The Journal of Thoracic Diseases is about to release a special edition on sepsis, edited by Paul Marik (open access). It’s deliciously rebellious – just look at the table of contents 👇
My article on using the complete blood count to diagnose sepsis is already available, so I’ll summarize it here (yes, I know, you don’t have time to read the whole article right now). This article was surprisingly difficult to write, for a few reasons. First, there’s relatively little high-quality evidence on this topic. Second, there are several definitions of sepsis and septic shock floating around, rendering these terms almost meaningless. But I tried hard, so please bear with me…
the white blood cell count (WBC)
- Sepsis can cause the WBC to increase or decrease!
- WBC is fundamentally a garbage parameter, used predominantly due to convenience and tradition.
- Very rarely, the WBC is actually helpful:
- If there is true neutropenia, that’s a big deal (e.g. patients may require empiric broad-spectrum antibiotic coverage for neutropenic fever).
- If the WBC is crazy high, something bad is probably going on. However, in this situation, the neutrophil/lymphocyte ratio will also be haywire, so the WBC count probably doesn’t add much additional information beyond the neutrophil/lymphocyte ratio (more on this below).
- Again, for the folks in the back: A normal WBC says nothing about septic shock. Lots of patients with septic shock will have a stone-cold normal WBC.
Sepsis triggers cytokine release, which causes immature leukocytes to be released from the bone marrow (e.g. bands, promyelocytes, and metamyelocytes). There are three major drawbacks to relying on the left shift:
- Left shift is often delayed following infection (e.g. a day or more).
- Left shift isn’t specific to sepsis (other causes may include tissue necrosis, surgery, myeloproliferative disorders, or exogenous granulocyte colony stimulating factor).
- There’s no terrific way to measure left shift… specifically…
…bandemia to measure left shift?
- Traditionally, left shift was measured based mostly on bandemia.
- This requires manual slide review, which delays results.
- There is no universal agreement about exactly what a band is! Different pathologists and different hospitals will yield variable results.
…immature granulocytes to measure left shift?
- Immature granulocytes are defined as promyelocytes, myelocytes, and metamyelocytes. Modern hematology analyzers are capable of precisely counting the number of immature granulocytes among thousands of leukocytes. This ought to eliminate the logistic and reproducibility issues of the band count.
- Immature granulocyte fraction is likely to replace bandemia over time. Currently it’s not reported out on all hemocytometers.
- Unfortunately, studies disagree on the appropriate cutoff and sensitivity/specificity values. Some studies found immature granulocyte fraction to be a valuable parameter, while others found it to be worthless.
- More work needs to be done on this. For now, if you notice an unequivocally elevated immature granulocyte fraction (perhaps >2-3%), this indicates a left shift.
Bottom line on left shift?
- The presence of a left shift suggests severe underlying illness, potentially sepsis. This should serve as a red flag to identify patients warranting further investigation.
- Left shift can be absent in septic shock – particularly during the early phases.
neutrophil/lymphocyte ratio (NLR)
Basic principles of the NLR:
- Physiologic stress (e.g. cortisol release & catecholamines) tend to increase the neutrophil count and decrease the lymphocyte count. This increases the NLR. Any type of physiologic stress can do this, but sepsis tends to cause the most dramatic elevations of NLR.
- NLR elevates rapidly in response to stress (within ~6 hours). This will often be the first indicator of severe illness on the blood count.
- Interpretation of NLR may be unreliable in the presence of the following confounders: exogenous steroid, HIV, active hematologic disorders, or chemotherapy.
NLR performance in sepsis:
- Normal NLR values are 1-3. A normal NLR (<3) argues against sepsis (but doesn’t exclude it entirely).
- NLR values in the range of 3-10 fall within a grey zone.
values >10 become increasingly suggestive of sepsis. The key here is the clinical context. For
- In a patient with obvious sources of physiologic stress (e.g. DKA with hypovolemic shock), an NLR of 14 doesn’t tell you anything you don’t already know.
- In a patient with infection and no obvious physiologic stressor (e.g. a patient who appears to have cellulitis or urinary tract infection), an NLR of 14 may reveal systemic inflammation – pointing to a possible diagnosis of sepsis.
- No component of the complete blood cell count is a perfect indicator of septic shock. However, since this data is available to us, we should use it to maximal advantage.
- White blood cell count is the least useful parameter. Persistent focus on the WBC isn’t evidence-based and should be curtailed.
- Left-shift is often a delayed finding in septic shock. Measurements of left-shift (bandemia and immature granulocyte count) have substantial drawbacks. However, if a left-shift is found, this is a red flag which warrants further attention.
- Neutrophil to lymphocyte ratio (NLR) might be the single most useful parameter. NLR responds rapidly to infection and is uniformly available across all laboratories. However, NLR is fundamentally an index of physiologic stress (not septic shock) – so clinical context and judgement are required to interpret this properly.
- JTD issue on septic shock (open access) – should be released next week.
- Rory et al's epic article on the Lacto-Bolo Reflex.
- Schinkel M et al: Timeliness of antibiotics for patients with sepsis and septic shock.
- The neutrophil/lymphocyte ratio (PulmCrit)
Image credit: Top image from Mare TA et al. Critical Care 2015