Introduction

Since 2015, antiretroviral therapy (ART) eligibility criteria have evolved from CD4-based clinical staging to a “Treat All” approach. This shift expanded treatment eligibility for all HIV-positive individuals, regardless of CD4 count, based on evidence supporting early ART initiation. The “Test and Treat” strategy recommends starting ART within two weeks of HIV diagnosis1,2,3.

The introduction of point-of-care (POC) CD4 testing, particularly the PIMA analyzer, has emerged as a promising solution for addressing the challenges in managing Advanced HIV Disease (AHD) care, especially in resource-limited settings4. Whereas CD4 testing is essential for the timely identification of advanced HIV disease and serves as a critical entry point into advanced HIV disease care pathways; clinical staging has demonstrated significantly low sensitivity for identifying individuals with CD4 counts below 200 cells/µL5.

Current WHO advanced HIV disease guidelines assert that “CD4 cell count testing at baseline for all people living with HIV remains important” and that, to detect advanced HIV disease among ART-experienced individuals, “CD4 cell count testing should be specifically prompted for people with a viral load exceeding 1000 copies/mL and for everyone whose clinical presentation suggests advanced HIV disease regardless of whether they have been exposed to ART or not”6. Since the publication of the WHO advanced HIV disease guidelines in 2017, several significant changes have occurred in the CD4 testing landscape. As the “treat all” approach has been implemented globally, there has been a decline in donor and program support for centralized laboratory CD4 testing infrastructure and commodities. Evidence from several African countries has demonstrated a marked decrease in the proportion of individuals initiating ART who receive baseline CD4 testing7,8,9, thereby complicating the diagnosis of advanced HIV disease.

Advanced HIV Disease remains a significant concern in many parts of the world, particularly in regions with limited healthcare infrastructure5,10. The ability to rapidly and accurately determine CD4 counts is crucial for identifying patients at high risk of opportunistic infections and mortality, as well as for guiding clinical decision-making regarding antiretroviral therapy (ART) initiation and prophylaxis for opportunistic infections.

The PIMA analyzer, a portable and user-friendly device that provides CD4 results within 20 min using a small blood sample. This rapid turnaround time can significantly reduce loss to follow-up and enable immediate clinical interventions. Furthermore, its portability allows for decentralized testing, bringing CD4 count capabilities closer to patients in remote or underserved areas.

Several studies have evaluated the performance of the PIMA analyzer in various settings, demonstrating generally good agreement with conventional laboratory-based CD4 testing methods11,12,13. However, the specific utility of PIMA in the context of AHD care, using the 200 cells/µL threshold, warrants further investigation to ensure its reliability and effectiveness in this critical aspect of HIV management.

By integrating PIMA testing into AHD care protocols, healthcare providers can potentially streamline patient assessment, expedite treatment decisions, and improve overall patient outcomes. This approach aligns with global efforts to reduce HIV related morbidity and mortality by ensuring timely identification and management of individuals with advanced disease. In this study, we determined the sensitivity, specificity, positive and negative predictive values, and likelihood ratios of PIMA against FACSCalibur flow cytometry as the gold standard for monitoring advanced HIV patients with CD4 threshold of ≤ 200 cells/µL using capillary blood.

Methods

A cross-sectional study was conducted between April and July 2015 at Mildmay Uganda, which is a specialist HIV/AIDS care, prevention, and training center located in Lweza along Entebbe Road Uganda. Mildmay Uganda provides care to over 16,000 HIV infected clients with close to 8,000 currently receiving various HAART combinations. We included 110 HIV-positive adult patients seeking care from Mildmay Uganda who were required to determine a CD4 cell count of ≤ 200 cells/µL of blood during the study period and provided informed consent to participate in the study. Individuals who were unable to provide suitable samples for analysis using both methods and those whose results were invalid for at least one of the methods were excluded from the analysis. The sample size was derived from Buderer’s formula based on both specificity and sensitivity14. Age and sex were recorded and CD4 counts were measured and recorded by a well-trained team of laboratory technicians.

Data collection and management

Patient demographic data were obtained from the Hospital Information Management System (HIMS) by pre-trained research assistants using pre-tested study tools, after which they were deleted from the subject by assigning a unique study ID. The results from the analysis of samples from the FACSCalibur were extracted from the hospital laboratory information system by the researcher. PIMA results were recorded in a laboratory notebook by pre-trained research assistants and verified against instrument printouts by the researcher. The results for CD4 counts were linked using a unique identifier, entered into EPI data V 3.1 and exported to Excel. Excel files were imported to the STATA software for analysis.

Laboratory methods

Sample collection

Two blood samples [venous and capillary] were collected from each subject using EDTA evacuated tubes and capillary tubes for venous and capillary blood, respectively, using Mildmay Uganda standard practice as detailed in the Mildmay Uganda sample collection manual.

Sample analysis

Venous blood was analysed using the BD FACSCalibur following the Mildmay Uganda standard operating procedure for the determination of CD4 + cell counts. Briefly, 50 µl of EDTA anti-coagulated blood was stained with 20 µl of BD MultiTest CD3/CD8/CD45/CD4 fluorescence labelled antibody mixture in BD true count tubes for 15 min in the dark at room temperature, and then lysed with BD lysing solution. Analysis was performed on the BD FACSCalibur using the BD MultiTest software, and the results were directly exported into LIMS. The BD FACSCalibur was calibrated using BS calbrite beads (BD Biosciences), and the quality of reagents and the specimen processing procedures were controlled using low and normal controls (BD Biosciences) on everyday samples. Capillary blood was analysed using the PIMA following the Mildmay Uganda standard operating procedure for the use of the PIMA point-of-care CD4 testing device. Briefly, 25 µl of capillary blood was drawn into the sample collector by holding the cartridge at 45° until the cartridge was filled. The filled cartridge was capped and inserted into an Alere Pima Analyzer. The Analyzer automatically started analysing the sample once the cartridge was inserted and displayed absolute CD4 counts after 20 min. The displayed results were printed, filled, and entered into a laboratory information system (LIMS). On each day, the samples were tested, and equipment performance was checked using valid Pima Bead Standards whose results were recorded in the LIMS. Samples were tested only if the results of the Pima Bead Standards passed. All methods were conducted following the relevant guidelines and regulations.

Statistical analysis

To determine the reliability of the PIMA POC in predicting a CD4 cell count of ≤ 200 cells/µL of blood, a classification based on the FACSCalibur results as the reference was used with subjects classified as True Positive (TP), True Negative (TN), False Positive (FP), and false negative (FN). We calculated the performance characteristics as follows: sensitivity (TP÷ (TP + FN) ×100), specificity (TN÷ (TN + FP) × 100), and diagnostic efficiency (TP + TN) ÷ (TP + TN + FN + FP) ×100. Positive and negative likelihood ratios were determined using standard methods. Upward and downward misclassification probabilities were calculated for a cutoff of 200 cells/µL. The upward misclassification probability was defined as one minus the sensitivity (1-Sn), and the downward misclassification probability was defined as one minus the specificity (1-Sp) of the PIMA CD4 analyzer.

Results

A total of 110 participants were enrolled, 69 (62.73%) were females and 41 (37.27%) were males. The median age in the study was 39 years (IQR 31–45). There was no statistically significant difference (p = 0.986) between the mean CD4 count using PIMA and that using BD FACSCalibur (Table 1).

Table 1 Demographic and clinical characteristics of the study participants.
Full size table

Performance characteristics of the PIMA

Using the results from the BD FACSCalibur as the reference, there were 4 false negatives and 2 false positives. At the threshold for advanced HIV Disease care (≤ 200 CD4 cells/µL), the performance characteristics of PIMA were sensitivity (55.56%), specificity (98.02%), and positive and negative predictive values;  71.43% (95% CI: 36.01–91.74%) and 96.12% (95% CI: 92.26–98.09%) respectively. The positive and negative likelihood ratios were 28.06 (95% CI: 6.31–124.66) and 0.45 (95% CI: 0.22–0.94) (Table 2).

Table 2 A 2 × 2 contingency table showing the performance characteristics PIMA POC at the ≤ 200 CD4 cells/µl threshold for HIV using the results of the BD facscalibur as the reference.
Full size table

The Bland-Altman analysis showed that PIMA CD4 analyzer had a slight negative bias of -0.1636 (95% CI − 194.8–194.5) in comparison with BD FACSCalibur (Fig. 1a). However, PIMA CD4 analyzer positively correlated with those of BD FACSCalibur (R2 = 0.8463) (Fig. 1b).

Fig. 1
figure 1

(a) Blant–Altman of FACSCalibur and PIMA for CD4 enumeration. (b) Linear regression analysis between FACSCalibur and PIMA.

Full size image

Discussion

In 2015, CD4 count was crucial for ART eligibility and HIV monitoring in Uganda, with a significant prevalence of advanced HIV disease at ART initiation2,15. The 2017 Universal Test and Treat policy shifted this, initiating ART regardless of CD4 count16. However, CD4 count remains important for identifying and managing advanced HIV disease, especially in late-presenting patients2,17. It’s diagnostic decisions for opportunistic infections remains crucial for initial clinical management where viral load testing is limited.

This study showed the performance characteristics of the PIMA CD4 analyzer using capillary blood compared to venous blood CD4 values determined using flow cytometry in a Ugandan. At the threshold of 200 CD4cells/µL, PIMA capillary blood had a sensitivity of 55.56% (95%CI 21.20–86.30%), specificity of 93.03% (95%CI 93.03–99.76%), NPV of 96.12% (95% CI 92.26–98.09%), and PPV of 71.43% (95% CI 36.01–91.74). The positive and negative likelihood ratios were 28.06 (95% CI 6.31-124.66) and 0.45 (95% CI 0.22–0.94) respectively. The sensitivity of PIMA at the threshold of ≤ 200 CD4 cells/µL has been reported as 96% (CI 95.2–96.9%) from a meta-analysis and 96.1% (94.4–97.8%) from a study in Uganda12,18,19,21. The low sensitivity of 55.56% (95%CI 21.20–86.30%) in our study could be attributed to differences in the type of samples and study subjects, as previous studies included ART-experienced subjects and used venous blood for comparison20,22,23.

Our study reported a significantly lower sensitivity of 55.56% at the 200 cells/µL threshold, which is considerably lower than the 96% sensitivity reported in a meta-analysis and the 96.1% sensitivity found in another Ugandan study12,18,19,21. The study found a high specificity of 93.03%, which is generally consistent with other studies, indicating good performance in correctly identifying individuals above the threshold. The Positive Predictive Value (PPV) of 71.43% suggests a moderate ability to correctly identify true positive cases, though this may be influenced by the prevalence of low CD4 counts in the population studied. The Negative Predictive Value (NPV) of 96.12% indicates a strong ability to correctly identify true negative cases, which is important for ruling out low CD4 counts. Likelihood Ratios: The positive likelihood ratio of 28.06 suggests that a positive test result is strongly indicative of a true low CD4 count. However, the negative likelihood ratio of 0.45 indicates a moderate ability to rule out low CD4 counts.

The discrepancy in sensitivity between this study and previous reports warrants further investigation. Possible factors contributing to this difference could be attributed to small sample size and population characteristics, operator training and experience, environmental factors affecting the PIMA analyser performance and differences in blood sample handling and processing.

Limitations

This study was conducted in an urban setting with well-trained laboratory technologists under ideal laboratory conditions. Thus, the results obtained may not represent performance characteristics in rural settings that may not have facilities and skilled labor, yet this is where the use point of care devices is needed. The absence of published real-world field data testing conditions and likelihood ratios limited our ability to compare findings with those of other studies.

Conclusion

This study provides valuable insights into the performance of the PIMA analyser for point-of-care CD4 testing in identifying Advanced HIV Disease patients in Uganda. While the analyser demonstrated lower sensitivity compared to previous research, the comprehensive statistical analysis offers important data for ongoing discussions about POC CD4 testing in AHD management. These findings contribute to the broader understanding of diagnostic tool evaluation in resource-limited settings and may inform future policy decisions regarding the implementation of such technologies in similar contexts.

Recommendations

Further research is warranted to address the limitations identified and to continue improving point-of-care diagnostics for HIV management in resource-constrained environments.