🧬 Genomics
A $2.7M Gene Therapy Buys Two Fewer Cornstarch Doses a Day, and for One in Three, the Night One.
The first gene therapy for glycogen storage disease type Ia removes about two of six daily cornstarch doses and, in uncontrolled 96-week data, the night dose for a third; its $2.7 million price is 92 years of the disease's measured excess medical cost.
Six doses a day, one of them in the middle of the night: that is the uncooked cornstarch that keeps a person with glycogen storage disease type Ia alive, because a missing liver enzyme blocks the release of stored glucose and a skipped dose can end in dangerous hypoglycemia. On August 19 the FDA approved Genglycos, the first therapy for patients 8 and older, and Ultragenyx priced it at $2.7 million. In the trial behind the approval, treated patients cut daily cornstarch 31 points more than placebo, about two of six doses.
This disease has a measured cost, and a 2025 claims analysis put mean annual healthcare spending for GSDIa patients at $33,910 against $4,410 for matched comparators, an excess of $29,500 a year. About ninety-two years of that excess is the list price. And the drug is approved only to reduce cornstarch, as an adjunct to nutritional management: the label makes no claim of fewer hypoglycemic episodes, of liver or kidney benefit, or of longer survival.
Where the Two Doses Come From
A review of 115 patients aged 10 to 62 found a mean of six doses a day at every age, and Genglycos, one infusion of an AAV8 vector carrying the G6PC gene into the liver, was tested in its trial's 48-week randomized phase. Twenty treated patients reduced daily cornstarch by 41.3 percent, 24 placebo patients by 10.3 percent. Subtract one from the other: 31 points, or 1.9 of six doses. Two doses a day, bought once.
At week 96, after the blinded crossover, the average reduction was 61 percent, and 33 percent of the original treatment group and 42 percent of the crossover group had stopped night-time dosing, a result the company reports came "while maintaining glycemic control". By week 96 everyone had the drug. They are uncontrolled figures; only the 31 points carry a placebo comparison.
What a Dose Costs, and for How Long
Durability so far runs four to six years. At different doses, 12 adults followed that long, 5.3 years on average, kept a mean 70 percent cornstarch reduction with no treatment-related serious adverse events in that cohort, and against that durability $2.7 million spread over 1.9 doses a day costs $970 an avoided dose at four years, $650 at six. Assume forty years and it costs $97.
| Quantity | Number | Against | Result | Basis |
|---|---|---|---|---|
| List price | $2.7 million | $29,500 a year of measured excess medical cost | About 92 years | Measured (claims analysis, 2016-20 dollars) |
| List price | $2.7 million | 1.9 doses a day avoided over 4 to 6 years | $970 per dose at 4 years, $650 at 6 | Observed durability, n=12 at different doses |
| List price | $2.7 million | 1.9 doses a day avoided over 40 years | $97 per dose | Assumption |
| Americans with GSDIa (company estimate) | 1,500 to 2,500 | each at the $2.7 million list price | $4.1 to $6.8 billion at list, a ceiling | Company estimate, all ages; label starts at age 8 |
| Consensus peak sales, 2033 | $294 million a year | $2.7 million per patient | About 109 patients a year | Analyst consensus at list; any net price would raise the count |
Less Starch, Not Fewer Lows
Accelerated approval rests on a surrogate, here cornstarch intake, and the benefit it stands in for, fewer dangerous drops in blood glucose, was not shown. TechTimes reported that the share of glucose readings below 70 milligrams per deciliter ran about 3 points higher in treated than in placebo patients at 48 weeks, a secondary account we could not check and not something 44 patients can resolve. Patients ate less starch, but whether they were safer, or less safe, is not established in anything we read, and until a confirmatory program measures time to hypoglycemia during a controlled fast, the product buys hours between doses.
The Site's Own Cure Math
A cure has to beat two numbers, what the drug it replaces would have earned and what the health it buys is worth, and this site put figures on both on September 5. On revenue, a one-time $356,000 cure earns its maker the same net present value as twenty years of a $60,000-a-year drug, assuming a 10 percent discount rate and a patient pool cleared in three years, or 5.9 times that drug's annual price. On value, eight extra quality-adjusted life years justify about $1.2 million.
Genglycos replaces no $60,000 drug, because Glycosade, the starch it displaces, retails near $100 a container, and the revenue yardstick does not apply, so run instead a 20-year annuity at 10 percent, a factor of 8.51, on the disease's measured excess cost: erasing every excess dollar is worth about $251,000. Genglycos costs 10.8 times that, is not approved to erase any of it, and whether it does is unmeasured, while on the value yardstick it costs 2.25 times the $1.2 million ceiling and claims no quality-adjusted life years at all.
Scale is the case for the price: 1,500 to 2,500 Americans and a Visible Alpha consensus compiled by S&P Global of $294 million a year in 2033. Over the observed durability, each removed dose costs $650 or more. A spoonful of starch. The drug sells avoided alarms, and no claims database prices those.
What You Can Do
Families weighing the infusion: the strongest published number is night-dose elimination in 33 to 42 percent of treated patients in uncontrolled 96-week data, and the weakest is hypoglycemia, where no benefit was shown and no safety finding was established either way. Ask the treating center which of those two the patient's case turns on and what the trial's own adverse events were, and whether the plan's coverage is tied to the confirmatory trial's fasting test. This is not medical advice; the decision belongs with your metabolic team.
Limitations
Every figure came through search summaries, not primary pages or the full paper. Because the trial measured cornstarch by amount, converting 31 points into 1.9 of six doses assumes equal doses dropped whole, and the dose count was a secondary endpoint we could not obtain, so 1.9 is our derivation. Six doses is one cohort's mean, range four to eight, held at every age; the amount per dose falls with age.
Our per-dose figures splice the phase 3 placebo-adjusted result with phase 1/2 follow-up, and the 3-point hypoglycemia difference is TechTimes's reading, not a tested superiority claim. Our excess-cost figure comes from a US commercial-claims database for 2016 to 2020, is not inflated to 2026 dollars, excludes uninsured and publicly insured patients, and a 20-year horizon at the earlier story's 10 percent rate truncates a lifelong disease. Genglycos's QALY gain is unmeasured, not zero, so the 2.25x compares the price with a benefit nobody has sized, and we approached neither the Association for Glycogen Storage Disease nor Ultragenyx, nor any payer or treating clinician.
The Strongest Case Against This Analysis
Cornstarch doses are the wrong unit and excess medical cost is the wrong denominator: 33 to 42 percent of treated patients lost the night dose, a gain no claims database captures. A 3-point difference in low readings among 44 patients is noise, the endpoint was designed as a non-inferiority test and no failure has been reported, and $2.7 million sits inside the range payers accept, from Zolgensma at $2.1 million to Hemgenix at $3.5 million. Most of that is right. We disagree on what the price is for: the label says cornstarch reduction, the payer pays for outcomes, and the outcome that would justify the number is the confirmatory trial's to show. Durability may yet vindicate the $97 row, which is an assumption, and the table labels it one.
The Bottom Line
Two doses, 92 years. Placebo-controlled, Genglycos takes about two of six daily cornstarch doses from the average patient; uncontrolled at 96 weeks, 3.7 of six, with 33 to 42 percent losing the night dose. That price is 92 years of the disease's measured excess cost, 10.8 times a 20-year annuity on it, and 2.25 times this site's value ceiling for eight good years, which Genglycos does not claim. To the families who get it, it may be worth every dollar; to everyone else, that turns on a number nobody has measured.