(245 days)
Diazyme Direct HbA 1c Assay (Enzymatic, On-Board Lysis) test kit is intended for use in the quantitative determination of stable HbA in venous whole blood samples with on-board blood lysis application in a clinical laboratory. This test is not to be used to diagnose or screen for diabetes. The measurement of HbA 1c concentration is for use in monitoring longterm glucose control of persons with diabetes. For in-vitro diagnostic use only.
Diazyme Direct HbA1c Assay Calibrator Set is intended to be used for calibration of Diazyme Direct HbA1c Assay (Enzymatic, On-Board Lysis). For in-vitro diagnostic use only.
Diazyme Direct HbA1c Assay Control Set is intended to be used for quality control by monitoring accuracy and precision of Diazyme Direct HbA1c Assay (Enzymatic, On-Board Lysis). For in-vitro diagnostic use only.
The Diazyme Direct HbA1c assay kit (Enzymatic, On-Board Lysis) is comprised of a Reagent 1, Reagent 2, Lysis Buffer.
Diazyme Direct HbA1c Assay Calibrator Set: 2 levels, human whole blood based in lyophilized form. It is intended to be used for calibration of Diazyme Direct HbA1c Assay (Enzymatic, On-Board Lysis).
Diazyme Direct HbA1c Assay Control Set: 2 levels, human whole blood based in lyophilized form. It is intended to be used for quality control by monitoring accuracy and precision of Diazyme Direct HbA1c Assay (Enzymatic, On-Board Lysis).
The verification and/or validation studies were performed on the automated chemistry analyzer Roche Modular P (K953239/005). Modular P analyzer is a high through-put automated chemistry analyzer and that can process 800 samples per hour. The predicate performance studies at submission were performed on Hitachi 917 which is an obsolete member of the Roche Mod P Roche family of instruments. The Control Unit of the Roche Modular P analyzer uses a graphical interface to control all instrument functions. The computer, keyboard, and touchscreen monitor allows users to navigate through the software, enter assay, calibrator, and control information, and make test selections. The Diazyme Direct HbA1c Assay Modular P application parameters provided are programmed into the Modular P analyzers. The reagents, calibrators and controls are loaded into the analyzer. The Roche Modular P stores the Diazyme Direct HbA1c Assay reagents in a refrigerated compartment. Reagent and sample pipettes automatically aspirate and dispense specified amounts of reagents or calibrators, controls, and samples into reaction cells. The change in absorbance is measured at specified wavelengths.
The document describes the Diazyme Direct HbA1c Assay (Enzymatic, On-Board Lysis) and its performance, comparing it to an earlier version of the device (K070734) and a predicate device (Tosoh G8 HPLC HbA1c method).
Here's an analysis of the acceptance criteria and the study proving the device meets them:
1. Table of Acceptance Criteria and Reported Device Performance
The document doesn't explicitly state "acceptance criteria" for all performance metrics in a single table, but the conclusion for each study indicates whether the criteria were met. Based on the "Conclusion" sections for Precision and Method Comparison, the implicit acceptance criteria appear to be:
| Performance Metric | Acceptance Criteria (Implicit) | Reported Device Performance |
|---|---|---|
| Internal Precision (CV%) | Less than 2% for within-run, between-run, between-day, between-lot, and total CV%. | Sample 1 (4.64%): Total CV: 1.7% |
| Sample 2 (5.36%): Total CV: 1.2% | ||
| Sample 3 (7.51%): Total CV: 0.9% | ||
| Sample 4 (9.61%): Total CV: 0.9% | ||
| Sample 5 (11.89%): Total CV: 1.0% | ||
| Con 1 Lot 1 (6.22%): Total CV: 1.0% | ||
| Con 2 Lot 1 (9.47%): Total CV: 0.8% | ||
| Con 1 Lot 2 (5.70%): Total CV: 1.1% | ||
| Con 2 Lot 2 (9.11%): Total CV: 0.7% | ||
| Con 1 Lot 3 (6.04%): Total CV: 1.2% | ||
| Con 2 Lot 3 (9.68%): Total CV: 0.7% | ||
| Cal 1 Lot 1 (6.22%): Total CV: 0.9% | ||
| Cal 2 Lot 1 (12.30%): Total CV: 0.7% | ||
| Cal 1 Lot 2 (5.68%): Total CV: 1.2% | ||
| Cal 2 Lot 2 (9.70%): Total CV: 0.8% | ||
| Cal 1 Lot 3 (6.04%): Total CV: 1.1% | ||
| Cal 2 Lot 3 (11.30%): Total CV: 0.6% | ||
| *(All met the 0.98) |
- Slope: Close to 1.0
- Intercept: Close to 0.0 | Combined (Linear Regression):
- Slope: 1.023
- Intercept: -0.090
- Correlation Coefficient (R): 0.9937
Combined (Deming Regression): - Slope: 1.030
- Intercept: -0.140
- Correlation Coefficient (R): 0.9937
(The conclusion states the results met acceptance criteria.) |
| Analytical Specificity (Interference) | 10%) may result in inaccurate values). |
2. Sample Size Used for the Test Set and the Data Provenance
-
Internal Precision:
- Sample Size: 5 unaltered whole blood specimens (4.6%, 5.4%, 7.5%, 9.7%, 11.9% HbA1c), 3 lots of calibrators (2 levels each), and 3 lots of controls (2 levels each). Each sample/control/calibrator was tested 2 runs per day, 2 replicates per run over 20 working days. This results in N=240 data points per sample/control/calibrator (2 replicates * 2 runs/day * 20 days * 3 lots).
- Data Provenance: The study was performed at Diazyme Laboratories. The whole blood samples were "IRB approved." Specific country of origin is not mentioned but Diazyme Laboratories is located in Poway, CA, USA. This appears to be prospective data collection, specifically for this study.
-
Multiple Site Precision:
- Sample Size: Same 5 whole blood samples (4.6%, 5.4%, 7.5%, 9.7%, 11.9% HbA1c), 3 lots of calibrators and 3 lots of controls. Tested in duplicates per run, 2 runs per day for 5 working days with the same lot of reagent. N=60 data points per sample/control/calibrator (2 replicates * 2 runs/day * 5 days * 3 sites).
- Data Provenance: Performed at Diazyme (internal site) and two external sites. Specific country of origin for external sites is not mentioned, but the overall study is assumed to be US-based given the FDA submission. This is prospective data collection.
-
LOD/LOB/LOQ: Based on three lots of reagents and analysis using CLSI EP17-A2, but specific sample numbers are not detailed beyond "three lots of reagents" tested with "5 concentrations" for LOQ.
-
Linearity:
- Sample Size: 11 levels of whole blood linearity set, prepared by dilution of two whole blood samples (Low 3.8%, High 12.3% HbA1c). Tested in triplicate.
- Data Provenance: Not explicitly stated, but likely from Diazyme Laboratories and prospective.
-
Method Comparison:
- Sample Size:
- Site 1: 124 unique K2 EDTA whole blood samples
- Site 2: 132 unique K2 EDTA whole blood samples
- Site 3: 120 unique K2 EDTA whole blood samples
- Total unique samples: 376 (124+132+120).
- Data Provenance: Not explicitly stated, but samples were "IRB approved" K2 EDTA whole blood samples, suggesting human subject data. Conducted at three different sites, likely US-based, and prospective for the purpose of this study.
- Sample Size:
-
Analytical Specificity (Interference): Three whole blood samples ("low", "medium", and "high" HbA1c concentrations), spiked with interfering substances and tested in triplicates. Data provenance not explicitly stated but likely internal and prospective.
3. Number of Experts Used to Establish the Ground Truth for the Test Set and the Qualifications of Those Experts
This information is not applicable in the context of this device. The device is an in-vitro diagnostic assay for measuring HbA1c concentration. The "ground truth" for these studies is typically established by:
- Reference Methods: Such as the Tosoh G8 HPLC HbA1c method, which is an NGSP certification method.
- Known Concentrations: For controls, calibrators, and linearity samples.
- Clinical Laboratory Standards Institute (CLSI) guidelines: For study design and statistical analysis.
There are no human experts subjectively interpreting results or establishing a ground truth in the way a radiologist would for an imaging device.
4. Adjudication Method for the Test Set
Not applicable. As described above, the ground truth is established by reference methods and known concentrations, not by human adjudication of qualitative results. The studies involve quantitative measurements and statistical comparisons.
5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study Was Done, If So, What Was the Effect Size of How Much Human Readers Improve with AI vs without AI Assistance
Not applicable. This device is an automated in-vitro diagnostic assay, not an AI-assisted diagnostic tool that aids human readers in making a diagnosis. There is no human-in-the-loop performance component or "human readers" in the context of this specific regulatory submission.
6. If a Standalone (i.e. algorithm only without human-in-the loop performance) Was Done
Yes, the studies presented (Precision, LOD/LOB/LOQ, Linearity, Method Comparison, Analytical Specificity) represent the standalone performance of the Diazyme Direct HbA1c Assay (Enzymatic, On-Board Lysis) on the Roche Modular P analyzer. The device operates automatically without human intervention during the measurement process, aside from loading samples and reagents.
7. The Type of Ground Truth Used (expert consensus, pathology, outcomes data, etc.)
The ground truth for the performance studies was established using:
- Reference Method: For the method comparison study, the Tosoh G8 HPLC HbA1c method was used as the comparator. This method is noted as an NGSP certification method used in NGSP reference laboratories, making it a highly reliable and recognized standard for HbA1c measurement.
- Known Concentrations: For precision, LOD/LOQ, and linearity studies, samples, calibrators, and controls were prepared with known or target HbA1c concentrations. The calibrators are traceable to NGSP/DCCT and IFCC.
- CLSI Guidelines: The studies followed established CLSI guidelines (EP5-A2, EP17-A2, EP6-A, EP9-A2), which specify rigorous methods for determining performance characteristics against accepted laboratory standards.
8. The Sample Size for the Training Set
The document does not explicitly mention a "training set" in the context of a machine learning or AI model development. This device is a chemical assay, not an AI algorithm that requires training data in the traditional sense. The data provided describes the validation and verification of the assay's analytical performance.
9. How the Ground Truth for the Training Set Was Established
As there is no "training set" for an AI model in this context, this question is not applicable. The device's performance characteristics are validated against established analytical methods and known standards (as described in point 7).
§ 864.7470 Glycosylated hemoglobin assay.
(a)
Identification. A glycosylated hemoglobin assay is a device used to measure the glycosylated hemoglobins (A1a , A1b , and A1c ) in a patient's blood by a column chromatographic procedure. Measurement of glycosylated hemoglobin is used to assess the level of control of a patient's diabetes and to determine the proper insulin dosage for a patient. Elevated levels of glycosylated hemoglobin indicate uncontrolled diabetes in a patient.(b)
Classification. Class II (performance standards).