Why do clinical laboratories need quantitative reagents? What can it do?

Although IVD reagents are mainly used in laboratory departments, they may also be used in other departments. Therefore, we use the word "clinical laboratory" to replace the application scenario of IVD reagents. Then we will test the samples in the clinical laboratory and diagnose the patients through the test results, so as to implement clinical management.

An important task in the clinical management of patients is to divide them into different management groups, so that they can be managed uniformly, reduce the corresponding costs, and prevent omissions. For example, we can divide people into 'little yang people' and 'little yin people' because of the COVID-19 pandemic.

However, clinically, whether a person is healthy or not is not a simple binary problem. For example, for the detection of COVID-19 nucleic acid, at the beginning, it will be judged as positive when the initial line is detected, then it will become a CT value less than 40, then it will become a CT value less than 35, and then it will be retested if it is less than 35. This is mainly because many of these patients are just carriers of COVID-19. The higher the viral load in their bodies, the more infectious they will be. But as long as they are below a certain threshold, they will not be infectious.

This situation is actually caused by the duality of in vitro diagnostic system (detection performance and clinical performance).

Therefore, simple binary judgment is difficult to meet the needs of diagnosis, which means that we need to express the patient's state through a series of continuously changing numbers, which is quantitative reagent.

For quantitative reagents, the most important performance is the ability to provide results that are proportional to the measured true value, which is also called linearity.

Why are linear attributes most important for quantification? This is determined by two aspects:

First, if the linearity of the quantitative reagent is not good, then the results of the whole quantitative reagent are incomparable in its own numerical system, which has lost the meaning of quantification itself;

Second, as long as the linearity of the quantitative reagent itself is OK, even if two reagents use different numerical systems (for example, different units, or different calibrators for traceability), we can still solve this problem through different conversion formulas, but other attributes cannot.

In addition, monitoring is another main field of quantitative reagent application, which is also another important task of patient management. Monitoring means that we need to test the same patient at different time points, and at the same time, we need to compare these results.

In this case, we must obtain a numerical value, not just a 'negative' or 'positive' answer. Moreover, these values should reflect the true values measured in the patient samples, so that our comparative study of the test results of the same patient at different time points can truly reflect the possible changes of the patients measured.

Symbol: ValueMP, measured mean value; ValueTRUE, the measured true value.

Figure 1 | Relationship between true value, measured value (represented by circle) and measured mean value

As shown in Figure 1, due to the influence of random error and systematic error, if we make measurements for many times (the measured values are those small blue circles), it is likely that such a normal distribution of the detection results will occur, so the true value (ValueTRUE) is within the range of the normal distribution, but we can also see that it does not overlap the measured mean value (ValueMP), This shows that there is systematic error (of course, it is unlikely that there is no systematic error at all).

For quantitative reagents, the error is clear and controllable. Both systematic error and random error can be reduced to an acceptable range by using certain methods, such as adjusting the coefficient in the formula to reduce the systematic error, and using the mean value to reduce the random error. These methods ensure the establishment of linearity.

In this case, for the target population, the measured value of any sample within a specific interval from the lower limit of the linear range (LLLI) of quantitative reagents to the upper limit of the linear range (ULLI) (we can also express it as the linear range [LLLI, ULLI]) is in direct proportion to the true value.

In this way, the results provided by the quantitative reagent can be used to monitor the change of the measured true value of the patient.

Sometimes, we will see that some manufacturers use absolute quantification as one of their technical selling points when they are making publicity. If you see such publicity, you can directly pat your ass and leave. Companies that can make such publicity are either stupid or bad. In fact, absolute quantification is nothing more than the number of measured molecules directly converted from the measured value of the machine (even small biochemical molecules can also be expressed as how many molecules are contained). Therefore, as long as it is a detection method, it can be said that it is absolute quantification.

However, absolute quantification cannot be used in the detection method (note that the definition of detection method I use here is not even a diagnostic method), because it violates the most basic detection principle, because the final measurement of absolute quantification is the absolute number of samples entering the detection reporting system (we don't consider the error of the detection reporting system itself), and does not include the previous sample processing, The previous sample processing is bound to have losses, and the existence of random errors makes the losses caused by each sample processing different. Therefore, most of the data obtained by the so-called absolute quantification are incomparable.

Of course, it is possible to ensure the availability of absolute quantification by reducing the total error of the entire detection system, but this also means the soaring cost and the substantial increase in requirements for operators, which is not worth the loss in the detection system.

This is why in all current detection systems, quantitative methods are used for relative quantification through comparison rather than absolute quantification.