Predicting acute respiratory distress syndrome in influenza pneumonia patients using delta mean platelet volume

This retrospective study was conducted using a health information system database of patients admitted to wards or the medical intensive care unit (ICU) of Songklanagarind Hospital (a university-affiliated, 800-bed tertiary hospital in southern Thailand) between January 2015 and December 2020.

Patients were included if they had been admitted with pneumonia and had influenza infection. Patients were excluded from this study if they were younger than 18 years old, had initial presentation of ARDS, or inadequate complete blood cell count data. The protocol of this study was approved by the ethics committee of the Faculty of Medicine, Prince of Songkla University (EC number: 63-482-14-1). The requirement for informed consent was waived due to the retrospective nature of the study.

The study participants’ epidemiologic and clinical data were obtained from the electronic medical record. The data included age, sex, height, body weight, comorbid conditions, illness severity scores, laboratory values, respiratory intervention and clinical outcomes. Complete blood cell counts including MPV values were determined using an automated haematology analyser (XT-1800i, Sysmex Corporation, Japan).

Influenza infection was confirmed by a positive result from one of the following tests: rapid antigen test, or nucleic reverse transcriptase polymerase chain reaction (RT-PCR) from nasopharyngeal swab, throat swab, sputum or bronchoalveolar lavage.

Pneumonia was defined as the presence of at least one of the following: fever (body temperature > 38.2 °C) or hypothermia (temperature < 35.0 °C), new cough with or without sputum production, dyspnoea or altered breathing sound on auscultation, or presence of new chest radiographic infiltrates [8, 9]. ARDS was diagnosed according to the Berlin definition: acute onset within 1 week, bilateral lung opacities, no evidence of cardiac failure-related hydrostatic oedema by echocardiography, and PaO₂/FiO₂ ratio < 300 mm Hg with positive end-expiratory pressure (PEEP) ≥ 5 cm H₂O [10].

Illness severity was determined using acute physiology and chronic health evaluation II (APACHE II) [11], sequential organ failure assessment (SOFA) score [12], national early warning score (NEWS) [13], CURB-65 (confusion, urea > 7 mmol/L, respiratory rate ≥ 30/min, blood pressure [systolic < 90 mm Hg or diastolic ≤ 60 mm Hg], and age ≥ 65 years) pneumonia severity score [14], and pneumonia severity index (PSI) [15]. All components of the variables of APACHE II, SOFA, NEWS, CURB-65 and PSI were recorded, capturing the highest and lowest value during the first 24 h of hospital admission (Additional file 1).

Day 1 was defined as the interval from admission to 5:00 am the next day; all other days were calendar days from 5:00 am to 4.59 am.

Delta MPV was defined as MPV on day 2 minus MPV on day 1. Two sets of CBC were selected to determine delta MPV, the first one at the time of admission, and the second one on the next calendar day when blood samples are routinely taken at 6 am. MPV/platelet count ratio (MPR) was defined as MPV (fL)/platelet count (^10⁹/L) × 100%.

The sample size was determined using sample size estimation for diagnostic test studies [16]. Estimating that 30% of influenza pneumonia patients have ARDS, we calculated that a sample size of 205 would provide 95% confidence with 80% power to detect a difference of 10% from the presumptive value of 80% for sensitivity.

The data were tested for normality using the Shapiro–Wilk test. Categorical data are expressed as percentages. Continuous data are shown as mean ± standard deviation or median with minimum and maximum interquartile range (IQR), depending on the distribution of the data. Continuous variables and proportions were compared between groups using Student’s t-test or the Mann–Whitney U test and chi-square tests, respectively.

Of 1240 patients with a positive test for influenza between January 2015 and December 2020, 212 were enrolled in the study (Fig. 1). The median (interquartile range, IQR) patient age was 71 (59–79) years, and 110 (51.9%) were female. The median SOFA score was 2 (1–3) and the median APACHE II score was 12 (8–15). The most common type of influenza was influenza A (80.7%). Overall, in-hospital mortality was 9%.

The baseline demographic, clinical, and biochemical data of each group, stratified according to ARDS, are presented in Table 1. Of 212 pneumonia patients, 56 (26.4%) patients were diagnosed with ARDS. The median number of days from hospital admission to develop ARDS was 3 (1–6) days (Additional file 1). Patients with ARDS had significantly higher severity scores, including SOFA score [3 (2–4) vs 2 (1–3), p < 0.001], APACHE II score [14 (12–18.8) vs 12 (7–13), p < 0.001], CURB-65 [2 (2–3) vs 2 (1–2), p < 0.001] and PSI [99.5 (87.2–108.8) vs 88 (69–103.8), p = 0.006], than the non-ARDS group. White blood cell count [10.7 (6.9–13.8) vs 9.2 (6.3–12.7) × 10³/µL, p = 0.024], MPV on day 2 [11 (10.3–11.5) vs 10.3 (9.6–10.9) fL, p < 0.001], MPR on day 2 [6.4 (4.4–8.4) vs 4.8 (3.6–6.9), p = 0.002] and delta MPV [1 (0.7–1.2) vs 0.2 (0–0.5) fL, p < 0.001] were significantly higher in the ARDS group, whilst the percentage of lymphocytes was significantly lower in the ARDS group compared with the non-ARDS group [8.9 (5.9–14.4) vs 12 (7–19.1), p = 0.008]. Comparison of delta MPV between the two groups is shown in Fig. 2. The ARDS group had a longer length of hospital stay and higher mortality rate [21 (12–23.8) vs 7 (4–11) days, p < 0.001; 26.8% vs 2.6%, p < 0.001] (Table 1).

Multivariable logistic regression analysis revealed that delta MPV, APACHE II scores and CURB-65 are independent predictors of ARDS in patients with influenza pneumonia (Table 2).

ROC curve analysis showed better performance of delta MPV than CURB-65 in predicting ARDS (AUROC 0.855 vs 0.696) (Fig. 3). A cut-off value of 0.7 fL for delta MPV (sensitivity 80.36%, specificity 80.77%, LR + 4.18, LR- 0.24) and ≥ 2 for CURB-65 (sensitivity 83.93%, specificity 42.95%, LR + 1.47, LR- 0.37) was used to predict ARDS in influenza pneumonia patients.