You are on page 1of 7

Lactate Measurements in Sepsis-Induced Tissue

Hypoperfusion: Results From the Surviving Sepsis
Campaign Database*
Brian Casserly, MD1,3,4; Gary S. Phillips, MAS5; Christa Schorr, RN, MSN6; R. Phillip Dellinger, MD6;
Sean R. Townsend, MD7; Tiffany M. Osborn, MD, MPH8; Konrad Reinhart, MD9;
Narendran Selvakumar, MD4; Mitchell M. Levy, MD2,3

Objective: The Surviving Sepsis Campaign guidelines recommend
obtaining a serum lactate measurement within 6 hours of presentation for all patients with suspected severe sepsis or septic shock.
A lactate greater than 4 mmol/L qualifies for administration of early
quantitative resuscitation therapy. We evaluated lactate elevation
(with special attention to values > 4 mmol/L) and presence or
absence of hypotension as a marker of clinical outcome.
Design and Setting: The Surviving Sepsis Campaign developed
a database to assess the overall effect of the sepsis bundles as
a performance improvement tool for clinical practice and patient
outcome. This analysis focuses on one element of the Surviving
*See also p. 705.
1
Division of Pulmonary, Critical Care and Sleep Medicine, Memorial Hospital of Rhode Island, Pawtucket, RI.
2
Rhode Island Hospital, Providence, RI.
3
Warren Alpert Medical School at Brown University, Providence, RI.
4
University of Limerick, Limerick, Ireland.
5
The Ohio State University Center for Biostatistics, Columbus, OH.
6
Division of Critical Care, Cooper University Hospital, Camden, NJ.
7
California Pacific Medical Center, San Francisco, CA.
8
Washington University at St. Louis, St. Louis, MO.
9
Friedrich-Schiller-University of Jena, Jena, Germany.
Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF
versions of this article on the journal’s website (http://journals.lww.com/
ccmjournal).
Since December 2010 the SSC has been funded with an unrestricted
grant from the Gordon and Betty Moore Foundation.
Dr. Phillips received support for participation in review activities from
Rhode Island Hospital, a Lifespan Partner. His institution received grant
support from a National Institutes of Health Grant and from the Murdoch
Children’s Research Institute. Dr. Reinhart served as a board member for
the Global Sepsis Alliance (unpaid chairman) and the International Sepsis
Forum (Council member) and received support for article research from
the German Ministry of Science and Education. The remaining authors
have disclosed that they do not have any potential conflicts of interest.
For information regarding this article, E-mail: Brian_Casserly@brown.edu
Copyright © 2015 by the Society of Critical Care Medicine and Wolters
Kluwer Health, Inc. All Rights Reserved.
DOI: 10.1097/CCM.0000000000000742

Critical Care Medicine

Sepsis Campaign’s resuscitation bundle, measuring serum lactate in adult severe sepsis or septic shock patients and its interaction with hypotension. This analysis was conducted on data
submitted from January 2005 through March 2010.
Subjects: Data from 28,150 subjects at 218 sites were analyzed.
Interventions: None.
Measurements and Main Results: Unadjusted analysis of the 28,150
observations from the Surviving Sepsis Campaign database demonstrated a significant mortality increase with the presence of hypotension in conjunction with serum lactate elevation greater than 2
mmol/L. On multivariable analysis, only lactate values greater than 4
mmol/L, in conjunction with hypotension, significantly increased mortality when compared with the referent group of lactate values less
than 2 mmol/L and not hypotensive. Mortality was 44.5% in patients
with combined lactate greater than 4 mmol/L and hypotension when
compared with 29% mortality in patients not meeting either criteria.
Conclusions: Serum lactate was commonly measured within 6
hours of presentation in the management of severe sepsis or septic shock in this subset analysis of the Surviving Sepsis Campaign
database in accordance with the Surviving Sepsis Campaign
guidelines. Our results demonstrate that elevated lactate levels are
highly associated with in-hospital mortality. However, only patients
who presented with lactate values greater than 4 mmol/L, with and
without hypotension, are significantly associated with in-hospital
mortality and is associated with a significantly higher risk than intermediate levels (2–3 and 3–4 mmol/L). This supports the use of
the cutoff of greater than 4 mmol/L as a qualifier for future clinical
trials in severe sepsis or septic shock in patient populations who
use quantitative resuscitation and the Surviving Sepsis Campaign
bundles as standard of care. (Crit Care Med 2015; 43:567–573)
Key Words: Surviving Sepsis Campaign; sepsis mortality; serum
lactate

A

significant amount of evidence accumulated in the
literature demonstrates that lactate is a useful predictor of outcome in severe sepsis and septic shock (1–3).
www.ccmjournal.org

567

the presence of hypotension was a strong effect modifier of dichotomized lactate’s association with in-hospital mortality.org performance data and the time of specific actions and findings. lactate was dichotomized at greater than 4 mmol/L (yes vs no). Consequently. We used a risk factor modeling approach to determine which covariates to add to the logistic regression model. Statistical Analysis The primary goal of the study was to evaluate lactate elevation (with special attention to values > 4 mmol/L) within 6 hours after identification of severe sepsis and septic shock and the presence or absence of hypotension as a marker of clinical outcome. we sought to evaluate the relationship between lactate and mortality in severe sepsis and septic shock and the choice of lactate greater than 4 mmol/L as the marker of greater risk and initiation of protocol-guided resuscitation in patients with severe sepsis and septic shock. or mean arterial pressure less than 65 mm Hg. Inclusion was limited to sites with at least 20 subjects and at least 3 months of data. We compared clinical characteristics of patients across four groups of lactate/hypotension where categorical variables are presented as frequencies and percentages. time zero is ED triage time. Differences across groups were determined using Pearson chi-square test of association for categorical variables and Wilcoxon rank-sum for continuous variables. unresolved controversies remain regarding the true role of lactate in severe sepsis (4–6). The first 3 months that a site entered subjects into the database was defined as the first quarter regardless of when those months occurred from January 2005 to March 2010. As the care of septic patients improve. Time of meeting severe sepsis and septic shock criteria was determined by the origin at time of presentation. Several logistic regression models were developed to assess the association between mortality and lactate/hypotension.000 patients and the intention to revise the SSC bundles with the publication of the 2012 SSC Guidelines.ccmjournal. and one or more organ dysfunction criteria (10. Institutional Review Board Approval Analysis of the data for the global SSC improvement initiative was approved by the Cooper University Hospital Institutional Review Board (Camden. IL) via file transfer protocol or as comma-delimited text files attached to e-mail submitted to the Campaign’s server. With the publication of the Surviving Sepsis Campaign (SSC) analysis of 15. or a decrease of 40 or more mm Hg from known baseline. will lactate remain a good marker of risk? As new trials involving early resuscitation have been published. In patients presenting from the emergency department (ED). METHODS Sites and Patient Selection The process of participation in the SSC is described in detail elsewhere (7). whereas continuous variables are presented as medians and the interquartile range. Data Collection Data were entered into the SSC database at participating hospitals into pre-established. Only covariates that acted either as a confounder or as an effect modifier were included. printed material. Therefore. Unanswered questions regarding lactate in sepsis do remain. A confounder was identified when its addition to the model March 2015 • Volume 43 • Number 3 . Hypotension is defined based on the patient’s having any of the following: systolic blood pressure less than 90 mm Hg. Eligible subjects were those having a suspected site of infection. is a lactate level greater than 4 mmol/L still accurate for identifying patients as significant mortality risk? The Surviving Sepsis Campaign database is the largest prospectively collected database in severe sepsis and septic shock patients that also records clinical practice patterns over time (9) and has been used to improve outcomes (7). as well as a lactate level greater than 4 mmol/L. and in-person educational sessions. we believe that it is important to re-evaluate the value of serum lactate measurement in the routine management of these patients (8). and the components of “usual” care for patients with severe sepsis and septic shock evolve. Clinical characteristics and the time of meeting severe sepsis and septic shock criteria were collected for analysis of time-based performance measures. Data collection instructions were provided to site data collectors through the SSC website. We also wanted to evaluate the relationship between patient mortality and routine measurement of lactate as part of the sepsis bundles and to re-assess the validity of and lactate level. 11). we felt it was important to use the large approximately 30. Furthermore.Casserly et al However. a single categorical risk factor with four levels was created using these two dichotomized variables and is nominally called lactate/hypotension. to identify patients at higher risk of death. two or more systemic inflammatory response syndrome criteria. All relevant patient characteristics recorded in the database were eligible for inclusion in the riskadjustment models. Analysis Set Construction The analysis set was constructed from the subjects entered into the SSC database from January 2005 to March 2010. Data stripped of private health information were submitted to the secure master SSC server at the Society of Critical Care Medicine (Mount Prospect. time zero is determined when the patient met screening tool criteria for severe sepsis or septic shock or time of beginning of resuscitation and severe sepsis and septic shock management from clinical annotation in the chart. For patients presenting with severe sepsis or septic shock in areas other than the ED. NJ) as meeting criteria for exempt status.000 patient SSC database to confirm the relationship between lactate and mortality to justify continue use of lactate measurement as a bundle element for risk assessment (7). Therefore. Data were organized by quarter through 4 years of participation in the Campaign. unmodifiable fields documenting 568 www. The US Department of Health and Human Services’ Office for Human Research Protections reiterated that quality improvement activities such as SSC often qualify for institutional review board exemption and do not require individual informed consent (12). More specifically.

RESULTS Subjects Between January 2005 and March 2010. ward. Even at lactate concentrations of 2–3 and 3–4 mmol/L. College Station.419 (15.26. Hospital mortality differs across location where severe sepsis and septic shock were identified.ccmjournal.1 (Stata Corporation. A consort diagram (Fig.001) in those www.org 569 . Patients who presented with lactate greater than 4 mmol/L with hypotension showed a significantly higher mortality of 44. All analyses were run using Stata 12. the SSC database consisted of 28.945 patients in the four different lactates groups who presented with presence or absence of lactate greater than 4 mmol/L in less than or equal to 6 hours with or without hypotension are detailed and compared in Supplemental Table 1 (Supplemental Digital Content 1. to determine whether a relationship exists between missing lactate and hypotension. TX). we used only those lactate values that were measured within 6 hours. the location of severe sepsis or septic shock identification. and the GEE logistic regression model was re-run. or ICU) found results consistent with the entire population (data not shown).104). respectively (p = 0. respectively). A covariate that had any statistically significant interaction (p ≤ 0. we tested whether the odds of hypotension changed over time for those missing and those without missing serum lactate values and found that the odds did not change with each site quarter increase (p = 0.076 and 0. This method takes into account the variability within and between ICUs and uses this inherent correlation when estimating the ses that are used to test model coefficients. Specifically. Similarly. The association between an elevated lactate and in-hospital mortality is strongest (odds ratio. greater than 2 to less than or equal to 3. however.05) with lactate/hypotension was considered to be an effect modifier. This association is consistent regardless of whether lactate is measured less than or equal to 6 or greater than 6 hours. Removal of 4.Clinical Investigations changed any of the odds ratios associated with lactate/hypotension by more than 10% in either direction. or number of organ dysfunctions appear among the groups. Consort diagram. United States. in the adjusted regression analyses. data reviewed by the location of severe sepsis and septic shock identification (ED.422. Critical Care Medicine Patient Characteristics and Effect on Mortality The clinical characteristics of the 19.419. The four-level risk factor was recreated. Thus. who had a mortality of approximately 29%. 1.6%. A sensitivity analysis used multiple imputation (n = 20) where missing lactate values (n = 4. 20 months). The common sites of infection include pneumonia. and greater than 3 to less than or equal to 4 mmol/L independent of hypotension status or when lactate was measured. The hierarchical nature of the SSC data lends itself to this type of analysis. a significant increase appears in the odds of mortality in the hypotensive patients compared with lactate less than 2 mmol/L. and abdomen. and South America). which comprised more than 80% of cases and with no significant differences among the lactate groups.419 subjects with missing lactate left 23.32. Unadjusted Lactates Associated With In-Hospital Mortality Statistical analysis on all 23. In addition.lww.5% compared with the other three groups. 1) illustrates that 4. Figure 1.15. For the primary analysis.com/CCM/B134). No significant differences with respect to hospital length of stay. all four of these groups had a statistically significant in-hospital mortality odds ratio for a 1-unit increase in serum lactate.7%) subjects were missing serum lactate measurements.945 subjects for the primary analysis.3% and 34. the patient characteristics were found to be consistent with the results above on the entire analysis population (data not shown). p < 0.731 observations without missing lactate values from the SSC database showed a statistically significant increase in the odds of in-hospital mortality in patients presenting with lactate values greater than 4 mmol/L compared with those with lactate values less than or equal to 2.7%) were regressed on 30 covariates using truncated linear regression. a generalized estimating equation (GEE) logistic regression was used because patients are nested within a particular ICU.21–1. location was entered into the regression model as a confounder and not as an effect modifier. Mortality was 33. http://links. without considering statistical significance. This analysis was determined using an unadjusted GEE population-averaged logistic regression model (Table 1). thus we removed 3. 95% CI. Whether the lactate was drawn in patients with severe sepsis or septic shock less than or equal to 6 or greater than 6 hours or in the presence or absence of hypotension. there were no statistically significant interactions between the four groups and location.786 observations leaving 19. Figure 2 illustrates the linear increase in predicted in-hospital mortality with increasing serum lactate measured on a continuous basis using a similar GEE model. When reviewed by region (Europe.150 patients at 218 hospitals (median equal to 57 and a range of 20–471 subjects per hospital). The subjects were assumed to be missing completely at random because we failed to detect a difference in the in-hospital mortality between those with a serum lactate and those without.731 subjects for the analysis of continuous serum lactate by hypotension status and whether or not it was measured within 6 hours per SSC resuscitation bundle guidelines. Each site contributed to the database between 1 and 21 quarters of data for a mean duration of 21 months (median. 1. urinary tract.

Count.00 5.6) 1.e.891] 2. Adjusted In-Hospital Mortality Table 2 shows the adjusted and unadjusted odds ratios from a GEE logistic regression analysis of the 19.16 (1. severe sepsis and septic shock admission source (ED.88) [0.21 (1.5) 1.41 (1.60) [0.ccmjournal.94–1. 95% CI. Scvo2 ≥ 70% compliance. lactate less than or equal to 4 mmol/L and hypotensive.11–1. the adjusted odds ratios were not statistically significant. A sensitivity analysis used multiple imputation (n = 20) to impute the 4.423/27.97 (95% CI.1) 1. inspiratory plateau pressure less than 30 cm H2O compliance. and lactate > 4 mmol/L and hypotensive) compared with lactate ≤ 4 mmol/L and not hypotensive are adjusted for the following confounding variables: fluids and vasopressor Figure 2. and 1.04 (0.org compliance.344/44.11.001] 107 (55/51.1) 1.03–1.5) 2. The odds ratios for the three groups (lactate > 4 mmol/L and not hypotensive.001] OR = odds ratio. The in-hospital mortality odds ratios were 1. patients with the so-called cryptic septic shock. central venous pressure ≥ 8 mm Hg compliance.64 (95% CI.00 369 (109/29. or South America) and point of severe sepsis and septic shock identification in the hospital (ICU.. p < 0. p = 0.94. ward.274 (718/31.6) 1.94–1.2) 1. p = 0. p = 0.2) 1.35) [< 0.721).Casserly et al Table 1.80–1. ED.153).39–2.302 (301/23.001] 719 (421/58.002] > 3 to ≤ 4 693 (158/22.16) [< 0.38 (1.6) 3.419 missing lactate values. Change in Measurement of Serum Lactate Over Time in the SSC Lactate compliance improved the number of quarters of participation in SCC and the percentage of missing lactate values March 2015 • Volume 43 • Number 3 .09–1. a Odds ratio based on generalized estimating equation population-averaged logistic regression model. or ward).001) in non-hypotensive patients measured less than or equal to 6 hours.001] 5.33 (1.20 (0. 0.17 (95% CI.85) [< 0.07. and Odds Ratios for In-Hospital Mortality Generated by the Four Different Lactate Levels Across Lactate Compliance and Presence or Absence of Hypotension Compliant Lactate Measured ≤ 6 hr No Hypotension Noncompliant Lactate Measured > 6 hr Hypotension No Hypotension Hypotension Lactate Group (mmol/L) Total.93–2.945 patients whose lactate was measured within 6 hours of severe sepsis and septic shock identification.73 (1. 1. n (Died n/%) ORa (95% CI) [p] Total. lactate ≤ 4 mmol/L and hypotensive. pulmonary organ failure.21) [0.87–4.0) 1.009 (242/24. mechanical ventilation.16–1. i. The cutoff of greater than 4 mmol/L with hypotension was the only level that was still significantly different from the referent group (≤ 4 mmol/L and not hypotensive) after adjustment using our multivariable analysis.83–1.002] 143 (46/32. 570 www. Europe.00 1. the interaction between mechanical ventilation and pulmonary organ failure. or ICU).018] 356 (150/42.6) 1.52) [0.661] 3.14.8) 0. Percent.07) [< 0. or South America).78–2.42 (2. In patients with high lactate levels (> 4 mmol/L) who were not hypotensive.143] 613 (222/36.6) 1. the results were consistent with the entire sample (results not shown).10 (1. n (Died n/%) ORa (95% CI) [p] Total. respectively. 1. hypotensive patients whose lactate was measured greater than 6 hours. n (Died n/%) ORa (95% CI) [p] Total.27) [< 0. geographic region (United States.4) 2.45.39–1.407 (417/29.0) 1. and lactate greater than 4 mmol/L and hypotensive.05–1.99 (0. n (Died n/%) ORa (95% CI) [p] ≤2 (referent) 1.272 (2.001] >4 996 (289/29. 0.65) [< 0.241 (991/30. and number of baseline organ failures.24) [0. United States. whereas the association is weakest (odds ratio.001] 72 (27/37. 0. When this analysis was repeated for specific subpopulations (Europe. In-hospital mortality compared to serum lactate when measured continuously.001) for lactate greater than 4 mmol/L and not hypotensive.00 > 2 to ≤ 3 1.87–1.25 (1.27) [0. 1. The adjusted odds ratios again indicated that only lactate greater than 4 mmol/L and hypotensive were significantly different from less than or equal to 4 mmol/L and not hypotensive.06–1.158 (1.5) 1.

ward.001).1–45.5–30. In addition. Other authors have focused on patients specifically with septic shock (15). fluids and vasopressor compliance. or ICU).50) < 0.5 (43.3 (21.. prospectively collected and recorded database of lactate values and clinical outcomes to confirm previous investigations about the prognostic use of lactate levels.23 (0. regardless of blood pressure.58) < 0.004 23. United States.2) 1. On the basis of analysis of 19.05–1.2–31. The unadjusted analysis justified the use of the cutoff value of lactate greater than 4 mmol/L in the subsequent multivariable analysis (Table 2).3 (28.332 No Yes 10.945 subjects from the SSC database. geographic region (United States.24–1.34 (1. our results were consistent in highlighting the association between rising lactate levels and in-hospital mortality. Howell et al (2) demonstrated an adjusted odds ratio of 7. a Adjusted for the following variables: Antibiotic compliance. Furthermore.63 (2. the adjusted analysis also demonstrated that the lactate cutoff of greater than 4 was the only cutoff to have a statistically significant association with in-hospital mortality in all four patient groups.8) 1. ED. The unadjusted odds ratio for overall lactate compliance is 1. The SCC database includes a large population of patients who were admitted to the hospital with severe sepsis and septic shock and used the SSC guidelines as standard of care. An unadjusted logistic regression model demonstrated that when combining the presence of hypotension and the timing of the lactate measurement. b Odds ratio based on generalized estimating equation population-averaged logistic regression model where the group lactate ≤ 4 mmol/L and no hypotension is the referent group.1 for patients with increased blood lactate of greater than 4 mmol/L (after adjustment for age and blood pressure). whereas the association is weakest in nonhypotensive patients measured less than or equal to 6 hours (Table 1). i. the association between an elevated lactate and in-hospital mortality is strongest (highest odds ratio) in those hypotensive patients whose lactate was measured greater than 6 hours.Clinical Investigations Table 2.00 1. and each individual location in the hospital where the severe sepsis or septic shock was initially identified.8) 1.38–2. Bakker et al.06 (0. DISCUSSION Seminal work by Broder and Weil (13) studied 56 patients in shock from a variety of causes and found a mortality rate of 89% when a single lactate value was greater than or equal to 4 mmol/L.e. Our results demonstrate that only patients who presented with lactate values greater than 4 mmol/L in the presence of hypotension were associated with in-hospital mortality after adjusting for the confounding variables. The application of the SSC bundles has already been demonstrated to improve outcomes in patients with severe sepsis and septic shock as a result we wanted to evaluate the prognostic value of lactate in those circumstances.06 (95% CI.001 1.58) 0. The present study uses a large. namely lactate drawn within 6 or after 6 hours and/or in the presence or absence of hypotension (Table 1).74) 0. Similarly. Europe. Our data support that a lactate value greater than 4 mmol/L is an appropriate cutoff for future clinical trials designed to improve www.e. or ward.004 OR = odds ratio.0 (26.. Vincent et al. They found that lactate levels were significantly higher in nonsurvivors. Scvo2 ≥ 70% compliance. This analysis was repeated for each individual geographic Critical Care Medicine location. or South America). Regardless of where the data were obtained.36 (1.272 44. i. our study is the largest comparative analysis of the association between lactate and in-hospital mortality in patients presenting with severe sepsis or septic shock. intermediate lactate levels were associated with increased risk of mortality (Fig. 1.787 Yes Yes 5. Only patients meeting both criteria (elevated lactate and hypotension) show a statistically significant association after risk adjustment (Table 2). sepsis admission source (emergency department.81–1.001 1.21–2. % (95% CI) Unadjusted Regression ORb (95% CI) p Adjusted regressiona ORb (95% CI) p No No 3.org 571 . mechanical ventilation. 2).07. Aduen et al (14) investigated lactate levels in a cross section of 180 undifferentiated ICU patients. without any changes in results. the interaction between mechanical ventilation and pulmonary organ failure. indicating lactate compliance increases 6% per site quarter.8–24. However. and Falk et al (16–18) found that serial lactate levels predicted mortality in ICU patients with septic shock. Our results demonstrate that lactate remains an effective risk-stratification biomarker in patients where the SCC guidelines are applied as standard of care.673 29. central venous pressure ≥ 8 mm Hg compliance. inspiratory plateau pressure < 30 cm H2O compliance. Hospital Mortality and Odds Ratio for the Four Different Lactate Groups Before and After Adjusted Regression Analysis Lactate > 4 mmol/L Hypotensive n Unadjusted Hospital Mortality. declined in the hospital sites participating for at least 2 years. Europe.71–1.91) < 0. Bakker et al (16) focused on 48 ICU patients with septic shock and found that initial and serial blood lactate levels were superior to pulmonary arterial catheter–derived data in predicting mortality.00 Yes No 996 29.8) 2.82 (1. p < 0. and number of baseline organ failures (1–5). but that initial levels had less of a prognostic ability. the study by Rivers et al (19) showed the importance of blood lactate concentrations greater than or equal to 4 mmol/L in the context of early goal directed therapy.14–1. and South America.001 1. pulmonary organ failure.ccmjournal. ICU.89) 0.

However. because of the small fraction of the total patient population in our database (n = 996). even in the absence of hypotension (Fig. In this study. this result may be. 24). categorical variables are often the only practical solution when creating inclusion criteria for clinical trials and their cutoff values are important in assessing therapeutic intervention or quality improvement projects (22). observational studies can provide valuable information on treatment effectiveness (31). This suggests that further studies are needed to confirm the mortality benefit from reducing blood lactate levels through the proposed resuscitation strategy that uses serial lactate measurements to guide therapeutic endpoints resuscitation for two reasons: First. Our results are in keeping with other studies that examined ED patient populations suggest that patients with intermediate lactate values between 2 and 4 may also benefit from quantitative resuscitation (20. there was an association with increased lactate values at presentation and an increase in inhospital mortality in patients presenting with severe sepsis or septic shock. Despite their inherent weaknesses. Lactate has been used not only as a marker of risk but also to guide therapy (21). this preliminary data will need to be verified in prospective clinical trials. our database does not allow us to make that distinction. which suggests that more stringent criteria may be required in clinical trials of severe sepsis and septic shock that include this subgroup of patients. Therefore. A study by Jansen et al (29) showed that a treatment protocol aimed at reducing lactate levels by 20% in 2 hours resulted in a significant reduction in in-hospital mortality. Ultimately. The SSC database was designed to provide a description of the management and outcomes of severe sepsis and septic shock patients. However. Our study is clearly limited by its observational design. Therefore. 26). although lactate served as a warning signal for more aggressive therapy that resulted in decreased mortality.ccmjournal. 26). 2 and Table 1). our results confirm that all lactate values greater than or equal to 2 mmol/L have a clinical value because their association with increased mortality is shown to increase linearly (Fig. However. This suggests dissociation between lactate measurement and in-hospital mortality. reflecting everyday clinical practice within the boundaries of the SSC guidelines (33). In this large population. 2). Contrary to the finding of Jansen et al (29) where failure to clear lactate did not separate patients who did or did not respond to quantitative resuscitation. Hernandez et al (27) highlighted the need to interpret hypotension in the context of lactate levels when they demonstrated that hypotension in septic patients does not portend a poor outcome if that lactate levels remain within normal range. It may also represent an improvement in patient outcome in the patients who have demonstrated a blood pressure response to quantitative resuscitation. Moreover. lactate appears to be more highly associated with hospital mortality when combined with hypotension. a cutoff of 4 mmol/L was not significant in the absence of hypotension after an adjusted multivariable analysis. However. Our data suggest that serum lactate levels are associated with mortality in septic shock. However.org which could also contribute to high mortality in severe sepsis and septic shock (28). it may also reflect a delay in initiating resuscitation efforts. which is strongly associated with increased mortality (25. Interestingly. more aggressive therapy did not succeed in faster lactate clearance compared with the control group (29). Our results show that the measurement of lactate in severe sepsis and septic shock increased over time. this association persisted whether the lactate was measured less than or equal to 6 and greater than 6 hours from presentation. observational studies maximize external validity and often reflect real-life practice more accurately than randomized controlled trials (32). This is a particularly important group of patients because compensated shock is not easily recognizable (23. patients admitted to the hospital who present with a clinically suspected significant infection and a single measurement of venous lactate value greater than 4 mmol/L in the presence of hypotension have been shown to be associated with in-hospital mortality in patients receiving the SSC care bundles. 2). This supports the continued use of greater than 4 mmol/L as the trigger for quantitative resuscitation in the SSC guidelines and bundles although the increasing mortality March 2015 • Volume 43 • Number 3 . our results present evidence of preservation of the association between increased lactate and increased inhospital mortality in a patient population who received quantitative resuscitation according to the SSC guidelines (Table 1 and Fig. Although we found the commonly used cutoff value of greater than 4 mmol/L for lactate with hypotension to be the only cutoff statistically associated with in-hospital mortality following multivariable analysis. This reflects the success of the sepsis bundles as part of a performance improvement initiative in improving physician compliance with a known quality indicator in the management of severe sepsis (30). 572 www. undermining its ability to predict in-hospital mortality. lactate levels may be helpful in deciding which patients could benefit from aggressive treatment strategies because lactate is shown to be independently associated with mortality (25. the therapeutic intervention succeeded in improving outcomes without an improvement in lactate clearance. Delayed identification of hypoperfusion may lead to inadequate or delayed resuscitation. We also examined the measurement of serum lactate in severe sepsis and septic shock in the SSC database over time. Unfortunately. to validate the findings of Jansen et al (29) and to identify how patient outcome improved in the absence of an effect on serial lactate levels. The lactate measurement of greater than 4 mmol/L after 6 hours may be a powerful indicator of in-hospital mortality because it reflects failure to clear lactate effectively in response to resuscitation efforts. 21). in part. Our article also looked at a population of patients with cryptic septic shock. Data resulting from secondary analysis of the SSC database will serve as an ongoing audit for the effectiveness of the individual components of the care bundles devised by the SSC (34).Casserly et al outcome in severe sepsis and septic shock where quantitative resuscitation is being applied as standard in adherence with SSC guidelines. In conclusion.

Vincent JL: The Surviving Sepsis Campaign sepsis change bundles and clinical practice. Intensive Care Med 2007. Nguyen B. 39:165–228 10. et al: Delayed lactate clearance in patients surviving circulatory shock. Salinas J. Dellinger RP. Castillo JL. Faude O. Levy M: The Surviving Sepsis Campaign: A history and a perspective. van Bommel J. Crit Care 2008. Marshall JC. Intensive Care Med 2012. et al: The prognostic value of blood lactate levels relative to that of vital signs in the pre-hospital setting: A pilot study. 99:956–962 Critical Care Medicine 17. 272:1678–1685 15. Loiacono LA. Chansky ME. 26:409–421. 31:1250–1256 12. Marik P. Crit Care Med 2003. controlled trials. Nguyen HB: Lactate in the critically ill patients: An outcome marker with the times. Intensive Care Med 2010. Kindermann W. et al: The Surviving Sepsis Campaign: Results of an international guideline-based performance improvement program targeting severe sepsis. Surviving Sepsis Campaign Guidelines Committee including The Pediatric Subgroup: Surviving Sepsis Campaign: International guidelines for management of severe sepsis and septic shock. Pharmacotherapy 2010. Gerlach H. et al: Blood lactate levels are superior to oxygen-derived variables in predicting outcome in human septic shock. Bakker J.S. Leavy J. Crit Care Med 2010. 30:973–984 33. 370:1683–1693 9. 182:752–761 30.gov/ohrp/ categories/1569 13. et al: Occult hypoperfusion and mortality in patients with suspected infection. 2:339–347 7. Suffoletto B. Eastridge BJ. Band RA. 345: 1368–1377 20. DiPietro NA: Methods in epidemiology: Observational study designs. et al. Crit Care Clin 2010. Shapiro DS: Detection of hypoxia at the cellular level. et al. SCCM/ESICM/ACCP/ATS/ SIS: 2001 SCCM/ESICM/ACCP/ATS/SIS International Sepsis Definitions Conference. Hartz AJ: A comparison of observational studies and randomized. Surg Infect (Larchmt) 2010. 2012. Benson K. Mulder PG.org 573 . Trends Mol Med 2014.ccmjournal. et al: Incremental benefit of individual American College of Surgeons trauma triage criteria. REFERENCES 1. Crit Care 2005. et al. Levy MM. A randomized trial of protocolbased care for early septic shock. Mikkelsen ME. Gaieski DF. Crit Care Med 2008. Intensive Care Med 2013. JAMA 1994. Broder G. Royston P. open-label. Stat Med 2006. 38:1045–1053 29. J Trauma 2007. Havstad S. Kellum JA. Howell MD. Henry MC. randomized controlled trial. Fink MP. Coffernils M. Meyer T: Lactate threshold concepts: How valid are they? Sports Med 2009. Huang DT. U. N Engl J Med 2000. 25:127–141 23. Townsend SR. et al: Serum lactate as a predictor of mortality in patients with infection. 38:1946–1954 11. et al. 9:653–654 31. 63:291–297 26. table of contents 4. et al. Phillips GS. Trzeciak S. et al: Impact of time to antibiotics on survival in patients with severe sepsis or septic shock in whom early goal-directed therapy was initiated in the emergency department. et al: Low-dose steroids in adult septic shock: Results of the Surviving Sepsis Campaign. Fuller BM. 12:R160 24. 11:275–281 34. et al. Early Goal-Directed Therapy Collaborative Group: Early goal-directed therapy in the treatment of severe sepsis and septic shock. Khastgir T. Emergency Medicine Shock Research Network (EMShockNet) Investigators: Lactate clearance vs central venous oxygen saturation as goals of early sepsis therapy: A randomized clinical trial. Schorr CA. et al: Persistent sepsis-induced hypotension without hyperlactatemia: A distinct clinical and physiological profile within the spectrum of septic shock. 70:782–786 2. Rivers E. 303:739–746 22. Hernandez G. et al: The use and clinical importance of a substrate-specific electrode for rapid determination of blood lactate concentrations. March 3. LACTATE Study Group: Early lactate-guided therapy in intensive care unit patients: A multicenter. et al. et al: Hypotension begins at 110 mm Hg: Redefining “hypotension” with data. Leon M. McManus JG. Rhodes A. 143:1457–1459 14. 33:970–977 16. Bernstein WK. 2012:536852 28. N Engl J Med 2014. Levy MM. Guyette F. van Bommel J. present and future. Schoonderbeek FJ. 11: 212–215 18. 39:469–490 5. Altman DG. Sauerbrei W: Dichotomizing continuous predictors in multiple regression: A bad idea. Acad Emerg Med 1996. Curr Opin Crit Care 2012. Falk JL. Science 1964. Garcia-Alvarez M. Sakr Y.hhs. Marshall JC. Crit Care Res Pract 2012. Am J Respir Crit Care Med 2010. Brown MD. Hollander JE. Jansen TC. Crit Care 2011. Levy MM. JAMA 2010. Acute Care 1985. Dellinger RP. Crit Care Med 2006. N Engl J Med 2001. J Trauma 2011. Trzeciak S. et al. Sepsis Occurrence in Acutely Ill Patients Investigators: Sepsis in European intensive care units: Results of the SOAP study. Lancet Diabetes Endocrinol 2014. Dellinger RP. Vincent JL. Rackow EC. Sprung CL. 34:344–353 19. 36:222–231 8. Dellinger RP. 3:992–1000 27. Aduen J. 20:192–194 www. Jansen TC. Yealy DM. 342:1878–1886 32. Chest 1991.Clinical Investigations associated with intermediate lactate levels (2–4 mmol/L) would suggest that this patient population would also derive benefit from quantitative resuscitation and highlights the need for those clinical trials to be performed. Alicandro JM. 36:2734–2739 21. Bruhn A. Available at http://answers. Jones AE. Donnino M. Casserly B. Dellinger RP. Dellinger RP: The Surviving Sepsis Campaign: Past. 18:267–272 6. 15:1016 25. 33:1892–1899 3. Emergency Medicine Shock Research Network investigators: The effect of a quantitative resuscitation strategy on mortality in patients with sepsis: A meta-analysis. et al: Prehospital serum lactate as a predictor of outcomes in trauma patients: a retrospective observational study. Intensive Care Med 2007. Weil MH: Excess lactate: An index of reversibility of shock in human patients. Shapiro NI. Department of Health & Human Services. Jones AE. Clardy P. Bellomo R: Stress hyperlactataemia: Present understanding and controversy. 2014 [cited 2014 october]. Castro R. Dellinger RP: Lactate as a hemodynamic marker in the critically ill. Trzeciak S.