Treatment by Posaconazole Tablets, Compared to Posaconazole Suspension, Does Not Reduce Variability of Posaconazole Trough Concentrations

The delayed-release tablet formulation of posaconazole (POS-tab) results in higher plasma POS trough concentrations (C_min) than the oral suspension (POS-susp), which raises the question of the utility of therapeutic drug monitoring (TDM). We aimed to compare the variability of the POS C_min for the two formulations and identify determinants of the POS-tab C_min and its variability.

ABSTRACT

The delayed-release tablet formulation of posaconazole (POS-tab) results in higher plasma POS trough concentrations (C_min) than the oral suspension (POS-susp), which raises the question of the utility of therapeutic drug monitoring (TDM). We aimed to compare the variability of the POS C_min for the two formulations and identify determinants of the POS-tab C_min and its variability. Demographic, biological, and clinical data from 77 allogeneic hematopoietic stem cell transplant patients (874 C_min) treated with POS-tab (n = 41), POS-susp (n = 29), or both (n = 7) from January 2015 to December 2016 were collected retrospectively. Interpatient and within-subject coefficients of variation (CVs) of the C_min adjusted to dose (D) were calculated for each formulation. Between-group comparisons were performed using a linear mixed effects model. The POS C_min was higher for the tablet than for the suspension (median [25th–75th percentile]: 1.8 [1.2–2.4] mg/liter versus 1.2 [0.7–1.6] mg/liter, P < 0.0001). Interpatient CVs for the tablet and suspension were 60.8 versus 63.5% (P = 0.7), whereas within-subject CVs were 39.7 and 44.9%, respectively (P = 0.3). Univariate analysis showed that age and treatment by POS-tab were significantly and positively associated with the POS C_min, whereas diarrhea was associated with a diminished POS C_min. Multivariate analysis identified treatment with POS-tab and diarrhea as independent factors of the POS C_min, with a trend toward a lower impact of diarrhea during treatment with POS-tab (P = 0.07). Despite increased POS exposure with the tablet formulation, the variability of the POS C_min was not significantly lower than that of the suspension. This suggests that TDM may still be useful to optimize tablet POS therapy.

INTRODUCTION

Posaconazole (POS) is an antifungal agent widely used for prophylaxis of invasive fungal infections in patients with hematological malignancies at high risk, especially allogeneic hematopoietic stem-cell transplant (AHSCT) patients with graft-versus-host disease. Since 2015, POS has been commercially available in Europe as a new formulation, i.e., delayed-release tablets (POS-tab), allowing a single 300-mg dose per day (1). This new formulation improves POS bioavailability relative to the former oral suspension formulation (POS-susp) (2–4). Hence, the proportion of patients achieving the recommended target goal of >0.7 mg/liter (5) while taking the tablet has increased over that of patients taking the oral suspension formulation (3, 6), which raises the question of the utility of POS therapeutic drug monitoring (TDM).

Increased POS exposure of patients using the tablet formulation over that of patients using the oral suspension is now well established (2–4). However, although the variability of the POS-susp trough concentrations (C_min) is known to be significant, with interindividual and within-subject coefficients of variation (CV) as high as 64% and 49%, respectively (7), little is known about the variability of the POS-tab C_min. A recent crossover study showed lower within-subject CV for the POS-tab C_min in lung-transplant patients treated with POS-susp and then with POS-tab (8). However, the number of patients was quite low (n = 24), and interindividual CVs were not studied. Moreover, determinants involved in such variability are yet to be clarified. Indeed, the impact of diarrhea (7), mucositis, and proton pump inhibitors (PPIs) on the POS C_min has been well described for POS-susp (9). However, the influence of these factors on the POS C_min during POS-tab therapy is not clear, as recent results are conflicting (3, 4, 10–14). For example, several studies reported a significant impact of diarrhea on the POS-tab C_min (11, 12), whereas others did not (4, 13–15).

The primary aim of this study was to compare the POS C_min, as well as its intra- and interindividual variability between the two galenic formulations (tablet versus suspension) in AHSCT patients. We further aimed to identify factors that influence the POS C_min and its variability.

RESULTS

Baseline characteristics of patients.

A total of 874 POS C_min, determined in 79 patients, were included in this study; 41 patients received POS-tab only, 31 received POS-susp only, and 7 switched from the suspension to the tablet formulation (see Fig. 1). The baseline characteristics of patients are shown in Table 1.

FIG 1.

Flow chart. For each group of patients (tablet, suspension, and switched patients), the number of posaconazole trough concentrations (POS C_min) for which potential factors that influence the POS C_min were known are specified at the bottom.

TABLE 1.

Baseline characteristics of patients^a

Characteristic	All patients (n = 77)
Demographics
Age (yrs)	53.0 (22.0–64.7)
Male (%)	41 (53.2)
Weight (kg)	64.0 (48.9–84.3)
Underlying hematological diseases
Acute myeloid leukemia	35
Acute lymphoblastic leukemia	16
Myelodysplastic syndrome	11
Lymphoma	8
Chronic myeloid leukemia	1
Chronic lymphoblastic leukemia	1
Othersb	5
Longitudinal follow-up of POS treatment
Duration of follow-up (days)	84 (41–376)
No. of dose adjustments/patient	1 (0–2)
No. of POS Cmin/patient	9 (4–18)

^aData are indicated as numbers (%) or median (25th–75th percentiles).

^bOthers include one idiopathic medullary aplasia, two nondifferentiated acute leukemia, and two acute biclonal leukemia.

Posaconazole trough concentrations and their determinants.

POS C_min were significantly higher during treatment with POS-tab than with POS-susp (Table 2; Fig. S1 in the supplemental material), resulting in a higher proportion of therapeutic POS C_min with POS-tab without reducing numbers of POS dose adjustment (Table 2). The results of univariate and multivariate analyses performed for all 874 POS-C_min are shown in Table 3. Treatment with POS-tab was associated with a higher POS C_min, whereas the presence of diarrhea was significantly associated with a lower POS C_min (Table 3). Age was positively associated with the POS C_min in the univariate analysis, but did not reach statistical significance in the multivariate analysis. Conversely, mucositis, cotreatment with a PPI, and body mass index (BMI) were not associated with the POS C_min. In addition, there was a trend toward an interaction (P = 0.07) between the presence of diarrhea and galenic formulation, suggesting a lower POS C_min in the presence of diarrhea with the suspension versus the tablet formulation (Fig. S2).

TABLE 2.

Description of plasma POS C_min according to galenic formulation

Characteristic	POS-susp Cmin	POS-tab Cmin	P value
No. of POS Cmin	373	501
POS Cmin (mg/l)
Mediana (25th–75th percentiles)	1.2 (0.7–1.6)	1.8 (1.2–2.4)	<0.0001
Minimum POS-Cmin	<0.1	<0.1
Maximum POS-Cmin	4.7	8.2
No. of therapeutic POS Cmin (%)	294 (79)	462 (92)	<0.0001
Median POS dosea (mg/day)	600 (600–800)	300 (200–300)	<0.0001
No. of POS dose adjustments (%)	39 (10.4)	48 (9.6)	0.7
Coefficient of variation of POS Cmin/dose
Interindividual CV (%)	63.5	60.8	0.7
Within-subject CV (%)	44.9	39.7	0.3
No. (%) of POS Cminb associated with:
Mucositis
Yes	18 (5.5)	73 (17.7)
No	307 (94.5)	338 (82.3)
Diarrhea
Yes	42 (12.9)	51 (12.7)
No	283 (87.1)	349 (87.3)
Proton pump inhibitor
Yes	141 (78.3)	352 (92.4)
No	39 (21.7)	29 (7.6)

^aData are indicated as median (25th–75th percentiles).

^bAs indicated in Fig. 1, some data are missing. Thus, the percentage of POS-C_min with these potential determinants could vary.

TABLE 3.

Mixed-effects model of potential determinants of POS C_min

Variable	Nonmissing data (%)	Univariate analysis		Multivariate analysis
		Estimate ± SE	P value	Estimate ± SE	P value
POS-tablet	100	0.92 ± 0.11	<0.001	1.13 ± 0.18	<0.001
Age	100	0.33 ± 0.07	<0.001	−0.19 ± 0.11	0.07
Body mass index	76	0.11 ± 0.03	0.72
Proton pump inhibitor	76	0.14 ± 0.19	0.47
Mucositis	82	0.11 ± 0.08	0.17
Diarrhea	80	−0.37 ± 0.08	<0.001	−0.30 ± 0.07	<0.001

Variability of posaconazole plasma trough concentrations and its determinants.

There was a large range of POS C_min among patients (Fig. 2 and 3), which translated into high intra- and interindividual variability of POS C_min/dose (D), regardless of the galenic formulation (Table 2). CVs and within-subject CVs of POS-tab C_min/D were slightly but not significantly lower than those of POS-susp (Table 2). Intra- and interindividual variability was not significantly associated with age, BMI, mucositis, diarrhea, or treatment with a PPI (Tables S1 and S2).

FIG 2.

Variability of plasma posaconazole (POS) trough concentration (C_min) in 36 patients treated with POS suspension. Each vertical series of crosses corresponds to repetitive POS trough concentrations in one patient, with the black diamond indicating the median POS C_min per patient. The dotted line indicates the prophylactic threshold of POS C_min (5).

FIG 3.

Variability of plasma posaconazole (POS) trough concentration (C_min) in 48 patients treated with POS tablet. Each vertical series of crosses corresponds to repetitive POS trough concentrations in one patient, with the black diamond indicating the median POS C_min per patient. The dotted line indicates the prophylactic threshold of POS C_min (5).

Onset of invasive aspergillosis.

Invasive aspergillosis (IA) occurred in three patients treated with POS-tab (3/48) and four treated with POS-susp (4/36). The characteristics of these patients and their POS C_min are shown in the supplemental material (Table S3). The mean POS C_min per patient before IA diagnosis was 2.3 mg/liter for patients treated with POS-tab and 0.65 mg/liter for those treated with POS-susp.

DISCUSSION

Retrospective analysis of 874 POS C_min measured in 77 AHSCT patients demonstrates that the POS tablet formulation results in higher POS exposure than the POS oral suspension, without reducing either inter- or intraindividual variability of the POS C_min.

Enhanced exposure with posaconazole tablets.

POS C_min determined during longitudinal TDM of AHSCT patients were significantly higher during POS-tab treatment than those obtained with POS-susp, in accordance with the results of a phase III study of POS-tab (4) and several other retrospective studies (2, 3, 6, 11, 13, 16). Logically, this increased POS exposure resulted in a higher percentage of therapeutic POS C_min with POS-tab but did not reduce the number of POS dose adjustments. The magnitude of the increase in POS C_min that we observed (mean increase of 0.6 mg/liter, which represents a 1.5-fold increase) for POS-tab is close to that previously reported by Pham et al. (1.4-fold increase) (6) and Jung et al. (2.5-fold increase) (2).

Determinants of posaconazole trough concentrations.

Among the potential determinants of POS C_min, we explored diarrhea, mucositis, age, BMI, and concomitant treatment with a PPI. We found a significant impact of diarrhea, which decreased the POS C_min in both univariate and multivariate analyses. This finding is in accordance with the previously described negative impact of diarrhea on POS-susp C_min (7, 9) but also with several retrospective studies conducted in hematological patients treated with POS-tab (11, 12, 15). However, the impact of diarrhea appeared to be lower during POS-tab therapy than POS-susp treatment (Figure S2), even if the interaction did not reach statistical significance. The lower magnitude of the effect of diarrhea with the tablet might explain why other studies did not find an influence of diarrhea on POS-tab C_min (13, 16). However, other factors could be involved, such as heterogeneous study populations, i.e., lung-transplant recipients and hematological patients (16) or hematological patients without a distinction between allografted or nonallografted patients (13).

In addition, we found a positive association between age and POS C_min in univariate analyses, which is in accordance with previous studies performed in hematological patients treated with POS-susp (7, 17) or POS-tab (18). However, this association did not reach significance in multivariate analyses, suggesting that the influence of age on POS C_min was dependent on another variable in our cohort. Conversely, we found no effect of BMI, mucositis, or PPI cotreatment on the POS C_min, whereas several studies did (3, 11, 19). However, these studies were not all comparable in terms of sample size and patient inclusion criteria, which could explain, at least in part, these discrepancies.

Variability of posaconazole trough concentrations.

The originality of our study consists of the comparison of intra- and interindividual variability of POS C_min/D for both galenic formulations. As dose adjustments occurred during longitudinal TDM (Table 1 and 2), we adjusted the POS C_min for the POS dose to overcome this factor of variability. Although the interindividual CV and within-subject CV of the POS C_min/D were lower for the POS-tab than for the POS-susp (Table 2 and Fig. 2 and 3), the difference was small and neither clinically relevant nor statistically significantly different, suggesting that the variability of POS-tab C_min is still as large as that of POS-susp C_min (20). Aside from one study, which reported lower POS C_min variability with the tablet regimen in 24 lung-transplant patients (8), such a comparison of the intra- and interpatient variability of POS C_min between the two oral galenic formulations of POS has never been performed. Several studies reported CVs for the POS-tab C_min, but comparing CV values is problematic since calculation methods of CV differ. Here, we used the method described by Bland (21) to calculate within-subject CV, whereas a different approach was used in other studies, i.e., the ratio between standard error and the mean POS C_min in each patient (18, 22). Finally, as explained above, we adjusted the POS C_min for the POS dose in our cohort, which was not done in the other studies (18, 22). Moreover, differences in study populations may also account for the differences observed between the studies. For example, the within-subject CV of the POS-tab C_min was 39.7% in AHSCT patients in our study, whereas it was estimated to be 16% in 24 lung transplant recipients and 48% in 15 patients with hematological malignancies (18). Despite the significant variability of POS C_min, we did not identify any variable associated with such variability in our cohort, suggesting that factors other than those studied here, such as genetic variants of UGT1A4 (23) or reduced hepatic function, still need to be investigated.

Efficacy of posaconazole.

Finally, the central question is probably the efficacy of POS-tab relative to that of POS-susp. In our study, we could not perform any statistical analysis to compare the number of treatment failures between groups, because episodes of IA were very rare during POS therapy, regardless of the galenic formulation used. In addition, the retrospective design of our study and heterogeneity in follow-up enhance the risk of bias. To date, only one study has demonstrated a lower number of IA episodes during treatment with POS-tab than POS-susp (15), whereas others have reported similar rates of invasive fungal infections between the two galenic formulations (3, 13, 24, 25). In our cohort, we reported treatment failure in 6.3% of patients receiving POS-tab (3/48), although all these patients had a therapeutic POS C_min. Such findings have already been described for POS-susp (13, 18), which suggests that some therapeutic failures may not be related to insufficient POS exposure.

Limitations.

Our study had several limitations. First, this study was retrospective and performed in a single center. Thus, some data were missing, especially for potential factors influencing the POS-tab C_min (allograft-versus-host disease, PPI, weight, and gradation of diarrhea and mucositis). In the same vein, POS safety was not evaluated, while high POS C_min could be associated with occurrence of adverse events as recently described in several case reports (14, 18, 26, 27). Despite these limitations, our study also had several strengths, since it is notably the first study to compare the variability of the POS C_min between POS-susp and POS-tab. In addition, our study was conducted in a homogeneous and large population of recipients of allograft stem-cell transplantation (77 patients and 874 POS C_min), which is not true of several previous studies (3, 11, 12, 16).

Conclusion.

Although the tablet formulation increased POS exposure, variability of the POS C_min was still high and not significantly lower than that observed with the oral suspension. Therefore, these results suggest that TDM may still be useful for AHSCT patients treated with POS-tab.

MATERIALS AND METHODS

Study design.

This retrospective study was conducted at the Grenoble University Hospital, France. Adult (>18 years old) AHSCT patients with graft-versus-host disease, treated with POS according to the European Conference on Infections in Leukemia 6 guidelines (5), with longitudinal TDM (defined by at least three POS C_min determinations per patient and per galenic form) from January 2014 to December 2016 were eligible (see Fig. 1 for details). Demographic (age, sex) and clinical data (diarrhea, mucositis, and weight), as well as records concerning POS therapy (daily dose, galenic form, and POS C_min) and concomitant PPI treatment were collected retrospectively. Concerning the factors that could potentially affect the POS C_min, we considered only those that are concomitant to the determination of the POS C_min. For clinical events (diarrhea, mucositis), we considered only those that occurred within 7 days before the POS C_min determination. Similarly, only PPI therapy initiated at least 7 days before measuring the POS C_min was considered. Due to the retrospective design of our study, some data are missing, especially for potential determinants of the POS-tab C_min (see Fig. 1 for details). Concerning comedications, only PPIs were considered to be a risk factor in our model, but the absence of any other drug that could interact with POS was carefully checked. All patients gave written consent for collection and use of their data.

Posaconazole therapeutic drug monitoring.

All patients received POS for prophylaxis of invasive fungal infection. Patients treated with POS-susp were asked to take their POS with an acidic beverage to improve POS absorption. All patients treated with POS received regular POS TDM. Blood samples were drawn just before drug intake or at least 20 h after intake for the POS-tab. After treatment initiation or dose adjustment, a 7-day period was considered necessary to obtain a pharmacokinetic steady state. POS C_min were excluded in case of inappropriate blood collection time (sample handled before pharmacokinetic steady state or less than 20 h after POS-tab intake) (see Fig. 1 for details). Plasma POS C_min were measured by validated liquid chromatography tandem-mass spectrometry, as previously described (28). The plasma drug standard curve ranged from 0.1 to 20 mg/liter with adequate between-day and within-day variabilities (CV < 15%) and accuracy. A therapeutic threshold of 0.7 mg/liter defined for prophylaxis was considered (5).

Evaluation of efficacy.

The efficacy of POS was assessed by the absence of invasive fungal infections (especially infections due to Aspergillus spp.) during the follow-up period for each patient. The follow-up period was defined as the time elapsed between the initiation of POS therapy and the last determination of the POS C_min. All cases were prospectively reviewed monthly by a multidisciplinary team including chest physicians, a microbiologist, a hygienist, and a pharmacologist (29). Cases were classified according to guidelines from the European Organization for Research and Treatment of Cancer/European Invasive Infections Cooperative Group and the criteria of the National Institute of Allergy and Infectious Diseases-Mycoses Study Group (EORTC/MSG) (30).

Statistical analysis.

Categorical variables are expressed as frequencies (and percentages) and quantitative variables, as medians (with the 25th to 75th percentiles). The analysis of POS exposure, depending on galenic formulation or potential determinants, was performed for the POS C_min, whereas the study of variability was performed on the POS C_min adjusted for dose (C_min/D) to overcome the influence of dosage adjustments occurring during longitudinal follow-up. CVs and within-subject CVs were calculated as previously described (21) for each POS formulation to assess inter- and intraindividual variability, respectively. They were subsequently compared using a method adapted from Levene’s test. As some patients received both formulations, we used mixed effects models, in which the formulation was a fixed factor and the patient a random factor. We performed a log transformation of the C_min/D ratio because the data were not normally distributed. A P value of 0.05 was considered statistically significant. Statistical analyses were performed with SPSS Statistics version 21 (IBM, Armonk, NY, USA).