Study Design, setting and Participants
This retrospective cohort study was designed by the investigators and reported its results in accordance with the STROBE(Strengthening the Reporting of Observational Studies in Epidemiology) guidelines and performed at the Third People’s Hospital of Shenzhen between January 14, 2020, and March 10, 2020. The data cutoff for the study was March 25, 2020. The Human Research Protection Office approved data collection and analysis at our institution with a waiver of informed consent. Data were obtained from 149 patients with COVID-19 hospitalized at the Department of Critical Care Medicine and Infection Third Ward during the study dates. A confirmed case of COVID-19 was defined as a positive result on real-time reverse-transcriptase–polymerase-chain-reaction (RT-PCR) assay of pharyngeal swab specimens by Shenzhen center for disease prevention and control (CDC). The discharges criterion was negative two times 24-hour interval result on RT-PCR assay of pharyngeal swab specimens by Shenzhen CDC. The study analyzed de-identified data from the hospital’s healthcare informatics group, which was supervised by the Shenzhen Municipal Health Commission. The Second People’s Hospital approved the study protocol of Shenzhen & First Affiliated Hospital of Shenzhen University (institutional review board number 202003009004).
Exclusion criteria included without lymphocyte count for three consecutive days on day 0, day 3, day7, and without lymphocyte count for two consecutive days on day 3, day 7.
Data Collection and definitions
Patient identification—a confirmed case of COVID-19 was defined as a positive result on real-time reverse-transcriptase–polymerase-chain-reaction (RT-PCR) assay of pharyngeal swab specimens by Shenzhen center for disease prevention and control (CDC) during the study period. The discharges criterion was negative two times 24-hour interval result on RT-PCR assay of pharyngeal swab specimens by Shenzhen CDC.
Leukocyte counts—The data for blood tests that were collected from the labs that were drawn as a standard of care. Every Complete Blood Cell count (CBC) (including the white blood cell count (WBC), neutrophil count, lymphocyte count, and monocyte count) that was obtained for day0, day3, day7, day14 from admission on these patients was extracted from the medical record. If multiple leukocyte counts were collected within any 24-h period, the nadir value was documented for that period.
Definitions—Lymphopenia was defined as an absolute lymphocyte count less than 1.0×109/L, which is lower than the limit of normal(1.1×109/L) at our institution. The absolute lymphocyte count was segregated into three groups. One consisted of patients who did not develop lymphopenia within the first seven days after admission. Another consisted of patients who developed lymphopenia, which never returned to normal within the first seven days after admission. The other is to develop a decreased lymphocyte count which returns to normal within the first 7 days after admission. Persistent lymphopenia was defined as whose lymphocyte count did not return to normal within the first seven days after admission. Shock and acute respiratory distress syndrome (ARDS) were defined in accordance with the WHO interim guidance(10). Vital organ injury—Acute kidney injury was defined based on Kidney Disease: Improving Global Outcomes Clinical Practice Guidelines (KDIGO)(11). Acute cardiac dysfunction was defined as the clinical syndrome characterized by typical symptoms that may be accompanied by signs or elevated biomarkers of myocardial injury. Acute hepatic injury was defined as a state in which the patient’s blood laboratory results met at least one of three criteria: total serum bilirubin (TBil) of 3.0 mg/dL or greater; aspartate aminotransferase (AST) of 41 IU/L or higher; alanine aminotransferase (ALT) of 41 IU/L or higher; The patients who met none of these criteria were classified as the “normal liver function” group(12).
Baseline characteristics—Detailed chart review was then performed to exclude patients based on inclusion and exclusion criteria and to collect demographic and outcome data. For every patient, all routinely collected vital signs and symptoms, and laboratory values were extracted from the electronic health records. Data included, but were not limited to, demographic data (e.g., age, gender, body mass index [BMI]), biochemical parameters (e.g., blood cell count, liver function, kidney function, coagulation function, blood gas analysis), mechanic ventilation and comorbidities including hypertension, diabetes, cerebrovascular disease, chronic obstructive pulmonary disease (COPD), and malignancy for the severity of illness. We calculated the Acute Physiology and Chronic Health Evaluation (APACHE) II score within the first 24 hours of hospitalization. The study analyzed de-identified data from the hospital’s healthcare informatics group, which was supervised by the Shenzhen Municipal Health Commission.
Main Measures and outcomes
The primary outcome was ARDS, and the Secondary outcomes included the development of vital organ dysfunction and hospital lengths of stay.
Statistical Analysis
The categorical data were summarized as numbers and percentages, and inter-group comparisons were performed using χ2 tests or Fisher’s exact test. Continuous variables were expressed as the arithmetic mean and standard deviation (SD) or as the median and interquartile range, depending on whether or not they showed a Gaussian distribution. Continuous data with Gaussian distribution were compared with the Student’s t-test or one-way ANOVA and those with a non-gaussian distribution, with the Wilcoxon rank-sum test. To compare the white blood cell, lymphocytes, neutrophils and monocyte count between ARDS and non-ARDS groups over time and to correlate for within-subject correlation, a linear mixed-effect model was used with an autoregressive covariance structure. To determine the independent effect of persistent lymphopenia on ARDS after accounting for significant confounders, multivariable logistic regression with a forced entry method was used with crude model and fully adjusted model: OR (odds ratio)and 95% confidence interval levels (95% CI). The predictive ability of lymphocyte count on day0, day3, and day7 were assessed using the AU-ROC curve method. Youden’s index determined the optimal cutoff value. Statistical analysis was performed using the statistical package SAS 9.4 (Windows, SAS Institute, Cary, North Carolina) and the R software (version 3.6.2). P values (two-tailed) below 0.05 were considered statistically significant.