Nevirapine Pharmacokinetics and Safety in Neonates Receiving Combination Antiretroviral Therapy for Prevention of Vertical HIV Transmission
Background: Nevirapine (NVP)-based combination antiretroviral therapy is routinely prescribed to infants deemed at high risk of vertical HIV infection in our centers. We evaluated NVP pharma- cokinetics and safety of this regimen.
Methods: Neonates were recruited prospectively between Septem- ber 2012 and April 2015 or enrolled retrospectively if treated similarly before prospective study initiation. NVP was dosed at 150 mg/m2 daily for 14 days, then twice daily for 14 days. NVP levels were drawn at weeks 1, 2, and 4 [target trough (NVP-T): 3–8 mg/L].
Results: Thirty-three neonates were included (23 prospectively). Median gestational age (GA) and birth weight were 38 weeks (32–41 weeks) and 2.9 kg (1.5–4.2 kg), respectively. Median NVP-Ts were 8.2 mg/L (1.6–25.1 mg/L), 3.5 mg/L (1.6–6.8 mg/L), and 4.3 mg/L (0.1–19.9 mg/L) at weeks 1, 2, and 4, respectively. The proportions with therapeutic NVP-T were 42%, 61%, and 73% at these same timepoints. Median apparent oral clearance (CL/F) increased from 0.05 L$kg21$h21 (0.01–0.50 L$kg21$h21) at week 2 to 0.18 L$kg21$h21 (0.01–0.78 L$kg21$h21) at week 4. Increased drug exposure [area under the curve (AUCt)] correlated with younger GA (r = 0.459, P = 0.032) and lower birth weight (r = 0.542, P = 0.009). The most common adverse events potentially attributable to combination antiretroviral therapy were transient asymptomatic hyperlactatemia (26%), anemia (24.7%), and neutropenia (22.1%).
Conclusions: Treatment dose NVP was generally well-tolerated and associated with normalization of trough levels over time in most cases without dose adjustment. Lower empiric dosing is recom- mended for infants ,34 weeks of GA. Routine therapeutic drug monitoring may not be required for infants $34 weeks of GA.
Key Words: HIV vertical transmission, pharmacokinetics, nevirapine, neonate, combination antiretroviral therapy
INTRODUCTION
The optimal management of newborns of HIV-infected mothers who were untreated or had detectable viral load before delivery despite receiving antenatal combination antiretroviral therapy (cART) remains controversial. The current U.S. Department of Health and Human Services (DHHS) guidelines recommend a 6-week course of zidovu- dine (ZDV) and 3 doses of nevirapine (NVP) in the first week
of life for such “high risk” newborns.1 However, many pediatric HIV specialists prefer to prescribe cART at treatment doses in this situation.2,3 Potential benefits of this approach include the possibility of increased prevention of vertical HIV transmission, and for neonates later found to be HIV infected, better preservation of immune function, limitation of the size of viral reservoirs, and reduction in antiretroviral medication resistance.4–6
There is limited data on the safety of cART for prevention of vertical HIV transmission in neonates and on the pharmacokinetics and appropriate dosing for some antiretroviral medications including NVP. Pharmacokinetic analysis of NVP in neonates has been largely restricted to prophylaxis studies involving term newborns receiving single-dose NVP or ongoing once-daily dosing in the context of breastfeeding (target trough 0.1 mg/L).7–12 The prophy- laxis NVP target is based on in vitro data (10 times the 50% inhibitory concentration)13 and is much lower than the 3.0 mg/L target considered appropriate for treatment of HIV infection.14 To date, there have been no published trials on the pharmacokinetics of treatment dose NVP for newborns, particularly during the first 2 weeks of life and in premature infants specifically. The primary objective of this study was to evaluate the pharmacokinetics and safety of treatment dose
NVP in the context of cART in such “high risk” newborns.
METHODS
Population and Study Design
This was a multicenter observational study involving 5 Canadian centers in Ontario and Quebec. The protocol was approved by the institutional review boards of all participat- ing institutions. The main arm of the study was prospective and recruited eligible newborns between September 2012 and April 2015. Written voluntary informed consent was obtained from participants’ legal guardians. Newborns born before Research Ethics Board’s approval who fulfilled the eligibility criteria and were managed in a similar manner were included retrospectively (April 2012–October 2013, depending on site). Consent for this latter group was waived.
Newborns were eligible if they were prescribed NVP, ZDV, and lamivudine (3TC) at treatment doses within 72 hours of birth for prevention of vertical HIV transmission. Two categories of newborns were eligible: (1) those whose mother was HIV-1 positive and had incomplete maternal virologic suppression at delivery or, in the absence of maternal viral load results, a history of nonadherence to cART, or late pregnancy initiation of cART and (2) those whose mother had unknown HIV status, no antenatal care, and significant risk factors for HIV, such as commercial sex work or intravenous drug use late in gestation, who could have been in the “window period” of HIV infection. Infants were excluded if they were born less than 32 weeks completed gestational age (GA), had life-threatening medical conditions, were unable to take oral medication, or were born to women considered at high risk of harboring NVP-resistant mutations.
NVP was dosed at 150 mg/m2 orally once daily for 14 days, then 150 mg/m2 orally every 12 hours for 14 days (total 4 weeks). This empiric NVP dose was derived from the treatment dose for older infants and children recommended in the DHHS guidelines at the time the study was conceived. ZDV and 3TC were administered at standard doses for 6 weeks.1
Therapeutic Drug Monitoring and Pharmacokinetic Sampling
Therapeutic drug monitoring (TDM) performed at 1 and 2 weeks of life was used to guide NVP dose adjustment if the NVP trough level was $20% above or below the target range of 3–8 mg/L. In such cases, doses were adjusted proportionally in relation to the target range, taking into consideration clinical factors including potential NVP-related adverse effects, prematurity, birth weight (BW), and degree of perceived risk of HIV transmission.
Consent to the full protocol included serum samples for trough NVP levels at 1, 2, and 4 weeks (concurrent with routine clinical bloodwork) as well as NVP levels 1 and 4 hours after the last NVP dose at age 4 weeks. Those who did not want additional bloodwork at 4 weeks were offered the option of study participation for trough levels only.
NVP plasma concentrations were determined using a validated liquid chromatography (LC) coupled to tandem mass spectrometry. The LC system used for the Toronto and Montreal sites was an Agilent 1200 system (Agilent Tech- nologies, Santa Clara, CA) coupled to a 4000 Q-Trap mass spectrometer (ABI SCIEX, Foster City, CA). In Ottawa, an HP1100 LC system (Agilent Technologies, Wilmington, DE) with a Supelcosil ABZ+ (15 cm · 4.6 mm, 3 mm) C18 column (Supelco, Bellefonte, PA) coupled to an API-2000 mass spectrometer (AB/MDS/Sciex, Concord, ON, Canada) was used. Quantification was performed using multiple reaction monitoring. The analytical measuring range and precision obtained from all 3 laboratory test results was acceptable and within the therapeutic ranges being targeted.
Safety Evaluations
At each study visit, patients underwent a complete history and physical examination. Routine laboratory monitoring included complete blood cell count (with differential), liver enzymes [alanine aminotransferase (ALT) and aspartate aminotransferase], and lactate measured at birth and 1, 2, and 4 weeks of age while on NVP and at 2 and 6 months of age after cART discontinuation. Toxicities were graded according to the National Institutes of Health Division of AIDS guidelines.15
Statistical and Pharmacokinetic Analysis
Baseline characteristics, pharmacokinetic parameters, and possible adverse events were described using frequencies and medians with ranges as appropriate. Potential associa- tions of NVP trough levels and pharmacokinetic parameters with GA, BW, sex, and race were assessed using the Kruskal–Wallis test, Pearson correlation coefficient, or x2 test. Laboratory parameters among those with high NVP trough levels vs. those without high levels were compared with the Mann– Whitney U test. Statistical significance was set at P , 0.05. Statistical analysis was performed using SPSS (version 20.0; SPSS Inc., Chicago, IL).
A modified area under the plasma concentration–time curve (AUCt) was calculated using the trapezoidal rule with noncompartmental analysis of trough levels and postdose levels of NVP at week 4 for the subset of 11 subjects with full pharmacokinetic sampling.16 Pharmacokinetic modeling using these data was used to compute the peak concentration (Cmax) as the highest postdose plasma NVP concentration and time to peak concentration (Tmax) as the time at which Cmax was observed for each subject. Individualized NVP pharmacoki- netic parameters were calculated using first-order, noncomp- martmental equations. Apparent oral clearance (CL/F) was calculated using the formula: CL/F = dose/AUCt, assuming that bioavailability (F) = 1 and normalized for weight.17 As there is no intravenous NVP formulation available, it is not possible to determine bioavailability, thus only apparent oral clearance (CL/F) could be estimated rather than true NVP clearance. Assuming steady state had been attained with multiple oral doses in subjects at weeks 1 and 2, the pharmacokinetic (PK) parameters obtained from the subset were used to derive CL/F and AUCt for all subjects, using trough levels drawn at weeks 1 and 2, respectively, and model- fitting with Phoenix WinNonlin 6.1.
RESULTS
Baseline characteristics of the 33 newborn infants with at least one NVP trough level are shown in Table 1. Twenty- three were recruited prospectively, and 10 were included retrospectively. Of those recruited prospectively, 13 consented to the full sampling protocol and 10 to trough levels only. The median GA (range) and BW (range) of the 33 newborns were 38 weeks (32–41 weeks) and 2906 grams (1540–4183 g), respectively. One infant, whose mother did not receive cART during pregnancy, was vertically infected in utero (based on a positive birth polymerase chain reaction).
The proportion of infants with therapeutic trough levels increased from 42.3% (11/26) at week 1, to 60.9% (14/23) at week 2, and 73.1% (19/26) at week 4. The distribution of NVP trough levels according to postnatal and GA are shown in Table 2. NVP doses were adjusted for 7 infants based on week 1 trough levels. Five infants (3 premature) required dose reductions because of supratherapeutic levels, with subsequent therapeutic levels in 2 infants (1 premature) and subtherapeu- tic levels in 3 infants (2 premature) at 4 weeks. Two term infants required an increase in dose frequency to twice daily before 2 weeks of age because of subtherapeutic levels, both with subsequent therapeutic levels at week 4. Six infants (5 term) with trough levels between 8.1 and 10.4 mg/L at week 1 had therapeutic or subtherapeutic levels in subsequent weeks despite no dose adjustment. At 2 and 4 weeks of life, the median trough level was in the therapeutic range for both premature and term infants (Table 3). In addition, all infants achieved trough levels above the standard prophylactic target of 0.1 mg/L throughout the entire study period.
The NVP concentration–time profile for the 11 infants with both trough and postdose levels at 4 weeks of age, and the trough levels and CL/F and AUCt for all patients are shown in Figure 1 and Table 3, respectively. At 4 weeks of age, increased drug exposure (AUCt) correlated with younger patients. Nonspecific clinical symptoms or signs potentially attributable to cART occurred in 15 cases and included tremor or irritability (n = 9; 11.7%), loose stools (n = 3; 3.9%), and rashes (n = 3; 3.9%). Clinical signs and symptoms were categorized as either grade 1 (93.3%; n = 14) or grade 2 (6.7%; n = 1). The grade 2 event consisted of an isolated maculo- papular rash (no fever and normal liver enzymes), potentially attributable to NVP in an infant with subtherapeutic NVP trough levels, which resolved on stopping NVP. Both grade 1 rashes were not thought to be NVP-related and resolved spontaneously without NVP discontinuation. Tremor and irritability occurred exclusively in infants of unknown HIV- status mothers with a history of illicit drug use and were attributed to neonatal abstinence syndrome in all cases. Laboratory abnormalities included hyperlactatemia (n = 20; 26%), anemia (n = 19; 24.7%), neutropenia (n = 17; 22.1%), hyperbilirubinemia, (n = 3; 3.9%) and elevated ALT/aspartate aminotransferase (n = 2; 2.6%). Laboratory abnormalities were grade 1 in 58.1% (n = 36), grade 2 in 30.6% (n = 19), grade 3 in 8.1% [n = 5; anemia (n = 3) and neutropenia (n = 2)], and grade 4 in 1.6% (n = 1; neutrophil count 0.44 ·109/L). Of the 6 with grade 3 or 4 laboratory abnormalities, only 2 had supratherapeutic NVP levels. By 6 months of age, all grade or improved significantly (n = 1, grade 1) (Table 4). There was no significant association between NVP trough level or GA with any laboratory parameter at any timepoint.
DISCUSSION
This study is the largest to date evaluating the pharmacokinetics of treatment dose NVP in neonates receiv- ing cART for prevention of vertical HIV transmission. Key findings were that treatment dose NVP was well-tolerated, exhibited delayed and highly variable absorption, and was associated with normalization of trough levels over time in most cases without dose adjustment. Prematurity and low BW were significantly associated with higher drug exposure based on AUCt.
During the first week of life, NVP trough levels were above 8 mg/L in 46% of infants, half of whom were ,37 weeks gestation at birth. If the upper limit cutoff of 12 mg/L was used, as suggested in a recent population PK model,18 only 23% (50% premature) would have been categorized as having supratherapeutic levels at 1 week of life. Furthermore, NVP levels decreased over the 4 week study period in all infants, and in 6 infants (5 term), this occurred without dose adjustment. Overall, our findings suggest that apparent oral drug clearance increases from week 2 to week 4 of life likely because of maturation and autoinduction of cytochrome P450 metabolizing enzymes.19
The increased drug exposure observed among younger GA/low BW infants is consistent with the previous litera- ture.7,11 In 2 separate studies, one of premature infants (mean GA 31.5 6 3.1 weeks) and one of term infants, both of which used prophylactic NVP dosing of 2 mg/kg for the first 14 days of life and 4 mg/kg thereafter, median trough levels were 4.2 mg/L (interquartile range 2.7–5.8 mg/L) before 14 days and 2.3 mg/L after 14 days in premature infants compared with 1.1 mg/L (interquartile range 1.0–1.4 mg/L) at 14 days in term infants.7,11 In a pharmacokinetic study of single-dose NVP in premature infants [median GA 34 weeks (range 28– 37 weeks)] in South Africa, AUCt was significantly higher and clearance significantly lower in those who were small for GA compared with those who were appropriate for GA.9 Our findings add to these data and suggest that NVP empiric dosing of 150 mg/m2 may be too high for premature infants, particularly those ,34 weeks of GA at birth.
One justification for using treatment dose NVP-based cART in high-risk exposed newborns is to provide immediate therapeutic benefit to infants subsequently found to be HIV infected. In this treatment context, a target NVP trough level $3 mg/L is recommended.14 In fact, 88.5% of infants in our cohort had trough levels above this cutoff at 1 week of age; those who did not were exclusively term infants. This finding is consistent with other studies of treatment dose NVP in term and late preterm infants where approximately 90% achieved trough levels $3 mg/L at 1 week of age.18,20 When compared with the weight-based dosing suggested by other investigators (6 mg/kg twice daily for term infants and 4 mg/kg twice daily for late premature infants),18,20–22 our empiric body surface area-based dosing provides a slightly lower total daily dose for term infants and slightly higher total daily dose for late preterm infants.
FIGURE 1. NVP concentration–time data (n = 11). Two of the original 13 infants whose care providers consented to the full protocol did not have results for all 3 timepoints and were, therefore, excluded from this analysis. Lines indicate individual samples for 11 infants taken predose (trough) and approxi- mately 1 and 4 hours postdose. The median GA of these in- fants was 38 weeks (range 32–41.7 weeks) and median BW was 3.1 kg (range 2.2–4.1 kg), respectively. Ten of the 11 infants were $37 weeks of GA, and 1 was ,34 weeks of GA, as indicated by (*). In 9 of 11 infants, concentrations continued to increase from 1 to 4 hours, indicating that peak concen- trations (Tmax) had not been reached yet.
There was high intrapatient and interpatient variability in NVP trough levels and pharmacokinetic parameters over time, possibly because of interplay between absorption and drug clearance with liver enzyme maturation.23 The relatively high apparent oral clearance at week 1 was most likely because of delayed or incomplete absorption, whereas at 2 and 4 weeks of life, the comparatively lower apparent oral clearance likely represented improved absorption more reflec- tive of steady-state conditions in newborn infants.18,22 In the subset of patients with both trough and postdose levels at 4 weeks, the concentration–time data showed a delayed and highly variable absorption/distribution phase. These findings suggest that routine TDM may need to be considered, par- ticularly for infants born prematurely.
The most common adverse effects potentially attribut- able to NVP-based cART that we observed were hyper- lactatemia, anemia, and neutropenia, which were mostly mild or moderate in severity and reversible by 2–6 months of age. We previously demonstrated that anemia, but not hyperlactatemia or neutropenia, was more common in cART recipients than in ZDV monotherapy recipients.24 As shown in previous studies, hematologic toxicity was most likely attributable to the combination of ZDV and 3TC rather than NVP.24–26 The safety of treatment dose NVP was further supported by the lack of association of NVP trough levels with hematologic parameters, serum ALT or lactate in our cohort. The low incidence of rashes and elevated trans- aminases that we observed mirrors findings of previous studies of neonates receiving prophylactic dose NVP.7,9,24 NVP at a starting dose of 6 mg/kg/dose twice daily in combination with ZDV and 3TC at standard treatment doses
was also well-tolerated in a preliminary report of 6 infants in the “Early Infant HIV Treatment in Botswana” study.22 Taken together, the current data suggest that treatment dose NVP- based cART is generally well-tolerated through the first 4 weeks of life in neonates with a GA of $32 weeks.
A major limitation of our study was the lack of postdose NVP levels at weeks 1 and 2 and the relatively small number of postdose NVP levels performed at 4 weeks. In addition, steady-state assumptions at weeks 1 and 2 were made based on week 4 data, which may have contributed to the interpatient and intrapatient variability in NVP apparent oral clearances. However, it was not deemed acceptable to request postdose measurement at all 3 timepoints because of the discomfort associated with additional bloodwork. Four weeks of age was selected as the most appropriate for this purpose because of the higher likelihood of phlebotomy being easier and, therefore, potentially less painful. The use of pharmacokinetic data from a small sample set, such as ours, for population modeling, while a limitation is considered an acceptable strategy given the inherent difficulties in conducting pharmacokinetic studies in children.27 The need for additional safety and pharmaco- kinetic data for NVP-based cART in resource-limited settings is acknowledged. Last, CYP2B6 genotype, an independent predictor of NVP clearance, was not determined.18,28
In conclusion, treatment dose NVP-based cART is generally well-tolerated in neonates with a GA at birth of $32 weeks. For neonates with a GA at birth $34 weeks, a starting NVP dose of 150 mg/m2 once daily for the first 2 weeks of life and twice daily thereafter achieves therapeutic levels in most at weeks 1, 2, and 4 of life. For those with a GA ,34 weeks at birth, our data suggest that a lower starting dose is warranted. If our pharmacokinetic and safety data are replicated by others, we believe that routine TDM would not be required for infants with a GA at birth $34 weeks; this could in turn allow for the application of NVP-based cART in resource-limited settings.