Hydroxyurea therapy in sickle cell disease: evaluating health care costs and use in a Brazilian tertiary setting Open Access

Sickle cell disease (SCD) is a chronic condition characterized by recurrent vaso-occlusive episodes, often causing intermittent pain crises.¹ Recurrent pain episodes significantly increase health care use, particularly emergency department (ED) inpatient care. Additionally, microcirculation vaso-occlusion (VOC) leads to end-organ damage, including splenic dysfunction, which causes functional asplenia, increasing risks of infections, including septicemia and bacteremia.² 

Hydroxyurea is a disease-modifying agent with multiple benefits in SCD; it increases fetal hemoglobin among other biological effects (eg, inflammation), thereby reducing VOC and other disease complications.^3-5 Studies have shown that hydroxyurea therapy leads to a significant reduction in hospitalizations, blood utilization, opioid use, and mortality.^6,7 In high-resource settings, hydroxyurea therapy has been associated with a reduction in total annual costs, including cost reductions in hospitalizations, ED visits, other ambulatory visits, and use of analgesia and transfusions.^8,9 

While high-income countries have broadly implemented hydroxyurea therapy, middle-income countries such as Brazil, although it has established national guidelines since 2002, initially applied more restrictive criteria for use. Despite the clinical benefits of hydroxyurea, mortality remains high, particularly among untreated patients. One study reported that only 1 out of 36 deaths involved a patient on hydroxyurea, with most deaths occurring in those with severe genotypes. Another national cohort reported a median age at death of 31.9 years, with mortality rates of 10.5% in children and 18.9% in adults.^9,10 This gap poses challenges for resource-limited health systems facing rising chronic health care costs.¹¹ Our group previously investigated the cost of acute care resource use in patients with SCD.² We now extend prior work by investigating the potential impact of hydroxyurea in reducing the overall costs of care for SCD. We compared the care expenditures of patients with SCD treated with hydroxyurea to those not treated at a specialized blood center in Rio de Janeiro, Brazil. We tested the hypothesis that the use of hydroxyurea decreases overall care costs for a cohort of children and adults with SCD.

This retrospective cohort consisted of children and adults who had a confirmed diagnosis of SCD and were treated at Instituto Estadual de Hematologia do Rio de Janeiro (HEMORIO), a state-funded tertiary hematology treatment referral facility serving ∼3000 people with SCD. Patients were included in the analysis if they had at least 1 ambulatory ED visit or hospital admission during the study period between 1 January 2018 and 30 June 2018. Ambulatory encounters were included if they were related to routine SCD management, such as dental care, laboratory-only visits, transfusion-only visits, and dressing changes. ED visits and hospital admissions were included if they were attributable to acute SCD-related complications (eg, vaso-occlusive crises, acute chest syndrome, or fever). We excluded encounters unrelated to SCD (eg, trauma or unrelated medical conditions). The data on acute care (ED visits and admissions) were previously published and included here to allow for the calculation of total costs.² 

Patients with at least 1 hydroxyurea prescription during the study period were classified as the hydroxyurea treatment group. All prescriptions for hydroxyurea filled during the analysis period were included in the data set. In accordance with national guidelines during the study period, hydroxyurea was typically initiated at 10 to 15 mg/kg per day and titrated to maximum tolerated dose while monitoring for hematologic toxicity and clinical response. These recommendations, established by the Brazilian Ministry of Health in 2002 and revised in 2018 and 2024, targeted patients aged ≥3 years with severe anemia, frequent VOCs, or a history of acute chest syndrome. Additionally, patients presenting with acute SCD-related complications followed local standards and guidelines set forth by the Brazilian Ministry of Health.^12,13 

Data were extracted from the institution’s clinical database using queries made through Microsoft SQL Server interfaced with the Visual FoxPro 9 system, consolidating patient care management across multiple hospital departments. The HEMORIO ethics committee approved the use of existing patient data, and all methods followed local institutional review board guidelines and regulations.

For each patient, the number of visits for each category (ambulatory, ED, or hospitalization) was calculated and aggregated across the study period, and the median was computed. Health care use was stratified by hydroxyurea adherence, which was estimated using the medication possession ratio (MPR).^14,15 The MPR was calculated by dividing the number of days of hydroxyurea possession by the total number of days in the study period.

Total cost included the costs of ambulatory visits, ED visits, and hospital admissions, whereas acute care costs referred to ED visits and admission costs only. Per hospital policy, all patients admitted are first treated in the ED. Thus, all hospital admissions are subsets of ED visits. Hence, ED visits that did not result in admission were classified as ED visits, and ED visits followed by admission were classified as admissions.

Costs were categorized as fixed or variable. Fixed costs were identified using institutional budgets, whereas variable costs were quantified using a microcosting approach.¹⁶ Fixed costs included (1) personnel time, (2) imaging and laboratory tests, (3) administrative time (including pharmacy costs for drug storage and dispensation), and (4) specialty consulting fees (eg, psychology and infectious disease specialists). For personnel costs, the required time for patient care was calculated by multiplying the time spent in direct patient care (ie, the estimated effort percentage of each staff member) by the annual salary of all staff with primary patient care responsibilities (eg, physicians, hematologists, nurses, and physical therapists). All other fixed costs were determined based on a combination of staff time spent treating patients with SCD and HEMORIO’s hourly salary rate. Variable costs included all consumables accounted for through primary data extraction from the HEMORIO database to estimate the costs of medications, laboratory tests, blood transfusions, and subspecialty consultations.

The total median costs were calculated by multiplying the number of visits (ambulatory, ED, and admission) by the total costs (including both fixed and variable costs) per ambulatory visit, per ED visit with subsequent discharge, and per ED visit with subsequent admission.

Per–ambulatory visit, per–ED visit, and per–admission visit costs were calculated for each group (prescribed hydroxyurea vs not prescribed) and stratified by age groups (<18 years, 18-25 years, and >25 years). The total cost per ambulatory, ED, or admission visit was calculated for each individual by dividing the total fixed and variable costs for all patients seen in ambulatory, ED, or admission visits during the study period. All costs are shown in 2018 US dollars, reflecting the study period, and calculated by converting from Brazilian reais (R$) to US dollars ($) using a conversion rate of $1 = R$3.18 (as of 30 June 2018).

Descriptive statistics are provided as total counts and percentages for categorical variables and median and interquartile ranges (IQR) for continuous variables. One-sample proportion test with continuity correction and the Wilcoxon rank sum test with continuity correction were used to compare the clinical outcome differences (ie, frequency of ED or admission visits) across discharge diagnosis groups. The Wilcoxon rank sum test with continuity correction was used to compare time and cost differences across the groups. The R software was used for all statistical analyses.¹⁷ 

There were 3331 active patients with SCD at HEMORIO during the study period. Of these patients, 3032 (91.0%) had at least 1 encounter in the ambulatory clinic or the ED or were admitted during the analysis period. Of these patients, 1561 (51.5%) were female, and 1471 (48.5%) were male, with a median age of 15 years (range, 0-76 years); a total of 614 (20.3%) were prescribed hydroxyurea, of whom 98 (16.0%) had a hydroxyurea MPR of >50% compared to 516 (84.0%) with a hydroxyurea MPR of ≤50%. The sickle cell genotype distribution was 2212 (73.0%) for homozygous sickle cell anemia (HbSS) or sickle beta-zero thalassemia (HbSβ⁰), 610 (20.1%) for hemoglobin SC disease (HbSC), 148 (4.9%) for sickle beta-plus thalassemia (HbSβ⁺), and 62 (2.0%) for hemoglobin SD disease (HbSD) (Table 1).

Of the 3032 patients, 2994 had a total of 18 418 ambulatory visits, with 5434 (29.5%) being from patients on hydroxyurea (Table 2). Patients on hydroxyurea had higher ambulatory service use (median, 6 vs 4 visits for patients not on hydroxyurea; P < .0001). This trend was consistent across all age groups (Table 3). Total cost of care in the ambulatory setting was ∼$3.2 million in the hydroxyurea-treated group compared to ∼$3.1 million in the nontreated group (Table 1). Costs per visit were significantly elevated for hydroxyurea-treated patients ($973.95 vs $212.70 for patients not on hydroxyurea; P < .0001; Table 4) driven by laboratory monitoring, transfusions, and chronic care (eg, wound management; supplemental Figure 1). Supplemental data also support higher costs among patients in the hydroxyurea group compared to the nontreatment group (supplemental Figure 2). Notably, patients with MPR of >50% incurred higher costs than those with MPR of ≤50% ($1120.84 vs $951.43, respectively; P = .0092; Table 5), particularly among adults aged >25 years ($1417.52 vs $867.58, respectively; P = .0004; Table 5). Iron chelation therapy further increased total costs (P = .0428), although its exclusion did not alter overall trends (supplemental Figure 2). No significant difference in median ambulatory transfusion-related costs was observed overall between patients treated on hydroxyurea compared to those not on hydroxyurea; however, among the patients aged <18 years, those receiving hydroxyurea incurred significantly higher costs compared to those not treated with hydroxyurea ($410.36 vs $205.18, respectively; P = .0477; supplemental Table 1).

There were 2152 ED visits across 1017 patients, of which 1655 (76.9%) were from patients not on hydroxyurea compared to 487 (22.6%) from patients on hydroxyurea (P < .0001; Tables 1 and 2). When stratified by age groups, the difference in median ED visits between patients on and off hydroxyurea was no longer statistically significant (Table 3). Total ED visit costs were higher among patients not treated with hydroxyurea, totaling $422 540.70 compared to $128 361.80 for the hydroxyurea-treated group (Table 1).

The median total cost per ED visit was similar for both groups ($364.03 for hydroxyurea-treated patients vs $338.17 for nontreated patients; P = .7520; Table 4). There was also no significant difference in costs based on hydroxyurea adherence (MPR >50% vs ≤50%), with costs of $266.73 and $371.65, respectively (P = .3124; Table 5). Costs were comparable across all age groups. The median transfusion-related costs did not differ significantly between patients receiving hydroxyurea vs those not treated with hydroxyurea (supplemental Table 1).

There were 519 admissions among 381 patients, with 392 (75.5%) admissions among patients not on hydroxyurea compared to 127 (24.5%) among patients on hydroxyurea (P < .0001; Tables 1 and 2). However, when the data were stratified by age group, this difference was no longer statistically significant (Table 3). Total admission costs were substantially higher among patients not receiving hydroxyurea, totaling $860 566.40 compared to $278 505.10 for hydroxyurea-treated patients (Table 1).

The median total cost of hospital admissions was similar for both groups: $1923.61 for hydroxyurea-treated patients and $1991.80 for nontreated patients (P = .7498; Table 4). Similarly, costs did not vary according to hydroxyurea adherence (MPR >50% and ≤50%), with costs of $1522.36 and $1942.10, respectively (P = .7224; Table 5). There was no statistically significant difference among the various age groups, and costs were similar for patients on hydroxyurea compared to those not on hydroxyurea. The median transfusion-related costs did not differ significantly between patients on hydroxyurea therapy vs nontreated patients (supplemental Table 1).

Patients on hydroxyurea had significantly higher overall medication costs compared to those not on hydroxyurea therapy. The median medication cost for patients on hydroxyurea was $694.50 compared to that for patients not prescribed hydroxyurea at $43.08 (P < .0001; Table 5). There were no differences in medication cost based on hydroxyurea adherence (MPR >50% vs ≤50%), with costs of $694.02 and $669.92, respectively (P = .1009; Table 6).

For ambulatory care, median medication costs for patients on hydroxyurea were higher than the median medication costs for those not on hydroxyurea ($658 vs $38.41, respectively; P < .0001; Table 5). Among hydroxyurea-treated patients, those with an MPR of >50% trended toward higher costs compared to those with an MPR of ≤50% ($690.88 vs $648.02, respectively; P = .0679; Table 7).

For ED visits, medication trended toward higher median costs among hydroxyurea-treated vs nontreated patients ($9.15 vs $7.15, respectively; P = .0547; Table 5). Costs did not vary relative to the MPR ($10.15 vs $9.11 for MPR >50% and ≤50%, respectively; P = .3646; Table 6).

The median medication costs for admission visits were higher for patients on hydroxyurea than for those not on hydroxyurea ($44.17 and $27.52, respectively; P = .0118; Table 5).

The total costs for patients on iron chelation therapy ranged from $18.81 to $136 933.25 among patients not on hydroxyurea compared to the total costs for those not on iron chelation therapy who were on hydroxyurea, which ranged from $18.81 to $52 540.88 (supplemental Figure 3).

SCD presents a complex and persistent challenge for health care systems worldwide, particularly in middle-income countries such as Brazil. The disease incurs substantial treatment costs driven not only by acute care needs but also by the management of chronic complications.^11,18 In our analysis, children and adults treated with hydroxyurea experienced fewer acute care visits and more ambulatory visits. Although the total cost of care was higher among those receiving hydroxyurea, it did not differ for acute care visits.

This study builds on our prior work, which explored the costs of acute care treatment for patients with SCD and examined the potential cost savings from hydroxyurea therapy. Our previous research identified that (1) anemia and end-organ damage were key cost drivers for ED visits and hospitalizations; (2) patients with severe sickle cell genotypes (HbSS/HbSβ⁰-thalassemia) had significantly more ED visits, hospitalizations, and complications and longer lengths of stay and experienced higher hospitalization costs compared to those with HbSC or HbSβ⁺-thalassemia; and (3) higher adherence to hydroxyurea was associated with lower acute care costs, suggesting that better adherence to hydroxyurea may lead to cost savings.² In this study, we confirmed that hydroxyurea therapy reduced ED visits and hospitalizations; however, this did not translate into significant cost savings across total health care expenditures.

Patients on hydroxyurea experienced higher costs in the ambulatory care setting, which included laboratory tests, transfusions, iron chelation therapy, specialty consultations, medication management, and wound care. Despite these higher costs, the reduction in acute care visits, specifically for vaso-occlusive crises, was consistent with previous literature and further supports the role of hydroxyurea in reducing acute episodes and hospitalizations.^7,9 The higher costs of care for patients with SCD associated with chronic care complications such as transfusion-related iron overload and wound management are less frequently discussed in the literature. In our study, chronic care complications were significantly higher in patients not on hydroxyurea compared to those on hydroxyurea. The associated costs, particularly for transfusions and wound care, were comparable between the 2 groups (treated vs nontreated) when assessed by median expenditure. The discrepancy between a higher clinical burden and similar median costs can reflect an uneven distribution of health care expenditures, a pattern often observed in real-world data sets where a small subset of patients incurs disproportionately high costs. In our study, the nonhydroxyurea group comprised ∼80% of the cohort, increasing the likelihood of capturing high-cost outliers. Although we applied appropriate statistical controls, the smaller sample size in the hydroxyurea-treated group necessitates caution when interpreting median cost estimates as these are less stable and less representative in smaller subgroups. Understanding both total and median cost metrics provides complementary perspectives: total costs reflect the overall burden on the health care system, whereas median costs provide insight into the typical patient experience.¹⁹ In this context, the analysis reveals that, although untreated patients had a higher burden of chronic complications, this did not translate into proportionally higher costs, particularly in the ambulatory setting, potentially offsetting cost savings typically associated with hydroxyurea use. Additionally, low adherence to hydroxyurea therapy (only 16% of patients achieving adherence of >50% as indicated by the MPR) complicates the interpretation of our results as it likely reduces the treatment’s potential to lower health care costs.

In our cohort, transfusion-related care and wound management, particularly dressing changes for leg ulcers, significantly contributed to ambulatory costs among hydroxyurea-treated and nontreated patients (supplemental Figure 2). Although toxicity monitoring was not captured in our data set, Brazilian national guidelines recommend monthly laboratory monitoring for cytopenias.¹³ Leg ulcers, sometimes associated with hydroxyurea use, were also prevalent across the treated and untreated groups, hence not being solely attributable to therapy. Leg ulcers, which affect an estimated 10% of patients with SCD during their lifetime, often require complex and prolonged wound care.^20,21 These findings align with global data highlighting wound care as a significant outpatient cost driver due to the specialized materials and frequent follow-up appointments required.²⁰ Our findings also align with data from high-income countries demonstrating the economic impact of transfusion therapy on SCD. In the United States, total transfusion-related costs increased by 907%, and out-of-pocket costs increased by 285% for patients with SCD compared to matched controls.²² In Brazil, standard secondary stroke prevention is primarily managed with chronic transfusion therapy rather than hydroxyurea.¹³ These findings underscore that transfusion-related costs are substantial and, even in middle-income country settings, can significantly contribute to the overall burden of ambulatory care.

Medication costs were another critical driver of increased costs among both the hydroxyurea-treated and nontreated groups. While the direct cost of hydroxyurea contributed to higher expenditures in the treated group, additional medications, including pain management and iron chelation therapy, to manage chronic complications remained a significant component of the cost in both treated and untreated patients in the cohort. The persistence of high medication costs across both the hydroxyurea-treated and non–hydroxyurea-treated groups suggests that hydroxyurea alone may not be sufficient to reduce the need for these adjunctive therapies in routine clinical practice. Notably, many hydroxyurea-treated patients also received iron chelation therapy, reflected in their ambulatory visits, as shown in supplemental Figure 1. Iron chelation therapy costs for patients in this study totaled ∼$150 000 per patient per year, similar to the average annual costs in the United States. Previous studies have shown that up to 66.9% of patients with SCD receiving frequent transfusions were also prescribed iron chelation, with most of these services delivered in the ambulatory setting.²³ Although iron chelation therapy is now considered the standard of care for specific SCD subgroups, it remains a significant cost driver in both low- and middle-income country and high-income country contexts, as demonstrated in our cohort.

While the role of hydroxyurea in reducing acute care events is well documented, only 20% of patients in this study received the drug, with very low adherence (16% achieving an MPR of >50%). This falls short of the adherence threshold cited in the literature (MPR >80%), which is necessary for optimal clinical and cost outcomes.²⁴ Low adherence may have prevented the hydroxyurea group from demonstrating the anticipated cost savings associated with reduced acute care use. Moreover, many patients in our cohort were likely in the chronic phase of their disease and initiated hydroxyurea later in life, in line with the Brazilian Ministry of Health’s national SCD guidelines.¹³ While these guidelines encourage early use, their implementation and uptake could have varied across the country. The timing of hydroxyurea initiation is critical as studies have demonstrated that early therapy is more effective in preventing the progression of SCD complications. For example, hydroxyurea initiation, especially before the age of 10 years at the maximum tolerated dose, can improve albuminuria.²⁵ Additionally, hydroxyurea administered at the maximum tolerated dose in children could improve neurocognition after adjusting for social vulnerability, sex, and treatment duration²⁶ Although hydroxyurea reduced acute complications, its impact on chronic pain remains less clear, with limited data available, highlighting that early initiation may delay or mitigate chronic pain.²⁷ 

Another insight from this research is that total medication costs for patients on hydroxyurea were significantly higher in the ambulatory setting than those incurred during ED visits or hospitalizations (Table 6). This highlights a critical dynamic in health care delivery in many low- and middle-income countries whereby inpatient medications may be subsidized or discounted through institutional coverage, bulk purchasing, or manufacturer discounts.²⁸ These are advantages that are often not extended to outpatient settings. As a result, patients may face higher out-of-pocket or system-level costs in ambulatory care, particularly for the management of chronic diseases. This underscores the need to understand further and address the structural cost differences across care settings to ensure more equitable and efficient access to essential medications such as hydroxyurea.

The greater frequency of ambulatory visits among patients receiving hydroxyurea can reflect the deliberate and valuable investment in patient health. Although ambulatory care costs were higher during the study period, this pattern likely indicates a societal willingness to incur increased short-term costs to support patients who are engaged in their care.²⁹ This interpretation is supported by Brazil’s updated 2018 national SCD guidelines, which recommend initiating hydroxyurea therapy from 9 months of age even in nonsevere cases as a proactive measure to reduce long-term complications.¹² Such policy guidance reflects an institutional shift toward earlier intervention. It aligns with a cost-effectiveness framework where up-front expenditures in ambulatory management are seen as necessary to improve long-term outcomes and reduce acute care demands in the long run.

Our study had limitations. The relatively short 6-month follow-up limits our ability to assess the cumulative transfusion burden for each patient, which may impact iron chelation use and long-term health care use and costs. An extended follow-up period would allow for a comprehensive evaluation of the long-term effects of hydroxyurea therapy on both clinical outcomes and health care costs. Furthermore, low adherence to hydroxyurea therapy was prevalent in our cohort. Future studies should focus on improving adherence to hydroxyurea as better adherence may lead to cost savings and improved clinical outcomes. Additionally, examining low adherence in future studies will help determine barriers within the health care system, such as long waiting times; travel distance to treatment centers; or economic factors, including the cost of transportation and patient education. We lacked data on baseline and follow-up fetal hemoglobin levels, preventing correlation between pharmacodynamic response and clinical outcomes or cost. Chronic complications such as stroke and priapism were also not captured and should be explored in future research. Nonmedical costs for patients with chronic diseases should also be assessed in future work to estimate indirect costs, such as lost income, travel expenses, and caregiver burden. These insights will provide a more comprehensive understanding of the total health burden, particularly in this group of patients. As patients with SCD age, chronic complications increase, which leads to a natural increase in health care costs, reinforcing the importance of long-term management strategies that address both acute and chronic disease components. Although our data did not capture metrics related to health care access and frequency of laboratory monitoring, we assessed prescription adherence using MPR. It is possible that patients on hydroxyurea were more engaged with the health care system due to the structured follow-up required for the safe administration of the drug. This may partially explain the higher ambulatory costs observed in this group as patients not on hydroxyurea could have received acute care outside the HEMORIO system. In addition, future studies with larger and more balanced samples will enable more robust comparisons using median cost metrics, thereby reducing the influence of outliers on total cost calculations.

In conclusion, our study aimed to assess the financial impact of hydroxyurea therapy on health care use by investigating overall costs for patients with SCD at a specialized tertiary center in Brazil. Hydroxyurea has demonstrated benefits in reducing acute care events; however, its overall potential for cost savings has not been realized, at least at the observed medication adherence rates and because of persistent costs associated with chronic care. Specifically, higher costs related to chronic care complications, such as transfusion-related iron overload, wound care, and other ambulatory services, remained substantial across both treated and untreated groups, which may have counterbalanced the savings from reduced ED visits and hospitalizations. Our findings suggest that, while hydroxyurea may offer potential cost savings in reducing acute care use, it is essential to consider the ongoing costs associated with managing chronic diseases. Future studies should investigate the long-term effects of hydroxyurea on complications and explore strategies to enhance adherence to and early initiation of therapy, thereby realizing its cost-saving potential.

This work was supported by the National Cancer Institute (NCI) Research Grant CA-016359 from the National Institutes of Health (NIH) (G.G.).

The contents of this publication are solely the responsibility of the authors and do not necessarily represent the official views of the funding sources. This manuscript is the result of funding in whole or in part by the NIH. It is subject to the NIH Public Access Policy. Through acceptance of this federal funding, the NIH has been given a right to make this manuscript publicly available in PubMed Central upon the official date of publication, as defined by the NIH.

Contribution: C.L. conceptualized and designed the study; E.M.d.N. extracted and prepared the clinical and cost data; L.L.A.B. performed data analysis and interpretation, drafted the initial manuscript, and completed the full manuscript; E.M.d.N. performed data analysis; L.L.A.B., G.G., J.S.H., C.L., P.M., D.F., J.D., R.B., and T.O. were responsible for critical review; E.M.d.N. and G.G. were responsible for critical review and methodology supervision; J.S.H. and C.L. were responsible for critical review, methodology supervision, and final approval of the version to be submitted; and all authors approved the final manuscript.

Conflict-of-interest disclosure: J.S.H. reports receiving royalties from UpToDate. The remaining authors declare no competing financial interests.

Correspondence: La’Ron L. A. Browne, Department of Global Pediatric Medicine, St. Jude Children’s Research Hospital, 262 Danny Thomas Place, Memphis, TN 38105; email: la'ron.browne@stjude.org.