How Primary Care Can Close DKA Monitoring Gaps Before Hospitalization

In the second episode of Patient Care Online’s Clinical Forum series on diabetes monitoring, clinicians focused on how primary care teams can recognize diabetic ketoacidosis (DKA) risk earlier, close gaps in ketone testing, and use continuous glucose monitoring to guide more proactive diabetes care. The central takeaway: DKA prevention depends not only on technology, but also on patient education, sick day planning, and standardized workflows that help clinicians act before patients require hospitalization.

The episode continued the forum’s broader discussion on gaps in diabetes monitoring and earlier detection of diabetic ketoacidosis. Javier Morales, MD, vice president at Advanced Internal Medicine Group ​in East Hills, New York, and Associate Professor of Medicine at the Hofstra Northwell School of Medicine​, led the discussion and framed the session around the clinical and economic burden of DKA, drivers of delayed recognition, opportunities for earlier identification outside the hospital, the role of continuous glucose monitoring in insulin and noninsulin-treated patients, and strategies to integrate ketone measurement into routine care.

Why does DKA remain difficult to recognize in primary care?

Morales emphasized that DKA can occur in any patient with diabetes, including patients with type 2 diabetes, even though it has historically been associated with type 1 diabetes. He cited data showing a 54% rise in DKA between 2009 and 2014, before sodium-glucose cotransporter-2 inhibitors became mainstream therapy, and noted that DKA may occur in approximately 30% of patients with type 2 diabetes and may be present at onset in some cases.¹

Several participants described how DKA risk can be obscured in real-world primary care. One clinician practicing in Jamaica, Queens, described patients with insurance instability, housing insecurity, food shortages, and competing cardiometabolic emergencies, noting that ketone testing can be crowded out when patients present with multiple urgent concerns. In one case, she described a patient with markedly elevated A1C, uncontrolled blood pressure, dyslipidemia, and reluctance to go to the emergency department, leaving little time to address glucose testing or ketone screening during the visit.

That kind of scenario, panelists suggested, is exactly why DKA prevention has to be operationalized. If ketone testing depends on a clinician remembering it at the end of an overloaded visit, it will be missed.

What should primary care know about euglycemic DKA?

A major focus of episode 2 was euglycemic DKA, particularly in patients receiving SGLT2 inhibitors. Morales explained that the class can lower glucose through urinary glucose excretion while contributing to metabolic conditions that favor ketosis. As a result, patients may present with DKA despite glucose values that are normal, near-normal, or only modestly elevated. He described diagnostic features including pH less than 7.3, bicarbonate around 18 or lower, anion gap greater than 12, and beta-hydroxybutyrate around 3 mmol/L or higher.

For clinicians, the practical point was direct: any patient using an SGLT2 inhibitor who presents unwell should prompt consideration of ketone testing. Morales noted that illness, stress, low-carbohydrate intake, and reassuring glucose values can all delay recognition.

Panelists also discussed the distinction between ketosis and ketoacidosis. Conan Tu, MD, said outpatient clinicians may be seeing mild ketosis more often than they realize, particularly among patients on SGLT2 inhibitors who are fasting for procedures or patients receiving glucagon-like peptide-1 receptor agonists who are nauseated and eating less. He added that a background level of ketosis may be physiologic for some patients, but it can narrow the margin of safety if illness or insulin deficiency develops.

Which patients are at higher risk?

Episode 2 outlined high-risk, moderate-risk, and situational-risk groups. High-risk patients included those with type 1 diabetes using multiple daily injections or insulin pumps, patients with prior DKA hospitalizations, those with ketosis-prone type 2 diabetes, and patients with poor sick day adherence. Moderate-risk groups included patients with type 2 diabetes using SGLT2 inhibitors, poorly controlled type 2 diabetes with HbA1C greater than 9%, Black and Hispanic adults with new-onset type 2 diabetes, and patients with psychiatric disorders. Situational risks included perioperative fasting, ketogenic or very-low-carbohydrate diets, pregnancy with preexisting diabetes, acute febrile illness, and new insulin initiation in type 2 diabetes.

The pediatric perspective added another layer. Pushpa Abraham, a pediatric nurse practitioner, said her practice is seeing more adolescents—and even children as young as 7 or 8 years—with type 2 diabetes. Although she said DKA remains less common in pediatric type 2 diabetes than type 1 diabetes in her experience, her health system created a chronic illness program in response to DKA hospitalizations among at-risk children, particularly those with depression, anxiety, eating disorders, and social determinants of health concerns.

Why are sick day plans essential?

The panel repeatedly returned to sick day education as a missing link in DKA prevention. Morales said patients need guidance on how to manage insulin, SGLT2 inhibitors, and other medications when they are sick, and clinicians need to move from reactive to proactive counseling. He also stressed that fragmented communication across the care team can increase DKA risk.

When asked whether clinicians should rely on resources from organizations such as the American Diabetes Association or Joslin Diabetes Center, Morales answered that sick day guidance should be individualized based on comorbidities, access to care, and health literacy.

Abraham described pediatric sick day management as more embedded into care than adult protocols: it is part of clinic visits, hospitalizations, and school health plans. However, she noted that adherence remains difficult, and families must be taught that ketones should still be checked even when glucose values are in range.

How does CGM support earlier intervention?

CGM was discussed as both a clinical tool and a behavior modification tool. Abraham called CGM a “game-changer” in pediatric diabetes care, particularly for insulin titration in the inpatient setting, where real-time data help clinicians make basal dosing decisions and support improved time in range.

Tu added that CGM tracings help clinicians customize treatment by identifying postprandial spikes, dietary patterns, and possible insulin deficiency. Other panelists noted that patients often change eating behavior when they can directly see the glucose impact of high-carbohydrate meals.

Morales highlighted time in range, time above range, time below range, coefficient of variation, and glucose management indicator as key CGM metrics. He noted that CGM can detect hypoglycemia, guide medication adjustment, and reduce hospitalizations compared with nonuse.

References:
1. Elendu C. Annals of Medicine and Surgery. 2023.85:2802–2807