|
The initial reporter stated they received questionable tina-quant hemoglobin a1cdx gen.3 patient sample results measured in two laboratories on two cobas c 513 analyzers.The first laboratory was evaluating the c 513 hba1c method against their existing biorad d-100 method.For this evaluation, 50 patient samples were initially tested in duplicate, then the samples were tested again after left standing for 3 hours.The following differences were observed in this evaluation: run 1 - when the 50 samples are mixed, then tested on the c 513, the average difference versus the biorad method was 2.4 mmol/mol.Run 2 - when the same 50 samples were re-tested on the c513 immediately after run 1, the average difference versus the biorad method was 2.7 mmol/mol.Run 3 - when the same 50 samples were left to stand for 3 hours and re-tested on the c 513 analyzer, the average difference versus the biorad method was 4.7 mmol/mol.For the evaluation in the first laboratory, an example of data for one patient sample is provided from among the 50.This sample had discrepant hba1c results.The sample was initially tested using the biorad method on three biorad systems, resulting with respective hba1c values of 51 mmol/mol, 49 mmol/mol, and 50 mmol/mol.For run 1, the c 513 result was 54.6 mmol/mol.For run 2, the c 513 result was 55 mmol/mol.For run 3, the c 513 result was 56.5 mmol/mol.In light of these results, a second laboratory carried our an additional experiment on their c 513 analyzer using 40 patient samples across the reporting ranges.For this experiment, there were 2 parts.For part 1 of the second laboratory experiment, the samples were run unmixed on the c 513 system and the results were compared to results reported the previous day from the c 513 system.The results were consistent.For part 2 of the second laboratory experiment, the samples were inverted manually 3 times and tested on the c 513 system.The results were compared to the results reported the previous day from the c 513 system.Based on the data of the second laboratory experiment, there is a clear difference between the unmixed and post-mixed runs.This ranges between 0.6 mmol/mol (lower end) to as much as 5.4 mmol/mol in the higher ranges.No consistent bias was observed across the tested range.An average difference post-mix of 2.76 mmol/mol was observed.This implies a difference on average of 3 mmol/mol.Values between 40 and 50 mmol/mol were affected more in some instances.An example of one patient sample used for the second laboratory experiment showed an hba1c value of 51.1 mmol/l without mixing and a value of 45.9 mmol/mol after inverting the sample 3 times.Both laboratories then commenced patient testing with pre-mixing of all patient samples prior to testing, with the first laboratory starting this on (b)(6) 2025.The first laboratory notified physicians and patients that patient hba1c results between on (b)(6) 2025 were elevated (these were tested on the c 513 system).Approximately 75000 patient results were affected.The first laboratory reviewed the data targeted the specific patient population where patients¿ diagnosis has changed particularly from non-diabetic to diabetic.On (b)(6) 2025, the first laboratory confirmed that the target patient groups are those with a hba1c of 41 to 52 mmol/mol, representing the move from pre diabetic to diabetic, and those with a hba1c from 53 to 65 mmol/mol representing a change in treatment.Patients were excluded if their delta was <5 mmol/mol from the previous results, and if they had been retested since on (b)(6) 2025 when mixing was implemented.It is estimated that approximately 15000 patients may be affected.
|
|
The serial number of the c 513 analyzer at the first laboratory is (b)(6).The serial number of the c 513 analyzer used at the second laboratory was requested, but not provided.The investigation is ongoing.Medwatch field d4: between 31-mar-2025 and 14-apr-2025, hba1c reagent lot: 79405801 was used.Between 14-apr-2025 and 21-may-2025, hba1c reagent lot: 80415401 was used.Between 21-may-2025 and 30-jun-2025, hba1c reagent lot: 80175501 was used.Between 30-jun-2025 and 28-jul-2025, hba1c reagent lot: 82015901 was used.The reagent expiration dates were not provided.
|
|
The roche internal investigation of sedimentation effects for tq hba1c gen.3 on cobas c 513 showed a small positive bias for samples in the prediabetic range up to the diagnostic cutoff.For diabetic samples a small negative bias was observed.When evaluating the data in analogy to ifcc certification and ngsp manufacturer certification and applying the same acceptance criteria, deviations were reliably within the acceptance criteria.Accordingly, the magnitude of the sedimentation effects is within acceptable limits and the tq hba1c gen.3 assay performs within its product specifications.No relevant impact of calibrator and reagent lot combination on the sedimentation effect has been observed in the investigation.Direct comparison of roche investigation data with data generated at the customer sites show a much more pronounced sedimentation effect with a 2-3 fold positive bias compared to the roche investigation.The comparison indicates that the root cause for the increased sedimentation effect compared to the roche investigation is likely to be associated with local factors.Local troubleshooting while identifying and fixing minor observations regarding analyzer condition and adjustments did not reveal a root cause for the increased sedimentation effect and increased positive bias observed at the customer sites.The investigation determined the issue is consistent with customer handling.The variance in hb is most likely down to workflow and that on days when the throughput is higher, samples are standing longer.
|
