If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
This prospective study investigated histamine and D-dimer levels in 72 patients with early stage bacteremia.
•
Histamine does not predict mortality in normal-ward inpatients with newly diagnosed bacteremia.
•
Early elevated d-dimer levels identify bacteremia patients at risk for in-hospital mortality.
Abstract
Introduction
Plasma histamine levels and D-dimer predict disease severity and mortality in advanced septic shock. We hypothesized that increased plasma histamine levels parallel coagulation activation and yield prognostic significance already at a very early stage of bacteremia.
Patients and methods
This prospective controlled cohort study enrolled 72 consecutive non-surgical non-ICU-ward inpatients with newly culture-diagnosed bacteremia and a Pitt Bacteremia score ≤2 to determine the extent of histamine and D-dimer release and their predictive role on outcome at the earliest stage of blood stream infection. Age-matched healthy adults served as internal controls (n = 36). A binominal logistic regression and a Cox proportional hazards regression analysis were performed to ascertain the effects of D-dimer and histamine on in-hospital mortality.
Results
In contrast to plasma histamine, D-dimer levels were significantly higher within hours of culture-proven bacteremia. In-hospital mortality occurred in 17%. Histamine levels were neither associated with D-dimer level (r = 0.04; p > 0.05) nor with ICU admissions (r = 0.06; p > 0.05) and outcome (crude OR 0.8, 95% CI 0.3–1.9; p = 0.6). In contrast, early-elevated D-dimer levels predicted mortality: the odds to die increased with the D-dimer level, and was 12.6 (crude OR, 95% CI 3–52; p = 0.001) in patients with a D-dimer ≥4 μg/mL (n = 13).
Conclusion
Histamine levels are elevated in only few patients (4%) with newly diagnosed bacteremia. Our findings suggest that D-dimer, but not plasma histamine, could be a promising marker of lethality already at a very early stage of blood stream infection.
] and a component of the disseminated intravascular coagulation (DIC) score of the International Society on Thrombosis and Haemostasis (ISTH). Coagulation activation is commonly based on endothelial damage with ensuing exposition of blood to extravascular tissue factor.
Histamine is a classic inflammatory mediator with various functions in antimicrobial host response. Among other actions, it is a potent vasodilator [
Neutrophils are the critical effectors of the human innate immune system and the first cells recruited into sites of inflammation. Previous studies showed that not merely mast cells but also neutrophils can release high amounts of histamine upon stimulation by bacterial endotoxins and through infection-related complement activation [
], neutrophil-derived histamine may early amplify inflammation, contribute to circulatory compromise and promote organ failure in blood stream infections.
] coincide with coagulation activation, a link between histamine and infection-related coagulopathy may be given by histamine's potency to open intercellular tight junctions [
Histamine differentially interacts with tumor necrosis factor-alpha and thrombin in endothelial tissue factor induction: the role of c-Jun NH2-terminal kinase.
A small prospective, controlled study investigated plasma histamine levels over five days in patients with overt sepsis, and indicated that histamine values on the first day were highest and best predicted sepsis severity and outcome [
Thus, we were interested in whether increased histamine and D-dimer levels are detected even at an earlier stage of infection and compared the predictive value of both in not yet critically ill inpatients with newly diagnosed culture-proven blood stream infection.
We hypothesized that increased plasma histamine levels parallel coagulation activation and may yield prognostic significance already at a very early stage of bacteremia.
2. Patients and methods
The study protocol was approved by the local Ethics Committee and all subjects provided oral and written informed consent. The study was conducted in compliance with the Declaration of Helsinki at the General Hospital of the Medical University of Vienna, a 2100 bed tertiary care facility, from February 2014 to February 2015.
One hundred eighteen consecutive non-surgical, non-intensive care unit (ICU)-ward inpatients ≥18 years with fever (body temperature ≥38.3 °C) and positive bacterial blood culture or LightCycler SeptiFast test result were screened for eligibility. Age-matched adults (n = 36) were used as internal controls to confirm the normal range of plasma histamine levels within the study and to define the assay-specific range of normal values. The primary outcome was the magnitude of histamine release in early bacteremia and its association with D-dimer levels. Secondary outcome was in-hospital mortality. Patients with anaphylaxis, those on histamine receptor blockers, vitamin K antagonists (VKA) or new oral anticoagulants (NOAC), patients with advanced disease severity assessed by the Pitt Bacteremia Score (PBS > 2) and those who underwent surgery within previous three weeks were excluded. Study-related blood sampling was performed immediately after receipt of positive blood-culture result. Parameters documented at time of enrolment are shown in Table 1. The frequency of ICU admissions was recorded and the patients were followed until hospital discharge or death. Coagulation tests were performed as previously described [
Scientific Subcommittee on Disseminated Intravascular Coagulation (DIC) of the International Society on Thrombosis and Haemostasis (ISTH).
Towards definition, clinical and laboratory criteria, and a scoring system for disseminated intravascular coagulation, Thromb Haemost.2001; 86: 1327-1330
]. D-dimer levels were measured on the STA-R Evolution analyzer (Diagnostica Stago for Roche Diagnostics GmbH, Germany; normal range < 0.5 μg/mL). Histamine EIA kits were purchased from Immunotech (Beckmann Coulter Company; normal range < 10 nM [nmol/L]). LightCycler SeptiFast test MGRADE (Roche Diagnostics, Germany) and blood cultures (BacT/ALERT FA/FN plus) were processed at the University's Department of Microbiology.
Table 1Baseline characteristics and outcome of study cohorts.
Antibiotic therapy present at the time point of study-related blood sampling. All patients received antibiotic treatment upon receipt of positive blood culture result.
One of remaining non-survivors was transferred to a palliative care clinic because of progressive incurable malignancy, and another patient suddenly died due to rupture of an aortic aneurysm. Three further patients (2 survivors and 1 non-survivor) with hematologic cancer were admitted to a specialized bone marrow transplant unit upon clinical deterioration, and one survivor with septic renal failure was temporarily treated at an intermediate care unit.
n.a.
5 (8)
9 (75)
Categorical data are presented as frequency count and percentage. Continuous variables are expressed as median and range (minimum–maximum).
Abbreviations: COPD, chronic obstructive pulmonary diseases; CRP, C-reactive protein; CVC, central venous catheter; DIC, disseminated intravascular coagulation; ICU, Intensive Care Unit; LMWH, low molecular weight heparin; WBC, white blood cell count; n.a., not applicable.
Study-related blood sampling was done 3.5 h (range: 1.0–19) after receipt of the positive blood culture result.
a The percentage distribution of bacteria causing blood stream infections is available with the supplement.
b Antibiotic therapy present at the time point of study-related blood sampling. All patients received antibiotic treatment upon receipt of positive blood culture result.
c Remaining two non-survivors died from multiple organ failure due to progressing underlying cancer (n = 1) and rupture of an aortic aneurysm (n = 1).
d One of remaining non-survivors was transferred to a palliative care clinic because of progressive incurable malignancy, and another patient suddenly died due to rupture of an aortic aneurysm. Three further patients (2 survivors and 1 non-survivor) with hematologic cancer were admitted to a specialized bone marrow transplant unit upon clinical deterioration, and one survivor with septic renal failure was temporarily treated at an intermediate care unit.
]. A sample size of n = 72 provided 80% power to detect elevated histamine levels in at least 5% of bacteremia patients. Based on a study on endocarditis patients [
], we determined a sample size of 9 patients per cohort to achieve a 90% power at an alpha error of 2.5% to detect increased D-dimer levels in non-survivors. Study variables are presented as absolute values (n) and relative frequencies (%), median and range (minimum to maximum) or proportions (odds ratio, OR; hazard ratio, HR) with 95% confidence intervals (95% CI). Healthy volunteers were used as internal controls to define the normal range of plasma histamine levels within the study. Since D-dimer levels increase with advanced age [
], we used an age-matched cohort of healthy adults, who were recruited from a volunteer-databank at the Department of Clinical Pharmacology, Medical University of Vienna.
Between-group comparisons of continuous and categorical variables were calculated using the Mann–Whitney U test and the chi-square test. Predefined subgroups included patients with low molecular weight heparin (LMWH) prophylaxis and those with renal impairment. Correlations were calculated using the Spearman rank correlation test. A binominal logistic regression and a Cox proportional hazards regression analysis (with mortality as dependent variable) were applied to ascertain the effects of D-dimer and histamine on in-hospital mortality. The Box–Tidwell procedure was performed to test for linearity assumption at a significance level of p < 0.017. Independent variables were log transformed or inversely transformed prior to analysis according to the skewness of the distributions. The Kaplan–Meier method was used to describe survival and the log-rank test was performed to determine differences in the survival distributions according to the D-dimer level. Regarding data missing at random, only available data were analyzed. Statistical calculations were performed with SPSS Statistical Software Version 22.0 (Armonk, NY: IBM Corp). A two-tailed p-value < 0.05 was considered significant.
3. Results
Of 118 consecutive bacteraemia patients screened, 72 were finally enrolled (male: 62.5%; median age: 64, range: 20–91). Remaining patients were excluded according to the predefined exclusion criteria (VKA/NOAC therapy: n = 21, recent surgical procedure: n = 10, Pitt Bacteraemia Score > 2: n = 9, histamine receptor blockers/anaphylaxis: n = 4). Two patients and 1 volunteer refused to participate in the study. All included patients (100%) were followed up until hospital discharge or death. Histamine and d-dimer levels were available for 100% of cases. Bacteraemia was verified by culture in 96% (n = 69) and by SeptiFast test in 4% (n = 3) of patients. The median Pitt Bacteraemia Score was 1 (range: 1–2) and the median duration from blood culture sampling to the first positive culture result (time to blood culture positivity) and the subsequent period to study-related blood sampling were 11.3 h (range: 2–59) and 3.5 h (range: 1.0–19), respectively. All patients (100%) received antibiotic treatment upon receipt of positive blood culture result. The most frequently identified organisms were E. coli (n = 20, 28%) and S. aureus (n = 17, 24%). Table 1 shows baseline characteristics and outcome of the study cohorts and the Supplemental Figure gives the percentage distribution of bacteria causing blood stream infections.
In 14% of patients no source of infection was identified (classified as primary bacteremia). These patients had a higher rate of cancer (70% vs. 30%), more often immunosuppressive therapy (70% vs. 18%) and pre-existing antibiotic treatment (30% vs. 13%) and had a slightly higher rate of death (20% vs. 16%), but numbers were small in these subgroups.
Plasma histamine levels were similar between patients and healthy controls and exceeded the normal range (i.e. >10 nM) in three patients, who all had an uncomplicated course of disease. D-Dimer was elevated above normal range (>0.5 μg/mL) in 68 patients (94%). Histamine levels did not correlate with coagulation tests (fibrinogen: r = −0.01, prothrombin time: r = −0.04, D-dimer: r = 0.04), laboratory markers of inflammation (CRP: r = −0.07, leukocytes: r = 0.12), source of infection (r = 0.19), or frequency of ICU admissions (r = 0.06; all: p > 0.05). Patients with renal impairment (n = 24, 35%) had only slightly higher median histamine levels (2.4 nM, range: 1.0–8.9) than those with normal kidney function (1.4 nM, range 0–15.1; p = 0.55). Prophylactic anticoagulation with LMWH (40 mg enoxaparin subcutaneously daily, n = 40, 59%) was neither associated with lower histamine (2.3 nM, range: 1.0–15.1 vs. 1.1 nM, range: 0–10.8 in patients without anticoagulation; p = 0.07) nor with lower D-dimer values (2.1 μg/mL, range: 0.5–32 vs. 2.2 μg/mL, range: 0.3–27; p = 0.79).
Among all patients, in-hospital mortality occurred in 17% (n = 12). Septic shock was the documented cause of death in 83% of non-survivors (n = 10). The rate of both S. aureus (42% vs. 20%; p = 0.11) and P. aeruginosa infection (17% vs. 10%; p = 0.26) was trendwise higher in non-survivors than in survivors, while rates of E. coli infection (17% vs. 30%; p = 0.70) were lower in those who died (Table 1, Supplemental Figure). Non-survivors were older (73 vs. 60 years; p = 0.03) and had higher d-dimer (4.6 vs. 1.7 μg/mL; p < 0.001) but similar creatinine (1.5 vs. 1.0 mg/dL; p = 0.83) and histamine levels (1.8 vs. 1.7 nM; p = 0.76). There was no difference in age between patients with a d-dimer level ≥4 μg/mL and those with a d-dimer below 4 μg/mL (63 vs. 67 years; p = 0.87).
The time to blood culture positivity was similar between survivors and non-survivors (10.9 vs. 12.4 h; p = 0.65). Across all patients the time to positivity negatively correlated with D-dimer levels (rho = −0.24; p = 0.06); this relationship was strongest in the subgroup of patients with E. coli bacteremia (rho = −0.40; p = 0.07). No correlation was found between time to blood culture positivity and plasma histamine levels (rho = 0.04; p = 0.76).
Fig. 1 shows Kaplan–Meier estimates of survival stratified according to the D-dimer level.
Fig. 1Kaplan–Meier curves showing survival according to the D-dimer level. Survival estimates were 92% for patients with D-dimer <4 μg/dL and 46% for patients with D-dimer ≥4 μg/mL, respectively. Blood sampling was done on the day of first positive blood culture result.
In a binominal logistic regression analysis the odds to die increased with the D-dimer level, and was 12.6 (OR, 95% CI 3–52; p < 0.001; Table 2) in patients with a D-dimer ≥4 μg/mL (n = 13). The model (χ2 = 23; p < 0.001) correctly identified 85% of cases with a negative predictive value of 92%. In contrast to D-dimer, histamine levels were not associated with mortality (OR 0.8, 95% CI 0.3–1.9; p = 0.6).
Table 2Univariable binominal logistic regression analysis for in-hospital mortality.
Variables
Crude OR (95% CI)
P-value
Age
1.05 (1.01–1.10)
0.029
Male sex
0.36 (0.10–1.27)
0.11
CRP
1.05 (0.98–1.13)
0.17
WBC
1.03 (0.98–1.09)
0.26
Histamine
0.90 (0.67–1.20)
0.48
D-dimer ≥ 4 μg/mL
12.6 (3.03–52.35)
0.001
Platelet count
1.0 (0.99–1.01)
0.73
Renal impairment
1.54 (0.43–5.49)
0.50
Time from hospitalization to first blood culture result
0.57 (0.94–1.03)
0.99
Staphylococcus aureus infection
2.86 (0.77–10.6)
0.12
Escherichia coli infection
0.47 (0.09–2.35)
0.36
95% CI, 95% confidence interval; CRP, C-reactive protein; OR, odds ratio; WBC, white blood cell count. P < 0.05 was considered significant.
Likewise, D-dimer (HR 2.5, 95% CI 1.5–4.9; p = 0.001) and age (HR 1.06, 95% CI 1.01–1.1; p = 0.017) were significantly associated with in-hospital mortality in univariable Cox regression analysis.
4. Discussion
Our study provides two major findings: (i) histamine levels are elevated only in a few patients with early bacteremia and did not predict outcome whereas (ii) increased D-dimer levels on the first day of positive blood culture are predictive of in-hospital mortality.
In contrast to our results, histamine levels in sepsis were shown to be significantly higher in non-survivors. Unfortunately this study did not provide any information on the frequency of end organ failure in the limited study population of 20 patients [
]. None of our patients had combined liver–kidney impairment, which could potentially lead to insufficient clearance of histamine levels in patients with advanced sepsis, because both organs clear histamine very efficiently [
]. Hence, histamine levels may only become elevated in systemic infection when multiple organ failure occurs, which may be addressed in future studies.
Although previous studies report a relationship between time to blood culture positivity and outcome [
], time to positivity was not associated with D-dimer levels in our study. Yet, we found a moderate negative correlation in the subgroup of patients with E. coli associated blood stream infection. Both the heterogeneity of bacteria causing blood stream infections and the low sample size of our study cohort may explain these findings.
Our study has two particular strengths. First, we investigated a homogenous population of non-surgical inpatients early after onset of bacteremia, but without need for intensive care at the time of enrolment. Studies on this specific population are rare compared to a vast amount of studies on septic critical illness. Secondly, blood samples were taken at a very early stage of disease — within 3.5 h after confirmation of bacteremia. This may be important because readily available parameters identifying patients with high in-hospital mortality could be of clinical relevance and may guide further hospital care. In this context, elevated D-dimer levels proved to be an early marker of mortality in early stage blood stream infection. In agreement, on-admission D-dimer levels predict disease severity and mortality in endocarditis [
The main limitation of our study is that we used healthy volunteers as controls. However, we hypothesized that any kind of infection-related systemic inflammation will increase histamine release. Therefore, we did not use inpatients with infection but negative blood culture as control group. We rather decided to include only patients with clinical signs of infection with culture-proven bacteremia to investigate a homogeneous population of inpatients with ascertained bacterial infection.
Another limitation is the relatively small sample size translating into wide confidence intervals and the low mortality rate allowing only one independent variable to be included in the regression analysis, so that further studies are required to confirm the predictive value of D-dimer in this population. However, the high negative predictive value is in good agreement with the performance of D-dimer to rule out pulmonary embolism [
], and a D-dimer cut-off value of 4 μg/mL is in line with results from a study investigating the predictive value of D-dimer in high-risk patients with infectious endocarditis [
We found increased histamine levels only in very few patients (4%) at an early stage of blood stream infection. In contrast, elevated D-dimer levels may be an early marker of increased mortality in non-ICU ward inpatients with culture proven bacteremia.
Learning points:
•
Studies investigating inpatients with newly diagnosed bacteremia are rare compared to a vast amount of studies on septic critical illness.
•
Our findings suggest that, in contrast to histamine, early-elevated D-dimer levels identify bacteremia patients at high risk for in-hospital mortality.
•
D-dimer could be a promising marker of lethality in this specific population.
The following are the supplementary data related to this article.
BJ conceived the study. MS, MMS, HL and NB identified eligible patients, obtained oral and written informed consent and performed study-related blood sampling. MS and BJ wrote the first draft of the manuscript and performed the statistical analysis. CS, HL, NB, PJS and TB collected data and substantially drafted the manuscript.
Conflict of interest
The authors declare no competing interests.
Acknowledgments
This work was supported by a grant from the Austrian Science Fund (SFB-54 grant: project number APF05404FW — Special Research Program: Cellular Mediators Linking Inflammation and Thrombosis, Medical University of Vienna). The Austrian Science Fund was neither involved in the conduct of the study nor in the writing of the report or in the decision to submit the article for publication.
References
Gris J.C.
Bouvier S.
Cochery-Nouvellon E.
Faillie J.L.
Lissalde-Lavigne G.
Lefrant J.Y.
Fibrin-related markers in patients with septic shock: individual comparison of d-dimers and fibrin monomers impacts on prognosis.
Histamine differentially interacts with tumor necrosis factor-alpha and thrombin in endothelial tissue factor induction: the role of c-Jun NH2-terminal kinase.
Scientific Subcommittee on Disseminated Intravascular Coagulation (DIC) of the International Society on Thrombosis and Haemostasis (ISTH).
Towards definition, clinical and laboratory criteria, and a scoring system for disseminated intravascular coagulation, Thromb Haemost.2001; 86: 1327-1330
00376-3&id=fx1.jpg){kind=link}
00376-3&id=gr1.jpg){kind=link}
00376-3&id=gr2.jpg){kind=link}