Evaluation of the CareStart™ glucose-6-phosphate dehydrogenase (G6PD) rapid diagnostic test in the field settings and assessment of perceived risk from primaquine at the community level in Cambodia

Abstract

Background: Primaquine is an approved radical cure treatment for Plasmodium vivax malaria but treatment can result in life-threatening hemolysis if given to a glucose-6-phosphate dehydrogenase deficient (G6PDd) patient. There is a need for reliable point-of-care G6PD diagnostic tests.

Objectives: To evaluate the performance of the CareStart™ rapid diagnostic test (RDT) in the hands of healthcare workers (HCWs) and village malaria workers (VMWs) in field settings, and to better understand user perceptions about the risks and benefits of PQ treatment guided by RDT results.

Methods: This study enrolled 105 HCWs and VMWs, herein referred to as trainees, who tested 1,543 healthy adult male volunteers from 84 villages in Cambodia. The trainees were instructed on G6PD screening, primaquine case management, and completed pre and post-training questionnaires. Each trainee tested up to 16 volunteers in the field under observation by the study staff.

Results: Out of 1,542 evaluable G6PD volunteers, 251 (16.28%) had quantitative enzymatic activity less than 30% of an adjusted male median (8.30 U/g Hb). There was no significant difference in test sensitivity in detecting G6PDd between trainees (97.21%), expert study staff in the field (98.01%), and in a laboratory setting (95.62%) (p = 0.229); however, test specificity was different for trainees (96.62%), expert study staff in the field (98.14%), and experts in the laboratory (98.99%) (p < 0.001). Negative predictive values were not statistically different for trainees, expert staff, and laboratory testing: 99.44%, 99.61%, and 99.15%, respectively. Knowledge scores increased significantly post-training, with 98.7% willing to prescribe primaquine for P.vivax malaria, an improvement from 40.6% pre-training (p < 0.001).

Conclusion: This study demonstrated ability of medical staff with different background to accurately use CareStart™ RDT to identify G6PDd in male patients, which may enable safer prescribing of primaquine; however, pharmacovigilance is required to address possible G6PDd misclassifications.

Fig 1. Project overview.

Fig 2. Distribution of G6PD enzymatic activity levels among male study population in Cambodia.

The bimodal distribution of G6PD activity levels was observed as was expected for male volunteers enrolled in the study. There were nine volunteers whose G6PD quantitative test results ranged between 30% and 60%. Among the nine, three had borderline low G6PD activity levels (30.1%, 31.0% and 33.9%) with corresponding G6PD deficient status based on CareStart^TM RDT. Six volunteers whose G6PD enzymatic activity ranged between 35.7% to 59.3% had a normal RDT result.

Fig 3. Knowledge scores for trainees who took part in the training workshop.

Significant improvement in the knowledge scores was observed following training, and maintained through week 8 follow up. There was additional improvement in the knowledge scores on week 6 and week 8 follow up. The list of questions utilized to assess knowledge are included in S2 Table.

Fig 4. Knowledge assessment score based on highest level of education achieved by the trainees and their level of experience.

Scores improved post training for all education levels: college (C), high-school (HS), primary-level education (P), and secondary-school education (S) (p<0.001) (Panel A). Scores improved post training for all levels of experience (p<0.001) (Panel B).

Fig 5. Acceptability assessment for using G6PD RDTs to guide treatment decisions with primaquine.

7-point Likert scale was used and scores of 1–2 and 6–7 were combined for classifying reliable vs. not reliable (Panel A, B), not comfortable vs. comfortable (Panel C), not acceptable vs. acceptable (Panel D), low likelihood of misclassification vs. high likelihood of misclassification (Panel E), and not willing vs. willing to give primaquine (Panel F). Stuart-Maxwell test was used to assess the difference in results between pre-training vs. Day 2, pre-training vs. Week 8, and Day 2 vs. Week 8 assessments. There was statistical difference in the responses between pre- and Day 2 follow up for all questions (p < 0.001) and pre- vs. week 8 assessments (p < 0.001), with no significant change observed between Day 2 and week 8 follow up. This highlights the positive impact of training on risk perception and acceptability of PQ, with no negative changes in perceived risk or acceptability observed with continued field experience.