Results
Data about surviving and not surviving patients’ characteristics, signs
and symptoms, PMH, physical examination, and outcomes are presented in
Table 1. Nearly 45% of patients were older than 65 years, 57.6% were
male. 16.6% of dead patients were severely obese, with a significant
difference between alive and dead patients (P-value<0.012).
Weakness was the most common sign. The most prevalent comorbidities were
HTN and DM (35.88% and 30.52%, respectively), and only 21% of
patients did not have any comorbidities. Since Shariati hospital is a
tertiary referral center, a remarkable portion of the patients
(approximately 20%) had degrees of immune deficiencies. Positive
COVID-19 RT-PCR was detected in 53.07% of our study population.
The not high-risk group (q-SOFA score: one) was associated with almost
four times greater odds of death (CI: 2.42-7.3,
P-value<0.001), and the high-risk group (q-SOFA score: two)
was associated with over ten times odds of death (CI: 2.98-36.04,
P-value<0.001). Four of the patients had a q-SOFA score of
three, whom all died during hospitalization. Systolic and diastolic
blood pressure were both significantly different between alive and dead
patients (P=0.001, and 0.024, respectively).
The average length of hospital stay was 5.98 days (SD:5.87). Twenty
hundred and sixty-five patients (74.8%) recovered. The average hospital
stay for recovered patients was 5.07 days (SD: 4.58), and the mean
number of days until death was 8.6 (SD:8.03). Of 354 patients, 64
patients (18.08%) did not require any oxygenation support. 117 patients
were admitted to ICU of which 82% were intubated during hospitalization
and 72% died. Twenty-eight patients (7.91%) received noninvasive
ventilation support.
Data about surviving and not surviving patients’ laboratory results and
lung CT scans are shown in Table 2. The most prevalent lung CT scan
finding was bilateral involvement and ground-glass opacities, with no
significant differences between surviving and not surviving patients.
Conversely, consolidation was significantly more common among
non-surviving patients. CT scan characteristics were not entered into
multivariate analysis since the p-value was borderline. Leukocytosis was
significantly more common among not surviving patients
(p<0.018). Also, NLR with a median of 4.23 (IQR; 2.8-7.45)
among surviving patients was significantly lower than non-surviving
patients with a median of 7.95 (IQR; 3.65-14.8) (p<0.001).
Although AST was significantly different between alive and dead patients
(p=0.002), indicators of liver function such as PTT and INR were not
significantly different between the groups.
Univariate and multivariate analysis of the association between age,
sex, anorexia, cancer, organ transplant, consciousness, vital signs, and
laboratory results and survival are given in Table 3.
Among patient’s characteristics, old age, impaired consciousness,
cancer, organ transplant, and BMI>35 were significantly
associated with higher mortality. Details are given in table 3. Impaired
consciousness increased odds of death around thirteen times (95%CI:
5.62-32.97, p<0.001), and organ transplant was associated with
almost eleven times higher risk of death (95%CI: 2.29-55.10, p=0.003).
The odds of death was 3.02 times higher in BMI above 35, and it reached
7.6 in the case of BMI exceeding 40 (p=0.016 and 0.0005, respectively).
Low SO2, high-grade fever, tachypnea, and tachycardia
were correlated with a higher death rate. Details are shown in table 3.
Additionally, not surviving patients had a higher frequency of SBP below
100 mmHg (OR: 2.68, 95%CI:1.18-6.09, p=0.019), and SBP above 140 mmHg
was associated with 0.45 times lower odds of death (95%CI:0.23-0.86,
p=0.016).
Among laboratory findings, anemia, leukocytosis, platelet count less
than 150,000 per microliter of blood, higher creatinine, and
pH<7.25 were significantly more frequent in non-surviving
patients. Details are presented in table 3. Finally, we chose the median
of CRP as the cutoff, and high CRP was correlated with greater mortality
(OR: 2.29, 95%CI: 1.9-5.69, p<0.001).
Given a large number of missing data, we did not enter BMI, and AST into
multivariate analysis. Also, we did not enter NLR, SBP, PCO2, and
bicarbonate into the multivariate analysis due to the correlation with
WBC count, DBP, and PH, respectively. The univariate analysis of the
variables which we did not enter into multivariate analysis because of
the reasons mentioned above is presented in the supplemental table.
The relationship between low consciousness, cancer, low
SO2, tachycardia, platelet < 150,000 per
microliter of blood, creatinine >1.2 mg/dL remained
statistically significant after multivariate analysis.