Authors: Brynn E. Marks (aChildren’s Hospital of Philadelphia, Division of Endocrinology and Diabetes, Philadelphia, PA, USA; bDepartment of Pediatrics, University of Pennsylvania School of Medicine, Philadelphia, PA, USA), Seema Meighan (aChildren’s Hospital of Philadelphia, Division of Endocrinology and Diabetes, Philadelphia, PA, USA), Emily E. Fivekiller (cBarbara Davis Center for Diabetes, University of Colorado Anschutz Medical Campus, Aurora, CO, USA), Estella Escobar (cBarbara Davis Center for Diabetes, University of Colorado Anschutz Medical Campus, Aurora, CO, USA), Cari Berget (cBarbara Davis Center for Diabetes, University of Colorado Anschutz Medical Campus, Aurora, CO, USA)
Categories: Article, Automated insulin delivery, Diabetes self-management, Diabetic ketoacidosis, Insulin pumps, Ketosis, Type 1 diabetes
Source: Hormone research in paediatrics
Doi: 10.1159/000541430
Authors: Brynn E. Marks, Seema Meighan, Emily E. Fivekiller, Estella Escobar, Cari Berget
Diabetic ketoacidosis (DKA) is the leading cause of mortality among youth with type 1 diabetes (T1D). Guidelines for DKA prevention exist; however, specific guidance about when to check ketones and how to manage youth using insulin pumps and automated insulin delivery (AID) systems is lacking.
A 35-item online survey exploring clinical ketone management practices for youth with T1D in the USA was distributed to diabetes healthcare professionals (HCPs). Survey responses, including multiple-choice and Likert scale questions, were summarized and rates of agreement and disagreement (Likert scale 4, 5 vs. 1, 2, 3) are reported.
In total, 123 HCPs (51% physicians, 26% diabetes educators, 19% nurse practitioners) from 47 institutions completed the survey. Seventy percent worked at academic specialty centers. Ninety-seven percent reported >50% continuous glucose monitoring use in their clinic and 72% reported >50% insulin pump use. Although 79% reported having ketone management protocols, the level and duration of hyperglycemia at which ketone monitoring was advised ranged from >200 to 350 mg/dL and from 0 min to >6 h of duration. While 72% had distinct ketone management protocols for pump users, only 29% had specific protocols for AID. Sixty-two percent agreed that DKA due to infusion site failure was a significant problem in their practice, and 70% agreed there was a need to standardize ketone management guidelines.
The preventable nature and high incidence of DKA highlight the need to build consensus for clinical ketone management and to develop tools to facilitate management, especially as the use of diabetes technologies continues to increase.
Diabetic ketoacidosis (DKA) is a life-threatening complication of diabetes defined by hyperglycemia, metabolic acidosis, and ketosis [1] resulting from relative insulin deficiency which is most often precipitated by insulin omission or interruption in insulin delivery through a pump [2]. The incidence of DKA in youth with established type 1 diabetes (T1D) is between 4 and 8 episodes per 100 person-years [2, 3] with 60% of these episodes occurring among just 5% of youth and a higher incidence among youth with poorer glycemic control, non-White race, and lower socioeconomic status [2, 4, 5]. DKA is the leading cause of hospitalizations, morbidity, and mortality among youth with T1D [6] accounting for approximately 50% of all deaths in youth with T1D less than 24 years of age [7–9]. DKA is also a significant healthcare burden with more than 500,000 hospital days per year and USD 2.4 billion in hospital costs annually in the USA [10].
Advances in technology have led to increased uptake of continuous glucose monitoring (CGM) and insulin delivery technologies, including automated insulin delivery (AID) devices. Data from the T1D Exchange Quality Improvement Collaborative show that in 2022, 78% of people with T1D used CGM, 62% used an insulin pump, and 33% used AID, an increase from 50%, 27%, and 0%, respectively, in 2016 [11]. These increases in technology use were associated with improvements in glycemic control, with average hemoglobin A1c levels declining from 8.7% to 8.4% in the same period. Despite these advances, DKA remains a common adverse event for youth using insulin pumps and infusion site failure is the primary cause. Infusion site failure leads to an interruption in insulin delivery due to a bent or dislodged cannula, or inadequate insulin absorption from the subcutaneous tissue, which may be due to local inflammation at the infusion site even in the absence of kinking in the cannula itself [12, 13]. It is often characterized by persistent hyperglycemia that does not resolve despite an increase in insulin delivery from the pump (e.g., administering a correction bolus). A recent retrospective analysis found that 57% of DKA admissions in pump users were caused by infusion site failure and another 14% due to other pump malfunctions [14]. Another analysis found a 10% rate of hospital admissions for pump users and identified that ketosis due to infusion site failure precipitated 81% of these admissions [15]. Whereas DKA was historically more common among youth using insulin pumps before the widespread use of CGM [16–19], more recent data have shown that DKA is more common among youth using multiple daily injection (MDI) therapy [3].
While there is a plethora of literature regarding the best ways to manage DKA, there is a lack of standardized guidelines for managing ketones in an effort to prevent DKA. Existing literature focuses on the prevention of DKA through education on sick day management [20–22]. There is consensus surrounding key educational points including the importance of insulin administration, frequent glucose and ketone monitoring, and hydration [1]. Educational programs to teach outpatient sick day management to youth with T1D and their caregivers have been shown to improve sick day management practices and reduce emergency room visits [22]. However, specific guidelines for when to check ketones, how to determine insulin doses, and the frequency of glucose and ketone monitoring are lacking. Moreover, current guidelines do not account for the intricacies of ketone management among pump and AID users, including troubleshooting infusion site failure and guidance on whether AID should be continued when giving insulin injections.
High rates of DKA in youth with T1D due to preventable causes, including infusion site failure and other pump malfunctions, highlight the need for greater understanding of current DKA prevention practices in clinical settings in the USA. The purpose of our study was to explore institutional and individual clinician approaches to ketone management as well as clinician’s perceptions about ketone management in the USA.
A survey was created by the study investigators (B.M. and C.B.) to gather information from pediatric diabetes healthcare professionals (HCPs) on their approach to ketone management in youth with T1D. The survey included 35 questions in addition to demographic information (online suppl. 1; for all online suppl. material, see https://doi.org/10.1159/000541430). Twelve multiple-choice questions addressed institutional guidelines for ketone management in the clinical setting and 10 assessed the HCP’s personal approach to ketone management. Questions addressed policies and practices around ketone self-management education, standardized protocols used to guide ketone management, and handling of urgent calls for ketones during and after office hours. Using a 5-point Likert scale, survey respondents indicated their level of agreement with 13 statements about ketone management.
We recruited nurses, certified diabetes care and education specialists (CDCES), nurse practitioners (NP), pediatric endocrinologists, and pediatric endocrinology trainees practicing in a variety of clinical settings in the USA. Participants were recruited via announcements posted in social media groups targeting pediatric diabetes professionals, during professional diabetes conferences (e.g., the American Diabetes Association and the Association of Diabetes Care and Education Specialists), and by email invitation. The Colorado Multiple Institutional Review Board approved the study with a waiver of documentation of consent and determined the study to be exempt from ongoing review, submission ID APP001-1. Information about the study and voluntary nature of participation were provided to participants at the beginning of the survey along with the instruction that completing the survey indicated consent to participate.
Survey data were collected and managed using REDCap electronic data capture tools hosted at the Barbara Davis Center [23, 24]. Data analysis was performed using Stata statistical software v.18 (StataCorp, College Station, TX, USA, 2023). Descriptive statistics are reported for demographic characteristics. Differences in agreement (Likert scores of 4 or 5) and disagreement (Likert scores 1, 2, or 3) between physicians and non-physicians were assessed using Pearson’s χ^2^ tests. Statistical significance was defined as two-tailed p values <0.05.
In total, 123 HCPs (age 42 ± 10 years; 86% female) from 47 unique institutions completed the survey; 63 were physicians and 60 were non-physicians (32 CDCESs, 23 NPs, 5 others). Characteristics of the individual respondents are displayed in Table 1, and data about their clinical practice settings are in Table 2. Seventy percent of participants were employed at an academic specialty clinic and 23% in a hospital-based diabetes clinic. Most participants (64%) worked in centers with ≥1,000 youth with T1D and had high rates of technology use. Ninety-seven percent and 73% of respondents, respectively, reported that more than half of their clinic population used CGM and insulin pumps. Whereas 59% reported >70% CGM use, only 16% reported >70% pump use.
Nearly all (95%) participants reported that their institution educates youth and families about ketone self-management. Most (79%) used educational handouts to provide this education, while 34% used smart phrases in the electronic medical record and 20% reported using websites. Most institutions (98%) offered supports for youth and families managing ketones at home. Seventy-six percent supported families with ketone management by phone, while 27% provided a pager service and 18% advised youth and caregivers to send a message through the electronic medical record. During business hours, 62% of HCPs indicated that CDCESs field urgent calls or messages related to ketone management; other team members including NPs, RNs without CDCES certification, and pediatric endocrinology fellows were involved at some centers. After hours, ketone management calls were more commonly fielded by physicians, with 68% and 45%, respectively, indicating that a pediatric endocrine fellow or attending pediatric endocrinologist was the first-line HCP.
Eighty percent of participants reported their institution had a ketone management protocol in place, 14% reported no ketone protocol and 7% were unsure if they had one. Most clinicians (56%) were unsure if their institution’s protocol was based on existing guidelines and 23% percent reported using the International Society for Pediatric and Adolescent Diabetes (ISPAD) guidelines [1, 20]. While 72% reported having distinct ketone protocols for youth using insulin pumps, only 29% had separate protocols for AID users. Most ketone protocols (84%) used either blood glucose (BG) or CGM data to make recommendations. Eighty-nine percent used either urine or blood ketone values for recommendations within their protocols.
Ninety-two percent of HCPs reported making ketone management recommendations interchangeably using either BG or CGM levels and either urine or blood ketone levels. There were a variety of glucose levels at which HCPs advised youth with T1D to check ketones. Fifty-four percent of HCPs recommended checking ketones for glucose levels >300 mg/dL, 31% for glucose levels >250 mg/dL, 12% for glucose levels >240 mg/dL, 2% for glucose levels >200 mg/dL, and 2% for glucose levels >350 mg/dL. The duration of hyperglycemia at which it was recommended to check ketones also varied widely, ranging from immediately (3%) to >6 h duration (3%). Twenty-nine percent advised checking ketones after a 2-h duration while thirty percent recommended checking after 3-h of hyperglycemia. Eleven percent reported that the duration of hyperglycemia did not influence their recommendations about when to check ketones; these individuals recommended checking in case of “persistent” hyperglycemia not responding to insulin administration. Most (73%) HCPs advised caregivers to contact the clinic for moderate urine ketones, and 53% for blood ketone values ≥1.0 mmol/L. However, there was significant variability with regard to the ketone level above which caregivers were advised to contact the clinic, ranging from 0.7 to 2.5 mmol/L. Sixty percent of respondents used the individual’s correction factor to recommend an insulin dose for the treatment of ketones. Others considered the degree of ketosis (36%) and weight-based total daily insulin dose (19%) in the decision.
Overall, 62% of HCPs agreed or strongly agreed that DKA due to infusion site failure was a problem in their clinic (Fig. 1). Eighty-seven percent reported following their institution’s ketone management policies and 99% reported they were comfortable managing ketones. While many HCPs agreed that DKA was a problem, only 12% believed their ketone protocols were too complicated and only 26% believed it was difficult to teach families how to manage ketones at home. Despite this, only 54% indicated that families start managing ketones before calling their office and only 48% agreed that families were able to manage ketones independently at home. Seventy percent agreed that there is a need to better standardize ketone management across pediatric diabetes centers and 69% agreed that separate ketone guidelines for youth using AID are needed. Responses were similar between physician and non-physician professionals with differences occurring only in the percentage who endorsed following institutional ketone guidelines in their personal practice (62% of physicians vs. 80% of non-physicians, p = 0.03), and the percentage who believed that their institutional ketone guidelines were too complicated (5% of physicians vs. 16% of non-physicians, p = 0.04).
DKA is the leading cause of hospitalizations, morbidity, and mortality among youth with known T1D, accounting for over USD 5 billion in healthcare costs per year in the USA [10]. While recent pediatric guidelines for sick day [20] and DKA management [1] are available from ISPAD, few institutions reported using these guidelines to inform protocols. The basic principles of ketone management, including frequent glucose and ketone monitoring, increased insulin doses, and hydration, remain unchanged; however, our survey findings demonstrate inconsistent clinical practices about when to check for ketones, what level of ketosis should prompt sick day management and guidance for youth using AID systems. With the increased use of AID systems [11] and the nuances of each individual AID algorithm, guidelines must provide specific advice for those using these technologies. Furthermore, youth with T1D and caregivers need simple, concrete education on the actions to take to manage ketosis to ensure they are able to implement ketone management protocols effectively. The preventable nature of DKA, poor health outcomes for youth with DKA, and the healthcare costs associated with DKA emphasize the need to develop guidelines and tools to facilitate outpatient ketone management and DKA prevention.
The 2022 ISPAD Clinical Practice Guidelines about sick day management in children and adolescents with diabetes are the only recently published pediatric guidelines on this topic. These guidelines emphasize the importance of frequent monitoring of either blood or interstitial glucose (every 1–2 h), ketone monitoring, preferably by blood, adequate hydration, and insulin administration [20]. Nearly 80% of respondents indicated the use of institutional ketone protocols; however, just 23% indicated they were based on ISAPD guidelines and 56% reported they were uncertain whether the protocol was based on guidelines. Nursing care team members are the most common initial point of contact for ketone management during business hours and these care team members are unable to make independent therapeutic decisions. As a result, clear institutional protocols are needed to ensure that youth with T1D receive appropriate care and guidance for managing ketosis.
In alignment with ISPAD guidelines HCPs overwhelmingly favored accepting BG or CGM readings. However, there was less consensus on the glucose threshold that should prompt ketone monitoring. More than 50% of respondents reported counseling youth and families to check for ketones when glucose levels exceeded 300 mg/dL and just over 40% recommended checking when glucose exceeds 240–250 mg/dL. The ISPAD guidelines recommend checking for ketones when glucose levels exceed 250 mg/dL for pump users; however, there is no guidance for users and no recommendations regarding the duration of hyperglycemia that should prompt ketone monitoring. Nearly 80% of people with T1D in the USA use CGM, with higher rates of use reported in pediatric populations [11]. Real world CGM data from AID users indicate that glucose levels exceed 250 mg/dL 12–13% of the time [25, 26] and glycemic control is poorer among non-AID users, highlighting the impracticality of basing ketone monitoring on a single glucose level. Recognizing this impracticality, nearly 90% of respondents reported that the duration of hyperglycemia, most commonly 2–3 h, is an important consideration when advising when to check for ketones. The increased uptake of CGM and greater awareness of periods of transient hyperglycemia call attention to the need for more specific guidance in this area.
There was nearly universal agreement among HCPs that either blood or urine ketone monitoring is acceptable. Studies have shown that youth are more likely to monitor blood ketones than urine ketones and that blood ketone are more effective in preventing DKA [27]; however, the lack of insurance coverage for blood ketone monitoring is a barrier for many youth. Most respondents also agreed that families should contact the office when youth have moderate urine ketones, although there was less consensus for blood ketones. More than half of all respondents agreed that ≥1.0 mmol/L is a reasonable cutoff for families to contact the office, though some preferred a threshold of >0.7 mmol/L or ≥1.5 mmol/L. While the ISPAD guidelines recommend providing higher insulin doses when blood ketones ≥1.0 mmol/L, the lack of consensus on this topic likely stems from mixed findings about usual levels of ketosis in youth with T1D [28, 29] and agreement between blood and urine ketone levels [28, 30, 31].
Most respondents reported having different protocols for pump users, but few institutions had separate guidelines for AID users. The ISPAD guidelines emphasize the importance of careful examination of the insulin pump and delivery system to ensure proper insulin delivery, as well as delivering insulin via injection until the infusion set, tubing, and insulin reservoir have been replaced [20]. However, specific guidelines for AID systems are lacking. Each AID algorithm is unique, with some systems using basal modulation alone to correct hyperglycemia while others incorporate automated correction boluses. The continued use of AID algorithms while managing ketosis has the potential to expedite the clearance of ketones without an increased risk for hypoglycemia if appropriate system-specific guidance is given. Furthermore, the newest AID system, iLet, cannot be used in an open-loop or manual mode [32], making ISPAD guidelines largely irrelevant for this system. Ketone management guidelines for the iLet include replacing the infusion set, increasing hydration, and monitoring ketone and glucose levels every 90 min at trace-moderate urine ketones (blood ketones 0.6–2.5 mmol/L).
There are many methods used to calculate insulin dosing for youth with ketosis without acidosis, including body weight (0.1–0.15 unit/kg), total daily dose (10–20% of total daily insulin dose), and a percentage increase from the usual correction factor (10–20% extra); however, there is little evidence to support any particular method [20]. Most HCPs reported using the individual’s personalized correction factor to calculate insulin dose for ketone management, though approximately one-third considered the degree of ketosis in their insulin dose calculation and nearly 20% reported using weight-based dosing. The lack of consensus on how to calculate insulin dosing reflects the lack of evidence and guidelines to recommend a particular approach. Furthermore, given that infusion set failure is the leading cause of DKA in insulin pump users, failure to replace the infusion set and/or deliver an insulin injection of any dose in these situations is much more of a problem than the calculation method for a correction dose. When educating youth and their caregivers on troubleshooting infusion site failure, replacing the infusion set is paramount and providing a simple method for calculating an insulin injection is important to ensure effective implementation and DKA prevention.
There are several limitations to this study. Although these data provide a thorough assessment of current practices, most participants (51%) were physicians. Although we recruited using a variety of means intended to reach HCPs from a variety of backgrounds, only 32 responses were obtained from CDCESs. Insights from CDCESs are invaluable as they are often responsible for educating youth and families about ketones and fielding ketone management phone calls. Most HCPs were employed at large academic centers that typically provide greater staffing and resources. Smaller practices may have unique insights into ketone management that we were unable to capture. Although just over half of HCPs agreed about the need for separate ketone management guidelines for AID users, this survey was conducted before the clinical launch of the iLet AID system and ongoing advances in fully closed-loop AID systems. Despite these limitations, responses from 123 HCPs at 47 unique institutions in the USA provides interprofessional insights into clinical practices spanning different insurance landscapes, catchment areas, and resources. Survey topics with a lack of consensus also call attention to the need for clearer guidelines and ketone management tools that keep pace with ongoing technological advancements.
Despite nuanced differences in ketone management, our survey data demonstrate HCP consensus on many aspects of care. While most institutions have protocols for the management of ketones, many HCPs were unsure whether they were based on guidelines. HCPs denied difficulties teaching families to manage ketones and felt that families could follow their protocols; however, less than half reported that families were able to start ketone management before contacting the office and many agreed that DKA due to infusion site failure remains a significant problem. Given the current shortage of pediatric endocrinologists [33] and CDCESs [34] consensus guidelines and simple tools to support families in developing the skills to manage ketones are needed.