Annals of Global Health (Feb 2023)

Point-of-Care Testing in Chronic Kidney Disease of Non-Traditional Origin: Considerations for Clinical, Epidemiological, and Health Surveillance Research and Practice

  • Miranda Dally,
  • Juan José Amador,
  • Jaime Butler-Dawson,
  • Damaris Lopez-Pilarte,
  • Alexandra Gero,
  • Lyndsay Krisher,
  • Alex Cruz,
  • Daniel Pilloni,
  • Joseph Kupferman,
  • David J. Friedman,
  • Benjamin R. Griffin,
  • Lee S. Newman,
  • Daniel R. Brooks

DOI
https://doi.org/10.5334/aogh.3884
Journal volume & issue
Vol. 89, no. 1

Abstract

Read online

Purpose: As the prevalence of chronic kidney disease of non-traditional origin (CKDnt) rises in low-resource settings, there is a need for reliable point-of-care creatinine testing. The purpose of this analysis was to assess the accuracy of two commonly used point-of-care creatinine devices, the i-STAT handheld (Abbott, Princeton, NJ, USA) and the StatSensor Creatinine (Nova Biomedical, Waltham, MA, USA) in comparison to venipuncture serum creatinine measures. The affordability, sensitivity, specificity, ease of use, and other considerations for each device are also presented. Methods: Three paired data sets were compared. We collected 213 paired i-STAT and venipuncture samples from a community study in Nicaragua in 2015–2016. We also collected 267 paired StatSensor Creatinine and venipuncture samples, including 158 from a community setting in Nicaragua in 2014–2015 and 109 from a Guatemala sugarcane worker cohort in 2017–2018. Pearson correlation coefficients, Bland-Altman plots, and no intercept linear regression models were used to assess agreement between point-of-care devices and blood samples. Results: The i-STAT performed the most accurately, overestimating creatinine by 0.07 mg/dL (95% CI: 0.02, 0.12) with no evidence of proportional bias. The StatSensor Creatinine performed well at low levels of creatinine (Mean (SD): 0.87 (0.19)). Due to proportional bias, the StatSensor Creatinine performed worse in the Nicaragua community setting where creatinine values ranged from 0.31 to 7.04 mg/dL. Discussion: Both devices provide acceptable sensitivity and specificity. Although adequate for routine surveillance, StatSensor Creatinine is less accurate as the values of measured creatinine increase, a consideration when using the point-of-care device for screening individuals at risk for CKDnt. Research, clinical, and screening objectives, cost, ease of use, and background prevalence of disease must all be carefully considered when selecting a point-of-care creatinine device. Conclusion: POC testing can be more accessible in resource-limited settings. The selection of the appropriate device will depend on the use-case.

Keywords