This prospective study investigates the impact of proton pump inhibitors (PPI) on mycophenolic acid (MPA) pharmacokinetics in heart transplant recipients receiving mycophenolate mofetil (MMF) and tacrolimus. MPA plasma concentrations at baseline (C0 h), 30 min (C0.5 h), 1(C1 h) and 2 h (C2 h) were obtained by high-performance liquid chromatography (HPLC) in 22 patients treated with pantoprazole 40 mg and MMF 2000 mg. Measurements were repeated 1 month after pantoprazole withdrawal. A four-point limited-sampling strategy was applied to calculate the MPA area under the curve (MPA-AUC). Predose MPA concentrations with PPI were 2.6 ± 1.6 mg/L versus 3.4 ± 2.7 mg/L without PPI (p = ns). Postdose MPA concentrations were lower with PPI at C0.5 h (8.3 ± 5.7 mg/L vs. 18.3 ± 11.3 mg/L, p = 0.001) and C1 h (10.0 ± 5.6 mg/L vs. 15.8 ± 8.4 mg/L, p = 0.004), without significant differences at C2 h (8.3 ± 6.5 mg/L vs. 7.6 ± 3.9 mg/L). The MPA-AUC was significantly lower with PPI medication (51.2 ± 26.6 mg × h/L vs. 68.7 ± 30.3 mg × h/L; p = 0.003). The maximum concentration of MPA (MPA-Cmax) was lower (12.2 ± 7.5 mg/L vs. 20.6 ± 9.3 mg/L; p = 0.001) and the time to reach MPA-Cmax (tmax) was longer with PPI (60.0 ± 27.8 min vs. 46.4 ± 22.2 min; p = 0.05). This is the first study to document an important drug interaction between a widely used immunosuppressive agent and a class of drugs frequently used in transplant patients. This interaction results in a decreased MMF drug exposure which may lead to patients having a higher risk for acute rejection and transplant vasculopathy.
Mycophenolate mofetil (MMF) has gained widespread acceptance as the predominant antiproliferative immunosuppressant of choice and has proven to be effective in preventing allograft rejection after heart transplantation (1,2). MMF is used in almost 76% of patients during the first year after transplantation (3). Following oral administration, MMF is rapidly metabolized to its active constituent mycophenolic acid (MPA), which acts as a potent and specific inhibitor of T- and B-cell proliferation by reversibly inhibiting inosine monophosphate dehydrogenase (IMPDH), the key enzyme of the de novo purine synthesis in activated lymphocytes (4). Therapeutic drug monitoring for MMF is well established in our center. We developed a variety of limited sampling strategies for the estimation of the MPA area under the curve (MPA-AUC) and abbreviated AUC measurements are routinely used (5). On reviewing our data, we found a correlation between PPI therapy and decreased MPA exposure. In this respect, previous investigations have shown a significant relationship between the MPA-AUC and an increased risk of acute rejection (6). Gastrointestinal side effects are common in patients after solid-organ transplantation, and a considerable proportion of transplant recipients receive proton pump inhibitors (PPI). The Mitos Study Group including 1788 heart transplant recipients reported that almost 40% of all patients suffered from gastrointestinal complications of which 86.3% were treated with gastrointestinal-protective comedication (7). Recently, a negative influence of proton pump inhibitor comedication on the MPA-AUC in renal transplant recipients has been published by Miura et al. (8). MPA plasma concentrations were significantly decreased by 30 mg lansoprazole but not by 10 mg rabeprazole especially in recipients having the CYP2C19 or the MDR1 C3435T polymorphisms.
The impact of PPI on the MPA-AUC after heart transplantation has not been examined so far. We decided to measure abbreviated MPA-AUCs in all maintenance patients who presented with PPI medication in our outpatient department and gave their consent for data acquisition and evaluation. Our prospective, case-controlled study was performed in order to investigate the impact of PPI therapy on MPA-plasma concentrations in a standardized setting in heart transplant patients.
Patients and Methods
From April to December 2008, a total of 22 heart transplant recipients (3 females, 19 males) with initial PPI comedication were followed. The immunosuppressive regimen of all patients consisted of MMF and tacrolimus. The oral dose of tacrolimus was adjusted to reach target trough levels of 5–14 ng/mL according to the point of time after the heart transplantation. The oral dose of MMF was 1000 mg twice daily in all patients. Table 1 shows the patients demographic data with and without PPI therapy.
Table 1. Clinical characteristics of 22 heart transplant recipients with and without PPI medication
+Pantoprazole (PPI medication)
Values are reported as mean ± SD or percentage (%).
No. of patients
45.1 ± 11.8
45.3 ± 11.8
Body weight (kg)
74.7 ± 15.3
75.1 ± 15.6
24.4 ± 4.5
24.4 ± 4.6
Time after transplantation (years)
2.5 ± 3.6
2.7 ± 3.5
Indication for transplantation (%)
Dilated cardiomyopathy (DCM)
Ischemic cardiomyopathy (ICM)
Congenital heart disease
Serum creatinine (mg/dL)
1.3 ± 0.5
1.3 ± 0.5
56.7 ± 30.7
48.4 ± 15.8
Glomerular filtration rate (GFR)
89.7 ± 47.0
83.6 ± 38.6
High sensitivity CRP (ref. <0.5 mg/dL)
0.3 ± 0.2
0.4 ± 0.5
6.4 ± 1.9
6.8 ± 2.3
Total bilirubin (mg/dL)
0.8 ± 0.3
0.8 ± 0.3
Current medication (%)
Simvastatin 12 mg/day
Furosimide 8 mg/day
Diltiazem 55 mg/day
Prednisolone 0.1 mg/kg/day
Patients presented to the routine follow-up visit after an overnight fast and without taking the morning dose of their medication. For each patient, we studied MPA-plasma concentrations with blood samples collected at the following time points: predose (C0 h), at 30 min (C0.5 h), 1 h (C1 h) and 2 h (C2 h). After the first measurements the PPI treatment was stopped, and the patients were analyzed again 1 month after pantoprazole withdrawal. MMF was continued with 1000 mg twice daily.
The predose MPA-serum concentration and routine laboratory parameters were obtained before the patients received their morning dose of 1000 mg MMF. Plasma was separated and stored at −20°C for all samples until MPA concentrations were measured by a validated high-performance liquid chromatography (HPLC) method employing solid-phase extraction and UV detection (Chromsystems, Munich, Germany; http://www.chromsystems.de). The method was fully automated by the use of the Gilson GX271 liquid handling system (Gilson, Middleton, WI, USA). Quality assurance was performed according to the guidelines of the German Medical Association, including internal and external quality control. The interassay coefficient of variation for quality control samples with an MPA concentration of 2.0 mg/L and 5.0 mg/L, respectively, was <4% (n = 12) (9).
For computer-assisted statistical data analysis, the SPSS software package was used (SPSS 16.0 for Windows; SPSS Inc., Chicago, IL). Values of continuous variables are expressed as mean ± standard deviation. Differences were considered as statistically significant at a p < 0.05. Statistical analysis was performed with the Wilcoxon signed-rank test comparing the measurements for MPA concentrations at the different time points, for the MPA-AUC, tmax and Cmax. An algorithm for a four-point limited-sampling strategy was applied to calculate the MPA-AUC (5).
The demographic data of the 22 patients are given in Table 1. Patients with PPI were examined at 2.5 ± 3.6 years and patients after PPI withdrawal at 2.7 ± 3.5 years after heart transplantation. Indications for transplantation were dilated cardiomyopathy (DCM), ischemic cardiomyopathy (ICM) or others (e.g. congenital heart disease or cardiac sarcoidosis). DCM was found predominantly in 59.1%, ICM in 22.7%, cardial sarcoidosis in 4.5% and congenital heart disease in 13.7%. Current medication included statins (a mean dose of simvastatin 12 mg/day in 61% of the patients with PPI vs. 71% without PPI), diuretics (a mean dose of furosemide 8 mg/day in 17% of patients at both time points) and calcium channel blockers (a mean dose of diltiazem 55 mg/day in 44% of patients at both time points). Furthermore, half of the patients were off steroids since they were routinely withdrawn 6 months after heart transplantation at our center. The patients included in the early phase between 2 and 6 months after heart transplantation were on a maintenance dose of low-dose steroids (0.1 mg/kg prednisolone). All 11 patients with steroid therapy had the same prednisolone dose at both time points (see Table 1). Therefore, there were no significant differences regarding the medication at both time points.
Laboratory tests revealed normal values for creatinine, urea, glomerular filtration rate (GFR), hsCRP, cholinesterase and bilirubin.
In Table 2 pharmacokinetic parameters are displayed. Under pantoprazole 40 mg/day, the daily mean tacrolimus dose was 3.8 ± 1.6 mg (ranging from 1.5 to 6.5 mg/day) with mean predose concentrations of 9.8 ± 2.7 ng/mL. Without PPI medication, the same patient group had a daily mean tacrolimus dose of 3.7 ± 1.9 mg (ranging from 1.0 to 6.5 mg/day) with mean predose concentrations of 8.9 ± 2.5 ng/mL. There was no statistical significance. Under a daily MMF dose of 2000 mg, mean predose serum MPA concentrations (C0 h) with PPI were 2.6 ± 1.6 mg/L versus 3.4 ± 2.7 mg/L without PPI (p = ns). Thirty minutes after the morning doses of MMF (C0.5 h), MPA plasma concentrations were significantly lower with PPI treatment than after PPI withdrawal (8.3 ± 5.7 mg/L vs. 18.3 ± 11.3 mg/L, p = 0.001). One hour after dosing (C1 h), MPA plasma concentrations were again significantly lower in the PPI group (10.0 ± 5.6 mg/L vs. 15.8 ± 8.4 mg/L, p = 0.004). Two hours thereafter, MPA plasma concentrations were persistently lower in the PPI group but without significant differences (8.3 ± 6.5 mg/L vs. 7.6 ± 3.9 mg/L; p = ns). Figure 1 summarizes the mean MPA plasma concentrations at the different time points comparing the patients with PPI treatment and after PPI withdrawal.
Table 2. Pharmacokinetic parameters of MPA with and without PPI medication in heart transplant recipients
+Pantoprazole (PPI medication)
Values are reported as mean ± SD or percentage (%). Each parameter was statistically compared with the control group.
No. of patients
MMF dose (mg/day) (mean)
Tacrolimus dose (mg/day) (mean)
3.8 ± 1.6
3.7 ± 1.9, p = ns
Tacrolimus plasma concentration (ng/mL)
9.8 ± 2.7
8.9 ± 2.5, p = ns
MPA plasma concentration (mg/L)
2.6 ± 1.6
3.4 ± 2.7, p = ns
MPA (0.5 h)
8.3 ± 5.7
18.3 ± 11.3, p = 0.001
MPA (1 h)
10.0 ± 5.6
15.8 ± 8.4, p = 0.004
MPA (2 h)
8.3 ± 6.5
7.6 ± 3.9, p = ns
Maximal concentration (Cmax, mg/L)
12.2 ± 7.5
20.6 ± 9.3, p = 0.001
Time to reach Cmax[tmax (minutes)]
60.0 ± 27.8
46.4 ± 22.2, p = 0.05
Total AUC (mg × h/L)
51.2 ± 26.6
68.7 ± 30.3, p = 0.003
The calculated mean MPA-AUC (four-point limited-sampling strategy) was significantly lower in patients with PPI medication (51.2 ± 26.6 mg h/L vs. 68.7 ± 30.3 mg h/L; p = 0.003, Figure 2). The maximum plasma concentration of MPA (MPA-Cmax) was significantly lower with PPI (12.2 ± 7.5 mg/L vs. 20.6 ± 9.3 mg/L; p = 0.001), and tmax was reached significantly later with PPI than without PPI medication (60.0 ± 27.8 min vs. 46.4 ± 22.2 min; p = 0.05).
All patients included in this study are alive. However, acute rejection episodes (ARE) and transplant vasculopathy (TVP) occurred in seven patients during their PPI treatment. We defined ARE as any treated rejection with ISHLT grade ≥ 1R (≥1B according to the old ISHLT classification). In the PPI group, ARE occurred in seven patients with 1R rejection. TVP was defined as coronary artery disease with >30% lumen narrowing. In the PPI group, seven patients developed TVP. The observation period after the PPI withdrawal was only 1 month. Therefore, to fully examine this phenomenon a more extensive study including a larger number of patients and a longer follow-up is necessary.
Drug interactions between PPI and tacrolimus have been reported. However, little is known about the interaction between PPI and MMF (8,10,11). In a former study, we retrospectively observed that patients with pantoprazole therapy had lower MPA levels than patients without PPI and reinvestigated potential drug interactions (5). Based on this observation, we started a prospective study to investigate the impact of PPI on MPA pharmacokinetics in heart transplant recipients receiving MMF and tacrolimus in a standardized setting.
PPI that inhibit gastric acid secretion through interaction with H+/K+-ATPase in gastric parietal cells can modify the intragastric release of other drugs by elevating pH value. Thereby they can influence drug absorption and metabolism by interacting with adenosine triphosphate-dependent P-glycoprotein or with the cytochrome P450 (CYP) enzyme system (12). The P450 CYP enzyme system seems to be important for the interactions with MMF (10). Although we did not measure the pantoprazole AUC, previous studies have shown that pantoprazole 40 mg has a stable AUC0–24 h of 9.93 μmol h/L, which correlates with the degree of acid suppression. Moreover, the bioavailability of 77% after the first dose does not change after repeated dosing (13,14).
Following oral administration, MMF is rapidly absorbed and undergoes extensive presystemic de-esterification by esterases in the plasma, liver and kidney. MPA, the active immunsuppressant metabolite, is inactivated in the liver via first pass glucuronidation. A significant part of the resulting phenolic glucuronide of MPA (MPAG) is not pharmacologically active. In addition, it undergoes extensive enterohepatic recirculation and its clearance is highly dependent on protein binding. Over 90% of the administered dose is excreted in the urine, mostly as MPAG (15–17).
The present study shows that the usual therapeutic dose of pantoprazole (40 mg) had a significant influence on the maximal MPA plasma concentration (1.9-fold higher after PPI withdrawal), and the total MPA-AUC could be increased by 34% after PPI withdrawal. MPA plasma concentrations 0.5 and 1 h after dosing were significantly reduced by comedication with pantoprazole 40 mg, but not the MPA plasma concentration 2 h after dosing. The gastric acid secretion inhibitory effect of PPI might decrease the elution and hydrolysis of MMF, subsequently diminishing the plasma concentration of MPA. Tomiyama et al. could show in rats that 2 h after administration of PPI, K+-ATPase activity is inhibited by about 40% and the acid secretory rate by about 94% (18). Therefore, within 2 h after dosing, pantoprazole almost completely inhibited the gastric acid secretion and subsequently the absorption of the MPA with reduced plasma concentrations in the first hour. Furthermore, the peak of the MPA time–concentration curve was reached 0.5 or 1 h after intake in the majority of patients throughout the literature (19,20). This peak represents the main contributor to the total MPA-AUC. IMPDH activity and, therefore, the immunosuppressive effect of MMF reveal a very good correlation with the total AUC as reported (21). When the main contributors to the total AUC (C0,5 h and C1 h) are reduced due to PPI, this should result in a significant decrease in the immunosuppressive effect. C2 h has less impact on total AUC and therefore on IMPDH-activity. An additional interesting question is whether the mechanism of altered MMF kinetics under PPI treatment can be modulated through the use of H2 blockers. In three patients treated with H2 blockers (ranitidine 300 mg/day), we found the same decreased MPA plasma concentration at C0.5 h, C1 h and MPA Cmax. However, we have to confirm this preliminary result with a larger number of patients. Until now there are no other data regarding this drug–drug interaction.
Time to reach Cmax (tmax) was significantly longer in patients with PPI than without PPI medication (60.0 ± 27.8 min vs. 46.4 ± 22.2 min; p = 0.05). Naesens et al. (22) demonstrated that a delayed gastric emptying rate in renal patients was associated with a significantly longer MPA tmax and a significant decrease in MPA Cmax. Therefore, one might assume that our data with reduced MPA Cmax and longer tmax under PPI are possibly also induced by a delayed gastric emptying rate. A counterargument to this assumption would be that the C2 h values do not reveal a significant difference after PPI withdrawal. Further investigations are necessary to prove that.
MMF kinetics are influenced by many factors such as renal function, liver function and comedications such as steroids (20,23,24). Renal and liver functions showed normal values and did not differ between the two measurement points. Half of the patients were on a maintenance dose of low-dose steroids (0.1 mg/kg prednisolone) and had the same prednisolone dose at both time points. Comparing both groups (with and without steroids) there is no difference in MPA plasma concentrations and MPA-Cmax. Therefore, we presume that steroids did not influence our results. Cattaneo et al. (24) showed that high-dose steroids in kidney transplantation induce the hepatic glucuronyltransferase (GT) expression enhancing the activity of uridine diphosphate-GT, the enzyme responsible for MPA metabolism. Therefore, high-dose steroids are another factor responsible for decreasing the MPA-AUC. Cattaneo et al. performed AUC measurements the first month with high-dose steroids and 6 months after transplantation with low-dose steroids. However, our patients did not receive high-dose steroids. In addition, the effects of low-dose steroids on MMF kinetics did not differ between each patient. We therefore feel that the effects of low-dose steroids in this study are marginal and not an explanation for the differences between the groups.
Miura et al. showed that lansoprazole and rabeprazole influence MPA pharmacokinetics due to multidrug resistance (MDR) 1 C3435T polymorphisms after renal transplantation in Japanese patients (10). He hypothesized that lansoprazole inhibits MPA absorption as a result of inducing a low-acid condition, similar to the effect of antacid coadministration (8). Hunfeld additionally demonstrated that healthy Caucasians with CYP2C19 mutation need higher concentrations of PPI e.g. pantoprazole for sufficient acid suppression (25).
Of central clinical interest in transplanted patients are the rates of ARE and the development of TVP. The study by Galiwango et al. showed that MMF dose reduction, e.g. for gastrointestinal intolerance, was associated with a significantly increased rate of sustained rejection in heart transplant patients (26). Kaczmarek et al. and Meiser et al. evaluated that lower therapeutic drug concentrations of MMF correlate not only with increased rates of ARE but also with the development of TVP in heart transplant recipients (27,28). However, in this study ARE and TVP occurred in seven patients during their PPI treatment with lower MPA plasma concentrations. The observation period after the PPI withdrawal was only 1 month; therefore, to fully examine this phenomenon a more extensive study including a larger number of patients and a longer follow-up is necessary. However, as previously mentioned, in our not yet published retrospective study with 21 patients a clear trend for more ARE and TVP was found in the PPI group. The follow-up in both groups (control group 34.5 ± 42 months and PPI group 21 ± 32 months) was not significantly different. Therefore, it is of central interest to further examine the mechanisms of PPI-induced lower MPA levels and to design a strategy for using PPI and MMF to preserve MPA plasma concentrations to increase safety in terms of freedom from ARE and TVP.
Not only increased rates of ARE and TVP but also cost-effectiveness are important issues in posttransplant therapy. One year of pantoprazole therapy (40 mg/day) costs 470 Euro when prescribed in Germany. Regarding the necessity of lower MMF doses for the patients after PPI withdrawal, average MMF costs of more than 2000 Euro per year can be saved in patients who are no longer on PPI. Increased laboratory costs for therapeutic drug monitoring are thereby ‘reimbursed’.
Several clinical trials have documented an increase in the MPA-AUC under fixed-dose regimens during the first month after transplantation, and this is widely accepted (29,30). Therefore, some transplant centers react to this phenomenon by increasing the initial MMF dose. Considering that we achieved an average increase in the MPA-AUC of 34% after the withdrawal of the PPI comedication, the widely accepted concept of increasing the MPA-AUC over time might has to be looked at from a different angle. PPI might have contributed to this phenomenon in previous investigations because they were withdrawn in the previous study populations over time resulting in increasing MPA-AUCs (29,30).
We conclude that the usual therapeutic dose of pantoprazole (40 mg) had a significant inhibitory influence on MPA-Cmax, the total AUC and tmax. MPA plasma concentrations 0.5 and 1 h after dosing were significantly reduced by comedication with pantoprazole. This is the first study to document an important drug interaction between a widely used immunosuppressive agent and a class of drugs frequently used in transplant patients. This interaction results in a decreased MMF drug exposure which may lead to patients having a higher risk for acute rejection and transplant vasculopathy. In addition, transplant physicians have to take into account that patients with PPI comedication need higher MMF dosages to reach equal immunosuppressive effects.