a b s t r a c t

Objectives: Our study aims to investigate the role of initial Venous lactate in predicting the probability of clinical de- terioration and 30-day mortality in nonelderly sepsis patients with Acute infections, without hemodynamic shock. Methods: We enrolled emergency department patients aged 18 to 65 years with acute major infections, but without organ hypoperfusion, and obtained a single venous lactate measurement at initial presentation. As the primary end point, the eligible patients were tracked for the need for vasopressor or mechanical ventilation in the next 72 hours. The patients’ venous lactate and related risk factors were analyzed. We also followed the cohort and the pre- dictors to investigate their prognostic role for 30-day mortality.

Results: Of 392 patients, 74 required vasopressor/MV, and 388 patients were available for mortality analysis. An ini- tial lactate greater than or equal to 2 mmol/L was the strongest independent predictor for the requirement of vaso- pressor/MV (adjusted odds ratio, 6.2; 95% confidence interval, 3.4-11.3). The other independent risk factors were immunosuppressive Drug users and positive blood culture. However, the initial lactate was not associated with 30-day mortality. The factors that were associated with mortality were the use of vasopressor/MV, active malignan- cy, Rapid Emergency Medicine Score greater than or equal to 6, and hospitalization within 90 days.

Conclusions: In nonelderly sepsis patients with stable hemodynamic, elevated venous lactate (>=2 mmol/L) was as- sociated with an increased probability of the need for vasopressor/MV. However, unfavorable medical histories and the severity of physiologic changes may be associated with short-term mortality to a greater extent than the single value of initial lactate.

(C) 2015

1. Introduction

Blood lactate is a biomarker that can predict organ failure and mor- tality in emergency department (ED) patients with sepsis [1]. Regarding its predictive value for clinical deterioration in normotensive sepsis pa- tients, a recent prospective observational study of patients with moder- ately high Serum lactate levels reported that one-fourth of these patients eventually developed progressive organ dysfunction or re- quired vasopressor treatment or mechanical ventilation (MV) [2]. How- ever, comparable results were not reported by others [3]. Furthermore, low serum lactate levels (b 2.5 mmol/L) were found in up to 50% of overt septic shock patients, even in vasopressor-dependent cases [4-6]. Be- cause of its prognostic value for mortality, several studies have shown that blood lactate can also predict death in sepsis patients, especially

? Conflict of interest declaration: This study was funded by KCMH and the Faculty of Medicine of Chulalongkorn University. The authors have no conflict of interest to declare.

* Corresponding author. Tel/fax: +66 2 256 5389.

E-mail addresses: khrongwong.m@chulahospital.org (K. Musikatavorn), docpup9@hotmail.com (S. Thepnimitra), andykrub@gmail.com (A. Komindr), patiputt@gmail.com (P. Puttaphaisan), na_nadol@hotmail.com (D. Rojanasarntikul).

¹ Principal investigator.

in elderly and critically ill patients [7-11]. Differences in lactate kinetics between age groups have been demonstrated in the literature [12,13]. Thus, clinical data on blood lactate’s predictive value for progression to overt shock or even mortality in those who are younger and not crit- ically ill at initial presentation remain limited. This specific subgroup of patients is typically considered to have a low risk of short-term mortal- ity. The primary aim of our study was to investigate the value of initial venous lactate for predicting the risk of clinical deterioration to an extent that requires intensive care unit (ICU) treatment among nonelderly, “nontoxic-appearing” sepsis patients in the ED who have acute major infection without hemodynamic shock. The secondary aim was to evaluate its prognostic value for 30-day mortality in this se- lected population.

1. Patients and methods
   1. Study design and setting

We conducted this prospective observational study in the ED of King Chulalongkorn Memorial Hospital (KCMH), an urban, 1500-bed, university-affiliated tertiary care referral hospital. Our ED sees more

http://dx.doi.org/10.1016/j.ajem.2015.04.010

0735-6757/(C) 2015

than 60 000 new cases per year. Our institutional review board ap- proved this study, and written informed consent was obtained for blood sampling. participant recruitment began in April 2013 and con- cluded at the end of October 2014. This cohort study was registered in ClinicalTrials.gov (identifier NCT01947127).

Population

Patients ranging from 18 to 65 years of age who presented to the ED and who had received a confirmed or suspected diagnosis of acute (onset within the past 7 days) major infection during the study period were consecutively enrolled in this study. Patient eligibility was assessed by the treating ED physicians for all work shifts. Major infec- tions included acute pyelonephritis, Acute bronchitis/pneumonia, acute hepatobiliary tract infections, Intra-abdominal abscesses, menin- gitis and other central nervous system infections, soft tissue infections involving tissue below the dermis or covering more than 10 cm² of the skin surface, significant tropical infection, dengue fever, leptospiro- sis, typhus fever, or high fever (N 38.5?C) from any infectious source. The eligible patients must have had stable hemodynamics, as indicated by at least 2 repeated measurements of systolic blood pressure (SBP) higher than 100 mm Hg and mean arterial pressure (MAP) higher than 70 mm Hg at initial presentation.

Patients were excluded based on the following criteria: (1) overt organ hypoperfusion, for example, cold, clammy or mottling skin, altered mental status; (2) Glasgow Coma Scale less than or equal to 12 or a decrease of larger than 1 compared with baseline; (3) peripheral pulse oximetry less than or equal to 92% in ambient air; (4) received more than 10 mL/kg intravenous fluid before lactate blood sampling;

(5) received an intravenous antibiotic for more than 1 hour before

blood lactate sampling; (6) SBP less than 100 mm Hg or required me- chanical ventilator support within 1 hour after ED presentation;

(7) minor infection (eg, uncomplicated upper respiratory tract infection, gastroenteritis, or minor skin infection); (8) current use of antiretroviral drugs; (9) experienced or suspected to have experienced seizures in the past 72 hours; (10) noninfectious disease as the final diagnoses; and

(11) duplicated cases that were previously included in this study.

Study definitions

We used the definitions of sepsis and other related spectra proposed in the literature [14]. Briefly, we used the systemic inflammatory re- sponse syndrome (SIRS) criteria, as follows: (1) core temperature great- er than 38.3?C or less than 36?C, (2) heart rate greater than 90 beats per minute, (3) respiratory rate greater than 20 breaths per minute,

(4) white blood cell count greater than 12 000 or less than 4000 per mm³ or greater than 10% immature band. The patients who met 2 or more of these SIRS criteria were defined as having sepsis. Uncomplicated sepsis was defined as sepsis without evidence of shock or end-organ dysfunction. Septic shock was defined as tissue hypoperfusion, including hypotension (SBP b 90 mm Hg or MAP b 65 mm Hg) persisting despite initial fluid challenge or a blood lactate concentration of 4 mmol/L or higher. In our protocol, we defined the need for ICU treatment as the development of shock requiring vasopressor treatment or MV within 72 hours after patient enrollment.

Data collection

After an eligible patient was identified and informed consent was obtained, venous blood lactate and other acid-base parameters (pH and Base deficit) were obtained (together with other blood work) via standard venipuncture within less than 2 minutes of tourniquet applica- tion before the administration of antibiotics or intravenous fluid, if re- quired. Complete blood count, blood urea nitrogen level, creatinine level, coagulograms, standard liver function tests, plasma glucose level, electrolyte levels, and aerobic blood cultures were the compulsory

workups included in our protocol. First, a blood sample was drawn within 1 hour after the emergency physicians’ orders. Our treatment protocol permitted treating physicians to treat patients who met 2 SIRS criteria or more with at least 20 mL/kg of normal saline, if feasible, based on their physiologic status and with empirically selected antibi- otics within 1 hour. For those whose initial lactate levels were greater than or equal to 4 mmol/L, prompt administration of 30 mL/kg normal saline and antibiotics was mandatory, along with Close monitoring, in all cases because they were at high risk for fatality. Other essential blood chemistry and microbiological investigations, imaging studies, treatments, invasive procedures (eg, central venous catheterization and Arterial lines), the threshold for administering vasopressors/MV, or admission of the patients were at the discretion of the treating phy- sicians according to the clinical profiles or diagnoses of the patients. The demographic data, initial physiologic vital signs, onset of infection, sites of infection, predisposing conditions, and laboratory and microbio- logical results were recorded (see Table). Standard blood pressure mea- surement was obtained from the patients at least twice, and the lowest value of SBP was used. The Infection sites, which were determined by study physicians who were blinded to the lactate results, were catego- rized as “infection likely” or “infection unlikely” according to the clinical summaries, imaging findings, and the identification of pathogens. Patients classified as “infection unlikely” were finally excluded from the study.

We used the Rapid Emergency Medicine Score (REMS) system to cat- egorize the disease severity of our patients because it can predict short- term mortality of sepsis patients in a nonsurgical department [15-17]. Furthermore, the REMSs could be promptly assessed and calculated from measurable physiologic parameters at initial ED triage. The patients were classified as low risk, intermediate risk, or high risk according to an REMS of less than 6, 6 to 13, or greater than 13, respectively.

Method and device used for blood lactate measurement

All blood lactate and other acid-base parameters were measured using a Food and Drug Administration-approved Epoc point-of-care blood analyzer (Epocal, Inc, Ottawa, ON, Canada) within 5 minutes of obtaining the venous blood sample without any preservative. The re- sults were consistently obtained within 30 seconds of sample delivery.

Outcome measurements

The venous Blood lactate levels were predefined before patient recruitment as normal (0-1.9 mmol/L), moderate (2-3.9 mmol/L), or high (>= 4 mmol/L). We defined the patients who clinically needed vaso- pressor treatment or MV within 72 hours after ED presentation that required ICU care as our primary outcome measure. The ED or inhospital medical records of the admitted patients were followed for 72 hours after ED admission to identify those who needed vasopressor treatment/MV to maintain their vital signs. Length of hospital stay and survival status were collected and recorded. For the 30-day mortality analysis, we searched the electronic database for inpatient and outpa- tient clinical records and telephone follow-ups of the patients or their contact personnel in every case at 30 days after the day of ED presenta- tion to identify the deceased cases. All telephone interviews and data input processes were completed by well-trained noninvestigators who were also blinded to the lactate levels of the patients.

Data analysis

We used SPSS software version 17.0 for Windows (Chicago, IL) for all analyses. Normally distributed and nonnormally distributed continuous data were analyzed using the 2-tailed independent t test and Mann- Whitney U test, respectively. The ?² with odds ratio (OR) calculation was used to compare the nonparametric values, and binary logistic re- gression was used to determine the adjusted ORs of the clinical factors associated with the primary outcome measure. The variables in the

Table

Demographic data of the cohort and comparison between patients requiring and not requiring ICU treatment

Parameters of interest at initial ED presentation	Total, n = 392	Patients not requiring ICU treatment,	Patients requiring ICU treatment,	P value, OR
	(100%)	n = 318 (100%)	n = 74 (100%)	(95% CI)
Male sex	208 (53.1)	161 (50.6)	47 (63.5)	.045?
Age (y), means +- SD	44.0 +- 14.2	42.8 +- 14.4	48.8 +- 12.1	.001?,a
Age N 40 y	240 (61.2)	184 (57.9)	56 (75.7)	.005?
Onset of infections (h), medians, (IQR)	24 (17-72)	24 (13.5-72)	48 (24-72)	.037?,b
Internal medicine cases	381 (97.2)	308 (96.9)	73 (98.6)
Sites of infections
Urinary	64 (16.3)	57 (17.9)	7 (9.5)
Respiratory	76 (19.4)	54 (17.0)	22 (29.7)
Hepatobiliary/gastrointestinal	57 (14.5)	51 (16.0)	6 (8.1)
Skin/soft tissue/bone/joint	19 (4.8)	11 (3.5)	8 (10.8)
Primary bacteremia	65 (16.6)	45 (14.2)	20 (27.0)
Other organ localizing infections	24 (6.1)	20 (6.2)	4 (12.6)
Other Systemic Infections	87 (22.2)	80 (25.2)	7 (9.5)
underlying conditions 264 (67.3)		206 (64.8)	58 (78.4) .025?, 2.0 (1.1-3.6)
Diabetes mellitus	71 (18.1)	53 (16.7)	18 (24.3)	.12
Chronic renal failure	36 (9.2)	32 (10.1)	4 (5.4)	.21
Chronic liver diseases	34 (8.7)	26 (8.2)	8 (10.8)	.47
Active malignancy	124 (31.6)	97 (30.5)	27 (36.5)	.32
Receiving chemotherapy	91 (23.2)	72 (22.6)	19 (25.7)	.58
Neutropenia	47 (12.0)	34 (10.7)	13 (17.6)	.10
Immunosuppressive drug users	45 (11.5)	31 (9.7)	14 (18.9)	.026?, 2.1 (1.1-4.3)
Hospitalization within 90 d	148 (37.8)	111 (34.9)	37 (50.0)	.016?, 1.9 (1.1-3.1)
Antibiotic treatment within 30 d	157(40.1)	127 (39.9)	30 (40.5)	.92
Temperaturec (?C), means +- SD	38.6 +- 1.1	38.6 +- 1.0	38.5 +- 1.3	.74a
Respiratory rate N 24 per min	54 (13.8)	33 (10.4)	21 (28.4)	.00?
Peripheral oxygen saturation at ambient air (%)	96.5 +- 3.3	96.8 +- 3.2	95.2 +- 3.2	.00?
Pulse rate per minute	110.4 +- 19.6	109.4 +- 19.7	114.7 +- 18.7	.037?,a
N 90 beats per minute	262 (66.8)	208 (65.4)	54 (73.0)	.21
SBP (mm Hg), means +- SD	127.7+-23.7	129.1+-24.2	121 +- 20.7	.02?,a
MAP (mm Hg), means +- SD	74.2 +- 15.0	92.9 +- 16.3	88.6 +- 17.4	.04?,a
White blood cell count (103/mm3), medians (IQR)	10.4 (5.7-15.6)	10.3 (5.7-15.0)	11.6 (5.9-17.3)	.42
serum bicarbonate (mEq/L)	22.1 +- 10.0	22.6 +- 10.7	19.7 +- 5.8	.02?,a
Venous lactate levels (mmol/L), means +- SD	2.0 +- 1.4	1.7 +- 1.1	3.1 +- 1.7	.000?,a
>= 2.0	143 (36.5)	90 (28.3)	53 (71.6)	.000?, 6.4 (3.6-11.2)
>= 4 (b4 mmol/L, reference)	36 (9.2)	16 (5.0)	20 (27.0)	.000?, 6.9 (3.4-14.3)
venous pH	7.42 +- 0.07	7.43 +- 0.06	7.41 +- 0.08	.12a
Base deficit (mmol/L), means +- SD	0.1 +- 4.4	-0.4 +- 3.9	2.8 +- 5.6	.000?,a
Positive blood culture growth	71 (18.2)	46 (14.5)	25 (33.8)	.000?, 3.0 (1.7-5.3)
Gram positive	24 (6.1)	15 (4.7)	9 (18.9)
Gram negative	47 (12.0)	31 (9.7)	16 (21.6)
SIRS criteria >= 3 points	257 (65.6)	204 (64.2)	53 (71.6)	.22
REMSd >= 6	53 (13.5)	32 (10.1)	21 (28.4)	.000?, 3.5 (1.9-6.6)
Days of hospital stay, medians (IQR)	4 (2-10)	3 (2-7.25)	10 (6-21.5)	.000?,b

Abbreviation: IQR, interquartile range (Q₁-Q₃).

^a By independent t test.

^b Mann-Whitney U test.

^c Via tympanic membrane route.

^d Maximal score found in this study was 9 (range, 0-9).

* P b .05.

univariate analysis showing a P value of less than .15 were selected for multivariate analysis, and the backward elimination method using a threshold P value of .05 was performed to remove variables from the multivariate model [18]. We identified the clinical factors associated with 30-day all-cause mortality in our cohort using a Cox proportional hazards model with corresponding curves. All tests were 2 sided, and P b .05 was considered to be statistically significant.

1. Results

In total, 458 patients were assessed for eligibility, revealing that 392 and 388 eligible cases were available for primary and secondary out- come analyses, respectively. The patient flow chart and reasons for ex- clusion are shown in Fig. 1.

Patients requiring and not requiring ICU treatment

In this study, 74 (18.9%) of 392 patients required ICU treatment and were administered vasopressors/MV within 72 hours after enrollment. The demographic and clinical characteristics were compared; these results

are summarized in Table. The patients requiring ICU treatment tended to be male and were slightly older than those not requiring ICU treatment. There were also slight differences in SBP, MAP, peripheral pulse oximetry, and bi- carbonate levels between the both groups, although these values had large overlapping ranges and were considered to be clinically insignificant. Based on the univariate analysis, the factors that were associated with the need for ICU treatment were male sex, older than 40 years, immunosuppressive drug use, hospitalization within 90 days before presentation, positive blood culture results, and REMS greater than or equal to 6. In the multivar- iate model, we found that initial venous lactate level greater than or equal to 2 mmol/L was the strongest independent predictor of the need for ICU treatment (adjusted OR, 6.2; 95% confidence interval [CI], 3.4-11.3). The other independent risk factors identified in this study were immunosup- pressive drug use (adjusted OR, 2.7; 95% CI, 1.2-6.1), and positive blood cul- ture results (adjusted OR, 2.3; 95% CI, 1.2-4.4).

Survivors and nonsurvivors

There were 31 fatal cases among the 388 patients eligible for mortal- ity analysis. In this study, we did not find an association between the

Fig. 1. The diagram shows patient flow during the study period.

initial venous lactate levels and 30-day mortality. The factors that were associated with 30-day mortality were treatment with vasopressors/ MV, active malignancy, REMS greater than or equal to 6 and hospitalization within the preceding 90 days (hazard ratio [HR], 5.7, 3.9, 3.6, and 3.6, re- spectively; all P b .05). The survival probability curves of the patients strat- ified according to the related clinical covariates are shown in Fig. 2A to E.

1. Discussion

Blood lactate is a well-established biomarker for the prediction of mortality in general sepsis patients, especially in the critically ill and the elderly. Advanced age, clinical severity at presentation, and comor- bidities definitely predispose patients to a higher risk of physiologic de- terioration and have detrimental effects on mortality. Therefore, these risks are likely to confound the consequences of occult organ hypoper- fusion. In contrast to previous studies, in this study, we deliberately re- cruited patients who were younger and whose physiologic parameters were initially in the “low risk” subgroup. We found that elevated venous lactate levels (>= 2 mmol/L) in these “low risk” sepsis ED patients were the strongest independent predictor of the need for vasopressors/MV, despite early treatment. This finding suggested the significant value of an elevated blood lactate level for predicting the need for ICU treatment in sepsis ED patients with low mortality risk at presentation.

However, we did find that the single value of the initial lactate level was not associated with 30-day mortality in our cohort. Our findings in- dicate that the mortality risk in nonelderly, non-shock sepsis patients is more likely due to unfavorable predisposing conditions such as active

malignancy and recent hospitalization and the severity of physiologic changes, as represented by the REMS and use of vasopressors/MV, than to the initial lactate level. A predisposing profile that may indicate susceptibility to more virulent pathogens or poor baseline status leading to repeated hospitalization, such as the acquisition of health casssociated infections, has been reported to be a significant factor asso- ciated with mortality in bacteremic patients [19-22]. Furthermore, a previous study has shown that hyperlactatemia with acidosis is, in fact, a predictive factor of mortality [23]. Most of our patients ulti- mately survived with early and aggressive treatment; however, those with elevated blood lactate levels did show a higher risk of requiring ICU support.

Occult hyperlactatemia in sepsis patients with stable hemodynamics is not uncommon. Several publications have demonstrated the potential negative impact of the “pre-shock” lactate level on the clinical outcomes of stable sepsis patients [2,24]. Although a previous study reported that mildly elevated lactate levels may have resulted from impaired lactate clearance rather than the overproduction from organ hypoperfu- sion [25], our study demonstrates that regardless of the extent, elevated blood lactate levels are not benign, even in stable patients. We advocate the measurement of venous lactate at the point of care as a rapid screen- ing tool to help identify sepsis patients with stable hemodynamics who may need future ICU treatment. However, because the single value of initial blood lactate may not be associated with short-term mortality, treating physicians should also consider other significant clinical vari- ables of their sepsis patients (eg, status of active malignancy; acquisition of health care-associated infections; overall physiologic manifestations

Fig. 2. Cox proportional hazards model estimates 30-day survival probability separated lines by the significant clinical factors listed as follows: use of vasopressor/MV (HR, 5.7; 95% CI, 2.5- 13.2) (A), active malignancy (HR, 3.9; 95% CI, 1.6-9.1) (B), REMS greater than or equal to 6 (HR, 3.6; 95% CI, 1.7-7.9) (C), and hospitalization within the preceding 90 days (HR, 3.6; 95% CI, 1.5-8.8) (D). E, Graph indicates that the initial lactate levels were not significantly associated with the survival (P N .05).

or scores; and, likely, the achievement of lactate clearance) to predict and optimize the survival probability.

Although patients with normal blood lactate levels were likely to have an uneventful clinical course, some of these patients developed septic shock during a later course in their illness. Therefore, for the early detection

of septic shock, close observation of not only non-shock sepsis patients with elevated lactate levels but also “non-lactatemic” patients is mandatory. Preventive therapeutic strategies to avoid clinical deterioration in non- shock sepsis patients witha modest elevation of blood lactate remain poor- ly established, and these strategies have not been standardized.

1. Limitations

Although most of the patients in this study were in a younger age group, were initially classified as having a “low risk of death” according our eligibility protocol, and were unlikely to experience clinical deterio- ration compared to those with a more catastrophic clinical presentation, we acknowledge the potential limitations of our study. First, this was an observational study that did not evaluate the effectiveness or aggres- siveness of the treatment strategies (eg, lactate clearance), and these factors may have affected the outcomes. Second, this study included se- lected patients in a single urban tertiary care teaching hospital, and the patients may not be representative of the general population or of those who receive other levels of medical care. Although the results of our study validate the predictive value of the initial level of blood lactate, the number of events (no. of patients who required ICU treatment or died) was relatively small in our cohort; thus, the magnitude of the dif- ferences may have been overestimated. Finally, although the predispos- ing conditions of the patients were analyzed in this study, the severity staging of those comorbidities (eg, staging of cancer cases) was not well categorized and, thus, may have affected mortality in addition to physiologic or metabolic pathway derangement.

1. Conclusions

In nonelderly sepsis patients with stable hemodynamics at ED pre- sentation, elevated venous lactate levels (>= 2 mmol/L) are associated with an increased risk of the need for ICU support, despite prompt ther- apies. Our study highlights the clinical measurement of venous blood lactate at the point of care in the ED as a screening tool for the prediction of the need for future ICU treatment in stable sepsis ED patients. Howev- er, unfavorable medical histories (active malignancy or recent hospital- ization) and the severity of physiologic changes (REMS and the use of vasopressors/MV) may be more strongly associated with short-term mortality than the initial lactate level alone. Further investigation of early effective treatment strategies for such patients with elevated blood lactate levels is required to prevent the clinical deterioration of these patients.

Acknowledgment

The authors would like to thank all personnel at the ED of KCMH for their helpful coordination, timely work, and dedicated medical treat- ment. The authors also express their deepest gratitude to the patients and their relatives who participated in this study.

References

1. Shapiro NI, Howell MD, Talmor D, Nathanson LA, Lisbon A, Wolfe RE, et al. Serum lactate as a predictor of mortality in emergency department patients with infection. Ann Emerg Med 2005;45:524-8.
2. Arnold RC, Sherwin R, Shapiro NI, O’Connor JL, Glaspey L, Singh S, et al. Multicenter observational study of the development of progressive organ dysfunction and

therapeutic interventions in normotensive sepsis patients in the emergency depart- ment. Acad Emerg Med 2013;20:433-40.

1. Jones AE, Shapiro NI, Trzeciak S, Arnold RC, Claremont HA, Kline JA. Lactate clearance vs Central venous oxygen saturation as goals of early sepsis therapy: a randomized clinical trial. JAMA 2010;24(303):739-46.
2. Wacharasint P, Nakada TA, Boyd JH, Russell JA, Walley KR. Normal-range blood lac- tate concentration in septic shock is prognostic and predictive. Shock 2012;38:4-10.
3. Dugas AF, Mackenhauer J, Salciccioli JD, Cocchi MN, Gautam S, Donnino MW. Preva- lence and characteristics of nonlactate and lactate expressors in septic shock. J Crit Care 2012;27:344-50.
4. Trzeciak S, Dellinger RP, Chansky ME, Arnold RC, Schorr C, Milcarek B, et al. Serum lactate as a predictor of mortality in patients with infection. Intensive Care Med 2007;33:970-7.
5. Glickman SW, Cairns CB, Otero RM, Woods CW, Tsalik EL, Langley RJ, et al. Disease progression in Hemodynamically stable patients presenting to the emergency department with sepsis. Acad Emerg Med 2010;17:383-90.
6. del Portal DA, Shofer F, Mikkelsen ME, Dorsey Jr PJ, Gaieski DF, Goyal M, et al. Emer- gency department lactate is associated with mortality in older adults admitted with and without infections. Acad Emerg Med 2010;17:260-8.
7. Khosravani H, Shahpori R, Stelfox HT, Kirkpatrick AW, Laupland KB. Occurrence and adverse effect on outcome of hyperlactatemia in the critically ill. Crit Care 2009; 13(3):R90.
8. Howell MD, Donnino M, Clardy P, Talmor D, Shapiro NI. Occult hypoperfusion and mortality in patients with suspected infection. Intensive Care Med 2007;33:1892-9.
9. Nichol AD, Egi M, Pettila V, Bellomo R, French C, Hart G, et al. Relative hyperlactatemia and hospital mortality in critically ill patients: a retrospective multi-centre study. Crit Care 2010;14(1):R25.
10. Tzankoff SP, Norris AH. Age-related differences in lactate distribution kinetics fol- lowing maximal exercise. Eur J Appl Physiol Occup Physiol 1979;42:35-40.
11. Beneker R, Hutler M, Jung M, Leithauser RM. Modeling the blood lactate kinetics at maximal short-term exercise conditions in children, adolescents, and adults. J Appl Physiol 2005;99:499-504.
12. Dellinger RP, Levy MM, Rhodes A, Annane D, Gerlach H, Opal SM, et al. Surviving sepsis campaign: international guidelines for management of severe sepsis and sep- tic shock: 2012. Crit Care Med 2013;41:580-637.
13. Olsson T, Terent A, Lind L. Rapid Emergency Medicine Score can predict long-term mortality in nonsurgical emergency department patients. Acad Emerg Med 2004; 11:1008-13.
14. Olsson T, Terent A, Lind L. Rapid Emergency Medicine score: a new prognostic tool for in-hospital mortality in nonsurgical emergency department patients. J Intern Med 2004;255:579-87.
15. Ghanem-Zoubi NO, Vardi M, Laor A, Weber G, Bitterman H. Assessment of disease- severity scoring systems for patients with sepsis in general internal medicine de- partments. Crit Care 2011;15:R95.
16. Sun GW, Shook TL, Kay GL. inappropriate use of bivariable analysis to screen risk factors for use in multivariable analysis. J Clin Epidemiol 1996;48:907-16.
17. Cheong HS, Kang CI, Kwon KT, Heo ST, Wi YM, Kim ES, et al. Clinical significance of healthcare-associated infections in community-onset Escherichia coli bacteraemia. J Antimicrob Chemother 2007;60:1355-60.
18. Bassetti M, Trecarichi EM, Mesini A, Spanu T, Giacobbe DR, Rossi M, et al. Risk factors and mortality of healthcare-associated and community-acquired Staphylococcus aureus bacteraemia. Clin Microbiol Infect 2012;18:862-9.
19. Horcajada JP, Shaw E, Padilla B, Pintado V, Calbo E, Benito N, et al. Healthcare-asso- ciated, community-acquired and hospital-acquired bacteraemic urinary tract infec- tions in hospitalized patients: a prospective multicentre cohort study in the era of Antimicrobial resistance. Clin Microbiol Infect 2013;19:962-8.
20. Hoenigl M, Wagner J, Raggam RB, Prueller F, Prattes J, Eigl S, et al. Characteristics of hospital-acquired and community-onset blood stream infections. South-East Austria. PLoS One 2014;9:e104702.
21. Lee SW, Hong YS, Park DW, Choi SH, Moon SW, Park JS, et al. Lactic acidosis not hyperlactatemia as a predictor of in hospital mortality in septic emergency patients. Emerg Med J 2008;25:659-65.
22. Puskarich MA, Illich BM, Jones AE. Prognosis of emergency department patients with suspected infection and intermediate lactate levels: a systematic review. J Crit Care 2014;29:334-9.
23. Levraut J, Ciebiera JP, Chave S, Rabary O, Jambou P, Carles M, et al. Mild hyperlactatemia in stable septic patients is due to impaired lactate clearance rather than overproduction. Am J Respir Crit Care Med 1998;157:1021-6.