weeks before (UK, US CDC), or from 1 week after (German). The German guideline also says vaccination with inactivated vaccines is possible during hormone therapy, with a possibly weaker response. None of these is a reason to delay ACTH (see the plan in §1).

Rotavirus is the time-critical one (Yonah's dates, born 30 Aug 2026): - First dose recommended by 14 weeks 6 days = 12 December 2026; UK and US rules never allow starting from 13 December. - Japan allows a later first dose after full explanation and with consent, within the routine age window (MHLW notice 感発0331第5号, kanpatsu, 31 Mar 2026 [G]): the last first dose that still completes the course is about 17–18 January 2027 (Rotarix) or about mid-February 2027 (RotaTeq). - Rotarix course must finish by 14 February 2027. - Japan's free catch-up rule excludes rotavirus — if the window passes, it is lost. - Japanese rotavirus labels list immunosuppressive treatment as a caution, not a contraindication; the UK says vaccinating clinically stable hospitalised infants "should not be delayed". But guidelines differ on live vaccines after steroids: wait 1 month (US), 4 weeks (Australia), 8 weeks (German spasms guideline), 3 months (UK general chapter), 6 months (Japanese MR/BCG/prednisolone labels). - In practice: giving rotavirus right before ACTH conflicts with the US, Australian and German 4-week rule. If ACTH runs about 1–29 November, a 4-week wait ends about 27 December, which is too late under US/UK rules but still possible under Japan's late-first-dose rule.

Pertussis (whooping cough) part of the 5-in-1: the US CDC lists "progressive neurologic disorder, including infantile spasms" as a precaution (not a contraindication): defer the DTaP vaccine "until neurologic status clarified and stabilized", decided case by case. In Japan whooping-cough vaccine comes only inside the 5-in-1, so following this would mean deferring the 5-in-1 (the researchers' inference). (A CDC sentence that therapy "should not be delayed because of past administration of vaccines" is about chronic inflammatory conditions; the same paragraph's default is to wait 2 weeks after non-live and 4 weeks after live vaccines before starting immunosuppressive therapy.) Japan treats a history of seizures as "needs caution", not "do not vaccinate".

Later live vaccines: BCG (5–8 months) — the Japanese BCG label forbids it during immunosuppressive treatment, and lists corticosteroids (prednisolone etc.) as not to be combined during, or within 6 months after, long-term or high-dose use (ACTH is not named; applying this to ACTH is an inference); three small Japanese series of ACTH started 2–8 weeks after BCG found no BCG infections (PMID 33838621 (May 2021); PMID 39381543 (Oct 2024)). MR and chickenpox vaccines come at 12 months.

Phenobarbital: no vaccine timing rule found in any label or guideline.

If vaccines are missed because of West syndrome, Kyoto City lists 点頭てんかん as a qualifying illness for free catch-up later (except rotavirus).