Plasma levels of hydroxychloroquine and seroconversion in health care workers during COVID prophylaxis: Retrospective evaluation using pharmacokinetic simulation

INTRODUCTION

With the abrupt emergence of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic in December 2019, drugs such as chloroquine (CQ) and hydroxychloroquine (HCQ) were repurposed urgently for prophylactic and therapeutic use in the absence of appropriate drug therapy. Despite inadequate pharmacokinetic/pharmacodynamic (PK/PD) studies, several dosage regimens were put into use for the prophylactic and therapeutic measures. Although both CQ and HCQ were propagated, preference was given to HCQ considering its safety over CQ. Nevertheless, several potential drug-drug interactions have been reported with HCQ for potential drug-induced toxicities.^[1]

Several clinical trials were initiated to evaluate HCQ efficacy and safety in patients with coronavirus disease-19 (COVID-19) throughout the world. Adequate literature is available for its use in malaria and other inflammatory situations like rheumatoid arthritis.^[2] However, no rationale is reported supporting the dosage and frequency of administration of HCQ to reach adequate levels to prevent the virus from entering into the lungs (prophylactic) and decrease viral load (therapeutic). Drug interaction of HCQ with co-administered drugs, cytochrome P450 (CYP) polymorphism,^[3] low body mass index,^[4] pre-existing cardiac problems, decreased renal clearance,^[5] dehydration associated with increased risk of cardiotoxicity^[6,7] and ocular toxicity^[8] are the key determinants of its plasma levels. Besides, the clinical efficacy of HCQ either as prophylactic or therapeutic in COVID-19 has been explained as a balance between inhibition of viral replication, immunosuppression and off-target side effects.^[9] In’t Ved et al. (2021) suggested clinical effect of prophylactic or therapeutic HCQ treatment in COVID-19 is probably dependent on the disease stage and severity.^[9]

The blood levels of the drugs such as CQ and HCQ are highly variable. Moreover, inter-individual variation accounts for their toxicity in a fixed dosage. These drugs are known to have a longer half-life and can be detected in the blood for up to 5 months after stopping the treatment. HCQ was recommended as prophylaxis for COVID-19 by various national and international agencies. On the recommendations of agencies, healthcare workers (HCWs) consumed HCQ prophylactically during the COVID-19. Interestingly, despite following the prophylactic dose of HCQ, many healthcare workers suffered from COVID-19 with few morbidities during the initial wave. Of note, till that time no credible data was available regarding the therapeutic drug levels of HCQ in our population for the extrapolation of its safety profile. Thus, rationalisation of the prophylactic dose of HCQ with respect to its inter-individual variation is required.

Therefore, this study was designed to monitor the HCQ blood levels in HCWs undergoing prophylactic therapy (Indian Council of Medical Research [ICMR] regimen) involved in COVID-19. Further, we aimed to assess the safe plasma levels of HCQ using PK simulation-based pharmacokinetic estimates for rationalising the dose of HCQ, based on the individual blood level for the COVID-19 prophylaxis.

MATERIALS AND METHODS

RESULTS

Simulation of HCQ pharmacokinetics in HWs

The pharmacokinetic simulations were performed using the 400 mg of HCQ dosing for 24 h for Tett et al., (1983);^[16] Fan et al., (2015);^[11] McChenskey et al., (1983)^[17] and Lim et al., (2009)^[18] [Figure 1]. Lim et al., (2009)^[18] reported 7–8 times higher HCQ concentration among all reports. The Tett et al., 1983^[16] data was utilised to derive the pharmacokinetic parameters for further simulations [Figure 2].

Close

Figure 1:

The plasma hydroxychloroquine levels after 400 mg oral dose reported in volunteers by Tett et al. (1989), Fan et al. (2015), McChesney et al. (1983) and Lim et al. (2008). The pharmacokinetic parameters for this study were derived from Tett et al. (1989).

Export to PPT

Close

Figure 2:

(a) The graph showing cumulative simulation of the 5% variance (n = 100 subjects) for the hydroxychloroquine (HCQ) plasma levels 200 mg dose. The lower and upper margins of the curve represent the 2.5^th and 97.5^th percentile. The middle green line represents the mean of the simulation. The dark dotted line shows the plasma levels of HCQ 200 mg from healthy volunteers from Tett et al. (1989). (b) The graph shows individual response simulation of the 5% variance (n = 100 subjects) for the HCQ plasma levels 200 mg dose. The dark dotted line shows the plasma levels of HCQ 200 mg from healthy volunteers from Tett et al. (1989). Simulation was performed in R-Package using Modviz Pop package.

Export to PPT

PK simulation of HCQ in LUNG

The pharmacokinetic simulations were performed using the^[14] variable dosing of the reported regimen in plasma [Figure 3]. The PK profiling showed highly variable predictions due to the lack of controlled data regarding the tissue concentrations obtained from recent publications. Moreover, due to the lack of pharmacodynamic data (as the SARS-CoV-2 infection model is not feasible), predictions for PK/PD are not being made. However, a lack of consistency in the pharmacokinetics and the levels attained was observed in the peripheral compartments [Figure 4] to achieve the prophylactic use.

Close

Figure 3:

The simulation of predicted plasma concentration for different regimens of hydroxychloroquine prophylactic dosing from various regimen. (a) Al-Kofahi et al. 2020 (regimen-1), (b) Rajasingam et al. (regimen-2), (c) Fan H et al. (regimen-3) and (d) Indian Council of Medical Research prophylaxis for coronavirus disease-19 (regimen-4). Simulation was performed in ‘R’ using Modviz Pop package.

Export to PPT

Close

Figure 4:

The graphs representing the simulation of predicted tissue concentration different regimens of prophylactic dosing of hydroxychloroquine tissue conc. (a) Al-Kofahi et al. 2020 (regimen-1), (b) Rajasingam et al. (regimen-2), (c) Rajasingam et al. (regimen-3) and (d) Indian Council of Medical Research prophylaxis for coronavirus disease-19 (regimen-4). Simulation was performed in R-Package using Modviz Pop package.

Export to PPT

PK simulation of HCQ in epithelial lining fluid

From the mean steady-state levels of HCQ from the simulation studies using ICMR prophylactic regimen, extrapolations were made for the evaluation of target organ levels of HCQ. For the extrapolation of HCQ levels in epithelial lining fluid, a factor of 21 was used based on the observations of Ruiz et al. (2021).^[19] For the extrapolation of lung tissue levels, the plasma steady-state HCQ levels were extrapolated using the factor 105 (median) derived from cynomolgus macaques reported by Maisonnasse et al. (2020)^[20] [Figure 5].

Close

Figure 5:

The graphs represent the simulation of Indian Council of Medical Research prophylaxis regimen for 4 months subjects. The round dotted lines represent the free drug levels, whereas the triangle depicting the lung levels derived from the Maisonnasse et al. A simulation was performed in R-Package using Modviz Pop package.

Export to PPT

Plasma levels of HCQ in HCWs

Plasma levels of HCQ in 246 HCWs were analysed using LC-MS/MS. Figure 5 shows the scatter plot of serum concentration versus the time of HCQ concentration. All the samples analysed for HCQ in HCWs were below the concentration of 300 ng/mL [Figure 6]. In addition, no HCWs reported any severe cardiotoxicity in this study. The observed and predicted HCQ plasma showed poor (r = 0.1147) as predicted through Pearson’s correlation [Figure 7].

Close

Figure 6:

The graphs represent the simulation of Indian Council of Medical Research (ICMR) prophylaxis regimen for 4 months (n = 100 subjects). Each dot represents the individual hydroxychloroquine levels of healthcare workers obtained at various time pointss following the ICMR prophylaxis regimen. A simulation was performed in R-Package using Modviz Pop package.

Export to PPT

Close

Figure 7:

The graphs represent the observed and predicted hydroxychloroquine plasma concentration ratio for 4 months of Indian Council of Medical Research (ICMR) proposed ICMR prophylactic treatment. The observed and predicted concentration had poor (r = 0.1147) as predicted through Pearson’s correlation.

Export to PPT

Quantitative assessment and seroconversion in HCWs prophylaxis

Table 2 represents the overall side effects reported by HCWs after taking HCQ. Majorly headache and dizziness were reported to account for 4.63 and 3.24%, followed by mild tachycardia (2.78%), diarrhoea (2.78%), abdominal pain (2.78%), gastritis (2.78%), nausea (2.31%) and acidity (1.39%). The drug levels of HWs were falling within the range of the pharmacokinetic profile^[17] [Figure 8]. The sero-index of all participating HCWs was compared in all three groups. The percentage of seroconversion (%) in Group-1 (No-HCQ), Group-2 (HCQ) and Group-3 (HCQ quantified) were found to be 13.1, 12.9 and 11 %, respectively, as shown in Figure 9. Although no correlation was found between sero-reactivity with HCQ concentration, seropositivity had a negative correlation with HCQ concentration in seropositive HCWs (r = 0.685).

Table 2: Side effects reported by health care workers after prophylactic use consumption of hydroxychloroquine.

Side effects experienced	Number (n)	Percentage
Dizziness	7	3.24
Nausea	5	2.31
Vomiting	1	0.46
Diarrhoea	6	2.78
Fatigue	1	0.46
Chest pain	2	0.93
Mild palpitation	6	2.78
Abdominal pain	6	2.78
Acidity	3	1.39
Gastritis	6	2.78
Anorexia	1	0.46
Join pain	1	0.46
Body pain	1	0.46
Low blood sugar	1	0.46
Heavy menstrual cycle	1	0.46
Mood changes	2	0.93
Skin rash with itching over neck, chest and arms	1	0.46
Sore throat	1	0.46
Dry mouth	2	0.93
Blurring vision	1	0.46
Tinnitus	1	0.46
Headache	10	4.63
Weakness	2	0.93
Insomnia	2	0.93
Drowsiness	2	0.93
No symptoms	115	53.24
Not reported	42	19.44

Close

Figure 8:

The line graph with blank dots represents hydroxychloroquine plasma concentration as reported by McChesney et al. (1983) for healthy volunteers. The black dots represent the individual data point for hydroxychloroquine plasma concentration of healthy volunteers from the present study.

Export to PPT

Close

Figure 9:

(a-c) Data showing the frequency of health care workers (HCW) with their seroconversion reactivity (1 or >1) and no sero-reactivity (<1) in all three groups. Of note, seropositivity was associated with the low hydroxychloroquine (HCQ) concentration in group 3. (d) The observed and predicted concentration had poor (r = 0.6857) as predicted through Pearson’s correlation.

Export to PPT

DISCUSSION

HCQ has been reported to exhibit multiple potential antiviral mechanisms in various pathogens (e.g. Chikungunya, Dengue virus, human immunodeficiency virus, poliovirus and Zika virus). In vitro studies on Vero cells have shown that HCQ effectively inhibited the SARS-CoV-2 by affecting viral entry and transport through endolysosomes. Its mechanism of action has created adequate assurance for using HCQ as a prophylactic and therapeutic agent in the initial phases of COVID-19.

During the pandemic of COVID-19, HCQ has been found to be effective as a prophylactic as well as a therapeutic option against SARS-COV-2 and has been recommended by several agencies worldwide. Therefore, the use of different prophylactic regimes was reported in different parts of the world. In India, ICMR, New Delhi, proposed a prophylaxis regimen of 400 mg BD followed by 400 mg weekly for Hospital Care Workers (ICMR-bulletin, 2020). Therefore, this study evaluated the HCQ levels in HCWs to understand the possibility of plasma level correlation with efficacy and adverse effects.

To find the dose for HCQ prophylaxis for COVID-19, AlKofahi et al. (2020)^[13] simulated pharmacokinetics to predict the levels above EC₅₀ for SARS CoV-2 in 1000 subjects for post-exposure and prophylaxis. As per their prediction, HCQ steady state above EC₅₀ was achieved by pre-exposure prophylaxis with the loading dose of 800 mg followed by 400 mg twice or 3 times weekly. In another exposure-driven, post-exposure prophylaxis regimen, an 800 mg loading dose followed by 6 hourly doses of 600 mg, then 600 mg daily for 4 more days could achieve daily troughs above EC₅₀ in >50% of subjects.

However, close to the ICMR prophylaxis regimen, only Fan et al.^[11] and Tett et al.^[16] reported HCQ pharmacokinetics (PK) with 200 mg HCQ, while McChenskey et al., (1983)^[17] reported PK of 400 mg (320 mg base HCQ). Although Lim et al. (2009)^[18] used 400 mg pharmacokinetic values in HWs, it showed 10 times higher AUC when compared with the pharmacokinetic profile reported for the same dose in other reports such as Tett et al.,^[16] McChesney et al., (1983)^[17] and Fan et al., (2015)^[11] [Figure 1]. While, McChesney et al., (1983)^[17] data was best suited for simulation studies to predict population kinetics, it was not used due to the deployment of a less specific technique (spectro-fluoroscence) for quantification of HCQ. Therefore, for this study, Tett et al., (1989)^[16] pharmacokinetic data was used for making pharmacokinetic estimates and simulation of multi-dose kinetic.

Yao et al.^[21] reported a prophylactic in vitro study by pre-exposing the cells with HCQ for 2 h followed by removal of the drug and exposing cells to SARS-CoV-2, and reported EC₅₀ values as 6.25 mM (2.1 mg/mL) and 5.85 mM (1.95 mg/mL) at 24 and 48 h, respectively. Whereas treatment model (drug exposure after infection), EC₅₀ values of HCQ were found to be 6.14 mM (2.0 mg/mL) at 24 h and 0.72 mM (0.25 mg/mL) at 48 h. In a similar model, Liu et al. (2020)^[22] reported EC₅₀ values of HCQ 48 h post-infection at a multiplicity of infection (MOI) of 0.01 as 4.51 mM (1.5 mg/mL). Similarly, Maisonnasse et al. (2020)^[20] reported inhibitory concentration (IC₅₀) in the in vitro activity of HCQ against SARS-CoV-2 in VeroE6 cells (MOI: 0.01) and a model of reconstituted human airway epithelium (MOI: 0.1). These values were found to be 2.19 mM (0.7mg/mL) and 4.39 mM (1.5 mg/mL) for 48 and 72 h, respectively. From all these studies, it is understood that an effective free drug concentration of 1–2 mg/mL is required for prophylactic as well as therapeutic to reach 50% efficacy. However, it is unclear whether the in vitro activity recorded for HCQ by the investigators is due to the concentration build-up at the plasma membrane (unbound drug) or lysosomal trapping (lysosomotropic) inside the cell (Tissue bound). Moreover, our simulation study has revealed that the prophylaxis regimen followed by HCWs could lead to effective free drug concentration. Our data shows that the ratio of plasma to lung (tissue levels) as well as to epithelial lining fluid levels (free drug levels) does not correlate with the benefit seen in clinical studies.

The findings of Rosenke et al. (2020)^[23] showed that prophylactic use of HCQ from a standard human antimalarial dose of 6.5 mg/kg to higher 50 mg/kg failed to stop SARSCoV-2 replication/shedding in the Syrian hamster disease model. Even in this model, only a high dose of 50 mg/kg mimicked (Hamster to human equivalent) of 6.75 mg/kg as this investigator failed to adopt metabolic conversion factors while conducting in vivo studies. At the dose of 6.5 mg/kg, HCQ (2 mg/kg of human dose) did not reduce SARS-CoV-2 replication/shedding in the upper and lower respiratory tract in the rhesus macaque disease model by Rosenke et al. (2020).^[23] As aforementioned, studies used the dose equivalent or less than the recommended human dose; therefore, not much of an inference can be drawn. In the prophylactic as well as therapeutic SARS-Cov-2-infected cynomolgus macaque animal model, Maisonnasse et al. (2020)^[20] showed that in spite of high HCQ concentration in the blood and lungs with the plasma exposure that was similar to those observed in COVID patients, HCQ did not show any benefit in vivo.

Our study compared the sero-index of all participating HCWs in all three groups. The percentage of seroconversion in no prophylactic group (Group-1), Group-2 (HCQ) and Group-3 (HCQ quantified) were found to be 13.1, 12.9 and 11%, respectively. Moreover, no correlation was found between seropositivity with HCQ concentration to explain the prophylactic role. Our data coincides with the observations of Goenka et al. (2020),^[24] where they have studied a 25% cross-section of 1122 HCW and reported a seroprevalence rate of 11.94%. They have attributed it to the BCG vaccination, HCQ prophylaxis and the job profile influencing the seroprevalence rate in HCW.

Other studies have also documented the prophylactic failure in HCQ among HCW and patients undergoing therapy with HCQ for auto-immunity-related conditions^[25-27] A meta-analysis of blinded placebo-controlled randomised clinical trials including 2668 patients to analyse the efficacy and safety of HCQ reported no clinical benefit of HCQ as pre- and post-exposure prophylaxis.^[28] Subsequent reports also supported the lack of efficacy in the HCQ prophylaxis. Nevertheless, it had no effect on hospital admission and mortality. In contrast, it probably increases the adverse effects and does not reduce the risk of SARS-CoV-2 infection.^[29]

Adverse drug reaction monitoring report on the prophylactic use of HCQ has been documented^[30,31] In a cross-sectional safety study reported from India for the HCQ prophylaxis in HCW, it was observed that the first dose of HCQ was well tolerated as evidenced by less discontinuation of the therapy.^[32] In the present study, HCWs self-reported mild tolerable side effects of HCQ prophylaxis which did not correlate with HCW plasma levels [Table 2]. In our study, after the first dose, an erratic response was observed in HCQ plasma levels as evidenced by a lack of any appreciable correlation in the observed versus predicted levels. This observed trend could be due to poor HCWs compliance with HCQ prophylactic regimen due to its side effects.

CONCLUSION

To conclude, this study compared the serum levels of HCQ used as a prophylaxis in HCWs with seroconversion and side effects. We further correlated the serum levels with the simulated population kinetics using appropriate pharmacokinetic estimates. This study showed that HCQ prophylaxis was expected to reach adequate levels to exhibit protection; however, lack of efficacy in the clinical settings adds to the existing paradox of in vivo-in vitro correlation.