Transfer of escitalopram and its metabolite demethylescitalopram into breastmilk

Abstract

Aims

To investigate the transfer of escitalopram and its demethyl metabolite into milk, the absolute and relative infant doses via milk and to assess any unwanted effects in the breastfed infant.

Methods

Multiple samples of blood and milk were obtained over a dose interval at steady state from eight women who were taking escitalopram for postnatal depression. Drug concentrations in plasma and milk were measured by high-performance liquid chromatography and milk/plasma ratio (M/P_AUC), absolute infant dose and relative infant dose were estimated by standard methods. Their breastfed infants were also examined clinically and in five infants a blood sample was taken for drug analysis.

Results

The median dose taken by the women was 10 mg day⁻¹. The mean (95% confidence interval) M/P_AUC was 2.2 (2.0, 2.4) for escitalopram and 2.2 (1.9, 2.5) for demethylescitalopram. Absolute infant doses were 7.6 µg kg⁻¹ day⁻¹ (5.2, 10.0) for escitalopram and 3.0 µg kg⁻¹ day⁻¹ (2.4, 3.6) for demethylescitalopram. The total relative infant dose for escitalopram plus its demethyl metabolite was 5.3% (4.2, 6.4) as escitalopram equivalents. All of the infants had met normal developmental milestones and no adverse effects were seen. Compared with average maternal plasma concentrations (24 µg l⁻¹), the concentrations of the parent drug and its metabolite in plasma from five infants were most commonly below the limit of detection (≤3 µg l⁻¹).

Conclusion

The study shows that escitalopram is safe for use during breastfeeding. Because its absolute infant dose is lower than that for an equivalent antidepressant dose of rac-citalopram, it may be preferred over rac-citalopram in treating depression in lactating women. Nevertheless, each decision to breastfeed should always be made on the basis of an individual risk:benefit analysis.

Introduction

Because of the 10–15% frequency of depression in the postnatal period, antidepressants are amongst the most frequently used drugs in lactation [1]. Children of depressed mothers may exhibit behavioural problems and delayed language development [2–5]. Thus it is well accepted that treatment of depression in the perinatal period has significant benefits for both mother and infant. We and others have previously studied the transfer of rac-citalopram and its demethyl-metabolite into milk [6–12]. Overall, both drugs had milk/plasma (M/P_AUC) transfer ratios of around 2–3, a combined relative infant dose of 2–6% of the maternal weight-adjusted dose and have been considered to be safe for use during breastfeeding. There is only one documented case report of adverse events (disturbed sleep) in a breastfed infant whose mother was taking 40 mg rac-citalopram daily [10], although, like most selective serotonin re-uptake inhibitors, citalopram use during pregnancy may result in a short-term serotonin toxicity syndrome in the neonate [12].

In recent years, pharmaceutical companies have been encouraged to market active drug enantiomers rather than racemates in order to produce greater selectivity in primary pharmacological actions and lower side-effects. This trend is important for psychotropic drugs, including antidepressants [13]. The major pharmacological activity resides in the S-(+) enantiomer of citalopram (escitalopram) [14–16]. In addition, R-(−)-citalopram counteracts the activity of S-(+)-citalopram [15], thereby providing a possible basis for the pharmacological and clinical differences observed between citalopram and escitalopram [17].

Most drugs are transferred into milk by passive diffusion processes and hence maternal drug concentration (controlled by maternal pharmacokinetics) and M/P are major determinants of infant dose via milk [18]. To our knowledge, the only published data for milk transfer of enantiomers are those for methadone following maternal administration of the racemate [19]. In this study, the (M/P) for R-methadone was 1.8 times higher than that for S-methadone (means of 0.68 and 0.38, respectively). Moreover, average concentrations of R-methadone in milk were 181 µg l⁻¹ compared with 121 µg l⁻¹ for S-methadone. The aim of the present study was therefore to investigate the transfer of escitalopram and its demethyl metabolite into milk, to estimate the absolute and relative infant doses via milk and to assess any unwanted effects in the breastfed infant.

Materials and methods

Patients

Eight breastfeeding women who were being treated with escitalopram (Lexapro®) for postnatal depression gave written informed consent and were enrolled in the study.

Study protocol and data collection

The study was approved by the Ethics Committee of the Women’s and Children’s Health Service, Subiaco, Australia. The mothers were admitted to the research ward at 08.00 h and had a venous catheter inserted in a forearm vein. Venous blood samples (5 ml each, lithium heparin tube) were taken just before the morning dose of escitalopram at around 08.30 h and also at 2, 4 and 6 h after dose. The women were discharged following the 6 h sample and further blood samples were taken by venepuncture at home (by K.F.I.) at 8, 12 and 24 h after dose. The mothers collected milk samples (8 ml each) by hand expression or manual breast pump just before the morning dose and again every time their infant fed during the next 24 h (usually six to seven feeds). For early feeds whilst in the research ward the women usually collected both fore- and hind-milk samples separately, and for all other samples they collected an equal mixture of fore- and hind-milk into a single tube.

Infant health and wellbeing were evaluated by enquiry of the mother, together with a full clinical examination including a Denver development assessment by a specialist neonatologist (R.K.) [20, 21]. Results for the Denver assessment were expressed as the quotient of Denver age/chronological age (as percentage).

Materials

Authentic escitalopram and demethylescitalopram standards were obtained from Lundbeck Australia Pty Ltd, and desipramine hydrochloride (internal standard for the high-performance liquid chormatography assay) from Ciba-Geigy Australia Ltd. All other solvents and chemicals were of analytical grade.

High performance liquid chromatography

Escitalopram and demethylescitalopram in plasma and milk were measured as previously described [9]. Intra- and interday relative standard deviations for the assays in milk and plasma were < 6.4% and < 13.2%, respectively, over the relevant concentration ranges, with a limit of quantification (LOQ) of 1 µg l⁻¹ in both matrices.

Data analysis

Areas under the plasma and milk concentration–time curves (AUC_{0–24 h}) were calculated using the mixed log-linear trapezoidal rule as appropriate [22] and the milk to plasma ratio (M/P_AUC) was calculated from the relevant AUC measurements. The average drug concentration in milk was calculated as C_av = (AUC_{0–24 h})/τ, where τ = the dose interval (nominally 24 h). In graphing the data in Figure 1 and estimating AUC, the concentration values for paired fore- and hind-milk samples were averaged and plotted/analysed against the average of their sampling times. Statistical analysis of data was performed using SigmaStat Ver. 3.1 (SPSS Inc., Chicago, IL, USA).

Figure 1.

Representative milk and plasma concentration–time profiles for escitalopram (▪, milk; •, plasma) and demethylescitalopram (□, milk; ○, plasma) in patients 2 (dose 10 mg), 3 (dose 20 mg), 5 (dose 10 mg) and 7 (dose 10 mg)

Absolute infant dose (µg kg⁻¹ day⁻¹) was calculated as the product of C_av and an average infant milk intake of 0.15 l kg⁻¹ day⁻¹ [23]. The relative infant dose was calculated as absolute infant dose/maternal dose (µg kg⁻¹ day⁻¹) and expressed as a percentage [18]. Maximum concentrations (C_max) of drug in milk and plasma and the times of these peaks (T_max) were interpolated from the primary data. Data have been summarized as mean [95% confidence interval (CI), or range] or median and range, as appropriate.

Results

The characteristics of the mothers and their infants are summarized in Table 1. The median single daily dose of escitalopram taken by the women was 10 mg (range 10–20 mg) and therapy had commenced a median of 55 days prior to the study day, and they were therefore considered to be at steady state. The infants were five females and three males and all had achieved normal weight for age milestones at the time of study. The paediatric assessment of the infants revealed no adverse findings and the mean Denver developmental age was 110% (95% CI 99, 121) of the actual age.

Table 1.

Characteristics of the mothers and their infants

Group	Parameter
Mothers
Age (years)	32 (24–42)*
Weight (kg)	68 (56–85)*
Escitalopram dose (mg day−1)	10 (10–20)†
Duration of treatment (days)	55 (23–240)†
Infants
Sex (M/F)	3/5
Gestational age (months)	39.1 (35.3–40.1)†
Chronological age (months)	3.6 (1.5–9.5)†
Birth weight (kg)	3.3 (2.5–3.8)†
Weight on study day (kg)	5.9 (3.7–10.5)†

^*Mean (range).

^†Median (range).

Representative milk and plasma concentration–time profiles for four of the eight women are shown in Figure 1. Maximum drug concentrations in milk and plasma, the times at which they occurred and the M/P_AUC values are summarized in Table 2. C_max and C_av for escitalopram and demethylescitalopram were consistently higher in milk than in plasma and the M/P for both drugs was 2.2. In addition, for escitalopram the median T_max was significantly greater (Wilcoxon rank signed test T = −34, P = 0.023) in milk than in plasma, indicating a delay in equilibration between the two compartments. By contrast, for demethylescitalopram, which exhibited a flatter concentration–time profile, T_max was similar in milk and plasma.

Table 2.

Escitalopram and demethylescitalopram C_max, T_max and C_av for milk and plasma and their M/P_AUC values

	Milk			Plasma
Patient	Cmax (µg l−1)	Tmax (h)	Cav (µg l−1)	Cmax (µg l−1)	Tmax (h)	Cav (µg l−1)	M/PAUC
Escitalopram
1	64	2.8	47	38	1.5	28	1.7
2	82	5.1	56	41	5.4	25	2.2
3	168	2.1	99	72	1.9	49	2.0
4	57	5.9	39	23	4.1	18	2.0
5	84	8.7	53	37	4.9	28	1.9
6	73	11.3	58	39	2	25	2.3
7	57	4.9	27	15	3.3	9	2.7
8	37	10.3	28	15	3.4	12	2.4
Mean (95% CI)	78 (51, 105)	5.5 (4.4, 9.1)*	51 (35, 67)	35 (22, 48)	3.4 (2.0, 4.3)*	24 (15, 33)	2.2 (2.0, 2.4)
Demethylescitalopram
1	36	1	28	9	1	20	2.3
2	29	2.1	25	11	5.4	23	2.4
3	41	2.1	49	14	3.9	31	2.3
4	22	5.9	18	10	7.1	18	1.9
5	19	8.7	28	7	4.9	13	2.0
6	26	4.7	25	14	4.2	23	1.8
7	27	4.8	10	9	4.3	18	2.1
8	17	16.1	12	6	2.1	15	3.1
Mean (95% CI)	27 (21, 33)	4.8 (2.1, 6.6)*	24 (16, 32)	10 (8, 12)	4.3 (3.5, 5.0)*	20 (16, 24)	2.2 (1.9, 2.5)

^*Median (25th and 75th percentiles).

Table 3 summarizes the weight-corrected maternal escitalopram doses, as well as the calculated infant doses and measured drug concentrations in the infant’s plasma. The total relative infant dose for the combination escitalopram and its metabolite was a mean of 5.3% of the maternal weight-adjusted dose. Overall, escitalopram contributed some 70% of the oral dose that the breastfed infant is calculated to receive. In five of the studies, a blood sample was able to be taken from the breastfed infants (Table 3) and analysed for drug content. The limit of detection (LOD) in these samples varied according to the volume of plasma that was available. Both escitalopram and its metabolite were undetectable in four of the samples and were present in one infant (patient 7) at very low levels of 3 µg l⁻¹ and 2 µg l⁻¹, respectively.

Table 3.

Maternal dose of escitalopram, absolute and relative infant doses, and infant plasma concentrations of escitalopram and demethylescitalopram

		Escitalopram			Demethylescitalopram
Patient	Maternal dose (µg kg−1 day−1)	Total relative infant dose*(%)	Absolute infant dose (µg kg−1 day−1)	Relative infant dose (%)	Infant plasma concentration (µg l−1)	Absolute infant dose (µg kg−1 day−1)	Relative infant dose*(%)	Infant plasma concentration (µg l−1)
1	179	7.0	3.9	ND	3.2	1.8	ND	5.7
2	139	8.3	6.0	<5	3.6	2.6	<3	6.1
3	235	14.9	6.3	<3	4.6	2.0	<2	8.3
4	299	5.8	1.9	ND	2.7	1.0	ND	2.9
5	180	8.0	4.5	ND	2.0	1.1	ND	5.6
6	267	8.7	3.3	<1	3.4	1.3	<1	4.6
7	143	4.1	2.9	3	2.7	2.0	2	4.8
8	154	4.3	2.8	<1	2.3	1.5	<1	4.3
Mean (95% CI)	199 (157, 241)	7.6 (5.2, 10.0)	3.9 (2.8, 5.0)		3.0 (2.4, 3.6)	1.7 (1.3, 2.1)		5.3 (4.2, 6.4)

^*Expressed in escitalopram equivalents. ND, No plasma sample available.

During the course of the study we collected 24 paired samples of fore- and hind-milk for drug analysis. For escitalopram there was a significant difference (paired t = 5.3, P < 0.001) between fore- and hind-milk means of 9 µg l⁻¹ (95% CI for difference 6.5, 12.4). For demethylescitalopram the median fore- (21 µg l⁻¹) and hind-milk (24 µg l⁻¹) concentrations were also significantly different (Wilcoxon rank signed test T = −13, P < 0.001). On average, the ratio of drug concentration in hind-milk to that in fore-milk was 1.2 (95% CI 1.12, 1.28) for escitalopram and 1.23 (1.1, 1.36) for demethylescitalopram.

Discussion

Escitalopram oral pharmacokinetics are characterized by a T_max of 4 h, plasma albumin binding of around 80%, a volume of distribution (V_d/F) of 12–36 l, a t_1/2β of 30 h and an oral clearance of 36 l h⁻¹[24]. Its main metabolite is demethylescitalopram (28–31%) and there is also a small amount of didemethylescitalopram (<5%) [24]. Comparative data for oral administration of R-(−) citalopram alone are not available but after oral administration of 40–60 mg day⁻¹ of rac-citalopram for 6 weeks, the average serum concentration of S-(+)-citalopram was 61% of that for R-(−)-citalopram, while the average serum concentration of S-(−)-demethylescitalopram was 71% of that for R-(+)-desmethylescitalopram [25]. Moreover, the mean t_1/2, steady-state volume of distribution and oral clearance for R-(−)-citalopram were 47 h, 16.4 l kg⁻¹ and 0.26 l kg⁻¹ h⁻¹, respectively, compared with 35 h, 18.3 l kg ⁻¹and 0.48 l kg⁻¹ h⁻¹, respectively, for S-(+)-citalopram. The mean t_1/2 for R-(−)-demethylescitalopram was significantly longer (70 h) than that for S-(+)-demethylescitalopram (51 h). These data show that the S-(+) enantiomer of citalopram has a higher oral clearance that that of the corresponding R-(−) enantiomer and suggest that there is a similar higher clearance for the S-(+)-desmethylescitalopram compared with that for its R-(−)-enantiomer. As the R-(−)-enantiomer accounts for some two-thirds of steady-state concentrations following administration of rac-citalopram [25], the total drug load to the breastfed infant would be markedly reduced with escitalopram compared with rac-citalopram. In addition, the recommended dose ranges of escitalopram (10–20 mg) [24] and citalopram (20–60 mg) [26] suggested better efficacy for escitalopram and led us to hypothesize that the absolute infant dose of escitalopram would be lower than that for rac-citalopram, conferring a higher safety margin when the drug is used to treat depression in breastfeeding women.

As expected, in the present study the median dose of escitalopram (10 mg) was significantly lower (Mann–Whitney rank sum test T = 76.5, P = 0.014) than that reported (20 mg) in our earlier study of the transfer of rac-citalopram into milk in lactating women [9]. In addition, the absolute infant dose for escitalopram [mean (95% CI) 7.6 µg kg⁻¹ day⁻¹ (5.2, 10.0)] was also significantly lower (52%) than that for rac-citalopram [14.5 µg kg⁻¹ day⁻¹ (8.3, 20.6)] in our earlier study (t-test, t = −6.9, P = 0.023). By contrast, the absolute infant doses of demethylescitalopram [3.0 µg kg⁻¹ day⁻¹ (2.4, 3.6)] and rac-demethylcitalopram [5.4 µg kg⁻¹ day⁻¹ (2.2, 8.6)] were not significantly different, indicating that exposure to this metabolite, which has a long t_1/2[25], is independent of the dose forms used in our present and previous studies. Overall, we conclude that absolute infant dose is lower when escitalopram is used instead of citalopram at clinically comparable antidepressant doses. The lower absolute infant dose results because the R-enantiomer is not present.

The relative infant doses for escitalopram [3.9% (2.8, 5.0)] and demethylescitalopram [1.7% (1.3, 2.1)] in the present study were, as expected, similar to those reported for rac-citalopram [3.7% (2.6, 4.8)] and rac-demethylcitalopram [1.4% (0.9, 1.9)] in our earlier study [9]. Hence, when comparing infant exposure from a pure enantioner drug formulation to that from a clinically comparable dose of its racemate formulation, the relative infant dose may not reveal the whole story and recourse to a comparison at the level of the absolute dose is more appropriate and informative.

In calculating total relative infant doses (escitalopram plus demethylescitalopram; see Table 3), we have valued the contribution of demethylescitalopram in escitalopram ‘equivalents’, using the parent:metabolite molecular weight ratio. Hence, our total relative infant dose calculation is very conservative in that S-(+)-demethylescitalopram has only 15% of the activity of escitalopram in inhibiting 5-hydroxytryptamine reuptake into rat brain synaptosomes [14]. Moreover, demethylescitalopram is less polar [9] and therefore less likely to enter the brain.

The M/P_AUC values for escitalopram [2.2 (2.0, 2.4)] and demethylescitalopram [2.2 (1.9, 2.5)] in the present study were similar to those we reported for rac-citalopram [1.8 (1.2, 2.3)] and rac-demethylcitalopram [1.8 (1.1–2.5)][9]. The lack of a significant difference suggests that the R-enantiomers of citalopram and its demethylmetabolite transfer to milk similarly to their respective S-enantiomers. Infant exposure is thus dependent primarily on maternal plasma concentration of the pure enantioner(s) and is unlikely to be influenced by chirality.

On the basis of the lower absolute infant dose calculations above, we conclude that escitalopram is preferred to rac-citalopram in the treatment of depression during lactation. Our assessment of the eight breastfed infants in our study revealed no adverse effects and low or undetectable plasma drug levels in five of the infants where a blood sample was able to be taken. The actual concentrations (or the LODs) for the infant sample as percent of the C_av in maternal plasma were 14% (4, 31) for escitalopram mean (range), and 20% (12, 36) for demethylescitalopram. These percentages are conservatively estimated, in that the LODs for two of the infant samples were high due to the small sample volumes available. While our positive clinical observations suggest that escitalopram has a low potential for causing adverse events in the breastfed infant, it should be remembered that our cohort is small and there is one case of an adverse event for rac-citalopram [10].

Overall, with a mean relative infant dose of 5.3% of the weight-adjusted maternal dose for escitalopram plus its major demethyl metabolite, our study shows that escitalopram is safe (notional level of safety < 10%) for use during breastfeeding and, because the absolute infant dose is lower, it should be preferred over rac-citalopram in treating depression in lactating women. Nevertheless, each decision to breastfeed should always be made on the basis of an individual risk:benefit analysis.