Vigabatrin for Infantile Spasms in Tuberous Sclerosis, Against a Visual Field Risk No Infant Can Report
A single infant with tuberous-sclerosis-associated infantile spasms. Vigabatrin is the most effective option for exactly this cause — its visual field risk is real, permanent, and unmeasurable in a patient this young.
Baby A.M., a 5-month-old girl, was diagnosed with tuberous sclerosis complex shortly after birth, following cardiac rhabdomyomas found on a prenatal ultrasound, and began having clusters of sudden flexor spasms three weeks ago — brief, repeated jackknifing movements her parents at first mistook for an exaggerated startle reflex, until a pediatrician recognized the pattern and an EEG confirmed hypsarrhythmia, establishing infantile spasms. Her parents, already managing a diagnosis they'd never heard of five months ago, have spent the weeks since her spasms began learning an entirely new vocabulary — hypsarrhythmia, cortical tubers, subependymal nodules — largely from a TSC-specific parent advocacy organization that connected them with other families within days of her diagnosis.
Vigabatrin and ACTH are the two standard first-line treatments for infantile spasms, and two separate bodies of evidence bear on A.M. — worth keeping apart, because they are often run together. The TSC-specific data, going back to Chiron et al. (1997) and the case series after it, show spasm cessation in TSC-associated spasms far above what vigabatrin achieves in infantile spasms generally, which is why guidelines name it first-line for these infants specifically. Separately, ICISS (O'Callaghan et al., 2016) showed across infantile spasms broadly that adding vigabatrin to hormonal therapy stopped spasms faster than hormonal therapy alone — though its own 18-month follow-up found no developmental or epilepsy-outcome advantage for the combination. It is the TSC evidence, not ICISS, that makes vigabatrin her drug. What vigabatrin also does, in a meaningful proportion of infants and children exposed for more than a few months, is cause permanent, irreversible bilateral peripheral visual field constriction — a real, well-documented risk that led to its own REMS program and a requirement for regular ophthalmologic monitoring throughout treatment.
The monitoring requirement runs directly into the limits of what a five-month-old can do. Formal visual field testing requires cooperation and communication no infant this age can provide, which leaves electroretinography and other indirect measures standing in for it. The REMS program is built around a test A.M. will not be able to take for years, in precisely the population where the drug's efficacy case is strongest.
Choosing first-line therapy for TSC-associated spasms
Vigabatrin, and I'd start it promptly rather than trial hormonal monotherapy first. Two distinct bodies of evidence point here and it's worth keeping them apart. The TSC-specific data — going back to Chiron et al. (1997) and the case series after it — show spasm cessation rates in TSC-associated spasms far above what vigabatrin achieves in infantile spasms generally, which is why guidelines name vigabatrin first-line specifically for these infants. Separately, ICISS (O'Callaghan et al., 2016) showed in infantile spasms broadly that vigabatrin added to hormonal therapy stopped spasms faster than hormonal therapy alone. I'd be honest that ICISS's 18-month follow-up found no developmental or epilepsy-outcome advantage for the combination — but time to spasm control still matters, and it's the TSC data, not ICISS, that make vigabatrin the drug for her.
I don't dispute the efficacy data, and I understand why time to control matters developmentally. But the REMS monitoring this drug's approval is built around assumes a patient who can undergo formal visual field testing — she can't, at five months, and won't be able to for years. We're prescribing a drug whose defining risk we can't actually measure in the population where we're now saying it works best.
Framing this as 'monitoring per REMS' when the actual test the REMS program specifies isn't available in an infant this age overstates how much real surveillance is happening — ERG is a real, useful surrogate, but it isn't the same test, and that gap deserves to be named plainly, not folded into a routine monitoring reassurance.
Both are right, and I don't think the monitoring gap changes which drug gives her the best real chance at spasm control — it changes what her parents need to understand about what “monitored” actually means here. Start vigabatrin — which the TSC-specific evidence supports on its own terms — alongside hormonal therapy, with ERG and available surrogate testing at the intervals the REMS program specifies, explained to her parents explicitly as an imperfect but real proxy, not equivalent to the formal test the program was designed around.
And set a firm, calendar-based reassessment of ongoing need once spasm control is achieved — vigabatrin exposure duration is itself a real driver of visual risk, so the moment continued use stops being clearly necessary is the moment to seriously revisit whether it's still the right drug to be on, not something to leave open-ended by default.
Agreed: vigabatrin started alongside ACTH per the TSC-specific evidence, with ERG and available surrogate visual monitoring at REMS-specified intervals, explicitly explained to her parents as a real but imperfect proxy for the formal testing she can't yet undergo. A calendar-based reassessment of ongoing vigabatrin need was set for once spasm control is confirmed, rather than left open-ended.
Not agreed: how frequently to escalate to more invasive or specialized visual-pathway assessment (available at a small number of tertiary centers, requiring sedation at her age) if surrogate monitoring stays inconclusive — the Ophthalmologist wanted a lower threshold for pursuing it; the Pediatric Neurologist was more cautious about subjecting an infant to sedation-requiring testing without a clearer indication. Left for reassessment once the first surrogate results are in hand.