Hausfater et al. Crit Care (2021) 25:227 https://doi.org/10.1186/s13054-021-03622-5 RESEARCH Open Access Monocyte distribution width (MDW) performance as an early sepsis indicator in the emergency department: comparison with CRP and procalcitonin in a multicenter international European prospective study Pierre Hausfater1,2,3* , Neus Robert Boter4,5, Cristian Morales Indiano5,7, Marta Cancella de Abreu1,2, Adria Mendoza Marin4,5, Julie Pernet1, Dolores Quesada5,8, Iris Castro9, Diana Careaga9, Michel Arock6, Liliana Tejidor9 and Laetitia Velly1,2 Abstract Background: Early sepsis diagnosis has emerged as one of the main challenges in the emergency room. Measure- ment of sepsis biomarkers is largely used in current practice to improve the diagnosis accuracy. Monocyte distribution width (MDW) is a recent new sepsis biomarker, available as part of the complete blood count with differential. The objective was to evaluate the performance of MDW for the detection of sepsis in the emergency department (ED) and to compare to procalcitonin (PCT) and C-reactive protein (CRP). Methods: Subjects whose initial evaluation included a complete blood count were enrolled consecutively in 2 EDs in France and Spain and categorized per Sepsis-2 and Sepsis-3 criteria. The performance of MDW for sepsis detection was compared to that of procalcitonin (PCT) and C-reactive protein (CRP). Results: A total of 1,517 patients were analyzed: 837 men and 680 women, mean age 61 ± 19 years, 260 (17.1%) categorized as Sepsis-2 and 144 patients (9.5%) as Sepsis-3. The AUCs [95% confidence interval] for the diagnosis of Sepsis-2 were 0.81 [0.78–0.84] and 0.86 [0.84–0.88] for MDW and MDW combined with WBC, respectively. For Sepsis-3, MDW performance was 0.82 [0.79–0.85]. The performance of MDW combined with WBC for Sepsis-2 in a subgroup of patients with low sepsis pretest probability was 0.90 [0.84–0.95]. The AUC for sepsis detection using MDW combined with WBC was similar to CRP alone (0.85 [0.83–0.87]) and exceeded that of PCT. Combining the biomarkers did not improve the AUC. Compared to normal MDW, abnormal MDW increased the odds of Sepsis-2 by factor of 5.5 [4.2–7.1, 95% CI] and Sepsis-3 by 7.6 [5.1–11.3, 95% CI]. Conclusions: MDW in combination with WBC has the diagnostic accuracy to detect sepsis, particularly when assessed in patients with lower pretest sepsis probability. We suggest the use of MDW as a systematic screening test, used together with qSOFA score to improve the accuracy of sepsis diagnosis in the emergency department. *Correspondence: pierre.hausfater@aphp.fr 1 Emergency Department, Hôpital Pitié‑Salpêtrière, APHP-Sorbonne Université, 83 Boulevard de l’hôpital, 75651 Paris Cedex 13, France Full list of author information is available at the end of the article © The Author(s) 2021. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/. The Creative Commons Public Domain Dedication waiver (http://​creat​iveco​ mmons.​org/​publi​cdoma​in/​zero/1.​0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Hausfater et al. Crit Care (2021) 25:227 Page 2 of 12 Trial Registration ClinicalTrials.gov (NCT03588325). Keywords: Monocyte volume distribution width, MDW, Procalcitonin, C-reactive protein, Emergency department, Sepsis Introduction Patients A significant proportion of patients developing sep- The study prospectively enrolled consenting adults (over sis enter the health system through an emergency 18-year-old) presenting to the ED, whose initial evalu- department (ED) [1–4]. Because the surviving sep- ation included a CBC-DIFF. Exclusion criterion is as sis campaign promotes targeting short-term specific, shown in Fig. 1. Emergency physicians were free to order goal-directed therapy bundles, early sepsis diagnosis necessary tests according to the standard of care. Treat- has therefore emerged as one of the main challenges ing physicians were blinded to MDW, PCT and CRP val- for emergency physicians [5]. Although the systemic ues obtained by protocol, but received PCT and/or CRP inflammatory response syndrome (SIRS) or quick results they ordered as current practice. Sequential Organ Failure Assessment (qSOFA) score Both PCT and CRP are used routinely at LPS, and CRP are helpful clinical tools for sepsis suspicion during is used routinely at UHG. early triage, sepsis diagnosis may still be delayed or misdiagnosed [4, 6]. So, measurement of sepsis bio- markers is largely used in current practice to improve Blood sampling the diagnosis accuracy [7]. Briefly, patients who were identified to have a CBC-DIFF MDW, a hematologic parameter measured as part of ordered by the physician were asked to participate. After the complete blood count with differential (CBC-DIFF) informed consent, baseline blood draws were performed. and describing the size distribution of circulating mono- An additional ­K3EDTA tube was drawn together with a cytes, was shown in two recent North American stud- sample for PCT and CRP measurement as well as rou- ies [8–10] to be a valuable new sepsis biomarker for the tine blood tests at the discretion of the treating physi- early detection of patients in the emergency department. cian. Samples for MDW measurement were analyzed on However, little is known about the performance of MDW a UniCel DxH 900 analyzer (Beckman Coulter, Inc., Brea, and how it compares to those of the most frequently used CA) with version 1.0.0.329 software within two hours of sepsis biomarkers: C-reactive protein (CRP) and procal- collection. This instrument measures specific cell volume citonin (PCT) [11]. variables and the distribution of cell volumes within a The primary objective of this study was to confirm group of white blood cells (WBC). Quality control was the clinical validity and performance of MDW to iden- performed daily with COULTER 6C Plus Cell Control to tify patients who developed sepsis in a European ED monitor the DxH 900 system performance. COULTER population. Secondary objectives were to compare the LATRON CP-X Control was used as part of the daily performance of MDW with CRP and PCT and to ver- quality control procedure to monitor volume, conductiv- ify MDW cutoff for tripotassium ethylenediaminetet- ity and light scatter measurements. PCT and CRP con- raacetic acid ­(K3EDTA) tubes. centrations were measured on Cobas analyzer (Roche Diagnostics, Meylan, France) at LPS, while UHG utilized Liaison XL (DiaSorin, Saluggia, Italy) and AU5800 (Beck- Methods man Coulter, Inc, Brea, CA, USA) analyzers for PCT and This was an international, blinded, prospective cohort CRP measurements, respectively. study enrolling patients presenting to two adult large EDs at the Pitié-Salpêtrière APHP-Sorbonne Uni- versité hospital in Paris (LPS), France, and the Hos- Clinical data, follow‑up, blinding and adjudication pital Universitari Germans Trias i Pujol in Badalona, Clinical data at presentation, including past medical his- Spain (UHG). The study assessed MDW’s ability to tory, assessment of vital signs, symptoms, SIRS criteria detect the development of sepsis in consecutive adult [12], qSOFA [13] and SOFA scores [14], microbiological patients (18–89 years) presenting to the ED and who testing and treatments were recorded on an electronic had a CBC-DIFF ordered. The study was registered in case report form and patients followed up for at least ClinicalTrials.gov (NCT03588325) and approved by 12 h. The clinical research team and physician adjudica- the ethical committees of the respective participating tors were blinded to MDW results at the time of clini- countries. cal data entry and during assignment of the patients to Hausfater et al. Crit Care (2021) 25:227 Page 3 of 12 Fig. 1 Flow diagram describing patient screening and enrollment. MDW, monocyte distribution width; SIRS, systemic inflammatory response syndrome; CBC-DIFF, complete blood count with differential; ED, Emergency Department; PCT, procalcitonin; CRP, C-reactive protein; TAT, Turnaround time a clinical category. Adjudicators were also blinded to the of infection, the patient was categorized as “not infected” results of PCT and CRP if not ordered by the treating ED or SIRS by the adjudicator. The test results were extracted physician. from the medical records 7–10 days later, including cul- Study subjects were classified by at least two independ- tures, molecular tests (e.g., polymerase chain reaction), ent physician adjudicators at each site. Discordances were antigens tests (immunoassay) and relevant imaging. A arbitrated by a third independent physician. Adjudicators subgroup, consisting of subjects for whom no CRP or categorized subjects based on the “Sepsis-2” consen- PCT was ordered by the emergency physician, was iden- sus criteria [15], such as non-SIRS or case controls (zero tified (post study) and defined as having low sepsis pre- or one SIRS criterion and no infection), SIRS (≥ 2 SIRS test probability. criteria and no infection), infection (suspected or con- firmed infection with 0—1 SIRS criteria), sepsis (infec- Statistical analysis tion plus ≥ 2 SIRS) (including sepsis [no organ failures], General descriptive statistics and box plots were cal- severe sepsis [sepsis with one or more organ failures] and culated for cell population variables. The sample size septic shock [sepsis with refractory hypotension]). Adju- calculations were based on 95% two-sided confidence dicated categories per Sepsis-3 criteria [16] included con- interval and 80% power. Proc power of SAS 9.3 was used trols, infection and sepsis (based on SOFA score criteria). for calculating sample size based on the approach pre- In order to characterize sepsis as being present upon ED sented in Johnson et al. [17] A target sensitivity of 75% admission, sepsis criteria had to be fulfilled within 12 h with the lower limit of the 95% two-sided confidence of the initial CBC-DIFF in patients with suspected infec- interval of 65% and a target specificity of 70% with the tion (as reflected by initiation of diagnostic infection lower limit of the 95% two-sided confidence interval of workup) and adjudication was based on the retrospective 65% were assumed. A minimum of 189 septic subjects chart review of tests ordered and clinical data available and a minimum of 817 non-septic subjects were needed within the first 12 h of ED presentation. If no infection from both sites combined. work-up was performed within 12 h, or if the adjudicator Diagnostic ability was evaluated in terms of the area believed that the infection work-up showed no evidence under the ROC curve (AUC), sensitivity, specificity, Hausfater et al. Crit Care (2021) 25:227 Page 4 of 12 positive predictive value (PPV), negative predictive value Probabilities and odds ratios (NPV), positive and negative likelihood ratios, along with Identification of sepsis in the ED relies on the physician’s their 95% confidence intervals (CIs). The score approach ability to assess probability of disease based on present- was used to calculate CI for sensitivity, specificity, PPV ing symptoms. To reflect this approach, we analyzed the and NPV. Differences in AUC were used to demonstrate probability and likelihood ratios for sepsis based upon the added value of MDW in comparison with WBC the values of MDW in combination with SIRS or qSOFA alone, calculated using a one-predictor variable logistic parameters determined during the initial patient encoun- model with WBC and a two-predictor variables logistic ter in the ED (typically within the first 2 h of ED admis- model with both WBC and MDW, as the predictor, and sion). Predicted probability of a positive sepsis diagnosis using sepsis status as the response. A similar approach was calculated from the positive likelihood ratios (LRs +) was applied to analyze the diagnostic ability of PCT and as previously described [19, 20]. In this approach, pre- CRP, and the combination with WBC and MDW for sep- dicted sepsis probability after receiving test results, or sis detection. AUC comparisons along with their CI were posttest probability P1, is based on an estimated pretest calculated as described by DeLong et al. [18] SAS 9.4 probability P0 and LR + and is calculated as: (SAS Institute, Cary, NC) statistical program was used P1 × LR+ for data analyses. P1 = (1 − P0 + P0 × LR +) Cutoff determination where P0 is the sepsis prevalence of the study cohort. PCT cutoff of > 0.25 µg/mL was based on the literature. The odds ratios (ratios of posttest probabilities) for sep- The determination of the cutoff for CRP (> 22 mg/L) sis diagnosis between parameter combinations with was based on the Youden index using the data from this abnormal and normal MDW values were calculated clinical study. The proposed MDW cutoff of 21.5 units based on prevalence of sepsis-2 or sepsis-3 for the study for ­K3EDTA was based on previously established MDW population. cutoff (in ­K2EDTA) for sepsis identification in the ED and internal testing demonstrating a shift of 1.5 units for Results blood specimens collected in ­K3EDTA versus ­K2EDTA As shown in the flow diagram of the study (Fig. 1), from (data not shown). This study validates the new MDW August 2, 2018, through June 27, 2019, 1,689 patients cutoff of 21.5 for ­K3EDTA. were screened and enrolled, and 172 were excluded due to patient not meeting inclusion criteria, inadequate Table 1 Demographic data. Data were available for all the patients, unless indicated by N. Quantitative data are expressed as Medians (interquartile range) LPS UHG Sites combined Sepsis-2 Sepsis-3 N = 260 N = 144 Clinical characteristics Female (No. %) 404 (48.3%) 276 (40.6%) 680 (44.8%) 101 (38.9%) 46 (31.9%) Male (No. %) 433 (51.7%) 404 (59.4%) 837 (55.2%) 159 (61.2%) 98 (68.1%) Age, years 57 (39–71) 72 (60–80) 64 (47–76) 66 (51–77) 72 (61–80) Temperature (°C) N = 1506 36.6 (36.2–37.0) 36.2 (36.0–36.7) 36.5 (36.0–36.9) 37.1 (36.5–38.2) 36.8 (36.2–38.0) Heart rate (beats/min) 85 (73–100) 82 (70–98) 84 (71–99) 100 (92–110) 93 (75–106) Respiratory (breaths/min) N = 865 20 (16–22) 18 (16–24) 19.0 (16–24) 22 (18–26) 23 (18–26) Immunosuppression 38 (4.5%) 101 (14.9%) 139 (9.2%) 38 (14.6%) 22 (15.3%) Laboratory parameters WBC (1 × ­103/μL) 8.33 (6.24–10.94) 9.17 (6.86–11.89) 8.71 (6.50–11.34) 13.02 (9.02–16.28) 10.82 (8.27–15.03) PMN (1 × ­103/μL) 5.80 (3.88–8.44) 6.74 (4.73–9.42) 6.27 (4.23–8.90) 10.65 (7.09–14.04) 8.99 (6.15–13.18) EO (1 × ­103/μL) 0.07 (0.02–0.15) 0.07(0.02–0.14) 0.07 (0.02–0.14) 0.02 (0.01–0.06) 0.02 (0.0–0.06) MDW (U) 19.28 (17.85–21.68) 21.09 (19.19–24.03) 20.15 (18.22–22.79) 23.92 (21.54–26.94) 24.61 (22.33–28.47) PCT (ng/mL) 0.04 (0.02–0.10) 0.06 (0.02–0.24) 0.05 (0.02–0.15) 0.21 (0.07–0.79) 0.34 (0.13–1.54) CRP (mg/L) 5.22 (0.85–31.76) 18.30 (4.75–82.50) 10.10 (1.91–52.89) 82.50 (36.90–185.50) 95.35 (44.13–197.98) CREAT (μmol/L) N = 1489 76.00 (64.00–95.00) 80.40 (62.80–113.20) 78.00 (63.60–102.00) 84.00 (65.40–123.00) 115.35 (74.30–186.50) LPS, Pitié-Salpêtrière APHP-Sorbonne Université Hospital; UHG, Hospital Universitari Germans Trias i Pujol; WBC, white blood count; PMN, polymorphonuclear; EO, eosinophils; MDW, monocyte distribution width; PCT, procalcitonin; CRP, C-reactive protein; CREAT, creatinine Hausfater et al. Crit Care (2021) 25:227 Page 5 of 12 sample collection or screening errors. The final analy- sis included 1,517 patients (LPS 837, UHG 680). Demo- graphic features are shown in Table 1. There were 837 men and 680 women, median age 64 (IQ 47–76 years). Using Sepsis-2 criteria, 819 patients (54%) were cat- egorized as non-SIRS/non-sepsis (case controls) while 260 (17.1%) had sepsis, including 38 with severe sepsis. According to Sepsis-3 criteria, 1,016 patients (67%) were adjudicated as case controls and 144 patients (9.5%) had sepsis. Eighty-five percent of Sepsis-2 patients had at least one microbiological test performed, of which 58% were positive. Bacterial cultures grew with Gram-posi- tive species in 16 patients, Gram-negative in 62 patients and mixed gram stain in 10. These cultures were 41 from urine, 24 from blood, 14 from sputum and 9 from other sources (Additional file 2: Sup. Table 5). Five hundred and fifty-three patients (36.5%) fell in the subgroup of the low pretest probability of sepsis (no CRP or PCT ordered by the emergency physicians). MDW performance for sepsis detection Figure 2A, B illustrates box whisker plots for a single baseline MDW measurement according to Sepsis-2 (A) or Sepsis-3 (B) criteria showing increases in MDW val- ues with infection and sepsis severity, regardless of sep- sis criteria used. In ­K3EDTA tubes, the MDW cutoff of 21.5 provides an optimum diagnostic power by balanc- ing the ability to detect positive sepsis patients (sensitiv- ity) and case controls (specificity). The area under the Fig. 2 Box plots of MDW baseline values conforming to sepsis classification by Sepsis-2 criteria (A) and Sepsis-3 criteria (B). MDW, receiver operating characteristic (AUROC) curves for monocyte distribution width; SIRS, systemic inflammatory response Sepsis-2 was 0.81 [95% confidence interval 0.78–0.84] syndrome for MDW vs 0.76 [0.72–0.79] for WBC and 0.86 [0.84– 0.88] when combined with WBC (Fig. 3a). For Sepsis-3, MDW AUROC was 0.82 [0.79–0.85] compared to 0.65 MDW increased the probability of Sepsis-2 for all ranges [0.60–0.70] for WBC (Fig. 3b). Adding WBC to MDW of WBC including in patients with low pretest probability assessment added no value for Sepsis-3 diagnosis. For with similar effect observed using Sepsis-3 adjudication the subgroup of patients with low sepsis pretest probabil- (Fig. 4a–d and Additional file 4: Sup. Figure 7). ity, AUROC for Sepsis-2 was 0.83 [0.75–0.91] for MDW alone and 0.90 [0.84–0.95] when combined with WBC Comparison of MDW with PCT and CRP performance (Fig. 3c). for sepsis detection The respective performance of PCT and CRP and com- MDW added value for sepsis detection bined with WBC and MDW is represented in Table 3. The value of MDW significantly modified the observed Overall, the diagnostic ability of MDW combined with posttest probability of sepsis vs. the baseline probabil- WBC was similar to CRP alone (AUROC: 0.86 [0.84– ity regardless of the sepsis definitions, specific clinical 0.88] vs. 0.85 [0.83–0.87], respectively) and performed risk scores (SIRS and qSOFA) and values of WBC, with better than PCT alone (AUROC: 0.78 [0.75–0.81]). Add- abnormal (> 21.5) MDW increasing and normal MDW ing either CRP or PCT to MDW/WBC analysis did not decreasing the posttest probability. Compared to normal improve the AUC. The sensitivity and specificity of using MDW, abnormal MDW increased the odds of Sepsis-2 by MDW and/or WBC in combination, based on their indi- a factor of 5.5 [4.2–7.1, 95% CI] and the odds of Sepsis-3 vidual cutoffs, is shown in Additional file 1: Sup. Table 4. by 7.6 [5.1–11.3, 95% CI] (Table 2), with highest increases If either parameter was abnormal (“or”) the sensitiv- occurring in patients with low SIRS and qSOFA scores ity was > 90%, if both parameters were abnormal (“and”) (Table 2 and Additional file 3: Sup. Figure 6). Abnormal Hausfater et al. Crit Care (2021) 25:227 Page 6 of 12 Fig. 3 MDW, WBC and MDW + WBC performance for sepsis diagnosis. A Sepsis-2. B Sepsis-3. C Sepsis-2, low pretest probability population (CRP or PCT not ordered by emergency physician). MDW, monocyte distribution width; WBC, white blood count then as expected the specificity of the combination of eightfold increase of Sepsis-3. Similarly, when WBC is MDW + WBC was high (> 90%). abnormal (< 4,000 or > 12,000/mm3) and MDW > 21.5, Figure 5a–d represents the posttest sepsis probabili- the probability of Sepsis-2 is twice the one when MDW is ties of the sequential results of baseline WBC, MDW and normal, and sevenfold for Sepsis-3. PCT or CRP tests at their specified cutoffs. When WBC When WBC is normal but MDW is abnormal is normal (in the 4,000–12,000/mm3 range), an MDW (> 21.5), addition of PCT increases probability of sep- value > 21.5 is associated with a tenfold increase of Sep- sis to the same level as when both (WBC, MDW) sis-2 probability (compared to MDW normal), and an parameters were abnormal, threefold above the pretest probability (see Fig. 5b, PCT > 0.25 and MDW > 21.5 Hausfater et al. Crit Care (2021) 25:227 Page 7 of 12 Table 2 Added value of MDW to SIRS criteria for Sepsis-2 diagnosis (pretest = 0.17), and to qSOFA for Sepsis-3 diagnosis (pretest = 0.9) Sepsis 2 & SIRS No SIRS 1 SIRS 2 SIRS 3 SIRS 4 SIRS Total Total patients 515 594 271 117 20 1517 Sepsis patients (#) 4 21 141 81 13 260 Sepsis probability, MDW unknown 0.8% 3.5% 51.8% 69.0% 64.8% 17.0% Sepsis probability, MDW normal 0.3% 1.7% 31.1% 37.6% 28.4% 6.6% Sepsis probability, MDW abnormal 2.4% 7.8% 71.2% 83.6% 84.5% 36.1% Sepsis odds (MDW abnormal/normal) [95% CI] 9.3 4.7 2.3 2.2 3.0 5.5[4.2–7.1] Sepsis 3 & qSOFA No qSOFA 1 qSOFA 2 qSOFA 3 qSOFA Total Total patients 1136 347 33 1 1517 Sepsis patients (#) 53 72 18 1 144 Sepsis probability, MDW unknown 4.4% 19.8% 53.1% 9.0% Sepsis probability, MDW normal 1.2% 8.1% 28.8% 2.7% Sepsis probability, MDW abnormal 12.0% 31.9% 63.9% 20.6% Sepsis odds (MDW abnormal/normal) [95% CI] 10.1 3.9 2.2 7.6 [5.1–11.3] MDW, monocyte distribution width; SIRS, systemic inflammatory response syndrome; qSOFA, quick Sequential Organ Failure Assessment; CI, Confidence Interval A B 80% 2.2x Sepsis-2 Probability 60% 60% 45% 40% 10.0x 27% 20% 20% 7% 2% 0% Normal WBC Abnormal WBC C 120004000 12000>WBC<4000 D Fig. 4 Added value of MDW on sepsis posttest probabilities according to WBC range at presentation. A Sepsis-2 (sepsis pretest probability = 0.17). B Sepsis-3 (pretest 0.09). C Sepsis-2, low pretest probability population (CRP or PCT not ordered by emergency physician. Sepsis pretest probability = 0.065). Cutoff: WBC < 4,000/mm3 or WBC > 12,000, MDW > 21.5, and PCT > 0.25 µg/L. D low pretest probability population per Sepsis-3 (pretest = 0.043). MDW, monocyte distribution width; WBC, white blood count; PCT, procalcitonin; CRP, C-reactive protein and WBC normal sepsis probability is 30%). A simi- PCT > 0.25 µg/L increases the Sepsis-2 probability lar pattern was observed with CRP. Similarly, when approximately threefold comparable to when both WBC is abnormal but MDW is normal, addition of WBC and MDW are both abnormal. Hausfater et al. Crit Care (2021) 25:227 Page 8 of 12 Table 3 Comparison of MDW with WBC, PCT and CRP performance for sepsis detection Cutoff: WBC < 4,000/mm3 or WBC > 12,000/ mm3, MDW > 21.5, PCT > 0.25 µg/L, CRP > 22 mg/L) Sepsis-2 Parameter Sensitivity Specificity PPV [95%CI] NPV [95%CI] LR + [95%CI] LR − [95%CI] AUC [95%CI] [95%CI] [95%CI] MDW 0.75 [0.69–0.80] 0.73 [0.70–0.75] 0.36 [0.32–0.40] 0.93 [0.92–0.95] 2.76 [2.46–3.09] 0.34 [0.28–0.43] 0.81 [0.78–0.84] WBC 0.69 [0.63–0.74] 0.83 [0.81–0.85] 0.45 [0.40–0.50] 0.93 [0.91–0.94] 4.00 [3.50–4.60] 0.40 [0.30–0.50] 0.76 [0.72–0.79] PCT 0.45 [0.39–0.51] 0.88 [0.86–0.90] 0.44 [0.38–0.50] 0.89 [0.87–0.90] 3.80 [3.10–4.65-] 0.62 [0.56–0.70] 0.78 [0.75–0.81] CRP 0.85 [0.80–0.88] 0.72 [0.70–0.75] 0.39 [0.35–0.43] 0.96 [0.94–0.97] 3.06 [2.76–3.39] 0.21 [0.16–0.28] 0.85 [0.83–0.87] MDW + WBC 0.86 [0.84–0.88] MDW + PCT 0.81 [0.78–0.84] MDW + CRP 0.85 [0.82–0.87] MDW + WBC + PCT 0.86 [0.84–0.89] MDW + WBC + CRP 0.87 [0.85–0.89] Sepsis-3 MDW 0.81 [0.73–0.86] 0.69 [0.67–0.72] 0.22 [0.18–0.25] 0.97 [0.96–0.98] 2.63 [2.35–2.94] 0.28 [0.20–0.39] 0.82 [0.79–0.85] WBC 0.49 [0.41–0.57] 0.77 [0.74–0.79] 0.18 [0.15–0.22] 0.94 [0.92–0.95] 2.10 [1.70–2.50] 0.70 [0.60–0.80] 0.65 [0.60–0.70] PCT 0.60 [0.52–0.68] 0.87 [0.85–0.89] 0.33 [0.27–0.39] 0.95 [0.94–0.96] 4.63 [3.83–5.60] 0.46 [0.37–0.56] 0.84 [0.80–0.87] CRP 0.89 [0.83–0.93] 0.68 [0.66–0.71] 0.23 [0.19–0.26] 0.98 [0.97–0.99] 2.79 [2.53–3.07] 0.16 [0.10–0.26] 0.85 [0.82–0.87] MDW + WBC 0.83 [0.79–0.86] MDW + PCT 0.82 [0.79–0.86] MDW + CRP 0.85 [0.82–0.88] MDW + WBC + PCT 0.83 [0.80–0.86] MDW + WBC + CRP 0.85 [0.82–0.87] MDW, monocyte distribution width; WBC, white blood count; PCT, procalcitonin; CRP, C-reactive protein; PPV, positive predictive value; NPV, negative predictive value; LR + : positive likelihood ratio. LR-, negative likelihood ratio; AUC: area under the ROC curve; CI confidence interval Discussion to the early clinical evaluation. Indeed, early sepsis detec- In this multicenter clinical trial conducted in two large tion is warranted to initiate specific goal-directed ther- European EDs, we confirmed the performance of MDW apy bundles [5]; hence, the systematic SIRS criteria and as an early sepsis biomarker with an AUC of 0.81 [0.78– qSOFA score calculation are highly recommended. How- 0.84] and 0.86 [0.84–0.88] when combined with WBC ever, both SIRS and qSOFA lack sensitivity and specificity for Sepsis-2 criteria. These results are in accordance and may falsely lead to over or under suspicion of sep- with those of two recent North American studies which sis [4, 6, 21, 22]. In the present study, we confirmed that reported very similar performance of MDW (AUC 0.79 MDW provided an added diagnostic power to SIRS and [0.73–0.84] and 0.79 [0.76–0.82], respectively) [8, 9] qSOFA variables, with an overall odd of 5.5 and 7.6 for which argues for the robustness and reproducibility of Sepsis-2 and Sepsis-3, respectively, when MDW > 21.5 this new biomarker in different populations. At an opti- (Table 2 and Fig. 5). As recently published by Crouser, mal cutoff of 21.5 in K ­ 3EDTA tubes, a sensitivity of 75% et al., because MDW value can be displayed with CBC [69–80, 95% CI] and specificity of 73% [70–75, 95% CI] results early in the ED’s course of the patient, it may be were observed which again agrees well with previous considered as a fifth SIRS or fourth qSOFA criteria [10]. studies in North America and demonstrates that the MDW combined with WBC had a similar AUC com- anticoagulant does not affect the diagnostic accuracy of pared to CRP (0.85 [0.82–0.87]) and performed better the biomarker. Moreover, we reported that MDW com- than PCT (0.78 [0.75–0.81]) (Table 3). To our knowl- bined with WBC was particularly useful to detect sepsis edge, this was the first report of the comparison between in a low pretest probability subgroup of patients (AUC: MDW, PCT and CRP in the emergency department, 0.90 [0.84–0.95]), that is patients for whom no sepsis bio- in contrast to the study by Polilli, et al., who reported marker was ordered by the emergency physicians upon nearly overlapping AUC for MDW and PCT (0.87 vs. presentation. MDW may be displayed routinely as part of 0.88, respectively) in a cohort of hospitalized patients for CBC with differential for most of the patients who have suspected infection or sepsis [11]. The performances of blood drawn in the ED and thus provides an added value PCT and CRP we observed were much higher than what Hausfater et al. Crit Care (2021) 25:227 Page 9 of 12 Fig. 5 Sequential assessment of sepsis probabilities according to WBC, MDW and PCT results. Cutoff: WBC < 4,000/mm3 or WBC > 12,000/mm3, MDW > 21.5, PCT > 0.25 µg/L, CRP > 22 mg/L). Pretest probabilities were 0.17 for sepsis-2 and 0.09 for sepsis-3. A PCT by Sepsis-2. B PCT by Sepsis-3. C CRP by Sepsis-2. D CRP by Sepsis-3. MDW, monocyte distribution width; WBC, white blood count; PCT, procalcitonin; CRP, C-reactive protein Hausfater et al. Crit Care (2021) 25:227 Page 10 of 12 had been reported to date in the literature, with AUCs Conclusion ranging from 0.64 to 0.79 and 0.57 to 0.79, respectively MDW in combination with WBC has the diagnostic [23–27]. The first explanation of the discrepancy is that accuracy to detect sepsis, particularly when assessed in our main inclusion criterion (ED patients for whom ini- patients with lower pretest sepsis probability. We sug- tial evaluation included a CBC-DIFF) resulted in a less gest the use of MDW as a systematic screening test, selected patient population than the criteria of previous used together with SIRS criteria and qSOFA score to studies (patients with a suspicion of infection or sepsis). improve the accuracy of sepsis diagnosis [29, 30] in The sepsis prevalence observed in current study (Table 1) the emergency department. The place of CRP or PCT was higher than the mean prevalence in ED popula- thereafter remains to be determined. tion [28], but much lower than what is reported in more selected populations [23, 24]. For example, Lungström, Supplementary Information et al., reported a Sepsis-2 prevalence of 42% in a cohort The online version contains supplementary material available at https://​doi.​ of 1,572 EDs patients with a suspicion of infection, along org/​10.​1186/​s13054-​021-​03622-5. with PCT and CRP AUCs of O.64 [0.61–0.67] and 0.57 [0.54–0.60], respectively, whereas Uusitalo-Seppälä, Additional file 1. Combining MDW and WBC to improve sepsis detection per Sepsis-2 and Sepsis-3 criteria. Abbreviations: MDW, monocyte distribu- et al., reported a Sepsis-2 prevalence of 57% and AUCs tion width; WBC, white blood count. of 0.77 [0.71–0.84] and 0.60 [0.51–0.69], respectively, for Additional file 2. Microbiological Tests per Sepsis-2 and Sepsis-3 Criteria. severe sepsis detection [23, 24]. The second hypothesis Additional file 3. Added value of MDW to SIRS criteria for Sepsis2 was that although the adjudicators were blinded to PCT diagnosis (pre-test = 0.17) and to qSOFA for Sepsis-3 diagnosis (pre-test and CRP results performed by protocol, they had access = 0.09). Abbreviations: MDW, monocyte distribution width; SIRS, systemic to PCT and CRP values obtained as standard of care, inflammatory response syndrome; qSOFA, quick Sequential Organ Failure Assessment. which may have contributed to an overestimation of the Additional file 4. Added value of MDW on sepsis post-test probabilities performance of these two sepsis biomarkers in our study. according to WBC range at presentation. A: Sepsis-2 (sepsis pre-test prob- On first analysis, our data did not support the addition ability = 0.17). B: Sepsis-3 (pre-test 0.09). C: Sepsis-2, low pre-test prob- of PCT and CRP testing to WBC plus MDW in the ED. ability population (pre-test = 0.065) D: low pre-test probability population per Sepsis-3 (pre-test = 0.043). However, as illustrated in Fig. 5, a sequential approach with WBC plus MDW as a systematic screening test, fol- lowed by PCT or CRP measurement in cases of discord- Acknowledgements The authors wish to acknowledge the support of the various clinical coordina- ant results (WBC abnormal–MDW normal, or WBC tors at UHG and LPS and the additional physician adjudicators at UHG that normal–MDW abnormal) may represent a pragmatic supported the study. LPS particularly thanks Ilaria Cherubini and Enfel Houas perspective. for excellent clinical research tasks, and Juliana Fernandes for dedicated MDW measurement. UHG also acknowledges Tania Bruno Garcia and Maite Calzado We acknowledge several limitations of the study. First, Flores for outstanding help with clinical data collection and monitoring. as there is no gold standard for the diagnosis of sepsis, The authors are also grateful to the following Beckman Coulter employees we cannot exclude the possibility of some misclassifica- Mohamad Hasan and Gang Xu, for statistical support, to Justin Rohrbach for creation of the eCRF & data management and to Ilya Grigorov for critical tion of patients that limits or biases the accuracy of the review and edits to this manuscript. biomarkers. However, we tried to limit this risk by adju- dicating all cases using two independent adjudicators and Authors’ contributions Study designwas decided by PH, LT, and LV. Patient’s recruitment and inclu- arbitration by a third physician in case of disagreement. sion were performed by NRB, AMM, MC, and JP. Laboratory testing including Second, as discussed above, the adjudicators were not MDW was organized and validated by MDQ, MA, and CM. Data management blinded to PCT and CRP values ordered as standard of was performed by IC and DC. Adjudication was performed by JP, MC, NRB, and AMM. Arbitration was performed by PH. Data analysis was completed care, which may have contributed to an overestimation by PH, LT, LV, MC, CM, and NRB. Manuscript writing was performed by PH of both the prevalence of sepsis and the performance of and LT. All authors participated in the critical review of the manuscript and these biomarkers. Similarly, sepsis biomarker ordering as approved the final version. a standard of care was different by sites (PCT and CRP Funding in LPS, CRP only in UHG). We think that this current This study was funded by Beckman Coulter. practice discrepancy rather enhances the validity of the Availability of data and materials combined sites results than biases it. Finally, the overall On request at Beckman Coulter, Inc. sepsis prevalence we reported was higher than the usual prevalence in EDs, arguing for a part of screening selec- tion bias toward sepsis patients at inclusion. Hausfater et al. Crit Care (2021) 25:227 Page 11 of 12 Declarations 8. Crouser ED, Parrillo JE, Seymour C, et al. Improved early detection of sep- sis in the ED with a novel monocyte distribution width biomarker. Chest. Ethics approval and consent to participate 2017;152(3):518–26. https://​doi.​org/​10.​1016/j.​chest.​2017.​05.​039. The study was registered in ClinicalTrials.gov (NCT03588325) and approved by 9. Crouser ED, Parrillo JE, Seymour CW, et al. Monocyte distribution width: the ethical committees of the respective participating countries. Patients were a novel indicator of sepsis-2 and sepsis-3 in high-risk emergency depart- included after informed consent. Study was approved by: Agence nationale ment patients. Crit Care Med. 2019;47(8):1018–25. https://​doi.​org/​10.​ de sécurité du médicament et des produits de santé (ANSM), July 02, 2018, 1097/​CCM.​00000​00000​003799. and by the Comité de Protection des Personnes (CPP) July 16, 2018. REF. 2018- 10. Crouser ED, Parrillo JE, Martin GS, et al. Monocyte distribution width A00892-53 Paris, France. Comités de Ética de la Investigación (CEI), Hospital enhances early sepsis detection in the emergency department Universitari Germans Trias i Pujol, June 22, 2018. REF.CEI PI-18-140 Badalona, beyond SIRS and qSOFA. J Intens Care. 2020. https://​doi.​org/​10.​1186/​ Spain. s40560-​020-​00446-3. 11. Polilli E, Sozio F, Frattari A, et al. Comparison of monocyte distribution Consent for publication width (MDW) and procalcitonin for early recognition of sepsis. Kou YR ed. NA. PLoS ONE. 2020;15(1):e0227300. https://​doi.​org/​10.​1371/​journ​al.​pone.​ 02273​00. Competing interests 12. Levy MM, Fink MP, Marshall JC, et al. 2001 SCCM/ESICM/ACCP/ PH receive consultant and advisory board fees from Beckman Coulter, ATS/SIS international sepsis definitions conference. Crit Care Med. educational support and congress communication fees from ThermoFisher 2003;31(4):1250–6. https://​doi.​org/​10.​1097/​01.​CCM.​00000​50454.​01978.​ Scientific, Radiometer and bioMérieux. IC, DC and LT are employed by Beck- 3B. man Coulter, Inc., the sponsor of the clinical study. The other authors have no 13. Seymour CW, Liu VX, Iwashyna TJ, et al. Assessment of clinical criteria for conflict of interest or competing interest to declare. sepsis: for the third international consensus definitions for sepsis and septic shock (sepsis-3). JAMA. 2016;315(8):762. https://​doi.​org/​10.​1001/​ Author details jama.​2016.​0288. 1 Emergency Department, Hôpital Pitié‑Salpêtrière, APHP-Sorbonne Université, 14. The SOFA (Sepsis-related Organ Failure Assessment) score to describe 83 Boulevard de l’hôpital, 75651 Paris Cedex 13, France. 2 Sorbonne Université, organ dysfunction/failure 4. GRC-14 BIOSFAST, Paris, France. 3 UMR INSERM 1166, IHU ICAN, Sorbonne 15. Levy MM, Fink MP, Marshall JC, et al. 2001 SCCM/ESICM/ACCP/ATS/ Université, Paris, France. 4 Emergency Department, Hospital Universitari SIS international sepsis definitions conference. Intens Care Med. Germans Trias I Pujol, Badalona, Spain. 5 Universitat Autònoma de Barcelona, 2003;29(4):530–8. https://​doi.​org/​10.​1007/​s00134-​003-​1662-x. Badalona, Spain. 6 Biochemisty and Emergency Biology Department, Hôpital 16. Shankar-Hari M, Phillips GS, Levy ML, et al. Developing a new definition Pitié‑Salpêtrière, APHP-Sorbonne Université, Paris, France. 7 Laboratory Medi- and assessing new clinical criteria for septic shock: for the third interna- cine Department, Laboratori Clinic Metropolitana Nord, Hospital Universitari tional consensus definitions for sepsis and septic shock (sepsis-3). JAMA. Germans Trias I Pujol, Badalona, Spain. 8 Microbiology Department, Labo- 2016;315(8):775. https://​doi.​org/​10.​1001/​jama.​2016.​0289. ratori Clinic Metropolitana Nord, Hospital Universitari Germans Trias I Pujol, 17. Herrendörfer G, Johnson NL, Kotz S, Kemp AW. Univariate discrete distri- Badalona, Spain. 9 Beckman Coulter, Inc., Miami, FL, USA. butions. 2nd ed. Wiley; 1992. 18. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under Received: 17 March 2021 Accepted: 31 May 2021 two or more correlated receiver operating characteristic curves: a non- parametric approach. Biometrics. 1988;44(3):837–45. 19. McGee S. Simplifying likelihood ratios. J Gen Intern Med. 2002;17(8):646– 9. https://​doi.​org/​10.​1046/j.​1525-​1497.​2002.​10750.x. 20. Deeks JJ, Altman DG. Diagnostic tests 4: likelihood ratios. BMJ. References 2004;329(7458):168–9. https://​doi.​org/​10.​1136/​bmj.​329.​7458.​168. 1. Page DB, Donnelly JP, Wang HE. Community-, healthcare-, and hospital- 21. Vincent J-L, Martin GS, Levy MM. qSOFA does not replace SIRS in acquired severe sepsis hospitalizations in the university healthsystem the definition of sepsis. Crit Care. 2016. https://​doi.​org/​10.​1186/​ consortium. Crit Care Med. 2015;43(9):1945–51. https://​doi.​org/​10.​1097/​ s13054-​016-​1389-z. CCM.​00000​00000​001164. 22. Hwang SY, Jo IJ, Lee SU, et al. Low accuracy of positive qSOFA criteria 2. EPISEPSIS: a reappraisal of the epidemiology and outcome of severe sep- for predicting 28-day mortality in critically ill septic patients during the sis in French intensive care units. Intens Care Med. 2004;30(4):580–588. early period after emergency department presentation. Ann Emerg Med. https://​doi.​org/​10.​1007/​s00134-​003-​2121-4 2018;71(1):1-9.e2. https://​doi.​org/​10.​1016/j.​annem​ergmed.​2017.​05.​022. 3. Leisman DE. Sepsis presenting in hospitals versus emergency depart- 23. Ljungström L, Pernestig A-K, Jacobsson G, Andersson R, Usener B, Tilevik ments: demographic, resuscitation, and outcome patterns in a multi- D. Diagnostic accuracy of procalcitonin, neutrophil-lymphocyte count center retrospective cohort. J Hospital Med. 2019. https://​doi.​org/​10.​ ratio, C-reactive protein, and lactate in patients with suspected bacterial 12788/​jhm.​3188. sepsis. Azevedo LCP ed. PLoS ONE. 2017;12(7):e0181704. https://​doi.​org/​ 4. Anand V, Zhang Z, Kadri SS, Klompas M, Rhee C. Epidemiology of 10.​1371/​journ​al.​pone.​01817​04. quick sequential organ failure assessment criteria in undifferentiated 24. Uusitalo-Seppälä R, Koskinen P, Leino A, Peuravuori H, Vahlberg T, Rintala patients and association with suspected infection and sepsis. Chest. EM. Early detection of severe sepsis in the emergency room: diagnostic 2019;156(2):289–97. https://​doi.​org/​10.​1016/j.​chest.​2019.​03.​032. value of plasma C-reactive protein, procalcitonin, and interleukin-6. 5. Rhodes A, Evans LE, Alhazzani W, et al. Surviving sepsis campaign: Scand J Infect Dis. 2011;43(11–12):883–90. https://​doi.​org/​10.​3109/​00365​ international guidelines for management of sepsis and septic shock: 548.​2011.​600325. 2016. Intensive Care Med. 2017;43(3):304–77. https://​doi.​org/​10.​1007/​ 25. Claessens Y-E, Mathevon T, Kierzek G, et al. Accuracy of C-reactive protein, s00134-​017-​4683-6. procalcitonin, and mid-regional pro-atrial natriuretic peptide to guide 6. Kalil AC, Machado FR. Quick sequential organ failure assessment is not site of care of community-acquired pneumonia. Intensive Care Med. good for ruling sepsis in or out. Chest. 2019;156(2):197–9. https://​doi.​org/​ 2010;36(5):799–809. https://​doi.​org/​10.​1007/​s00134-​010-​1818-4. 10.​1016/j.​chest.​2019.​06.​003. 26. Hausfater P, Juillien G, Madonna-Py B, Haroche J, Bernard M, Riou B. 7. Pierrakos C, Velissaris D, Bisdorff M, Marshall JC, Vincent J-L. Biomarkers Serum procalcitonin measurement as diagnostic and prognostic marker of sepsis: time for a reappraisal. Crit Care. 2020. https://​doi.​org/​10.​1186/​ in febrile adult patients presenting to the emergency department. Crit s13054-​020-​02993-5. Care. 2007;11(3):R60. https://​doi.​org/​10.​1186/​cc5926. 27. Hausfater P, Garric S, Ayed SB, Rosenheim M, Bernard M, Riou B. Useful- ness of procalcitonin as a marker of systemic infection in emergency Hausfater et al. Crit Care (2021) 25:227 Page 12 of 12 department patients: a prospective study. Clin Infect Dis. 2002;34(7):895– 30. Henning DJ, Puskarich MA, Self WH, et al. An emergency department vali- 901. https://​doi.​org/​10.​1086/​339198. dation of the SEP-3 sepsis and septic shock definitions and comparison 28. Wang HE, Jones AR, Donnelly JP. Revised national estimates of emer- with 1992 consensus definitions. Ann Emerg Med. 2017;70(4):544-552.e5. gency department visits for sepsis in the United States. Crit Care Med. https://​doi.​org/​10.​1016/j.​annem​ergmed.​2017.​01.​008. 2017;45(9):1443–9. https://​doi.​org/​10.​1097/​CCM.​00000​00000​002538. 29. Long B, April MD. Are qSOFA criteria better than the systemic inflamma- tory response syndrome criteria for diagnosing sepsis and predicting Publisher’s Note inhospital mortality? Ann Emerg Med. 2018;72(4):470–2. https://​doi.​org/​ Springer Nature remains neutral with regard to jurisdictional claims in pub- 10.​1016/j.​annem​ergmed.​2018.​03.​008. lished maps and institutional affiliations. Ready to submit your research ? Choose BMC and benefit from: • fast, convenient online submission • thorough peer review by experienced researchers in your field • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations • maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions