Patient cohort baseline characteristics

The study was performed using a total of 89 plasma samples taken within the first 6 h after ICU admission from patients with clinical suspicion of sepsis, and subjects with spontaneous hemorrhagic stroke (used as ICU controls). Patients were classified into three different groups consisting of 9 (10%) ICU controls and 80 (90%) cases, among which 11 samples (14%) belonged to patients with sepsis diagnosis and 69 samples (76%) belonged to SS patients. It should be noted that two patients categorized as septic were undergoing vasopressor treatment, and required hemodynamic support in the form of orotracheal intubation and invasive mechanical ventilation (IMV) with intravenous sedation, instead of the SS condition (Table 1). Microbiological documentation was possible in 84% of cases. Regarding infection source, respiratory and abdominal focus predominated, and 42% of cases presented bacteriemia. In the SS group, 13 out of 69 patients (19%) finally died, of whom 12 (92%) had bacteremia (p Chi square < 0.001). In overt DIC patients’ mortality reached 30%. Clinical characteristics of the patients included in the study are shown in Table 1.

Table 1 Baseline clinical features of patients at ICU admissionPotential early diagnostic value of circulating histones

To evaluate the potential of circulating histones as biomarkers for diagnosis of septic processes, H2B and H3 levels were compared between ICU controls and septic cases. Note that H2B is one of the prevalent histones released during NETosis [25], and H3 has been linked to organ failure and coagulopathy in septic patients [26]. Levels of both histones were quantified through MRM-MS-based assays, which quantify circulating histones in plasma samples, using standard curves with different concentrations of light peptides for H2B and H3 and stable isotopic labeled spike-in peptides [24], as described in the methods section. This approach allowed the absolute protein quantification of both circulating histones to evaluate inter-group differences.

The median levels of circulating histone H2B found in the 9 plasma samples from ICU-controls was 0.00 ng/mL (interquartile range: 0.00), which increased up to 106.60 ng/mL (interquartile range: 423.44) in the 80 plasma samples from cases. The observed differences were statistically significant (p = 0.001 Mann–Whitney test) for histone H2B, but not histone H3 (Fig. 1).

Fig. 1

Levels of circulating histones H2B and H3 in plasma of ICU-controls and septic patients. Circulating histones were measured using LC–MS/MS using spike-in isotopic labelled peptides for absolute quantification. Groups were compared by U Mann Whitney test. *P < 0.001. The number of subjects analyzed were ICU Controls (n = 9) and cases [Sepsis + Septic Shock] (n = 80)

The correlation analysis for both H2B and H3 levels revealed a Spearman´s Rho correlation index of 0.752 (p = 0.001), indicating that despite the differing absolute levels of each histone, their levels increased in parallel.

Receiver Operation Characteristic curve (ROC curve) analysis also was performed to evaluate the diagnostic performance of histone H2B levels as a biomarker to distinguish septic cases from ICU-controls. The AUC (area under ROC curve), confidence interval (CI), optimal concentration cut-off value and sensitivity and specificity for histone H2B are shown in Fig. 2. The positive predictive value (PPV) of the diagnostic test based on a positive test cut-off point at 0.53 ng/mL was 98%, and the negative predictive value (NPV) was 30%.

Fig. 2

AUROC for histone H2B as a biomarker for diagnosis of septic processes. A detailed summary of the AUC, 95% CI lower and upper limit, sensitivity and specificity of the circulating histone H2B are shown in the inserted Table. The AUCs for circulating histone H2B in plasma were above 0.8, indicating that histone H2B showed high diagnostic value

Biomarker

AUC (95% CI)

P value

Cut-off value ng/mL

Sensitivity (%)

Specificity (%)

H2B

0.848 (0.754—0.942)

 < 0.0001

0.53

77.0

89.0

Circulating histones H2B and H3 are early diagnostic predictors of septic shock

Next, we compared the ability of circulating histone levels to differentiate patient severity in an ICU setting.

In patients with sepsis, H2B median concentration levels were 22.40 ng/mL (interquartile range: 176.90). However, levels were much higher in the SS group, with a median concentration of 121.40 ng/mL (interquartile range: 446.70) [p = 0.001].

For circulating histone H3, no statistical differences were found due to the number of samples with a concentration around 0.00 ng/mL in the different groups (Fig. 3).

Fig. 3

Concentration of circulating histones H2B and H3 in plasma of ICU-controls, sepsis, and septic shock patients. Circulating histones were measured using LC–MS/MS using spike-in isotopic labelled peptides for absolute quantification. Groups were compared by the Kruskal–Wallis test. *P < 0.01. The number of subjects analyzed were ICU Controls (n = 9); Sepsis (n = 11); septic shock (SS) (n = 69)

ROC curve analysis was performed to evaluate the diagnostic performance obtained by measuring the concentration of histones H2B and H3 as relevant biomarkers for patients suffering from sepsis or SS classification. The results obtained from curve analysis are shown in Fig. 4.

Fig. 4

AUROCs for histones H2B and H3 levels as biomarkers for classification of septic processes (sepsis vs. septic shock). A detailed summary of the AUC, 95% CI lower and upper limit, sensitivity and specificity of the circulating histones H2B and H3 are shown in the inserted Table. The AUCs for circulating histones H2B and H3 in plasma were above 0.6, indicating that histones H2B and H3 showed a moderate classification power

Biomarker

AUC (95% CI)

P value

Cut-off value ng/mL

Sensitivity (%)

Specificity (%)

H2B

0.670 (0.519–0.821)

0.028

74.66

63.2

72.7

H3

0.669 (0.533–0.804)

0.014

93.48

33.8

100

For histone H2B, the PPV of the diagnostic test based on a positive test cut-off point of 74.66 ng/mL was 92.50%, and the NPV was 24.24%. For histone H3 the PPV of the diagnostic test based on a positive test cut-off point for 93.48 ng/mL was 100%, and the NPV was 21.05%.

Circulating histones correlate with sepsis severity-related clinical and biological parameters

To assess the relationship between circulating histones and different parameters of disease severity, we performed a Spearman’s correlation study between circulating histones H2B and H3 and different clinical variables and biochemical parameters in septic cases. First, as mentioned earlier, a strong correlation was found with both histones H2B and H3 (Spearman’s ρ = 0.771, p < 1·10–16). Our results demonstrated moderate, positive and statistically significant correlation in septic patients between lactate levels in the first hour after ICU admission and H2B and H3 levels (Spearman’s ρ = 0.410, p < 0.001; Spearman’s ρ = 0.509, p < 0.001, respectively).

Regarding organ failure, a significant finding was a correlation between H2B and H3 histone levels and SOFA score at ICU admission (Spearman’s ρ = 0.378, p < 0.001; Spearman’s ρ = 0.491, p < 0.001, respectively). Moreover, circulating histones showed a strong correlation specifically with coagulation-related parameters: H2B (Spearman’s ρ = 0.725, p < 0.001) and H3 (Spearman’s ρ = 0.541, p < 0.001) correlated positively with D-dimer. H2B (Spearman’s ρ = − 0.269, p = 0.017) and H3 (Spearman’s ρ = -0.285, p = 0.011) were negatively correlated with platelet plasmatic count (Additional file 1: Tables S1 and S2). Finally, histone H3 showed a moderate negative correlation with functional protein C (Spearman’s ρ = − 0.477, p = 0.001).

In addition, we performed a correlation analysis focused on SS patients (Fig. 5), finding a strong correlation in this subgroup with both H2B and H3 (Spearman’s ρ = 0.812, p < 0.001). In addition, we found significant positive correlations of histones H2B and H3 with lactate (Spearman’s ρ = 0.409, p < 0.001; Spearman’s ρ = 0.475, p < 0.001, respectively), and H3 with SOFA (Spearman’s ρ = 0.466, p < 0.001). Regarding the correlations of circulating histones with coagulation parameters, we found that in SS patients H2B and H3 showed a strong positive correlation with D-dimer (Spearman’s ρ = 0.742, p < 0.001; Spearman’s ρ = 0.519, p < 0.001, respectively), whereas both histones correlated negatively with the Quick index (Spearman’s ρ = -0.485, p < 0.001; Spearman’s ρ = − 0.385, p < 0.001, respectively). Interestingly, H2B and H3 also correlated positively with LoS (Spearman’s ρ = 0.298, p = 0.014; Spearman’s ρ = 0.315, p = 0.009, respectively). Detailed information of correlations among the different parameters analyzed can be seen in Additional file 1: Tables S1 and S2.

Fig. 5

Heatmap representing the Spearman correlations coefficients (− 1 to + 1) among the clinical and analytical variables measured in septic shock cases. Red indicates negative and blue positive correlation between compared parameters. APACHE II: Acute Physiology and Chronic Health disease Classification System II; SOFA: Sequential [Sepsis-related] Organ Failure Assessment; APTT: activated partial thromboplastin time. The number of subjects analyzed were 80 septic patients (11 sepsis, 69 SS)

Afterwards, we evaluated the correlation of circulating histones H2B and H3 with PCT and PCR, currently widely used as biomarkers in sepsis. In septic cases, we found significant mild correlations between histone H2B and PCT (Spearman’s ρ = 0.319, p = 0.004), while H3 showed a negative moderate correlation with PCR (Spearman’s ρ = 0.477, p < 0.001). In SS cases (Fig. 5), we found only mild positive correlations between histone H2B and PCT and PCR (Spearman’s ρ = 0.286, p = 0.002; Spearman’s ρ = 0.342, p = 0.005, respectively), and H3 showed a moderate positive correlation with PCR (Spearman’s ρ = 0.329, p = 0.006). Black dots indicate statistical significance p < 0.05.  

In summary, our correlation analysis shows that levels of circulating histones H2B and H3 correlate with several classic biomarkers of sepsis progression, as well as with biochemical and clinical parameters related to specific processes associated with clinical management of septic patients, such as coagulation and organ failure. We therefore sought to determine whether levels of circulating histones could be used in the management of specific septic patient subsets.

Circulating levels of histones H2B and H3 were higher in septic shock patients with organ failure requiring invasive organ support therapies

To further evaluate the potential use of histones as early biomarkers for organ failure assessment, as well as their capability to predict the need of organ support therapies within the following 24 h, we analyzed H2B and H3 histone levels in SS patients requiring organ support therapy, specifically under RRT and IMV.

A total of 15 (22%) of 69 SS patients required RRT, 28 (40%) needed IMV support and 14 (20%) received both (RRT plus IMV). The median level of histone H2B in patients with RRT was 393.19 ng/mL (interquartile range, 2289.37) vs. 93.92 ng/mL in patients who did not require RRT support [interquartile range 404.73] (p = 0.030, Mann–Whitney test). For histone H3, in patients with RRT the median concentration was 261.36 ng/mL (interquartile range, 590.08) vs. 0.00 ng/mL (interquartile range, 190.54) in patients without (p = 0.008, Mann–Whitney test). For patients requiring early IMV support, differences in histone levels were observed but were not statistically significant (Fig. 6). In patients receiving both RRT + IMV support, H2B and H3 concentrations were 435.61 ng/ml (IR 2407.10) and 300.61 ng/ml (IR 912.77), respectively (p = 0.050), compared to 93.92 ng/mL of H2B and 0.00 ng/mL of H3 in patients not requiring any organ support therapy.

Fig. 6

Levels of circulating histones H2B and H3 in plasma of septic shock patients with and without support therapies. Circulating histones were measured using LC–MS/MS with spike-in isotopic labelled peptides for absolute quantification. Groups were compared via Kruskal–Wallis test. *P < 0.001. The number of subjects analyzed were septic shock cases without vital support therapy (n = 41); septic shock cases with IVM (n = 14) and septic shock cases with IVM + RTT (n = 14)

High levels of circulating histones are associated with overt DIC in septic shock patients

To evaluate the possible relationship between levels of histones H2B and H3 in septic processes and the presence of DIC, we proceeded to construct a new variable using the ISTH DIC scoring algorithm [27]. This score requires analytical parameters that individually give insufficient information to assess coagulation state as dynamic organ failure. Identifying a feasible DIC-related biomarker may therefore provide useful information to evaluate the coagulation state of critically ill patients in an early stage at ICU admittance and indicate overt DIC severity. This enables us to identify a subset of SS patients who could potentially benefit from specific management. Among the 80 samples from septic subjects (SS = 86%) only 10/69 (14.5%) SS cases were classified as overt DIC (ISTH scale ≥ 5), so values of circulating H2B and H3 were restricted to the SS group (Table 2).

Table 2 Values of biomarkers in septic shock patients according to ISTH Diagnostic Scoring System

To evaluate the diagnostic performance of histones H2B and H3 in SS patients who were admitted to ICU with associated overt DIC, a ROC curve was constructed and analyzed. The ROC curve obtained was not significant for histone H3, yet we were able to identify a cut-off for H2B with a good sensitivity and specificity percentage, even higher than the threshold described above for other associated organ failures in the SS population (Fig. 7). It should be noted that H2B was measured immediately after patient ICU admission, a time when coagulation might either still not be reflected by common analytical parameters, or be in the so called pre-overt or non-overt DIC state, with the possibility to progress or not to overt DIC.

Fig. 7

AUROC for histone H2B a as biomarker for DIC identification in septic shock patients. A detailed summary of the AUC, 95% CI lower and upper limit, sensitivity and specificity of circulating histones H2B and H3 are shown in the inserted Table. The AUCs for circulating histones H2B and H3 in plasma were above 0.7, indicating moderate diagnostic power for histone H2B

Biomarker

AUC (95% CI)

P value

Cut-off value ng/mL

Sensitivity (%)

Specificity (%)

H2B

0.703 (0.564–0.842)

0.041

335.50

70.0

71.0

Outcome prediction in septic shock patients based on circulating H2B and H3 levels

To assess the potential of histones for early prediction of fatal outcome in SS patients, we compared their values in cases diagnosed as SS with fatal outcome during the ICU stay (n = 13; 19%) and cases of SS who survived the SS episode (n = 56; 81%). For histone H2B the concentration in plasma samples from SS-surviving patients was tenfold lower than in those who died. The results obtained are shown in Table 3. The mean value for histone H3 concentration was under 163 ng/mL in all survivors, demonstrating that H3 presence in patient plasma was associated with fatal outcome in SS cases.

Table 3 Values of septic shock patient biomarkers according to outcome

The potential use of these biomarkers as outcome predictors in SS patients was assessed by a ROC curve analysis (Fig. 8). Accordingly, we found a 32% PPV of the diagnostic test based on a positive test cut-off point at 190.85 ng/mL and a 95% NPV for the histone H2B. Similarly, we obtained a 54.5% PPV of the diagnostic test based on a positive test cut-off point at 486.84 ng/mL and 89.5% NPV for the histone H3.

Fig. 8

AUROCs for histones H2B and H3 as biomarkers of outcome prognosis in cases of septic shock. A detailed summary of the AUC, 95% CI lower and upper limit, sensitivity and specificity of circulating histones H2B and H3 are shown in the inserted Table. The AUCs for circulating histones H2B and H3 in plasma were above 0.7, indicating that histone H2B and H3 showed moderate classification power

Biomarker

AUC (95% CI)

P value

Cut-off value ng/mL

Sensitivity (%)

Specificity (%)

H2B

0.743 (0.581–0.904)

0.008

190.85

83.3

63.1

H3

0.720 (0.546–0.895)

0.016

486.84

66.7

73.9

In summary, our results indicate that circulating histone concentration increases according to patient severity. Figure 9 represent a schema showing how levels of circulating histones can contribute to diagnosis and prognosis, classify patients requiring vital support therapies, identify pre-DIC patients, and even predict fatal outcome. Therefore, circulating histones may help physicians in deciding on the best therapeutic options and management for septic patients.

Fig. 9

Decision thresholds according to levels of circulating histones H2B and H3 in plasma for early management of patients with sepsis. Values below 190 ng/mL for H2B are associated with increased survival probability. This result is reinforced if patients showed a value below 163 ng/mL of circulating histone H3. However, values higher than 190 ng/mL of H2B are associated with high risk of renal failure in septic shock patients and predictive of fatal outcome. When the levels of circulating histone H2B were higher than 335.50 ng/mL, there was an increased risk of developing DIC in patients. In line with this idea, we found that values for histone H2B above 435.61 ng/ml were indicative of need for vital support therapies