Procalcitonin (PCT) is an acute-phase protein released into the bloodstream mainly in response to bacterial infection.1–4 Its sensitivity and specificity for bacterial infections is around 0.77 and 0.79, respectively, with an area under the curve (AUC) that has been reported as high as 0.85.5 However, serum PCT concentrations can be increased during inflammatory conditions in the absence of bacterial infection.6 PCT is eliminated by kidney clearance, thus, high concentrations of PCT in patients with chronic or acute kidney dysfunctions are frequently interpreted as non-bacterial infection related elevations.2–7

Several studies have found increased PCT levels in patients presenting with chronic kidney disease, peritoneal dialysis and oliguria in the absence of bacterial infections.8–13 A recent meta-analysis,14 including studies with high heterogeneity, reported that in patients with chronic kidney disease (CKD), PCT might have poor sensitivity (73%) but acceptable specificity (88%) for the diagnosis of bacterial infection. Nevertheless, the best cut-off value of PCT in patients with acute kidney injury (AKI) remains unknown and the real relation among creatinine/urea and PCT concentration are not clear. Moreover, the underlying mechanisms of PCT elevation in these patients are not fully understood.

Our hypothesis is that low concentrations of PCT are accurate in order to rule out bacterial coinfection in patients with influenza pneumonia admitted to the intensive care unit (ICU), but higher PCT levels should be carefully interpreted in patients with acute kidney injury (AKI). In this context, the primary goal of this study was to determine the relation between AKI (using creatinine and urea as surrogate markers of AKI) and PCT concentrations in critically ill patients with influenza pneumonia without bacterial co-infection.

Two thousand one hundred and thirty-two patients with PCR-confirmed influenza pneumonia were admitted to the 148 ICUs. PCT was measured in 972 (45.6%) patients at the time of ICU admission. A total of 633 (65%) patients were included in the study (Fig. 1). Among them, 571 (90.2%) did not develop AKI, whereas 52 (8.2%) and 10 (1.6%) were categorized as AKI I and II, respectively. Further details regarding renal function and clinical characteristics are presented in Table 1 and Table 1e (electronic material). PCT concentrations were higher in patients with AKI (2.62ng/mL [IQR=0.6–10.]) than in non-AKI patients (0.40ng/mL [IQR=0.13–1.20], p=0.002). AKI patients were older, had more major comorbidities, and were more severely ill, and had higher white cell counts upon ICU admission than those without AKI (Table 1e).

Figure 1.

Flow chart of patients (PCT: procalcitonin; ICU: intensive care unit; CARC: community-acquired respiratory coinfection; CKD: chronic kidney disease; RRT: renal replacement techniques; AKI: acute kidney injury).

(0.21MB).

Renal dysfunction and PCT relationship

Patients had increased serum creatinine and urea levels according to PCT cut-off values (A/B and Table 2e [electronic material]). In overall population, we observed a significant, but weak correlation between both creatinine (rho=0.18; p<0.001, Fig. 3) and urea (rho=0.19; p=0.001, Fig. 4) levels and PCT concentration upon ICU admission.

Figure 2.

A/B: Boxplots showing the relationship between creatinine (A) and urea (B) levels according PCT cut-off.

(0.24MB).

Figure 3.

Scatterplot of creatinine (x axis) vs. procalcitonin (PCT). Each point in the scatterplot represents the value of two variables for a given observation. The low rank Spearman correlation coefficient (rho=0.18) confirms that Cr and PCT are not correlated strongly.

(0.34MB).

Figure 4.

Scatterplot of urea (x axis) vs. procalcitonin (PCT). Each point in the scatterplot represents the value of two variables for a given observation. The low rank Spearman correlation coefficient (rho=0.19) confirms that urea and PCT are not correlated strongly.

(0.36MB).

The main conclusion of our study is that the contribution of renal dysfunction (assessed by creatinine and urea levels) on increased PCT concentration is poor. Although PCT values were increased in patients with AKI and creatinine, and urea correlated significantly with PCT, this correlation was very weak. Therefore, the relative contribution of renal dysfunction to PCT values is not greater than 5% in critically ill patients with influenza infection.

Very little is known about the elimination mechanisms of PCT; however, it is likely that renal function interferes with PCT values. Herget-Rosenthal et al.13 observed that serum PCT progressively increased in parallel to the deterioration of renal function. Importantly, although elevated, PCT levels in patients with chronic kidney disease stages I–IV did not markedly differ from those in healthy volunteers. Only in patients with chronic kidney disease stage V and those undergoing peritoneal dialysis, PCT concentrations were significantly higher when compared to healthy controls or patients with stages I–IV. The authors suggested that the progression of renal dysfunction and reduction in urine output could reduce the renal elimination of PCT. Dahaba et al.24 observed in non-septic uremic patients a mean PCT plasma concentration of 1.82 (0.39)ng/mL prior to the start of the first haemodialysis session. The authors suggested that uremia “per-se” is the principal origin of the increased concentrations of PCT; due to increased levels of serum cytokines. Importantly, this PCT concentration was lower than identified in our study, but our patients have a viral infection and the median value of PCT in patients without AKI was 0.4ng/mL as previously reported in another study.25 In contrast, Meisner et al.26 reported that renal elimination of PCT was not the mainstay mechanisms of PCT elimination.

PCT synthesis has been identified in various tissues and cell types such as leukocytes, specifically in peripheral blood mononuclear cells.27,28 Activated peripheral blood mononuclear cells and pro-inflammatory cytokines, such as tumor necrosis factor alpha (TNF)-α and interleukin 1 (IL)-1, are considered to play a key role in the inflammatory process and were also shown to induce PCT production.27 Some authors have observed a strong correlation between serum PCT and PCT released by peripheral blood mononuclear cells cultured in vitro28: we did not observe any correlation between white cell counts and PCT levels (rho=0.03; p=0.45 data not shown). It is likely that the increase of pro-inflammatory cytokines may be the link between inflammation and elevated PCT levels in patients with chronic kidney disease, or those receiving renal replacement therapy. However, the role of this pathway in AKI requires further characterization.

According to our results, we can expect predicted PCT values of 1.06ng/mL, 1.67ng/mL, 2.28ng/mL and 2.89ng/mL for creatinine values of 1, 2, 3, and 4mg/dL, respectively. Although we observed a median value of PCT (1.93ng/mL) similar to predicted PCT (1.67ng/mL) in AKI I patients, the median value of PCT in AKI II patients (5.40ng/mL) was higher than predicted PCT (2.28–2.89ng/mL) by creatinine concentration. We cannot explain this finding, but we might hypothesize that unknown variables other than renal dysfunction, such white blood cells, severity of illness, chronic heart disease or presence of bacterial infection, might play a pivotal role in PCT release/elimination. Although the relative contribution of renal dysfunction to PCT levels is low according to our data, the use of a different PCT cut-off, adjusted to the level of renal dysfunction could increase its diagnostic utility when assessing the probability of bacterial coinfection in this special population of patients with influenza and AKI.

Some limitations in the design of our study must be acknowledged. First, we only assessed renal function using creatinine and urea. Even though these biomarkers are well accepted subrogates of renal function, international consensus recommends using urinary outputs,20 unfortunately, these data was not collected during this study. Second, we examined the PCT levels only upon ICU admission. However, our main goal was to evaluate the relative contribution of AKI to basal PCT levels and not changes over time. Third, we excluded patients with AKI stage III and those who receive continuous renal replacement therapy because the PCT levels in these patients are difficult to interpret due to PCT release associated with microinflammation and/or blood contact with biocompatible membranes.3,4,8,22,24 Fourth, bacterial co-infection was ruled out by microbiology testing in respiratory samples and urinary antigens in patients under mechanical ventilation. However, we cannot certainly rule out bacterial coinfection in non-ventilated patient with only negative urinary antigens. Importantly, less than 20% of patients included in our study were non-ventilated. Finally, we considered only patients with influenza without bacterial infection and therefore our results cannot be generalized to all patients admitted to the ICU.

Our study suggests that even though PCT levels in patients with AKI were higher, physicians should not continue thinking that high levels of PCT in patients with AKI are secondary to kidney dysfunction and should be aware of other causes of PCT elevation, including bacterial infections.

This study was supported by the Spanish Intensive Care Society (SEMICYUC) and Ricardo Barri Casanovas Foundation. The study sponsors have no role in the study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author (AR) had full access to all the data in the study and ultimate responsibility for the decision to submit it for publication.

AR, IML, ED, MB, JSV, MIR, DB, GS and LFR conceived and designed the study. All authors, apart from MIR, LFR and DB, contributed to the acquisition and local preparation of the constituent database.

AR, IML and MB contributed to database creation and standardization, design of statistical analyses, and data analysis.

AR, IML, JSV, MIR, AGE, LFR, ED, GS, and BS interpreted the data and wrote the paper.

All authors contributed to the critical examination of the paper for important intellectual content and approval of the final manuscript. Each author, apart from MIR, FLR and DB, acts as the guarantor of data from their individual center. AR acts as overall guarantor for the pooled analysis and the report.

The authors declare no conflict of interest.