Children in Intensive Care With Diabetic Ketoacidosis: Landmark Analysis Reveals Who Is Most at Risk

When a child arrives in a pediatric intensive care unit with diabetic ketoacidosis, the metabolic emergency that can be the first and most dramatic sign of type 1 diabetes, clinicians face a cascade of dangerous derangements at once: soaring blood glucose, acid-spilling ketones, dehydration that can tip into shock, and the ever-looming threat of cerebral edema, the leading cause of death in this population. A new systematic review and meta-analysis published in BMC Endocrine Disorders has now pulled together the global evidence on these critically ill children, offering one of the most detailed pictures to date of what actually happens to them in the PICU and what laboratory markers on admission should set alarm bells ringing. The study, led by Shahab Noorian of Alborz University of Medical Sciences in Karaj, Iran, together with colleagues across several Iranian research institutions, analyzed 41 studies encompassing 3,792 pediatric patients under 21 years of age who required intensive care for diabetic ketoacidosis.

The headline finding is a paradox that will resonate with anyone who has cared for these children. Mortality in the pediatric intensive care unit is strikingly low, pooled at just one percent, yet the burden of critical illness is enormous. More than half of the children, 53.7 percent, presented with severe diabetic ketoacidosis, the most dangerous tier of the condition, and a remarkable 58 percent arrived at the intensive care unit with diabetic ketoacidosis as their very first presentation of diabetes. In other words, for the majority of these children, the intensive care admission was not a complication of a known chronic disease but the brutal unveiling of one. The figure underscores a persistent challenge in pediatric medicine: the diagnosis of type 1 diabetes is still too often delayed until the body’s compensatory mechanisms collapse entirely and ketoacidosis forces the issue.

The clinical signature of these admissions is remarkably consistent across studies and continents. Vomiting and polydipsia, the profound thirst that drives children to drink compulsively, each appeared in 63 percent of patients, while polyuria, the excessive urination that results when glucose drags water out of the body through the kidneys, was documented in 61 percent. Weight loss, a consequence of the catabolic state in which the body burns fat and muscle because it cannot use glucose, affected 43 percent of the children. Tachypnea, the rapid deep breathing known as Kussmaul respiration that serves to blow off carbon dioxide and partially compensate for the metabolic acidosis, was present in 42 percent. Loss of consciousness, a red flag for cerebral involvement, was recorded in 41 percent, a sobering reminder of how close many of these children already are to neurological catastrophe by the time they reach intensive care.

The remaining symptoms paint the full picture of a body in systemic distress. Fever accompanied 38 percent of admissions, abdominal pain 32.5 percent, a finding that frequently misleads clinicians toward surgical diagnoses such as appendicitis, enuresis, or new-onset bedwetting in previously dry children, affected 31 percent, shock was documented in 30 percent, and nausea in 25 percent. Acute kidney injury emerged as the single most common complication, affecting 30 percent of the children, a statistic that speaks to the profound volume depletion and hypoperfusion these patients sustain before admission. The kidneys, caught between dehydration and the osmotic load of glucose, are among the first organs to falter, and their injury on admission is one of the strongest practical signals that a child’s metabolic crisis has been brewing for some time.

What sets this analysis apart is its use of meta-regression, a statistical technique that examines how continuous laboratory variables track with the risk of specific complications across the pooled studies. The results are technically revealing. Children with higher levels of glycated hemoglobin, or HbA1c, the long-term marker of glucose control, and those with lower bicarbonate, lower carbon dioxide partial pressure, and lower pH on admission, were significantly more likely to develop cerebral edema. This cluster makes physiological sense: deeper acidosis reflects more advanced ketone accumulation and more severe cellular starvation, while a lower partial pressure of carbon dioxide marks more vigorous compensatory hyperventilation. The correlation between very high HbA1c and brain injury has long been suspected clinically, and this pooled analysis now places the association on firmer statistical footing across thousands of patients.

The patterns around hypoglycemia tell an equally instructive story from the opposite direction. Children who presented with higher blood glucose or, counterintuitively, lower HbA1c had fewer hypoglycemic events during treatment. The paradox of the HbA1c finding likely reflects the fact that a lower HbA1c in a child presenting with diabetic ketoacidosis suggests a more abrupt metabolic decompensation, often with less chronic ketone burden and a different treatment trajectory, whereas the relationship between high presenting glucose and subsequent hypoglycemia risk is shaped by how aggressively insulin therapy drives glucose downward in the first hours of care. Meanwhile, a higher carbon dioxide partial pressure at admission was statistically associated with fewer recurrent ketoacidosis episodes but, intriguingly, with a higher risk of both hypokalemia, the dangerous depletion of serum potassium, and acute kidney injury, tying the acid-base picture directly to the electrolyte and renal consequences of treatment.

One of the most clinically consequential comparisons in the study addresses a question that intensive care teams confront every week: do children who present with diabetic ketoacidosis as the first sign of diabetes fare differently from those who already carry the diagnosis? The pooled data show meaningful biochemical differences between the two groups. Newly diagnosed patients arrived with higher levels of bicarbonate, potassium, and carbon dioxide partial pressure, suggesting somewhat less severe acidosis at presentation, yet their blood ketone levels were significantly lower, which is a curious signature that the authors’ pooled data capture clearly. Newly diagnosed children also required a significantly shorter duration of insulin infusion. In contrast, and perhaps surprisingly, the two groups showed no significant differences in renal function markers, other electrolytes, glycemic indices, neurological status, or the length of their intensive care stay. The bottom line for clinicians is that the intensity of supportive care needed appears largely independent of whether the diabetes was previously known.

That equivalence of outcomes carries an important public health message. The slightly more favorable acid-base profile of newly diagnosed children does not translate into shorter intensive care stays or better neurological outcomes, which means that prevention efforts aimed at families already living with type 1 diabetes, namely education about sick-day management, ketone monitoring, and the dangers of missed insulin doses, remain as critical as ever, because recurrent ketoacidosis in a known diabetic child is every bit as dangerous as a first presentation. At the same time, the finding that 58 percent of intensive care admissions represent first presentations is a loud argument for earlier recognition of the classic warning signs, polyuria, polydipsia, weight loss, and rapid breathing, in primary care settings, where a simple fingerstick glucose test could intercept many of these emergencies days before they reach the intensive care unit.

For a condition that remains one of the most feared emergencies in pediatrics, the one percent mortality figure is genuinely reassuring, and it reflects decades of refinement in fluid therapy protocols, low-dose insulin infusion strategies, and neurological surveillance. But the analysis makes clear that survival is only part of the story. With more than half of these children presenting with severe disease, thirty percent sustaining acute kidney injury, and cerebral edema risk rising with every millimeter of acidosis, the morbidity of pediatric diabetic ketoacidosis in the intensive care unit remains a formidable challenge. The authors registered their protocol prospectively in the PROSPERO database and conducted their searches across PubMed, Web of Science, Embase, and Scopus through July 23, 2025, lending methodological weight to conclusions that intensive care physicians, pediatric endocrinologists, and emergency clinicians can now use to sharpen risk stratification the moment a child in ketoacidosis crosses the threshold. The message distilled from nearly 3,800 critically ill children is ultimately one of measurable, actionable physiology: the numbers on the admission blood gas and the HbA1c assay are not merely laboratory values, but predictors of which children will fight their way through and which will face the most dreaded complications of this disease.

Subject of Research: Clinical outcomes and risk factors in children admitted to pediatric intensive care with diabetic ketoacidosis

Article Title: Clinical factors associated with pediatric intensive care unit outcomes in children admitted with diabetic ketoacidosis: a systematic review and meta-analysis comparing newly diagnosed and previously known diabetes

Article References: Clinical factors associated with pediatric intensive care unit outcomes in children admitted with diabetic ketoacidosis: a systematic review and meta-analysis comparing newly diagnosed and previously known diabetes. (n.d.). https://doi.org/10.1186/s12902-026-02517-2

Image Credits: AI Generated

DOI: 10.1186/s12902-026-02517-2

Keywords: diabetic ketoacidosis, pediatrics, intensive care, cerebral edema, acute kidney injury, type 1 diabetes, meta-analysis, systematic review, HbA1c, hypoglycemia, acid-base balance, mortality