Cancer Chemotherapy and Pharmacology

, Volume 74, Issue 2, pp 349–357 | Cite as

Third-space fluid distribution of pemetrexed in non-small cell lung cancer patients

  • Per Hartvig Honoré
  • Sigrid Jóhansdóttir Joensen
  • Michelle Olsen
  • Steen Honoré Hansen
  • Anders Mellemgaard
Original Article



Hydrophilic drugs particularly those with low plasma protein binding may accumulate in third-space fluid in the body. Cytotoxic drugs like methotrexate (MTX) cause damage in the tissue, and evacuation of the third-space fluid in pleura is strongly recommended before new dosing. Pemetrexed (PEM) is a multi-targeted antifolate similar to MTX approved for the treatment for malignant pleural mesothelioma and non-small cell lung cancer. Current recommendations for patients receiving treatment with PEM prescribe draining of the pleural fluid. This is based upon the recommendations for MTX and not directly to any specific findings relating to PEM. The recommendations are the same because PEM is an analogue of MTX; the molecular structures and pharmacokinetic parameters are similar. However, since draining the pleural fluid is painful and cancer patient are particularly susceptible to infection subsequently, it is relevant to examine the recommendations for PEM explicitly.


Eight patients treated with a 500 mg/m2 PEM combined with platinum salt were examined. Plasma samples were first collected in relation to the start of PEM infusion. Thereafter, plasma and pleura samples were taken at various times after drug infusion from each patient; in two patients, sampling was done twice but on different occasions. The quantitative determination of PEM was performed with reversed-phase high-performance liquid chromatography, and sample preparation was performed using protein precipitation with perchloric acid. Pharmacokinetic analysis was performed using a non-compartment method as well a two-compartment model.


The results were calculated from 10 samples taken from eight patients, where data from one patient point were excluded as the patient had impaired renal function, and three samples were reported as below limit of quantification. The plasma PEM pharmacokinetics calculated showed an elimination half-life (t½ elimination) of 3.2 h and distribution half-life (t½-distribution) of 6 min. Clearance (CL) was 5.1 L/h, central volume of distribution (Vcentral) 23.2 L and peripheral volume distribution (Vperipheral) 10.6 L, and the area under the curve was 186 μg h/mL. Using non-compartment methods, an elimination half-life of 3.1 h and an apparent CL of 3.2 L/h were measured, whereas an apparent steady-state volume became 14.2 L. The pleura concentrations were only half of simultaneous plasma concentrations, and elimination half-life was 3.15 h.


Pemetrexed is not likely to accumulate in the pleural fluid, and evacuation of fluid might not be necessary. Further investigation is needed to recommend no drainage of the fluid, i.e., in patients with renal impairment.


Pemetrexed Pharmacokinetics Pleura Plasma Non-small cell lung cancer 


Third-space fluid is the assembly of watery reservoirs in the tissue. Drugs may penetrate into these reservoirs to different extent depending on their physicochemical properties and cause harm. The cytotoxic drug methotrexate (MTX) has a high uptake and slow effusion in the pleura sac in patients with lung cancer and may cause damage to the lung tissue. The dilution of MTX into a larger than normal volume in patients with serosal effusions has been shown to result in prolonged plasma half-lives and was associated with subsequent myelosuppression [1]. The increased risk of toxicity following high-dose MTX in patients with pleural effusions was shown to be due to changes in MTX kinetics resulting in delayed excretion [2].

The lungs and thorax cavity are covered with two thin membranes, e.g., the visceral, covering the lungs, and the parietal pleura that covers the thorax cavity. During the various phases of respiration, the surfaces of parietal and visceral pleura float over each other facilitating proper lung movements [3]. The visceral and parietal pleura are separated by the pleural space in which the pleural fluid is contained [4]. The normal volume of fluid is low, usually <10 mL in each pleural space (0.1–0.2 mL/kg body weight) [5]. Pleural fluid is under normal conditions primarily secreted by the mesothelial cells in the parietal layer of the pleura originating from the systemic circulation [6, 7]. Approximately 80–90 % of the pleural fluid is reabsorbed mainly through the lymphatic drainage on the parietal pleural side [8].

Pleural effusion is a result of accumulation of fluid caused by imbalance in the equilibrium of absorption and reabsorption of fluid [6]. This can be attributed to a number of mechanisms, such as capillary permeability, hydrostatic and colloidal osmotic pressure, interstitial osmotic pressures, and lymphatic drainage, depending on the pathological characteristic of the individual patient [9].

Approximately 15 % of all lung cancer patients are initially diagnosed with pleural effusion, 50 % later being diagnosed with pleural effusion [10]. Pleural diffusion in lung cancer patients can be grouped into the following: malignant pleural effusion, a pleural effusion indirectly caused by the cancer or its treatment, or pleural effusion unrelated to the cancer [2]. Although most of these effusions are determined to be malignant, about one-half are initially classified as negative cytology [5]. Malign pleural fluid can be exudate or transudate, although most commonly exudate [11]. The difference between exudate and transudate can be assessed according to Light´s criteria, where the pleural protein/serum protein ratio is higher than 0.5, the pleural LDH is higher than two-thirds part of the serum LDH upper limit, and pleura LDH/serum LDH is higher than 0.6. An exudate is defined if one of these criteria is met [12]. Protein content is usually an adequate indicator in clinical practice, with a protein content of more and <30 g/L for exudate and transudate, respectively [13]. It has been suggested that the accumulation of malignant pleural effusion is due to a combination of two mechanisms [14, 15]. The first one is explained by a decreased drainage of pleural fluid as a result of disruption or obstruction by tumor cells in the lymphatic system. This mechanism has been confirmed by necroscopic studies [7]. The second mechanism is explained by the presence of pleural metastases that may increase the permeability of the capillaries in the visceral and/or parietal pleura, which can lead to increased amounts of fluid [14]. Tumor characteristics that determine the degree and type of blockage cannot be expected to be uniform among patients.

Different malignancy in the pleural space will therefore affect the pharmacokinetics of each drug differently for each patient [6].

Movement of drugs in and out of pleural fluid may be presumed to follow the same route as pleural fluid [6]. Most drugs move out of plasma into pleura by passive diffusion [6]. In order to be eliminated from plasma, drugs must move back into the systemic circulation by diffusion through the capillary pores, penetration of capillary membrane, and/or by bulk flow or diffusion through the lymphatic system [6]. Pleural fluid is thought to enter the pleural space mainly via the parietal pleura; meanwhile, drugs may enter from both the visceral and parietal systemic capillaries [6].

The ability of a drug to reach the pleura is influenced by several factors such as plasma drug concentration, plasma and tissue protein binding, pleural membrane penetration, blood perfusion of the pleura, the effect of pleural disease on vascular permeability, and by the physicochemical properties of the drug (molecular size and weight, lipophilicity, and ionization) [14]. Only the unbound form of protein-bound drugs is thought to enter the pleura under normal conditions because of molecular size and lipophilicity [14].

Current recommendations for patients receiving treatment with MTX prescribe draining the pleural fluid. However, since draining the pleural fluid is painful and cancer patients are particularly susceptible to infection subsequently, it is highly relevant to examine for third-space distribution of drugs explicitly.

Pemetrexed (PEM) is a multi-targeted antifolate approved for the treatment for malignant pleural mesothelioma and non-small cell lung cancer (NSCLC). Its mechanisms of action are similar to MTX and can be used on similar indications in the clinic. Current recommendations for patients receiving treatment with PEM prescribe draining the pleural fluid as for MTX. This is based upon the recommendations for MTX and not directly to any specific findings relating to PEM. The recommendations are the same because PEM is an analogue of MTX; the molecular structures and pharmacokinetic parameters are similar. A pharmacokinetic study of PEM in plasma and pleura was undertaken in lung cancer patient to explicitly assess the risk for cytotoxic harm from third-space distribution of the drug.

Materials and methods

Patient material

Data used in this analysis were pooled from an open, non-randomized population pharmacokinetic study of patients with non-small cell lung cancer, NSCLC, and with pleural effusion. Eight patients were included on distribution of PEM to third-space fluid in the lungs following a short infusion. Two of the patient participated twice after two doses of PEM separated in time. Patients were included in the study if they met the following criteria: histologically or cytologically confirmed NSCLC regardless of size and stage, with pleural effusion as a result of the disease and to be treated with 500 mg/m2 of PEM alone or in combination with other cytotoxic drugs, regardless of previous treatment and being male or female ≥18 years old that was able to give informed consent. Patients were excluded from the study if they did not met the following criteria contraindications to thoracentesis for example bleeding disorders or severe respiratory dysfunction corresponding to a walking distance of <50 m.

The study was conducted in collaboration with the Department of Oncology at Herlev Hospital, in accordance with the Guidelines of Good Clinical Practice and Declaration of Helsinki. Approval was obtained from The Danish Medicine Agency, The Danish National Committee on Biomedical Research Ethics, and the Danish Data Protection Agency. Potential participants were identified at the Department of Oncology at Herlev Hospital. If they met the inclusion criteria, they were invited to participate in the study. To select patients for the study, it required that a CT showed a pleural effusion and that the presence of effusion could be confirmed by clinical examination at the time of puncture.

All patients were informed both verbally and in writing about the purpose of the study, and after reflection, patients who participated had given both verbal and written consent before enrollment. Patients’ characteristic data collected were gender, age, height, weight, BSA, exact dose of PEM and platinum salt, courses of PEM, eGFR (estimated glomerular filtration), serum creatinine, performance status as well as tumor classification, TNM status (Table 1).
Table 1

Patient characteristics

Patient no.


Height (cm)

Weight (kg)

BSA (m2)


PEM dose (mg)

Courses of PEM

Platinum salt + dose (mg)

TNM status

PS status









Cisplatin 140











Cisplatin 140











Carboplatin 600











Carboplatin 600











Carboplatin 570











Carboplatin 500











Carboplatin 600











Cisplatin 130











Cisplatin 140











Carpoplatin 450














Body surface area (BSA), the dose, the courses of PEM, which is the number of times the patient has received treatment: 0 refers the first time and 1 refers to the second time. Platinum salt drug PEM combined with PEM. TNM states fx T4N2M1B T is the primary tumor where T4 is the description of the tumor, size, and extent, N refers node involvement where N2 is where it is located, M refers to metastasis where M1B refers to the distant metastases

Patients were administered 500 mg/m2 during a 10-min intravenous infusion. A plasma sample was collected after the end of infusion. At predetermined times (1–48 h), after drug infusion, a second plasma sample was drawn. Pleura fluid samples were drawn at the same time as the second plasma sample using sterile technique and a 210 cannula. Plasma and pleura samples were stored at −20 °C.


The concentrations of PEM in plasma and pleura were determined using a validated reversed-phase high-performance liquid chromatography (HPLC) method described by Rinaldi et al. [16] and modified. The chromatographic system used for the assay compromised an Agilent technologies 1,120 compact liquid chromatograph from Agilent technologies (Agilent Technologies Denmark ApS, Glostrup, Denmark) coupled to an autosampler, pump, and an UV detector set at 250 nm from the same manufacturer. The chromatographic data handling was accomplished using the EZChrom elite software. Samples of 10 µL were injected into an Accelaim ®120: C8; 3 µm; 120 Å; 4,6 × 150 mm column (Thermo Scientific Inc., USA) and operated at 40 °C. The flow rate was set to 0.8 mL/min. The mobile phase A consisted of 5 % acetonitrile and 0.1 % formic acid, and mobile phase B of 99.9 % acetonitrile and 0.1 % formic acid. The separation was performed using gradient elution with the time program, 0 min 90 % A and 10 % B, 2 min 90 % A and 10 % B, 10 min 50 % A and 50 % B, 12 min 20 % A and 80 % B, 13 min 20 % A and 80 % B, and 14 min 90 % A and 10 % B. Total run time was 18 min and with 4 min between the samples.

Patients’ plasma samples and drug plasma samples obtained were stored at −28 °C and were thawed at room temperature. Plasma proteins were precipitated with 7 % perchloric acid, where 400 μL of plasma and 400 μL (1:1) 7 % perchloric acid were mixed thoroughly before centrifugation. Afterwards, the denatured proteins were removed by centrifugation and left a clear supernatant that was injected into the HPLC. Working standards were prepared in the following concentrations: 200, 100, 50, 25, 5, 1, and 0.4 μg/mL in water or drug-free human plasma.

Validation of the bioanalytical method

Validation of the bioanalytical method was based on the FDA guideline “Guidance for Industry—Bioanalytical Method Validation” [17]. Precision, accuracy, recovery, yield, and lower limit of quantification, LLOQ, were validated. The precision was measured using a minimum of three QC samples at low, middle, and high concentrations representing the entire range of the standard curve. The precision determined at each concentration should not exceed 15 % of the coefficient of variation (CV) except for the LLOQ, where it should not exceed 20 % of the CV.

Accuracy was determined at a minimum of three QC samples at low, middle, and high concentration with minimum five determinations at each concentration. The mean value was expressed in percentage (%) and should stay within ±15 % of the theoretical/true value, except at LLOQ, where it should not deviate by more than ±20 %.

Recovery was assayed by comparing results from extracted samples for a minimum of three concentrations (low, medium, and high). The obtained peak areas of the spiked plasma standards at these concentrations after sample preparation were compared with the peak areas resulting from aqueous standard solutions at the same concentrations without sample preparation to represent 100 % recovery.

The lowest standard on the calibration curve was accepted as LLOQ as the analyte response was at least five times the response compared to blank response and was reproducible with an imprecision of maximum 20 % and an accuracy of 80–120 %.

The standard deviation of the response was determined based on the standard deviation of y-intercepts of regression lines [17]. The standard error of the intercept was obtained from the regression analysis by multiplying the standard error with the square root of the number of patient samples [18].

Pharmacokinetic calculations

The PK calculations from concentrations measured in plasma were done as both a non-compartment and two-compartment models. For the two-compartment model, the method of residuals was used [19]. The elimination rate constant was determined by log-linear regression analysis of at least the last three terminal points on the plasma concentration–time curve and was used to back-extrapolate to zero plasma concentration. The central volume of distribution (Vcentral) was calculated by dividing CL with the slope (λ2) of the terminal phase. The peripheral volume of distribution (Vperipheral) was calculated by dividing average dosage with C1 + C2. The PK parameters area under the curve (AUC) and area under the first moment curve (AUMC) in the non-compartment model were calculated by the trapezoidal method in the non-compartmental method [19, 20]. The elimination rate (kelimination) obtained from calculations in the two-compartment model was used in the calculation of AUC0 and AUMC0. The mean residence time (MRT) was obtained by dividing AUMC0 by AUC0 in the non-compartment model, and the mean elimination rate (k) was obtained by dividing one with MRT. The steady-state volume (Vss) was obtained multiplying MRT with CL. Clearance (CL) was in both alternatives calculated by dividing AUC with dose.

Calculation of PK parameters for pleura was done by using absorption kinetics for uptake of PEM into pleura. The method of residuals was used to calculate the elimination (kelimination) and the absorption rates (kabsorption). The elimination rate constant was determined by log-linear regression analysis of at least the last three terminal points on the plasma concentration–time curve. The points were then used to back-extrapolate the uptake data points.

The PEMpleural/PEMplasma ratio was determined for each patient by dividing the concentrations of PEM in pleura and plasma at each data point. In addition, the overall PEMpleural/PEMplasma ratio was determined by dividing AUCpleural with AUCplasma. Mean and ± SD is given for the parameters.


Patients characteristics

Eight patients were enrolled in this study, where a total of twenty plasma samples and ten pleura samples were collected between March 2012 and June 2013. Two patients participated twice; patient samples 1.1 and 1.2 were taken at two different occasions from the same patient, and patient samples 1.20 and 2.20 were taken from another patient at two different occasions. Out of the eight patients, there was only one woman. The average age was 69.3 (±9.2) years. All patients were diagnosed with adenocarcinoma NSCLC and received PEM combined with cisplatin or carboplatin. All patients were in their first or second cycle of PEM treatment. The performance status (PS) was 0–1. Patients’ characteristics and demographics are listed in Table 1. The serum creatinine concentration was ranging from 0.58 to 0.98 mg/dL. Normal range is, respectively, 0.5–1.1 mg/dL for women and 0.6–1.2 mg/dL for men [20]. For two of the male patients (patient sample 1.1–1.2 and 2.4), the serum creatinine concentration was slightly lower. The creatinine CL in the patients ranged from 76.0 to 216 mL/min. Patient 1.48 can according to FDA guidelines [21] be classified having slightly impaired renal function and therefore not used in the analysis, the others were normal. Not any clinical adverse effects due to pleural effusion were seen in any of the patients.

Plasma and pleura concentrations of PEM

The plasma and pleura concentrations of PEM from the eight patients and two samples from the two patients studied at two occasions after intravenous PEM infusion over 10 min are given in Fig. 1 with exact sampling times. In the analysis, all samples coming from the patient population were treated as coming from one patient. The initial plasma concentration of PEM (POI) varied from 70.5 to 191 μg/mL with a mean 109.2 μg/mL. There were a large spread between the initial concentrations but sample 1.48 from a patient with a low creatinine CL and GFR had the highest one. The plasma concentration fell rapidly, and no PEM was detected 20 h after administration.
Fig. 1

Plasma, squares, and pleura, circles, concentrations in patients after an intravenous infusion of PEM, 500 mg/m2

The PEM concentrations were higher in plasma than in pleura in the time period 0–8 h. From 20 to 48 h, it seems likely that pleura has a higher concentration of PEM. The PEM pleura concentrations thus indicate a delay for entrance into the third-space pleura fluid. For three samples, the analysis indicated concentrations below LLOQ appearing at late sampling times.

Pharmacokinetics of PEM in plasma and pleura

The pharmacokinetic profile of PEM in plasma displays a biphasic profile. There were only seven data points in the plasma concentration–time curve as three concentrations were below limit of quantification. One patient’s plasma sample was excluded, as it graphically seemed to be an outlier compared to the rest of the data points. The pharmacokinetics were calculated both using non-compartmental and compartmental analyses. The pharmacokinetic parameters are summarized in Table 2. CLs of parenteral PEM were 3.2 and 5.1 L/h in the non-compartment and two-compartment models, respectively. AUC was higher for the non-compartment model probably because all data points are included in the measurements. The elimination half-live was 3.1 and 3.2 h in the non-compartment and the two-compartment models, respectively.
Table 2

PK parameters for plasma from a two-compartment and a non-compartment model

PK parameters

Two-compartment model

Non-compartment model

C1 (mg/L)

49.9 mg/L

C2 (mg/L)

38.5 mg/L

kmean (h−1)

0.22 h−1

λ1 (h−1)

7.4 h−1

λ2 (h−1)

0.217 h−1

t½-elimination (h)

3.2 h

3.1 h

t½-distribution (hr)

5.6 min

CL (L/h)

5.1 L/h

3.2 L/h

Vcentral (L)

23.2 L

Vperipheral (L)

10.6 L

Vss (L)

14.2 L

AUC (mg h/mL)

185.9 mg h/mL

297.0 mg h/L

AUMC (mg h2/L)

1,335.6 mg h2/mL

k12 (h−1)

0.5 h−1

k21 (h−1)

3.4 h−1

k10 (h−1)

3.8 h−1

MRT (h)

4.7 h

C1 and C2 concentrations at zero time; λ1 and λ2 rates of the distribution and elimination, respectively; kmean mean residence time; t½ elimination and t½ distribution halftime for elimination and distribution; Cl, clearance; Vcentral, Vperipheral, and Vss, volume of central and peripheral compartments and volume at steady state; AUC area under the curve; AUCM area under the first moment curve; k12 rate from the central to the peripheral compartment; k21 rate from the peripheral to the central compartment, k10 elimination rate from the central compartment; MRT mean residence time

The pharmacokinetics of PEM in pleura was calculated using non-compartment analysis. Patient sample 1.48 was excluded in the PK analysis in plasma, and the pleural sample for this patient was therefor also excluded in the PK analysis in pleura. The derived pharmacokinetic parameters are displayed in Table 3. The elimination was expected to be one-phasic and is assumed to be separated from the central compartment. The uptake kinetic model based on the assumption that elimination and uptake only occur from and into the central compartment [19] was used. The half-life of the uptake, taken as the time for the residual to diminish by one-half, is 13.9 min. The uptake half-life reflects the time it takes for PEM to be transferred from plasma to reach the pleural fluid. The elimination half-life for PEM from pleura was 3.15 h. The AUC was calculated in both plasma and plerura that were calculated using the trapezoidal method. PEM in pleura is lower than the AUC for PEM in plasma, the ratio being 0.48.
Table 3

Pharmacokinetics of PEM in pleura (non-compartment) showing halftime of uptake and elimination and area under the curve (AUC)

t½-elimination (h)

3.15 h

t½-uptake (h)

13.9 min

AUC (µg h/mL)

158.2 mg h/L

Validation of the bioanalytical method

Validation of the bioanalytical method was performed to demonstrate that the described procedure was suitable to use for the quantification of PEM in human plasma and pleural effusion. The results from the validation are summarized in Table 4. Calibration/standard curves were constructed both in water and drug-free plasma with four to six samples covering the expected concentration rang, including LLOQ. Both calibration curves showed linearity over the range of 0.4–200 μg/mL. The requirements in the FDA guideline [17] for precision were met, as all the intra-run precisions were <15 %. This means that a good repeatability and consistency can be attributed. The accuracy was based on spiked water samples as the true measure. All the results are negative, suggesting a higher concentration measured in water than in plasma, and the criteria for acceptance were not met at all concentrations (Table 4). The mean recovery value obtained for PEM in plasma was 75.5 %. According to FDA [17], only the extraction effect was studied, which tells the extent of impact sample preparation has on PEM. Complete recovery is not demanded, but a consistency in results is required. The LLOQ was 2.9 µg/mL.
Table 4

Summary of the results of the validation of the analysis of PEM

C (μg/mL)

Intra-run precision (water)

Intra-run precision (plasma)

Accuracy (%)

Recovery (%)

½0.4 μg/mL





1 μg/mL





25 μg/mL





100 μg/mL





200 μg/mL











The pharmacokinetic properties of PEM with a rapid elimination and also a comparably high protein-bound fraction favorably prevent its accumulation into third-space accumulation in pleura [16, 22, 23, 24].

The pleura concentrations were generally lower than those of plasma, and the estimated elimination rates were of the same magnitude. Protein-bound drug would have a limited distribution over the pleura membrane, and the rapid elimination further prevented any of the limited amount taken up to reside in the pleura. The pleura concentrations were below detection after less than five half-lives and long before a second dose was administered. The pharmacokinetic profiles were similar in these patients with lung tumors compared to others without lung tumors indicating that third-space fluid accumulation had a very limited influence on the disposition of PEM [25, 26].

The pharmacokinetic parameters for PEM in plasma were similar to previously reported and validated in a non-compartment model [16, 22, 23]. PEM showed a relative small volume of distribution suggesting a limited tissue distribution. Approximately 75–81 % of PEM is bound to plasma proteins [27] and is primarily eliminated unchanged by renal excretion via glomerular filtration and active secretion by the renal tubule, and 70–90 % of the PEM dose is excreted unchanged within 24 h. Clearance has been reported to 90 mL/min [28, 29], which is higher than found in the present study. The plasma concentrations could also be fitted a two-compartment model, which changed some of the parameters defined differently in the model as, e.g., volume of distribution. The plasma sampling of PEM was sparse up to the first hour after administration, which gives a poor description of the distribution phase of PEM.

This study pooled plasma and pleura from eight patients out of which two participated twice as several samplings of pleura not being feasible from the same patient. Therefore, interindividual variability adds to the inprecision in the analysis of PEM in plasma and pleura and the variations might also to a large extent be attributable to patient characteristics. In the case of PEM, these characteristics would especially be kidney function, body surface area (BSA), disease and performance status, albumin, and age, although most of these seemed in the same range for this patient sample. The content of pleural fluid and the amount of proteins in the pleural fluid are widely varying among cancer patients [6]. The uptake of PEM in pleural fluid is dependent on the membrane penetration. Exudative effusion in the pleural cavity will affect the membrane, by altering the structure in the membrane [6]. Obstructive tumors may block lymphatic drainage of the pleural space, and the drug can therefore be trapped inside pleura. The characteristics of the tumor, which determines the degree and type of blockage, cannot be expected to be uniform among patients. Different tumor status for patients in the pleural space will therefor affect the pharmacokinetics differently for each patient. The penetration characteristics of the drug, the tumor status for patient, and also the general condition of the patient will all contribute to a variance that can affect the pharmacokinetics of the drug in pleural space [6].

Variation in kidney function is an important factor for PEM elimination. Higher exposure of cytotoxic drugs also contributes to variability in the kidney function of cancer patients [30, 31]. Sample 2.48 from a patient with a lower creatinine clearance CrCL was compared to the others, which was judged as an intermediate kidney status. This patient had an obvious slower CL of PEM from pleura and plasma and was excluded from the pharmacokinetic analyses. The concentration of albumin is also a contributing factor to variability, and the protein binding in blood as well as in pleura may influence third fluid distribution.

Pemetrexed and methotrexate (MTX) have similar structures and physicochemical characteristics. Their pharmacokinetic profiles are also similar, and both display a multi-phasic plasma concentration–time curve [33]. Their primary route for elimination is renal excretion where the drugs are eliminated mostly unchanged. Studies with PEM and MTX have shown that impaired kidney function resulted in lowered CL [29]. There are differences between these two drugs concerning toxic effect. MTX toxicity correlates with the time of plasma concentration above a threshold of value 0.1 μM while PEM toxicity may correlate with total systemic exposure. This can be expressed through their mechanism of action where high doses of polyglutamated MTX are needed to hold down DRFH to be effective [34].

Since prolongation of MTX elimination will increase the risk of severe toxicity, it is important to establish those clinical variables related to malignant effusions that alter MTX disposition. Pleural effusion has been the problem for patients when administrating high doses of MTX, and patients with pleural effusion may have an increased risk for developing toxicity to high-dose MTX as a result of accumulation of MTX in the pleural cavity or MTX trapping in the infusion causing a slow release, which leads to sustained MTX concentrations in plasma.

Plasma concentrations of MTX with and without pleural effusion have been simulated [32]. The effect of changing the fluid characteristic was tested with MTX by a physiologic-based pharmacokinetic method This was done by evaluating the muscle tissue binding, which resulted in, when increasing the muscles tissue binding for MTX, volume of distribution at steady-state Vss becoming lower and the rate of transport in and out of the pleural space decreasing. The study showed that drugs with a low protein binding and a low Vss are more likely to accumulate in pleural fluid. PEM was compared to MTX using the same model, and an assumption was made on how protein binding and tissue binding affected the Vss and elimination half-life for PEM. The volume of distribution seemingly was halved by the presence of pleural fluid for PEM. This means that the presence of pleural fluid had less effect on PEM than MTX [32]. This can be related to that PEM apparently being less hydrophilic than MTX, and the volume of distribution is therefore expected to be larger for PEM compared to MTX. Further, the protein-bound drugs are also less likely to penetrate over membranes into (as well as out) from the pleura effusion. As PEM has a higher protein binding, and hence larger volume of distribution compared to MTX, it is not expected to transfer to the pleural space to the same extent as MTX.

Hence, it is important to have an estimate of the volume of pleura to predict how pleural fluid affects PEM distribution. It was not possible to get the volume of pleural fluid, as it cannot be measured with a good reliability. A lower unbound fraction of drug in the blood (fu) would result in a higher volume of distribution and a higher elimination rate, but CL remains unchanged [33].

There are certain limitations in the present study, which flaw the results and also may influence the conclusion. The analytical method was not able to quantify the low pleura and plasma concentrations of PEM 20 h after drug administration. This together with the absence of sampling in the period 12–20 h after administration makes that few data points were available for calculation making pharmacokinetic analysis more uncertain. A full pharmacokinetic profile from the patients with matched pleural fluid samples would have been a better trial design. A pooled analysis of population data has certain substantial limitations. The key period of interest is the late elimination phase where accumulated drug is released back into the systemic circulation. The data would have been better analyzed simultaneously with data from the individual plasma pharmacokinetics. The parameters derived from the patients would have been better estimated and allowing a better estimate of the pleura/plasma ratio.

The bioanalytical method was validated to meet the acceptance criteria of the FDA [17]. The acceptance criteria for intra-precision were met but not met for accuracy, which might be due to low and variable recovery. The reason for the relative low recovery can be caused by the sample preparation and explained by the extraction effect. Recovery following protein precipitation is frequently incomplete.

The analytical method was not able to quantify the low pleura and plasma concentrations of PEM 20 h after drug administration. Key period of interest is the late elimination phase where accumulated drug is released back into the systemic circulation. Though, the behavior of late-phase pharmacokinetics of PEM could not be adequately captured in the study.

Common approaches for handling of concentration measurements reported as below quantification limit, BQL, such as dismiss the data point or by dividing LLOQ with two, have been shown to introduce bias in parameter estimates [35, 36]. It can be understood that concentrations below LLOW cannot reliably measure a true concentrations below LLOQ [35]. Studies have shown that severe parameter bias can occur when data are omitted from a pharmacokinetic analysis when missing data may attribute to wrong assumptions [37, 38]. It will probably have a little impact on the model parameter estimates in a one-compartment model but will have a larger impact on analysis with a two-compartment model.

However, these limitations, the outcome of this study points to that PEM, do not seem to be accumulating in the pleural cavity. Further study with a larger population is though needed to fully clarify this with an acceptable significance as there in this study were too few data points to explain the relationship between the factors that are responsible for the variability and particularly the hazard risk for PEM pleura toxicity in renal impairment.


Conflict of interest

None of the authors declare any conflict of interest.


  1. 1.
    Frei E III, Jaffe N, Tattersall MH, Pitman S, Parker L (1975) New approaches to cancer chemotherapy with methotrexate. New Engl J Med 292:846PubMedCrossRefGoogle Scholar
  2. 2.
    Evans WE, Pratt CB (1978) Effect of pleural effusion on high-dose methotrexate kinetics. Clin Pharmacol Ther 23(1):68–72PubMedGoogle Scholar
  3. 3.
    Cotes JE (2006) Lung function: physiology, measurement and application in medicine, Chapter 1 and 2, 6th edn. Blackwell publishing, Australia. ISBN 978-0-632-06493-9CrossRefGoogle Scholar
  4. 4.
    Thomas JT, Musani AI (2013) Malignant pleural effusions: a review. Clin Chest Med 34:459–471PubMedCrossRefGoogle Scholar
  5. 5.
    Hyeon Y (2011) Management of pleural effusion, empyema and lung abscess. Semin Intervent Radiol 28(1):75–86CrossRefGoogle Scholar
  6. 6.
    Light R, Lee GY (2008) Textbook of pleural diseases, 2nd edn. Hodder Education, LondonGoogle Scholar
  7. 7.
    Pass HI, Carbone DP, Johnson DH, Minna JD et al (2010) Principles and practice of lung cancer, 4th edn. Lippincott Williams & Wilkins, USA. ISBN 978-0-7817-7365-2Google Scholar
  8. 8.
    Zarogoulidis K, Zarogoulidis P, Darwiche K, Tsakiridis K et al (2013) Malignant pleural effusion and algorithm management. J Thorac Dis 5(54):S413–S419PubMedCentralPubMedGoogle Scholar
  9. 9.
    Putnam JB (2002) Malignant pleural effusions. Surg Clin North Am 82:867–883PubMedCrossRefGoogle Scholar
  10. 10.
    Hung TL, Chen FF, Liu JM, Lai WW et al (2003) Clinical evaluation of HER-2/neu protein in malignant pleural effusion-associated lung adenocarcinoma and as tumor marker in pleural effusion diagnosis. Clin Cancer Res 9:2605–2612PubMedGoogle Scholar
  11. 11.
    Gonlugur TE, Gonlugur U (2008) Transudates in malignancy: still a role for pleural fluid. Ann Acad Med Singapore 37(9):760–763PubMedGoogle Scholar
  12. 12.
    Thomas R, Lee G (2013) Causes and management of common benign pleural effusions. Thorac Surg Clin 23:25–42PubMedCrossRefGoogle Scholar
  13. 13.
    Clementsen PS (2001) Pleuracentese. Ugeskr Læger 163:2639–2640PubMedGoogle Scholar
  14. 14.
    Light RW, Hamm H (1997) Malignant pleural effusion: would the real cause please stand up? Eur Respir J 10:1701–1702PubMedCrossRefGoogle Scholar
  15. 15.
    Kaifi JT, Toth JW, Gusani NJ, Kimichi ET et al (2012) Multidisciplinary management of malignant pleural effusion. J Surg Oncol 105:731–738PubMedCrossRefGoogle Scholar
  16. 16.
    Rinaldi DA, Burris HA, Dorr FA, Woodworth JG et al (1995) Initial phase I evaluation of the novel thymidylate synthase inhibitor, LY231514, using the modified continual reassessment method for dose escalation. J Clin Oncol 13(11):2842–2850PubMedGoogle Scholar
  17. 17.
    Guidance for Industry—Bioanalytical Method Validation (2010) U S Department of Health and Human Services, Food and Drug Administration, Center for Drug Evaluation and Research 2001. pp 1–22Google Scholar
  18. 18.
    Altman D, Bland MJ (2005) Standard deviations and standard errors. Statistics note. BMJ 331(7521):903PubMedCentralPubMedCrossRefGoogle Scholar
  19. 19.
    Rosenbaum SE (2011) Basic pharmacokinetic and pharmacodynamic: an integrated textbook and computer simulations. Wiley, New Jersey. ISBN 978-0-470-56906-1Google Scholar
  20. 20.
    Rowland M, Tozer TN (2011) Clinical pharmacokinetics and pharmacodynamics: concepts and applications, 4th edn. Williams & Wilkins, LippincottGoogle Scholar
  21. 21.
    Guidance for Industry (1998) Pharmacokinetics in patients with impaired renal function—study design, data analysis, and impact on dosing and labeling.
  22. 22.
    Rinaldi DA, Kuhn JG, Burris HA, Dorr FA et al (1999) A phase I evaluation of multi-targeted antifolate (MTA, LY231514), administered every 21 days, utilizing the modified continual reassessment method for dose escalation. Cancer Chemother Pharmacol 44:372–380PubMedCrossRefGoogle Scholar
  23. 23.
    McDonald AC, Vasey PA, Adams L et al (1998) A phase I and pharmacokinetic study of LY231514, the multitargeted antifolate. Clin Cancer Res 4:605–610PubMedGoogle Scholar
  24. 24.
    Quellet D, Periclou AP, Johnson RD, Woodworth JR et al (2000) Population pharmacokinetics of pemetrexed disodium (ALIMTA) in patients with cancer. Cancer Chemother Pharmacol 46:227–234CrossRefGoogle Scholar
  25. 25.
    Dickgreber NJ, Fink TH, Latz JE, Hossain AM et al (2009) Phase I and pharmacokinetic study of pemetrexed plus cisplatin in chemonaive patients with locally advanced or metastatic malignant pleural mesothelioma or non-small cell lung cancer. Clin Cancer Res 15:382–389PubMedCrossRefGoogle Scholar
  26. 26.
    Dickgreber NJ, Sorensen JB, Paz-Ares LG, Schytte TK et al (2010) Pemetrexed safety and pharmacokinetics in patients with third-space fluid. Clin Cancer Res 16(10):2872–2880PubMedCrossRefGoogle Scholar
  27. 27.
    Nakagawa K, Kudoh S, Matsui K, Negoro S et al (2006) A phase I study of pemetrexed (LY231514) supplemented with folate and vitamin B12 in Japanese patients with solid tumours. Br J Cancer 95:677–682PubMedCentralPubMedCrossRefGoogle Scholar
  28. 28.
    Latz JE, Chaudhary A, Ghosh A, Johnson RD (2006) Population pharmacokinetic analysis of ten phase II clinical trials of pemetrexed in cancer patients. Cancer Chemother Pharmacol 57:401–411PubMedCrossRefGoogle Scholar
  29. 29.
    Hughes A, Calvert P, Azzabi A, Plummer R et al (2002) Phase I clinical and pharmacokinetic study of pemetrexed and carboplatin in patients with malignant pleural mesothelioma. J Clin Oncol 20(16):3533–3544PubMedCrossRefGoogle Scholar
  30. 30.
    Li J, Gwilt P (2002) The effect of malignant effusions on methotrexate disposition. Cancer Chemother Pharmacol 50:373–382PubMedCrossRefGoogle Scholar
  31. 31.
    Mita AC, Sweeney CJ, Baker SD, Goetz A et al (2006) Phase I and pharmacokinetic study of pemetrexed administered every 3 weeks to advanced cancer patients with normal and impaired renal function. J Clin Oncol 24(4):552–562PubMedCrossRefGoogle Scholar
  32. 32.
    Undevia SD, Gomez-Abuin G, Ratain MJ (2005) Pharmacokinetic variability of anticancer agents. Nat Rev Cancer 5:447–458PubMedCrossRefGoogle Scholar
  33. 33.
    Huffman DH (1973) Pharmacokinetics of methotrexate. Clin Pharmacol Ther 14:572–579PubMedGoogle Scholar
  34. 34.
    Chabner BA, Young RC (1973) Threshold methotrexate concentration for in vivo inhibition of DNA synthesis in normal and tumorous target tissues. J Clin Invest 52:1804–1811PubMedCentralPubMedCrossRefGoogle Scholar
  35. 35.
    Beal SL (2001) Ways to fit a PK model with some data below the quantification limit. J Pharmacokinet Pharmacodyn 28(5):481–504PubMedCrossRefGoogle Scholar
  36. 36.
    Bergstrand M, Karlsson MO (2009) Handling data below the limit of quantification in mixed effect models. AAPS J 11(2):371–380PubMedCentralPubMedCrossRefGoogle Scholar
  37. 37.
    Duval V, Karlsson MO (2002) Impact of omission or replacement of data below the limit of quantification on parameter estimates in a two-compartment model. Pharm Res 19:1835–1840PubMedCrossRefGoogle Scholar
  38. 38.
    Bonate PL (2011) Pharmacokinetic-pharmacodynamic modeling and simulation. Chapter 8. Nonlinear mixed effects models: practical issues, 2nd edn. Springer, Deerfield, pp 303–358. doi: 10.1007/978-1-4419-9485-1-8

Copyright information

© Springer-Verlag Berlin Heidelberg 2014

Authors and Affiliations

  • Per Hartvig Honoré
    • 1
  • Sigrid Jóhansdóttir Joensen
    • 1
  • Michelle Olsen
    • 1
  • Steen Honoré Hansen
    • 2
  • Anders Mellemgaard
    • 3
  1. 1.Systems Pharmacology, Department of Drug Design and Pharmacology, Faculty of Health and Medical SciencesUniversity of CopenhagenCopenhagenDenmark
  2. 2.Department of PharmacyUniversity of CopenhagenCopenhagenDenmark
  3. 3.Department of OncologyHerlev HospitalCopenhagenDenmark

Personalised recommendations