Serum Lactic Acid (LA) as a Predictor of Septic Shock in Patients with Hematologic Malignancies (HM) Who Develop Febrile Neutropenia

Introduction: Febrile neutropenia, a commonly occurring entity in patients (pts) with HM, is treated empirically with antibiotics due to a high risk of developing septic shock. Currently, there is no way to predict which hospitalized pts with febrile neutropenia are at greatest risk of developing septic shock. Clinicians have traditionally used systemic inflammatory response syndrome (SIRS) criteria (≥ 2 of the following: RR > 20, HR >90, Temp > 38C or <36C, and WBC >12k or <4k) to facilitate the diagnosis of sepsis and to risk-stratify septic pts. However, it is unknown if SIRS criteria have any discriminatory value in pts with febrile neutropenia who, by definition, meet 2 of 4 of these criteria. Critical care guidelines recommend measuring serum lactate to assess sepsis severity. However, since malignancy itself may produce lactate, the utility of its measurement in this population is not clear. The purpose of this study was to determine the incremental predictive value of serum lactate in addition to the SIRS criteria to predict septic shock in HM pts with febrile neutropenia.

Methods: We tested the association of lactate, tachypnea, and tachycardia with the development of septic shock in hospitalized adult HM pts (age 18–83) in a nested, case-control design within a prospective, single-center, cohort study. Vital signs and lactate were measured during episodes of febrile neutropenia (Temp > 100.4 & ANC < 1000) during hospitalization. We defined an elevated lactate as a level ≥ 2.0 mmol/L. The study endpoint was the development of septic shock, defined as a document or suspected infection resulting in hypotension unresponsive to 1.5 L of IV fluids. The case-control design with incidence-density sampling allowed for maximal efficiency and minimal bias. Controls were matched on length of stay at the time of septic shock in a 4:1 ratio to achieve 80% power to detect an OR (odds ratio) of 2.5. Using multivariable logistic regression and receiver operating curve (ROC) analysis, we evaluated the association of SIRS and lactate with the development of septic shock within 48 hrs following their measurement.

Results: Of the 547 pts enrolled, 46 (8.4%; 95% confidence interval [CI]: 6.2–10.9) developed septic shock. Baseline characteristics including age, vital signs, ANC, medications, and lactate were similar between the groups. The mean baseline lactate (1.0 mmol/L; 95% CI: 0.5–2.0) for the cohort was similar to our hospital’s normal range (.7–2.1mmol/L). In univariate analysis of HR, RR and lactate, only tachypnea (RR>20) (OR 5.9; 95% CI: 2.0–16.9, p =.001) and elevated lactate level (OR 18.4; 95% CI: 4.1–81.6, p <.001) were significantly associated with the subsequent development of septic shock. In multivariate analysis, lactate and RR remained independent predictors of septic shock (lactate, OR=12.1; 95% CI: 2.1–70.1, p=.005; tachypnea, OR=7.6; 95%CI: 1.6–35.5, p=.01). The area under the ROC curve for the SIRS+LA model was significantly higher than that of the SIRS model alone (0.75 and .69 respectively, p=.02).

Conclusions: In febrile neutropenic pts, measurement of tachypnea is the only SIRS criteria with independent prognostic importance for the development of septic shock. In addition to RR, measurement of serum lactate (a valid, inexpensive, readily available laboratory test) at the time of febrile neutropenia, adds significant value in the prediction of which HM pts are at increased risk of developing septic shock. Although further study is required to validate these results, routine measurement of lactate may help identify pts who may benefit from increased monitoring and early intervention strategies.