Sepsis is the 10th leading cause of death in the United States, with mortality rates for severe sepsis estimated to be 30% or higher (1). Severe sepsis has also been estimated to be the primary cause for at least 500,000 emergency department (ED) visits annually in the United States, with average ED stay times of almost 5 h (2). The current consensus definition of overt septic shock requires suspicion of infection and systemic inflammation associated with persistent hypotension after adequate fluid challenge or a vasopressor requirement (3). Furthermore, hyperlactatemia has been shown to be a predictor of mortality in sepsis that is independent of organ failure and shock, with even modest elevations in lactate (2–4 mmol/L) having an impact on clinical outcomes (4, 5). The objective of this study was to determine if there are differences in clinical characteristics and outcomes among patients with vasoplegic septic shock, defined as overt shock and a normal lactate, as compared with patients with tissue dysoxic septic shock, defined as overt shock and an elevated lactate greater than 2 mmol/L.
We conducted a secondary analysis of a recently completed, large, multicenter randomized controlled trial (6). The objective of the trial was to evaluate the noninferiority of lactate clearance versus central venous oxygen saturation (ScvO2) as a marker of adequate oxygen delivery during early quantitative resuscitation of septic patients in the ED.
The methodology of the trial was previously reported (6). In short, the trial was conducted at Carolinas Medical Center, Charlotte, NC; Beth Israel Deaconess Medical Center, Boston, Mass; and Cooper University Hospital, Camden, NJ, and occurred between January 2007 and January 2009. The institutional review board at each institution (090602A) approved the study, and all participants or their surrogate provided written informed consent. The trial was registered on Clinicatrials.gov identifier NCT00372502.
Patients presenting to participating EDs with septic shock were eligible for enrollment if they were older than 17 years, had confirmed or suspected infection, had two or more systemic inflammatory response criteria, and had systolic blood pressure less than 90 mmHg after a 20-mL/kg rapid volume challenge or a whole blood lactate concentration of greater than 4 mmol/L. Patients were excluded from participation if they were pregnant, had any primary diagnosis other than sepsis, had an expected surgical requirement within 6 h of diagnosis, had an absolute contraindication to chest or neck central venous catheterization, had cardiopulmonary resuscitation, or had advanced directive orders that would conflict with the study procedure.
Once enrolled, patients were randomized to one of two study groups. While in the ED, each group had a structured quantitative resuscitation protocol, which has been previously described and published (6). The ScvO2 group was resuscitated by directing therapy required to meet threshold values of central venous pressure, mean arterial pressure, and ScvO2. The lactate clearance group had similarly targeted goals in central venous pressure, mean arterial pressure, and then lactate clearance (decrease in lactate of at least 10% over at least 2 h) instead of ScvO2 as a measure for adequate oxygen delivery. Protocols were followed until all end points were met or a maximum time of 6 h was reached. The results of this study confirmed the primary hypothesis of noninferiority, with a 6% (95% confidence interval [CI], 3% to 14%) in-hospital mortality difference between the two study groups (6). A total of 300 patients were randomized, 53 of whom demonstrated an elevated lactate and normotension, and were therefore excluded from the present analysis. This group of patients has been analyzed, and the results published previously (7) and were outside the scope of the current analysis, given the focus on patients with hypotension in sepsis.
For the present study, we categorized patients enrolled in the trial into one of two groups, vasoplegic shock or tissue dysoxic shock, as previously defined. Patient demographics and clinical characteristics were evaluated between the two groups using t, Mann-Whitney U, or χ2 test, as appropriate. The primary outcome of in-hospital mortality was evaluated using proportion differences and Kaplan-Meier curve with log-rank test. To assess for potential confounding, we conducted a logistic regression model using in-hospital mortality as the dependent variable. Candidate variables were chosen based on known predictors of mortality, such as age and degree of organ dysfunction, and those variables found to be different between groups in the bivariate analysis, and were maintained in the multivariate model if P < 0.10 while maintaining the event to independent variable ratio of approximately 8:1 to 10:1 (8). The model was refined using reverse stepwise elimination. Model fit was determined using Hosmer and Lemeshow’s goodness-of-fit test. All statistical tests were two-sided, with P < 0.05 considered significant. Data were analyzed using STATA (10.0; StataCorp, College Station, Tex) or StatsDirect statistical software (StatsDirect 2.7.7, Cheshire, England).
A total of 247 patients were included in this study. Ninety patients (36%) met criteria for vasoplegic shock versus 157 patients (64%) with tissue dysoxic shock. Patient demographics and initial clinical characteristics are shown in Table 1. There were no significant differences in demographics between the vasoplegic and tissue dysoxic shock groups. The group with vasoplegic shock had a lower initial Sequential Organ Failure Assessment (SOFA) score (5.5 points) than did the group with tissue dysoxic shock (7.0 points) (P = 0.0007).
The primary outcome of in-hospital mortality occurred in 49 (20%) patients of 247, including 8 of 90 patients (9%; 95% CI, 4%–17%) in the vasoplegic shock compared with 41 of 157 patients (26%; 95% CI, 19%–34%) in the tissue dysoxic shock group (proportion difference, 17%; 95% CI, 7%–26%; P < 0.0001). Figure 1 shows survival to hospital discharge curve for the two groups using a Kaplan-Meier format. A significant difference in survival between the two groups was noted (log-rank test, P = 0.02). In addition, we found significant differences in intensive care unit (ICU) and hospital length of stay and multiple organ failure between the groups (Table 2).
To attempt to control for potential confounding, we developed a logistic regression model using in-hospital mortality as the dependent variable. The final model included age, total SOFA score at enrollment, treatment with norepinephrine versus other vasopressors, and shock type (vasoplegic versus tissue dysoxic shock). Increasing age and SOFA score were associated with adverse outcome, whereas choice of norepinephrine versus other vasopressors was associated with improved outcomes in the multivariate model. The adjusted odds ratio to predict in-hospital mortality for tissue dysoxic was 3.0 (95% CI, 1.3–7.2). The final model demonstrated goodness of fit by the method of Hosmer and Lemeshow (P = 0.83).
In this study, we sought to evaluate the differences in outcome for patients with vasoplegic shock and tissue dysoxic shock. Our results indicate a significant difference in mortality between the two groups and that the presence of tissue dysoxic shock is an independent predictor of in-hospital mortality, even after accounting for potential confounding variables between cohorts. These results highlight the importance of potentially considering these subgroups of shock differently, particularly during enrollment into clinical trials where homogeneous populations are necessary to decrease the likelihood of random error and maintain the power to detect differences between study groups.
In this study, we evaluated only patients with septic shock by consensus definition and categorized those patients as having a normal lactate or an elevated lactate. In clinical practice, patients with persistent hypotension would be treated similarly, with fluids, vasopressors, and inotropic agents, regardless if lactate levels were elevated or normal. Thus, the implications of our results for clinicians are more prognostic than therapy changing. However, for the clinical trialist, these results could potentially impact the design of inclusion criteria for trials and may serve to ensure less heterogeneity in septic shock populations.
Previous authors have questioned if patients with sepsis-induced hypotension and a normal lactate should be considered as having septic shock (9). Hernandez et al. (9) examined the addition of hyperlactatemia to the consensus definition of septic shock, citing significant differences in mortality between septic shock patients without an elevated lactate compared with those with an elevated lactate (7.7% and 42.9%, respectively). These authors went on the question if hypotension without lactate elevation should even be considered septic shock. In follow-up of their initial study, Hernandez and colleagues (10) evaluated the outcomes and microcirculatory profiles of septic shock patients with and without hyperlactatemia. They again found significant differences in mortality and microcirculatory blood flow in patients with sepsis-induced hypotension and a normal lactate when compared with those with an elevated lactate (10).
Our findings confirm these findings and suggest a differential prognosis for patients with biochemical evidence of tissue dysoxia as opposed to just vasoplegia without evidence of tissue dysoxia. In the previous study by Hernandez et al., patients (9) were evaluated with serial lactates during the pre-ICU and ICU resuscitation periods, and patients with any abnormal lactate measurement during that time were included in the hyperlactatemia subgroup. Our study grouped ED patients based on initial lactate levels and further shows inherent differences apparent on presentation, despite disease course or resuscitation efforts.
In addition, previous research by Puskarich et al. (7) evaluated the outcomes of the varying presentations of septic shock. In that study, cryptic shock (suspected infection with a lactate >4 mmol/L and normotension) and overt shock (suspected infection and persistent hypotension) had similar in-hospital mortality rates of 20% and 19%, respectively. Our study extends these findings to further show that there are significant differences within the overt shock group when incorporating lactate as a variable.
Our study does have some important limitations. The initial study took place at experienced hospitals that perform high volumes of acute sepsis resuscitations and thus have resources that may not be available at other hospitals, so the results of this study may not be generalizable. Also, patients were placed into one of two treatment groups with different protocols in the initial study. The ScvO2 treatment group’s protocol did not target lactate clearance, which has been associated with worse outcomes versus lactate nonclearance and discordance from ScvO2 optimization (11). However, both treatment groups had protocols with similar goals and outcomes, and both the vasoplegic and tissue dysoxic shock groups had similar rates of achieving their predefined resuscitation goals (Table 3), mitigating this potential concern. Furthermore, when added to the multivariate model, treatment arm did not impact the results of our study. Finally, our study can only draw associations and cannot show cause and effect.
In this analysis, we found a significant difference in in-hospital mortality between vasoplegic and tissue dysoxic septic shock groups. These findings suggest a need to consider such outcomes when designing future studies of septic shock therapies.
1. Angus D, Linde-Zwirble W, Lidicker J, Clermont G, Carcillo J, Pinsky M: Epidemiology of severe sepsis
in the United States: analysis of incidence, outcome, and associated costs of care. Crit Care Med
29: 1303–1310, 2001.
2. Wang HE, Shapiro NI, Angus DC, Yealy DM: National estimates of severe sepsis
in United States emergency departments. Crit Care Med
35: 1928–1936, 2007.
3. Levy MM, Fink MP, Marshall JC, Abraham E, Angus D, Cook D, Cohen J, Opal SM, Vincent JL, Ramsay G, et al. 2001 SCCM/ESICM/ACCP/ATS/SIS International Sepsis
Definitions Conference. Crit Care Med
31: 1250–1256, 2003.
4. Mikkelsen M, Miltiades A, Gaieski D, Goyal M, Fuchs BD, Shah CV, Bellamy SL, Christie JD: Serum lactate
is associated with mortality
in severe sepsis
independent of organ failure and shock. Crit Care Med
37: 1670–1677, 2009.
5. Song YH, Shin TG, Kang MJ, Sim MS, Jo IJ, Song KJ, Jeong YK Predicting factors associated with clinical deterioration in sepsis
patients with intermediate levels of serum lactate
38: 249–255, 2012.
6. Jones AE, Shapiro NI, Trzeciak S, Arnold RC, Claremont HA, Kline JA Emergency Medicine Shock Research Network (EMShockNet) Investigators: Lactate
clearance vs central venous oxygen saturation as goals of early sepsis
therapy: a randomized clinical trial. JAMA
303: 739–746, 2010.
7. Puskarich MA, Trzeciak S, Shapiro NI, Heffner AC, Kline JA, Jones AE; Emergency Medicine Shock Research Network (EMSHOCKNET) Outcomes of patients undergoing early sepsis
resuscitation for cryptic shock compared with overt shock. Resuscitation
82: 1289–1293, 2011.
8. Peduzzi P, Concato J, Kemper E, Holford TR, Feinstein AR: A simulation study of the number of events per variable in logistic regression analysis. J Clin Epidemiol
49: 1373–1379, 1996.
9. Hernandez G, Castro R, Romero C: Persistent sepsis
-induced hypotension without hyperlactatemia
: is it really septic shock
? J Crit Care
26: 435, 2011.
10. Hernandez G, Bruhn A, Castro R: Persistent sepsis
-induced hypotension without hyperlactatemia
: a distinct clinical and physiological profile within the spectrum of septic shock
. Crit Care Res Pract
2012: 536852, 2012.
11. Pope JV, Jones AE, Gaieski DF, Arnold RC, Trzeciak S, Shapiro NI Emergency Medicine Shock Research Network (EMShockNet) Investigators: Multicenter study of early lactate
clearance as a determinant of survival in patients with presumed sepsis
32: 35–39, 2009.