Sigma Metric Analysis
Siemens Atellica Hema 580 in South Africa, multimode analysis
Siemens Alinity HQ is one of the newest hematology analyzers out there. How does it stack up when you benchmark it against the latest CLIA goals, EFLM desirable and minimum goals?
Siemens Atellica Hema 580 in South Africa, multimode analysis
Sten Westgard, MS
March 2024
Other articles in the multimode series:
- Beckman Coulter DxC 700
- Abbott Alinity
- Abbott Alinity in Turkey
- Abbott HbA1c in USA
- Siemens Atellica
- Siemens Atellica in Romania
- Siemens Atellica in Spain
- Siemens ADVIA 2120i
- Roche c501 in Turkey
- Roche c501 in Saudi Arabia
- MicroLab RX-50 in India
- Roche cobas 6000 immunoassays in Turkey
- Sysmex XN 350 in India
- Dymind DH 76 in Bulgaria
- Mindray 7500 in China
- Mindray BS 2000M in China
Sigma Score 1.5 out of 6. Misleading [The only positives of this study are the use of patient samples for comparison purposes, and the application of EFLM goals in the assessment.]
This study is not ideal. In fact, there isn't enough here to give a full picture of performance. But it's the first analysis we've seen of a new Atellica hematology instrument. Using manufacturer controls over a short period of time will probably generate imprecision estimates that are too optimistic. Thus, if this study is misleading, it is likely to be misleadingly positive. As we will see below, however, even the most optimistic interpretation of this data is damning.
Performance evaluation of the new automated Atellica Hema 580 hematology analyzer. Schapkatiz E, Baiden A, Raburabu S. Int J Lab Hematol. 2024:46:63-71. DOI: 10.1111/iljh.14170
The Imprecision and bias data
"Imprecision was performed with manufacturer's control material (high, low and normal for blood samples[)]..All levels were processed as five replicates per run over five days."
"Method comparison studies were performed to compare the performance of the Atellica Hema 580 to the Sysmex XN-1000 for FBC, DIFF and Retic in whole blood." The paper doesn't state the exact number of patient samples that were run, but by the graphs appearances, it appears to be a considerable number.
The precision data is presented without noting the level at which each control is being run. In order to calculate the bias at each of those levels, we had to extrapolate the position based on the working range stated in the study. Bias was not calculated at the extreme ends of the range.
Parameter | Level | Slope | Y-intercept | % Bias | CV |
Hemoglobin | 11.3 | 1 | 0.2 | 1.77 | 0.26 |
8 | 1 | 0.2 | 2.50 | 0.86 | |
15 | 1 | 0.2 | 1.33 | 0.27 | |
Hematocrit | 0.34 | 0.96 | 0 | 4.00 | 0.77 |
0.2 | 0.96 | 0 | 4.00 | 0.77 | |
0.45 | 0.96 | 0 | 4.00 | 0.89 | |
MCV | 84.4 | 1.09 | -8.76 | 1.38 | 0.54 |
70 | 1.09 | -8.76 | 3.51 | 0.26 | |
95 | 1.09 | -8.76 | 0.22 | 0.26 | |
MCH | 28.2 | 1 | 1 | 3.55 | 0.48 |
25 | 1 | 1 | 4.00 | 0.58 | |
32.5 | 1 | 1 | 3.08 | 0.69 | |
MCHC | 33.9 | 0.4 | 20.76 | 1.24 | 0.64 |
31 | 0.4 | 20.76 | 6.97 | 0.62 | |
36 | 0.4 | 20.76 | 2.33 | 0.84 | |
Platelets -Impedance | 350.5 | 1.01 | -1.23 | 0.65 | 2.07 |
50 | 1.01 | -1.23 | 1.46 | 4.74 | |
600 | 1.01 | -1.23 | 0.79 | 2.07 | |
Platelets - Optical | 68.5 | 0.98 | -0.05 | 2.07 | 1.96 |
10 | 0.98 | -0.05 | 2.50 | 3.66 | |
120 | 0.98 | -0.05 | 2.04 | 1.58 | |
RBC | 4.155 | 0.93 | 0.23 | 1.46 | 0.5 |
2.5 | 0.93 | 0.23 | 2.20 | 0.67 | |
5.5 | 0.93 | 0.23 | 2.82 | 0.91 | |
WBC | 93.27 | 0.96 | -0.05 | 4.05 | 1.11 |
10 | 0.96 | -0.05 | 4.50 | 2.28 | |
170 | 0.96 | -0.05 | 4.03 | 0.66 | |
Lymphocytes | 2.935 | 0.97 | 0.01 | 2.66 | 3.61 |
1 | 0.97 | 0.01 | 2.00 | 5.23 | |
5 | 0.97 | 0.01 | 2.80 | 4.14 | |
Monocytes | 1.225 | 1.09 | -0.03 | 6.55 | 6.89 |
0.5 | 1.09 | -0.03 | 3.00 | 15.23 | |
2 | 1.09 | -0.03 | 7.50 | 3.89 | |
Eosinophils | 0.57 | 0.97 | 0.01 | 1.25 | 7.98 |
0.25 | 0.97 | 0.01 | 1.00 | 8.34 | |
1 | 0.97 | 0.01 | 2.00 | 4.52 | |
Basophils | 0.05 | 0.75 | -0.01 | 45.00 | 1.67 |
0.02 | 0.75 | -0.01 | 75.00 | 2.14 | |
0.07 | 0.75 | -0.01 | 39.29 | 0.89 | |
Neutrophils | 13.47 | 0.98 | 0.02 | 1.85 | 1.59 |
5 | 0.98 | 0.02 | 1.60 | 3.01 | |
23 | 0.98 | 0.02 | 1.91 | 1.18 |
Sigma-metrics according to EFLM-derived DESIRABLE performance specifications
As set forth by the Milan Heirarchy, performance goals based on biological variation are considered superior to those set by regulators and proficiency testing or external quality assurance programs. However, the hematology goals are known to be very demanding.
So, it's clear now why EFLM no longer recommends using desirable performance specifications. Instead, they recommend the minimum performance specifications. So let's see that.
Sigma-metrics according to EFLM-derived MINIMUM performance specifications
There's a definite improvement, but still there's a vast proportion of the assay performance is below 2 Sigma.
Sigma-metrics according to CLIA 2024 performance specifications
For US labs and labs that are governed by CAP accreditation, there's a respite: the EFLM biological rules aren't mandatory. CLIA 2024 goals are instead. One shortcoming is that the CLIA goals don't cover all the differential parameters, they only cover a few analytes. The 1992 CLIA goals were fairly forgiving, but there's a rude wake up for labs once the new goals kick in.
Here's the most optimistic assessment of the Hema 580. A few of these paramters can hit the bull's-eye when the CLIA 2024 goals imposed. But not all. The CLIA goals are tougher than the EFLM minimums for a few parameters.
Conclusion
The authors end the paper with this assessment: "In conclusion, to the best of our knowledge, for the first time we describe here the performance characteristics of the new Atellica Hema 580 hematology analyzer. The measurement technologies are comparable to the reference Sysmex XN-1000, precise and efficient for FBC, DIFF and Retic measurement in whole blood."
The precision reported here is probably optimistic, yet it does not lead to a favorable assessment. Using the global standards of the EFLM, much of this performance would be judged unacceptable. The CLIA 2024 are the most forgiving, but they do not give an overall clean bill of health to this instrument.
If this is the first look at the Hema 580, there is a great need to get a better look at this instrument.