Serum calcium as a biomarker of clinical severity and prognosis in patients with coronavirus disease 2019

Serum calcium and clinical variables

The median serum calcium levels were 2.12 (IQR, 2.04-2.20) mmol/L on admission. We divided the patients into three groups based on the serum calcium values: ≤2.0 mmol/L (defined as group A, n = 43), 2.0-2.2 mmol/L (defined as group B, n = 137), and >2.2 mmol/L (defined as group C, n = 61). As shown in Table 2, significant differences in the clinical variables except for serum creatinine were found among the three groups, and the same differences were found between groups A and B (P <0.05). There were also differences in the clinical variables except for WBC count (P =0.07) and serum creatinine (P =0.244) between groups A and C, whereas differences in the variables except for WBC count (P =0.60), ALT (P =0.839), the lowest SpO2 (P =0.328), and serum creatinine (P =0.635) were found between groups B and C. These results indicated that patients with lower serum calcium levels had worse clinical variables.

Of the 241 patients, 26 were tested to determine levels of parathyroid hormone (PTH) and 25-hydroxy-vitamin D (VD) according to clinical needs. The median serum calcium level of the 26 patients was 2.13 (IQR, 2.03-2.16) mmol/L. The median PTH level was 55.27 (IQR, 42.73-73.15) pg/mL, and the median VD level was 10.20 (IQR, 8.20-12.65) ng/mL. All of these patients had low levels of VD (VD deficiency).

The SPSS scatterplots and correlation analyses of serum calcium and the blood biomarkers are shown in Figures 1 and 2. The serum calcium levels were significantly positively correlated with lymphocyte count (Figure 1A, P <0.001), albumin levels (Figure 1B, P <0.001), VD levels (Figure 1C, P =0.004), and lowest SpO2 (Figure 1D, P< 0.001), whereas they are significantly negatively correlated with CRP (Figure 2A, P< 0.001), D-dimer (Figure 2B, P< 0.001) and PTH (Figure 2C, P =0.048) levels. These results indicated that hypocalcemia may be associated with imbalanced VD and PTH in the acute phase of COVID-19.

Serum calcium and clinical severity

ARDS developed in 19 of 241 (7.9%) patients, and liver injury developed in 16 (6.6%), AKI developed in 14 (5.8%), and cardiac injury developed in 12 (5.0%) patients during the research period. Twelve patients received MV, and 7 patients received CRRT. As shown in Table 3, significant differences in the clinical severity and outcome variables were found among the abovementioned three groups (A, B, and C) (P <0.001) and between groups A and C (P <0.001). There were also differences in all these variables except for liver injury incidence (P =0.201). No differences were found between groups B and C (P >0.05). These results indicated that patients with serum calcium values ≤2.0 mmol/L had higher 28-day mortality, and a higher incidence of organ injury. Moreover, the serum calcium values were significantly lower in patients who died and in patients with MODS, septic shock, and organ injury, requiring MV or CRRT (P <0.001) (Table 4).

ROC curves were also performed to assess the associations between serum calcium and MODS, septic shock, and 28-day mortality. As shown in Figure 3, the area under the curves (AUCs) of MODS (Figure 3A), septic shock (Figure 3B), and 28-day mortality (Figure 3C) were 0.923 (P <0.001), 0.905 (P =0.001), and 0.929 (P <0.001), respectively. Optimal cut-off points of serum calcium values were derived from the ROC curves. The optimal cut-off point for MODS was 2.035 mmol/L, the sensitivity was 88.2%, and the specificity was 82.6%. The optimal cut-off point for septic shock was 2.01 mmol/L, the sensitivity was 100.0%, and the specificity was 84.3%. The optimal cut-off point for 28-day mortality was 2.01 mmol/L, the sensitivity was 100.0%, and the specificity was 85.7%. Figure 4 shows the significant differences in the 28-day mortality among groups A, B, and C. The 28-day mortality of group A was significantly higher than that of groups B or C (P <0.001). No difference was found between groups B and C (P >0.05).

Patients

From February 10 to February 28, 2020, adult patients (age ≥18 years) with confirmed COVID-19 admitted to our specialized isolation units, Tongji Hospital of Huazhong University of Science and Technology in Wuhan, were enrolled in this clinical retrospective study. Patients with chronic organ dysfunction (e.g., hepatic or renal dysfunction), terminal cancer, immunodeficiency, and a history of long-term use of hormones were excluded. Written informed consent was waived by our institutional review board because this was a retrospective study that assessed deidentified data and included no potential risk to patients. The diagnosis of COVID-19 was made according to the WHO interim guidance and the recommendations of the National Health Commission of China [4, 5], and confirmed by RNA detection of SARS-CoV-2 in the clinical laboratory of Tongji Hospital.

Definitions

An identified case of COVID-19 was defined as a positive finding by real-time reverse transcriptase–polymerase chain reaction (RT-PCR) assay of nasal and pharyngeal swab specimens [4, 5, 7]. Only laboratory-confirmed cases were enrolled in the analysis. The clinical classifications of COVID-19 were in accordance with the Chinese recommendations [5]: Mild, with minor clinical symptoms (e.g., fever, cough) without imaging manifestations. Moderate, with fever or respiratory tract infection symptoms with imaging indicating pneumonia. Severe, met any of the following, I—respiratory distress and respiratory rate ≥30 breaths/min; II—pulse oxygen saturation (SpO2) ≤93% at rest; or III—arterial partial pressure of oxygen (PaO2)/ fraction of inspired oxygen (FiO2) ≤300 mmHg (1 mmHg =0.133 kPa). Critical, met any of the following, I—respiratory failure with mechanical ventilation (MV); II—shock; or III—multiple organ failure requiring ICU treatment. Hypocalcemia was defined as a serum calcium level less than 2.2 mmol/L in our clinical laboratory. Sepsis was defined as life-threatening organ dysfunction caused by a dysregulated host response to infection, and septic shock was defined as a subset of sepsis with circulatory and cellular/metabolic dysfunction that is associated with a higher risk of mortality [11]. The diagnostic criteria of acute respiratory distress syndrome (ARDS) were in accordance with the Berlin definitions [12]. The definitions of acute kidney injury (AKI) were based on the 2012 Kidney Disease: Improving Global Outcomes (KDIGO) guidelines [13]. Cardiac injury was defined if serum levels of cardiac biomarkers (e.g., troponin I) were more than twice the reference upper limit or new abnormalities were found in electrocardiography and echocardiography [2]. Liver injury was defined if serum levels of hepatic biomarkers (e.g., alanine aminotransferase) were more than twice the reference upper limit or if there was disproportionate elevation of alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels compared with alkaline phosphatase levels [14]. Multiple organ dysfunction syndrome (MODS) was defined as the combined dysfunction of two or more organs.

Data collection

The baseline clinical characteristics, including age, sex, days from onset to admission, initial symptoms or signs, and clinical classifications were collected from electronic medical records, and all laboratory tests were performed according to the clinical needs of patients. The levels of serum calcium, C-reactive protein (CRP), ALT, albumin, creatinine, troponin I (TNI), and plasma D-dimer and white blood cell (WBC) count, lymphocyte count, and the lowest SpO2 within 24 hours of admission were recorded. The hormone levels associated with blood calcium (e.g., parathyroid hormone, 25-hydroxy-vitamin D) were also recorded. All blood parameters were detected by the clinical laboratory of Tongji Hospital. Moreover, the numbers of patients with ARDS, AKI, cardiac injury, liver injury, septic shock and MODS and patients receiving noninvasive ventilation (NIV), high-flow nasal cannula (HFNC), MV, and continuous renal replacement therapy (CRRT) were also recorded. The primary endpoints were the development of septic shock, MODS, and 28-day mortality. The secondary endpoints were the other disease severity parameters (e.g., organ injury or not).

Statistical analysis

The Kolmogorov-Smirnov test was first performed to test the normal distribution of the data. Normally distributed data were expressed as the means ± standard deviation and were compared by t tests. Abnormally distributed data were expressed as the medians (interquartile ranges, IQR) and were compared by the Mann-Whitney U test or the Kruskal-Wallis test. Categorical variables were presented as absolute numbers or percentages and were analyzed using the χ² test or Fisher’s exact test. To take into account the repeated nature of the variables, analysis of variance (ANOVA) for repeated measurements of the general linear model was implemented. Receiver operating characteristic (ROC) curves were used to evaluate the associations between serum calcium and septic shock, MODS, and 28-day mortality. IBM Statistical Package for the Social Sciences (SPSS, version 22.0, NY, USA) software was used for statistical analysis, and P <0.05 was considered statistically significant. SPSS scatterplots and a correlation analysis were performed to evaluate the relevance between serum calcium and blood biomarkers. The statistical methods of this study were reviewed by Qiao Liu, a biostatistician from the Center for Disease Control and Prevention of Jiangsu Province in China.