Abstract
Background
This study was to examine the association between treatment with angiotensin-converting enzyme inhibitors (ACEIs) or angiotensin receptor blockers (ARBs) and the risk of developing ventilator-associated pneumonia (VAP) among patients receiving mechanical ventilation (MV) in the intensive care unit (ICU).
Methods
Utilizing a retrospective cohort approach, the data were extracted from the Medical Information Mart for Intensive Care IV database. VAP diagnoses were ascertained through the international classification of disease codes recorded in the database. Both univariate and multivariable logistic regression analyses were conducted to assess the association between ACEI or ARB use and VAP. Subgroup analyses were performed to evaluate the impact of comorbidities (AKI, renal failure, diabetes, hypertension, and sepsis), simplified acute physiology score II (SAPS II), as well as the use of vasopressors and antibiotics on this association. Odds ratios (ORs) with 95% confidence intervals (CIs) were used as the evaluation metrics.
Results
The study comprised 8,888 patients, with 897 (10.09%) experiencing VAP. The analysis revealed that patients on ACEI or ARB therapy had a lower risk of developing VAP (OR: 0.79, 95% CI: 0.62–0.99, P = 0.047). Subgroup analyses revealed that the protective effect was observed in patients with AKI (OR: 0.70, 95% CI: 0.52–0.94, P = 0.020), renal failure (OR: 0.14, 95% CI: 0.02–0.84, P = 0.032), and diabetes (OR: 0.64, 95% CI: 0.43–0.94, P = 0.024), as well as in those receiving vasopressors (OR: 0.67, 95% CI: 0.49–0.92, P = 0.012), and antibiotics (OR: 0.74, 95% CI: 0.57–0.96, P = 0.021). No significant difference in VAP development was observed between patients treated with ACEI versus ARB (OR: 0.84, 95% CI: 0.49–1.47, P = 0.547).
Conclusion
This study’s findings suggest a substantial association between the use of ACEIs or ARBs and reduced development of VAP, particularly among patients with specific comorbidities and those on vasopressor and antibiotic therapy. This study may educate the ICU team on the potential benefits of ACEIs and ARBs in preventing VAP, emphasizing the importance of considering these medications in the overall treatment plan.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12890-024-03386-y.
Keywords: Angiotensin-converting enzyme inhibitors, Angiotensin receptor blockers, Ventilator-associated pneumonia, Intensive care unit, MIMIC database
Background
The application of mechanical ventilation (MV) in the intensive care unit (ICU) for critically ill patients represents a standard therapeutic intervention [1, 2]. However, ventilator-associated pneumonia (VAP) is one of the most prevalent healthcare-acquired infections in the ICU environment [3, 4], defined as pneumonia that develops in ICU patients who have been mechanically ventilated for at least 48 h [5]. The reported incidence of VAP is around 55.3% [6]. VAP leads to prolonged stays in the ICU and increased hospitalization costs, and it is a significant cause of mortality among critically ill patients [7]. Therefore, exploring the factors associated with the risk of developing VAP in the ICU is critical for prevention, early diagnosis, and effective treatment.
Angiotensin-converting enzyme inhibitors (ACEIs) and angiotensin receptor blockers (ARBs) are widely used antihypertensive drugs that play a crucial role in the management of hypertension [8]. Beyond blood pressure-lowering effects, ACEI and ARBs have also improved short-term outcomes in hospitalized and ICU patients [9–12]. Previous studies have indicated that the use of ACEI and ARBs was associated with a reduced risk of community-acquired pneumonia (CAP) occurrence [13] and hospitalization [14, 15]. In a study conducted among patients with chronic obstructive pulmonary disease (COPD), compared to ACEI, the utilization of ARBs was associated with a lower risk of pneumonia and severe pneumonia occurrence [16]. The potential mechanisms by which ACEI and ARBs reduce the risk of pneumonia may involve the modulation of immune responses [17] and attenuation of lung injury [18]. By inhibiting the effects of angiotensin II, both ACEIs and ARBs can reduce the production of pro-inflammatory cytokines [19], which are key mediators in the inflammatory response and have been implicated in the pathogenesis of pneumonia. Although the benefits of ACEI and ARB in the treatment of several diseases are widely recognized, studies on whether these drugs have an impact on the risk of developing VAP remain relatively limited.
Therefore, this study aims to investigate the association between ACEI or ARB use and the risk of VAP in patients with MV in the ICU. This study may provide clinicians with guidelines on the use of ACEI/ARB in ICU patients, as well as directions for future research aimed at improving strategies for the prevention and management of VAP.
Methods
Study design and selection of patients
The study was a retrospective cohort design, and collected data from the Medical Information Mart for Intensive Care IV (MIMIC-IV) database from the period 2008 to 2019: https://mimic.mit.edu/docs/iv/. The MIMIC database is a collaborative effort published by the Computational Physiology Laboratory at the Massachusetts Institute of Technology (MIT, Cambridge, MA, USA), Beth Israel Deaconess Medical Center (BIDMC, Boston, MA, USA), and Philips Medical. This database compiles and organizes clinical diagnosis and treatment information from over 40,000 actual patients, in the ICU at BIDMC, since the year 2001. The inclusion criteria were: (1) Patients aged ≥ 18 years old; (2) Patients who have undergone continuous MV for a duration exceeding 48 h; (3) Information for usage of ACEI or ARB at admission. Exclusion criteria included patients who were not evaluated for VAP. All patient information within the database was stripped of identifiers to ensure privacy protection. As a result, the requirement for obtaining informed consent was waived. Utilization of the MIMIC database for this study was approval by the review boards of MIT and the BIDMC. Consequently, it was determined that there was no requirement to seek supplementary ethical approval from the People’s Hospital, Yanliang District.
Data collection
The dataset extracted from the MIMIC database encompassed a range of information, including (1) Baseline characteristics:age (years), gender, race, comorbidities [diabetes, hypertension, trauma, liver failure, renal failure, sepsis, and acute kidney injury (AKI)]; (2) Vital signs: heart rate (bpm), temperature (Deg.C), mean blood pressure (MBP, mmHg), respiratory rate (bpm); (3) Laboratory value: peripheral capillary oxygen saturation (SpO2, %), hematocrit (%), white blood cell (WBC, K/uL), platelet (K/uL), red cell distribution width (RDW, %), creatinine (mg/dL), glucose (mg/dL), blood urea nitrogen (BUN, mg/dL), phosphate (mg/dL), anion gap (mEq/L), urine output mL); (4) Scoring systems: simplified acute physiology score II (SAPS II); (5) Intervention: MV duration (hours), antibiotics, Propofol, Midazolam, Dexmedetomidine, and vasopressor use.
Definitions and measurements
In the MIMIC-IV database, patients diagnosed with VAP were identified using the following classification codes: international classification of diseases (ICD)-9 codes 4957 and 99,731, as well as the ICD-10 code J95851. The use of ACEI (benazepril, captopril, enalapril, fosinopril, lisinopril, moexipril, monopril, quinapril, ramipril, trandolapril) and ARB (irbesartan, valsartan, candesartan, losartan, irbesartan, olmesartan, and valsartan, telmisartan) was identified by keywords in the database, with the period from admission to the ICU until the first 48 h of continuous MV (baseline). The definitions and measurements of other potential covariates are shown in Supplementary File 1.
Statistical analysis
Continuous data that were normally distributed were described using the mean (standard deviation) [Mean ± SD], and comparisons between groups were checked using the independent samples t-test. Non-normally distributed continuous data were presented as the median and interquartile range [M (Q1, Q3)], with group comparisons performed using the Mann–Whitney U test. Categorical data were expressed as the number of cases and the proportion [N (%)], and comparisons between groups were conducted using the chi-square test. Rank data were compared by the rank sum test. To address missing data in the variables, multiple imputation techniques were applied (Supplementary Table 1). Subsequent sensitivity analyses (Supplementary Table 2) were carried out to evaluate the consistency of the outcomes pre- and post-imputation, thereby determining the influence of the imputation process on the research results.
For the selection of covariates for the association between the use of ACEI or ARBs and the risk of VAP, a univariate logistic regression model was employed to identify variables that could potentially influence the risk of VAP. The variables with P < 0.05 were considered as the covariates including age, gender, race, ventilation duration, antibiotics, Propofol, Midazolam, Dexmedetomidine, vasopressor use, SAPSII, Trauma, liver failure, sepsis, MBP, respiratory rate, glucose, anion gap, and urine output. Subsequently, the cohort was narrowed down by excluding patients who did not receive ACEI or ARB. Another round of univariate logistic regression was conducted to select covariates for the outcome of the comparative analysis of association with VAP in patients using ACEI and ARB.. The variables with P < 0.05 were included as covariates, which involved age, ventilation duration, Propofol, Midazolam, Dexmedetomidine, SAPSII, trauma, liver failure, temperature, MBP, and platelet count. Additionally, considering the clinical significance of gender and race, these variables were also included as covariates. The univariate (adjusted for none) and multivariable (adjusted for covariates) logistic regression models were employed to explore the association between the use of ACEI or ARBs and the risk of VAP. To ensure the reliability of our multivariable models, we employed the variance inflation factor (VIF) to test for multicollinearity among the variables. A VIF threshold of 10 was used to identify variables that may be collinear and to guide decisions on variable retention in the model (Supplementary Table 3). Subgroup analyses were conducted based on whether the patients suffered from AKI, renal failure, diabetes, hypertension, sepsis, or SAPSII, and based on whether the patients used vasopressors and antibiotics. The likelihood ratio (LR), Wald test, and score test were employed to evaluate the statistical significance of the models, with their respective P-values reported in Supplementary Table 4. These tests demonstrated statistical significance, affirming the model's validity. Additionally, the area under the curve (AUC) value of 0.784 (Supplementary Fig. 1) provided an indication of the model's discriminative power. In addition, patients were matched on the calculated propensity scores using a 1:1 ratio. Evaluation indexes were reported as odds ratios (ORs) with 95% confidence intervals (CIs). All statistical tests were two-tailed, with a significance level set at α = 0.05. Data cleaning, handling of missing values, and statistical analysis of the models were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA).
Results
Population selection process and characteristics of included populations
A total of 53,538 records of patients at ICU admission were identified from the MIMIC IV database. After exclusions, the final sample was 8,888. Figure 1 illustrates the patients’ selection process. Among the 8,888 individuals included, 897 (10.09%) developed VAP. In the group not receiving either ACEI or ARB, 800 out of 7,486 patients (10.69%) developed VAP. In the group receiving either ACEI or ARB, 97 out of 1,402 patients (6.92%) developed VAP. Table 1 presents a comparative analysis of various indicators based on the grouping of patients who did not use either ACEI or ARBs in conjunction with those who used ACEI or ARBs. The mean age was 65.05 ± 16.33 years. The sample comprised 3,946 females (44.40%) and 4,942 males (55.60%). Within the group not receiving either ACEI or ARBs, there were 3,359 females (44.87%) and 4,127 males (55.13%). In the group receiving ACEI or ARBs, the count was 587 females (41.87%) and 815 males (58.13%). For the overall sample, the median ventilation duration was 70.05 (57.00, 99.00) hours. Out of the total patient population, 1,749 (19.68%) did not receive antibiotics, while a substantial majority, 7,139 patients (80.32%), were treated with antibiotics. There were significant differences between patients who used ACEI or ARBs and patients who did not use either ACEI or ARB in age, gender, race, ventilation duration, antibiotics use, Midazolam use, vasopressor use, VAP risk, SAPSII, diabetes, hypertension, trauma, liver failure, sepsis, heart rate, temperature, MBP, respiratory rate, WBC, platelet, RDW, phosphate, anion gap, and urine output (each P < 0.05). The characteristics of included populations after propensity score matching (PSM) are shown in Supplementary Table 5. Supplementary Fig. 2 illustrates the distribution of the Logit of Propensity Score in the treatment and control groups with respect to patients with the use of ACEI or ARB and patients without the use of ACEI or ARB before and after PSM. Supplementary Fig. 3 shows the distribution of the Logit of Propensity Score in the treatment and control groups that compared the use of ACEI and ARB.
Fig. 1.
The flow diagram of the patient’s selection process
Table 1.
Basic characteristics of included patients
| Variables | Total (n = 8888) | ACEI = No and ARB = No (n = 7486) | ACEI = Yes or ARB = Yes (n = 1402) | Statistic | P |
|---|---|---|---|---|---|
| Age, ages, Mean ± SD | 65.05 ± 16.33 | 64.08 ± 16.77 | 70.24 ± 12.55 | t = -15.91 | < 0.001 |
| Gender, n (%) | χ2 = 4.310 | 0.038 | |||
| Female | 3946 (44.40) | 3359 (44.87) | 587 (41.87) | ||
| Male | 4942 (55.60) | 4127 (55.13) | 815 (58.13) | ||
| Race, n (%) | χ2 = 40.010 | < 0.001 | |||
| Black | 647 (7.28) | 534 (7.13) | 113 (8.06) | ||
| White | 5874 (66.09) | 4870 (65.05) | 1004 (71.61) | ||
| Other | 919 (10.34) | 787 (10.51) | 132 (9.42) | ||
| Unknown | 1448 (16.29) | 1295 (17.30) | 153 (10.91) | ||
| Ventilation duration, hours, M (Q1, Q3) | 70.05 (57.00, 99.00) | 71.21 (57.57, 101.28) | 66.50 (56.00, 90.00) | Z = -6.030 | < 0.001 |
| Antibiotics, n (%) | χ2 = 26.337 | < 0.001 | |||
| No | 1749 (19.68) | 1403 (18.74) | 346 (24.68) | ||
| Yes | 7139 (80.32) | 6083 (81.26) | 1056 (75.32) | ||
| Propofol, n (%) | χ2 = 0.087 | 0.768 | |||
| No | 3880 (43.65) | 3273 (43.72) | 607 (43.30) | ||
| Yes | 5008 (56.35) | 4213 (56.28) | 795 (56.70) | ||
| Midazolam, n (%) | χ2 = 74.744 | < 0.001 | |||
| No | 6169 (69.41) | 5059 (67.58) | 1110 (79.17) | ||
| Yes | 2719 (30.59) | 2427 (32.42) | 292 (20.83) | ||
| Dexmedetomidine, n (%) | χ2 = 2.234 | 0.135 | |||
| No | 7748 (87.17) | 6543 (87.40) | 1205 (85.95) | ||
| Yes | 1140 (12.83) | 943 (12.60) | 197 (14.05) | ||
| Vasopressor use, n (%) | χ2 = 35.138 | < 0.001 | |||
| No | 3917 (44.07) | 3198 (42.72) | 719 (51.28) | ||
| Yes | 4971 (55.93) | 4288 (57.28) | 683 (48.72) | ||
| VAP, n (%) | χ2 = 18.476 | < 0.001 | |||
| No | 7991 (89.91) | 6686 (89.31) | 1305 (93.08) | ||
| Yes | 897 (10.09) | 800 (10.69) | 97 (6.92) | ||
| SAPSII, M (Q1, Q3) | 44.00 (34.00, 56.00) | 44.00 (34.00, 56.00) | 42.00 (33.00, 52.00) | Z = -4.929 | < 0.001 |
| Diabetes, n (%) | χ2 = 122.243 | < 0.001 | |||
| No | 6316 (71.06) | 5492 (73.36) | 824 (58.77) | ||
| Yes | 2572 (28.94) | 1994 (26.64) | 578 (41.23) | ||
| Hypertension, n (%) | χ2 = 193.898 | < 0.001 | |||
| No | 5119 (57.59) | 4548 (60.75) | 571 (40.73) | ||
| Yes | 3769 (42.41) | 2938 (39.25) | 831 (59.27) | ||
| Trauma, n (%) | χ2 = 18.395 | < 0.001 | |||
| No | 7482 (84.18) | 6248 (83.46) | 1234 (88.02) | ||
| Yes | 1406 (15.82) | 1238 (16.54) | 168 (11.98) | ||
| Liver failure, n (%) | χ2 = 68.905 | < 0.001 | |||
| No | 8120 (91.36) | 6759 (90.29) | 1361 (97.08) | ||
| Yes | 768 (8.64) | 727 (9.71) | 41 (2.92) | ||
| Renal failure, n (%) | χ2 = 0.094 | 0.759 | |||
| No | 8539 (96.07) | 7190 (96.05) | 1349 (96.22) | ||
| Yes | 349 (3.93) | 296 (3.95) | 53 (3.78) | ||
| Sepsis, n (%) | χ2 = 34.145 | < 0.001 | |||
| No | 7574 (85.22) | 6308 (84.26) | 1266 (90.30) | ||
| Yes | 1314 (14.78) | 1178 (15.74) | 136 (9.70) | ||
| AKI, n (%) | χ2 = 1.261 | 0.261 | |||
| No | 3129 (35.20) | 2617 (34.96) | 512 (36.52) | ||
| Yes | 5759 (64.80) | 4869 (65.04) | 890 (63.48) | ||
| Heart rate, bpm Mean ± SD | 91.15 ± 20.53 | 92.01 ± 20.83 | 86.52 ± 18.22 | t = 10.12 | < 0.001 |
| Temperature, Deg.C, Mean ± SD | 36.79 ± 1.89 | 36.78 ± 1.99 | 36.87 ± 1.20 | t = -2.30 | 0.021 |
| MBP, mmHg, Mean ± SD | 82.89 ± 17.02 | 82.70 ± 16.99 | 83.88 ± 17.16 | t = -2.39 | 0.017 |
| Respiratory rate, bpm, M (Q1, Q3) | 19.00 (15.00, 23.00) | 19.00 (15.00, 23.00) | 20.00 (16.00, 24.00) | Z = 3.706 | < 0.001 |
| SpO2,%, Mean ± SD | 96.42 ± 10.06 | 96.45 ± 10.86 | 96.27 ± 3.54 | t = 1.14 | 0.255 |
| Hematocrit, %, Mean ± SD | 31.71 ± 6.38 | 31.73 ± 6.47 | 31.59 ± 5.83 | t = 0.82 | 0.411 |
| WBC, K/uL, M (Q1, Q3) | 11.80 (8.60, 15.90) | 11.90 (8.50, 16.20) | 11.50 (8.80, 14.60) | Z = -3.369 | < 0.001 |
| Platelet, K/uL, M (Q1, Q3) | 183.00 (127.00, 256.00) | 182.00 (125.00, 254.00) | 190.00 (135.00, 264.00) | Z = 3.591 | < 0.001 |
| RDW, %, Mean ± SD | 15.26 ± 2.30 | 15.32 ± 2.38 | 14.94 ± 1.80 | t = 6.94 | < 0.001 |
| Creatinine, mg/dL, M (Q1, Q3) | 1.00 (0.70, 1.70) | 1.00 (0.70, 1.70) | 1.00 (0.80, 1.40) | Z = -1.498 | 0.134 |
| Glucose, mg/dL, M (Q1, Q3) | 131.00 (108.00, 165.00) | 131.00 (108.00, 166.00) | 131.50 (110.00, 163.00) | Z = -0.057 | 0.954 |
| BUN, mg/dL, M (Q1, Q3) | 21.00 (14.00, 36.00) | 21.00 (14.00, 36.00) | 22.00 (15.00, 33.00) | Z = 0.815 | 0.415 |
| Phosphate, mg/dL, M (Q1, Q3) | 3.50 (2.80, 4.40) | 3.50 (2.80, 4.50) | 3.40 (2.80, 4.10) | Z = -3.521 | < 0.001 |
| Anion gap, mEq/L, Mean ± SD | 14.66 ± 4.31 | 14.82 ± 4.43 | 13.80 ± 3.49 | t = 9.58 | < 0.001 |
| Urine output, mL, M (Q1, Q3) | 1600.00 (930.00, 2488.50) | 1555.00 (897.00, 2465.00) | 1770.50 (1145.00, 2625.00) | Z = 6.561 | < 0.001 |
Notes: ACEI Angiotensin-converting enzyme inhibitors, ARB Angiotensin receptor blocker, VAP Ventilator-associated pneumonia, SAPSII Simplified acute physiology score II, AKI Acute kidney injury, MBP Mean blood pressure, SpO2 Peripheral capillary oxygen saturation, WBC White blood cell, RDW Red cell distribution width, BUN Blood urea nitrogen
Analysis of the association between the use of ACEI or ARBs and the development of VAP
Table 2 presents the analysis of the association between the use of ACEI or ARB, and the development of VAP. For patients using either ACEI or ARB, model 1 showed a significantly reduced risk of VAP compared to the reference group (patients not using either ACEI or ARB) (OR: 0.62, 95% CI: 0.50–0.77, P < 0.001). After adjusting for covariates (model 2), the results also showed that the use of ACEI or ARB was associated with a decreased risk of VAP development (OR: 0.79, 95% CI: 0.62–0.99, P = 0.047). After PSM, the results were consistent (Supplementary Table 6).
Table 2.
Association between the use of ACEI or ARB and the risk of VAP
| Indicators | Model I | Model II | ||
|---|---|---|---|---|
| OR (95% CI) | P | OR (95% CI) | P | |
| ACEI = NO and ARB = NO | Ref | Ref | ||
| ACEI = YES or ARB = YES | 0.62 (0.50–0.77) | < 0.001 | 0.79 (0.62–0.99) | 0.047 |
Notes: ACEI Angiotensin-converting enzyme inhibitors, ARB Angiotensin receptor blocker, VAP Ventilator-associated pneumonia, OR Odds ratio, CI Confidence interval; Model I was an unadjusted model; Model II adjusted for age, gender, race, ventilation duration, antibiotics, Propofol, Midazolam, Dexmedetomidine, vasopressor use, SAPSII, trauma, liver failure, sepsis, MBP, respiratory rate, glucose, anion gap, and urine output
Subgroup analyses of the association between the use of ACEI or ARBs and the development of VAP
The subgroup analyses indicated that in patients with AKI (OR: 0.70, 95% CI: 0.52–0.94, P = 0.020), in patients with renal failure (OR: 0.14, 95% CI: 0.02–0.84, P = 0.032), in patients with diabetes (OR: 0.64, 95% CI: 0.43–0.94, P = 0.024), in patients used vasopressor (OR: 0.67, 95% CI: 0.49–0.92, P = 0.012), and antibiotics (OR: 0.74, 95% CI: 0.57–0.96, P = 0.021), the use of ACEI/ARB was associated with a reduced risk of VAP. Subgroup analyses of the association between the use of ACEI or ARBs and the risk of VAP are presented in Table 3.
Table 3.
Subgroup analyses of the association between the use of ACEI or ARB and the risk of VAP
| Indicators | OR (95% CI) | P | OR (95% CI) | P |
|---|---|---|---|---|
| Subgroup I: AKI | No | Yes | ||
| ACEI = NO and ARB = NO | Ref | Ref | ||
| ACEI = YES or ARB = YES | 0.98 (0.66–1.45) | 0.922 | 0.70 (0.52–0.94) | 0.020 |
| Subgroup II: Renal failure | No | Yes | ||
| ACEI = NO and ARB = NO | Ref | Ref | ||
| ACEI = YES or ARB = YES | 0.82 (0.64–1.04) | 0.094 | 0.14 (0.02–0.84) | 0.032 |
| Subgroup III: Vasopressor use | No | Yes | ||
| ACEI = NO and ARB = NO | Ref | Ref | ||
| ACEI = YES or ARB = YES | 1.10 (0.77–1.57) | 0.612 | 0.67 (0.49–0.92) | 0.012 |
| Subgroup IV: Antibiotics | No | Yes | ||
| ACEI = NO and ARB = NO | Ref | Ref | ||
| ACEI = YES or ARB = YES | 1.19 (0.65–2.20) | 0.572 | 0.74 (0.57–0.96) | 0.021 |
| Subgroup V: Diabetes | No | Yes | ||
| ACEI = NO and ARB = NO | Ref | Ref | ||
| ACEI = YES or ARB = YES | 0.90 (0.67–1.21) | 0.473 | 0.64 (0.43–0.94) | 0.024 |
| Subgroup VI: Hypertension | No | Yes | ||
| ACEI = NO and ARB = NO | Ref | Ref | ||
| ACEI = YES or ARB = YES | 0.63 (0.42–0.93) | 0.020 | 0.85 (0.63–1.15) | 0.295 |
| Subgroup VII: Sepsis | No | Yes | No | Yes |
| ACEI = NO and ARB = NO | Ref | Ref | ||
| ACEI = YES or ARB = YES | 0.81 (0.63–1.04) | 0.103 | 0.65 (0.34–1.24) | 0.191 |
| Subgroup VIII: SAPSII | < 34 (Q1) | > = 34 (Q1) | ||
| ACEI = NO and ARB = NO | Ref | Ref | ||
| ACEI = YES or ARB = YES | 0.94 (0.44–2.02) | 0.876 | 0.79 (0.62–1.00) | 0.054 |
Notes: ACEI Angiotensin-converting enzyme inhibitors, ARB Angiotensin receptor blocker, VAP Ventilator-associated pneumonia, OR Odds ratio, CI Confidence interval, AKI Acute kidney injury, SAPSII Simplified Acute Physiology Score II; Subgroup I of AKI adjusted for age, gender, race, ventilation duration, antibiotics, Propofol, Midazolam, Dexmedetomidine, Vasopressor use, SAPSII, trauma, liver failure, sepsis, MBP, respiratory rate, glucose, anion gap, and urine output; Subgroup II of renal failure adjusted for age, gender, race, ventilation duration, antibiotics, Propofol, Midazolam, Dexmedetomidine, vasopressor use, SAPSII, trauma, liver failure, sepsis, MBP, respiratory rate, glucose, anion gap, and urine output. Subgroup III of vasopressor use adjusted for age, gender, race, ventilation duration, antibiotics, Propofol, Midazolam, Dexmedetomidine, SAPSII, trauma, liver failure, sepsis, MBP, respiratory rate, glucose, anion gap, and urine output; Subgroup IV of antibiotics adjusted for age, gender, race, ventilation duration, Propofol, Midazolam, Dexmedetomidine, vasopressor use, SAPSII, trauma, liver failure, sepsis, MBP, respiratory rate, glucose, anion gap, and urine output; Subgroup V of diabetes adjusted for age, gender, race, ventilation duration, antibiotics, Propofol, Midazolam, Dexmedetomidine, vasopressor use, SAPSII, trauma, liver failure, sepsis, MBP, respiratory rate, glucose, anion gap, and urine output; Subgroup VI of hypertension adjusted for age, gender, race, ventilation duration, antibiotics, propofol, midazolam, dexmedetomidine, vasopressor use, SAPSII, trauma, liver failure, sepsis, MBP, respiratory rate, glucose, anion gap, and urine output; Subgroup VII of sepsis adjusted for age, gender, race, ventilation duration, antibiotics, Propofol, Midazolam, Dexmedetomidine, vasopressor use, SAPSII, trauma, liver failure, MBP, respiratory rate, glucose, anion gap, and urine output; Subgroup VIII of SAPSII adjusted for age, gender, race, ventilation duration, antibiotics, Propofol, Midazolam, Dexmedetomidine, vasopressor use, trauma, liver failure, sepsis, MBP, respiratory rate, glucose, anion gap, and urine output; SAPS II values were divided at the first quartile (Q1) of the study population's distribution
Comparison analysis of association with VAP development in patients using ACEI and ARB
The result showed that there was no statistical difference in the risk of VAP development between patients who used ACEI and ARB (OR: 0.84, 95% CI: 0.49–1.47, P = 0.547) (Table 4). The findings from the subgroup analyses also indicated no significant difference in the risk of VAP development among patients treated with ACEI compared to those treated with ARBs (Table 5).
Table 4.
Comparison analysis of the association with VAP risk in patients using ACEI and ARB
| Indicators | Model I | Model II | ||
|---|---|---|---|---|
| OR (95% CI) | P | OR (95% CI) | P | |
| Drug type | ||||
| ACEI | Ref | Ref | ||
| ARB | 0.80 (0.48–1.32) | 0.379 | 0.84 (0.49–1.47) | 0.547 |
Notes: ACEI Angiotensin-converting enzyme inhibitors, ARB Angiotensin receptor blocker, VAP Ventilator-associated pneumonia, OR Odds ratio, CI Confidence interval; Model I was an unadjusted model; Model II adjusted for Age, Gender, Race, Ventilation duration, Propofol, Midazolam, Dexmedetomidine, SAPSII, trauma, liver failure, temperature, MBP, and platelet
Table 5.
Subgroup of the comparison analysis of the association with VAP risk in patients using ACEI and ARB
| Indicators | OR (95%CI) | P | OR (95%CI) | P |
|---|---|---|---|---|
| Subgroup I:AKI | No | Yes | ||
| ACEI | Ref | Ref | ||
| ARB | 1.03 (0.41–2.59) | 0.956 | 0.83 (0.40–1.72) | 0.618 |
| Subgroup II: Renal failure | No | Yes | ||
| ACEI | Ref | Ref | ||
| ARB | 0.84 (0.48–1.48) | 0.556 | - | - |
| Subgroup III: Vasopressor use | No | Yes | ||
| ACEI | Ref | Ref | ||
| ARB | 0.57 (0.22–1.47) | 0.244 | 1.11 (0.54–2.29) | 0.777 |
| Subgroup IV: Antibiotics | No | Yes | ||
| ACEI | Ref | Ref | ||
| ARB | 0.26 (0.03–2.21) | 0.217 | 1.01 (0.56–1.84) | 0.968 |
| Subgroup V: Diabetes | No | Yes | ||
| ACEI | Ref | Ref | ||
| ARB | 0.94 (0.47–1.89) | 0.863 | 0.75 (0.28–2.01) | 0.565 |
| Subgroup VI: Hypertension | No | Yes | ||
| ACEI | Ref | Ref | ||
| ARB | 0.93 (0.39–2.22) | 0.875 | 0.81 (0.38–1.70) | 0.570 |
| Subgroup VII: Sepsis | No | Yes | ||
| ACEI | Ref | Ref | ||
| ARB | 0.88 (0.49–1.60) | 0.679 | 0.35 (0.04–3.16) | 0.352 |
| Subgroup VIII: SAPSII | < 34 (Q1) | < 34 (Q1) | ||
| ACEI | Ref | Ref | ||
| ARB | 0.77 (0.07–8.89) | 0.836 | 0.91 (0.52–1.59) | 0.740 |
Notes: ACEI Angiotensin-converting enzyme inhibitors, ARB Angiotensin receptor blocker, VAP Ventilator-associated pneumonia, OR Odds ratio, CI Confidence interval, AKI Acute kidney injury, SAPSII Simplified Acute Physiology Score II; Subgroup I of AKI adjusted for age, gender, race, ventilation duration, Propofol, Midazolam, Dexmedetomidine, SAPSII, trauma, liver failure, temperature, MBP, and platelet; Subgroup II of the renal failure adjusted for age, gender, race, ventilation duration, Propofol, Midazolam, Dexmedetomidine, SAPSII, trauma, liver failure, temperature, MBP, and platelet; Subgroup III of the vasopressor use adjusted for age, gender, race, ventilation duration, Propofol, Midazolam, Dexmedetomidine, SAPSII, trauma, liver failure, temperature, MBP, and platelet; Subgroup IV of the antibiotics adjusted for age, gender, race, ventilation duration, Propofol, Midazolam, Dexmedetomidine, SAPSII, trauma, liver failure, temperature, MBP, and Platelet; Subgroup V of diabetes adjusted for age, gender, race, ventilation duration, Propofol, Midazolam, Dexmedetomidine, SAPSII, trauma, liver failure, temperature, MBP, and platelet; Subgroup VI of hypertension adjusted for age, gender, race, ventilation duration, Propofol, Midazolam, Dexmedetomidine, SAPSII, trauma, liver failure, temperature, MBP, and platelet; Subgroup VII of sepsis adjusted for age, gender, race, ventilation duration, Propofol, Midazolam, Dexmedetomidine, SAPSII, trauma, liver failure, temperature, MBP, and Platelet; Subgroup VIII of SAPSII adjusted for age, gender, race, ventilation duration, Propofol, Midazolam, Dexmedetomidine, trauma, liver failure, temperature, MBP, and platelet
Discussion
In this study, we focused on exploring the association between the utilization of ACEI or ARBs and the risk of developing VAP among patients who were in the ICU. The use of ACEI/ARB was found to be associated with a reduced risk of VAP, specifically in patients with AKI, renal failure, and diabetes. Additionally, the association between ACEI/ARB use and decreased VAP risk was observed in patients receiving vasopressor support and those treated with antibiotics. The comparison analysis of the association between the use of ACEI and ARB and the risk of VAP development did not reach statistical significance.
Previous studies have confirmed the association between ACEI/ARB and pneumonia. A case-crossover study conducted using the Taiwan National Health Insurance Research Database (NHIRD) revealed a significant protective effect of ACEI use on pneumonia hospitalization in stroke patients, exhibiting a dose–response relationship [20]. In a study evaluating patients aged fifty years or older, initially diagnosed with Parkinson’s disease (PD) and comorbid hypertension, the use of ACEI was associated with a reduced risk of pneumonia [21]. A systematic review suggests that ACEI may be beneficial in reducing the incidence of CAP in the elderly [22]. A retrospective cohort study conducted in Taiwan found that the use of both ACEI and ARB was associated with a reduced risk of CAP infection [13]. A cohort study also found that, compared to non-users of ACEI, users of ACEI had a lower likelihood of developing pneumonia [14]. This study was the first study revealing the association between the use of ACEI/ARB and a reduced risk of VAP in the ICU. This association, not previously reported, opens up new avenues for further research into the potential protective effects of these medications beyond their established cardiovascular benefits. Our findings may also have implications for clinical practice, guiding the consideration of ACEI/ARB use in the ICU setting for patients who could benefit from a reduced risk of VAP. The potential mechanisms by which ACEI and ARBs may be associated with a reduced risk of VAP are multifaceted and not yet fully elucidated. However, several hypotheses can be proposed based on current understanding. ACEI and ARBs are known to modulate the renin–angiotensin–aldosterone system (RAAS), which plays a role in immune function [23]. By inhibiting the production of angiotensin II, ACEI and ARBs may reduce inflammation and enhance the immune response to pathogens, potentially decreasing the risk of pneumonia [24]. The use of ACEI and ARBs may improve pulmonary vascular integrity by reducing endothelial dysfunction and inflammation [25], which could limit the spread of infection within the lungs.
The findings from the study indicate that the association between the use of ACEI or ARBs and a reduced risk of VAP was significant in patients with comorbidities such as AKI, renal failure, and diabetes, as well as in patients who are receiving vasopressors and antibiotics. However, in patients without these comorbidities and those not receiving vasopressors or antibiotics, the use of ACEI/ARB does not show a statistically significant association with a reduced risk of VAP. Further research is needed to elucidate the underlying mechanisms and to determine whether the observed associations are due to the direct effects of ACEI/ARBs on VAP risk or if they are related to other factors associated with the complex care of patients with multiple comorbidities in the ICU. These results highlight the importance of considering patient-specific factors and treatment regimens when evaluating the potential benefits of ACEI/ARB use in the ICU.
Previous research has compared the association between the use of ACEI and ARBs with the risk of pneumonia. A systematic review and meta-analysis indicated that ACEI (but not ARBs) have a presumed protective effect on the risk of pneumonia [26]. A study that recruited patients with COPD who had used ACEI and ARBs for more than 90 days between 2000 and 2005 found an association between the use of ARBs and pneumonia, including severe pneumonia, in patients with COPD when compared to ACEI use [16]. A nationwide cohort study within the Taiwan National Health Insurance database found no association between ACEI treatment and a reduced risk of pneumonia incidence compared to losartan (an ARB with similar indications) [16]. Our study discovered no statistically significant difference in the association between the use of ACEI and ARBs with the risk of VAP. The results indicate that, within the confines of our research and the parameters examined, the impact of ACEI and ARBs on the incidence of VAP is comparable, with neither class demonstrating a distinct advantage in terms of risk reduction. Further investigation may be required to check other factors that could influence VAP outcomes among ICU patients.
The clinical significance of our study encompasses several key aspects. First, our findings may guide intensivists in considering the use of ACEI or ARBs as adjunctive therapy in patients requiring MV, with the aim of reducing the risk of VAP. Second, given that VAP treatment often necessitates additional antibiotic use and prolonged MV, this could lead to increased healthcare resource utilization. A reduction in VAP risk may consequently help to mitigate the associated medical costs. Third, the results of this study underscore the potential benefits of using ACEI/ARBs in patients with specific comorbidities, supporting the concept of personalized medicine that tailors treatment plans according to the individual patient’s conditions. Fourth, our findings may stimulate further research into the role of ACEI/ARBs in preventing other complications in the ICU, as well as a broader assessment of the safety and efficacy of these medications in critically ill patients.
This study is the first to explore the relationship between the use of ACEI or ARBs and the risk of VAP in mechanically ventilated patients in the ICU, providing a basis for drug decision-making aimed at reducing VAP risk in this patient population. The large sample size ensures adequate power to detect significant associations, and the study comprehensively considers various factors, including comorbidities and laboratory indicators, that could influence outcomes within the ICU setting.
However, this study still has limitations. First, the retrospective cohort design of the study is subject to inherent biases characteristic of this type of research. Second, the data was collected from the MIMIC-IV database, which is a single medical center, potentially limiting the generalizability of the findings. Large-scale, multicenter studies are needed to validate the results. Third, due to database limitations, pre-admission medication use could not be included in the analysis, which may affect the comprehensiveness of the study’s conclusions. Fourth, one of the primary constraints of this research is the inability to differentiate between patients who were newly prescribed ACEIs or ARBs and those who were already on these medications prior to their ICU admission. This limitation stems from the constraints inherent in the available database, which does not provide comprehensive medication histories for all patients. The absence of detailed pre-ICU admission medication data may obscure the nuances in patient responses to ACEIs and ARBs, as the effects of de novo therapy could differ significantly from those of chronic use. Given that some medications require an accumulation period to achieve therapeutic efficacy, the timing of initiation relative to ICU admission could be a pivotal factor influencing the study outcomes. Furthermore, this limitation underscores the need for caution when interpreting the results, as the observed associations may not be generalizable to all patients receiving ACEIs or ARBs, particularly those who have been on these medications for an extended period prior to ICU admission. Fifth, the reduced sample sizes within certain subgroups may impact the robustness of our model, particularly in terms of variable selection and significance. While we performed rigorous statistical testing to select covariates, smaller subgroup sizes may limit the power to detect associations, potentially introducing variability in the model outcomes. Future studies with larger sample sizes are needed to validate these findings and enhance generalizability.
Conclusion
While the findings from this study suggest a significant association between the use of ACEIs or ARBs and a lower risk of VAP in the ICU setting, it is crucial to consider the limitations regarding medication usage. The incomplete data on certain medications may introduce variability that affects the precision of our estimates. Therefore, these results should be interpreted with caution and serve as a foundation for further investigation. Future research with more comprehensive medication data is needed to confirm these findings and to fully assess the potential benefits of incorporating ACEIs or ARBs into clinical strategies aimed at improving patient outcomes in the ICU.
Supplementary Information
Acknowledgements
Not applicable.
Abbreviations
- MV
Mechanical ventilation;
- ICU
Intensive care unit
- VAP
Ventilator-associated pneumonia
- ACEIs
Angiotensin-converting enzyme inhibitors
- ARBs
Angiotensin receptor blockers
- CAP
Community-acquired pneumonia
- COPD
Chronic obstructive pulmonary disease
- MIMIC-IV
Medical Information Mart for Intensive Care IV
- MIT
Massachusetts Institute of Technology
- BIDMC
Beth Israel Deaconess Medical Center
- AKI
Acute kidney injury
- MBP
Mean blood pressure
- WBC
White blood cell
- RDW
Red cell distribution width
- ICD
International classification of diseases
- SD
Standard deviation
Authors’ contributions
HC designed the study and wrote the manuscript. HS and XC collected, analyzed, and interpreted the data. HC critically reviewed, edited, and approved the manuscript. All authors read and approved the final manuscript.
Funding
Not applicable.
Data availability
The datasets generated and/or analyzed during the current study are available in the MIMIC-IV database, https://mimic.mit.edu/docs/iv/.
Declarations
Ethics approval and consent to participate
The requirement of ethical approval for this was waived by the Institutional Review Board of People’s Hospital, Yanliang District, because the data was accessed from MIMIC-IV (a publicly available database). The need for written informed consent was waived by the Institutional Review Board of People’s Hospital, Yanliang District due to retrospective nature of the study. All methods were performed in accordance with the relevant guidelines and regulations.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Luo W, Xing R, Wang C. The effect of ventilator-associated pneumonia on the prognosis of intensive care unit patients within 90 days and 180 days. BMC Infect Dis. 2021;21(1):684. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Arn Ng Q. Yew Shuen Ang C, Shiong Chiew Y, Wang X, Pin Tan C, Basri Mat Nor M, Salwa Damanhuri N, Geoffrey Chase J: CAREDAQ: Data acquisition device for mechanical ventilation waveform monitoring. HardwareX. 2022;12:e00358. [DOI] [PMC free article] [PubMed]
- 3.Bassetti M, Mularoni A, Giacobbe DR, Castaldo N, Vena A. New Antibiotics for Hospital-Acquired Pneumonia and Ventilator-Associated Pneumonia. Semin Respir Crit Care Med. 2022;43(2):280–94. [DOI] [PubMed] [Google Scholar]
- 4.Semet C. The ongoing challenge of ventilator-associated pneumonia: epidemiology, prevention, and risk factors for mortality in a secondary care hospital intensive care unit. Infection Prevention in Practice. 2023;5(4):100320. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Hasan MJ, Sumi CD, Huq SMR, Anam AM, Rabbani R. Aerosolized Plus Intravenous Polymyxin B Versus Colistin in the Treatment of Pandrug-Resistant Klebsiella Pneumonia-mediated Ventilator-Associated Pneumonia: A Retrospective Cohort Study in Bangladesh. J Crit Care Med (Targu Mures). 2023;9(2):106–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Schreiber PW, Sax H, Wolfensberger A, Clack L, Kuster SP. The preventable proportion of healthcare-associated infections 2005–2016: Systematic review and meta-analysis. Infect Control Hosp Epidemiol. 2018;39(11):1277–95. [DOI] [PubMed] [Google Scholar]
- 7.Li Y, Liu C, Xiao W, Song T, Wang S. Incidence, Risk Factors, and Outcomes of Ventilator-Associated Pneumonia in Traumatic Brain Injury: A Meta-analysis. Neurocrit Care. 2020;32(1):272–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Rysz J, Franczyk B, Rysz-Górzyńska M, Gluba-Brzózka A. Pharmacogenomics of Hypertension Treatment. Int J Mol Sci. 2020;21(13):4709. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Christiansen CF, Heide-Jørgensen U, Rasmussen TB, Bodilsen J, Søgaard OS, Maeng M, Vistisen ST, Schmidt M, Pottegård A, Lund LC, et al. Renin-Angiotensin System Blockers and Adverse Outcomes of Influenza and Pneumonia: A Danish Cohort Study. J Am Heart Assoc. 2020;9(19): e017297. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.ElAbd R, AlTarrah D, AlYouha S, Bastaki H, Almazeedi S, Al-Haddad M, Jamal M, AlSabah S. Angiotensin-Converting Enzyme (ACE) Inhibitors and Angiotensin Receptor Blockers (ARB) Are Protective Against ICU Admission and Mortality for Patients With COVID-19 Disease. Front Med (Lausanne). 2021;8:600385. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Fang YP, Zhang X. A propensity score-matching analysis of angiotensin-converting enzyme inhibitor and angiotensin receptor blocker exposure on in-hospital mortality in patients with acute respiratory failure. Pharmacotherapy. 2022;42(5):387–96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Hasegawa D, Lee YI, Prasitlumkum N, Chopra L, Nishida K, Smith RL, Sato R. Premorbid angiotensin converting enzyme inhibitors or angiotensin II receptor blockers in patients with sepsis. Am J Emerg Med. 2022;62:69–77. [DOI] [PubMed] [Google Scholar]
- 13.Lin SY, Chang SS, Lin CL, Lin CC, Hsu WH, Chou CH, Chi CY, Lin CD, Tu CY, Hsu CY, et al. Association between angiotensin-converting enzyme inhibitors or angiotensin receptor blockers and community-acquired pneumonia: A nationwide population propensity-score matching study. Int J Clin Pract. 2021;75(10):e14476. [DOI] [PubMed] [Google Scholar]
- 14.Alexander DP, Nickman NA, Chhibber A, Stoddard GJ, Biskupiak JE, Munger MA. Angiotensin-converting enzyme inhibitors reduce community-acquired pneumonia hospitalization and mortality. Pharmacotherapy. 2022;42(12):890–7. [DOI] [PubMed] [Google Scholar]
- 15.Shah S, McArthur E, Farag A, Nartey M, Fleet JL, Knoll GA, Kim SJ, Garg AX, Jain AK. Risk of hospitalization for community acquired pneumonia with renin-angiotensin blockade in elderly patients: a population-based study. PLoS ONE. 2014;9(10):e110165. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Lai CC, Wang YH, Wang CY, Wang HC, Yu CJ, Chen L. Comparative effects of angiotensin-converting enzyme inhibitors and angiotensin II receptor blockers on the risk of pneumonia and severe exacerbations in patients with COPD. Int J Chron Obstruct Pulmon Dis. 2018;13:867–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Bryniarski P, Nazimek K, Marcinkiewicz J. Immunomodulatory Activity of the Most Commonly Used Antihypertensive Drugs-Angiotensin Converting Enzyme Inhibitors and Angiotensin II Receptor Blockers. Int J Mol Sci. 2022;23(3):1772. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Jiang JS, Wang LF, Chou HC, Chen CM. Angiotensin-converting enzyme inhibitor captopril attenuates ventilator-induced lung injury in rats. J Appl Physiol (1985). 2007;102(6):2098–103. [DOI] [PubMed] [Google Scholar]
- 19.Weber J, Tiriveedhi V, Takenaka M, Lu W, Hachem R, Trulock E, Patterson GA, Mohanakumar T. Inhibition of renin angiotensin aldosterone system causes abrogation of obliterative airways disease through inhibition of tumor necrosis factor-α-dependant interleukin-17. J Heart Lung Transplant. 2012;31(4):419–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Liu CL, Shau WY, Wu CS, Lai MS. Angiotensin-converting enzyme inhibitor/angiotensin II receptor blockers and pneumonia risk among stroke patients. J Hypertens. 2012;30(11):2223–9. [DOI] [PubMed] [Google Scholar]
- 21.Wang HC, Lin CC, Lau CI, Chang A, Kao CH. Angiotensin-converting enzyme inhibitors and bacterial pneumonia in patients with Parkinson disease. Mov Disord. 2015;30(4):593–6. [DOI] [PubMed] [Google Scholar]
- 22.Rafailidis PI, Matthaiou DK, Varbobitis I, Falagas ME. Use of ACE inhibitors and risk of community-acquired pneumonia: a review. Eur J Clin Pharmacol. 2008;64(6):565–73. [DOI] [PubMed] [Google Scholar]
- 23.Oosthuizen D, Sturrock ED. Exploring the Impact of ACE Inhibition in Immunity and Disease. J Renin Angiotensin Aldosterone Syst. 2022;2022:9028969. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Lopes RD, Macedo AVS, de Barros ESPGM, Moll-Bernardes RJ, Feldman A, D'Andréa Saba Arruda G, de Souza AS, de Albuquerque DC, Mazza L, Santos MF et al. Continuing versus suspending angiotensin-converting enzyme inhibitors and angiotensin receptor blockers: Impact on adverse outcomes in hospitalized patients with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)--The BRACE CORONA Trial. Am Heart J. 2020;226:49–59. [DOI] [PMC free article] [PubMed]
- 25.Silva IVG, de Figueiredo RC, Rios DRA. Effect of Different Classes of Antihypertensive Drugs on Endothelial Function and Inflammation. Int J Mol Sci. 2019;20(14):3458. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Caldeira D, Alarcão J, Vaz-Carneiro A, Costa J. Risk of pneumonia associated with use of angiotensin converting enzyme inhibitors and angiotensin receptor blockers: systematic review and meta-analysis. BMJ. 2012;345:e4260. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The datasets generated and/or analyzed during the current study are available in the MIMIC-IV database, https://mimic.mit.edu/docs/iv/.