Predicting Maternal and Infant Tetrahydrocannabinol Exposure in Lactating Cannabis Users: A Physiologically Based Pharmacokinetic Modeling Approach

Abstract

A knowledge gap exists in infant tetrahydrocannabinol (THC) data to guide breastfeeding recommendations for mothers who use cannabis. In the present study, a paired lactation and infant physiologically based pharmacokinetic (PBPK) model was developed and verified. The verified model was used to simulate one hundred virtual lactating mothers (mean age: 28 years, body weight: 78 kg) who smoked 0.32 g of cannabis containing 14.14% THC, either once or multiple times. The simulated breastfeeding conditions included one-hour post smoking and subsequently every three hours. The mean peak concentration (C_max) and area under the concentration-time curve (AUC_{(0-24 h)}) for breastmilk were higher than in plasma (C_max: 155 vs. 69.9 ng/mL; AUC_{(0-24 h)}: 924.9 vs. 273.4 ng·hr/mL) with a milk-to-plasma AUC ratio of 3.3. The predicted relative infant dose ranged from 0.34% to 0.88% for infants consuming THC-containing breastmilk between birth and 12 months. However, the mother-to-infant plasma AUC_{(0-24 h)} ratio increased up to three-fold (3.4-3.6) with increased maternal cannabis smoking up to six times. Our study demonstrated the successful development and application of a lactation and infant PBPK model for exploring THC exposure in infants, and the results can potentially inform breastfeeding recommendations.

Keywords: PBPK; THC in breastmilk; cannabinoids; cannabinoids in breastmilk; cannabis smoking; infant THC exposure; marijuana and breastfeeding; pediatric cannabis exposure; smoking and lactation; weed; weed and breastfeeding.

Figure 1

Workflow of model development for the maternal lactation and infant PBPK model. IV, intravenous; PBPK, physiologically based pharmacokinetic model.

Figure 2

Schematics of the maternal and infant PBPK model structure, highlighting modes of drug input, distribution, and elimination. ET₂, extrathoracic (excluding the nose); BB, bronchial; bb, bronchiolar; and AL, alveolar region; CL_H, hepatic clearance; Q_bt, blood flow to the breast; CL_btmk, drug clearance from the breast into milk; CL_mkbt, drug clearance from milk into the breast.

Figure 3

Schematics of particle deposition following (a) inhalation and exhalation and (b) absorption of deposited particles through the bronchial region. fDE, fraction deposited; ${\mathrm{k}}_{{\mathrm{t}\mathrm{r}}_{\mathrm{B}\mathrm{B}}}$, ${\mathrm{k}}_{{\mathrm{t}\mathrm{r}}_{\mathrm{b}\mathrm{b}}}$, transit rate constant at bronchial and bronchiolar regions (1/h), respectively; ${\mathrm{k}}_{{\mathrm{d}}_{\mathrm{B}\mathrm{B}}}$, dissolution rate constant at bronchial region (1/h); ${\mathrm{k}}_{{\mathrm{m}\mathrm{c}\mathrm{c}}_{\mathrm{B}\mathrm{B}}}$, ${\mathrm{k}}_{{\mathrm{m}\mathrm{c}\mathrm{c}}_{\mathrm{B}\mathrm{B}}}$, mucociliary transit rate constant at bronchial and bronchiolar regions (1/h), respectively; ${\mathrm{P}\mathrm{S}}_{\mathrm{B}\mathrm{B}}$, permeability surface area product of bronchial epithelium (L/h); ${\mathrm{Q}}_{\mathrm{l}\mathrm{u}}$, blood flow to the lungs.

Figure 4

Plasma concentration profiles following a short intravenous infusion. The open blue circle represents the observed data. The solid line depicts the median PBPK predicted concentrations, while the shaded area (5th to 95th percentile) indicates the 90% prediction interval. The purple and blue profiles display the overlay of PBPK simulated values with the observed values in the training and verification datasets, respectively [85,86,87,88,89,90,91].

Figure 5

Median (interquartile range) predicted/observed ratios for AUC and Cmax following intravenous administration of THC. The purple median (IQR) corresponds to the results from the training datasets, while the blue median (IQR) represents the results from the verification datasets. Each median (IQR) on the plot is accompanied by the absolute average fold errors (AAFE), which provide an overall measure of the deviation between simulated and observed values. The acceptance criterion for the ratio is set between 0.5 and 1, while the criterion for the AAFE is set between 1 and 2. Multiple datasets were included from certain studies where both chronic (a) and casual (b) cannabis use scenarios were investigated within the same study [85,86,87,88,89,90,91].

Figure 6

(a) Plasma concentration profiles following THC inhalation after smoking cannabis. The open blue circle represents the observed data. The solid line depicts the median PBPK predicted concentrations, while the shaded area (5th to 95th percentile) indicates the 90% prediction interval. The purple and blue profiles display the overlay of PBPK simulated values with the observed values in the training and verification datasets, respectively. (b) Plasma concentration profiles following THC inhalation after smoking cannabis. The open blue circle represents the observed data. The solid line depicts the median PBPK predicted concentrations, while the shaded area (5th to 95th percentile) indicates the 90% prediction interval. The blue profile displays the overlay of PBPK simulated values with the observed values in the verification datasets [85,86,92,93,94,95,96,97].

Figure 7

Median (interquartile range) predicted/observed ratios for AUC following inhalation of cannabis smoke. The purple median (IQR) corresponds to the results from the training datasets, while the blue median (IQR) represents the results from the verification datasets. Each median (IQR) on the plot is accompanied by the absolute average fold errors (AAFE). The acceptance criterion for the ratio is set between 0.5 and 1, while the criterion for the AAFE is set between 1 and 2. Multiple datasets were included from certain studies that examined a range of scenarios, including chronic (a) and casual (b) cannabis use, as well as variations in THC content such as low (l), medium (m), and high (h) THC content in cannabis joints within the same study [85,86,92,93,94,95,96,97].

Figure 8

Breastmilk concentration profiles following THC inhalation after smoking cannabis. The open blue circle represents the observed data. The solid line depicts the median PBPK predicted concentrations, while the shaded area (5th to 95th percentile) indicates the 90% prediction interval. Eight lactating mothers participated in the Baker et al. study, and their profile is shown on the graph. However, in the Bertrand et al. study, 50 mothers who reported recent marijuana use contributed one breastmilk sample at different times [13,98].

Figure 9

The predicted median (IQR) plasma (Purple) and breastmilk (blue) concentration of mothers who smoked one, two, three, four, five, or six times per day. Multiple smoking sessions were spaced evenly within a 24 h period.

Figure 10

Simulated plasma concentration profile from the infant PBPK model and scaled Klumpers et al. [108] population pharmacokinetic model.

Figure 11

Simulated plasma concentration profile for infants up to one year of age. It was assumed baby feeds every three hours, and the breast milk concentration at the time of feeding depends on the number of times the lactating mother smoked. One to six times smoking frequency per day was tested and represented by different colors, as shown in the legend.

Figure 12

Sensitivity analysis of exposure metrics to variations in fixed model input parameters. The viscosity and hygroscopicity of THC smoke were assessed in relation to the fraction deposited, while solubility was evaluated against the plasma area under the curve (AUC). Additionally, the breast partition coefficient was examined in correlation with breast milk AUC.