FDA approves the first treatment for TK2 deficiency, on evidence from 78 matched pairs

On 3 November the Food and Drug Administration approved Kygevvi, a combination of the pyrimidine nucleosides doxecitine and doxribtimine, for thymidine kinase 2 deficiency in adults and children whose symptoms began on or before age 12 [s1]. It is the first approved treatment for the condition. The application, from UCB, was reviewed as a priority new molecular entity [s1].

TK2 deficiency is an ultra-rare, progressive mitochondrial myopathy caused by pathogenic variants in the thymidine kinase 2 gene, in which patients often lose the ability to walk, eat and breathe independently [s2]. Until now there were no approved therapies for it [s2].

What the approval rests on

The interesting part of this approval is its evidence architecture, because a placebo-controlled trial in a disease this rare was never realistic.

FDA's label states that efficacy was established from one Phase 2 study, two retrospective chart review studies, and an expanded access programme, with survival in treated patients compared against an untreated external control group assembled from published literature and from one of the chart reviews [s1]. In total, 82 patients with genetically confirmed TK2d and symptom onset at or before age 12 received Kygevvi or pyrimidine nucleosides [s1]. Seventy-eight of them were matched to untreated patients by age of symptom onset — two years or younger, versus older than two to 12 — to produce 78 matched pairs [s1].

In that comparison, three of 78 treated patients (3.8%) died, against 28 of 78 matched untreated patients (35.9%) [s1]. The hazard ratio for death from treatment start was 0.14 (95% CI 0.04 to 0.39), which the label describes as an approximately 86% reduction in the risk of death (95% CI 61% to 96%) [s1]. Restricted mean survival time at four years post treatment start was 3.8 years in treated patients versus 2.6 years in untreated ones; at ten years, 9.6 versus 5.7 years [s1].

Median age of TK2d symptom onset among treated patients was 1.5 years (range 0.01 to 12), median duration of treatment was four years (range one day to 12 years), and the median dose received was 762 mg/kg/day [s1].

The honest limits

An external control group is not a randomised comparator. The untreated patients came from published literature and from a retrospective chart review, not from a concurrently enrolled arm, which means differences in era of care, referral pattern, diagnostic certainty and reporting bias all sit inside the comparison and cannot be fully adjusted away. Published untreated cases tend to be the more severe ones — that is often why they get written up — and matching on age of symptom onset does not neutralise that.

The peer-reviewed study behind one arm of the evidence package is explicit about where it sits. A multicentre retrospective chart review of pyrimidine nucleos(t)ide therapy in TK2d, published in Neurology on 5 September, compared 38 treated patients with 69 untreated controls drawn from the literature [s2]. None of the 38 treated patients died; 58% (40 of 69) of untreated patients did [s2]. Before treatment, 71.1% of patients (27 of 38) had lost at least one motor milestone and only one of 27 had regained one; during treatment, no patients lost milestones and 65.4% (17 of 26) regained at least one [s2]. Of 21 patients on ventilatory support, 28.6% had their support duration decrease and none increased [s2]. The authors classified their own work as Class III evidence — a designation that means exactly what it appears to mean about the strength of causal inference available here [s2].

That is a consistent, biologically plausible, large effect measured with a study design that cannot prove it. Regulators accepted it because the alternative in an ultra-rare fatal disease is no treatment at all.

How it is used, and what to watch

Kygevvi is a powder for oral solution, given in three equally divided doses with food [s1]. Dosing starts at 260 mg/kg/day, titrates to an intermediate level of 520 mg/kg/day, and reaches a maintenance dose of 800 mg/kg/day, with a minimum of two weeks at each level before moving up [s1]. It must be used only with the ZX2000 administration kit [s1]. There are no contraindications [s1].

Two safety areas carry specific label instructions. Baseline alanine and aspartate aminotransferase and total bilirubin must be measured before starting treatment, and if signs of liver injury appear, treatment is interrupted until levels return to baseline or stabilise; the label notes that in one of the supporting studies two patients permanently discontinued after elevated liver enzymes recurred on rechallenge at a reduced dose [s1]. Liver enzymes and bilirubin are monitored yearly thereafter [s1]. Separately, diarrhoea and vomiting leading to hospitalisation, dose reduction and permanent discontinuation were reported, and the label directs dose reduction or interruption based on severity [s1].

The most common adverse reactions, at an incidence of 5% or more, are diarrhoea, abdominal pain, vomiting, and increased ALT and AST [s1]. In one of the retrospective studies, six of 18 treated patients (33%) discontinued because of an adverse reaction [s1].

This article is informational and is not medical advice; treatment decisions in TK2 deficiency belong with the specialist centres that manage it.

Sources

• [s1] KYGEVVI (doxecitine and doxribtimine) powder, for oral solution — Prescribing Information, U.S. Food and Drug Administration, issued November 2025. https://www.accessdata.fda.gov/drugsatfda_docs/label/2025/219792s000lbl.pdf
• [s2] Pyrimidine Nucleos(t)ide Therapy in Patients With Thymidine Kinase 2 Deficiency: A Multicenter Retrospective Chart Review Study, Neurology, 5 September 2025. https://doi.org/10.1212/WNL.0000000000213908