Triple combination cystic fibrosis transmembrane receptor modulator effects on glycaemia and insulin kinetics in cystic fibrosis with and without diabetes

Abstract

Background

Greater insight into the effects of cystic fibrosis (CF) transmembrane modulators such as elexacaftor‐tezacaftor‐ivacaftor (ETI) on glucose metabolism can support a more dynamic and individualised approach to CF‐related dysglycaemia.

Aim

To assess changes in glycemia, body weight and pulmonary function in adults with CF with and without cystic fibrosis related diabetes (CFRD) following ETI.

Methods

This single‐centre retrospective cohort study included adults with at least one F508del mutation (23 with CFRD, 38 without CFRD). Exploratory outcomes were HbA1c and total daily insulin doses in the CFRD group and glucose and insulin profiles (0–120 min oral glucose tolerance test [OGTT]), HOMA2‐derived indices of insulin resistance (IR), secretion (%S), and β‐cell function (%B) in the non‐CFRD group. Pulmonary function and weight were assessed in both groups.

Results

Mean age was 33.8 ± 10.0 years (CFRD) and 32.8 ± 8.8 years (non‐CFRD), with median ETI exposure of 12.3 and 13.4 months, respectively. In the CFRD cohort, HbA1c decreased non‐significantly (6.6% to 6.1%, p = 0.2) with no significant change in insulin dose (0.28 to 0.20 IU/kg/day, p = 0.1). In non‐CFRD participants, ETI significantly reduced 1‐ and 2‐hour OGTT glucose levels (9.6 vs. 9.2 mmol/L at 1 h, p < 0.01; 6.1 vs. 4.8 mmol/L at 2 hours, p = 0.01), with 13/38 participants improving OGTT classification; no progression to CFRD occurred. Lung function improved in both groups.

Conclusions

Adults with at least one F508del mutation, ETI was associated with a clinically relevant reduction in HbA1c in those with CFRD and improved glucose tolerance in those without, supporting favourable metabolic effects beyond established respiratory benefits.

Cystic fibrosis‐related diabetes (CFRD) affects up to 50% of people with cystic fibrosis (pwCF) and is associated with pulmonary decline, weight loss and increased mortality. ¹ As survival rates improve in the era of highly effective cystic fibrosis transmembrane receptor (CFTR) modulator therapy, including elexacaftor–tezacaftor–ivacaftor (ETI), focus is increasingly shifting toward long‐term metabolic health. Greater insight into the effects of CFTR modulators on glucose tolerance can support an individualised approach to dysglycaemia in this population.

Emerging evidence suggests that ETI may reduce the burden of abnormal glucose metabolism in pwCF by improving glucose tolerance, lowering haemoglobain A1c (HbA1c) CFRD, and decreasing exogenous insulin requirements. ² , ³ , ⁴ , ⁵ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² There are no glycaemic outcome data from Australia's CF population. In this context, our study aimed to elucidate the impact of ETI on glycaemia, pulmonary function and body weight in individuals with and without CFRD, with separate analyses of glucose tolerance and insulin kinetics in those without CFRD.

We conducted a single‐centre, retrospective cohort study at an urban tertiary referral hospital in Sydney which provides a multidisciplinary service for pwCF ≥16 years. Exploratory outcomes for the CFRD group were HbA1c and total daily insulin dose. In the non‐CFRD group, outcomes were glucose and insulin concentrations from 2‐h 75 g oral glucose tolerance tests (OGTTs), including 30‐min glucose concentration, insulin resistance (HOMA2‐IR), insulin secretion (HOMA2‐%S) and β‐cell function (HOMA2‐%B). Other outcomes in both groups were change in pulmonary function, measured by forced expiratory volume in 1 s (FEV1) and per cent predicted FEV1 (ppFEV), body weight (kg) and body mass index (BMI, kg/m²).

CFRD was defined as either a fasting plasma glucose concentration >7.0 mmol/L and/or >11.0 mmol/L at 2 h. ¹ Normal glucose tolerance (NGT) was defined as fasting glucose ≤6.0 mmol/L, <11.1 mmol/L at 1 h and <7.8 mmol/L at 2 h; indeterminant glucose tolerance as fasting glucose ≤6.9 mmol/L, >11.1 mmol/L at 1 h and ≤7.8 mmol/L at 2 h; impaired glucose tolerance (IGT) as fasting glucose ≤7.0 mmol/L and 7.8–11.1 mmol/L at 2 h.

Descriptive statistics were calculated and presented as mean ± standard deviation if parametrically distributed, as assessed by quantile‐quantile plots, or median (interquartile range (IQR)) for non‐parametric data. Fasting glucose and insulin concentrations were used to calculate HOMA2 values, obtained using the HOMA calculator. ¹³ Continuous variables were analysed by the paired Student's t test (parametric) or Wilcoxon test (non‐parametric). As HOMA2 indices are internally correlated, P‐values were adjusted using the Benjamini–Hochberg procedure to minimise the probability of a false positive finding. P‐values <0.05 indicated significance. Analyses were conducted using the statistical software Jamovi. ¹⁴

Between August 2018 and August 2023, 126 adults with CF and at least one copy of the F508del mutation were eligible for the study. Of these, 66 participants were excluded: 57 had incomplete OGTT, one had type 1 diabetes and seven were pregnant. Baseline clinical characteristics of enrolled participants (23 CFRD, 38 non‐CFRD) are summarised in Table 1. All glycaemic parameters were recorded with participants off steroids without intercurrent illness. Participants receiving pancrelipase were classified as having pancreatic insufficiency. Baseline OGTTs were performed 24.7 ± 15.2 months before ETI and repeated 14.3 ± 7.4 months after initiation.

Table 1.

Baseline demographics

	CFRD (n = 23)	Non‐CFRD (n = 38)	P‐value
Age, years (SD)	33.8 ± 10.0	32.8 ± 8.9	0.8
Females, %	47.8	47.1	0.9
Duration on ETI, months (IQR)	12.3 (11.7, 15.6)	13.4 (10.0, 16.2)	0.8
Duration of CFRD at ETI commencement, years ± SD	13.4 ± 6.2	—	—
F508del mutation status
Homozygous, n (%)	11 (47.8)	20 (52.6)	0.7
Heterozygous, n (%)	12 (52.1)	17 (44.7)	0.5
Unknown, n (%)	0	1 (<0.1)	1.0
Previous CFTR modulator use, n (%)	8 (34.8)	20 (52.6)	0.2
Pancreatic insufficiency, n (%)	23 (100)	33 (86.8)	0.1
Weight, kg (SD)	65.5 ± 14.3	67.6 ± 13.1	0.6
BMI, kg/m2 (SD)	23.5 ± 3.8	23.3 ± 4.1	0.8
FEV1, L (SD)	2.0 ± 0.9	2.9 ± 1.1	<0.01
FEV1 predicted, % (SD)	55.8 ± 21.0	75.4 ± 24.9	<0.01

Abbreviations: BMI ‐ body mass index; CFRD ‐ cystic fibrosis‐related diabetes; CFTR ‐ cystic fibrosis transmembrane receptor; ETI ‐ elexacaftor‐tezacaftor‐ivacaftor; FEV1 ‐ forced exporatory volume at 1s; IQR ‐ interquartile range; SD ‐ standard deviation.

CFRD outcomes were assessed median 12.3 months (IQR 11.7–15.6) after ETI initiation (Table 2). The change in HbA1c of 0.5% did not reach statistical significance (6.6% (49 mmol/mol) vs 6.1% (43 mmol/mol), P = 0.2). There was a non‐significant decrease in total insulin dose (19.0 vs 16.0 IU, P = 0.5) and weight‐adjusted dose (0.28 vs 0.20 IU/kg/day, P = 0.1); 3/23 ceased insulin. Mean weight increased by 3.7 kg (65.5–69.2 kg, P = 0.02) and BMI by 1.2 kg/m² (23.5–24.7 kg/m², P = 0.04). Mean FEV1 increased by 0.3 L (P = 0.03) and ppFEV1 by 6.9% (55.8%–62.8%, P = 0.02).

Table 2.

Outcomes on ETI

CFRD group	Pre‐ETI	Post‐ETI	P‐value
Time between HbA1c result and ETI, months (IQR)	1.8 (0.5, 4.4)	12.3 (11.7, 15.6)	—
HbA1c, % (IQR)	6.6 (5.9, 7.6)	6.1 (5.8, 7.3)	0.19
Insulin total daily dose, IU (IQR)	19.0 (5.3, 41.0)	16.0 (2.3, 37.8)	0.50
Weight‐adjusted insulin dose, IU/kg (IQR)	0.3 (0.1, 0.7)	0.2 (0.1, 0.5)	0.14
FEV1, L (SD)	2.0 ± 0.9	2.2 ± 0.8	0.03
FEV1 predicted, % (SD)	55.8 ± 21.0	62.8 ± 20.8	0.02
Weight, kg (SD)	65.5 ± 14.5	69.2 ± 15.2	0.02
BMI, kg/m2 (SD)	23.5 ± 3.9	24.7 ± 4.8	0.04

Non‐CFRD group	Pre‐ETI	Post‐ETI	P‐value
Age, years (SD)	28.7 ± 9.0	32.0 ± 8.8	—
OGTT glucose and insulin level, mmol/L (IQR) and mIU/L (IQR)
0 min	4.8 (4.1, 5.1) 6.0 (4.0, 8.0)	4.8 (4.5, 5.2) 5.0 (1.0, 6.0)	0.14 0.23
30 min	9.2 (7.6, 10.0) 27.5 (17.3, 40.0)	9.0 (7.6, 10.6) 22.0 (22.0, 40.0)	0.69 0.18
60 min	9.6 (7.4, 11.2) 43.5 (37.5, 63.5)	9.2 (6.4, 11.2) 32.0 (41.5, 67.5)	<0.01 0.32
120 min	6.1 (4.7, 8.1) 32.5 (27.5, 45.8)	4.8 (4.0, 6.1) 22.5 (18.0, 35.0)	0.01 0.01
Insulinogenic index, pmol/mmol (IQR)	6.1 (3.2, 9.2)	5.0 (2.7, 10.3)	0.94
FEV1, L (SD)	2.9 ± 1.1	3.2 ± 1.1	0.01
FEV1 predicted, % (SD)	75.9 ± 26.0	82.5 ± 25.9	<0.01
Body weight, kg (SD)	69.0 ± 13.1	72.7 ± 13.0	<0.01
BMI, kg/m2 (SD)	23.3 ± 4.0	24.6 ± 3.8	<0.01

BMI, body mass index; CFRD, cystic fibrosis‐related diabetes; ETI, elexacaftor‐tezacaftor‐ivacaftor; FEV1, forced expiratory volume at 1 s; IGT, impaired glucos tolerance; IQR, interquartile range; OGTT, oral glucose tolerance test; SD, standard deviation; HbA1c, haemoglobain A1c.

In the non‐CFRD group, 0‐ and 30‐min glucose levels were unchanged on OGTT. Median glucose concentration decreased at 1 h (9.6 vs 9.2 mmol/L, P < 0.01) and 2 h (6.1 vs 4.8 mmol/L, P = 0.01), with a corresponding reduction in 2‐h insulin level (32.5 vs 22.5 mIU/L, P = 0.01). Improved glycaemia was reflected in re‐classification of OGTT categories on ETI: 13/38 participants improved, 20 unchanged and five previously NGT deteriorated. Notably, 13/14 participants with indeterminate or IGT improved, with 10/14 participants reverting to normoglycaemia. No participants progressed to CFRD on ETI.

Insulin kinetics showed reduced HOMA2‐%B (88.1 to 78.4), significant on unadjusted analysis (P = 0.04) but not after Benjamini–Hochberg correction (P = 0.08). There were no changes in HOMA2‐IR (0.8 vs 0.7, P = 0.4) or HOMA‐%S (131.2 vs 151.2, P = 0.6). FEV1 improved by 0.26 L (2.91 vs 3.17 L, P = 0.01) and ppFEV1 by 6.6% (75.9% vs 82.5%, P < 0.01). Weight increased by 3.7 kg (69.0 vs 72.7 kg, P < 0.01), BMI by 1.3 kg/m² (23.3 vs 24.6 kg/m², P < 0.01).

Discussion

This study demonstrated that among individuals with abnormal glucose tolerance without CFRD, 34% improved OGTT categories following initiation of ETI. On OGTT, 2‐h glucose and insulin concentrations decreased significantly, with improvement in β‐cell function without change in insulin sensitivity. There was no change in fasting glucose. Within the CFRD cohort, ETI was not associated with significant changes in HbA1c or insulin requirements.

In the literature, ETI have been associated with modest but statistically significant HbA1c reductions of 0.1%–0.4% in individuals with and without impaired glucose tolerance or CFRD, sustained for up to 2 years on ETI. ² , ⁶ , ⁷ , ⁹ , ¹⁰ , ¹⁵ , ¹⁶ However, HbA1c is a relatively blunt measure of glycaemic change potentially impacted by anaemia. ¹⁷ Although the HbA1c reduction in our cohort did not reach statistical significance, a 0.5% decrease is considered clinically relevant and has been accepted as evidence of therapeutic efficacy in intervention studies. ¹⁸

The clinical impact of ETI is further supported by emerging data that suggest ETI can reduce insulin requirement, potentially mediated by improved insulin secretion. In our cohort, 3/23 ceased insulin and 13/23 reduced their weight‐adjusted insulin requirement. Insulin dose changes in the literature are mixed. Similar to the present study, a retrospective analysis involving 107 adults with CFRD observed a total insulin dose reduction of 0.07 IU/kg/day, driven by a decrease of bolus insulin, indicating a change in post‐prandial glucose levels rather than a change in fasting concentrations. ⁵ While not significant, the insulin dose decrement in our study was similar, 0.1 IU/kg/day. In contrast, one study found no change in basal insulin after 6 months on ETI, ¹¹ while others have reported reductions in basal insulin doses. ⁶ The heterogeneity in outcomes reflects differences in cohorts with a mix of CFRD and non‐CFRD participants, variation in study design and follow‐up duration.

Improvements in glucose levels observed during OGTT may be explained by improved first phase insulin secretion, leading to suppression of prandial glucose excursion at 1 and 2 h. This was without change in insulin resistance or sensitivity measures, suggesting enhanced β‐cell secretion within the pancreas. A recent meta‐analysis showed significantly reduced 2‐h glucose (mean difference −1.35 mmol/L) on ETI, without affecting fasting levels. ¹⁶ Other studies similarly report improved insulin secretion and lower 2‐h glucose following a glucose load. ³ , ¹⁹ By contrast, the PROMISE Endocrine substudy prospective observational study found only fasting glucose decreased significantly, postulated to be caused by decreased hepatic gluconeogenesis and glucagon. ²

Reports of changes in HOMA2 indices are varied in the literature. Chan et al. evaluated 20 pwCF for 12 months on ETI and found measures of insulin secretion and insulin resistance increased without net change in β‐cell function. ¹⁰ Other studies reported a non‐significant decrease in HOMA2‐%B but not HOMA2‐%S. ⁷ , ⁹ Mixed results reflect competing physiological processes, where increased adiposity from improved nutrition may offset gains in insulin sensitivity. These discrepancies also reflect limitations of HOMA2, whereby fasting measures do not capture dynamic changes in insulin secretion.

Among participants without CFRD, most individuals (10/14) with indeterminate glucose tolerance or IGT reverted to NGT. Similar improvements in the OGTT category following ETI have been reported. In PROMISE Endocrine substudy, 25% demonstrated improvement in glucose category at 12–18 months, in a mixed cohort which included those with and without CFRD. ² Galderisi et al. found 31% of adolescents with indeterminate glucose tolerance and 44% with IGT returned to NGT at 12 months. ⁴ Observed glycaemic improvements likely stem from enhanced insulin secretion and reduced systemic demand following improved lung function and decreased inflammation. While ETI‐related weight gain may counteract benefits by promoting insulin resistance, this was not evident in our study. ²⁰

Several limitations should be acknowledged. Although our sample size is comparable to prior studies, the lack of statistical significance in some outcomes may reflect limited power. Post hoc analysis indicated 73% power to detect a 0.5% HbA1c change in the CFRD group; 27 participants would be required for 80% power. However, our sample size was limited due to the retrospective design. Additionally, the study was conducted during COVID, lengthening the interval between baseline OGTT and after ETI initiation. Prior exposure to dual CFTR modulators may have attenuated baseline differences. Continuous glucose‐monitoring devices would potentially provide different glycaemic insights, but they are not subsidised for individuals aged 21 years or older, preventing use within this cohort.

Despite this, our findings add to growing evidence that glucose metabolism can improve with ETI in pwCF. Rather than assuming inevitable progression, clinicians should consider reassessing glycaemic status after modulator initiation, including repeat oral glucose tolerance testing and careful insulin titration to minimise hypoglycaemia risk. In some cases, insulin therapy may be safely de‐escalated, helping to reduce treatment burden.

In summary, adults with at least one F508del mutation treated with ETI showed improved glycaemia and pulmonary function despite weight gain, including a clinically meaningful reduction in HbA1c in the CFRD cohort and improved glucose tolerance with reduced β‐cell secretory demand in those without CFRD. If these findings can be consistently shown in larger cohorts, screening and treatment guidelines may need to evolve to reflect a more dynamic understanding of glucose metabolism in CF.

Data availability statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.