Background Frequent and severe gastrointestinal disturbances have been reported with the use of diazoxide in adults and older children. However, no studies have investigated the incidence of necrotising enterocolitis (NEC) in diazoxide-exposed newborns.
Objective To evaluate a possible association between diazoxide treatment for neonatal hypoglycaemia and the occurrence of NEC.
Design Multicentre retrospective cohort study.
Setting Three tertiary neonatal intensive care units in Toronto, Canada.
Patients All patients treated with diazoxide for persistent hypoglycaemia between July 2012 and June 2017 were included. Overall incidence of NEC during those years on the participating units was obtained for comparison from the Canadian Neonatal Network database.
Main outcome Incidence of NEC after diazoxide exposure.
Results Fifty-five neonates were exposed to diazoxide during the study period. Eighteen patients (33%) showed signs of feeding intolerance, and 7 developed NEC (13%). A diagnosis of NEC was more prevalent in the diazoxide-exposed, as compared with non-exposed infants of similar gestational age (OR 5.07, 95% CI 2.27 to 11.27; p<0.001), and greatest among infants born at 33–36 weeks’ gestation (OR 13.76, 95% CI 3.77 to 50.23; p<0.001). All but one of the neonates diagnosed with NEC developed the disease within 7 days from initiation of diazoxide treatment.
Conclusion The present data suggest a possible association between diazoxide exposure and the development of NEC in neonates. Further evaluation of the diazoxide-associated risk of NEC in neonates treated for persistent hypoglycaemia is warranted.
- intensive care
Data availability statement
Data are available upon reasonable request.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
What is already known on this topic?
Diazoxide gastrointestinal side effects are frequent in children and adults. However, there is a paucity of data regarding its effect on newborns.
What this study adds?
This study reports a possible association between diazoxide treatment for hypoglycaemia and the onset of necrotising enterocolitis.
Prolonged neonatal hypoglycaemia is a common condition in infants of a diabetic mother and small for their gestational age. In infants with congenital disorders such as nesiodioblastosis and Beckwith-Wiedemann syndrome, hypoglycaemia can be persistent.1 Although new therapies have been proposed, glucagon, diazoxide and octreotide are the current pharmacological options to treat persistent hypoglycaemia in neonates.2–4
Diazoxide is a potassium channel activator that impedes beta-cell depolarisation and, therefore, lessens insulin secretion and hypoglycaemic events.5 Since its introduction in the 1960s, it has remained the first-line pharmacological agent for the treatment of congenital hyperinsulinism in multiple countries. Hypertrichosis, fluid retention and pulmonary oedema are well-known diazoxide side effects6; however, the safety profile of this drug has not been completely investigated in newborns.
Necrotising enterocolitis (NEC) is one of the leading gastrointestinal complications in neonates. Although more frequent in preterm infants, NEC also occurs in term and near-term babies. The different clinical presentations, severities and outcomes have led the belief that NEC is not a single entity but rather the result of distinct offending factors.7 This neonatal disease is associated with high morbidity and mortality rates.
Octreotide usage was reported to be associated with NEC.8–13 The aetiology of NEC in neonates receiving this drug has been linked to its vascular side effects, given its reducing action on splanchnic blood flow.14 Interestingly, most of the octreotide-exposed reported cases of NEC had also received diazoxide before NEC was diagnosed.8–13
Frequent gastrointestinal disturbances have been reported in both adults and infants exposed to diazoxide.15–18 Nonetheless, to our knowledge, a link between diazoxide and NEC has not been studied.
This study aimed to describe the gastrointestinal effects of diazoxide-exposed infants, compared with similar postnatal and gestational age neonates not therapeutically exposed to the drug. We hypothesised that diazoxide exposure may increase the risk for NEC in neonates.
We conducted a retrospective cohort study. Patients treated with diazoxide for persistent hypoglycaemia admitted to one of the neonatal intensive care units (NICUs) in Toronto (The Hospital for Sick Children, Mount Sinai Hospital and Sunnybrook Health Science Centre) were included. There were no exclusion criteria. The study period comprised 1 July 2012–30 June 2017. At the time the study began, the institutional protocol recommended that diazoxide be started at a dose of 10 mg/kg/day divided in three doses.
Data collection included gestational age, birth weight, comorbidities, type of milk, fortification and/or energy supplementation, underlying cause of hypoglycaemia, mean postnatal age of exposure, initial dose, glucose infusion rate at initiation of diazoxide, incidence of gastrointestinal side effects, dose and time frame between exposure and side effects, exposure to octreotide, as well as their clinical outcomes. The overall incidence of gestational age-specific NEC for the Toronto NICUs during the study period was obtained from the Canadian Neonatal Network (CNN) database.
We defined diazoxide exposure as an infant who received at least one dose of the drug. Re-exposure was defined as reinitiation of diazoxide therapy 24 hours after medication discontinuation.
Feeding intolerance was defined as any gastrointestinal sign leading to feeds being held at least once. NEC was defined as radiological or ultrasonographic findings (pneumatosis, portal venous gas and pneumoperitoneum) in a patient with abnormal abdominal clinical signs. NEC staging was assigned as per Bell’s modified classification.19
Descriptive data are presented as medians (IQRs), means (ranges) or percentages, depending on the variable. When data are missing, we report the incidence according to the available data denominator. Univariable analysis was performed using the Mann-Whitney test for continuous quantitative variables or Fisher exact test for categorical variables. A p value of <0.05 was considered statistically significant. Statistical analysis was performed with STATA V.13.1 statistical software.
Over the study period, there were 15 460 neonates, of whom 447 (3%) had NEC and 55 (0.4%) were treated with diazoxide. In the diazoxide-exposed group, 35 patients (64%) were premature (<37 weeks’ gestation) and 25 (45%) were low-birthweight infants. In 16 cases (29%), being small for gestational age was believed to be the cause of the prolonged hypoglycaemia. Two infants had nesidioblastosis (4%) and two had prolonged hypoglycaemia after postnatal dexamethasone exposure (4%). In 35 patients (64%), the cause of the hypoglycaemia remained unknown at discharge. The main demographic and clinical data are summarised in table 1.
Data regarding the clinical conditions associated with hypoglycaemia and its treatment are provided in table 2.
Seven (13%) patients developed NEC after diazoxide treatment. The mean time between the beginning of diazoxide treatment and the onset of NEC was 5 days (range 1–12 days). In one neonate, abdominal signs recurred after the second course of diazoxide was initiated. This infant required prolonged fasting, high-glucose infusion rates and octreotide for 17 days following the initial surgically managed NEC episode. The infant did not present with abdominal complications during the exposure to octreotide. However, feed intolerance and bloody stools recurred 4 days after a low dose of diazoxide was reintroduced (2.5 mg/kg/day). The episode was classified as NEC stage IB.
Regarding Bell’s stages, four episodes were classified as NEC stage IIA, and the remaining three had stage IIIB NEC.
In four cases, medical treatment was sufficient, while one patient required surgical intervention. Two cases (29%) met surgical criteria but were not stable enough for the procedure and died from fulminant NEC. Table 3 provides a comparison according to NEC diagnosis for the diazoxide-treated infants.
Although no significant associations were found regarding gestational age, age at diazoxide treatment and birth weight, infants that developed NEC tended to have a lower gestational age, to have a lower birth weight and to be younger than those who did not.
The NEC incidence in diazoxide-exposed infants in this study was compared against the overall rate of NEC during the same years, in infants of the same gestational age admitted to Toronto’s NICUs that were not exposed to the drug. The diagnosis of NEC was more frequent in diazoxide-exposed infants when compared with non-exposed neonates (OR 5.07, 95% CI 2.27 to 11.27; p<0.001). NEC was diagnosed more commonly in all gestational age groups greater than 26 weeks following diazoxide exposure. The ORs were notably higher for infants in the 33–36 weeks’ gestation group (OR 13.76, 95% CI 3.77 to 50.23; p<0.001) (data shown in table 4).
Four patients in this study were exposed to octreotide. One of these babies presented with two episodes of NEC over the observation period. Octreotide was introduced during the initial NEC treatment fasting period and continued until full enteral feeding was achieved. Octreotide was discontinued after 17 days of therapy as diazoxide was reintroduced; 4 days after reinitiating diazoxide, a second episode of NEC developed. The remaining three infants did not develop NEC.
Diazoxide is a non-diuretic benzothiadiazide that activates KATP channels in various cells, including pancreatic B cells.5 Its first therapeutic use was in adults with hypertension due to its potent vascular smooth muscle relaxant effect.20–23 Given the hyperglycaemic response noted in these patients, its clinical indications were expanded, in the late 1960s, to include persistent hypoglycaemia early in life.16–18
Despite the lack of universal approval by National Regulatory Authorities, the availability of an oral preparation led to an increase in the therapeutic use of diazoxide for neonates with prolonged hypoglycaemia. Hypertrichosis, fluid retention and cardiorespiratory failure are known side effects of diazoxide.6 24–26 Yet, only limited documentation of other potentially serious side effects has been reported in neonates.24 27
The diazoxide effects on the neonatal gut have not been adequately studied, but there is reason to suspect their occurrence. Data about the diazoxide-related gastrointestinal effects in children subjects are available. Gong et al described gastrointestinal adverse effects in 56% of 55 patients aged 6 months–5 years treated with diazoxide for hypoglycaemia.16 In that study, diazoxide treatment was interrupted in eight patients because of profound gastrointestinal disturbances. Similarly, Hu et al reported that a rate of 28% transient gastrointestinal side effects in a cohort of 44 paediatric subjects with hypoglycaemia treated with diazoxide. The side effects disappeared 1–2 weeks following discontinuation of the drug.17 More recently, Welters et al 18 showed a 12% incidence of gastrointestinal side effects in a large cohort of infants being treated for hyperinsulinaemic hypoglycaemia. In the present study, we documented gastrointestinal disturbances in 33% of neonates receiving diazoxide, and these were serious enough to result in temporary enteral feeding discontinuation.
Limited data on NEC as a complication of diazoxide treatment for neonatal hypoglycaemia is available. A recent retrospective study on the clinical outcome of 1066 neonates treated with diazoxide for hypoglycaemia reported less than 1% of neonates developing NEC.28 However, the incidence of NEC among diazoxide-treated infants in the present study suggests a possible association between drug exposure and the disease. Between 2012 and 2017, the CNN reported an incidence of NEC of 2.97% in Toronto’s tertiary units. In contrast, a fourfold higher NEC prevalence was documented in diazoxide-treated infants in the present study.
In support of a causal relationship between diazoxide exposure and the disease, the clinical characteristics of the NEC cases in our cohort differ significantly from the ‘usual’ NEC presentation. While Yee et al 29 recently described that the incident of NEC is higher in those infants born at less than 28 weeks’ gestation, all NEC cases in our cohort involved infants with a gestational age at birth greater than 30 weeks.
The timing of NEC signs presentation also differed. In the previously cited publication, NEC occurred more frequently between 32 and 34 weeks’ corrected gestational age. In our cohort, the mean time of presentation was 36 weeks, with all babies being over 33 weeks’ corrected gestational age at the time of diagnosis.
In our study, the odds of developing NEC while receiving diazoxide was highest in infants born at 33–36 weeks’ gestational age. Interestingly, the SGA prevalence was also greatest in this gestational age bracket. Thus, when compared with adequate for gestational age, SGA preterm infants may be at greater risk of developing NEC following therapeutic exposure to diazoxide.
For persistent neonatal hypoglycaemic treatment, diazoxide is recommended, given its tolerability, when compared with octreotide. NEC has previously been reported in association of octreotide treatment. However, while the incidence of NEC in octreotide-exposed infants is 2%,11 13 in this study, the incidence of NEC in diazoxide-exposed neonates is much higher. Also, interestingly, while some of the reported octreotide-associated cases of NEC were described to have been exposed to diazoxide, none of our patients developed NEC while receiving octreotide. In addition, in the 17 reported cases of NEC in octreotide-treated infants, the timeline between the drug administration and onset of NEC signs was not described.8–13 In our cohort, all infants developed NEC within 12 days from diazoxide treatment initiation (mean time of 5 days). This relatively short time lag between drug initiation and disease signs further supports a potential association between diazoxide and the development of NEC.
Although NEC is known to be multifactorial, gut dysmotility and increased wall permeability have recently shown to have an important role in the physiopathology of this disease. Moreover, in animal models, it has been demonstrated that the incidence and severity of the bowel damage increase with duration and magnitude of intestinal dysmotility.30 It is well known that diazoxide hyperpolarises the smooth muscle cells, leading to their relaxation. By hyperpolarising the smooth muscle cells and closing the L-type calcium channels, diazoxide was reported to abolish the spontaneous and rhythmic intestinal activity in animals.31 32 We propose that the mechanism accounting for the diazoxide-associated NEC in neonates involves gut dysmotility.
This study has limitations. Aside from being a retrospective study, we are unable to exclude the possibility that persistent hypoglycaemia is a risk factor for NEC development. Although hypoglycaemia was listed as a risk factor for NEC two decades ago,33 current studies make no mention of this metabolic disturbance as having a causative effect on the disease.34–36
In conclusion, we documented an increased incidence of NEC in a neonatal cohort treated with diazoxide for persistent hypoglycaemia. The risk of NEC associated with diazoxide was particularly greater in late preterm intrauterine growth restricted (IUGR) infants. Given the high morbidity and mortality associated with this disease, further investigation of its safety and caution use of diazoxide in newborns is warranted.
Data availability statement
Data are available upon reasonable request.
This study was approved by the institutional research ethics boards (MSH REB: 17–0291-C, SBHC: REB 017–2018 and HSC REB 1000059064).
Authors gratefully acknowledge the data abstractors from the Canadian Neonatal Network (CNN) and all of the staff at the CNN coordinating centre.
Contributors JB conceptualised and designed the study. He was the primary investigator in The Hospital for the Sick Children and collaborated with obtaining the Research Ethic Board approval and the Data Transfer Agreement among sites. He also reviewed and revised the manuscript for important intellectual content. LAP conceptualised and designed the study, designed the data collection instruments, collected data, carried out the initial analysis, drafted the initial manuscript, and reviewed and revised the manuscript. MC designed the data collection instruments, collected the data, and reviewed and revised the manuscript. AJ and DEW were the primary investigators in Mount Sinai Hospital and Sunnybrook Health Sciences Center, respectively. They collaborated with obtaining the Research Ethic Board approval and the Data Transfer Agreement among sites. They also coordinated and supervised the data collection, and critically reviewed the manuscript for important intellectual content. All authors approved the final manuscript as submitted and agreed to be accountable for all aspects of the work.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.