Early-onset, treatment-cycle-linked hyperinsulinaemic insulin resistance during capivasertib therapy: a case report

Abstract

Introduction

Capivasertib is an oral AKT inhibitor used with fulvestrant for hormone receptor-positive, human epidermal growth factor receptor 2-negative advanced breast cancer with relevant pathway alterations. Hyperglycaemia is a recognised on-target adverse effect, but its early biochemical phenotype has not been fully characterised in patients without known diabetes.

Case presentation

A 43-year-old Japanese woman with metastatic breast cancer and no history or family history of diabetes started capivasertib on a 4-days-on/3-days-off schedule. Before treatment, fasting plasma glucose was 89 mg/dL and glycated haemoglobin was 5.6%. Although this glycated haemoglobin was within the non-diabetic range, impared glucose tolerance could not be excluded because a pre-treatment oral glucose tolerance test was not performed. On the day after treatment initiation, the 2-h postprandial plasma glucose level increased to 256 mg/dL. On day 2, fasting plasma glucose was 103 mg/dL, fasting insulin was 22.8 µU/mL, C-peptide was 3.40 ng/mL, and HOMA-IR was 5.80, indicating hyperinsulinaemic insulin resistance with preserved endogenous insulin secretion. During a subsequent drug-free interval, fasting plasma glucose decreased to 87 mg/dL, fasting insulin to 5.4 µU/mL, and HOMA-IR to 1.16 without glucose-lowering medication. HOMA-IR measured on day 4 declined form 3.16 in the first cycle to 2.50 and 1.52 in the 10th and 19th cycles, respectively, suggesting that the metabolic effect was greatest early after treatment initiation and may have partially attenuated during repeated treatment. Glycoalbumin gradually increased from 14.9% to 17.9%.

Conclusion

Capivasertib may induce very early hyperinsulinaemic insulin resistance, even in patients without known diabetes. In this patient, the initial metabolic disturbance improved during the drug-free interval and appeared less pronounced during later cycles. Early, cycle-aware glucose monitoring, including postprandial assessment, should be considered from treatment initiation.

Plain Language summary

Capivasertib is a cancer drug that blocks AKT signalling. Because AKT is also involved in insulin action, this drug can affect glucose metabolism. We describe a woman without previously diagnosed diabetes who developed high post-meal glucose very soon after starting capivasertib. Her baseline glycated haemoglobin was 5.6%, although as oral glucose tolerance test was not performed before treatment.

Blood tests showed high insulin levels and preserved C-peptide levels, suggesting that the glucose abnormality was mainly related to reduced insulin sensitivity rather than impaired insulin secretion. The changes were most marked during the first treatment medication. Similar but less pronounced changes were observed during later treatment cycles.

This case suggests that glucose abnormalities can occur soon after capivasertib is started, even in patients without known diabetes. Glucose monitoring should therefore begin at treatment initiation and should take account of both post-meal glucose levels and the intermittent treatment schedule.

Key Message

Why carry out this study?:

Capivasertib-associated hyperglycaemia is recognised, but the early biochemical features of insulin resistance and their changes across treatment cycles remain incompletely characterised in patients without previously diagnosed diabetes.

What was learned from this case?:

A woman without previously diagnosed diabetes developed postprandial hyperglycaemia soon after starting capivasertib, accompanied by high fasting insulin, preserved C-peptide, and increased HOMA-IR, suggesting reduced insulin sensitivity rather than impaired insulin secretion.

What was distinctive?:

The metabolic changes were most marked during the first treatment cycle and improved during the drug-free interval without glucose-lowering medication. Similar but less pronounced changes were observed during later treatment cycles.

What are the clinical implications?:

Glucose monitoring should begin when capivasertib is initiated and should include postprandial measurements and assessments during both dosing and drug-free days, even in patients without known diabetes.

Introduction

Capivasertib is a selective oral AKT inhibitor that, in combination with fulvestrant, improves progression-free survival in patients with hormone receptor-positive, human epidermal growth factor receptor 2 (HER2)-negative advanced breast cancer after progression during or after endocrine therapy [1, 2]. In this context, hormone receptor-positive disease generally refers to tumours expressing the oestrogen receptor and/or progesterone receptor, and fulvestrant acts as an oestrogen receptor antagonist and degrader. Because AKT is a central mediator of insulin signalling, inhibition of this pathway can reduce insulin-stimulated glucose uptake and alter hepatic glucose handling, resulting in hyperglycaemia as an on-target adverse effect [3–5].

In clinical trial and safety analyses, hyperglycaemia has been recognised as a common adverse event of capivasertib, with a median time to onset of approximately 15 days, although onset may occur as early as day 1 [2, 6]. Recent case reports and case series have described severe capivasertib-associated hyperglycaemia, including diabetic ketoacidosis, life-threatening hyperglycaemia, and characteristic glucose excursions during dosing days captured by continuous glucose monitoring (CGM) [7–11]. However, the early biochemical phenotype and its evolution during repeated treatment remain incompletely characterized.

We report a patient without known diabetes who developed rapid glycaemic deterioration immediately after starting capivasertib. Serial measurements demonstrated an early increase in fasting insulin and HOMA-IR, improvement during a drug-free interval, and lower HOMA-IR values during later treatment cycles.

Case presentation

A 43-year-old Japanese woman with a history of right breast cancer, previously treated with surgery and chemoradiotherapy, developed multiple bone metastases. No liver metastases or other visceral metastases were detected. Her tumour was oestrogen receptor-positive, progesterone receptor-positive, and HER2-negative. Genomic testing identified a PIK3CA mutation. She had no history or family history of diabetes and no previously documented dysglycaemia. Her height was 155.1 cm, body weight 48.2 kg, body mass index 20.0 kg/m², blood pressure 110/73 mmHg, and body temperature 36.2 °C. Baseline fasting plasma glucose was 89 mg/dL and glycated haemoglobin was 5.6%. Although the glycated haemoglobin value was within the non-diabetic range, it was at the upper end of normal and did not exclude impaired glucose tolerance; a pre-treatment oral glucose tolerance test was not performed. Renal function was preserved, with serum creatinine 0.43 mg/dL and estimated glomerular filtration rate 123 mL/min/1.73 m². Capivasertib was initiated on an intermittent schedule of 4 days on treatment followed by 3 days off treatment in combination with endocrine therapy. Soon after initiation, postprandial hyperglycaemia was detected, and she was referred to our department for evaluation.

Investigation

Laboratory findings before and shortly after starting capivasertib are summarised in Table 1. On the day after starting treatment initiation, the 2-h postprandial plasma glucose level was 256 mg/dL. On day 2, fasting plasma glucose was 103 mg/dL, fasting insulin was 22.8 µU/mL, C-peptide was 3.40 ng/mL, glycoalbumin was 14.9%, HOMA-IR was 5.80, HOMA-β was 205.2, and the C-peptide index was 3.30. The C-peptide index, calculated as fasting C-peptide (ng/mL) × 100 / fasting plasma glucose (mg/dL), is an index of endogenous insulin secretory capacity rather than insulin sensitivity. These findings were compatible with hyperinsulinaemic insulin resistance and preserved endogenous insulin secretion. No infection or clinically relevant deterioration in hepatic or renal function was observed that could otherwise explain the abrupt metabolic change.

Table 1.

Laboratory findings before and shortly after initiation of capivasertib

Parameter	Before treatment	Day 1 after meals	Day 2 fasting	Unit	Reference range
WBC	3,100	3,700	3,300	/µL	3,600-8,900
Haemoglobin	13.3	12.1	13.6	g/dL	11.1–15.2
AST	26	22	55	IU/L	13–33
ALT	12	13	19	IU/L	8–42
Creatinine	0.43	0.52	0.57	mg/dL	0.47–0.79
eGFR	123	100	90	mL/min/1.73 m²	≥ 60
Plasma glucose	89	149	103	mg/dL	70–109 (fasting)
2-h postprandial glucose	-	256	-	mg/dL	< 140
HbA1c	5.6	-	-	%	4.7–6.2
Glycoalbumin	-	-	14.9	%	11.0–16.0
Fasting insulin	-	-	22.8	µU/mL	2.0–10.0
C-peptide	-	-	3.40	ng/mL	0.60–1.80

Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; eGFR, estimated glomerular filtration rate; HbA1c, glycated haemoglobin; WBC, white blood cell count

Serial changes in glucose metabolism across treatment cycles are shown in Table 2. During the first cycle, HOMA-IR was 5.80 on day 2 and 3.16 on day 4. At the beginning of the third cycle, following a drug-free interval, fasting plasma glucose was 87 mg/dL, insulin 5.4 µU/mL, C-peptide 1.15 ng/mL, and HOMA-IR 1.16. On day 4 of the 10th and 19th cycles, HOMA-IR was 2.50 and 1.52, respectively. Thus, the largest increase in HOMA-IR occurred immediately after initial exposure, whereas the day-4 values were progressively lower during later cycles. Glycoalbumin increased from 14.9% during the first cycle to 17.9% in the 19th cycle. HbA1c was not measured during follow-up. Apart from the initial 2-h postprandial glucose value of 256 mg/dL, serial postprandial glucose measurements were not available.

Table 2.

Serial changes in glucose metabolism and insulin resistance indices across capivasertib treatment cycles

Cycle/ time point	Treatment day	Capivasertib	FPG (mg/dL)	Insulin (µU/mL)	C-peptide (ng/mL)	HOMA-IR	HOMA-β	CPI	GA (%)
Baseline	Before cycle 1	Not started	89	-	-	-	-	-	-
1st cycle	Day 2	On	103	22.8	3.40	5.80	205.2	3.30	14.9
1st cycle	Day 4	On	90	14.2	3.07	3.16	189.3	3.41	15.1
3rd cycle	Day 1 (after drug-free interval)	On	87	5.4	1.15	1.16	81.0	1.32	15.8
10th cycle	Day 4	On	81	12.5	2.68	2.50	250.0	3.31	17.3
19th cycle	Day 4	On	77	7.5	1.45	1.52	192.9	1.97	17.9

Capivasertib was administered on a 4-days-on/3-days-off schedule. No glucose-lowering medication was used during the observation period. HOMA-IR was calculated as fasting plasma glucose (mg/dL) × fasting insulin (µU/mL) / 405. CPI was calculated as fasting C-peptide (ng/mL) / fasting plasma glucose (mg/dL) × 100

Abbreviations: CPI, C-peptide index; FPG, fasting plasma glucose; GA, glycoalbumin; HOMA-β, homeostasis model assessment of beta-cell function; HOMA-IR, homeostasis model assessment of insulin resistance

Treatment

The patient continued oncological treatment with capivasertib under metabolic observation. No glucose-lowering medication, including insulin or oral antidiabetic agents, was administered during the observation period.

Outcomes and follow-up

The most prominent metabolic deterioration occurred immediately after capivasertib initiation. Fasting insulin and HOMA-IR subsequently improved during a drug-free interval. HOMA-IR values measured on day 4 were lower in the 10th and 19th cycles than in the first cycle, indicating that the early effect was not sustained at the same magnitude during repeated administration. The mechanism of this apparent attenuation could not be determined from this single case. Despite relatively stable fasting plasma glucose, glycoalbumin gradually increased, raising the possibility of persistent or recurrent postprandial dysglycaemia; however, this could not be confirmed because serial postprandial glucose measurements and CGM data were unavailable.

Discussion

This case provides clinically relevant observations regarding the timing and evolution of capivasertib-associated dysglycaemia.

First, the glycaemic disturbance occurred very early. The 2-h postprandial plasma glucose level increased to 256 mg/dL on the day after treatment initiation. The baseline glycated haemoglobin level of 5.6% was within the non-diabetic range, but it was relatively high within that range and did not exclude pre-existing impaired glucose tolerance. Because an oral glucose tolerance test was not performed before capivasertib therapy, normal glucose tolerance cannot be proven. Nevertheless, the normal fasting plasma glucose, absence of a history or family history of diabetes, and abrupt temporal association support a substantial treatment-related component.

Second, the biochemical pattern was compatible with acute insulin resistance rather than insulin deficiency. On day 2, fasting insulin was 22.8 µU/mL and C-peptide was 3.40 ng/mL, with HOMA-IR 5.80 and C-peptide index 3.30. The C-peptide index reflects endogenous insulin secretion and does not directly measure insulin sensitivity. AKT signalling promotes insulin-stimulated glucose transport, including GLUT4 translocation in skeletal muscle and adipose tissue, and contributes to hepatic glycogen synthesis through downstream regulation of glycogen synthase kinase 3 [3–5]. AKT inhibition can therefore impair peripheral glucose disposal and hepatic glucose handling, producing compensatory hyperinsulinaemia and postprandial hyperglycaemia.

Third, the magnitude of insulin resistance appeared to change over time. HOMA-IR was highest on day 2 of the first cycle and decreased by day 4. The day-4 HOMA-IR values were 3.16 in the first cycle, 2.50 in the 10th cycle, and 1.52 in the 19th cycle. This pattern suggests that the metabolic effect was greatest during initial exposure and may have partially attenuated during repeated treatment. Potential explanations include physiological adaptation to intermittent AKT inhibition, changes in diet or physical activity, altered drug exposure, changes in cancer-related inflammation or tumour burden, and day-to-day variability in fasting insulin. Because measurements were obtained at limited and non-identical time points and no mechanistic testing was performed, these possibilities remain speculative. The declining HOMA-IR values should therefore not be interpreted as proof of pharmacological tolerance.

No liver metastases were present in this patient. This is relevant because hepatic tumour involvement could independently alter hepatic glucose metabolism. Nevertheless, hepatic glucose uptake and endogenous glucose production were not directly assessed. Future studies combining metabolic tracers, hyperinsulinaemic-euglycaemic clamps, or dynamic 18 F-fluorodeoxyglucose positron emission tomography could help distinguish hepatic from peripheral effects of AKT inhibition [12].

Recent reports have expanded the recognised spectrum of capivasertib-associated dysglycaemia. Diabetic ketoacidosis has been described in patients without prior diabetes, severe hyperglycaemia with high insulin requirements has been reported, and CGM has demonstrated marked glucose excursions during dosing days [7–11]. The present case complements these reports by showing serial insulin and C-peptide measurements, a marked early increase in HOMA-IR, improvement during a drug-free interval, and an apparent reduction in the magnitude of HOMA-IR elevation during later cycles.

Glucose monitoring should begin at capivasertib initiation, including in patients without known diabetes. Because fasting plasma glucose may remain relatively stable despite postprandial abnormalities, structured postprandial measurements or CGM may be more informative. Baseline risk assessment may also include glycated haemoglobin, fasting glucose, family history, body mass index, and, when clinically appropriate, an oral glucose tolerance test.

This report has several limitations. It describes a single patient. A pre-treatment oral glucose tolerance test was not performed, and therefore impaired glucose tolerance before therapy cannot be excluded. HbA1c was not measured during follow-up. Serial postprandial glucose measurements and CGM were unavailable, so the amplitude and duration of glucose excursions could not be assessed. Insulin resistance was inferred from fasting insulin and HOMA-IR and was not directly measured by a hyperinsulinaemic-euglycaemic clamp. HOMA-IR measurements were limited to selected time points, and possible changes in diet, physical activity, body composition, cancer status, concomitant therapy, and capivasertib exposure were not systematically evaluated. Accordingly, the apparent attenuation of HOMA-IR during later cycles cannot be attributed to a specific mechanism.

Conclusions

Capivasertib was associated with very early hyperinsulinaemic insulin resistance in a patient without known diabetes. The metabolic disturbance improved during a drug-free interval, and HOMA-IR values were lower during later treatment cycles than during the first cycle. Although the mechanism of this apparent attenuation is uncertain, the case supports early and cycle-aware glucose monitoring, with attention to postprandial glucose levels.

Author contributions

YM and HS managed the patient and wrote the manuscript. SN, SK, and DS contributed to the acquisition and interpretation of the data. HS conceived the report, supervised clinical management, critically reviewed the manuscript, and approved the final version. All authors approved the submitted version.

Funding

No funding or sponsorship was received for this case report or publication of this article.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethical approval and consent to publish

Written informed consent for publication of this case report was obtained from the patient.

Competing interests

The authors declare no competing interests.