High Resolution Computed Tomography of Thorax- Is it an Early Predictor of Hypoxaemia in COVID-19 Patients?

Novel Coronavirus-2019 (nCoV-2019) is capable of human-to-human transmission and can lead to acute respiratory distress syndrome similar to Middle East Respiratory Syndrome (MERS) due to lung parenchyma destruction. Some patients with COVID-19 consistently demonstrated no hypoxaemia, however, some patients develop sense of difficulty in breathing due to increased airway resistance. Aim: To assess the potential of High Resolution Computed Tomography (HRCT) thorax as an early predictor of hypoxaemia in COVID-19 patients. Materials and Methods: A prospective longitudinal cohort study of 1000 Reverse Transcription Polymerase Chain Reaction (RT-PCR) confirmed COVID-19 and HRCT thorax positive patients, who were monitored simultaneously for SpO2 levels, were undertaken. HRCT findings were graded into Computerised Tomography Severity Index (CTSI) and correlated with patient’s SpO2 levels, at the time of scan on admission. Patients, who had normal SpO2 levels (≥95%) at the time of initial scan, were monitored upto five days. Pearson’s correlation test was used to find correlation between CTSI and SpO2 levels. Results: In present study group there was male predominance (4:1). Fever was the most common clinical presentation followed by cough. HRCT thorax features were categorised as Typical 769 (76.9%), Indeterminate 176 (17.6%) and atypical 55 (5.5%). 371 (82.8%) patients with SpO2 >95% were having CTSI between 0-7, similarly 189 (54.4%) patients with SpO290-94% were having CTSI between 8-15 and 133 (64.8%) patients with SpO2 <90% were having CTSI between 16-25. So, the present study categorised the patients into three groups- Category 1 (CTSI 0-7), Category 2 (CTSI 8-15) and Category 3 (CTSI 16-25) for better and prompt identification of clinical severity and their management. Majority of patients in CTSI category 1, 2 and 3 were having SpO2 levels ≥95%, 90-94% and <90%, respectively. Statistical correlation between CTSI and SpO2 levels at the time of initial scan was significant (Pearson’s correlation coefficient (r)=-0.261 and p-value <0.01). Number of patients who developed hypoxaemia (SpO2 <95%) on follow-up in CTSI Category 1, 2 and 3 were 42 (11.32%), 10 15.87%) and 2 (14.28%), respectively. The association between CTSI and development of hypoxaemia based on follow-up SpO2 levels was statistically found to be insignificant (chi-square value=1.21, degree of freedom (d.f.) 2 and p-value=0.570). Conclusion: In present study group, a negative correlation was established between CTSI and SpO2 levels. The association between CTSI and development of hypoxaemia on follow-up SpO2 monitoring was found to be non-significant statistically. So, HRCT thorax cannot be relied upon as an early predictor of hypoxaemia in COVID-19 patients.