Neonatal hypoxic-ischemic encephalopathy (HIE) remains the leading cause of death and lifelong disability among newborns. However, therapeutic hypothermia, the only established postnatal treatment, improves rates of survival without disability, yet death or disability still occurs and it offers no benefit once its narrow therapeutic window has closed or where the cooling infrastructure is unavailable [1]. Thus, there is growing interest in preventing or at least preconditioning against brain injury before it occurs. Unfortunately, antenatal conditioning to prevent HIE is underexplored compared with other postnatal neuroprotection strategies (Table 1) [2-6].

Antenatal magnesium sulfate is the only intervention with high-certainty evidence of fetal neuroprotection; however, its benefit is limited to reducing cerebral palsy or death up to 2 years of corrected age after a very preterm birth [2]. However, beyond 2 years of corrected age, little or no difference was found compared to the placebo group. Melatonin, allopurinol, and creatine have antioxidant, antiapoptotic, and energy-buffering effects in animal models of perinatal HIE; however, none have demonstrated a clinically meaningful reduction in the incidence of HIE when administered to mothers. Moreover, most studies focused primarily on postnatal administration [3-5,7].

Parallel efforts have been made to identify placental dysfunction and impending fetal compromise using Doppler indices, circulating biomarkers, and artificial intelligence–assisted cardiotocography to optimize timing and mode of delivery. However, these approaches may prompt iatrogenic preterm delivery, and prospective evidence of HIE prevention remains limited [1].

In recent years, metformin has emerged as a potential neurorestorative and neuroprotective agent supported by a growing body of preclinical evidence [7]. Metformin indirectly activates AMP-activated protein kinase (AMPK), which may regulate autophagy, mitochondrial homeostasis, and inflammatory responses, and has extensive clinical use during pregnancy, oral availability, and low cost, making it an attractive candidate for repurposing as an antenatal neuroprotective agent [8,9]. In this issue, Kim et al. [6] provide important preclinical evidence supporting the potential role of metformin-mediated metabolic preconditioning. Using a rat model, the authors showed that maternal metformin exposure during late gestation preserved auditory brainstem responses and partially rescued neuronal and oligodendrocyte-lineage populations. A preliminary unquantified western blot finding was consistent with the partial preservation of AMPK signaling, although this proposed mechanism requires further confirmation.

Metformin has several advantages over the other candidate agents listed in Table 1. First, unlike those other agents, metformin has been used extensively for gestational diabetes and polycystic ovary syndrome for decades, with its well-characterized pharmacokinetics and placental transfer [8,9]. Second, metformin showed therapeutic potential for neonatal HIE in a small number of animal studies [10]. Third, its mechanism, AMPK-dependent metabolic and anti-inflammatory modulation, targets multiple stages of the energy-failure cascade underlying HIE rather than a single pathway. Finally, metformin is inexpensive, orally administered, and widely available. In particular, its low cost and oral administraAntenatal tion make its use feasible even when therapeutic hypothermia is limited or unavailable, especially in low- and middle-income countries [1].

However, the study of Kim et al. [6] has notable limitations. First, the AMPK western blot analysis was performed in a single animal per group and not statistically tested. No sham or uninjured metformin-exposed group was included to exclude direct effects on normal auditory development. Moreover, the maternal and offspring safety parameters of metformin were not systematically recorded, which is important because gestational metformin exposure has been associated with altered maternal-placental- fetal metabolism in nondiabetic pregnancies in addition to growth and obesity [8,9]. While metformin exposure does not appear to affect serious metabolic or neurodevelopmental outcomes of the offspring, long-term human data remain limited [8,9]. In addition, antenatal metformin was started on gestational day 16. However, the risk of HIE is usually recognized only shortly before or even during labor, making its clinical implementation a major challenge.

Several questions must be addressed before antenatal metformin administration can be translated into clinical practice. First, does metformin retain its efficacy when started acutely or closer to delivery rather than from mid- to late gestation? Two complementary studies could address this issue. Targeted short-course prophylaxis could be evaluated in pregnancies with objectively identifiable antenatal risk, such as severe fetal growth restriction or placental insufficiency. Alternatively, intrapartum rescue dosing can be initiated based on predefined evidence of fetal compromise.

Second, can optimal dosing and placental pharmacokinetics be established for its antenatal administration? A phase II clinical study of postnatal metformin in 3- to 6-month-old infants who received therapeutic hypothermia for neonatal HIE is ongoing (NCT06429007). This study is expected to provide valuable infant pharmacokinetic data that would facilitate antenatal administration, although its postnatal design cannot directly address maternal dosing, placental transfer, or antenatal safety.

In summary, antenatal metformin is a mechanistically distinct, clinically relevant, and inexpensive candidate for antenatal neuroprotection. The immediate priority is to determine whether short-term maternal treatment close to delivery provides reproducible neuroprotection in clinically relevant high-risk models and establish maternal- fetal safety, placental pharmacokinetics, and compatibility with postnatal therapeutic hypothermia.