Point-of-Care CBC Analyzer with 5-Part Differential

In the roughly 500-specimen comparison, all parameters met the pre-defined acceptance criteria except the basophil parameters, whose range in that sample set was limited to roughly 0 to 2%. All-sites correlation coefficients were above 0.9 for WBC, RBC, HGB, HCT, PLT, NEUT# and LYMPH#, with slopes close to 1 (PLT about 1.1). Monocytes correlated less closely (MONO# r of about 0.7, slope of about 0.8), and the 510(k) summary reports that the monocyte correlation met its acceptance limit when flagged results were excluded. The supplemental elevated-basophil study met its pre-defined acceptance criteria, with r of about 0.7 for basophil percentage versus light microscopy and about 0.5 for absolute basophils versus the laboratory analyzer.

Overall WBC flagging against the laboratory analyzer showed PPA of about 95%, NPA of about 82% and overall agreement of about 92%. Against blood-film review, PPA was about 93%, NPA about 71% and overall agreement about 87%.

Capillary and venous results were comparable and met the acceptance criteria. Repeatability and multi-site reproducibility met their pre-defined criteria; for example, total reproducibility CV for WBC was about 8%, 5% and 4% at the low, normal and high control levels, and for HGB about 2% at each level. FDA found the analyzer substantially equivalent and cleared it under 510(k).

Correlation coefficients depend on the spread of the data as much as on agreement. MCHC (r of about 0.7 across a narrow range of roughly 28 to 37 g/dL) and basophils (r near zero for BASO#, over a range from near zero to about 0.2 ×10³/µL) show low r largely because their ranges are narrow. Basophils did not meet the acceptance criteria in the main set for this reason and were evaluated separately in an enriched sample set. The 510(k) summary also reports that for every parameter and both comparators, the 95% confidence intervals of mean bias were within the acceptance limits, a more direct test of agreement than r.

The flagging NPAs (about 82% against the analyzer and about 71% against blood films) mean that a meaningful share of samples the comparator called normal were flagged by the candidate. Sensitivity to abnormal samples was high (PPA above 90% against both comparators). In practice, a lower NPA means more samples sent for smear review, while a high PPA limits missed abnormal samples. FDA's public documents do not publish numeric acceptance criteria for the flagging study, so these figures should be read as the performance FDA accepted for this device rather than a published threshold.

The analyzer is intended for enumeration and classification of WBC, RBC, HGB, HCT, MCV, MCH, MCHC, RDW, PLT, MPV and the absolute counts and percentages of neutrophils, lymphocytes, monocytes, eosinophils and basophils, in capillary and venous K2EDTA whole blood. It is intended for in vitro diagnostic use in clinical laboratories and/or point-of-care settings for adults and children at least 2 years of age, and is labeled for prescription use.

The clinical evidence had to show agreement with a laboratory automated hematology analyzer across the analytical measuring ranges and at medical decision points, using normal and pathological samples tested at point-of-care sites by operators resembling the intended users. Because a 5-part differential is reported, the studies also had to address low-frequency cell types, WBC abnormality flagging against a reference blood-film differential, and the comparability of capillary and venous specimens from the same subject.

Precision also had to hold outside the laboratory. Short-term repeatability needed to include point-of-care operators, and multi-site reproducibility had to cover low, normal and high levels for all 20 parameters.

Automated hematology analyzers that identify formed elements of the blood are classified under 21 CFR 864.5220, automated differential cell counter, as Class II devices with special controls. The regulation names as its special control an FDA guidance on premarket notifications for automated differential cell counters for immature or abnormal blood cells. That guidance matters because these devices may also flag, count or classify immature or abnormal hematopoietic cells.

In practice, the evidence for a CBC with differential is organized around CLSI standards that FDA cites in its decision summaries. CLSI H26 covers validation, verification and quality assurance of automated hematology analyzers: precision, linearity, carryover, method comparison, and reference intervals. CLSI H20 defines the reference leukocyte differential, a manual or digitally assisted count of about 400 cells from duplicate blood films, against which instrument differentials and abnormal-cell flags are judged. EP05, EP06, EP07, EP17 and EP28 address precision, linearity, interference, detection capability and reference intervals.

The category is demanding because one instrument reports many measurands with very different distributions. Rare cell types such as basophils and eosinophils sit near zero in most patients, so ordinary sample sets cannot test them across a meaningful range. Flagging performance depends on morphology review. A point-of-care claim adds capillary samples and non-laboratory operators to the picture.

• Running the candidate at point-of-care sites, with operators whose roles resembled intended users (the 510(k) summary lists phlebotomists, laboratory assistants, nurses and medical assistants for the precision work), tested the claim in the setting where it would be used.
• Deliberately including leukemias, anemias, hemoglobinopathies and other pathological samples stretched results across the analytical measuring ranges and medical decision points. A random outpatient draw would cluster near normal.
• A 400-cell film differential per CLSI H20-A2 gave an independent reference for the differential and for flagging, so the candidate was not judged only against another automated analyzer.
• The separate basophil study responded directly to a known limitation. When the main sample set could not produce a usable basophil range, a targeted set enriched for elevated basophils was analyzed against light microscopy.
• Paired capillary and venous draws from the same donors justified listing both specimen types, even though the main method comparison used residual venous specimens.