Effectiveness of Pharmacokinetic-Guided Hydroxyurea Dose Individualization in Patients with Sickle Cell Anemia: A Mini-Review

Inconsistent therapeutic responses have been observed among patients with sickle cell anemia (SCA) undergoing hydroxyurea (HU) following the adoption of the standardized protocol. Moreover, this treatment regimen necessitates a prolonged period to reach the maximum tolerated dose in which beneficial therapeutic effects are observed in most SCA patients. To overcome this limitation, several studies have performed HU dose adjustments in SCA patients based on individualized pharmacokinetic profiles. The present systematic mini-review aims to select and analyze published data to present an overview of HU pharmacokinetics studies performed in SCA patients, as well as evaluate the effectiveness of the dose adjustment strategy. A systematic search was performed in the Embase, Pubmed, Scopus, Web of Science, Scielo, Google Scholar, and the Virtual Library of Health databases from December 2020 to August 2022, with a total of five studies included. Inclusion criteria consisted of studies in which the dose adjustment was performed in SCA patients based on pharmacokinetic parameters. Quality analyzes were performed using QAT, while data synthesis was performed according to the Cochrane Manual of Systematic Reviews of Interventions. Analysis of the selected studies revealed improved HU treatment effectiveness using personalized dosages in SCA patients. Moreover, several laboratory parameters were utilized as biomarkers of the HU response, and methods designed to simplify the adoption of this practice were presented. Despite the scarcity of studies on this topic, HU-personalized treatment based on individualized pharmacokinetic profiles represents a viable alternative for SCA patients who are candidates for HU therapy, especially for pediatric patients. Registration number: PROSPERO CRD42022344512.


Introduction
Sickle cell anemia (SCA) is an autosomal recessive genetic disease characterized by the inheritance of the hemoglobin S variant (HbS) from both parents. HbS results from the substitution of adenine for thymine in the sixth codon of the beta (β)-globin gene (HBB), leading to the replacement of glutamic acid with valine [1]. Due to the polymerization of HbS under recurring conditions of hypoxia, red blood cells thereby assume a sickle shape [2], and suffer two main events: (1) hemolysis, characterized by the rupture of red blood cells with a release of hemoglobin, reduction in the concentration of nitric oxide (NO), and increase in the inflammatory process and activation of endothelial cells [3], and (2) the vaso-occlusion, in which sickled red blood cells adhere to endothelial cells along with reticulocytes, platelets, and leukocytes, leading to obstruction of blood vessels [1]. These events are responsible for several clinical manifestations as acute chest syndrome, priapism, pain crises, malleolar ulcers, stroke, and pulmonary hypertension [2]. Hydroxyurea (HU), commercially known as HYDREA ® , is one of the drugs approved by the U. S. Food and Drug Administration to treat SCA patients with more severe clinical profiles [4]. Despite the broad indication and usage, the mechanism of action associated with HU has yet to be completely elucidated. Reports showed that HU inhibits ribonucleotide reductase, maintaining cells in the S phase of the cell cycle and preventing the continuation of cell division. Moreover, it activates soluble guanylate cyclase through the production of nitric oxide (NO), leading to an increase in the production of cyclic guanylate monophosphate, reactivation, and transcription of the gamma (γ)-globin gene (HBG), and, consequently, the production of fetal hemoglobin (HbF) [5]. Another point that needs to be fully elucidated is its pharmacokinetics. Reports have shown that HU is hydrophilic and has a volume of distribution close to that of water, with approximately 37% of the orally administered dose being excreted through urine without being metabolized [6]. Some authors defend the partial excretion of the drug by the kidneys and metabolism by the liver via enzymatic reduction catalyzed by the cytochrome P450 monooxygenases (CYPs) [4], which transform HU into urea and carbon dioxide [7], in addition to a portion which is converted into NO in both the liver and blood [8]. However, other studies have argued that HU is mostly excreted by non-renal mechanisms [9,10].
Currently, the standard therapeutic strategy is to start treatment with an initial fixed dose of 15 mg/kg/day, with increases of approximately 5 mg/kg/day every 4 to 8 weeks until reaching a dosage of 30 to 35 mg/kg/day, determined as the maximum tolerated dose (MTD) identifying the absolute neutrophil count of 1.5 to 3.0 × 10 9 /L, where evidence of the therapeutic effectiveness of HU has been noted, such as increases in the hemoglobin concentration and mean corpuscular volume (MCV), as well as the HbF level [11]. Despite the known beneficial effects of HU therapy in SCA patients, the need for dosage adjustments demands a higher number of medical visits, and, therefore, an elevated frequency of laboratory parameter monitoring. Furthermore, the variability in the therapeutic responses represents a significant challenge and may arise due to several factors, including variability in the pharmacokinetic profile of the patients.
Accordingly, several studies have sought to investigate the pharmacokinetic profile of HU in SCA patients, and based on an established profile, administer a personalized HU treatment regime for reducing variability in the therapeutic response [6]. The present review aims to select and analyze these studies to present an overview of HU pharmacokinetics in SCA patients, as well as analyze the effectiveness of personalized dosing.

Selection and Qualitative Analysis of Articles
Articles were selected following the protocol illustrated in the flowchart by the Preferred Reporting Items for Systematic Reviews and Meta-Analyzes (PRISMA) study ( Figure 1). Initially, 283 articles corresponding to the selected search terms were identified in different databases, of which 48 were duplicates. Two hundred and thirty articles were excluded based on title and abstract screening, as these were unrelated to the topic of the present review (Supplementary Material) Using the Quality Assessment Tool for Quantitative Studies (QAT) [12], we obtained the results of the qualitative analysis performed, analyzing aspects related to the methodology of the studies: selection bias, study design, and the data collection method. Dong

Main Findings from the Selected Articles
All included studies performed HU dose adjustments based on pharmacokinetic parameters in SCA patients and demonstrated treatment effectiveness using laboratory parameters. Table 1 presents the objectives and main results of each study.

Identification
Articles identified by searching in databases (n = 283) Additional articles found via other research sources (n = 0)

Selection
Articles left after removing duplicates (n = 235)

Eligibility Inclusion
Articles selected based on title and abstract (n = 5) Articles excluded (n = 0) Articles evaluated by reading the full text (n = 5) Articles excluded by reading the full text (n = 0) Studies included in the qualitative synthesis (n = 5) Figure 1. Flowchart detailing the present study protocol in accordance with the PRISMA model.

Main Findings from the Selected Articles
All included studies performed HU dose adjustments based on pharmacokinetic parameters in SCA patients and demonstrated treatment effectiveness using laboratory parameters. Table 1 presents the objectives and main results of each study.
Dong et al. conducted a study involving 96 SCA patients treated with 20 mg/kg of HU. The median age of the patients was 9 years of age [13]. These authors investigated individual pharmacokinetic profiles using a non-compartmental pharmacokinetic model. Data revealed an association between the AUC 0-∞ of 115 mg L −1 h −1 and the MTD of HU, suggesting that this value could be achieved using the MTD. In addition to the HU population pharmacokinetic model, an optimal sampling strategy for determining an individual's pharmacokinetic profile using only three sampling times was developed, with serum cystatin C and body weight identified as significant predictors of HU clearance.
The study by McGann et al. was conducted in children with SCA (n = 51) (age range: 6 months to 21 years of age) who were treated with 20 mg/kg of HU [14]. A pharmacokinetic-guided dose individualization strategy based on the population pharmacokinetic model permitted MTD achievements more efficiently, and a Bayesian adaptive control strategy was employed to measure the systemic exposure to HU via a spaced sampling schedule. Data demonstrated the influences of age and renal hyperfiltration on the pharmacokinetic profile of HU. Moreover, the time to reach MTD was found to be shorter, and dose determination, in accordance with pharmacokinetics produced better results, which was assessed via biomarkers including HbF and hemoglobin. The study hypothesis is that pharmacokinetic-guided dosing will improve laboratory responses compared to weight-based dosing, as the study outcome will be relevant for optimizing treatment in children with SCA. Recently, Quinn et al., published a study that is the continuation of that performed by McGann et al. They selected 48 children from the 51 included in the previous study who had F-cell analysis performed at least once. Early initiation of HU at the predicted MTD using pharmacokinetics-guided starting doses yielded promising results. Children showed sustained and near-pancellular or pancellular HbF distribution, indicating the effectiveness of this therapeutic strategy in meeting or exceeding the goals intended for SCA "curative treatment" [15].
Similar to the studies cited above, Meier et al., are planning to include a minimum of 104 pediatric SCA patients, with the age ranging between 6 months and 21 years of age [16]. These children will be divided into two groups: one will receive an initial dose of 20 mg/kg/day, while the second will receive an initial dosage guided by pharmacokinetics. In their study, the amount of blood to be collected will be optimized using a volumetric absorptive microsampling (VAMS) device to enhance the therapeutic response.
Another study performed by Paule et al. considered SCA patient data from two previous clinical studies. The first was a PK-PD study that analyzed blood samples from 81 adults aged 18 to 54 years old who received an initial HU dose of 20 mg/kg/day. The second was a study that analyzed the bioequivalence of a new formulation of coated tablets in 16 adult SCA patients [17]. Based on a population pharmacokinetic model, the authors observed variability in the pharmacokinetic parameters, such as the volume in the central compartment, clearance, absorption constant, and distribution constant.

Dose Adjustment Using the Population Pharmacokinetic Model
Considering the search terms chosen, the intervention employed in all selected articles was HU dose adjustment based on pharmacokinetic parameters. This conduct was prompted by the ineffectiveness of conventional HU treatment as was previously observed in some patients who received HU at an initially fixed dose followed by escalation to MTD, which may be due to inter-individual pharmacokinetic variations [18,19]. Moreover, other disadvantages, such as recurrent laboratory testing and prolonged time to reach MTD, further contribute to the non-adherence to pharmacotherapy, together with family members' fear of cancer development as well as other adverse effects [18].
Pharmacokinetic analyzes in the selected studies were performed using the population pharmacokinetic model, which requires fewer samples per individual and enables the analysis of the studies which generate dispersed and uneven data, e.g., those performed in pediatric populations [20]. Moreover, Dong et al. implemented non-compartmental pharmacokinetics in their study.

Sample Collection for Pharmacokinetic Analysis
For their study, Dong et al. performed plasma sample collection at seven different times, ranging from before HU therapy to 8 h after administration [13]. Moreover, these authors also developed an experimental design, which identified more informative collection times, thereby reducing the number of samples that needed to be collected for future HU pharmacokinetic studies: 15-20 min, 50-60 min, and 3 h after HU administration, respectively. These identified times help to determine the HU dose by pharmacokinetics as they reduce the number of collections, which is a relevant challenge, especially in pediatric patients. Due to this determination and the establishment of a standard AUC for predicting MTD, this study has been used as a basis in the design of subsequent studies performed by other authors including Estepp et al. [21], McGann et al. [14], and consequently Quinn et al. [15], and Meier et al. [16].
In addition to the collection periods determined by Dong et al., Meier et al., used VAMS devices following the development of a method based on tandem mass spectrometry. This required smaller sample volumes and may serve to facilitate collection in pediatric patients as well as dose adjustment by pharmacokinetics.
In the case of the study conducted by Paule et al., samples analyzed from the first study were collected on the first day of HU treatment, and then at fifteen days, followed by after the first, second, fourth, and sixth subsequent months. In the second study, samples were collected before HU administration and 45, 90, 120, 150, 180, 240, 360, and 480 min later [17]. It is worth mentioning that among the included articles, this was the only study that was performed in adult patients, which may have resulted in a more straightforward collection. Analysis of laboratory parameters showed that patients treated with an HU-personalized dose presented significantly increased HbF levels compared to patients who followed a traditional dosage regimen.

Pharmacokinetics and Pharmacodynamic Parameters
Several important pharmacokinetics parameters encountered in the included studies are presented in Table 2. Dong et al. demonstrated a mean AUC 0-∞ value of 115.7 ± 34.0 mg L −1 h −1 at MTD, establishing 115 mg L −1 h −1 as the predicted AUC 0-∞ at MTD. They also reported MTD values ranging between 14.2 and 35.5 mg/kg/day [13]. This finding corroborates the McGann et al. study, which identified a narrower MTD range (26.7 ± 4.8 mg/kg/day) [14]. On the other hand, Paule et al. presented dosages ranging between 500 and 2000 mg, with 78% of the patients receiving 1000 mg [17]. Another aspect that is also of utmost importance to be considered is that the MTD in patients undergoing dose adjustment according to the standard protocol was reached after 8 months of dose escalation [21], while the patients in the McGann study achieved MTD in just 4.8 months [14]. Although, in the later study, the variation of the time was large (from 3.3 to 9.3 months), the findings demonstrated the effectiveness of the pharmacokinetic dose adjustment regimen in shortening the time necessary for patients to reach the ideal dosage (MTD) when compared to the traditional dose adjustments (4.8 months vs. 08 months, respectively). Moreover, these studies demonstrated clearance values that were remarkably similar except for Dong et al., which observed a slightly higher value than the others. However, the authors considered this value to be close to that described in the literature for both adults and children [13].
Regarding pharmacodynamics, the studies included in the present review identified some covariates, such as body weight, which was associated with clearance and apparent volume of distribution [13,14]. This finding was also identified in a previous study [22] and demonstrates the importance of body weight in assessing the pharmacokinetics of HU in individuals. Dong et al., also reported cystatin C but not serum creatinine as a covariate, as they observed its association with the glomerular filtration rate, kidney injury, and impaired HU excretion, and was thereby considered as a crucial predictor of HU clearance [13]. Moreover, cystatin C has been considered as an important biomarker of renal function in patients with SCA, as pathology can lead to glomerular hypertrophy, and levels of cystatin C have been found to correlate with the glomerular filtration rate [23].
Another covariate considered as predictor of the HU response was HbF, in addition to other laboratory parameters such as hemoglobin and reticulocyte count, which were found to be more pronounced in younger patients who received individualized dosage [14]. Likewise, Meier et al. observed an increase in HbF levels in patients who underwent dose adjustment via pharmacokinetics compared to those who received an initial dose of 20 mg/kg/day [16]. Moreover, these authors also considered pharmacodynamic parameters, such as hemoglobin, MCV, and total reticulocyte and neutrophil counts.
The study of Paule et al. revealed increases in HbF levels and MCV, and reductions in bilirubin, LDH, and neutrophil count after six months of treatment [17]. These authors also reported that HbF achieved a steady state long after MCV, with higher inter-individual variability and responses observed in the simulation of a dosing regimen in which HU was administered/day compared to a five-day per week consecutive regimen.  In addition to the laboratory biomarkers, Quinn et al. demonstrated that nearpancellular and pancellular HbF expression can be used as a biomarker of HU therapy effectiveness. They established the total F-cell fraction of 80% and the F/F-cell values of 10 pg as optimal values corresponding to better pharmacological or genetic responses.
As shown in Table 2, the Paule et al. study is the only one which involved adults. This is an important variable that needs to be considered, since there are existing differences in the pharmacokinetic processes between children and adults, such as the decrease in the intestinal transit time that affects absorption, and the increase in the distribution of hydrophilic drugs as HU, due to the greater volume of extracellular water in preschool children [20]. Moreover, drug metabolism can be influenced by age and liver clearance, which is higher in children than in adults due to elevated liver blood flow. The expression level of liver enzymes such as CYPs, which were found to be positively correlated with age, is another factor that may affect drug metabolism [19]. Despite this, we observed similarities in the pharmacokinetic and pharmacodynamic parameters in the presented studies.

Limitations and Perspectives
The main limitation of the present study is the number of included studies, based on the search terms we established to meet our objective. Moreover, none of these studies stratified patients according to genetic factors such as β S -globin gene haplotypes, which are associated with variations in HbF levels, given that this parameter is used when assessing the HU response. Another limiting factor is the relatively fewer participant number in some of these studies. However, this may be due to the complexity of pharmacokinetic studies themselves, principally those involving children, in addition to other issues related to therapy adherence, withdrawal during the study, etc.
According to the above, there is a need for studies involving different populations, considering aspects such as the genetic background of geographical regions, age groups, etc. These studies will shortly enable, the application of this therapeutic strategy in patient candidates for HU therapy, bringing them substantial benefits. Figure 2 clearly illustrates this perspective.
Pharmaceuticals 2023, 16, x FOR PEER REVIEW 11 of 12 tient candidates for HU therapy, bringing them substantial benefits. Figure 2 clearly illustrates this perspective.

Methods
The present systematic mini-review investigated previously published studies based on the following question: "Does individualized dosing in accordance with pharmacokinetic analysis contribute to the effectiveness of HU treatment in SCA patients?".
Inclusion criteria were clinical trial-type articles performed in SCA patients where the intervention was HU dose adjustment, regardless of HU formulation and participant age, and without limitations as to the publication language or date. Searches between December 2020 and August 2022 on the Embase, Pubmed, Scopus, Web of Science, Scielo, Google Scholar, and the Virtual Library of Health databases were performed using the following terms: ("sickle cell disease" OR "sickle cell anemia" OR "sickle cell anaemia") AND ("hydroxyurea" OR "hydroxycarbamide") AND pharmacokinetics AND ("dosing regimen" OR "guided dose" OR "guided dosing") and their equivalents in Portuguese, Spanish, and French. Characteristics of each study were listed and compared with the inclusion criteria to assess their eligibility.

Methods
The present systematic mini-review investigated previously published studies based on the following question: "Does individualized dosing in accordance with pharmacokinetic analysis contribute to the effectiveness of HU treatment in SCA patients?".
Inclusion criteria were clinical trial-type articles performed in SCA patients where the intervention was HU dose adjustment, regardless of HU formulation and participant age, and without limitations as to the publication language or date. Searches between December 2020 and August 2022 on the Embase, Pubmed, Scopus, Web of Science, Scielo, Google Scholar, and the Virtual Library of Health databases were performed using the following terms: ("sickle cell disease" OR "sickle cell anemia" OR "sickle cell anaemia") AND ("hydroxyurea" OR "hydroxycarbamide") AND pharmacokinetics AND ("dosing regimen" OR "guided dose" OR "guided dosing") and their equivalents in Portuguese, Spanish, and French. Characteristics of each study were listed and compared with the inclusion criteria to assess their eligibility.
Data selection and study quality analysis were performed by two independent authors using the QAT [12] that evaluates the aspects as risk of bias and study design, while data synthesis was performed according to the Cochrane Manual of Systematic Reviews of Interventions. We extracted the data from the included studies and summarized them in a table (Table 1) including variables such as country, objectives, participant's characteristics, pharmacokinetic model, and principal results. We also sought to investigate and report pharmacokinetic parameters such as the AUC and MTD, and pharmacodynamic parameters including HbF, MCV, and hemoglobin. The protocol has been registered in PROSPERO CRD42022344512.

Conclusions
Analysis of the included studies highlights the effectiveness of dose adjustment based on pharmacokinetic parameters, which may contribute to greater patient adherence in addition to improvements in the HU response, as evidenced by the pharmacodynamic parameters, and the reduction of clinical manifestations in comparison with conventional HU treatment. Several crucial studies presented strategies that resulted in less blood volume and fewer collection events to enable the adoption of this practice. They also identified an AUC value as an MTD predictor. However, additional studies are required, since only a few studies were identified as consistent with the objectives of this review.