The use of angiotensin-converting enzyme inhibitors in hospitalized patients with COVID-19 is associated with a lower risk of mortality Authors: Mykola Khalangot, Vitaliy Gurianov, Tamara Zakharchenko, Victor Kravchenko, Olena Kovzun, Volodymyr Shupachynskiy, Mykola Tronko Published: 2025 ABSTRACT The relationship between the use of angiotensin-converting enzyme inhibitors (ACEIs) and angiotensin II receptor blockers (ARBs), diabetes mellitus (DM), and the risk of death in patients with COVID-19 remains controversial. We hypothesized that DM and certain characteristics of the COVID-19 course during hospital treatment may alter the assessment of the association of ACEIs/ARBs with COVID-19 outcomes. The records of 153 COVID-19 inpatients admitted to a municipal clinic in Kyiv, Ukraine, between October and December 2021 were reviewed. To assess the association of ARBs/ACEIs and other hypotensive drugs with outcomes, a no hypotensives group was used for comparison. A multivariable logistic regression model was employed to assess the odds ratio (OR) of death. If DM was known at the time of hospitalization (n = 28), there was a higher proportion of deaths compared to the group without DM (n = 125): 53.6% vs. 12.8%, p < 0.001. After adjusting for age, minimal O2 saturation, DM, and antihypertensive therapy, the ACEIs-associated OR was 0.10 (0.02–0.69). The DM-associated OR was 8.25 (1.92–35.42). The use of ACEIs in the treatment of COVID-19 inpatients is associated with a lower risk of mortality compared to those not using hypotensive treatment, regardless of the presence of DM.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-025-29757-x. Introduction The end of the SARS-CoV-2 pandemic has not ended the state of uncertainty regarding the effect of angiotensin-converting enzyme inhibitors (ACEIs) and angiotensin II receptor blockers (ARBs) on the risk of mortality in patients with COVID-19: an observational study conducted in China1 based on data from several hospitals found that ARB use was associated with lower all-cause mortality (adjusted HR, 0.53; 95% CI, 0.38–0.73; p < 0.001), while ACEI treatment did not affect mortality. However, in a similar study in Japan, ARBs did not significantly reduce in-hospital mortality over ACEIs in patients admitted for COVID-19, (adjusted OR, 0.95; 95% CI 0.69–1.3)2. Furthermore, the conclusions drawn by different reviews are markedly diverse. Xu J. et al.3 assert that “there has been no evidence for initiating an ACEI/ARB regimen as prevention or treatment of COVID-19.” Conversely, Baral R. et al.4, in their systematic review and meta-analysis of 52 studies involving 101,949 patients with COVID-19, identified a significantly lower risk of multivariable-adjusted mortality and severe adverse events among patients who received ACEIs or ARBs compared to those who did not. These authors suggest that ACEIs and ARBs may offer protective benefits, advocating for future randomized clinical trials to establish causality. However, it is worth noting that the study by Baral R. et al. did not distinguish between the separate association of ACEIs and ARBs with COVID-19 mortality. It is also concerning that in the meta-analysis by Xu J. et al., the criteria for excluding 4,415 out of 4,969 records were not clearly explained. The RCT5 conducted in Indian and Australian patients with mild COVID-19 who were randomly assigned to receive a low-dose ARB (40 mg/day telmisartan) failed to demonstrate a benefit of experimental treatment for COVID-19. The impact of diabetes both on the association of ACEIs/ARBs treatment in the context of COVID-19 mortality and on COVID-19 outcomes unrelated to ACEIs/ARBs remains unclear. Study in Ukraine found a very significant reduction in the risk of death for hospitalized patients with COVID-19 without DM, who received ACEIs/ARBs compared with the corresponding risk of normotensive COVID-19 inpatients who did not receive antihypertensive treatment: OR 0.22 (95% CI 0.07–0.72) adjusted for age, gender and fasting plasma glucose6. A recent retrospective multicentre European study7 revealed no association between mortality and renin-angiotensin-aldosterone system inhibitor therapy in adults with diabetes admitted to hospital with COVID-19. In the last 2 studies, there was no separate assessment of the association of ACEIs or ARBs treatment on the outcomes of COVID-19. The assessment of the relationship between DM and the risk of death in patients with COVID-19 is still controversial. On the one hand, a large population study in England confirmed the presence of such positive relationship: the odds ratios (ORs) for in-hospital COVID-19-related death were 3·51 (95% CI 3·16–3·90) in people with type 1 diabetes and 2·03 (1·97–2·09) in people with type 2 diabetes8. At the same time, another single-center study in England found no difference in mortalities based on the diabetes status9, a similar result was reported by the authors in Morocco10 and Ghana11. There is still no obvious explanation for the controversial association with DM and mortality from COVID-19. We suggest that it is advisable to change the approaches to assessing the severity of patients, which may change during hospital treatment. This may refer to oxygen saturation, glycemia, and other indicators. We also hypothesized that such approaches to characterization of the disease may change the assessment of the effect of ACEI/ARBs on the outcomes of COVID-19 in patients with DM. Thus, our aim was to evaluate the final outcomes of treatment of COVID-19 patients requiring hospitalization, taking into account the presence of DM before hospitalization and the use of either ACEIs or ARBs during treatment. Methods Study design and setting The study was conducted by analyzing the archives of one of the multidisciplinary city hospital’s: Municipal Clinic – (MC), Kyiv, Ukraine. From October 2021 patients with clinical symptoms similar to Covid-19 were hospitalized in cardiologic and intensive care departments of this Hospital. Diagnosis of Covid-19, selection of patients for hospitalization, clinical examinations and treatment were performed according to relevant national standards12, which were updated according to WHO recommendations. Detailed criteria for hospitalization were presented earlier6. A retrospective, observational study was carried out at the MC. The medical files of 153 patients, with a definite in-hospital outcome, admitted at MC between the periods of October 1 and December 31, 2021, were consecutively reviewed and the relevant data extracted. Multiple logistic regression was used to assess the effects of ACEIs on COVID-19 outcomes. Minimal sample size was estimated using G*Power v.3.1.9.713. Calculations were made on the assumption α = 0.05, Power = 0.8, probability of outcome = 0.3 (null hypothesis). The available sample size (n = 153) allows us to determine the presence of a strong effect (OR ≤ 0.2 see Supplement Table 1A, B). Patients’ characteristics The diagnosis of Covid-19 was confirmed by polymerase chain reaction (PCR test, n = 145) or by detection of specific antibodies (IgM, n = 7), or specific antigen (express method, n = 1). One patient with severe symptoms of COVID was included in the analysis despite lack of laboratory confirmation of SARS-CoV-2. Information on some clinical and anthropometric characteristics of patients: fasting plasma glucose (FPG), peripheral capillary oxygen saturation (SpO2); c-reactive protein (CRP), procalcitonin, D-dimer, sodium levels, body mass index (BMI); systolic and diastolic blood pressure (BP) level was extracted from patients’ paper documents. We used the highest (maximum) levels of BP, body temperature, CRP, D-dimer, Procalcitonin and the lowest (minimum) levels of SpO2 recorded during the hospital stay are presented. Regarding FPG and Sodium, we present both their maximum and minimum levels. This also applies to glucose-adjusted14 sodium levels. Each patient’s treatment data were extracted from the paper files by two members of our research team and then checked and confirmed by a third person. Similar to our recent study of the association between the use of certain antihypertensive drugs and COVID-19 outcomes6 performed in Donetsk region of Ukraine, we categorized only previously known cases as DM. When evaluating COVID-19 outcomes depending on the use of certain types of hypotensive treatment, we used the results of patients who did not need hypotensive treatment as a reference group. However, in Kyiv study we were able to present more detailed treatment categories, i.e., we separated the results of ACEIs treatment from the results of ARBs. Leaving the hospital due to recovery or death was considered as a dichotomous treatment outcome. 10 patients were transferred from MC to other city hospitals, their outcomes were also identified and taken into account in this study. Statistical analysis has been performed using MedCalc® Statistical Software version 22.009 (MedCalc Software Ltd, Ostend, Belgium; https://www.medcalc.org; 2023). For the quantitative data mean and standard deviation (± SD) for normal distribution or median (Me) and interquartile range (QI – QIII) for non-normal distribution were presented. Comparisons of two groups were performed by t-test for normal distribution or Mann–Whitney test for non-normal distribution. For qualitative data the proportion (%) was presented, comparisons of two groups were performed by Fisher’s exact test. For the analysis association of the dependent variable with risk factors logistic regression analysis was used. To estimate the diagnostic efficiency of the logistic regression models area under the ROC curve (AUC) and its 95% CI were calculated. To estimate the effect of risk factors on the dependent variable the odds ratios (OR) and its 95% CI were calculated. In all of the tests the p value < 0.05 was considered significant. Results In a group of patients with COVID-19, where DM was known at the time of hospitalization (n = 28, Table 1), there was a significantly higher proportion of deaths than in the group without information on the presence of DM (n = 125): 53.6% vs. 12.8%, p < 0.001. Both groups were equally dominated by women. A lower level of minimal O2 saturation, a higher level of maximal Systolic BP, FPG, CRP, D-Dimer and Sodium corrected were noted in the DM group. These groups did not differ in terms of gender, BMI, myocardial infarction, stroke history, or maximum body temperature. The median ages in the group without DM and with DM did not differ significantly (p = 0.079) and were equal to 70.0 and 73.25 years, respectively. Table 1Anthropometric and laboratory characteristics and outcomes of hospitalized patients with COVID-19 depending on the presence of a history of diabetes mellitus (DM).VariablesWithout DM (n = 125)DM (n = 28) p Age, years70 (61.5–76.45)73.25 (67.250–81.15)0.079women, n (%)79 (63.2)17 (60.7)0.831Height, cm165 (162.250–170)165 (162.250–175)0.659Weight, kg80 (70–90)82.5 (72–96.5)0.279BMI, kg/m229.1 (25.325–31.6)28.4 (26.550–34.6)0.341Systolic BP, maximal, mmHg140 (130–160)170 (140–180) 0.003 Diastolic BP, mmHg90 (80–90)90 (80–107.5)0.076Myocardial infarction history, n (%)13 (11.0)5 (19.2)0.322Stroke history, n (%)7 (6.0)0 (0.0)0.347O2 saturation, minimal, %88 (79.5–92)78 (65.5–90.5) 0.006 Temperature, maximal, ° C38.09 (0.89)38.11 (0.82)0.882FPG maximal, mmol/L6.8 (5.8–8.525)15.8 (11.6–22.850) < 0.001 FPG minimal, mmol/L4.8 (3.9–5.7)4.7 (3.9–6.750)0.29CRP, maximal, mg/L50.2 (11.650–106)110.3 (49.750–126.250) 0.009 Procalcitonin, maximal, ng/ml0.1 (0.1–0.2)0.23 (0.075–0.525)0.227D-Dimer, maximal, mg/L0.64 (0.320–1.450)0.84 (0.750–3.907) 0.047 Sodium minimal, mmol/L141 (139–144)141.5 (137.5–142)0.256Sodium maximal, mmol/L143 (141–145)143.5 (142–145.5)0.175Sodium minimal, corrected mmol/L141 (139–143)141 (138–142)0.3Sodium maximal corrected, mmol/L144 (141–146)148 (145.5–152) < 0.001 Duration of treatment, days13 (9–18)15 (10–20)0.471In hospital mortality, n (%)16 (12.8)15 (53.6) < 0.001 Data are presented by medians and inter quartile ranges (QI-QIII) or ± SD. The comparison was based on the Mann-Whitney test or t-test. For qualitative variables frequency are presented, Fisher’s exact test was used.BMI body mass index, O2 saturation oxygen saturation, FPG fasting plasma glucose, CRP C-reactive protein.Significant values are in bold. Table 1Anthropometric and laboratory characteristics and outcomes of hospitalized patients with COVID-19 depending on the presence of a history of diabetes mellitus (DM).VariablesWithout DM (n = 125)DM (n = 28) p Age, years70 (61.5–76.45)73.25 (67.250–81.15)0.079women, n (%)79 (63.2)17 (60.7)0.831Height, cm165 (162.250–170)165 (162.250–175)0.659Weight, kg80 (70–90)82.5 (72–96.5)0.279BMI, kg/m229.1 (25.325–31.6)28.4 (26.550–34.6)0.341Systolic BP, maximal, mmHg140 (130–160)170 (140–180) 0.003 Diastolic BP, mmHg90 (80–90)90 (80–107.5)0.076Myocardial infarction history, n (%)13 (11.0)5 (19.2)0.322Stroke history, n (%)7 (6.0)0 (0.0)0.347O2 saturation, minimal, %88 (79.5–92)78 (65.5–90.5) 0.006 Temperature, maximal, ° C38.09 (0.89)38.11 (0.82)0.882FPG maximal, mmol/L6.8 (5.8–8.525)15.8 (11.6–22.850) < 0.001 FPG minimal, mmol/L4.8 (3.9–5.7)4.7 (3.9–6.750)0.29CRP, maximal, mg/L50.2 (11.650–106)110.3 (49.750–126.250) 0.009 Procalcitonin, maximal, ng/ml0.1 (0.1–0.2)0.23 (0.075–0.525)0.227D-Dimer, maximal, mg/L0.64 (0.320–1.450)0.84 (0.750–3.907) 0.047 Sodium minimal, mmol/L141 (139–144)141.5 (137.5–142)0.256Sodium maximal, mmol/L143 (141–145)143.5 (142–145.5)0.175Sodium minimal, corrected mmol/L141 (139–143)141 (138–142)0.3Sodium maximal corrected, mmol/L144 (141–146)148 (145.5–152) < 0.001 Duration of treatment, days13 (9–18)15 (10–20)0.471In hospital mortality, n (%)16 (12.8)15 (53.6) < 0.001 Data are presented by medians and inter quartile ranges (QI-QIII) or ± SD. The comparison was based on the Mann-Whitney test or t-test. For qualitative variables frequency are presented, Fisher’s exact test was used.BMI body mass index, O2 saturation oxygen saturation, FPG fasting plasma glucose, CRP C-reactive protein.Significant values are in bold. Anthropometric and laboratory characteristics and outcomes of hospitalized patients with COVID-19 depending on the presence of a history of diabetes mellitus (DM). Data are presented by medians and inter quartile ranges (QI-QIII) or ± SD. The comparison was based on the Mann-Whitney test or t-test. For qualitative variables frequency are presented, Fisher’s exact test was used. BMI body mass index, O2 saturation oxygen saturation, FPG fasting plasma glucose, CRP C-reactive protein. Significant values are in bold. Table 2Use of some types of treatment of COVID-19 inpatients depending on the history of diabetes.VariablesWithout DMn = 125 (100)DMn = 28 (100) p Corticosteroids Dexamethasone79 (63.2)17 (60.7)0.388 No steroids13 (10.4)1 (3.6) Methylprednisolone33 (26.4)10 (35.7)Antibiotics Intravenously86 (72.9)19 (70.4)0.741 No antibiotics24 (20.3)5 (18.5) Orally8 (6.8)3 (11.1)Insulin No120 (96.0)14 (50.0) < 0.001 Yes5 (4.0)14 (50.0)Antihypertensives No antihypertensives39 (32.5)1 (3.6) 0.009 ARBs19 (15.8)9 (32.1) ACEIs37 (30.8)9 (32.1) Other’s25 (20.8)9 (32.1)Other antidiabetes treatments No121 (96.8)15 (53.6) < 0.001 Yes4 (3.2)13 (46.4)Mechanical ventilation (any type) No112 (89.6)18 (64.3) 0.002 Yes13 (10.4)10 (35.7)Invasive mechanical ventilation No114 (91.2)18 (64.3) 0.001 Yes11 (8.8)10 (35.7)Noninvasive mechanical ventilation No119 (95.2)22 (78.6) 0.009 Yes6 (4.8)6 (21.4)Significant values are in bold.Percentages are given in parentheses.DM diabetes mellitus, ARBs angiotensin II receptor blockers, ACEIs angiotensin-converting-enzyme inhibitors. Table 2Use of some types of treatment of COVID-19 inpatients depending on the history of diabetes.VariablesWithout DMn = 125 (100)DMn = 28 (100) p Corticosteroids Dexamethasone79 (63.2)17 (60.7)0.388 No steroids13 (10.4)1 (3.6) Methylprednisolone33 (26.4)10 (35.7)Antibiotics Intravenously86 (72.9)19 (70.4)0.741 No antibiotics24 (20.3)5 (18.5) Orally8 (6.8)3 (11.1)Insulin No120 (96.0)14 (50.0) < 0.001 Yes5 (4.0)14 (50.0)Antihypertensives No antihypertensives39 (32.5)1 (3.6) 0.009 ARBs19 (15.8)9 (32.1) ACEIs37 (30.8)9 (32.1) Other’s25 (20.8)9 (32.1)Other antidiabetes treatments No121 (96.8)15 (53.6) < 0.001 Yes4 (3.2)13 (46.4)Mechanical ventilation (any type) No112 (89.6)18 (64.3) 0.002 Yes13 (10.4)10 (35.7)Invasive mechanical ventilation No114 (91.2)18 (64.3) 0.001 Yes11 (8.8)10 (35.7)Noninvasive mechanical ventilation No119 (95.2)22 (78.6) 0.009 Yes6 (4.8)6 (21.4)Significant values are in bold.Percentages are given in parentheses.DM diabetes mellitus, ARBs angiotensin II receptor blockers, ACEIs angiotensin-converting-enzyme inhibitors. Use of some types of treatment of COVID-19 inpatients depending on the history of diabetes. Significant values are in bold. Percentages are given in parentheses. DM diabetes mellitus, ARBs angiotensin II receptor blockers, ACEIs angiotensin-converting-enzyme inhibitors. The treatment of patients in the groups without a history and with a history of diabetes did not differ with regard to the use of corticosteroids and antibiotics, while the use and structure of antihypertensive medications differed significantly between these groups (Table 2). More than a third of patients without a history of DM did not need treatment for arterial hypertension, while in the group with diabetes this situation applied to only one patient (3.6%). The proportion of ARB treatment was 15.8% and 32.1% for the non-DM and DM groups, respectively. A small proportion of patients without a history of DM were treated with insulin (4%) or other antidiabetic agents because of hyperglycemia detected during treatment for COVID-19. The use of mechanical lung ventilation in patients with DM occurred three times more often (35.7%) than in the group without DM (10.4%): RR = 3.4 (95% CI 1.7–7.0). A similar ratio in the groups of COVID-19 inpatients without diabetes and with diabetes occurred regarding the share of use of both invasive and non-invasive mechanical ventilation (Table 2). Table 3Risks of death of COVID-19 inpatients (n = 153) estimated using univariable logistic regression models.FactorsModel coefficient b ± m p OR (95% CI)Age, years0.043 ± 0.019 0.02 1.04 (1.00–1.08) Gender WomenReference Men0.08 ± 0.410.851–Myocardial infarction history− 0.27 ± 0.670.687–Stroke history0.48 ± 0.860.579–Diabetes mellitus history2.06 ± 0.46 < 0.001 7.86 (3.17–19.52) BMI, kg/m20.089 ± 0.039 0.022 1.09 (1.01–1.18) Systolic BP maximal, mmHg0.033 ± 0.009 < 0.001 1.03 (1.02–1.05) Diastolic BP, mmHg0.036 ± 0.016 0.024 1.04 (1.00–1.07) Temperature, maximal, °C0.23 ± 0.230.327–O2 Saturation minimal, %− 0.15 ± 0.03 < 0.001 0.86 (0.81–0.91) FPG maximal, mmol/L0.12 ± 0.03 < 0.001 1.12 (1.06–1.19) Procalcitonin maximal, ng/ml0.039 ± 0.0290.1881.04 (0.98–1.10)CRP maximal, mg/L0.0091 ± 0.0045 0.046 1.01 (1.00–1.02) D-Dimer maximal, mg/L0.20 ± 0.10 0.042 1.22 (1.01–1.49) Sodium minimal, mmol/L− 0.075 ± 0.0420.0730.93 (0.86–1.01)Sodium maximal, mmol/L0.12 ± 0.05 0.029 1.12 (1.01–1.25) Sodium minimal, corrected mmol/L− 0.064 ± 0.0410.116–Sodium maximal corrected, mmol/L0.17 ± 0.05 < 0.001 1.19 (1.087–1.30) Insulin1.79 ± 0.52 0.001 5.98 (2.17–16.48) Other anti-diabetes treatments1.76 ± 0.54 0.001 5.83 (2.03–16.76) Corticosteroids− 19.5 ± 6300> 0.05–Antibiotics IntravenouslyReference None− 1.39 ± 0.770.0690.25 (0.06–1.12) Orally0.23 ± 0.710.750–Antihypertensives NoneReference ACEIs− 0.72 ± 0.600.234– Other’s0.36 ± 0.540.498– ARBs0.47 ± 0.560.402–Significant values are in bold.ARBs angiotensin II receptor blockers, ACEIs angiotensin-converting-enzyme inhibitors, BMI body mass index, O2 saturation oxygen saturation, FPG fasting plasma glucose, CRP C-reactive protein. Table 3Risks of death of COVID-19 inpatients (n = 153) estimated using univariable logistic regression models.FactorsModel coefficient b ± m p OR (95% CI)Age, years0.043 ± 0.019 0.02 1.04 (1.00–1.08) Gender WomenReference Men0.08 ± 0.410.851–Myocardial infarction history− 0.27 ± 0.670.687–Stroke history0.48 ± 0.860.579–Diabetes mellitus history2.06 ± 0.46 < 0.001 7.86 (3.17–19.52) BMI, kg/m20.089 ± 0.039 0.022 1.09 (1.01–1.18) Systolic BP maximal, mmHg0.033 ± 0.009 < 0.001 1.03 (1.02–1.05) Diastolic BP, mmHg0.036 ± 0.016 0.024 1.04 (1.00–1.07) Temperature, maximal, °C0.23 ± 0.230.327–O2 Saturation minimal, %− 0.15 ± 0.03 < 0.001 0.86 (0.81–0.91) FPG maximal, mmol/L0.12 ± 0.03 < 0.001 1.12 (1.06–1.19) Procalcitonin maximal, ng/ml0.039 ± 0.0290.1881.04 (0.98–1.10)CRP maximal, mg/L0.0091 ± 0.0045 0.046 1.01 (1.00–1.02) D-Dimer maximal, mg/L0.20 ± 0.10 0.042 1.22 (1.01–1.49) Sodium minimal, mmol/L− 0.075 ± 0.0420.0730.93 (0.86–1.01)Sodium maximal, mmol/L0.12 ± 0.05 0.029 1.12 (1.01–1.25) Sodium minimal, corrected mmol/L− 0.064 ± 0.0410.116–Sodium maximal corrected, mmol/L0.17 ± 0.05 < 0.001 1.19 (1.087–1.30) Insulin1.79 ± 0.52 0.001 5.98 (2.17–16.48) Other anti-diabetes treatments1.76 ± 0.54 0.001 5.83 (2.03–16.76) Corticosteroids− 19.5 ± 6300> 0.05–Antibiotics IntravenouslyReference None− 1.39 ± 0.770.0690.25 (0.06–1.12) Orally0.23 ± 0.710.750–Antihypertensives NoneReference ACEIs− 0.72 ± 0.600.234– Other’s0.36 ± 0.540.498– ARBs0.47 ± 0.560.402–Significant values are in bold.ARBs angiotensin II receptor blockers, ACEIs angiotensin-converting-enzyme inhibitors, BMI body mass index, O2 saturation oxygen saturation, FPG fasting plasma glucose, CRP C-reactive protein. Risks of death of COVID-19 inpatients (n = 153) estimated using univariable logistic regression models. Significant values are in bold. ARBs angiotensin II receptor blockers, ACEIs angiotensin-converting-enzyme inhibitors, BMI body mass index, O2 saturation oxygen saturation, FPG fasting plasma glucose, CRP C-reactive protein. Univariable logistic regression modeling (Table 3) indicates a positive relationship between age, history of diabetes, blood pressure, BMI, maximal levels of FPG, Sodium, Sodium corrected, CRP, D-Dimer and the risk of dying for COVID-19 inpatients. Minimal O2 saturation negatively and expectedly is associated with the risk of dying. Assessment of treatment categories reveals a positive association with the odds of dying only for insulin and other anti-diabetes treatments (Table 3). Table 4Analysis of a multivariable logistic regression model for predicting the risk of death for inpatients with COVID-19.FactorsModel coefficient b ± m p OR (95% CI)Age, years0.057 ± 0.027 0.033 1.06 (1.00–1.12) O2 saturation minimal, %− 0.17 ± 0.03 < 0.001 0.84 (0.79–0.90) Diabetes mellitus history2.11 ± 0.74 0.005 8.25 (1.92–35.42) Antihypertensives NoneReference ACEIs− 2.26 ± 0.97 0.019 0.10 (0.02–0.69) Other’s− 0.17 ± 0.780.826– ARBs− 0.65 ± 0.960.499–ARBs angiotensin II receptor blockers, ACEIs angiotensin-converting-enzyme inhibitors, O2 saturation oxygen saturation.Significant values are in bold. Table 4Analysis of a multivariable logistic regression model for predicting the risk of death for inpatients with COVID-19.FactorsModel coefficient b ± m p OR (95% CI)Age, years0.057 ± 0.027 0.033 1.06 (1.00–1.12) O2 saturation minimal, %− 0.17 ± 0.03 < 0.001 0.84 (0.79–0.90) Diabetes mellitus history2.11 ± 0.74 0.005 8.25 (1.92–35.42) Antihypertensives NoneReference ACEIs− 2.26 ± 0.97 0.019 0.10 (0.02–0.69) Other’s− 0.17 ± 0.780.826– ARBs− 0.65 ± 0.960.499–ARBs angiotensin II receptor blockers, ACEIs angiotensin-converting-enzyme inhibitors, O2 saturation oxygen saturation.Significant values are in bold. Analysis of a multivariable logistic regression model for predicting the risk of death for inpatients with COVID-19. ARBs angiotensin II receptor blockers, ACEIs angiotensin-converting-enzyme inhibitors, O2 saturation oxygen saturation. Significant values are in bold. Multivariate logistic regression analysis was performed for evaluation the risk of death. A stepwise method was used (entering a variable if p < 0.05 and removing a variable if p > 0.1). Selected significant characteristics in the multivariate model: Age, Antihypertensives, O2 Saturation minimal, DM history (Table 4). Area under ROC curve of the model AUC = 0.94 (95% CI 0.89–0.97) which indicates a strong association of selected variables with the risk of death (Nagelkerke R2 = 0.62, p < 0.001). Regardless of age, O2 Saturation, and Diabetes mellitus history, treatment of inpatients with COVID-19 with ACEIs was associated with a significant reduction in the risk of death compared to no antihypertensive treatment: OR (95% CI) 0.10 (0.02–0.69). However, the estimated OR (95% CI) associated with diabetes mellitus history within this model was very high: 8.25 (1.92–35.4). The addition of other variables, the risk of which was significant when assessed using univariable logistic regression models (Table 3) in the multivariable model led to the loss of their significance and did not lead to an improvement in the prognostic characteristics of the model, which indicates their association with the already applied variables. Since two variables in this multivariate model were continuous and the other four categorical, we ran a similar model that included only categorical variables: age > 70 years and O2 saturation minimal ≤ 82% (cut-offs based on median level and better Joden index, respectively). This model demonstrated the same odds directions: higher mortality risk for DM and lower mortality risk for ACEIs (Figure S1). The AUC of this model was 0.937 (95% CI 0.86–0.97), which was very close to the previous model. Additionally, we conducted an analysis using a multivariable logistic regression model to predict the risk of death exclusively for inpatients with COVID-19 who tested positive via PCR, adding the maximal systolic BP variable. Excluding several patients without PCR confirmation and adding this variable did not significantly alter the association between ACEI treatment and the risk of COVID-19-related death (Table S2). The area under the ROC curve of the model, AUC = 0.95 (95% CI 0.89–0.98), indicates a strong association of the selected variables with the risk of death. Discussion In our relatively small (n = 153) sample of older (median age 70 yrs) COVID-19 inpatients with a significant proportion of patients with DM history (18.3%), a rather large proportion of deaths (20.26%) was found. Recent studies from China report mortality in an age-matched population of 5.2% and 8.6% (ARB- and ACEI -treated COVID-19 inpatients respectively)1, similar data from Japan were 5.2% and 7.7%2. However, in 2020, one of the hospitals in England reported a mortality rate of 38.4% for all hospitalized patients with COVID-19 (n = 232), and 46% for patients with diabetes9. Most likely, these discrepancies indicate different criteria or practice of their application for hospitalization of COVID-19 in different countries or clinical centers. Our analysis revealed that the majority (53.6%) of hospitalized patients with DM history unfortunately died, while for patients without pre-existing DM the mortality was 12.8%. Diabetes-related risk of death estimated using a one-way logistic regression model, OR (95% CI) was 7.86 (3.17–19.52). A similar regression analysis indicated a statistically significant mortality risk associated with increasing age and BMI (Table 3), that is, it is impossible not to consider at least these factors when assessing the contribution of DM to the mortality risk of patients with COVID-19. To a large extent, this may also apply to systolic and diastolic maximal BP and indicators reflecting the course of DM and COVID-19 (minimal SpO2, maximum levels of FPG, Sodium, Sodium corrected, CRP, D-Dimer) that have revealed a connection with the risk of death in patients with COVID-19. A peculiarity of this study is that we used the ‘worst’ clinical indicators recorded in each patient’s medical documents as variables. Thus, when evaluating COVID-19 death risks, we evaluated SpO2, FPG, sodium and some cytokines like minimal and/or maximal levels, instead of their levels at the time of admission, which was used by most other studies. Regarding FPG (but not SpO2), an approach similar to ours was used by Italian researchers15. According to our assessment, the higher the minimum SpO2 level for a particular patient, the lower the risk of dying, which is the expected result (Table 3). Logistic regression analysis was applied to evaluate the effects of the independent variables on the risk of death. Stepwise method was used to find the best fitting model which described the relationship between the risk of death and a set of independent variables in a multifactor logistic regression model. As a result of this analysis, we can conclude that regardless of age, minimum SpO2 and history of DM, treatment of inpatients with COVID-19 with ACE inhibitors was associated with a significant reduction in the risk of death compared to no antihypertensive treatment: OR (95% CI) 0.10 (0.02–0.69). However, the estimated risk associated with a history of diabetes in this model was very high: OR (95% CI) 8.25 (1.92–35.42). It should be noted that due to the use of such a statistical model, we were able to show a relationship between the use of ACEIs and a decrease in the chances of dying for patients with COVID-19, while this relationship was not detected during univariate analysis (Table 3). To our knowledge, a similar analytical approach has not previously been applied to assess COVID-19 outcomes. Researchers from Morocco reported an increase in CRP and D-Dimer levels in COVID-19 inpatients with DM, but unlike us, they did not find a significant difference in the mortality of COVID-19 inpatients depending on DM. To clarify the relationship between COVID-19 mortality and DM, they suggest investigating the interaction of medications with ACE-2 receptors (that is, the use of ARBs and ACEIs in treatment) and the role of co-morbidities10. Resent multivariable logistic regression mortality analysis of COVID-19 patients with type 2 diabetes mellitus in Istanbul (Turkey) confirmed an independent mortality risk related with increasing age, male gender, obesity, lymphopenia and insulin treatment, but did not reveal such a relationship with ARBs / ACEIs treatment, high HbA1c and some comorbidities16. It should be noted that the study in Istanbul concerned the entire population, and not only those patients who, due to the severity of their condition, required hospitalization. A large retrospective study in China has just presented an analysis of adjusted hazard ratios for risk factors associated with in-hospital all-cause mortality in the multivariate Cox regression model. According to this model, treatment of hypertension with ARBs (but not ACEIs) was found to be associated with a significant reduction in the risk of death for patients hospitalized for COVID-191: “After adjusting for age, gender, comorbidities, and in-hospital medications (such as metformin and statins), the use of ARBs was associated with lower all-cause mortality (adjusted HR 0.54; 95% CI, 0.39–0.74; p < 0.001). Besides, the adjusted HRs for association of other variables for all-cause mortality were 1.91 (95% CI, 1.63–2.23), 1.53 (95% CI, 1.35–1.75), 0.59 (95% CI, 0.47–0.73), 0.51(95% CI, 0.31–0.85), and 1.63 (95% CI, 1.43–1.86) for age, gender, statins, metformin, and hypercholesterolemia, respectively”. In our opinion, an interesting result of this study, which its authors for some reason did not pay enough attention to, is the assessment of DM and its treatment with metformin as a protective factor for mortality in COVID-19 inpatients, while insulin treatment is associated with a significant increase in the risk of death in these patients. These results are consistent with our data on the risk associated with the treatment of patients with type 2 diabetes with insulin: OR 5.98 (2.17–16.48) using one-way logistic regression model. Risk connected with DM history: OR 10.14 (3.45–29.83) using a multifactor (age and antihypertensive treatment) logistic regression model for predicting the risk of death for inpatients with COVID-19. If the high insulin-related risk of mortality in patients with COVID-19 is confirmed by studies from England17 and China18, the protective effect of DM shown by authors from China1 is not at all an expected result. A study from England is rather coincident with our inpatients-based study: the odds ratios (ORs) for in-hospital COVID-19-related death were 2.03 (1.97–2.09) in people with type 2 diabetes8. Using a multivariable logistic regression model (age, SpO2, DM history, stratified antihypertensive therapy) to predict the risk of death for hospitalized patients with COVID-19 found ACEI-associated ORs of 0.10 (0.02–0.69) or 0.18 (0.04–0.70) depending on SpO2 on/off. Our estimate of the association of treatment with ARBs or ACEIs with the risk of death in inpatients with COVID-19 does not agree with corresponding estimates recently made in Japan2 and China1. Using a logistic model of analysis, Japanese authors found that among hypertensive patients hospitalized for COVID-19, ARB was associated with lower crude rate of in-hospital mortality, but not in multivariate-adjusted analysis2. The described sample of data from Japan is unique in that there is no use of non-invasive mechanical ventilation and no data on antibiotics. The proportion of deaths was at least 2 times lower than in the Ukrainian sample of hospitalized patients, which may indicate less strict criteria for hospitalization and partially explain the different results of the studies. A study from China1 reported almost the same mortality of hospitalized patients as in Japan, but ARB administration was significantly associated with a lower risk of all-cause mortality in patients with COVID-19. This conclusion was made by multivariate Cox regression analyzed and Kaplan–Meier curves. The use of Cox regression in this case may have flaws that could affect the result. The use of Cox proportional-hazards regression method to analyze the influence of factorial features on mortality risk, in this case may lead to estimate bias (the best method for analyzing these data would be using logistical regression models). Italian researchers (Piarulli, F., et al. 2023) reported mortality among 389 COVID-19 patients with diabetes hospitalized in the early period of the pandemic in the Veneto region: 22.89%15. Using multivariable logistic regression analysis, these authors reported that previous use of RAAS-i was associated with both reduced hospitalization and best outcomes in the diabetic population hospitalized for COVID-19. This finding is largely consistent with our results, but we were able to conduct separate analyses for ARBs and ACEIs. In addition, we used a stepwise method to find the best fitting in a multifactor logistic regression model. Besides, we analyzed the relationship of treatment outcomes with the actual intake of medication during hospitalization, but not with the treatment that preceded it. One of the features of our approach is that the effect of using ARBs / ACEIs and other hypotensive drugs was not compared between these groups, but the results of “No hypotensives” were used as a comparison group. We used a similar approach earlier when analyzing the results of hospital treatment of COVID-19 (n = 1097) depending on the use of RAAS inhibitors (ARBs / ACEIs). We found a reduction in the risk of death for COVID-19 inpatients without DM, who received RAAS inhibitors compared with the corresponding risk of normotensive inpatients, who did not receive antihypertensives: OR 0.22 (95% CI 0.07–0.72) adjusted for age, gender and FPG. The study mentioned did not distinguish the effect of ARBs from the effect of ACEIs and the protective effect was shown only for patients without DM6. The limitations of our study include its observational design, which prevents us from making definitive conclusions about causality, the study being conducted at a single center, and the small number of subjects studied. A small sample size can be considered as a weakness of our study. However, large sample size in the studies of Xu K, et al., 20231 and Hamada, S., et al., 20232 did not prevent them from obtaining conflicting results. A possible explanation may be that in the case of working with large medical databases and not with paper medical documents, often only the medications that patients took before hospitalization are considered. In our study, only actual antihypertensive treatment in the hospital was compared. In addition, we believe that the small sample size may be a source of doubt rather in the case of reporting a negative result, while in our case the null hypothesis is rejected with statistically sufficient probability. Severe sepsis in COVID-19 can cause hypotension, potentially limiting the use of antihypertensive drugs. However, since only one patient in our dataset had mild hypotension at admission, this likely did not affect the treatment categorization we analyzed. We emphasize that one of the strengths of our study is the separate investigation of the association between ACEIs/ARBs and COVID-19 mortality in the Ukrainian population of patients with severe COVID-19. In a recent review, Sodhi P. et al., 202319, highlight the role of ACE2 as a regulator of the RAAS and as a receptor for SARS-CoV-2. They discuss the significance of the two forms of ACE2, mACE2 and sACE2, and suggest that the interplay between SARS-CoV-2, mACE2, sACE2, and ANG II warrants further investigation in relation to COVID-19 severity. Additionally, they point out that ACE2 variants, which have been identified as specific to different populations, highlight interindividual variability. A recently published opinion (Tsukamoto S. 2024 )20 explains the mechanism, due to which ACEIs have advantages over ARBs in the prevention and treatment of viral pneumonias: “One mechanism of ACE-I prevention of pneumonia is speculated to be an increase in the cough reflex. Bradykinin and substance P, which are normally metabolized by angiotensin-converting enzyme, accumulate in the airways after ACE-I administration, stimulating vagal afferent fibers that control the cough reflex, thereby promoting the cough reflex. Increased coughing may promote expectoration of pathogens, including microaspiration, and may contribute to the prevention of pneumonia. This mechanism is not present in angiotensin II receptor blockers (ARBs), which are also renin-angiotensin system (RAS) inhibitors”. Another proposed mechanism of this protective effect is that RAAS hyperactivation is involved in severe pneumonia and systemic inflammation, and RAAS inhibitors may suppress RAS hyperactivation20. Our recent data can confirm the protective effect for COVID-19 mortality of ACEIs alone, so antiinflammation protection through cessation of RAAS hyperactivation may look questionable. Fastbom, J., et al. & Nyman Iliadou, A., 2024 conducted a cohort study, based on record linked individual-based data from national registers, of all Swedish inhabitants 50 years and older (n = 3,909,321) at the start of the first SARS-CoV-2 wave in Sweden. Lower risk of dying in COVID-19 was observed for ACEIs: HR 0.92 (0.86–0.99). For ARBs no significant associations was found21. Why do studies of the potential effect of RAAS inhibitors on the development and course of pneumonia often not match1,2,6,15,22,23? Possible explanations can be: separate assessments of ACEIs and ARBs were not conducted4,6,15,22,23; the impact of COVID-19 severity may not have been taken into account due to different hospitalization criteria; treatment data obtained from electronic databases1,2 were more likely to reflect prior use of RAAS inhibitors, whereas data manually extracted from paper records in hospital archives indicated their current use during acute COVID-19. The latter approach was used by researchers in Italy15 and Ukraine6, who found a reduction in COVID-19-related in-hospital mortality when RAAS inhibitors were used during hospitalization. Thus, our results probably received a theoretical explanation20 and partial confirmation at a large population level21. We highlight the distinct aspects of our research, which include: The analyzed population of patients with a very severe course of COVID-19 and high mortality, especially among those with a history of diabetes mellitus (DM). The identification of a DM-independent association between ACEIs treatment during hospitalization and COVID-19 outcomes. The results of the research that we present now are the basis for the following conclusion: regardless of age, minimal O2 saturation, DM history, treatment of inpatients with COVID-19 with ACEIs was associated with a significant reduction in the risk of death compared to no antihypertensive treatment. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1