Comorbidities Associated with Complicated Hospital Course and Death in COVID-19 Patients: A Retrospective Study from Iran


 Objective There are limited data regarding the impact of comorbidities on hospitalized patients with coronavirus disease 2019 (COVID-19) in Iran. Methods We evaluated the risk of serious adverse outcomes in 1368 Iranian COVID-19 patients, admitted to five academic hospitals in Tehran between February-June 2020. The composite end-points were defined as admission to an intensive care unit, invasive ventilation, or death. The Cox proportional survival model determined the potential comorbidities associated with death. Results Overall, 576 patients (42.3%) reached the composite end-point (280 death). Adjusted for age, sex, duration of hospitalization, and the presence of the other comorbidities, patients with diabetes (RR=1.25, 95%CI; 1.08-1.44), heart failure (RR=1.45, 95%CI; 1.10-1.91), chronic kidney disease (RR=1.32, 95%CI; 1.04-1.67), malignancy (RR=1.79, 95%CI; 1.41-2.28), and lung diseases (RR=1.53, 95%CI; 1.27-1.84) were more likely to reach the composite end-point than those without the very comorbidity. Moreover, patients aged less than 65 years had a greater risk of death in the presence of two (HR=2.68, 95%CI; 1.46-4.95, p=0.002) or more (HR=3.47, 95%CI; 1.69-7.12, p=0.001) comorbidities, compared to those without any comorbidity. Conclusion To conclude, having two or more comorbidities in patients less than 65 years is associated with a greater risk of death during hospitalization.


Introduction
The novel coronavirus SARS-CoV-2, presenting with heterogeneous clinical manifestations [1][2][3], easily predisposes the susceptible patients to the respiratory failure and death [1,4,5]. Previous studies have demonstrated such underlying chronic disease as diabetes mellitus (DM), hypertension (HTN), cardiovascular disease (CVD), and chronic obstructive pulmonary disease (COPD) are more likely to complicate the clinical course of coronavirus disease 2019   [6,7]. The affected patients are at a greater risk of developing respiratory failure, admitting to an intensive care unit (ICU), and eventually death [3,8,9]. However, there are signi cant differences in population demographics, smoking rates, and prevalence of comorbidities among the countries [10,11]. Thus, comprehensive reports on early clinical outcomes of COVID-19 from involved patients of various ethnicities seem to be essential for identi cation of the sub-populations with poorer prognosis.
The rst con rmed case of COVID-19 in Iran was reported from Qom in February 2020 [12]. Soon after, Tehran and the other provinces reported outbreak, and now Iran ranked the tenth position globally [13]. Few studies investigating the association of some underlying chronic diseases and COVID-19 in Iran have certain limitations including the relatively small sample size and being single center observation [14]. On the other hand, the rate of infection in Tehran, with its high population density, has exceeded every other province [15]. This study evaluates the early clinical course of patients with COVID-19 hospitalized at several academic centers in Tehran, strati ed by the number and type of comorbidities.

Data source and data collection
This is a retrospective cohort study conducted at 5 academic hospitals a liated to Iran Universities of Medical Sciences (IUMS) in Iran. The study was approved by the ethics committee of the IUMS. Considering the minimal-risk research using data collected in routine clinical practice, the requirement for informed consent was waived. We enrolled all hospitalized patients with con rmed COVID-19 admitted between February 20, 2020 and June 13, 2020. Patients were considered to have con rmed COVID-19 if the polymerase chain reaction (PCR) assay of a nasopharyngeal specimen was positive. Those without laboratory documents for COVID-19 who presented the clinical manifestations of COVID-19 along with typical chest computed tomography (CT) ndings were also considered as having COVID-19. For patients with a readmission during the study period, data from the rst admission are presented.
Medical records of all admitted patients including demographic data, smoking status, vital signs at the time of admission, baseline comorbidities, chest CT ndings, as well as hospital course data including admission to ICU, treatment with invasive mechanical ventilation, duration of hospitalization, and patient status at the time of discharge (alive or dead) were collected.
All clinical pro les from ve academic hospitals were centrally merged. Some experienced technicians entered data into a computerized database for further double checking of all cases.
Comorbidities were determined based on patients' self-report on admission. A total of 8 comorbidities, namely DM, HTN, coronary artery disease (CAD), chronic kidney disease (CKD), heart failure (HF), lung diseases including asthma and COPD, cerebrovascular accidents (CVA), and malignancy, were included in the nal analysis. Comorbidities were initially considered as a categorical variable (yes versus no) and subsequently classi ed based on the number (single versus multiple).
Admission to ICU, treatment with invasive mechanical ventilation, and death were the primary outcomes of this study measured individually and as a composite outcome. All clinical outcomes are presented for patients who completed their hospital course (discharged alive or dead).

Statistical analysis
The continuous variables are presented as median (interquartile ranges (IQR)), and the categorical ones are presented as the number and (percentages).
To measure the impact of each of the eight mentioned comorbidities on the occurrence of either of the response variables namely death, ICU admission, invasive mechanical ventilation, and the composite outcome, the adjusted risk ratios were computed, applying the method suggested by Norton  Finally, a survival analysis was performed to estimate the probability of death from the date of hospital admission. For this purpose, the Kaplan-Meyer survival curve was drawn separately for the patients with 0, 1, 2, or ≥ 3 comorbidities, and the Log-rank test assessed the difference between these categories. Then, the cox proportional hazard survival models were tted at the hospitalization times to obtain the hazard ratios of death between those with and without any of the eight mentioned comorbidities. These models were adjusted for age, sex, and the other comorbidities. Statistical analyses were performed applying Stata software (version 13).The signi cant level was set as 0.05.

Survival analysis
We further investigated time-to-death during the hospital course separately for each type of comorbidity and the number of comorbidities. Kaplan-Meier survival curve demonstrates the probability of death during the hospital course in the patients with one, two, or more than two comorbidities ( Figure 1  The most prevalent comorbidities in this large cohort of COVID-19 patients were DM, HTN, and CAD. These comorbidities remained the most common underlying diseases in the studies with different sample sizes conducted in various populations [1,4,17]. Moreover, meta-analyses of several studies, although most of them had been conducted in china, reported the same ndings [18, 19)]. Considering the relatively large sample size, we investigated a spectrum of comorbidities including CKD, HF, CVA, COPD, asthma, chronic liver disease, HIV, and TB. Consistent with the results of previous studies [1,4,17], the percentage of patients with these comorbidities was relatively low. This observation could be explained by the facts that such cardiometabolic diseases as DM, HTN, and CAD are highly prevalent across the world, and they are more likely to be reported by the patients. Furthermore, the observed frequency of comorbidities might also be representative of the transmission of disease within the particular sub-groups.
Similar to the previous studies [17,20,21], our ndings suggested that comorbidities such as malignancy, lung disease (COPD and asthma), HF, CKD, and DM are associated with a complicated hospital course in patients with COVID-19. However, in this cohort patients with known HTN did not reach the composite end-point. While HTN was associated with a signi cant greater risk of the same composite end-point in a large cohort of Chinese patients with COVID-19 [17]. This discrepancy might be explained by the fact that no adjustment for other comorbidities was made in this study while we considered all the other co-existence underlying disease such as malignancy, lung disease, HF, CKD, and DM that could confound the results.
Moreover, patients in our cohort had well-controlled blood pressure (BP) at the time of admission (with median of 120/80 mmHg Although the coexistence of different underlying diseases has been frequently reported in the cohorts of COVID-19-patients [4,21,27], a few investigated the impact of coexisting comorbidities on the clinical outcome of COVID-19 [17]. Our study con rmed the coexistence of two or more comorbidities incrementally increased the hazard of the death. However, there was no signi cant difference in the prognosis of those with one comorbidity compared to the ones without any comorbidity. Considering the strong independent role of age in the complicated hospital course and in-hospital death [14,25], we performed a sensitivity analysis stratifying the patients according to the age (< 65 vs ≥ 65 years). We found the adverse impact of the coexisting comorbidities on death is considerable in patients less than 65 years. This nding provides some evidence in that fatality of COVID-19 patients more than 65 years is independent of the presence of any comorbidity.

Strengths and limitations
To the best of our knowledge, it was the rst multicenter cohort study on Iranian people with COVID-19 exploring the association between a spectrum of comorbidities and a complicated hospital course. Moreover, data on comorbidities were available for more than 90% of the admitted patients. However, this study has several limitations. First, due to the urgency of data extraction, random sampling could not be applied in our study. Second, comorbidities were self-reporting. Underreporting of comorbidities due to the lack of awareness or diagnostic testing might confound the strength of their association with the clinical outcomes.
In this cohort of Iranian people with COVID-19, the presence of malignancy, lung disease, HF, CKD, and DM were predictors of a complicated hospital course. DM and CKD were also risk factors of in-hospital mortality. Moreover, there is an incremental risk of in-patient mortality with increased burden of comorbidity in hospitalized patients less than 65 years old.  Chronic kidney disease. Figure 3