HemoTypeSC Demonstrates >99% Field Accuracy in a Sickle Cell Disease Screening Initiative in Children of Southeastern Uganda

Sickle Cell Disease (SCD) is a widely prevalent hemoglobinopathy in sub-Saharan Africa that is frequently deadly in early life, killing up to 90% of afflicted children before their fifth birthday.1 SCD accounts for ~20% of annual childhood deaths in Uganda2 - one of the first African countries surveyed for SCD prevalence. This health danger is currently addressable: treatment with hydroxyurea and prophylactic intervention programs including antibiotics and immunizations have drastically reduced SCD mortality in developed areas3 and could be implemented in sub-Saharan Africa in a cost-effective manner.4 However, the cornerstone of such programs, newborn diagnostic screening, has not been widely implemented in Uganda, or elsewhere in Africa. This dilemma springs from the cost and logistical problems with accepted diagnostic methodologies, which require laboratory equipment, highly-trained operating personnel, uninterrupted electrical supply, and sample transport away from the point-of-care (POC). Clearly, a low-cost, low-complexity, POC diagnostic test for SCD could help diminish this public health crisis.

HemoTypeSC is a competitive lateral-flow immunoassay that uses monoclonal antibodies to detect hemoglobins A, S, and C in a 1.5-μL droplet of whole blood in about ten minutes. This test requires no equipment or refrigeration and has displayed a sensitivity of ≥99.5% and a specificity of ≥99.9% in both low-resource screening clinics5 and higher-resource laboratories.6 HemoTypeSC is also inexpensive (<$2.00 per test) and highly accurate in newborns and young children with high fetal hemoglobin (HbF) level.5 In this study we implemented HemoTypeSC as a standard of SCD diagnosis while simultaneously validating the test's accuracy in children of southeastern Uganda, with the study aimed at providing evidence for the applicability of the test in widespread newborn screening programs in the region.

This study was designed as a blinded, prospective diagnostic accuracy trial of HemoTypeSC as an investigational test compared to capillary zone electrophoresis (CZE) as a reference method. Jinja Regional Referral Hospital, the largest hospital in southeastern Uganda, was selected as the study center due to its relatively high daily patient volume, as well as its suitability as a prototype center for organized newborn SCD screening within Uganda. CZE cost was $12 per test at this study site. The study was designed and managed by the Sickle Cell Association of Uganda, and HemoTypeSC tests were donated by Silver Lake Research Corporation. Exactly 1000 individuals were prospectively recruited from the hospital wards, outpatient clinics, and referrals from nearby health centers, with an emphasis on children under five years old. Both sexes and a range of ages were represented in the study as shown in Supporting Information Table S1. Importantly, 19.7% of all participants were between one month and one year old, an age at which HbF levels may be elevated and at which diagnostic accuracy is crucially dependent on high test sensitivity for HbA and HbS. Furthermore, this age group would be targeted by newborn and immunization clinic screening programs and is thus logistically informative.

From each participant, a 1.5-μL sample of blood was first collected for the HemoTypeSC test using the included blood collection device from a heel-prick (newborns, infants, and small children) or venipuncture (older children). A larger blood sample of 2.0-mL was obtained in a collection tube for confirmatory analysis. HemoTypeSC testing was performed by a small team of health professionals immediately upon sample collection from each patient, while samples were transported to an accredited diagnostic laboratory at the Uganda Cancer Institute (Kampala, Uganda) for CZE reference testing. HemoTypeSC and reference test results were logged into separate spreadsheets, and study administrators were blinded to results until the data analysis stage. Individuals conducting the HemoTypeSC test required minimal training and referred to an instructional video and package insert instruction sheet (available at www.hemotype.com). No issues were reported with performing the procedure or interpreting the test results.

The HemoTypeSC test detects HbA, HbS, and HbC through a combination of three distinct, highly-specific, competitive immunoassays. HemoTypeSC sensitivities and specificities were calculated for independent detection of HbA and HbS (HbC is not prevalent in Uganda and was not detected by either test method). As shown in Supporting Information Table S2, 904 out of 904 samples containing HbA (phenotype HbAA or HbAS) and 280 out of 280 samples containing HbS (phenotype HbAS or HbSS) were correctly identified by HemoTypeSC when compared to CZE, for a sensitivity and specificity of 100% for both variants.

HemoTypeSC was next compared to CZE for detection of the Hb phenotypes HbAA (normal), HbAS (sickle cell trait), and HbSS (SCD). Initial data analysis indicated that HemoTypeSC correctly identified 998 out of 1000 phenotypes, for an overall accuracy of 99.8% (Table 1). This included 720/720 (100%) specimens correctly identified as HbAA, 182/182 (100%) correctly identified as HbAS, and 96/98 (98%) correctly identified as HbSS. The two discordant samples were both identified as HbSS by CZE and HbAS by HemoTypeSC. To resolve these discrepancies and determine which testing method delivered the accurate result, the patients with the discordant HbSS/HbAS results were recalled for repeat testing with a secondary reference method (HPLC was intended). During the ensuing phone contact for scheduling, it was discovered that both individuals were in fact previously diagnosed with SCD by HPLC, and both had been recently transfused. The previous diagnostic result and transfusion reports for these participants were subsequently requested and viewed by the Uganda ground team, with previous diagnoses confirmed as HbSS. Because HbA was indeed present in the blood of these two patients at the time of sampling, the HbAS result was determined to be true. The result discrepancies may be due to the relatively low overall sensitivity of the CZE method, along with a high sensitivity for each Hb of HemoTypeSC.6 A secondary data analysis was therefore conducted, in which these two specimens were included as true positives for HbAS and true negatives for HbSS (bottom half of Table 1). Following this adjustment, it was ultimately determined that HemoTypeSC correctly identified 1000 out of 1000 phenotypes across all patients screened, including 96/96 HbSS specimens, for an overall sensitivity, specificity, positive predictive value, and negative predictive value of 100%.

SCD is one of the largest causes of childhood mortality across sub-Saharan Africa, and most of these deaths are preventable. Newborn screening initiatives in high-income countries combined with proper treatment and counseling have been highly effective at diminishing SCD-related mortality, but the cost and logistical difficulties associated with current gold-standard diagnostic methodologies have made early-life SCD screening unfeasible in resource-constrained regions with high disease prevalence. The United Nations General Assembly and World Health Organization have highlighted SCD as a priority threat to global health, but African governments and Ministries of Health have thus far been hesitant or unable to devote adequate budget and resources in order to and establish diagnostic screening efforts. Clearly, an accurate, low-cost test that can rapidly determine hemoglobin type at the point-of-care could form the foundation of such screening efforts.

In this study we implemented a sustainable screening effort for SCD at a typical resource-constrained medical center in a high disease prevalence region of southeastern Uganda. The HemoTypeSC test was performed efficiently, results were highly accurate, and the cost-per-test of HemoTypeSC ($2.00) was approximately 1/6th that of the existing gold-standard diagnostic method (~$12.00). The results from this study indicate that HemoTypeSC performs at least as accurately as the gold-standard method of CZE in detecting SCD and sickle cell trait at the POC in a resource-limited setting. This to our knowledge represents the first ever report of a rapid test for SCD displaying 100% sensitivity and specificity in a field validation study. While our detected frequency of HbS variant and of SCD trait is in line with previous reports for southeastern Uganda, the 9.6% prevalence of homozygous disease (confirmed by both the reference and investigational test methods) is unexpectedly high. We hypothesize that this may be a result of both consanguinity and self-selection of the patient pool, as some recruited individuals may have presented at the Jinja Regional Referral Hospital with illnesses and symptoms due to SCD along with their family members.

In summary, HemoTypeSC represents a promising tool that can presently enable newborn and general population screening. Widespread combination of HemoTypeSC newborn and population screening programs with appropriate treatment, prophylaxis, and health counseling systems in countries most affected by the disease could save the lives of millions of children over the coming decades.

AUTHORSHIP CONTRIBUTIONS

R.N., S.K, A.T., and S.B. designed the research, performed the research, and analyzed the data. M.G. and E.S. approved the study design and donated the HemoTypeSC tests. All authors read and approved the final draft of the manuscript.

DISCLOSURES

Authors Erik Serrao and Mark Geisberg are employees at Silver Lake Research, which sponsored this study and donated HemoTypeSC tests.