Elexacaftor-tezacaftor-ivacaftor pharmacokinetics with concurrent tacrolimus administration after lung transplant

Abstract

BACKGROUND:

CFTR modulators in post-transplant people with cystic fibrosis (pwCF) are less frequently used due to uncertainty regarding effectiveness and interactions with immunosuppressive agents. Elexacaftor/tezacaftor/ivacaftor (ETI) is a triple combination cystic fibrosis (CF) therapeutic with benefits in multiple organ systems where complications can impact lung transplant (LTx) outcomes, including malnutrition, diabetes, and sinus disease. ETI use in LTx recipients is variable.

METHODS:

We conducted a pharmacokinetics (PK) study of concentrations of ETI parent compounds and the four major metabolites (M23-ELX, M1-TEZ, M1-IVA, M6-IVA) in a prospective non-randomized observational study, with all transplant participants concomitantly taking tacrolimus for LTx immunosuppression and excluded if taking any other medication with known interactions (e.g., azole antifungals) and compared to a non-transplant group of pwCF. We completed non-compartmental analysis (NCA) for both groups and compared the transplant to non-transplant PK parameters, as well as to published data from the manufacturer for non-transplant pwCF. Area under the curve (AUC), average concentrations (C_avg), minimum and maximum concentrations, clearance, and other parameters were determined.

RESULTS:

Twelve transplant and fourteen non-transplant participants with CF completed the study. There were no significant differences between the mean values for any PK parameters for the transplant and non-transplant groups and no substantial differences in frequency of concentrations outside the therapeutic ranges in the two groups.

CONCLUSIONS:

Our data suggest there are not significant differences in concentrations of ELX, TEZ, IVA, or their major human metabolites in LTx recipients compared to non-transplant pwCF.

Introduction.

Cystic fibrosis transmembrane conductance regulator (CFTR) modulators are a major advance in the care of people with cystic fibrosis (CF) and contribute to life expectancy estimates that now exceed 60 years of age.¹ Despite the significant improvements seen among the general CF population with elexacaftor/tezacaftor/ivacaftor (ETI), a highly effective CFTR modulator therapy (HEMT), there is uncertainty regarding use of these therapies in CF recipients of solid organ transplant. CF is a multi-system disease, and HEMT has shown significant improvements in many organ systems^2–10, suggesting that patients who receive transplant in one system would still be likely to benefit from HEMT treatment in others. While transplanted lungs may not benefit from HEMT, non-pulmonary manifestations of CF including diabetes, sinus disease, GI symptoms, and malnutrition remain relevant after lung transplant.

Transplant recipients were omitted from drug development studies of ETI.^11–15 Transplant recipients are typically treated with multiple medications including calcineurin inhibitors (CNI), other agents for immunosuppression, and medications for prevention of opportunistic infections (such as azole anti-fungal medications), which pose significant risk for potential drug interactions with ETI. There are few studies looking at patients who have had a solid organ transplant from which to draw conclusions, and those that have been published are limited to case reports, case series, single-center observations, and an observational cohort study that did not directly address drug interactions.^16–21 Several recent studies of patients taking tacrolimus provided evidence that concurrent use of ETI with tacrolimus results in higher tacrolimus concentrations which can be managed with standard of care monitoring of tacrolimus and adjustments of tacrolimus dosing, but none evaluated ETI concentrations to conclude therapeutic ETI concentrations were maintained.^22–26 While some studies report benefit of ETI in transplant patients, others report intolerance and uncertain benefit,^{21, 26} which could be related to suboptimal ETI concentrations below a therapeutic range. Without a clear understanding of ETI pharmacokinetics in this population, there is an uncertain risk/benefit balance, leading many to avoid the combination.^{27, 28}

This study was designed as a prospective pharmacokinetic (PK) study of ETI concentrations to evaluate whether they differ in the transplant compared to non-transplant group and whether they may be outside of the range of expected effective concentrations. Concentrations that are too low place patients at risk for suboptimal response, whereas concentrations that are too high risk the development of side effects.^{21, 29}

Methods.

Human subjects protections and patient population.

Enrollment was completed at two primary sites (University of Washington, UW, and University of Alabama at Birmingham, UAB) after informed consent was obtained under IRB approval # 300001194 (UAB) and #STUDY14660 (UW). Transplant participants were diagnosed with CF, were lung transplant recipients, and were clinically prescribed ETI by their providers. Non-transplant participants had CF but were not post-transplant. Transplant participants were all taking tacrolimus for inclusion in the study. Among transplant and non-transplant participants, those prescribed systemic azole antifungals or other drugs that could interact with ETI (e.g., diltiazem, sirolimus) were excluded. Azoles were specifically excluded because they are inhibitors of the cytochrome P450 system (CYP3A) which metabolize tacrolimus and ETI compounds. All medications used by potential participants were assessed during screening. Participants were instructed to take ETI and tacrolimus (if applicable) as per manufacturer’s recommendations (every 12 hours) at the standard dose during the run up to our study, which included recommendations to consume ETI at regular intervals with a fat-containing meal. Participants were also instructed to remain on the same dose of tacrolimus, if applicable, for at least two weeks to ensure steady state levels were obtained and to continue taking all other prescribed medications (including pancreatic enzymes). There was one study visit. We collected 9 samples over 12 hours (0, 1, 2, 4, 5, 6, 8, 10, 12h after the morning dose), which was feasible for collection and prior data showing it was possible to flip the 0h concentration to estimate PK parameters despite q24h dosing for two of the compounds (elexacaftor and tezacaftor). We collected clinical outcome data including sweat chloride (SC) post-ETI use and clinical data from the clinic visit closest to the time of sample collection.

Pharmacokinetic analysis.

Quantitation of ivacaftor (iva), elexacaftor (elx), tezacaftor (tez) and their major human metabolites iva-M1, iva-M6, elx-M23, and tez-M1 were performed as previously described.³⁰ Briefly, two LC-MS/MS assays fully validated to FDA standards were used to simultaneously quantify iva, elx, and tez; a separate fully validated assay was used for the four metabolites. After quantitation, we performed non-compartmental analysis (NCA) using Phoenix 8.1 (Certara, Princeton, NJ). Because we sampled only 12 hours due to feasibility, we flipped the 0h concentration to 24h for elx and tez. We calculated the Area Under the Curve at the end of the dosing interval, at steady state (AUC_τ,ss), time to reach maximum concentration (T_max), minimum and maximum concentrations in plasma (C_min and C_max), half-life (t_1/2), oral clearance (CL/F), and distribution volume (V/F). We also calculated the clearance for each metabolite (adjusted for molecular weight), and the metabolite:parent ratio (MPR) for each compound (a measure of P450 metabolism activity).

Statistical analysis.

Initial sample size determination for this study was determined from previously published work.31, 32 Means and standard deviations were obtained for each parameter among transplant recipients, non-transplant recipients, and compared to those described from the manufacturer. Welsch’s t-test with Sattherthwaite correction for degrees of freedom, which does not assume equal variance, are used to determine if the transplant and non-transplant group are significantly different from the manufacturer. Equivalence between manufacturer and the study groups was determined by using the two one-sided tests (TOST) procedure.³³ After non-compartmental analysis, we obtained PK parameters for the non-transplant and transplant cohorts to directly compare. Normality testing was applied. Measures of concentration (C_max, C_min, C_average, and AUC) were overall normally distributed for most compounds; however, one or more variables failed normality testing for some compounds. Other parameters were not normally distributed. We performed two-tailed t-tests for all parameters comparing the transplant group to the non-transplant group, and statistical significance is determined using a two tailed t-test of p<=0.05. We performed non-parametric comparisons of all parameters between non-transplant and transplant groups, given the inconsistent distributions, which showed the same results. As a sensitivity analysis, we also conducted linear mixed effect modeling using all available concentrations at each time point to determine if the concentrations of either parent or metabolites were different between non-transplant and transplant groups.

Results.

Demographics of the study groups.

We recruited 14 PwCF who had not undergone transplantation and 12 PwCF who had undergone lung transplantation. Demographic details are described in Table 1. The transplant population was older, with a greater proportion of male participants than the non-transplant group. Both groups were evenly divided as to CFTR mutation status, and the majority of all participants in both groups were pancreatic insufficient. Racial distribution, lung function, sweat chloride pre-modulator and post-modulator were similar. For transplant group participants for whom tacrolimus levels were available to study investigators at the time of the study visit (n=10), levels were at goal.

Table 1.

Demographics.

	All (n=26)	Non-Transplant (n=14)	Transplant (n=12)
Age, median (range)	37.5(14.3–52.3)	30.3(14.3–52.3)	40.05(32–51)
Female, n (%)	17(63.38)	11(78.5)	6(50)
Genotype
F508del CFTR homozygote, n (%)	13(50)	7(50)	6(50)
F508del heterozygote, n (%)	13(50)	7(50)	6(50)
Other, n (%)	0	0	0
Pancreatic insufficient, n (%)	25(96.1)	13(92.85)	12(100)
Race, n (%)
Non-Hispanic White	25(96.1)	13(92.85)	12(100)
Hispanic White	1(3.84)	1(7.14)	0
Non-Hispanic Black/African American	0	0	0
FEV1 % Predicted mean (SD)	73.25(25.90)	75.76(28.86)	72.85(18.78)
Sweat chloride (mmol/L), mean (SD)
Pre-modulator (baseline)	111(19.99)	111.55(13.94)	111.45(24.17)
Post-modulator (ETI)	53.18 (25.26)	51.17 (28.56)	55.6 (21.93)
Time on ETI, median days (range)	735(31–5552)	789(306–5552)	579(49–1607)
Tacrolimus concentration (ng/mL), mean (SD)	N/A	N/A	6.5 (1.5)
Lung transplant years, median (range)	N/A	N/A	12.5 (4–20)

Comparing transplant and non-transplant pharmacokinetic parameters determined by NCA.

Figure 1 shows the concentrations over time throughout the sampling window. The PK parameters of both transplant and non-transplant groups are shown in Table 2 and Figure 2. In this study we assessed Cmin, Cmax, Cavg, AUC_τ, Tmax, half-life, clearance, and volume of distribution for all compounds. No significant differences were seen with these comparisons, as shown in Figure 2. Insufficient data is publicly available to analyze the differences from the manufacturer’s population PK model estimates, but for qualitative comparison, the manufacturer’s data for AUCτ,ss were 162 ug*h/mL, 93.4 ug*h/mL, and 11.7 ug*h/mL for ELX, TEZ, and IVA after standard dosing at steady state.³⁴ These values are similar to what we see in our study. Supplemental Table 1 shows all NCA data for metabolites. Direct comparison of the transplant and non-transplant groups were made for all PK parameters (Figure 2). There were no significant differences between groups for any parameter for any compound. Although not statistically significant, it is possible that there are true differences for some of these data, as some of these comparisons are underpowered (see Results section on sample size re-estimation below) due to higher than expected variability in this real-world population. Linear mixed modeling also showed no significant differences between the transplant and non-transplant groups for any of the compounds (parent or metabolite). There were no significant differences in metabolite PK parameters or MPR between transplant and non-transplant (supplemental Table 1).

Figure 1. Concentration-time curves for (A, D) ELX, (B, E) TEZ, and (C, F) IVA.

All participants were recruited for 9 samples over 12 hours due to feasibility. Solid lines show non-transplant participants, and dotted lines show transplant participants. The 0h concentration for ELX and TEZ was flipped to 24h to demonstrate the overall time curve and used for NCA. The shaded areas represent the area between the EC₅₀ and EC₉₀ for each compound derived from published data.³¹

Table 2.

Non-compartmental analysis parameter results.

	Elexacaftor		Tezacaftor		Ivacaftor
	Transplant	Non-transplant	Transplant	Non-transplant	Transplant	Non-transplant
Cmin (ng/mL)	5632(2664)	4327(2359)	2016(904)	1746(1050)	467 (203)	525 (325)
Cmax (ng/mL)	11373(4336)	9083(4813)	6523(2124)	7460(2296)	1227 (505)	1382 (746)
Cavg (ng/mL)	8339(3152)	6398(3233)	3403(1240)	3294(1314)	833 (342)	947 (507)
AUCτ	200(76)	154(78)	82(30)	79(32)	9.9 (4.2)	11.4 (6.1)
Tmax (h)	7.33(2.96)	4.79(2.22)	3.08(3.15)	2.21(1.42)	4.6 (1.5)	4 (1.7)
T1/2 (h)	23.6(11.9)	35.0(26.6)	21.0(16.5)	19.0(13.9)	7.8 (3.3)	9.4 (3.7)
CLss/F (L/h)	1.23(0.79)	2.05(2.29)	1.54(1.07)	1.53(0.83)	19.9 (15.5)	27.9 (47.4)
V/F	37.4(15.0)	107(155)	47.5(59)	37.1(24.4)	168 (71)	337 (441)

All values are mean (standard deviation); min minimum value; max maximum value; CV coefficient of variation; T_1/2 half life; T_max time to maximum plasma concentration; C_max maximum plasma concentration; C_min minimum plasma concentration; AUC_T area under the plasma concentration time curve from time of administration to 12 or 24 hours after dosing; CL/F oral clearance; V/F apparent distribution volume. All comparisons shown in this table were made and are non-significant.

Figure 2. Comparison of NCA parameters.

All NCA parameters were compared, including C_min, C_max, C_avg, T_1/2, T_max, Clearance, Volume of Distribution, and AUC₂₄ (ELX, TEZ) or AUC₁₂ (IVA). All parameters are shown in detail in Table 2, but for simplicity and to demonstrate the variability within the data, the minimum and maximum concentrations as well as the average concentration are shown here for all three parent compounds and the four major metabolites. The dotted lines represent the EC₅₀ and EC₉₀ for each compound derived from published data,³¹ although these thresholds for metabolites are unknown. NT = Non-transplant, T = transplant.

Comparison of C_max and C_min to manufacturer’s reported data.

Details of the C_max and C_min for all three parent compounds were available from the manufacturer.³² Welsch’s t-test for mean differences and TOST for equivalency between the transplant and manufacturer group and the non-transplant and manufacturer group are calculated. For the transplant group, C_min of ELX and TEZ, as well as the C_max for TEZ and IVA, were not significantly different, and determined to be equivalent, to the manufacturer. For the non-transplant group, the C_min for ELX was not significantly different, and determined to be equivalent, to the manufacturer. C_min for IVA for both groups were determined to be significantly different (lower) and not equivalent to the manufacturer’s values (428ng/mL compared to reported 750ng/mL) for transplant group and (525ng/mL compared to reported 750ng/mL) for non-transplant group. However, all other comparisons (C_min for TEZ and the C_max for all compounds in the non-transplant group, and C_max of ELX for the transplant group) were unable to be determined as they were both non-significantly different, but also not equivalent, using these statistical tests, which were likely due to larger than expected variability within the tested groups. While some of the differences suggest meaningful differences, these data do not provide information on the potential clinical relevance. Therefore, we undertook additional comparisons to better contextualize these data.

Proportions of concentrations outside of estimated effective concentrations.

Precise values for effective concentrations are not publicly available, however, can be estimated from prior publications.³¹ We estimated EC₉₀ (and EC₅₀) for compounds: elexacaftor 4500 (500) ng/mL; tezacaftor 3500 (400) ng/mL; and ivacaftor 350 (40) ng/mL when used concurrently from published data.³¹ We evaluated the proportions of values for each parent compound C_avg that were above the EC₉₀ or below EC₉₀ and EC₅₀. No values were below the EC₅₀ for any compound in either the transplant or non-transplant group. While there were some values below the EC₉₀, as shown in Fig. 1 and 2, there were no differences in the proportions below these thresholds in the transplant versus non-transplant groups. Overall, all participants appear to have sufficient concentrations for therapeutic effect. The majority of participants had concentrations above the EC₉₀, but the potential threshold for “excessive” concentrations has not been established at this time and cannot be estimated.

Sample size re-estimation.

Our study utilized the manufacturer’s data to estimate initial sample size needed for this study. After comparing our data to the manufacturers, we found that our current sample size was sufficient to determine equivalency and significant differences for some but not all tested values. For the C_min of TEZ, the C_max of ELX, TEZ, and IVA for the non-transplant group, and the C_max of ELX for the transplant group, we re-calculated the sample size required to determine equivalency for TOST. We determined that a sample size for these undetermined comparisons were between 21–35 per group assuming p=0.05, power=0.8, and the lower and upper equivalency bounds defined by the standard deviation reported by the manufacturer and taking into account the variance of the non-transplant and transplant groups data. These data can provide some insight into future study design utilizing single values such as the C_max or C_min, which reflects the increased variability of real-world data.

Discussion.

In this study we identified the PK parameters for the non-transplant and transplant group for ETI compounds and the major human metabolites. These data provide preliminary evidence that there are not significant differences in ETI concentrations to explain differences in side effects in lung transplant recipients that lead to high discontinuation rates,²¹ and no substantial differences in concentrations outside the therapeutic ranges that could explain a lack of benefit. We also performed sensitivity testing using linear mixed modeling, evaluating both the overall differences and rate of change of each compound between transplant and non-transplant participants, which also revealed no significant differences between groups. Furthermore, there were no significant or meaningful differences in metabolite concentration or MPR, which speaks to one mechanism by which transplant recipients could have a different experience of ETI. If the parent compounds are similar, but metabolite concentrations differ, that finding could suggest a potential etiology for greater side effects or lower tolerance of ETI, but these data suggest transplant recipients do not have meaningfully different metabolism of ETI. Although there has been substantial interest in the use of ETI after solid-organ transplant,^{16, 20–27, 29, 35, 36} there have been no detailed reports of ETI PK parameters or concentrations over time in this population with concomitant tacrolimus use. Our study provides ETI concentration data to add to what is already known about tacrolimus concentrations in transplant recipients taking both medications.^{16, 23, 25, 35}

Multiple extra-pulmonary complications improve with the use of ETI, including sinus disease in non-transplant population^{3, 37} and improved sinonasal treatment of Pseudomonas aeruginosa³⁸, which could contribute to slowed or halted development of CLAD in lung transplant recipients.^{39, 40} Improved CFTR function has been associated with improved median post-transplant survival⁴¹ with a 39% lower risk of death or re-transplant in those with residual function mutations (adjusted HR 0.61, 95% CI = 0.40, 0.91) compared to those with a more severe genotype. HEMT improves CFTR function to residual function levels, or better. Lastly, the use of ETI increases body mass index (BMI) in the non-transplant population.^42–44 Recovery to normal BMI has been shown by our group to be associated with improved post-transplant survival compared to those who remain underweight, among lung transplant recipients with a low pre-operative BMI.⁴⁵ Collectively, the management of extra-pulmonary complications is key to better outcomes in post-lung transplant recipients, leading to the hypothesis that ETI may improve post-transplant outcomes. However, the uncertainty of potential drug interactions remains a concern for using ETI after lung transplant. This study was initiated due to concerns in our prior study revealing discrepant practices across lung transplant centers in the use of ETI, which showed that as many as 35% of patients at some centers and as few as 0% at others were prescribed ETI, despite known extrapulmonary benefits of ETI use.²¹ In addition, a much higher percentage of lung transplant recipients (42%) stopped ETI due to side effects (~32%) or no benefit (10%), compared to the pivotal clinical trials in non-transplant patients.^{21, 46} Twenty percent of the post-transplant patients who stopped ETI went on to restart it during the study period. Some lung transplant recipients reported significant improvement in symptoms after initiating ETI, including some of those in this study. We hypothesized that there may be a significant difference in ETI concentrations that could contribute to side effects or limit benefit given other prior reports showing poor tolerability and limited benefit.^{22, 26, 29, 35, 36} In this study, we did not find significant differences in concentrations of the parent compounds in ETI or their major human metabolites that could explain these differences in clinical response between the transplant and non-transplant groups.

Concern for drug interactions is high, especially for narrow therapeutic index drugs such as tacrolimus. Tacrolimus is a substrate for the cytochrome P450 isozymes (CYP3A superfamily) and for drug transporter P-glycoprotein (P-gp), both of which are involved with the metabolism or transport of one or more compounds in ETI.⁴⁷ Combined with three different compounds in ETI, each with different potential to inhibit or induce these enzymes or P-gp, the effects on circulating concentrations of either ETI components or tacrolimus are challenging to predict. Supra-therapeutic increases in tacrolimus concentrations result in toxicity (e.g., acute kidney injury), while sub-therapeutic concentrations could result in allograft rejection. Sub-therapeutic concentrations of ETI would result in treatment failure, whereas the impact of supra-therapeutic ETI concentrations are not known but may result in toxicity as well.⁴⁸ On the other hand, mild changes in tacrolimus have been reported with concurrent use of ETI, but are manageable with appropriate therapeutic drug monitoring.^{16, 17, 20} Our data do not support differences in ETI concentration in either direction (too high or too low) that may substantially impact therapy. Similar to other studies^{16, 35}, tacrolimus concentrations were manageable among our study participants.

This study can also serve as a pilot to generate intensive data that may be used to inform larger studies. The variability of the concentrations of all components of ETI was higher than expected based on manufacturers’ data. This is typical for real-world pharmacokinetic studies, especially those performed in populations that have comorbidities or characteristics that were excluded from the original PK studies. We calculated power for our study and determined that we are powered to detect differences and equivalency in certain parameters for certain compounds but are not powered to detect differences and equivalency in all compounds and all parameters, which reflects the increased variability in real-world studies and reinforces the need for appropriately powered studies to evaluate the association of drug concentrations with response. We also evaluated concentrations outside of the estimated therapeutic range, which did not reveal any significant differences compared to the non-transplant group, suggesting that any differences are not likely to be clinically impactful. We specifically excluded participants taking additional drugs that are known to impact ETI concentrations, such as azole antifungals. Additional studies evaluating concomitant use of azoles may be warranted to evaluate ETI concentrations in that population. Our study was also limited by feasibility. We were not able to obtain additional blood sampling between 12 and 24 hours for ELX and TEZ, which may have limited our dataset. However, despite this, our NCA parameters were similar to those reported by the manufacturer, which limits this concern. Furthermore, our effective concentrations are derived from modeled data published by the manufacturer and are not explicitly published anywhere. The clinical value of these estimates are uncertain. The estimated values derived for our study provide context, but may not be applicable in all populations or for all dose adjustments.

This pharmacokinetic study of ETI in lung transplant and non-transplant recipients did not identify clinically meaningful differences in ETI concentrations in these two groups, and it highlights the variability of concentrations of these compounds and provides data to inform study design including sample size to ensure robust results may be obtained in future, larger studies.

Highlights.

• ETI is sometimes prescribed in transplant recipients concurrently taking tacrolimus

• Interaction of tacrolimus (immunosuppressant) with ETI presents a theoretical risk

• ETI PK data were not significantly different in non-transplant vs transplant patients

• Real-world drug concentrations for ETI were more variable than in clinical trials

Abbreviations:

ETI

elexacaftor-tezacaftor-ivacaftor

PK

pharmacokinetics

NCA

noncompartmental analysis

ELX

elexacaftor

TEZ

tezacaftor

IVA

ivacaftor

pwCF

people with CF

LTx

lung transplant

HEMT

highly effective modulator therapy

CF

cystic fibrosis