Nevirapine hair and plasma concentrations and HIV-1 viral suppression among HIV infected ante-partum and post-partum women attended in a mother and child prevention program in Maputo city, Mozambique

Introduction Prevention of mother to child transmission of HIV (PMTCT) is frequently challenged by irregular access to more effective anti-retroviral therapy. Nevirapine single dose (sdNVP), sdNVP+AZT+3TC for MTCT prophylaxis and NVP+ AZT+3TC for treatment and PMTCT were withdrawn due to low genetic resistance barrier and low efficacy. However current PMTCT lines in Mozambique include DTG+3TC+TDF, TDF+3TC+EFV, DTG +ABC+3TC, and AZT + NVP syrup prophylaxis for exposed babies. We assessed NVP hair and plasma concentrations and association with HIV-1RNA suppression among HIV+ ante-partum and post-partum women under PMTCT in Maputo, Mozambique. Methods From December 2013 to November 2014, prospectively were enrolled 200 HIV+ ante-partum women on 200mg nevirapine and zidovudine 300 plus lamivudine 150mg twice daily at least with 3 months treatment and seen again at 24 weeks post-partum. Self-reported pill-taking adherence, NVP concentrations in hair, plasma, hemoglobin, CD4 cell count, HIV-1 RNA load was evaluated. NVP concentration in hair and plasma was analyzed as categorical quartile variable based on better data fit. NVP concentration was set between ≤3.77 ng/ml in plasma and ≤17,20 ng/mg in hair in quartile one to ≥5.36 ng/ml in plasma and ≥53.21 ng/mg in hair in quartile four. Logistic regression models for repeated measures were calculated. Following the World Health Organization (WHO) guidelines we set viral suppression at HIV-1RNA < 1000 c/mL. Outcome was HIV-1 RNA<1000 copies/ml. Predictor was NVP concentration in hair categorized in quartiles. Results In total 369 person-visits (median of 1.85) were recorded. Self-reported adherence was 98% (IQR 97–100%) at ante-partum. In 25% person visits, NVP concentrations were within therapeutic levels (3.77 ng/ml to 5.35 ng/ml) in plasma and (17.20 ng/mg to 53.20 ng/mg) in hair. In 50% person visits NVP concentrations were above 5.36 ng/ml in plasm and 53.21 ng/mg in hair. HIV-1 RNA suppression was found in 34.7% of women with two viral loads, one at enrollment and another in post-partum. Odds of HIV-1 RNA suppression in quartile 4, was about 6 times higher than in quartile 1 (p-value = 0.006) for NVP hair concentration and 7 times for NVP plasma concentration (p-value = 0.012). Conclusions The study results alert for potential low efficacy of current PMTCT drug regimens in use in Mozambique. Affordable means for individual monitoring adherence, ART plasma and hair levels, drug resistant and HIV-1 RNA levels monitoring are recommended for prompt identification of inadequate drug regimens exposure patterns and adjust accordingly.

a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 Introduction Prevention of mother to child transmission of HIV (PMTCT) is frequently challenged by irregular accessibility to more effective anti-retroviral therapy (ART) [1,2] and limited routine monitoring of HIV-1 RNA [3,4] particularly in HIV/AIDS most affected and low resourced sub-Saharan African countries like Mozambique [5,6] Timely identification of HIV-1 viral unsuppressed ante-partum and post-partum women and implementation of suitable measures is key for prevention of MTCT of HIV infection [7].
The HIV prevalence in Mozambique is one of the highest, including among ante-natal attendees with an estimated of 15.8% in 2019 [8] Nevertheless there are over 90% of HIV positive pregnant women under ART for PMTCT in the country, but appropriate HIV-1RNA load monitoring was available to less than 50% [9] and HIV-1 RNA suppression was estimated at 45% at 12 months post-partum with decreasing tendency thereafter [10]. Similarly, the rate of mother to child transmission of HIV (MTCT) increased from 11% in 2014 to 14% in 2019 [11,12]. Pregnant and post-partum women should continuously access effective ART and routine monitoring of HIV-1 RNA load to prevent reduce MTCT of HIV infection [7].
In 2011, based on WHO guidance for PMTCT outcomes improvement [2,18] Mozambique moved from starting ART at 28 weeks to 14 weeks of pregnancy or sooner, with NVP either Sd NVP intra-partum combined to AZT + 3TC for mothers, option A, or NVP as a composite of maternal triple-drug provided throughout ante-partum and breastfeeding the option B, as well as NVP or AZT syrup for exposed infants until 4-6 weeks post-partum [18].
Adequate treatment for HIV positive mothers to achieve HIV1 RNA suppression is key for attaining 2030 UNAIDS PMTCT targets [9] Nevertheless, similar to other settings, in Mozambique, monitoring of ART exposure and early detection of HIV1-RNA unsuppressed women on MTCT program is still a challenge [1,4]. Therapeutic drug monitoring (TDM) is the gold standard that shows the drug concentrations in plasm [3,4,[25][26][27] But its usefulness is limited to some hours to few days and there is wide intra-individual variability [25]. Additionally, the TDM in plasma is relatively expensive [28] as it requires specialized technology for sampling and testing, hampering routine use in low-income countries [25][26][27].
In contrast, the evaluation of ART concentration in hair is relatively affordable [2,3]. Results have shown hair concentration to be a strong independent predictor of HIV-1 RNA load suppression [4,27,29]. The ART concentrations in hair reflects trends of historic medication taken by a patient over a long period from days up to months [29][30][31][32][33]. This study aimed to assess self-reported adherence, measurements of NVP in hair and plasma and effects on HIV viral suppression among ante-partum and post-partum women under treatment with NVP+ AZT+ 3TC provided at routine PMTCT in three urban primary health centers in Maputo city, Mozambique. We hypothesized that ante-partum and post-partum women would have low self-reported adherence, and as consequence a low NVP levels in hair and plasma and low HIV-1RNA suppression. We also considered that NVP levels would enable predictions of HIV-1RNA suppression.

Material and methods
An observational prospective study was conducted from December 2013 to November 2014 in three primary public health facilities (Primeiro de Maio Health Center, Primeiro de Junho Health Center and Mavalane Health Center) located in the Mavalane health area in Maputo City, Mozambique. The three health facilities attend 60% of ante-partum and post-partum women on ART with estimated HIV prevalence among ante-partum women of 24% in 2018 [23]. In 2019, 90.000 HIV positive women became pregnant more than 60% were under TDF + 3TC+ EFV and 10% of ante-partum and post-partum women were still on NVP based ART regimens. The remaining were under DTG +3TC+ ABC and other alternative ART lines.

Sample size
Conveniently and prospectively, we selected 200 HIV positive ante-partum women from the three ANC/PMTCT services at the above-mentioned healthcare facilities. Ante-partum women who met the selections criteria were enrolled after they provided informed consent.
All HIV positive ante-partum women aged 18 years or older on at least three months on ART with AZT+3TC+NVP regardless of the gestational age period and willing to be seen again at 24 weeks post-partum were eligible for the study. All women arriving for antenatal visit were invited to participate in the study and consenting women were enrolled consecutively until sample size being achieved for the study. All women were seen at two time points: ante-partum and six months after delivery.
A structured questionnaire was used to collect socio-demographic and self-reported adherence data. We used four days adherence tool recommended by the AIDS Clinical Trials Group (ACTG) to assess ART regimen and number of pills taken modified by the authors to include plasm and hair NVP concentrations levels [34] Hair and blood samples were collected to measure NVP concentrations in both study time periods in enrolled women.

Hair sample collection and measurement of nevirapine concentration
Approximately 10 to 20 strands (around 1-3 mg) of hair were obtained from occipital region from each woman [27,29]. The hair samples were individually labeled and assembled in plastic bags and stored at the room temperature at the health facility before being sent to the National Institute of Health (INS) laboratory located in Maputo city, Mozambique. Hair samples were sent for analysis to the laboratory at the University of California San Francisco (UCSF) in United States of America (USA). The therapeutic drug monitoring (TDM) was performed by using liquid chromatography/tandem mass spectrometry (LC/MS/MS) [3,27,29,35]. In brief, NVP was extracted in methanol/trifluoroacetic acid in a proportion of 9:1 and shaken at 37˚C in a water bath overnight and submitted to liquid-liquid extraction in alkaline conditions [29,35]. The lower limit of quantification of NVP was 0.5 ng/mg (LLOQ). NVP hair therapeutic levels are ranged from 3.77 ng/mg to 5.35 ng/mg.

Blood sampling and measurement of nevirapine concentration
Five ml of whole blood was collected in EDTA tubes to measure NVP concentration in woman during ante-partum and at post-partum period. The collected blood samples were centrifuged at 3,000 rpm for 3 minutes within six hours after collection at the INS laboratory. Plasma samples were stored in cryovials at the -20˚C until final shipment to the University of Stellenbosch in South Africa for analysis. NVP concentration in plasma were measured through validated high performance liquid chromatography (HPLC) [36] NVP lower limit of quantification (LLOQ) was 25 ng/ml. NVP blood therapeutic levels was set between 3,000-8,000 ng/mL.

HIV-1 viral load, CD4+ T cells and full blood count
An additional 5ml of the whole blood sample was collected in EDTA tube and sent to INS laboratory within 6 hours after collection for HIV-1 RNA viral load using COBAS Amp li Prep /COBAS Taq Man HBV test, v2.0 (Roche Diagnostics, Germany), a fully-automated system that employs real-time PCR technology with a limit of detection of �50 copies/ml. T-lymphocyte (CD4 cells) quantification was performed by using FACS Caliber flow cytometer (Becton Dickinson, USA) using the Multistep software (Becton Dickinson, USA) within 24 hours after sample collection at INS laboratory. Full blood count including three-part differential of hematological parameters was performed within 6 hours of blood collection in Vacutainer 1 tubes with K2 (EDTA) anticoagulant (Becton Dickinson, USA) using the hematology analyzer Sysmex KX21N (Sysmex Corporation, Japan).

Statistical analysis
The primary outcome was HIV-1 RNA load<1000 copies/mL following the World Health Organization (WHO) guidelines [13,15]. And primary predictors were NVP levels in hair and plasma at each study visit. Summary statistics were computed for a set of baseline characteristics. A log-transformed viral load was calculated to compare the viral loads between visits using a Wilcoxon signed-rank test. Mixed-effects logistic regression models for repeated measures were used to estimate and compare the association of NVP concentration levels in hair, plasma and HIV1-RNA. Following previous model [27]. NVP levels in hair and plasma were analyzed as categorical variable and quartiles. The best estimation of categorization was based on data fit and NVP therapeutic levels in hair and plasma. The model was adjusted by age groups and number of visits. We were unable to adjust for time of ART initiation due to unclear dates. All analyses were performed using SAS/STAT software, version 9.4.
Ethical Approval: Mozambican National Bioethics Committee approved the study protocol (IRB 357/CNBS/12). The Ministry of Health, Maputo City Directorate also provided administrative clearance. Written and oral informed consent was obtained from each enrolled woman.

Socio-demographic characteristics of pregnant and post-partum women
The " Table 1" presents the demographic characteristics of the study participants, as well as NVP concentration in hair and plasma for the 200 ante-and post-partum women enrolled in this study. All women were under AZT+3TC+NVP as part of their own health and PMTCT. Each patient contributed on average with 1.85 visits to the analysis (range: 1-2) for a total of 369 person-visits.
Of the 200 women, 86.5% were younger than 35 years, 27.5% had either secondary or higher education and 10.5% were employed at the time of study roll out. Median hemoglobin was 10.5 g/dl (IQR: 9.6-11.4), median CD4 count 377. As shown in " Table 2" in 44% of person visits women had viral load counts equal or above 1000 copies/ml. NVP concentration varied from � 3.77 ng/ml in plasm and � 17,20 ng/mg in hair in quartile1, to � 6.98 ng/ml in plasma and � 106.89 ng/mg in hair in quartile 4. However 25% of women in quartile 1 had NVP levels below therapeutic threshold of (� 3.77 ng/ml) in plasm and (� 17.20 ng/ml) in hair.
In approximately 25% and 50% of person visits, women had NVP concentration in hair and in plasma within and higher therapeutic levels respectively.
" Fig 1" displays the percent of person-visits in each quartiles of NVP concentration level in hair and plasma related to HIV suppression (HIV-1 viral load <1,000 copies/ml). There was a strong association between NVP concentration levels in the hair with HIV-1 viral load < 1,000 copies/mL (p <0.01) than NVP concentration levels in the plasma (p = 0.030).

Median and interquartile range (IQR) of nevirapine concentration levels in the plasma and hair by viral load count at visit 1 and 2
The median nevirapine concentration levels were slightly high for women with viral load count <1000 cp/ml for both plasma and hair measurements at any visit. The overall median nevirapine concentration in the plasma for women with HIV-1RNA load count <1000 cp/ml was 5ng/ml (IQR: 4-85 ng/ml) and 6 ng/ml (IQR: 2-75 ng/ml) for HIV-1RNA load count >1000 cp/ml was 3ng/ml (0-6ng/ml) ante-partum and 6ng/ml(2-7n/ml) no post-partum. The

PLOS ONE
Nevirapine hair and plasma concentrations and HIV-1 viral suppression among ante-partum and post-partum women

Association of NVP in hair and plasma concentration with HIV-1 viral load suppression
" Table 3" presents the results of the mixed-effects model for HIV-1 viral load < 1,000 copies/ mL, adjusted for age and visit number. We could not adjust for time of ART initiation due to mostly unclear date for ART initiation.
For a random patient, increased NVP concentration levels either in the hair or in plasma were significantly associated with a higher odds ratio for HIV-1RNA load < 1,000 copies/ml. For person-visits with NVP hair concentration in quartile 4, the odds of HIV-1 viral

PLOS ONE
Nevirapine hair and plasma concentrations and HIV-1 viral suppression among ante-partum and post-partum women load < 1,000 copies/ml were about 6 times higher than that of NVP hair concentration in quartile 1 (p-value = 0.006). Similarly, for NVP plasma concentration in quartile 4, the odds of HIV-1RNA load < 1,000 copies/ML was about 7 times higher than that of NVP blood concentration in quartile 1 (p-value = 0.012).
The " Fig 3" also display plots of adjusted odds ratio with 95% confidence interval for the association of NVP hair and blood concentration with HIV-1RNA viral load < 1000 copies/ mL (for all person-visits). The odds of HIV-1RNA load < 1,000 copies/mL was significantly lower for visit number two when compared with visit number one. Age wasn't a significant factor in the virology outcome.

Discussion
Effective ART exposure is critical for virology suppression [7,27,33] and PMTCT. In 2019, DTG+3TC+TDF was introduced and alternative line included TDF+3TC+EFV [24]. The AZT and NVP syrup were still used for prophylaxis among HIV exposed children [24]. In the same year the PMTCT coverage in Mozambique was over 90% among HIV positive pregnant women. However, HIV-1RNA suppression was 45%, MTCT was 14%, a very high rate when compared to the neighbouring countries [8,[37][38][39] Maternal HIV-1RNA load at delivery is a high predictor of MTCT of HIV infection [40,41]. The low rates of HIV1-RNA suppression and high MTCT are suggestive of the need of affordable HIV-1RNA /TDM monitoring and DRM testing for personalized clinic decisions particularly for childbearing women.
We carried out a TDM and HIV-1RNA load trends. We hypothesized, ante-partum and post-partum women would have low self-reported adherence, low NVP concentrations in plasma and in hair, low HIV-1RNA suppression with consequent increased risk for MTCT of HIV infection. Differently to our hypothesis we found a short term high self-reported adherence, but 25% of women have low NVP level in plasma and in hair, these findings support a long term low adherence at least among these women as hair concentration gives a long period up to months of drug intake behaviour. Also, among other women we cannot roll out the Table 3. Odds ratio (95% confidence interval) for the association of NVP hair and blood concentration with viral load < 1000 copies/mL (for all person-visits).

Hair Blood
Age groups (years)

PLOS ONE
Nevirapine hair and plasma concentrations and HIV-1 viral suppression among ante-partum and post-partum women possibility of instable adherence throughout ante and post-partum period particularly, worsening at post-partum. The possibility of resistant mutations also is present However50% person visits NVP levels were above 5.36 ng/ml in plasm and 53.21 ng/mg in hair and in 25% person-visits NVP concentration were within therapeutic ranges, but HIV-1RNA suppression occurred only in 17 (34.7%) of women with two measurements. Consequently, failing to show adequate NVP hair concentration as an independent predictor of HIV-1 RNA suppression. Perhaps, the failure to demonstrate our predictions may be due to small sample size and convenient sampling. Nevertheless, our findings were consistent in demonstrating an increase in NVP plasma and hair levels positivity in random patients associated with increased odds of HIV-1 RNA suppression. The odds of HIV-1 RNA suppression in quartile four of NVP levels in hair was over six-time higher than in quartile 1; 6.16 (95% 1.73-21.99; p-value = 0.006) [23,29,43]. Although the odds of HIV-1 RNA suppression of NVP in plasma was higher, 7.15 (1.57-32.60; p-value = 0.012), the confidence interval was wider, favouring comparatively weaker association. Also, it is important to acknowledge that NVP plasma concentration levels provide very short information of drug availability in the blood stream, from several minutes to some days and is prone to the bias of time gap between drug uptake and blood withdraw [29,42] Whereas hair concentration represents a lengthier period of drug exposure in the most body fluids [25,27] Furthermore, among the failure group, the majority (22 or 66.6%) of women had HIV-1 RNA at post-partum significantly higher than at enrolment (p<0.0001) and at any visit the median nevirapine levels were slightly low in the failure group (HIV-1RNA count >1000 cp/ ml) for both plasma (6 ng/ml (IQR: 2-75 ng/ml) and hair (19 ng/mg (IQR: 0-73 ng/mg) and median NVP concentration levels differed significantly between HIV-1RNA load count groups only for hair measurements at postpartum visit (p<0.001). Earlier identification of HIV-1RNA and therapeutic status would have enabled early detection of HIV-1RNA failure in at least 66.6% women and better guided drug switch on 31% to recover from probable HIV-1RNA failure and therefore, reducing the risk of MTCT of HIV infection.
Technical constraints limited the possibilities for assessing other current PMTCT ART components such as EFV, AZT and 3TC, TDF levels including the possible impact of NVP/ EFV pharmacogenetics on HIV-1RNA drug exposure [43,44]. However, we acknowledge HIV-1RNA low suppression, may be also related to AZT monotherapy intake in (sdNVP +AZT+3TC) PMTCT line and consequently HIV-1 resistant mutations selection. Additionally pre-ART drug-resistance to multiple antiretroviral classes like NVP +AZT+3TC is reported to increase the risk of virological failure particularly to EFV [45]. Also low HIV-1RNA failure has been associated to NVP/EFV perhaps due to HIV-1 cross resistance mutations and also related to AZT/3TC and 3TC [45] Additionally, intermittent and discontinued exposure related to sdNVP+AZT+3TC intake has been appointed as important contributor to HIV-1RNA low suppression particularly among HIV + women with high natality rates and short pregnancy spacing as is the case for most Mozambicans of childbearing age [46,47].
Recent surveys among pregnant women in south rural Mozambique HIV-1 drug resistant mutation (DRM) was around 10% mainly due to NNNRTI and NRTI including NVP, EFV, 3TC, ABC, TDF [48]. In a survey of general population that included childbearing women as participants in south urban and central rural Mozambique [48] 83% were on tenofovir (TDF)/ lamivudine (3TC)/efavirenzes (EFV) and the pre-treatment drug resistance PDR (NNRTI) varied from 16.8% to 31.2% in south urban and central rural setting, respectively and mainly due to EFV/NVP drug resistance, mostly among pre-exposed. Acquired drug resistance (ADR) varied from 8.3% to 15.5% and the majority had NNRTI-NRTI dual resistance. The most common NRTI (ADR) was to 3TC/Emtricitabine and Abacavir [48]. These results support the probability of existence of drug resistance for the most first line drugs currently used for PMTCT reinforcing the need for individual HIV1-RNA and drugs resistance status assessment.
The CYP2B6 NPSs rates in Mozambique and others countries support the likelihood of influence of CYP2B6 SNPs on NVP/EFV metabolism and HIV-1RNA response among our study population [27,43,52]. Nevertheless, this hypothesis needs further investigations.
The study had some limitations that included a small sample size and thus may not reflect ART behaviours and virology outcomes of all pregnant and postpartum women across the country as a whole. The convenient sampling of study sites and selection of women, the issue of lost to follow-up (LTFU), and some missing information also impacted negatively on study results and do not allow appropriate extrapolations for a larger population of the country.

Conclusion
With the results of this study, we can conclude that for a random participant NVP in hair NVP predicted HIV-1RNA suppression. High NVP levels in hair and plasma, and low HIV-1 RNA suppression, such as what was documented in this study likely suggests emerging HIV-1 RNA NVP resistant mutations and virology and therapeutic failure and possible effect of CYP2B6 NPSs. These findings alert for potential low efficacy of current and future NVP/EFV based ART treatment that are still being used in childbearing women and infants in Mozambique. Though cannot be generalized in overall results reinforce the need of individual HIV-1RNA monitoring, DRM and pharmacogenetic testing for prompt identification of erratic drug exposure patterns and adjust accordingly, to improve NVP/EFV based ART outcomes and prevent MTCT of HIV infection.
A larger study to determine the magnitude of hair NVP exposure, DRM, virology outcomes at PMTCT settings is needed because NVP/EFV, 3TC and AZT based regimes are still in use throughout the country. Whenever possible, HIV-1RNA, load measurement, assessment of DRM at ante-partum and post-partum period could be helpful to identify women at risk and adequate measures are taken to minimize MTCT of HIV infection. Pregnant women unique identifier number could uplift PMTCT efforts and improved follow-up.