**3. Results**

With the LIAISON ® BRAHMS PCT ® II GEN method, 52% of the results were between 0.0 and 0.5 ng/mL, 18% between 0.5 and 2.0 ng/mL, and 30% between 2.0 and 100 ng/mL, and the maximum and minimum values were 97.2 and 0.02 ng/mL, respectively. The mean and median values were 4.09 ng/mL and 0.456 ng/mL, respectively.

With the Brahms PCT sensitive Kryptor method instead, 55% of the results were between 0.0 and 0.5 ng/mL, 21% between 0.5 and 2.0 ng/mL, and 24% between 2.0 and 100 ng/mL, and the maximum and minimum values were 103 ng/mL and 0.01 ng/mL, respectively. The mean and median values were 3.72 and 0.39 ng/mL, respectively.

There are no significant di fferences between the mean and the median of the results obtained with the two methods (mean 3.717 for Kryptor and 4.094 for LIAISON ®; median 0.39 for Kryptor and 0.45 for LIAISON ®), with the distribution of values following normality (Normal distribution: <0.001) (Figure 2A).

**Figure 2.** Analysis results. ( **A**) Summary table of our data: mean (95% CI), median (95% CI), minimum and maximum of the two methods; (**B**) Linear regression line independent of the sample distribution Passing–Bablok (PB): *y* = −0.0102086 + 1.172143*x* (Pearson coe fficient = 0.99); ( **C**) PB residual chart; (**D**) Comparison of Bland–Altman graph with representation of the di fference, in terms of absolute value, between the two measurements shown according to the mean of the measurements; (**E**) Mountain plot graph or representation of the empirically folded cumulative distribution, obtained by calculating the distribution of percentiles relative to the di fferences between the two methods placed in an orderly manner.

The analysis shows an excellent correlation between the results obtained on the LIAISON ®XL analyzer and on the Brahms Kryptor analyzer, with the respective reagents, both in terms of slope and intercept (Pearson coe fficient: 0.99) (Figure 2B).

The regression model is further validated by the residual analysis (Figure 2C).

The Bland–Altman graph obtained by comparing the measurement di fferences between the two methods as a function of the average of the measurements shows a reduced dispersion of values around the mean and within the standard deviations, without significant di fferences (Figure 2D). The absence of significant di fferences is also reflected in the mountain plot graph (Figure 2E), where the graph is in fact centered on zero, missing values along the two tails.

In the literature, there is growing evidence about the fact that the determination of PCT in the monitoring of antibiotic therapy is significantly useful [6]. In this context, the PCT assay, once the appropriate antibiotic therapy has been set, allows the clinician to highlight the patient's e ffective response to the antibiotic, thus reducing the risk of unnecessarily prolonged treatments, the development of resistance, and other side e ffects.

In this regard, we have followed the course of PCT concentrations in some patients undergoing antibiotic therapy over time.

These patients, coming respectively from the emergency medicine department (patient 1, man 78 years old), a long-term care setting (patient 2, man 83 years old), and medical oncology (patient 3, woman 55 years old), were part of the population used for comparing the two methods.

The PCT determination was performed with the two assay methods throughout the antibiotic administration period, obtaining the following results.

Figures 3–5 show the trend of PCT concentrations in the three patients (PCT LIAISON ® blue line; Kryptor orange line). In all three cases, the PCT values measured with LIAISON ® and Kryptor are completely similar to each other, as shown by the overlap of the graphs. Furthermore, it is possible to observe a clear decrease in PCT values over time, in response to antibiotic therapy.

**Figure 3.** Patient 1 (Internal medicine, man 78 years old).

**Figure 4.** Patient 2 (Long-term hospitalization, man 83 years old).

**Figure 5.** Patient 3 (Medical oncology, woman 55 years old).
