#411 - [Journal Club] - 📌 Is Diazoxide Use Increasing for NICU Hypoglycemia?

In this Journal Club episode, Ben and Daphna review a large cohort study from the Journal of Perinatology on the prevalence and safety of diazoxide in the NICU. With neonatal hypoglycemia seemingly on the rise, they discuss off-label use for transient hyperinsulinism and evaluate real-world data from over 340 Pediatrix units. They dive into the rates of concurrent diuretic therapy, respiratory support, and the dreaded risk of pulmonary hypertension. Tune in for a clinical breakdown of when and how this medication is being utilized across centers, plus Ben's echocardiography struggles with cranky term babies on diazoxide!

----

Prevalence and safety of diazoxide in the neonatal intensive care unit. Collins LC, Daniel KB, Tolia VN, Parikh P, Gray KD, Greenberg RG.J Perinatol. 2026 Feb 3. doi: 10.1038/s41372-026-02568-2. Online ahead of print.PMID: 41634357

Transcript

Ben Courchia MD (00:01.528) Hello everybody, welcome back to the Incubator podcast. We're back today for another episode of Journal Club.

Daphna Yasova Barbeau MD (00:08.09) Daphna, good morning. Good morning. I am sticking with the medication theme this week. Most of my papers are regarding medications in the NICU. This is from the Journal of Perinatology. It is called The Prevalence and Safety of Diazoxide in the Neonatal Intensive Care Unit. Lead author Lucas Collins, senior author Rachel Greenberg. This is coming through, as we briefly spoke about in the intro yesterday,

Ben Courchia MD (00:11.905) Okay, this week.

Daphna Yasova Barbeau MD (00:37.115) the Pediatrix Center for Research using their data warehouse. This paper caught my eye because my sense is that our unit has had a real increase in hypoglycemic infants, and diazoxide has come up as a point of conversation for many of these babies. As a reminder for units who don't use a lot of diazoxide—which is a lot of units, as they'll talk about in this paper—it's a non-diuretic benzothiadiazide derivative. It's the only medication approved by the US FDA for infantile hyperinsulinism. It binds to the pancreatic islet cells where it finds the potassium ATP receptors. It hyperpolarizes the cell membrane and prevents the release.

Ben Courchia MD (01:13.4) Thiazide?

Daphna Yasova Barbeau MD (01:34.747) That's why it's a great medication for babies with hyperinsulinism. The FDA approval is limited to a very specific subset of conditions: symptomatic hyperinsulinemic hypoglycemia caused by leucine sensitivity, extrahepatic malignancies, islet cell hyperplasia, islet cell adenoma, and adenomatosis. It's frequently used as a first-line treatment for all types of infantile hyperinsulinemic hypoglycemia. Its most common use, however, is off-label for transient infantile hyperinsulinism. As part of its story, in 2015, the FDA issued a black box warning for the development of pulmonary hypertension in infants receiving diazoxide. Furthermore, the medication has a tendency to reduce free water clearance through sodium retention. It often requires diuretic therapy, which is another potential adverse effect. I think after that happened, units were even more concerned about initiating diazoxide. The study wanted to look at updated data regarding diazoxide use across NICUs, also looking at some of the demographic and clinical factors associated with diazoxide treatment.

Ben Courchia MD (02:45.773) Mm-hmm.

Daphna Yasova Barbeau MD (02:55.899) This is a cohort study. The infants were discharged from any of the 345 NICUs managed by PMG between 2017 and 2022. As I said, all data was extracted from the PMG clinical data warehouse. Let's talk about some definitions as we move in. For hypoglycemia status, I'm chuckling because a lot of it was based on clinician diagnosis. If you made the diagnosis, that's how it was defined. Hypoglycemia was determined based on clinician diagnosis in the EMR. The etiology of hypoglycemia—congenital or transient hyperinsulinism—was also based on clinician documentation. They used the Olsen definitions to classify infants as SGA or LGA. New exposure to diuretics, oxygen, or ventilators was defined as initiation after the start date of diazoxide. Only new exposures occurring at least one day after the initiation of diazoxide and before its discontinuation were included. Infants who had prior exposure to diuretics, oxygen, or mechanical ventilation were included only if they were not exposed on the day of diazoxide initiation. They really wanted to distinguish babies who were needing new respiratory or medication support after the initiation of diazoxide. Multiple diuretics were used, and infants were classified as continuing diazoxide at discharge if they received the medication on the day of or the day before discharge. One of the things they were looking for was pulmonary hypertension. This was also defined by clinician diagnosis documented in the notes or the initiation of a new pulmonary antihypertensive medication like inhaled nitric oxide, sildenafil, or bosentan, during the period between diazoxide initiation and discontinuation. They did not have the echocardiographic data. In terms of results, they had 545,065 infants admitted to the NICU between 2017 and 2022. Of these, 22% were diagnosed with hypoglycemia.

Daphna Yasova Barbeau MD (05:12.122) Among all NICU admissions, not 16%, 0.16%. So less than 0.5% were exposed to diazoxide. Of those exposed to diazoxide, 95% were also diagnosed with hypoglycemia. The percentage of hypoglycemic infants exposed to diazoxide varied widely across the 272 centers that discharged at least 50 hypoglycemic infants during the study period. They've got a figure here. You can see that about 50% of units had 0% diazoxide use, going up to 16% in the units with the most use. The percentage of infants diagnosed with hypoglycemia and the percentage exposed to diazoxide slightly decreased from 2017 to 2022, but these changes were not statistically significant. Actually, it's kind of interesting. These two graphs mirror one another. There was a peak in 2020 and then a downtrend to 2021 and 2022, both for hypoglycemia diagnosis and diazoxide exposure.

Ben Courchia MD (06:18.487) Mm-hmm.

Ben Courchia MD (06:32.748) I know you're describing the pattern of the curve, but these are very mild. Even based on the units on the x-axis. They haven't seen what we've seen. We feel like every kid's hypoglycemic. Maybe it's our machines.

Daphna Yasova Barbeau MD (06:40.032) Right. They're pretty flat. I feel like we've got a bunch right now that are out of proportion. It's not like we put them on a normal amount of D10 and they recover after one bad glucose. These kids are down and out.

Ben Courchia MD (07:04.073) One bolus, yeah.

Daphna Yasova Barbeau MD (07:09.114) Okay, so that's the overtime. For the group of infants that were hypoglycemic, the median gestational age and birth weight in the group exposed to diazoxide were 36 weeks and 2.53 kilos, as compared to those babies who were not exposed, at 36 weeks and 2.62 kilos. The median postnatal age and PMA at diazoxide initiation was nine days and 38 weeks, with a median exposure duration of four days, which I thought was interesting. Documented hyperinsulinism diagnoses also differed between groups. Among the hypoglycemic infants who got diazoxide, 23% were reported to have congenital hyperinsulinism and 16% had transient hyperinsulinism, compared with less than 1% for both diagnoses among the non-exposed hypoglycemic infants. On univariable analysis, the median length of NICU stay was significantly longer for hypoglycemic infants exposed to diazoxide, 23 days compared to nine days. Significant differences were also observed between the two groups. The hypoglycemic infants exposed to diazoxide were more likely to be male (61% versus 56%), more likely to be growth restricted (SGA 30% versus 15%), or LGA (24% versus 19%) compared to those not exposed. There were no significant differences in prenatal steroid exposure or mortality rates between the two groups. They wanted to look at whether infants who had diazoxide were more likely to have diuretics, a ventilator, and oxygen use. Among the hypoglycemic infants treated with diazoxide, 1% were already receiving at least one diuretic prior to initiation of diazoxide, and 26% were started on a diuretic on the same day as diazoxide. These babies were not classified as new exposures. It seemed like the clinical team started diazoxide and the diuretic at the same time, given the risk.

Daphna Yasova Barbeau MD (09:26.298) Additionally, 17% had a pre-existing oxygen requirement and 4% were already on a ventilator at the time of initiation. Again, this respiratory support was not considered part of the new exposure after diazoxide therapy. Of the infants treated with diazoxide, 2% developed pulmonary hypertension at a median of six days during exposure. Another 2% were diagnosed after diazoxide had been discontinued. And 7% carried a diagnosis of pulmonary hypertension prior to diazoxide initiation. Those babies weren't counted in the previous statement. NEC was diagnosed in 2% and seizures occurred in only one infant (less than 1%) following diazoxide initiation. Among those infants exposed to diazoxide, 87% were discharged home, 11% were transferred, and 1% died. Of the infants, 36% were missing an end date for the diazoxide course, which was interesting. 260 were discharged home, 42 were transferred, and three died. It was hard to determine exactly when the medication discontinued or what the cause of discontinuation was.

Ben Courchia MD (10:45.816) So from the database standpoint, you're saying that they didn't really have that information, got it.

Daphna Yasova Barbeau MD (10:48.9) Correct. But there were nearly 500 infants discharged home with known end dates for diazoxide. 11% of those babies were still receiving diazoxide at discharge. 6% of these infants were also receiving diuretics at discharge. The median total duration of diazoxide exposure prior to discharge for those 53 infants was five days, with a range of four to seven days before the decision to discharge home on the medication. Infants discharged home while receiving diazoxide were more likely to have been born at greater than 36 weeks gestation (64% versus 40%), more likely to have a birth weight greater than 2,500 grams (77% versus 51%), and be of white race or ethnicity compared to those who had diazoxide discontinued prior to discharge. They don't speak to this specifically, but I think that makes sense. I think people are most comfortable with older, more mature babies going home on diazoxide. Infants discharged home while receiving diazoxide were less likely to have been exposed to prenatal steroids. There were no significant differences between the groups in other variables examined, including sex or SGA or LGA status. Their overall conclusion...

Ben Courchia MD (11:55.022) Mm-hmm.

Daphna Yasova Barbeau MD (12:10.738) ...is that diazoxide use has been pretty stable across time in US NICUs, though there is much variability in use across centers. Some infants exposed to diazoxide did require additional therapies, most commonly diuretics. Most were able to wean off of diazoxide prior to discharge. And again, we need larger trials to better characterize the efficacy, specifically for the various types of hypoglycemia that we see.

Ben Courchia MD (12:44.386) Very interesting. I'm painfully aware of this data because since starting a hemodynamics fellowship, I get consulted when I'm in Montreal for kids on diazoxide for pulmonary hypertension rule-out. It's a pain because the babies that usually get started on diazoxide are the full-term ones. They're usually very cranky and it's a very hard echo to do. Purely from a technical standpoint, I know that when I see a

Daphna Yasova Barbeau MD (12:50.382) Hmm. Mm-hmm.

Daphna Yasova Barbeau MD (12:55.716) The end.

Daphna Yasova Barbeau MD (13:02.683) Mm-hmm. They don't want you there.

Ben Courchia MD (13:12.878) Yeah. former 35, 36 weeks, now two weeks old, needs an echo... It's like, oh man. I usually try to ask Gabriel to come with me so that he can capture some of the other views, but it's tough. My first reaction to all this was, diazoxide equals more hypertension. I don't know. I didn't know that. There it is. All right, buddy. That was great.

Daphna Yasova Barbeau MD (13:15.287) Thank you for doing that. It's like, man.

Daphna Yasova Barbeau MD (13:34.574) Well, there it is.

Ben Courchia MD (13:42.518) I'll see you tomorrow. Bye.