Long-lasting hypoglycemia through insulin degludec during acute kidney injury after diabetic ketoacidosis: A case report

Abstract.

Diabetic ketoacidosis (DKA) may be complicated by acute kidney injury (AKI). While the pharmacokinetics of insulin degludec were previously suggested to be stable even in chronic kidney disease, its safety profile in acute or dynamic renal dysfunction remains unclear. We herein report a 13-yr-old boy with severe DKA complicated by AKI, who developed prolonged hypoglycemia after switching from continuous intravenous regular insulin to subcutaneous insulin aspart and insulin degludec. Insulin degludec was initiated, although the patient was no longer acidotic, during a period when his serum creatinine level was still rising. The patient repeatedly became hypoglycemic, which required the interruption of basal insulin for 8 d. In retrospect, the decision to initiate long-acting insulin during AKI was made in accordance with standard DKA protocols, without sufficient consideration of the deteriorated clearance of insulin degludec under impaired renal function. This case highlights the need for caution when initiating long-acting insulin analogues in patients with AKI, even in the post-DKA phase. Further studies are needed to elucidate insulin pharmacokinetics during non-steady-state renal impairment.

Highlights

● Insulin degludec may cause long-lasting hypoglycemia during acute kidney injury.

● Initiating long-acting basal insulin during acute kidney injury requires caution.

Introduction

Diabetic ketoacidosis (DKA) may be complicated by acute kidney injury (AKI) due to volume depletion. A previous study reported AKI in approximately 47% of pediatric DKA cases, with 28% being classified as severe AKI (1).

Insulin degludec is a long-acting basal insulin that forms multi-hexamers in subcutaneous tissue and gradually separates into monomers (2). This insulin analogue acts slowly and stably because the monomers are gradually absorbed into the circulation (2). Due to its stability, severe hypoglycemia occurs less frequently in patients with diabetes using insulin degludec than in those using insulin glargine (3). While insulin degludec provides stable glycemic control due to its long half-life, overdosing results in prolonged hypoglycemia. (4, 5). Similarly, reduced insulin degludec clearance may prolong hypoglycemia for a long time; however, there are currently no reports on how long it may persist.

Exogenous insulin is mainly cleared by the kidneys; therefore, the decreased clearance of exogenous insulin in patients with impaired renal function may result in prolonged exposure to insulin (2, 6). Impaired renal function may increase the risk of hypoglycemia in patients with diabetes (7). AKI is reported to be a strong risk factor of hypoglycemia in hospitalized patients with type 2 diabetes (8). Insulin degludec and insulin aspart were previously shown to maintain their clearance through single administration in patients with renal failure (2, 9). However, it remains unclear whether repeated injections of insulin degludec or insulin aspart in patients with renal failure maintain their clearance. Furthermore, it is not known whether AKI during DKA impairs the clearance of exogenous insulin.

We herein report a case of severe DKA complicated by AKI in which hypoglycemia persisted after transitioning to long-acting insulin therapy, possibly due to the decreased renal clearance of insulin degludec. This case highlights the need to re-evaluate insulin strategies in patients with DKA complicated by AKI.

Case Report

A 13-yr-old boy was emergently transported to the hospital due to the loss of consciousness. He had experienced thirst and fatigue for two weeks. He tested positive for influenza B at a clinic four days before admission. The day before admission, the doctor at the clinic tried to insert a needle into a vein for an intravenous drip, but was unable to do so. No blood tests, blood glucose measurements, or urine tests were performed at that time. The patient was drowsy and found it difficult to walk by himself. On the day of admission, he spoke inconsistently and became unconscious. His parents called for an ambulance, and he was transported to the hospital. The Glasgow Coma Scale was 4 (eye opening 1, verbal response 2, motor response 1), the respiratory rate was 30 breaths/min, and rectal temperature was 32.5°C. The pupil diameter was 3 mm on both sides, and the light reflex was dull. Body weight was 46 kg, which was 12 kg lower than two weeks before. Since blood pressure was unmeasurable and the patient was in a state of shock, rapid infusion was performed. The patient had severe metabolic acidosis (pH 6.886, pCO₂ 28.4 mmHg, HCO^3– 5.4 mmol/L, BE –27.9 mmol/L, and lactate 1.5 mmol/L) and hyperglycemia (glucose 780 mg/dL), and was diagnosed with DKA. Since SpO₂ was unstable (65–85%), the patient was intubated. After initiating a regular human insulin continuous intravenous infusion, the patient was transferred to our hospital.

He still had metabolic acidosis (pH 7.069, pCO₂ 20.2 mmHg, HCO^3– 5.6 mmol/L, BE –23.9 mmol/L, and lactate 1.0 mmol/L), hyperglycemia, and ketosis (Table 1) at admission to our hospital. Furthermore, BUN and creatinine levels were elevated (Table 1), which was attributed to dehydration. The patient was extubated on day 2 of hospitalization due to improvements in consciousness and breathing. Metabolic acidosis improved on day 3. On day 4 of admission, a subcutaneous injection of rapid-acting insulin (Insulin aspart) for meals was initiated. The continuous intravenous infusion of regular human insulin was stopped and followed by an injection of long-acting insulin (insulin degludec). At our institution, the standard practice was to initiate long-acting insulin analogues after resolution of acidosis and upon resumption of oral intake. The initial dose of basal insulin was estimated based on the patient’s final intravenous insulin infusion rate. In this case, the patient required continuous intravenous regular insulin at 2 units/h on day 4, suggesting a basal requirement of approximately 48 units/d. To ensure safety during the post-acute phase of DKA, we started with a conservative dose of degludec (20 units), approximately 40% of the estimated daily basal need, and titrated upward based on subsequent glucose profiles. Our institution had no clear guidelines regarding safety measures when AKI occurs during DKA. Blood glucose levels were monitored by self-monitoring of blood glucose (SMBG) seven times a day; before breakfast (8 am), two hours after breakfast (10 am), before lunch (12 pm), two hours after lunch (2 pm), before dinner (6 pm), before bedtime (9 pm), and at 11 pm. Supplementary food was considered when his blood glucose level fell below 80 mg/dL during the daytime or below 120 mg/dL after 9 pm. Insulin degludec was injected subcutaneously at approximately 6 pm every day from day 4 of admission (Figs. 1 and 2). The patient became hypoglycemic between 11 pm and 4 am (approximately five to ten hours after dinner) and between 3 pm and 6 pm (approximately three to six hours after lunch) from day 8 of admission and, thus, supplementary food was required or insulin administration was postponed until after meals because of hypoglycemia three to five times a day (Fig. 2). During some of the hypoglycemic episodes, the patient experienced numbness in the legs (Fig. 2). He did not develop seizures or coma. The intravenous administration of glucose was not needed. After stopping insulin degludec on day 9 of admission, hypoglycemia between 3 pm and 6 pm persisted for a few more days (Fig. 2), which we attributed to the delayed effect of insulin caused by acute renal failure and the removal of glucotoxicity. To reduce the risk of recurrent hypoglycemia, a subcutaneous injection of insulin glargine U300, which has a shorter half-life than degludec, was initiated at two units per day from day 17 of admission and gradually increased. After confirming that hypoglycemia did not frequently occur with long-acting insulin, insulin glargine U300 was switched to insulin degludec on day 25 of admission (Fig. 1). Insulin level was measured only after his blood glucose levels had stabilized, on day 17 of admission. The fasting insulin level was 3.3 µIU/mL, and the simultaneously measured fasting blood glucose level was 180 mg/dL. Insulin secretory disorder detected by the glucagon test on day 17 of admission (basal and peak C-peptide levels: 0.87 and 1.74 ng/mL, respectively) and a positive test result for the anti-glutamic acid decarboxylase antibody confirmed the diagnosis of type 1 diabetes. The patient was discharged on day 27 of admission.

Table 1. Blood test results (at admission to our hospital).

Fig. 1.

Daily changes in blood glucose levels and serum creatinine concentrations during the clinical course. Black circles show glucose levels (mg/dL). Blue squares indicate serum creatinine concentrations (mg/dL). Basal insulin doses (gray striped/green/purple bars) and bolus insulin doses (orange bars) are shown at the top. Gray striped bars represent regular human insulin; green bars represent insulin degludec; and purple bars represent insulin glargine U300 (units/d). Orange bars represent insulin aspart as bolus insulin (units/d). The dashed horizontal line represents both the lower limit of the blood glucose target range (80 mg/dL) and the upper limit of the serum concentration of creatinine (0.80 mg/dL).

Fig. 2.

Glucose trends and insulin administration from day 5 to day 19. This figure illustrates the hourly blood glucose levels and insulin administration from day 5 to day 19 of hospitalization. The vertical axis represents time (0–23 h), and the horizontal axis corresponds to each hospital day. For each day, the left column shows glucose values, and the right column lists insulin doses. The background color of each glucose value indicates the glycemic range: red for ≥ 400 mg/dL, orange for 300–399 mg/dL, pale yellow for 200–299 mg/dL, light green for 100–199 mg/dL, sky blue for 80–99 mg/dL, and blue for ≤ 79 mg/dL. In the insulin column, “A” followed by a number indicates the dose (in units) of subcutaneous insulin aspart (rapid-acting), “D” indicates insulin degludec (long-acting), and “G” indicates insulin glargine U300 (long-acting). An asterisk (*) following “A” denotes that insulin aspart was administered before a planned snack. Black circles (●) indicate hypoglycemic episodes requiring carbohydrate supplementation, and black squares (■) represent symptomatic hypoglycemia (e.g., numbness). The total administered units of insulin degludec, serum creatinine levels, and eGFR for each day are shown at the bottom of the Figure. The values above the reference ranges are boldfaced. NE, not evaluated.

The patient was diagnosed with pre-renal acute renal failure due to renal dysfunction on the day of admission to our hospital and a low fractional excretion of sodium (FENa) level (Table 2). The intravenous fluid volume was adjusted and urine output was maintained at ≥ 0.5 mL/kg/h. However, serum creatinine increased, reaching a peak on day 5 of hospitalization (Fig. 1, Table 2). FENa and the fractional excretion of uric acid also increased and urinary β2MG was markedly elevated (Table 2), suggesting acute tubular necrosis due to renal ischemia caused by shock. Ultrasound showed that renal blood flow was maintained, and there was no decrease in fragmented red blood cells or haptoglobin, ruling out thrombotic microangiopathy. Renal failure gradually improved after day 6 of admission. The patient has not developed diabetic nephropathy seven years after DKA and AKI.

Table 2. Results of laboratory test related to renal function and urine volume.

Discussion

The present case highlights a paradoxical phenomenon: repeated hypoglycemic episodes occurring soon after the resolution of severe DKA complicated by AKI, despite the expected insulin resistance due to glucotoxicity. We hypothesize that the impaired renal clearance of insulin degludec during AKI might contribute to prolonged insulin activity and unexpected hypoglycemia.

Insulin degludec is a long-acting basal insulin that exerts its prolonged effects by binding to albumin after converting from a hexamer to a monomer and gradually being released into the circulation (5). While renal clearance plays a role in insulin elimination, the pharmacokinetics of insulin degludec are generally considered to be stable even in individuals with renal impairment, based on single-dose administration studies on adults with chronic kidney disease or end-stage renal disease (2). In a type 1 diabetes patient with CKD undergoing hemodialysis, switching from insulin glargine to insulin degludec provided better glycemic control without developing severe hypoglycemia (10). However, these findings may not necessarily translate to repeated administration under non-steady-state conditions, such as AKI, in pediatric DKA patients. In such situations, the choice of basal insulin may be important. Since insulin degludec—characterized by stability—may pose a risk of too prolonged action, using other long-acting insulin analogs, such as insulin glargine, insulin glargine U300, or insulin detemir—which have shorter durations than degludec—might have allowed for more flexible blood glucose control.

In the present case, the administration of insulin was switched to subcutaneous injections of insulin aspart and insulin degludec despite renal function continuing to deteriorate, as indicated by rising serum creatinine levels. This decision, made in accordance with conventional DKA protocols, was not preceded by sufficient consideration of the ongoing renal impairment. In hindsight, greater caution may have been warranted, particularly when initiating a long-acting insulin analogue during the dynamic phase of AKI recovery. Frequent hypoglycemic episodes during renal impairment mainly occurred more than six hours after the injection of rapid-acting insulin (aspart), suggesting that basal insulin might have played a major role. Of course, the involvement of rapid-acting insulin analogues cannot be entirely excluded. However. the most suggestive episodes of degludec involvement occurred during the late-night hours (e.g., around 3 am on day 8 and day 9), when the influence of meals and prandial insulin could be reasonably excluded. In addition, on day 9 and day 10, hypoglycemia recurred several hours after recovery from postprandial hypoglycemia, indicating a possible overlap of basal and bolus insulin effects under impaired renal clearance (Fig. 2). While it is difficult to precisely disentangle the contribution of each insulin type, these findings underscore the need for cautious dosing and strict glucose monitoring during AKI and its recovery phase. Although insulin degludec was discontinued on day 9, hypoglycemia persisted for several days, which may support the hypothesis of a prolonged duration of action due to impaired clearance. More flexible adjustment of basal insulin, using continuous subcutaneous insulin infusion (CSII) or continuing intravenous insulin infusion, could have made it safer to control blood glucose levels until improvement of renal clearance.

Under physiological conditions, insulin is freely filtered through the glomeruli. Most of the filtered insulin is then reabsorbed in the proximal tubules and subsequently degraded in proximal tubular cells (11). Therefore, in AKI, insulin filtration may decrease because of reduced glomerular filtration. Furthermore, tubular injury may reduce insulin degradation in tubular cells. However, direct evidence linking AKI to altered insulin pharmacokinetics in humans is limited, and this mechanism remains speculative.

A previous study demonstrated that an overdose of insulin degludec in a patient without renal insufficiency induced persistent hypoglycemia for more than one week, aligning with the half-life of albumin (5). Although the present case did not involve an overdose, the repeated administration of insulin degludec during deteriorated renal function may have resulted in accumulation, functionally resembling an overdose in terms of the persistence of hypoglycemic effects.

It is important to note that other mechanisms may have contributed to repeated hypoglycemia, such as decreased insulin aspart clearance or the rapid reversal of glucotoxicity. However, the consistent timing and prolonged course of hypoglycemia suggest that the impaired clearance of basal insulin was a major contributor. High dose of insulin degludec may have contributed to prolonged hypoglycemia. While the initial degludec dose was conservatively set at 20 units (approximately 40% of the estimated basal requirement), it was gradually increased based on the glycemic profile. Recently, international guidelines states that total daily dose of insulin at pubertal age is 0.7 to 1.0 IU/kg/d, and basal/long-acting analog dose is approximately 30 to 50% of the daily insulin (12). Although the max dose may seem relatively high for a 46 kg patient in this case, it was during a period when insulin resistance due to glucotoxicity would typically be expected. In retrospect, we cannot entirely exclude the possibility that the total basal dose was excessive relative to his unexpectedly high insulin sensitivity during renal recovery. This case underscores the need for individualized insulin titration in the context of dynamically changing renal function.

Continuous glucose monitoring (CGM) is now widely recognized as a superior tool for detecting nocturnal and asymptomatic hypoglycemia, and current guidelines recommend CGM use to reduce severe hypoglycemia (13). However, CGM was not utilized during the acute phase in this case, and blood glucose level was instead monitored by SMBG, which occasionally detected asymptomatic hypoglycemia. Incorporating CGM may help reduce the risk of nocturnal or unnoticed hypoglycemia in cases with hypoglycemic risks, such as during AKI.

This case underscores the importance of considering the renal status when initiating long-acting insulin analogues in patients recovering from DKA complicated by AKI. Although insulin degludec is generally regarded as safe in chronic kidney disease, its behavior in the context of AKI—particularly during dynamic changes in renal clearance—remains poorly understood.

While prior studies have focused on insulin pharmacokinetics in stable renal dysfunction, this case highlights an under-characterized scenario: repeated degludec administration during non-steady-state AKI in pediatric DKA. The observed persistent hypoglycemia after degludec discontinuation calls for greater awareness of basal insulin accumulation risks in dynamic renal contexts.

In conclusion, while the present case suggests a potential risk of prolonged hypoglycemia related to basal insulin in the context of pediatric DKA with AKI, multiple mechanisms may be involved. We recommend close monitoring and careful insulin adjustment in such settings. Further studies are needed to clarify the pharmacokinetics of insulin analogues under conditions of impaired renal function.

Conflict of interests

The authors have no conflicts of interest to declare.