• 
• 

Three Part Question

In [a patient presenting to the Emergency Department] is [lactate clearance better than the initial lactate level] at [predicting mortality]

Clinical Scenario

A 98 year old woman is brought to the Emergency Department on an unseasonally chilly summer's morning. The paramedics had picked her up from home in a peri-arrest state with a respiratory rate of 3, bradycardic, hypotensive, hypoglycaemic and hypothermic. Despite some degree of ridicule from Sister, you decide to resuscitate her. Her initial lactate, on an arterial blood gas, is 12.
4 hours later, she is sat up chatting with a lactate of 6. You wonder whether her rapid rate of lactate clearance, rather than the initial absolute level, is a predictor of survival.

Search Strategy

Medline 1950-Feb 2008 using the OVID interface
[(Lactate.mp. or exp Lactic Acid/) OR (exp Lactic Acid/ or exp Lactates/ or Lactate levels.mp. or exp Acidosis, Lactic/) OR (exp Lactates/ or exp Acidosis, Lactic/ or Lactic acidaemia.mp. or exp Lactic Acid/) OR (Lactic acidosis.mp. or exp Acidosis, Lactic/) OR (exp Lactic Acid/ or exp Lactates/ or Lactate determinants.mp.)] AND [(serial lactate.mp.) OR (Lactate clearance.mp.) OR (Lactime.mp.)]

Search Outcome

81 papers were found, of which 3 were relevant to the three part question

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Kleigel et al, 2004 Austria	394 post-cardiac arrest patients surviving 48 hours in an urban tertiary care emergency department. Inclusion criteria: survival at 48 hours from a cardiac arrest of cardiac, pulmonary, or cerebral origin, or caused by acute intoxication or electrolyte disorders; and if serum lactate levels had been repeatedly measured within the first 48 hours. Exclusion criteria: known liver disease, neurological deficits, or severe disabilities before the arrest, or if arrest was caused by trauma, severe bleeding, hypothermia, metabolic disorders, or unknown etiology.	Retrospective cohort (2b)	Survival at 6 months	Differences in lactate levels between survivors and non-survivors on admission, at 24 and at 48 hours reached statistical significance (p<0.01, p<0.01 and p<0.001 respectively). After 48 hours, levels were normal in 86% of survivors c.f. 69% of non-survivors (p<0.001). Patients who died within 6 months were more likely to still have abnormal lactate levels at 48 hours after ROSC (OR 2.7). A multivariate regression model found serum lactate levels after 48 hours to be an independent predictor of mortality (OR 1.49 increase per mmol/l). A serum lactate > 2mmol/l after 48 h had a specificity for mortality of 86%.	Retrospective analysis. No allowance made for the passage of time over the 10 year period. No sample size estimate. Only looked at those surviving at least 48 hours. Excluded patients with incomplete data and those lost to follow-up.
Nguyen et al, 2004 USA	111 adult patients presenting to the emergency department of an urban tertiary care centre with severe sepsis or septic shock. Exclusions: MI, pulmonary oedema, haemorrhagic shock, trauma, seizures, pregnancy, DNR order, or requiring immediate surgery.	Prospective observational case series / cohort (1b)	Primary: In-hospital mortality	Initial lactate levels in survivors and non-survivors were 6.1 +/- 4.4 v. 8.0 +/- 4.7 (p=0.01). Survivors had a lactate clearance at 6 h of 38.1 +/- 34.6% v. 12 +/- 51.6% for non-survivors (p<0.05). Multivariate logistic regression modelling found only lactate clearance to be significantly associated with a decreased mortality rate (p=0.04). There was an approximate 11% decrease in likelihood of mortality for each 10% increase in lactate clearance. The high clearance (>10% clearance at 6 h) group had a 52% relatively lower in-hospital mortality rate c.f. the low clearance (<10% clearance at 6 h) group (p<0.001). This difference in mortality rate was also observed at 30 (37.5 v. 67.7%) and 60 (42.5 v. 71%) days (p=0.004 and p=0.007, respectively).	No sample size estimate. Well resourced department with a high level of patient acuity and an ICU admission rate double the national average, therefore results may not be generalizable to other emergency departments.
Donnino et al, 2007 USA	79 post-cardiac arrest patients >18 years of age, admitted to an urban emergency department. Exclusions: traumatic arrest, successful resuscitation by bystanders prior to arrival of EMS, cardiac arrest in the presence of pre-hospital or in-hospital providers.	Retrospective cohort study (2b)	Primary: 24-h survival	24-h survival: There was no difference between survivors and non-survivors with respect to initial lactate levels. Survivors had a mean lactate clearance at 6 h of 57.7 +/- 23% c.f. non-survivors who had a mean clearance of 39.6 +/- 37.5% (p<0.05). Survivors had a mean lactate clearance at 12 h of 70.7 +/- 19% c.f. non-survivors who had a mean clearance of 40.1 +/- 27% (p<0.05). A multivariate regression model showed high lactate clearance at 12 h to be singularly predictive of 24-h survival (p<0.05).	Retrospective analysis. Some missing data. No sample size estimate. Long down times with most common rhythms being PEA and asystole, so may not be applicable to VF and pulseless VT arrests.
			Secondary: survival to hospital discharge	Survival to hospital discharge: Clearance was significantly higher in survivors at 6 h (62.8 +/- 17.2% v. 49.4 +/- 31.2%, p=0.05), 12 h (77.7 +/- 12.1% v. 59.6 +/- 29.6%, p<0.01), and 24 h (85.5 +/- 7.3% v. 75.5 +/- 17.8%, p=0.01).

Comment(s)

Although there were statistical differences in the initial lactate level between survivors and non-survivors according to the first two papers, subsequent multivariate logistical modelling found only lactate clearance to be significantly associated with a decreased mortality rate. Furthermore, Nguyen et al. were able to quantify this as an 11% decrease in in-hospital mortality for each 10% increase in lactate clearance at 6 hours; Kleigel et al. also quantified this as an increased mortality rate at 6 months with an odds ratio of 1.49 per mmol/L of lactate at 48 hours. The most recent paper by Donnino et al. found no difference between survivors and non-survivors with respect to initial lactate level. Multivariate regression analysis again confirmed that lactate clearance at 12 hours was singularly predictive of 24 hour survival. In addition, lactate clearances at 6, 12 and 24 hours were the only factors significantly associated with survival to hospital discharge.

Clinical Bottom Line

Multivariate regression showed lactate clearance, rather than the initial lactate level per se, to be the only independent predictor of mortality in all 3 papers.

References

1. Kleigel A., Losert H., Sterz F. et al. Serial lactate determinations for prediction of outcome after cardiac arrest. Medicine 83 (5) Sept 2004; 274-9.
2. Nguyen H.B., Rivers E.P., Bernhard P. et al. Early lactate clearance is associated with improved outcome in severe sepsis and septic shock. Crit Care Med 2004 Aug;32(8):1637-42.
3. Donnino M.W., Miller J., Goyal N. et al. Effective lactate clearance is associated with improved outcome in post-cardiac arrest patients. Resuscitation 75(2) Nov 2007: 229-34