With interest we read the article by Jesmani et al. about a retrospective study on the clinical presentation and outcome of 665 COVID-19 patients admitted to Kermanshah hospital between 22 October 2020 and 22 November 2020.1 In this study, 85% of patients had oxygen saturation (SPO2) <93% on admission. The most common comorbidities were arterial hypertension, diabetes, and cardiovascular disease. Among 655 of the included COVID-19 patients, 82 died within 7 days after admission. The in-hospital fatality rate was 12.33%. Each one unit increase in SPO2 increased the odds of survival by 0.88 times while each one day increase in the length of the intensive care unit (ICU) stay decreased the odds of death by 0.49 times.1 The study is excellent, but has limitations that are cause of concern and should be discussed.
The main limitation of the study is its retrospective design. Retrospective studies rely on data collected by many others and can usually neither be controlled nor corrected. There is also the disadvantage that missing data cannot be supplemented and that open questions concerning individual data cannot be inquired and obtained.
Other limitations of the study are that none of the patients that died underwent autopsy and that there is no definition in the methodology for “death from COVID-19”.1 Autopsy is critical particularly in overlap cases, in which the cause of death could be due to several different diseases that were present simultaneously. As to the lack of definition of specific cause of death being COVID-19, and because causes of death are usually not easy to identify unequivocally, it is crucial to assess how myocardial infarction or ischemic stroke, unrelated to the current SARS-CoV-2 infection, were delineated from SARS-CoV-2 pneumonia as the cause of death. Were complications from SARSCoV-2 infection, such as venous sinus thrombosis (VST), ischemic stroke, encephalitis, status epilepticus regarded as death due to COVID-19 or only SARS-CoV-2 pneumonia or acute respiratory distress syndrome (ARDS)?
Besides, there is no mention of treatment administered to these patients during hospitalisation. It is known that several drugs administered to SARS-CoV-2 infected patients (e.g. chloroquine) can be harmful and can increase severity of the disease.2 Therefore, the medications given during hospitalisation should be provided.
Statistics strongly depend on the quality of data and comprehensiveness of its collection. How did the authors ensure that diagnosis of SARS-CoV-2, documentation of SARS-CoV-2 infection in the records, number of comorbidities, and outcome were correctly registered and reported?
Because the study discerned between confirmed, suspected and probable COVID-19, we should be able to know in how many patients the infection was confirmed by a positive RT-PCR, in how many by lung computerized tomography (CT), and in how many was COVID-19 diagnosed upon clinical suspicion alone. Lung CT can be suspicious of COVID-19 but does not definitively confirm the infection. The number of positive lung CT without a positive RT-PCR should be available and these cases should have been excluded from the evaluation.
Overall, this interesting study has limitations that question the results and their interpretation. Addressing these issues would strengthen the conclusions and could improve the status of the study. To assess mortality due to COVID-19 in patients with comorbidities, it is crucial to perform autopsy.
Conflict of interestThere is no conflict of interest.