Dear Editor,

Hyperinsulinaemic hypoglycaemia (HH) is a rare condition with elevated and unregulated levels of insulin, resulting in low blood glucose concentrations. It is the most common cause of persistent hypoglycaemia in infants and children, causing a high risk of developing brain injuries such as epilepsy, cerebral palsy, or neurological impairment.¹ Diazoxide remains the first-line medication used to treat HH.¹ It binds to the SUR1 subunit of KATP channels to inhibit β-cell depolarisation and thus insulin secretion. However, it is important to note that fluid retention, hypertrichosis, and feeding problems are common side effects of diazoxide. Rare severe consequences can also occur like pulmonary hypertension and congestive heart failure. Apart from a single study reporting the use of low-dose diazoxide in small for gestational age (SGA) infants, diazoxide has been reported to be used in doses of 5−20 mg/kg/day.²

Chandran et al. highlighted the efficacy and safety of using ≤5 mg/kg/day of diazoxide in 27 SGA babies.² Twenty-six (97%) of these passed a fasting study before discharge from the hospital, establishing normal glucose control on low-dose diazoxide. Furthermore, diazoxide was discontinued at a median age of 63 days, and resolution of HH was confirmed in 26/27 (96%) infants on passing a fasting study. Their study benefits from a robust trial protocol with a modest sample size considering that HH is a rare disease and that Singapore has a small population.

Similarly, we conducted a retrospective analysis evaluating the effectiveness and outcomes of using low-dose diazoxide (≤5 mg/kg/day) that have successfully managed HH. We have identified 34 patients with biochemically confirmed HH that were treated with low-dose diazoxide at two tertiary children hospitals in London from April 2020 to March 2023. Patient characteristics (birth weight and gestational age) and treatment details were collected from electronic patient records. For the comparative analysis, the patients were stratified into two groups based on their birth weight: SGA and non-SGA (appropriate for gestational age–AGA and large for gestational age–LGA). SGA was defined as birth weight <10th centile. The patients were also differentiated by their gestational age: preterm (<37 weeks) and term (≥37 weeks). Patient outcomes that were assessed included: (1) median age of starting and stopping treatment; (2) median dosage of diazoxide on discharge from hospital; and (3) follow-up outcomes: side effects from diazoxide, adjustment of diazoxide dose, and neurodevelopmental outcomes.

Of the 34 infants, 15 were SGA and 19 were non-SGA. The patient characteristics and details of diazoxide treatment are summarised in Table 1. All babies had their total fluid volume adjusted to approximately 130 mL/kg/day when initiating diazoxide. All infants underwent an age appropriate controlled fast successfully before discharge from the hospital, showing normal glucose control whilst on diazoxide.

We did not gain ethical approval specifically to start diazoxide as is standard of care, however, every patient was reviewed by the hospital pharmacist who agreed to the use of low-dose diazoxide. Furthermore, a general consensus exists on the use of a lower dose of diazoxide which is used in many centres and is reported in the United Kingdom consensus guidelines (due to be published soon). This would be the first published report on the efficacy and outcomes from the use of low dose diazoxide in these cohort of patients.

During follow-up, there were 6 (17.6%) babies who had documented episodes of hypoglycaemia (i.e., ≤3.5 mmol/L); one each had three and two episodes of hypoglycaemia respectively, and the remaining four babies had one episode. The lowest documented blood glucose reading was 3.3 mmol/L. Three babies in the SGA group (two preterm) and three babies in the non-SGA group. All patients required only once increment of diazoxide <1 mg/kg/day in view of these observations, with the total daily dosage still remaining ≤5 mg/kg/day given the infant's weight gain. It is worth noting that there were no reports of symptomatic hypoglycaemia. Five (14.7%) infants were identified to have minimal fluid retention during follow-up which resolved on increasing the dose of diuretics. Two infants had a genetic confirmation of HNF4A mutation, done in view of the strong family history of diabetes. No significant neurodevelopmental concerns have been identified on follow-up so far.

HH may be transient or persistent, lasting from a few days to being life-long.¹ The patient cohort that responded to low dose diazoxide in our study had mostly transient HH, with the median age of stopping diazoxide being 4 months. These observations are encouraging to note as some babies may only need to be on medications for a few months, lowering the risk of the side effects from diazoxide.

Our study is the first report to show the effectiveness of low-dose diazoxide in stabilising blood glucose across all groups (SGA, non-SGA, preterm, and term). In addition, we report two cases of congenital HH caused by a mutation of HNF4A. The HNF4A gene encodes for the transcription factor hepatocyte nuclear 4a which controls the expression of genes involved in insulin secretion.³ The sibling of one of these patients has been reported previously and was known to have HH that was very sensitive to diazoxide treatment, requiring lower doses⁴ and developed hyperglycaemia on traditional doses of diazoxide. It may be that this particular mutation in HNF4A (p.Ser419Ter; c.1256C>G) leads to exceptional sensitivity to diazoxide or that mutations in HNF4A in general require smaller doses of diazoxide. Notably, HNF4A is associated with maturity-onset diabetes of the young (MODY). These children will need to be continuously monitored for signs of diabetes so that early treatment can be commenced.

To conclude, low-dose diazoxide can be an effective treatment for babies with HH, independent of birth weight, and may avoid prolonged stay in the hospital. Certain genetic forms of HH may also be suitable for treatment with low dose diazoxide. Hence, lower doses of diazoxide should be considered in infants with HH before using traditionally published doses (>5 mg/kg/day) with close and frequent blood glucose monitoring. Although generally well-tolerated, fluid retention can develop in a minority of patients, confirming a need for regular follow-ups and vigilance even at lower doses. The use of low-dose diazoxide would benefit from a larger cohort and multicentre study.

亲爱的编辑，

高胰岛素性低血糖 (HH) 是一种罕见的疾病，胰岛素水平升高且不受调节，导致低血糖浓度。它是婴儿和儿童持续低血糖的最常见原因，导致发生癫痫、脑瘫或神经功能障碍等脑损伤的高风险。¹二氮嗪仍然是治疗 HH 的一线药物。¹它与KATP通道的 SUR1 亚基结合，抑制 β 细胞去极化，从而抑制胰岛素分泌。然而，值得注意的是，液体潴留、多毛症和喂养问题是二氮嗪的常见副作用。罕见的严重后果也可能发生，如肺动脉高压和充血性心力衰竭。除了一项报告在小于胎龄 (SGA) 婴儿中使用低剂量二氮嗪的研究外，据报道二氮嗪的使用剂量为 5−20 毫克/公斤/天。²

钱德兰等人。强调了对 27 名 SGA 婴儿使用 ≤ 5 毫克/公斤/天二氮嗪的有效性和安全性。²其中二十六人 (97%) 在出院前通过了禁食研究，建立了低剂量二氮嗪的正常血糖控制。此外，二氮嗪在中位年龄 63 天时停用，并且通过禁食研究后，26/27 (96%) 的婴儿证实 HH 得到缓解。考虑到 HH 是一种罕见疾病且新加坡人口较少，他们的研究受益于稳健的试验方案和适度的样本量。

同样，我们进行了一项回顾性分析，评估使用低剂量二氮嗪（≤5 mg/kg/天）成功控制 HH 的有效性和结果。我们确定了 2020 年 4 月至 2023 年 3 月期间在伦敦两家三级儿童医院接受低剂量二氮嗪治疗的 34 名生化确诊的 HH 患者。患者特征（出生体重和孕龄）和治疗详细信息从电子患者记录中收集。为了进行比较分析，根据出生体重将患者分为两组：SGA 和非 SGA（适合胎龄 – AGA 和大胎龄 – LGA）。SGA 定义为出生体重<10%。患者还根据胎龄进行区分：早产（<37 周）和足月（≥37 周）。评估的患者结果包括：（1）开始和停止治疗的中位年龄；(2) 出院时二氮嗪的中位剂量；(3) 随访结果：二氮嗪的副作用、二氮嗪剂量的调整和神经发育结果。

在 34 名婴儿中，15 名是 SGA，19 名是非 SGA。表 1 总结了患者特征和二氮嗪治疗细节。开始二氮嗪治疗时，所有婴儿的总液体量调整至约 130 mL/kg/天。所有婴儿在出院前都成功地进行了适当年龄的禁食控制，在服用二氮嗪期间显示出正常的血糖控制。

我们没有获得伦理批准专门开始使用二氮嗪作为护理标准，但是，每位患者都经过医院药剂师的审查，同意使用低剂量二氮嗪。此外，对于使用较低剂量的二氮嗪存在普遍共识，许多中心都使用这种剂量，并在英国共识指南中进行了报告（即将发布）。这将是第一份关于在这些患者群体中使用低剂量二氮嗪的疗效和结果的报告。

随访期间，有 6 名（17.6%）婴儿出现过低血糖（即≤3.5 mmol/L）；其中 1 名婴儿分别出现 3 次和 2 次低血糖，其余 4 名婴儿分别出现 1 次低血糖。记录的最低血糖读数为 3.3 mmol/L。SGA 组有 3 个婴儿（2 个早产），非 SGA 组有 3 个婴儿。鉴于这些观察结果，所有患者仅需要一次<1 mg/kg/天的二氮嗪增量，考虑到婴儿的体重增加，每日总剂量仍保持≤5 mg/kg/天。值得注意的是，尚无症状性低血糖的报告。随访期间发现 5 名（14.7%）婴儿有极少的液体潴留，最终解决了增加利尿剂剂量的问题。鉴于糖尿病家族史明显，两名婴儿进行了HNF4A突变基因确认。迄今为止，随访中尚未发现重大的神经发育问题。

HH 可能是暂时性的，也可能是持续性的，从几天到终生不等。¹在我们的研究中，对低剂量二氮嗪有反应的患者队列大多有短暂性 HH，停止二氮嗪的中位年龄为 4 个月。这些观察结果令人鼓舞，因为一些婴儿可能只需要服药几个月，从而降低了二氮嗪副作用的风险。

我们的研究是第一份显示低剂量二氮嗪在稳定所有群体（SGA、非 SGA、早产儿和足月）血糖方面的有效性的报告。此外，我们报告了两例由HNF4A突变引起的先天性 HH 病例。HNF4A基因编码转录因子肝细胞核 4a，控制与胰岛素分泌有关的基因的表达。³其中一名患者的兄弟姐妹之前曾有报道，已知其患有 HH，对二氮嗪治疗非常敏感，需要较低剂量⁴，并且在传统剂量的二氮嗪治疗后出现高血糖。HNF4A中的这种特殊突变（p.Ser419Ter；c.1256C>G）可能会导致对二氮嗪异常敏感，或者HNF4A中的突变通常需要较小剂量的二氮嗪。值得注意的是，HNF4A与青少年发病的成人糖尿病 (MODY) 相关。需要持续监测这些儿童是否有糖尿病迹象，以便尽早开始治疗。

总之，低剂量二氮嗪可以成为 HH 婴儿的有效治疗方法，与出生体重无关，并且可以避免长时间住院。某些遗传形式的 HH 也可能适合用低剂量二氮嗪治疗。因此，对于 HH 婴儿，在使用传统公布的剂量（> 5 mg/kg/天）之前，应考虑降低二氮嗪的剂量，并密切频繁地监测血糖。尽管通常耐受性良好，但少数患者可能会出现液体潴留，这表明即使在较低剂量下也需要定期随访和保持警惕。低剂量二氮嗪的使用将受益于更大规模的队列和多中心研究。