sinaiem.org

I gave a talk last week on the workup of nonsevere sepsis that referenced a bunch of little papers, and a few big ones… We’ll leave the discussion of the landmark 2001 EGDT severe-sepsis talk for another time (sigh). Right now I just wanted to go over an Annals paper (AEM Vol 45, No 5, May ‘05) by Shapiro et al from Beth-Israel Deaconess, about lactate in the ED. They were looking at the value of ED serum lactate levels as a predictor of later mortality – echoing studies on ICU lactate and mortality for patients with septic shock, burns, or trauma. This was a prospective cohort study of 1278 consecutive adult urban ED patients that were admitted with any kind of infection-related diagnosis, who had serum lactacte levels drawn. They looked at mortality within 3 days and 28 days.

They found 8.2% of these patients died, with just over half (4.3%) dying within the first three days. Lactate levels and mortality levels correlated: those with a lactate <2.5 mmol/L had a 4.9% mortality rate, those between 2.5  and 4.0 mmol/L had a 9% mortality rate, and in those unfortunate patients with a lactate greater than 4.0 mmol/L, death occurred in 28.4% of them (38 of 134). They found lactate > 4 had a 92% specificity for death (95%CI was 90-93%) and 36% sensitivity (95%CI was 27-45%), and more sensitive for death within 3 days.

Of note: they say they tried to raise awereness of ordering lactate on any patient suspected of infection, encouraging a lactate level with every blood culture. It was always venous lactate. Also, suspicion of infection was the criterion of inclusion, not final diagnosis (some people probably had CHF instead of pneumonia, and vice versa).

Also of note: they looked at platelet count, WBC count, and band percentage as predictor of mortality — as well as age! Platelets < 150k was a significant independent predictor of mortality, as was bands > 5% and of course, lactate greater than 4.0 mmol / L. WBCs didn’t matter. The odds ratio of death in a lactate > 4 was 4.9 (and 9.5 within 3 days), compared to an OR of 1.9 with bands > 5%. Some comments clarified the findings.

The biggest limitation is lack of comorbidities: maybe lactate > 4.0 is a good predictor of death, but hey, maybe everyone who died was hypotensive and had multi-organ failure, too, which kind of makes lactate’s clinical utility questionable. Outpatients weren’t included (maybe a lot of people are discharged from the ED with a lactate of 2.6, what happens to them? should we call them back, like we do for positive blood cultures?)

Other groups have shown that lactate in the ICU is a better predictor of mortality than blood pressure or other measures, with nearly 90% dying when lactate > 4, in any ICU patient, regardless of cause.

Overall, though, I think this paper makes a good case for obtaining lactate levels — though they don’t say it, I think the most benefit will come in younger patients with greater physiological reserve, who don’t look too sick. Just order lactate when you order blood cultures. (and remember — anion gap is not sufficient to diagnosis lactic acidosis) A low lactate won’t change your (probably conservative) treatment, but a high value should prompt more aggressive intervention. Like what? Well, remember the criteria for sepsis and septic shock. EGDT is only proven to benefit patients with severe sepsis or shock, but some of the guidelines and monitoring tools from EGDT can be applied to otherwise stable patients with high lactate levels.

Other related information (because sometimes, clicking on emcrit for this is just too hard) :

According to the ACP and SCC, in Chest 101:1644, 1992, sepsis/SIRS was defined as having two of the following:

1. Temperature > 38, or < 36
2. heart rate over 90
3. respiratory rate over 20, or a PaCO2 of less than 32 mmHg
4. WBC over 12, or under 4, or more than 10% bands

With sepsis, it’s two of the above, plus infection. “Severe” sepsis means sepsis, plus organ dysfunction. Septic shock is a systolic BP less than 90, or a drop of 40 mmHg from baseline.

The EGDT / “surviving sepsis campaign” measures are available online, also in convenient “bundle” form. They recommend initiating EGDT on any patient with a lactate over 4, even if they’re normotensive and have no organ dysfunction. Between 2.5 and 4, what to do? It’s a gray area: maybe start a foley, oxygen, cardiac monitoring. Heck, maybe even CVP if they look suspicious.

But all the intensive monitoring and care can’t begin, without a lactate. The Surviving Sepsis people are also responsible for this quote:

“Obtaining serum lactate is essential to identifying tissue hypoperfusion in patients who are not yet hypotensive but who are at risk for septic shock.”

I think Shapiro’s data supports that.  

Posted on Sunday, July 9th, 2006 at 2:52 am by Nick. Filed under Risk Stratification, Sepsis, Journal Club.
You may post a comment.