Assessment of ENI Gemeni microprocessor-controlled centrifugal analyser

for each step of the analysis cycle. At the end of each step the external clock is reset as is the 8 bit control pattern for the next function. The final step is to activate the AA analysis cycle and to wait for the furnace to cool before injection of the next sample. We have found that programming a waiting time longer than the cooling time extends the life of the graphite cuvettes. The program for operation in the pulsed mode is given in Table 1. The second injection is a critical part of the analysis cycle for non-metals such as selenium and arsenic. The addition of a co-analyte such as Ni2+ has been shown to increase the sensitivity for these elements and reduces loss due to the volatility of these elements 1, 5]. The syringe pump used for the second injection delivers 11.0 + 0.1 /.tl/sec of coanalyte. The same syringe pump is used for the total consumption analysis mode. In the total consumption mode, however, the analyte is pumped out of the holding tube into the sampling valve and the dispensing of the sample then proceeds in the same fashion as in the pulsed mode of operation. The software modification for total consumption analysis involves only increasing the second injection time to 4-8 seconds and performing the second injection prior to the sample injection. The graphite cuvette volume in the currently employed AA system is approximately 50 /al and the standard pulsed mode experiment employs 37/al + 1.7 ktl analyte from the sampling loop and 1.0/al + 0.1 /al of co-analyte from the syringe pump. The total consumption mode pumps 100 of eluent into the sampling loop and 37/.tl are then dispensed into the graphite furnace. This ’overrun’ ensures the complete filling of the sample loop and causes no problem in subsequent interpretation of data. Conclusion The use of an inexpensive microprocessor system adds a great deal of versatility to a previously "hard wired" LCAA sampling system [1]. Both pulsed and total consumption analyses are possible with only minor plumbing and software changes. The use of other microprocessor systems would require only a change in machine language. The system has the advantage of being inexpensive. The interface sampling system can be assembled for a component cost of approximately $500 (depending on the availability of surplus equipment). With the many advantages of performing LCAA analysis on trace level metal-containing compounds, hopefully this technique will find widespread use with investigators in the clinical, environmental, and inorganic biochemistry fields.

the sampling valve and the dispensing of the sample then proceeds in the same fashion as in the pulsed mode of operation. The software modification for total consumption analysis involves only increasing the second injection time to 4-8 seconds and performing the second injection prior to the sample injection.
The graphite cuvette volume in the currently employed

Conclusion
The use of an inexpensive microprocessor system adds a great deal of versatility to a previously "hard wired" LCAA sampling system [1]. Both pulsed and total consumption analyses are possible with only minor plumbing and software changes. The use of other microprocessor systems would require only a change in machine language. The system has the advantage of being inexpensive. The interface sampling system can be assembled for a component cost of approximately $500 (depending on the availability of surplus equipment).
With the many advantages of performing LCAA analysis on trace level metal-containing compounds, hopefully this technique will find widespread use with investigators in the clinical, environmental, and inorganic biochemistry fields. [

Materials and Methods
The Gemeni is a miniature centrifugal analyser consisting of an analyser module, microprocessor and work station. The methods recommended by the manufacturer for use on the Gemeni were run in parallel with routine methods used in this laboratory. These routine methods were:  [5] and modified in the light of our own experience.  (b) Interbatch imprecision was assessed by analysing control materials at three levels, twice a day for a period of ten days. The materials were selected to coincide with decision levels, and to cover the linear range of the analyte to be determined.

Recovery
Constituents were added to pooled sera so that the final concentrations represented typical ranges of normal and abnormal values encountered in patient sera.

Accuracy
Bias was checked by comparing patient samples against a reference method or one of known bias. These were used in   [7].

Miscellaneous tests
The precision and accuracy of the reagent dispenser was checked.
The precision and accuracy of delivering a pre-set volume of reagent was determined at two volume settings. This was achieved by dispensing set volumes of distilled water into a weighing tray and weighing the amount of water dispensed.
Several days were allowed in which to become familiar with the instrument prior to the commencement of the evaluation. The familiarisation protocol consisted of the   Recovery Table 3 shows the recovery data for glucose determination.
Recovery data were made for glucose only for two reasons. The report is primarily an assessment of the instrument, and not the kits, and the instrument had to be returned to the distributor before a 'complete evaluation' could be completed.
The recoveries were acceptable by Logan's criteria, i.e. acceptable recoveries average between 90% and 110% with no single recovery less than 85% or greater than 115%.

Accuracy (a) Patient comparison
The computer plot of the patient comparison data is shown in Figures 1, 2, 3 and 4. Applying the "t" test, all the Gemeni results except alkaline phosphatase are significantly different from the comparative methods.
Results of the "t" test were as follows: glucose 3.30, cholesterol 11.00, calcium 3.68 and alkaline phosphatase 1.5. This must be viewed in its proper context; a result may be statistically significant but clinically acceptable.
If the "t" test is significant then further criteria are applied, that is a kit is unacceptable if: (1) The bias exceeds-the allowable limits of error for precision.
(2) The number of false clinical decisions exceeds 5%.   Based on these criteria of acceptability, the results for glucose, cholesterol and alkaline phosphatase are acceptable, while those obtained for calcium.are unacceptable.
(b) Comparison of results on commercial control material Bias less than allowable limits of error for precision from a carefully verified reference serum value which relates to all parameters tested, indicates that the test method gives a satisfactory measurement of the "true value".
Results obtained for control sera are shown in Tables 4, 5, 6 and 7. These indicate that accuracy is acceptable for glucose, cholesterol and alkaline phosphatase but unacceptable for the calcium methodology.
Linearity Linearity for the methods tested glucose linear to 20 mmol/1 cholesterol linear to 15 mmol/1 alkaline phosphatase linear to 400 U/1 calcium linear to 3.75 mmol/1 was as follows: Literature score One mark was assigned for each point covered in the information supplied by the manufacturer. One half mark was awarded for each point that had received only partial attention. In case of doubt the scoring favoured the manufacturer.
The Gemeni's literature scored 13 out of a possible 15 points.

Miscellaneous tests
Precision and accuracy of Clay.Adams Selectapette.
Results obtained were as follows: (a) Volume pre-set to deliver 1.0 ml Mean volume delivered A feature of the instrument which will aid quality control programmes is the print out of a 'quality control factor' and a reagent drift at the end of each run. A draw back with the instrument is its lack of flexibility. The microprocessor program is selected by coded holes in the test card. This makes it impossible to alter parameters in the program and, therefore, difficult to use kits produced by other manufacturers.
The instrument performed creditably throughout the evaluation and proved to be easy to operate, giving precise and accurate results for cholesterol glucose and alkaline phosphatase. However the quality of results from the calcium kits were unacceptable. When the results for the other analytes are considered this would appear to be a fault of the method rather than the instrument.