Introduction

Chronic obstructive pulmonary disease (COPD) is a leading cause of morbidity and mortality worldwide [1]. COPD exacerbation (ECOPD) is a major outcome of this disease and ECOPD severity is associated with future mortality risk [2], a decline in FEV1 [3], diminished quality of life [4] and future ECOPD [5, 6]. Moreover, ECOPD hospitalizations are associated with high mortality rates [7, 8]. Nonetheless, few tools are available to help clinicians to determine ECOPD prognosis in hospitalized patients. These tools include C-reactive protein (C-RP), the only biomarker recommended by GOLD [9] to evaluate ECOPD severity, and the dyspnea–eosinopenia–consolidation–acidemia–atrial fibrillation (DECAF) prognostic score, which uses several clinical variables (dyspnea, eosinopenia, consolidation, acidemia and atrial fibrillation) to establish the prognosis of severe ECOPD [10]. Therefore, new parameters must be investigated in clinical practice to determine prognosis in patients with severe ECOPD.

The inflammatory response during ECOPD is mediated by the activation of neutrophilic and lymphocytic inflammation [11]. Circulating neutrophils and monocytes are involved in the initial response to pathogenic organisms that cause ECOPD [12]. Monocyte distribution width (MDW) reflects the degree of change in circulating monocyte volume (“heterogeneity”) in response to proinflammatory signals elicited by infectious organisms. MDW determination is easily automated, inexpensive and quickly obtained as part of routine blood counts. Recent data have suggested that an increase in MDW may be useful in early detection of all-cause sepsis [13,14,15,16,17,18,19,20,21]. However, few studies have been performed in acute respiratory conditions; one example is a study in a limited number of patients with COVID-19 [22,23,24]. Some studies have suggested that MDW may be associated with the neutrophil/lymphocyte ratio, a biomarker associated with ECOPD prognosis [11].

We hypothesized that high MDW at hospital admission might serve as a prognostic inflammatory biomarker in patients with ECOPD and that combining the MDW with the DECAF score might further increase the prognostic value. To our knowledge, this aspect has not previously been studied in this setting.

Methods

This was a multicenter observational retrospective study in which clinical records from patients admitted for ECOPD between March 1st, 2020, and March 1st, 2023, to three public hospitals of the Servicio Cántabro de Salud network in the Cantabria community in northern Spain were reviewed. The ethics committee of our institution (2023.061) approved the study.

Participants

We recruited patients hospitalized for ECOPD between March 2020 and March 2023.

The inclusion criteria were as follows: (1) Patients older than 40 years previously diagnosed with COPD according to the GesEPOC guidelines [25] and (2) Patients hospitalized because of ECOPD. ECOPD was defined by an increase in more than one of the following respiratory symptoms: dyspnea, sputum purulence, increased sputum, cough or wheezing; symptoms persisting for at least 2 consecutive days; and symptoms requiring treatment with antibiotics and/or systemic steroids [9].

The exclusion criteria were as follows: (1) Patients with a diagnosis other than ECOPD at discharge; (2) Patients without clinical, biochemical or microbiological data consistent with ECOPD; (3) Patients with active cancer, leukemia, lymphoma, lymphoproliferative disorders, bone marrow diseases or AIDS; (4) Patients with vitamin B12 or folic acid deficiency; and (5) Patients treated with immunosuppressants (including systemic corticosteroids) or drugs causing macrocytosis.

Measurements

Stable spirometry was performed according to the Spanish Society of Pulmonology and Thoracic Surgery (SEPAR) protocol [26] less than 1 year before admission. Patient age; sex; smoking status; number of moderate and severe ECOPD events 1 year before admission; comorbidities included in the Charlson index [27]; basal mMRC dyspnea score; and dates of hospital admission, ICU admission and/or death during hospitalization were retrospectively recorded. The DECAF score [25] was calculated for all patients. The MDW and results of routine blood tests performed after arrival at the emergency department of each center were retrospectively collected.

Routine hematological and biochemical analytes were measured with automated assays. Specifically, C-RP was measured with Siemens traceable enzymatic method assays (Atellica Analyzer, Siemens, Germany).

MDW was measured in the emergency laboratory of each center with the same DxH 900 analyzer (Beckman Coulter. Inc., Brea, California, USA), according to the manufacturer´s instructions.

Statistical analysis

Data are presented as mean ± SD for normally distributed data or median (interquartile range) for nonparametric data. We calculated sample sizes in Stata Statistical Software: Release 15. College Station, TX: StataCorp LLC.), with an α level of 0.05 and a β level of 0.2. Differences between groups were analyzed with unpaired t tests for parametric data or Mann–Whitney tests for nonparametric data. Normality of distribution was evaluated with the Kolmogorov–Smirnov test. We evaluated the correlation between MDW and other variables with Spearman tests. Evaluation of MDW as a dichotomized variable was established with a cutoff at 21.5 units, according to previous studies [28,29,30] performed in the same setting and using similar laboratory protocols as our study. We evaluated cross-sectional associations with high versus low MDW through univariate and multivariate logistic regression, with outcome variables of mortality; ICU admission; and a composite end point including in-hospital all-cause mortality and escalation to ICU admission. We used Kaplan–Meier estimates to calculate the proportion of participants experiencing mortality; ICU admission; and a composite end point including in-hospital all-cause mortality and escalation to ICU admission due to ECOPD over time. We performed univariate and multivariate Cox proportional risk analysis in SPSS version 25.00. A receiver operating characteristic (ROC) curve and the area under the ROC curve (AUC) were used to assess the diagnostic value of MDW for a composite end point including in-hospital all-cause mortality and escalation to ICU admission as the outcome variables. ROC curve analysis was performed in MEDCALC version 11.6.1.0 (MedCalc Software, Mariakerke, Belgium). Differences with p values < 0.05 were considered significant. All reported p values were two sided.

Results

Patient characteristics

A total of 474 pw ECOPD were ultimately included in the study (flowchart for patient selection in Fig. 1; demographic, clinical and biochemical data in Table 1). The median patient age was 75 (67–82) years, and most patients were men (67.7%). The prevalence of current smokers (32.9%) was high. Most patients had a previous admission for ECOPD [178 (37.6%)] and moderate or severe obstruction, on the basis of FEV1 (%) 50 (35–67). A total of 26 patients died (5.5%); although ICU admission was not frequent [31 (6.5)], a composite end point including mortality or ICU admission reached 11% (52 patients). The blood MDW levels were 19.2 (17.2–21.2) units, the blood leucocyte levels were 9900 (7300–13525) cells/µL and the serum C-RP levels were 4.4 (1.5–11.7) mg/dL.

Fig. 1
figure 1

Flowchart for patient selection

Full size image
Table 1 Demographic, clinical and biochemical characteristics of all patients and patients with MDW ≤ 21.5 units vs MDW > 21.5 units
Full size table

The group of patients with ECOPD and high MDW tended to have high ECOPD severity, on the basis of the DECAF score, mortality, ICU admission rate and elevated inflammatory parameters, such as C-RP and the neutrophil to lymphocyte ratio. Notably, the group with high MDW had slightly higher FEV1 values. However, the two groups had similar age, sex distribution, smoking habits, basal mMRC scores and ECOPD history.

MDW correlations

MDW was positively correlated with the DECAF score (r = 0.184, p < 0.001), FEV1 (L) (r = 0.153, p = 0.001), FEV1 (% predicted) (r = 0.149, p = 0.001), FEV1/FVC (r = 0.106, p = 0.022), C-RP (mg/dL) (r = 0.571, p < 0.001) and neutrophil to lymphocyte ratio (r = 0.135, p = 0.003).

MDW was negatively correlated with basophil count (r = −0.111, p = 0.016) and eosinophil count (r = −0.288, p < 0.001).

MDW did not correlate with age (r = 0.025, p = 0.588), Charlson score (r = 0.056, p = 0.291), BMI (r = −0.02, p = 0.978), FVC (L) (r = 0.069, p = 0.145), FVC FEV1 (% predicted) (r = 0.077, p = 0.099), paO2 (r = -0.084, p = 0.069), paCO2 (r = −0.088, p = 0.056), HCO3 (r = −0.065, p = 0.173), pH (r = 0.088, p = 0.056), leucocyte count (r = 0.024, p = 0.597), neutrophil count (r = 0.063, p = 0.172), monocyte count (r = 0.050, p = 0.274) or lymphocyte count (r = −0.080, p = 0.081).

Association of MDW with ECOPD characteristics

Table 2 highlights the associations of MDW with baseline ECOPD characteristics. Multivariate logistic regression analysis showed that high MDW was positively associated with C-RP (OR 1.115 95% CI 1.076–1.155, p < 0.001), death (OR 9.831 95% CI 2.981–32.417, p < 0.001) and ICU admission (OR 11.204 95% CI 3.173–39.562, p < 0.001); however, no associations with other ECOPD characteristics were observed.

Table 2 Univariate and multivariate logistic regression analysis of the associations between ECOPD characteristics and high MDW
Full size table

MDW as a predictor of mortality, ICU admission or both

Among 474 patients with ECOPD included in the study, 109 had a high MDW. Table 1 shows the clinical characteristics of both groups. A total of 26 patients died during hospitalization (18 in the high MDW group), 31 patients were admitted to the ICU (17 in the high MDW group) and 52 patients met the composite end point including mortality or ICU admission (30 in the high MDW group).

Predictors of mortality

Univariate Cox proportional risk analysis indicated that age (p = 0.04), sex (women) (p = 0.292), MDW (p = 0.001) and high MDW (higher than 21.5) (p = 0.002), but not smoking status (p = 0.064), FEV1 (p = 0.626), Charlson index (p = 0.101), DECAF score (p = 0.069), C-RP (p = 0.127) or neutrophil to lymphocyte ratio (p = 0.416) were predictors of mortality during hospitalization due to ECOPD. Multivariate Cox proportional risk analysis revealed that the absolute values of MDW (HR 1.171, CI 95% 1.073–1.277, p < 0.001) (Table 3) and high MDW (HR 3.647, CI 95% 1.313–10.136, p = 0.013) (Table 3, Fig. 2) were independent risk factors for mortality during hospitalization for ECOPD.

Table 3 Cox regression analysis showing absolute and dichotomized values of MDW as a predictor of death, ICU admission and the composite end point (mortality or ICU admission)
Full size table
Fig. 2
figure 2

High MDW (> 21.5 units) as a predictor of A ICU admission, B mortality and C ICU admission or mortality. MDW monocyte distribution width

Full size image

Predictors of ICU admission

Univariate Cox proportional risk analysis indicated that age (p < 0.001), smoking status (p = 0.007), DECAF score (p < 0.001), C-RP (p = 0.002), MDW (p < 0.001) and high MDW (higher than 21.5) (p < 0.001), but not sex (p = 0.292), FEV1 (p = 0.802), Charlson index (p = 0.995) or neutrophil to lymphocyte ratio (p = 0.654) were predictors of ICU admission. Multivariate Cox proportional risk analysis revealed that absolute values of MDW (HR 1.086, CI 95% 1.009–1.169, p = 0.027) (Table 3) and high MDW (HR 2.550, CI 95% 1.131–5.753, p = 0.024) (Table 3, Fig. 2) were independent risk factors for ICU admission.

Predictors of the composite end point (mortality or ICU admission)

Univariate Cox proportional risk analysis indicated that age (p = 0.047), DECAF score (p < 0.001), C-RP (p < 0.001), MDW (p < 0.001) and high MDW (higher than 21.5) (p < 0.001), but not smoking status (p = 0.072), sex (p = 0.767), FEV1 (p = 0.786), Charlson index (p = 0.100) or neutrophil to lymphocyte ratio (p = 0.754) were predictors of the composite end point (mortality or ICU admission). Multivariate Cox proportional risk analysis revealed that absolute values of MDW (HR 1.122, CI 95% 1.065–1.182, p < 0.001) (Table 3) and high MDW (HR 3.084, CI 95% 1.624–5.858, p = 0.001) (Table 3, Fig. 2) were independent risk factors for the composite end point (mortality or ICU admission).

Potential utility of the MDW and DECAF–MDW score in predicting the composite end point

According to our findings, we created a new tool, the MDW–DECAF score, based on the DECAF score and including the same variables as DECAF and the MDW, with a relative weight assigned according to the regression coefficient for the composite end point (3 points). In ROC analysis (Fig. 3), the MDW–DECAF score’s AUC for differentiating patients who died or were admitted to the ICU from the rest of the patients (AUC 0.777 95% IC 0.708–0.845, p < 0.001) had the best diagnostic power and was followed by the DECAF score (AUC 0.710 95% IC 0.639–0.782, p < 0.001) and MDW (AUC 0.705 95% IC 0.618–0.791, p < 0.001) (Fig. 3, Supplementary file 1). Youden’s index for MDW was > 21.7, with a sensitivity of 57.69 and a specificity of 81.95. Youden’s index for the MDW–DECAF score was > 2, with a sensitivity of 84.62 and a specificity of 63.01. MDW–DECAF had a statistically significantly higher AUC than the DECAF score (p = 0.023), MDW (p = 0.026), C-RP (p = 0.002) and neutrophil to lymphocyte ratio. No statistically significant differences were found among the AUC values of the remaining variables.

Fig. 3
figure 3

Receiver operator characteristic curve showing the discrimination ability of MDW, C-reactive protein, neutrophil to lymphocyte ratio, the DECAF score and the MDW–DECAF score for in-hospital mortality or ICU admission. MDW monocyte distribution width, DECAF dyspnea, eosinopenia, consolidation, acidemia and atrial fibrillation score

Full size image

Discussion

This study provides the first demonstration that MDW is associated with the severity of severe ECOPD and can be used as a predictor of mortality and ICU admission. Moreover, it introduces the new MDW–DECAF score.

Patients with high MDW, as reported by previous studies in patients with systemic inflammatory response syndrome [28, 29] or COVID-19 [30], had elevated C-RP, DECAF scores and neutrophil to lymphocyte ratios, and lower blood eosinophils; these prognostic factors are well known to be associated with both ECOPD inflammatory response and severity [10, 11, 31,32,33]. We also evaluated variables associated with higher MDW values in a multivariate logistic regression analysis, which indicated that C-RP, mortality and ICU admission were associated with high MDW values. Our findings indicate the importance of inflammation in MDW levels, as discussed in other clinical contexts, such as sepsis [16, 19, 20, 24, 28, 29], COVID-19 [22,23,24], influenza [34] and complicated diverticulitis [35].

Our findings provide the first evidence that the MDW values according to blood tests performed after emergency department arrival predict death and ICU admission for ECOPD. This novel finding has not previously been described in the context of COPD, but has been reported in other settings, such as COVID-19 [23] and sepsis [28]. Furthermore, our results were obtained by using Cox regression analysis considering the time to the event, whereas other studies have evaluated ECOPD prognosis by using only logistic regression analysis [10, 11]. Additionally, the cutoff point of the test was determined on the basis of previous studies [28,29,30]. This was similar to Youden’s index in our study.

The AUC of MDW for predicting the prognosis of severe ECOPD in terms of mortality or ICU admission was comparable to that of other well-known inflammatory markers of COPD, such as C-RP [9] or the neutrophil to lymphocyte ratio [11]. MDW is biomarker that can be routinely measured rapidly, easily and inexpensively in emergency departments. Further studies are necessary to evaluate the limitations and benefits of each biomarker in COPD. The DECAF score [10] is a well-known risk stratification tool for patients with severe ECOPD, but its AUC can be improved by using novel biomarkers. Because MDW is a different predictor from the DECAF score for mortality or ICU admission, we created a new score, the MDW–DECAF score, which had a better AUC than the other biomarkers or the DECAF score alone. Further studies are required to evaluate the potential roles of these scores in assessing specific etiologies of ECOPD or inflammatory conditions.

This study has several strengths. First, it is a novel study reporting the first evaluation of MDW in ECOPD in real-world circumstances. Second, this was a multicenter study in patients from three hospitals. Finally, the patients included in this study were carefully selected and well characterized, and patients with diseases or therapies that might have influenced the results were excluded.

However, our study also has several limitations. First, this was a retrospective study and therefore was subject to a risk of information bias. Although all routine blood tests were performed minutes after arrival at the emergency department, the retrospective nature of this study could imply different timing for analysis from whole blood venous sample collection. Furthermore, our results cannot be extrapolated to all patients with COPD since we excluded patients with hematological and nutritional conditions. Our findings should be replicated in other settings, in a larger number of patients, with standardized therapy and multiple samples, to evaluate the time course of the responses of MDW. Future specifically designed prospective studies should be performed to evaluate the utility of MDW and externally validate MDW–DECAF.

Conclusion

Our study provides the first evidence that MDW is associated with ECOPD severity and predicts mortality and ICU admission with a diagnostic accuracy similar to that of DECAF and C-RP. Furthermore, on the basis of our results, we created a new tool, the MDW–DECAF score, which has better diagnostic accuracy than the DECAF score in identifying mortality or ICU admission.