Patients’ characteristics and labworks

In total, 543 cases of lung (n = 207) or liver (n = 336) recipients were analysed. The median age of the patients was 52 years and the majority were male (m: 326; f: 217). Of 5,072 lab works performed, marker availability sorted by frequency was as followed: leukocytes: 99.9%; neutrophils: 99.9%; CRP: 98.4%; IL-6: 95.7%; NLR: 73.8%, lymphocytes: 73.8%; and PCT 7.2% (Table 1).

Table 1 Baseline characteristics of patients and availability of inflammation marker. Baseline characteristics and inflammation markers in cases of lung or liver transplant patients in the ICU. References for markers are in […]The NLR is increased in patients with positive bacterial and fungal samples

To evaluate the discriminating ability of the NLR, inflammation markers were analysed in patients who underwent liver or lung transplantation with multiple time positive microbial testing. Depending on the kind of microbe, different values for the NLR were obtained. The highest values were found in patients with positive test results for the combination of Gram-positive and Gram-negative bacteria (v) (18.4 [14.1, 24.2]), followed by patients with bacteria and fungi (iv) (15.1 [12.0, 18.4] (Fig. 1). In patients positive for only fungi (iii), the NLR was 13.5 [10.9, 17.0], and in patients with only Gram-positive bacteria (i) the NLR was 11.1 [9.4, 13.9]. The lowest NLR was found in patients with Gram-negative bacteria only (ii) (9.1) [8.5, 10.6].

A NLR value of 6.2 (9.03), was proposed as a threshold with both sensitivity and specificity > 90% for bacterial infection (fungal infection) [20, 21]. The frequency of NLR > 6.2 was the highest in patients tested positive for the combination of bacteria and fungi (iv) (87.1%), followed by the combination of Gram-positive and Gram-negative bacteria (v) (86.0%). In patients positive for fungi alone (iii), NLR > 6.2 was 86.0%, for Gram-positive bacteria alone (i) 74.8% and Gram-negative bacteria alone (ii) 70.8%.

The NLR shows higher sensitivity and specificity than other inflammatory markers

A NLR value of 6.2 (9.03), was proposed as a threshold with both sensitivity and specificity > 90% for bacterial infection (fungal infection) [20, 21]. To evaluate the performance of inflammation markers in lung and liver transplant recipients, ROC analysis was performed. NLR, neutrophils, lymphocytes, and IL-6 were the inflammation markers that demonstrated usefulness in terms of sensitivity and specificity in the detection of bacterial and fungal specimens (Fig. 2). The largest AUC was observed for the NLR (0.631; p < 0.001). The optimal NLR threshold for infection, based on the Youden index, was 12.0 (sensitivity: 74.0%, specificity: 50.7%, J = 0.246,). In ICU patients who underwent liver or lung transplantation, overall sensitivity for the detection of microbial samples with an NLR cutoff of 6.2 was 90.5% with a specificity of 21.8% (J = 0.122). The AUC for IL-6 was 0.579 (p < 0.001). A cutoff of 22.6 pg/mL (J = 0.147) showed the best combination of sensitivity (50.4%) and specificity (64.4%). The AUCs for neutrophils and lymphocytes were 0.550 (p = 0.001) and 0.608 (p < 0.001), and cutoffs were determined at 7.745 G/L (J = 0.121, sensitivity: 64.5%, specificity: 47.6%) and 0.505 G/L (J = 0.202, sensitivity: 57.0%, specificity: 63.2%).

Fig. 2

The NLR showed the highest sensitivity and specificity of inflammation markers in patients with multiple times positive microbial results. NLR showed highest sensitivity to specificity relationship in ROC analysis (AUC: 0.632; p < 0.001. The AUC for IL-6 was smaller (AUC: 0.580; p < 0.001). AUC of neutrophils and lymphocytes were 0.550 (p = 0.001) and 0.608 (p < 0.001). AUC of CRP was not significant and therefore not included in the figure

Detection frequency of positive bacterial and fungal samples is increased with the addition of leukocytes

In order to increase the detection frequency of bacterial and fungal samples a two-step analysis of inflammation markers was evaluated. If leukocytes were increased (≥ 9.8 G/L), the NLR was tested against the threshold of 6.2. The NLR was highest in patients tested positive for a combination of Gram-positive/Gram-negative bacteria (v) (24.2; [16.5, 38.9]), and lowest in patients with Gram-negative bacteria only (ii) (16.3; [13.8, 26.8]) (Fig. 3).

Fig. 3

Frequency of elevated NLR as an inflammation marker is increased, if leukocytes are elevated. The NLR was highest in patients with the combination of Gram-positive/Gram-negative bacteria (v) (24.2; [16.5, 38.9]), but lowest in patients infected with Gram-negative bacteria only (ii) (16.3; [13.8, 19.1]. The dotted line indicates the proposed threshold for NLR (6.2) for infection. (i: Gram-positive; ii: Gram-negative; iii: fungus; iv: combination of bacteria and fungus; v: combination of Gram-positive and Gram-negative)

NLR values > 6.2 were more often associated with positive microbiology results, if leukocytes were added to the decision process. The frequency could be increased for all specimen categories (i-v) from 78.2 to 93.1% The frequency of NLR values ≥ 6.2, if Gram-positive bacteria (i) were detected, was 92.3% (+ 17.5%); it was 91.8% (+ 21.0%) for Gram-negative bacteria (ii), 94.6% (+ 8.6%) for fungi (iii), 95.1% (+ 7.9%) for both bacteria and fungi (iv) and 93.1% (+ 14.9%) for both Gram-positive and Gram-negative bacteria (v).

The initial screening for elevated CRP (or IL-6) did not lead to significant changes in the detection frequency for microbial specimens for NLR values > 6.2. The frequency of NLR values ≥ 6.2 for the detection of Gram-positive bacteria was 73.8% [-1.0%] (76.8% [+ 2.0%]); it was 69.7% [-1.1%] (70.8% [± 0.0%]) for Gram-negative bacteria, 85.2% [-0.8%] (85.7% [-0.3%]) for fungi, 87.4% [+ 0.3%] (87.6% [+ 0.5%]) for both bacteria and fungi and 86.0% [± 0.0%] (81.1% [-5.1%]) for both Gram-positive and Gram-negative bacteria.

The NLR is independent of organ type or patient gender

To investigate whether the NLR is a stable and reliable marker, the covariables patient gender and type of organ transplant were analysed separately. NLR values were similar in patients who received lung or liver transplant (11.7 [10.6, 12.6] vs. 13.4 [10.9, 16.5]; p = 0.156) (Fig. 4.a). Stratification by gender showed no difference in NLR values (women: 10.9 [9.6, 13.1] vs. men: 12.2 [11.1, 14.2]]; p = 0.178) (Fig. 4.b). In contrast, IL-6 concentration differed significantly depending on the type of organ (liver: 57.9 pg/mL [48.8, 69.4] vs. lung: 28.2 pg/mL [26.0, 31.7]; p < 0.001) (Fig. 4.c) and between female and male patients (28.1 pg/mL [25.4, 32.0] vs. 41.1 pg/mL [35.6, 47.4]; p < 0.001) (Fig. 4d). In addition, the NLR showed only a moderate to no correlation with other inflammation markers ((A) WBC vs. NLR: r = 0.552; p < 0.001, B) CRP vs. NLR: r = 0.240; p < 0.001, C) IL-6 vs. NLR: r = 0.173; p < 0.001, D) PCT vs. NLR: r = 0.460; p = 0.012)) (Additional Fig. 2).

Fig. 4

Concentration of the NLR is independent of type of organ transplant and patients gender in contrast to Interleukin-6. The NLR values were similar in patients who are (a) lung or liver recipients (11.7 [10.6, 12.6] vs. 13.4 [10.9, 16.5]; p = 0.156) and (b) female or male (10.9 [9.6, 13.1] vs. 12.2 [11.1, 14.2]; p = 0.178). IL-6 values were higher in (c) liver recipients (57.9 pg/mL [48.8, 69.4] vs. 28.2 pg/mL [26.0, 31.7]; p < 0.001) and in (d) male patients (41.1 pg/mL [35.6, 47.4] vs. 28.1 pg/mL [25.4, 32.0]; p < 0.001)