Introduction
Although the rate has been reduced by compliance with the bundles recommended by the Surviving Sepsis Campaign, reported septic shock mortality still varies from 29 % to 38 % [
1]. Both pro- and anti-inflammatory responses occur early and simultaneously in septic shock. A rapid and early upregulation of genes of the innate immune response occurs, along with downregulation of genes of the adaptive immune response [
2‐
4]. This acute dysregulation may result in death in the first week (“early death”). This initial dysregulation may evolve to a complex state combined with an unabated innate immune response, which can lead to persistent, nonresolving inflammation, resulting in organ dysfunction and an impaired adaptive immune response that leaves the host unable to react to any assault [
5‐
7]. This other type of dysregulation may lead to death within or beyond the first month (“late death”). In addition, several studies have suggested an even longer–term mortality due to sepsis with persistent inflammation associated with mortality over the course of 1 year [
8‐
10]. The distinction between early (before day 5) and late (after day 5) death is increasingly being studied and may involve different underlying mechanisms [
11‐
17]. Patients who die quickly must be taken into account in any study of sepsis and must not be excluded, because they represent a sizable ratio and are distinguished from those who may die after a longer period of time [
18].
Polymorphonuclear neutrophils are the first cellular defense against infection and the key cell type of the innate immune system. The response to bacterial infection involves neutrophil recruitment and extravasation into the infected tissues. It can be seen as equilibrium between bone marrow release and tissue migration [
19]. Lymphocytes are heterogeneous cell populations with different functional and phenotypical properties involved in adaptive immunity. Lymphopenia has been proposed as an indicator of mortality in severe sepsis, mainly because of its activation of apoptotic processes [
20‐
22]. The physiological immune response of circulating leukocytes to various stressful events is often characterized by an increase in neutrophil counts and a decline in lymphocyte counts. Zahorec demonstrated a correlation between the severity of the clinical course and the grade of neutrophilia and lymphocytopenia [
23]. He proposed the neutrophil-to-lymphocyte count ratio (NLCR), which is an easily measurable parameter to express injury severity. In the context of infections, the NLCR has been proven to predict bacteremia more accurately than routine parameters [
24]. Using administrative data, researchers in a recent study suggested that the NLCR is associated with 28-day mortality in unselected intensive care unit (ICU) patients; however, they did not find this association when they focused on patients with sepsis [
25]. Moreover, it seems that there is some association between the source of the infection and hospital mortality in patients with septic shock. Leligdowicz et al. [
26] indicated that sepsis originating from the abdomen was associated with the highest rate of hospital mortality and that obstructive uropathy-associated urinary tract infections were associated with the lowest rate.
Accordingly, in the present study, we investigated whether the timing of death was related to a particular NLCR. In addition, we compared a group of patients with septic shock originating from the abdomen with a group who had sepsis of extra-abdominal origin.
Discussion
In the present study, we found that at the onset of septic shock, a low NLCR was associated with nonsurvival. This low NLCR was due to a higher lymphocyte count. When distinguishing time until death in the early-death (before day 5) and late-death (on or after day 5) groups, the NLCR of the early-death patients at admission was significantly lower than that of late-death patients and survivors. The risk of late death was associated with a neutrophil increase, lymphocyte count decrease, and subsequent increase in the NLCR from day 1 to day 5.
Early death (before day 5) versus late death (on or after day 5) is increasingly being studied [
5,
6,
11,
14‐
17]. An overly strong immune response with a cytokine storm [
2], mitochondriopathy, and massive tissue damage may lead to early death [
3]. This dysregulation may cause a complex state combining an unabated innate immune response that leads to persistent inflammation, resulting organ dysfunction, and an impaired adaptive immune response that leaves the host unable to react to any assault [
2,
6]. In our study, clinical characteristics did not differ between early- and late-death patients, particularly age, severity scores, and comorbidities. The most remarkable clinical difference was that of hospital-acquired infections, which was 2.7 times higher in nonsurvivors. Although nonsignificant, the ICU length of stay was longer in nonsurvivors. Whether the hospital-acquired infection was a consequence of alteration of adaptive function related to lymphopenia, whether it was responsible for neutrophilia, or both remains unclear.
Most of the prognostic scores mention leukocytosis (>12,000/mm
3) or leukopenia (<4000/mm
3) as a severity index, but none considers the leukocyte subpopulations [
30‐
34]. Significant differences between neutrophil and lymphocyte counts and, consequently, their ratio have been shown to predict the severity or outcome in different pathological circumstances, such as surgical stress, systemic inflammation and sepsis [
23], bacteremia [
24], community-acquired pneumonia [
35], ischemic events [
36], and cancer [
37,
38]. We must be cautious because the prognostic value of the NLCR in our study is below the value of 0.8 that is felt to be indicative of a reliable prognostic marker.
The evolution of these leukocyte populations may differ based on their respective role in the inflammatory response. Neutrophils are the first cellular line of defense in the innate immune system and have a short half-life, while lymphocytes are the major cells of the adaptive immune system.
In our study, all patients were lymphopenic. Lymphopenia is due first to a huge recruitment from peripheral circulation to the nidus of infection, followed by lymphocyte apoptosis caused by several stimuli [
20]; persistent lymphopenia of the late-death patients is most likely due to ongoing sepsis-induced lymphocyte apoptosis secondary to the continued release of proapoptotic stimuli and the nonresolution of inflammation.
Among the patients with lymphopenia, the lymphocyte count was higher in all nonsurvivors than in survivors. This difference was not found when all nonsurvivors were separated into early- and late-death groups, probably because of the small number of patients in each group. Nevertheless, the NLCR was surprisingly reduced in patients who died early, independent of the origin of the sepsis. This result might be explained by a decrease in neutrophils associated with an increased lymphocyte count. Some researchers have not reported this result, likely because they included only patients who survived until day 4 hospital mortality without considering death in the first 5 days [
18]. Accounting for death in the first 5 days is a significant issue. In the present study, more than two-thirds of the patients died in the first 5 days, which is a high rate and cannot be ignored. Although lymphocyte apoptosis is well described in septic shock, some authors have demonstrated that the early phase of human sepsis is characterized by a combination of apoptosis and the proliferation of T cells. A rapid recovery of total, CD4
+, and CD8
+ T lymphocytes could indicate their intense trafficking between tissues and the lymphatic system during the acute phase of illness [
12,
39‐
42]. Moreover, patients who die early might produce higher levels of stress hormones such as adrenaline, which increases lymphocyte counts [
43]. Although the exact underlying mechanism for a higher lymphocyte count in nonsurvivors remains unclear and requires further investigation, the NLCR appears to be a simple parameter to detect patients at risk of early death.
An increase in the NLCR from day 1 to day 5, with an increase in the neutrophil count and a smooth drop in the lymphocyte count, was associated with late death in our study population. Interestingly, in other studies researchers have found that late mortality, at 6 months or 12 months after emergency abdominal surgery, was associated with a persistent high neutrophil count [
44]. Increased neutrophils might indicate that the site of the infection has not been eradicated and that there is still pus or abscess in the peritoneal cavity or other nidus. Thus, the bone marrow continues to produce large amounts of neutrophils to fight the infection. We also suggest that the neutrophils might not become apoptotic. In contrast to lymphocytes, neutrophil apoptosis is beneficial in sepsis. The apoptosis of these cells initiates and facilitates the resolution of inflammation, tissue repair, and reestablishment of homeostasis. Apoptotic neutrophils are cleared by macrophages and prompt the macrophages to switch from a proinflammatory to an anti-inflammatory phenotype [
21,
45,
46]. This neutrophilic pattern could reflect a nonresolving inflammation [
47].
Finally, while some studies have indicated that the source of infection is important to consider because the immune response differs according to the site [
26], we did not find any significant differences between the abdominal and extra-abdominal sepsis groups concerning lymphocytes, neutrophil counts, or the NLCR in the early- or late-death groups. However, the size of the extra-abdominal sepsis group was quite small compared with the abdominal sepsis group.
Limitations of the study
The present study has several limitations. First, we conducted a single-center observational study, and thus, as with any observational study, the potential remains for residual confounding. The results have to be confirmed in other centers. Second, for some patients, several measures were available on the same day, and we always used the first one to maximize to consistency among the patients. However, we could have missed information related to intraday cell count variations. Third, we analyzed circulating neutrophil and lymphocyte counts and did not explore the different subsets of lymphocytes. Phenotypic markers are missing, but this is the subject of a forthcoming study. Fourth, our sample size was limited, and our results should be confirmed in a larger population. Given our limited sample size, we were not able to study other time points to define late death, such as death after 14 days or 28 days. Finally, no information about inflammatory biomarkers (such as C-reactive protein or procalcitonin) was available in our patients; the role of these markers in this context and their potential interaction with NLCR should be assessed.
Competing interests
DP acknowledges receiving lecture fees from Eli Lilly, Toray Industries, Tokyo, Japan, Vytech Health, Italy, and Meditor Dialysis ; LA WANTZENAU, France in the past 5 years. The other authors declare that they have no competing interests.
Authors’ contributions
All authors fulfill the requirements for authorship. FR contributed to the study conception and design, acquisition and interpretation of data, and drafting of the manuscript. EG performed statistical analysis and participated in the design of the study, interpretation of data, and drafting of the manuscript. RB contributed to acquisition and interpretation of data and revision of the manuscript. MLD and JM were involved in interpretation of data and revision of the manuscript. DP made a substantial contribution to interpretation of data and was involved in revision of the manuscript for important intellectual content. All authors read and approved the final manuscript.