Table 2 Studies investigating predictive value of single or serial blood lactate assessment.

Shapiro et al., 2005 [2]

Prospective, observational

Moderate

Acceptable external validity. Acceptable internal validity.

Sample size of uncertain adequacy.

1278

ED patients with infection-related diagnosis

(≥ 18 years).

Venous lactate at admission to ED.

3-day and 28-day in-hospital mortality.

0-2.5; 2.5-4.0; > 4.0

Mortality rate increased with lactate levels: 0-2.5 mM: 4.9%; 2.5-4.0 mM: 9%; > 4.0 mM: 28%. Area under ROC curve for 3-day mortality was 0.80; for 28-day mortality it was 0.67. For 28-day mortality lactate between 2.5-4.0 mM had sens. 59% and spec. 71%. Lactate > 4.0 mM had sens. 36% and spec. 92%. For 3-day mortality lactate between 2.5-4.0 mM had sens. 76% and spec. 71%. Lactate > 4.0 mM had sens. 55% and spec. 91%.

Callaway et al., 2009 [3]

Retrospective

Moderate

Acceptable external validity. Acceptable internal validity.

Sample size of uncertain adequacy.

588

Normotensive blunt trauma patients

(≥ 65 years).

Venous lactate at admission to ED.

In-hospital mortality.

2.5

Compared to patients with admission lactate < 2.5 mM, lactate > 4.0 mM had OR 4.2 (2.4-7.5) for death. Area under ROC curve was 0.60.

Howell et al., 2007 [4]

Prospective, observational

Moderate

Acceptable external validity. Good internal validity.

Sample size of uncertain adequacy.

1287

ED patients with infection-related diagnosis.

Venous lactate at admission to ED.

28-day in-hospital mortality.

2.5-4.0; > 4.0

Admission lactate predicted 28-day mortality independently of blood pressure (p < 0.0001). Compared to lactate < 2.5 mM, lactate between 2.5-4.0 mM had OR 2.2 (1.1-4.2). Lactate > 4.0 mM had OR 7.1 (3.6-13.9). Area under ROC curve was 0.87.

Khosravani et al., 2009 [11]

Retrospective

Moderate

Acceptable external validity. Acceptable internal validity.

Sample size of uncertain adequacy.

9036

Intensive care patients, unspecified

(≥ 18 years).

Arterial or venous lactate at admission to ICU.

ICU mortality.

2.0

Lactate was an independent predictor of mortality:

2-5 mM: OR 1.94 (1.62-2.32); 5-10 mM: OR 3.38 (2.64-4.33); 10-15 mM: OR 4.41 (2.99-6.5); 15-20 mM: OR 7.58 (3.93-14.6); 20-max: OR 10.89 (4.85-24.48). All compared to control group with lactate < 2 mM

Nichol et al., 2010 [12]

Retrospective

Moderate

Acceptable external validity. Good internal validity.

Sample size of uncertain adequacy.

7155

Intensive care patients, unspecified.

Arterial lactate at admission to ICU.

Serial: interval unspecified.

In-hospital mortality.

2.0

Compared to lactate < 0.75 mM, admission lactate > 2.0 mM had OR for mortality at 2.1 (1.3-3.5, p = 0.01).

Sustained lactate between 0.75-1.0 mM had OR = 2.0 (p < 0.0001). Sustained lactate > 2.0 mM had OR = 3.7 (1.9-7.0, p < 0.0001).

Smith et al., 2001[13]

Prospective, observational

Moderate

Acceptable external validity. Acceptable internal validity.

Sample size of uncertain adequacy.

148

Intensive care patients, unspecified.

Arterial lactate at admission to ICU.

Serial: 24 hours later.

28-day in-hospital mortality.

1.5

Admission lactate > 1.5 mM was associated with 28-day mortality (p < 0.0001). Area under ROC curve = 0.78.

Patients with lactate > 1.0 mM at 24 hours have significantly higher mortality (p = 0.0001).

Suistomaa et al., 2000 [14]

Prospective, observational

Moderate

Acceptable external validity. Acceptable internal validity.

Sample size of uncertain adequacy.

98

Intensive care patients, unspecified.

Arterial lactate at admission to ICU.

Serial: every 2 hours the first 24 hours.

In-hospital mortality.

2.0

Median peak lactate for non-survivors was 5.3 mM (IQR, 1.9-7.5) vs. 1.9 mM (IQR, 1.3-2.9) for survivors, p = 0.003. Hyperlactatemia at admission to ICU was associated with higher mortality than hyperlactatemia that developed after admission (29.0% vs. 5.9%, p = 0.003).

Persistent hyperlactatemia (> 6 hours) was associated with higher mortality than transient hyperlactatemia (36.8% vs. 0%, p = 0.008).

Hatherill et al., 2000 [15]

Prospective, observational

Low

Acceptable external validity. Uncertain internal validity.

Probably underpowered study.

50

Children at ICU with shock, and initial hyperlactatemia.

Arterial lactate at admission to ICU.

Serial: at 24 hours, and additionally at unspecified intervals.

ICU mortality.

2.0

The area under the ROC curve for all values of lactate > 2 mM on admission was 0.59.

Persistent hyperlactataemia > 2 mM after 24 hours was associated with 93% mortality, as compared to 30% in those children whose lactate level had normalised. Persistent hyperlactataemia at 24 hours identified mortality with a likelihood ratio of 7, sens. 78%, and spec. 89%. The area under the ROC curve for lactate > 2 mM at 24 h after admission was 0.86.

Cerovic et al., 2003 [16]

Prospective, observational

Low

Acceptable external validity. Good internal validity.

Probably underpowered study.

94

Seriously injured patients defined as ISS ≥ 16 who survived ≥ 12 hours.

Arterial lactate at admission to ICU.

Serial: every 12 hours during the first 48 hours after admission.

In-hospital mortality.

2.0

Admission lactate was not a significant predictor for mortality.

Survivors exhibited a progressive decline in lactate levels, while lactate in the non-survivors remained broadly unchanged from the 12th hour after the first sampling.

del Portal et al., 2010 [17]

Retrospective

Moderate

Acceptable external validity. Good internal validity.

Sample size of uncertain adequacy.

1442

ED patients

(≥ 65 years).

Lactate at admission to ED.

30-day and 60-day in-hospital mortality.

2.0

Admission lactate were linearly associated with mortality (RR = 1.9 to 3.9) depending on lactate levels (p < 0.01).

Jansen et al., 2008 [18]

Prospective, observational

Moderate

Acceptable external validity. Acceptable internal validity.

Sample size of uncertain adequacy.

124

Patients who required urgent ambulance dispatching with systolic blood pressure < 100 mmHg, or respiration rate < 10, or > 29, or GCS < 14.

Pre-hospital venous or capillary lactate arrival on the site of injury and at admission to the hospital.

In-hospital mortality.

3.5

Mortality was significantly higher in patients with lactate ≥ 3.5 mM at the site of injury (41% vs. 12%; p < 0.001) or at admission to the hospital (47% vs.15%; p < 0.001). Lactate, on average, increased 0.1 mM in non-survivors, whereas in survivors, it decreased 0.6 mM (p = 0.044) Pre-hospital lactate had better prognostic value than vital signs alone.

Kaplan et al., 2004 [19]

Retrospective

Moderate

Acceptable external validity. Acceptable internal validity.

Sample size of uncertain adequacy.

282

Trauma patients with vascular injury (torso or extremity).

Arterial lactate at admission to trauma center.

28-day in-hospital mortality.

Admission lactate could discriminate survivors from non-survivors (3.6 mM vs. 11.1 mM, p < 0.001).

Pal et al., 2006 [20]

Retrospective

Moderate

Acceptable external validity. Acceptable internal validity.

Sample size of uncertain adequacy.

5995

Trauma patients, unspecified.

Arterial lactate at admission to trauma center.

In-hospital mortality.

2.0

Survivors had 3.0 mM, and non-survivors had 5.2 mM (p < 0.0001). Sens. and spec. of an elevated lactate was 85% and 38%, respectively. Area under ROC curve was 0.72. PPV was 4%.

Vandromme et al., 2010 [21]

Retrospective

Moderate

Acceptable external validity. Acceptable internal validity.

Sample size of uncertain adequacy.

2413

Trauma patients with systolic blood pressure between 90 and 110 mmHg.

Capillary or venous lactate at admission to hospital.

Need for ≥ 6 units packed red blood cells within 24 hours. In-hospital mortality.

2.5

Admission lactate was a better predictor for mortality and need for blood transfusion than systolic blood pressure (p < 0.0001). Lactate had area under ROC curve = 0.76 and systolic blood pressure had area under ROC curve = 0.61, p < 0.0001.

Arnold et al., 2009 [22]

Retrospective

Low

Acceptable external validity. Good internal validity.

Probably underpowered study.

166

ED patients diagnosed with severe sepsis.

(> 17 years)

Venous lactate at admission to ED.

Serial: interval unspecified.

In-hospital mortality

4.0

The mean initial lactate for survivors was 4.3 mM (SD = 2.6), while non-survivors had 4.7 mM (SD = 2.8), p = 0.41.

The mean serial lactate for survivors was 2.2 mM (SD = 1.6), while non-survivors had 3.6 mM (SD = 2.8), p < 0.001.

Guyette et al., 2011 [23]

Retrospective

Moderate

Acceptable external validity. Good internal validity.

Sample size of uncertain adequacy.

1168

Trauma patients transported by air.

(≥ 18 years)

Pre-hospital venous or capillary lactate was measured

In-hospital mortality

2.0

Pre-hospital lactate was median 3.8 mM (IQR, 2.8-6.1) in those who died and median 2.3 mM (IQR, 1.3-3.4) in those who survived to discharge, p < 0.0001.

Trzeciak et al. 2007 [24]

Prospective, observational

Low

Acceptable external validity. Uncertain internal validity.

Sample size of uncertain adequacy.

1177

ED patients with diagnosis of sepsis or infection.

(> 18 years).

Venous lactate at admission to ED.

In-hospital mortality and death within 3 days from measurement

4.0

Compared to baseline 0.0-2.0 mM, patients with lactate ≥ 4.0 mM had OR 6.1(3.7-10.5) for dying within 3 days from lactate measurement. Sens. 35%. Spec. 92%. Area under ROC curve 0.63. Equivalently lactate ≥ 4.0 mM had OR 3.0(2.0-4.6) for in-hospital death. Sens. 19%. Spec. 93%. Area under ROC curve was 0.56.

Kaplan et al., 2008 [25]

Retrospective

Low

Acceptable external validity. Good internal validity.

Probably underpowered study.

78

Patients with blunt, or penetrating trauma requiring intensive care

Arterial lactate at admission to ED.

28-day in-hospital mortality.

2.2

Admission lactate could not discriminate survivors from non-survivors (2.3 mM vs. 2.9 m, p = 0.24). Area under ROC curve was 0.6.

Van Beest et al., 2009 [27]

Prospective, observational

Moderate

Acceptable external validity. Acceptable internal validity.

Sample size of uncertain adequacy.

135

Patients with at least 2 symptoms of shock.

(≥ 18 years)

Pre-hospital capillary or venous lactate.

Length of stay, and in-hospital mortality.

4.0

Hyperlactamic patients had significantly higher mortality (12.2% vs. 44.3%, p = 0.002), longer LOS at ICU (p = 0.03), and LOS in hospital (p = 0.04). Area under ROC curve was 0.775. Lactate > 3.2 mM was the optimal cut off with sens. 75% and spec. 72%.

Nguyen et al., 2004 [28]

Prospective, observational

Moderate

Acceptable external validity. Good internal validity.

Sample size of uncertain adequacy.

111

Patients with severe sepsis or septic shock.

(> 18 years).

Arterial lactate at admission to ED.

Serial: at 6 hours.

60-day in-hospital mortality.

**

Lactate clearance less than 10% in 6 hours was associated with a higher 60-day mortality, than lactate clearance more than 10% (p = 0.007). Sens. 44.7% and spec. 84.4%.

Claridge et al., 2000 [29]

Prospective, observational

Moderate

Acceptable external validity. Good internal validity.

Sample size of uncertain adequacy.

381

Trauma patients requiring intensive care.

Lactate at admission to ED.

Serial: every 4-6 hours.

In-hospital mortality.

2.4

**

Patients with sustained hyperlactatemia (> 12 hours) had higher risk of infection (pneumonia etc.) than controls (69.8% vs. 40%, p < 0.001). The mortality rate for patients who developed infection was 7.9% vs. 1.9% (p < 0.05).

Jansen et al., 2009 [30]

Prospective, observational

Low

Acceptable external validity. Good internal validity.

Probably underpowered study.

394

Intensive care patients with sepsis, hemorrhage, or other conditions of low oxygen transport.

Arterial lactate at admission to ICU.

Serial: 12 and 24 hours after admission.

In-hospital mortality.

2.5

**

Reduction of lactate within 24 hours was associated with significantly lower mortality in the septic group (p = 0.003), but not in the other groups, (p = 0.42).

Kliegel et al., 2004 [31]

Retrospective

Moderate

Acceptable external validity. Acceptable internal validity.

Sample size of uncertain adequacy.

394

Patients resuscitated after cardiac arrest who survived > 48 hours.

Arterial lactate. First sample at admission to ED.

Serial: every 4-8 hours.

In- hospital mortality.

2.0

On admission survivors had 7.8 mM (IQR, 5.4-10.8) and non-survivors had 9 mM (IQR, 6.5-11.9), p < 0.01.

Lactate > 2 mM after 48 hours predicted mortality with a spec. of 86%, and poor neurologic outcome with a spec. of 87%. Sens. for both were 31%.

Jansen et al., 2010 [32]

Randomised, controlled trial

High

Acceptable external validity. Good internal validity. Premeditated and sufficient study size.

348

Intensive care patients, unspecified.

(≥ 18 years).

Arterial lactate at admission to ICU. (venous and capillary lactate was also allowed).

Serial: every 2 hours.

In-hospital mortality.

3.0

**

Hazard ratio in the intervention group was 0.61 (0.43-0.87, p = 0.006). Reduction of lactate below 2.0 mM was not associated with better outcome.

Jones et al., 2010 [33]

Randomised, controlled trial

High

Acceptable external validity. Good internal validity. Premeditated and sufficient study size.

300

Patients with severe sepsis and hypoperfusion or septic shock.

(≥ 17 years).

Venous lactate at admission to ICU.

Serial: every hour.

In-hospital mortality.

2.0

**

No difference in mortality was observed in patients treated by a protocol aiming at normalizing blood lactate compared to normalizing SvO2.

Blow et al., 1999 [43]

Prospective, intervention

Low

Acceptable external validity. Uncertain internal validity.

Probably underpowered study.

79

Hemo-dynamic stable trauma patients with ISS ≥ 20 and survival > 24 hours.

Lactate at admission to trauma centre.

Serial: interval unspecified.

In-hospital mortality.

2.5

**

Persistent hyperlactatemia after 24 hours was associated with increased mortality (p < 0.05).

Lee et al., 2008 [44]

Prospective, observational

Low

Acceptable external validity. Good internal validity.

Probably underpowered study.

126

Patients with severe sepsis, or septic shock.

(≥ 20 years).

Arterial lactate at admission to ED.

Serial: 4 hours later.

In-hospital mortality.

2.0

**

No significant difference in mortality was found between patients with elevated lactate compared to normal lactate, as long as pH was within normal limits.