Potential effects of regular use of antihypertensive drugs for in-hospital delirium in geriatric patients with trauma

Although the regular administration of antihypertensive drugs is a risk factor for falls in older adults, whether their anti-inflammatory effects confer a survival benefit in older adults remains unknown. This single-center retrospective cohort study examined patients with trauma aged ≥ 65 admitted to our hospital between January 2018 and December 2020. Patients who had not received antihypertensive drugs before admission (i.e., AHT(−) group) and those who had received the drugs (i.e., AHT(+) group) were compared using a 1:1 propensity score-matched analysis. The primary outcome was 28-day mortality, and the secondary outcomes were in-hospital mortality and the incidence of complications during the hospital stay. In total, 637 patients were analyzed. After propensity score matching, each study group had 223 patients. No significant difference was observed in the primary outcome (28-day mortality: AHT(−) group, 3.6% vs. AHT(+) group, 3.6%; adjusted relative risk: 1.00, 95% confidence interval (CI): 0.38–2.62); only the in-hospital incidence of delirium was significantly low in the AHT(+) group (25.1% vs. 13.9%; adjusted relative risk: 0.55, 95% CI: 0.37–0.82). Overall, the regular use of antihypertensive drugs did not affect outcomes in geriatric trauma patients; however, the incidence of delirium was reduced in those regularly receiving antihypertensive drugs.

Data collection. Collectors who were not among the authors used a medical chart review to collect data, including patient demographics, medical history, injury severity score (ISS), vital signs upon arrival (i.e., Glasgow coma scale score, systolic blood pressure, heart rate, respiratory rate, and body temperature), complications during hospital stay (i.e., acute kidney injury [AKI], acute respiratory distress syndrome [ARDS], arrhythmia, bleeding, cardiovascular disease, delirium, infections, stroke, and venous thromboembolism [VTE]), "do not attempt resuscitation (DNAR)" order, invasive treatments received within 24 h of admission (i.e., blood transfusion, emergency operation under general anesthesia, interventional radiology, tracheal intubation), and 28-day in-hospital mortality. The white blood cell (WBC) count and C-reactive protein (CRP) values collected as inflammation values were the highest three days after injury. Supervisors checked and anonymized the data collected. The authors subsequently analyzed the data.
Definition. We defined "antihypertensive medication" as a prescription of antihypertensive drugs and their administration within 28 days before the injury. Antihypertensive drugs were defined as angiotensin-converting enzyme (ACE) inhibitors, angiotensin II receptor blockers (ARBs), β blockers, calcium channel blockers (CCBs), and thiazides. We ascertained antihypertensive information from pharmacy notebooks or medical information documents from patients' family doctors.
Outcomes. The primary outcome was 28-day mortality, which was selected with reference to the results of previous studies [17][18][19][20] . The secondary outcomes were in-hospital mortality and the incidence of complications (i.e., AKI, ARDS, arrhythmia, bleeding, cardiovascular disease, delirium, infections, stroke, and VTE) during the hospital stay. These variables were selected based on their impact on the prognosis of trauma patients after admission.
Statistical analysis. AHT(-) group comprised individuals who did not take any antihypertensive drug (as per their medical records) for > 28 days before admission and those who were not prescribed antihypertensive drugs. AHT(+) group comprised those who had clear information of antihypertensive drugs and had not recorded information of drug withdrawal before admission. These AHT(−) and AHT(+) groups were compared using a 1:1 propensity score-matched analysis. Logistic regression analysis was used to estimate the propensity scores. The variables included in the model were age, sex, type of injury, vital signs on arrival, ISS, DNAR order, treatments received within 24 h upon admission, and comorbidities. We used a caliper with a width of 0.2 of the standard deviation of the logit of the propensity score. The balance between both groups was evaluated using the standardized mean difference (SMD), with an SMD > 0.1 indicating a significant imbalance. Data are presented as the median and interquartile range or numbers and percentage, as appropriate. Thereafter, outcomes were compared in the matched cohort. The relative risk (RR) was calculated, and the differences and 95% confidence intervals (CIs) were reported.
In the sub-analysis, to assess the effects of each antihypertensive drug on delirium, multiple-logistic regression analysis was performed in the cohort group. Odds ratio (OR) was calculated, and the differences and 95% CIs were reported. A univariate analysis was performed to investigate the relationship between delirium and the number of antihypertensive drugs.
As a post-hoc analysis to evaluate the association between delirium and inflammation, we collected inflammation data, and the Kruskal-Wallis and Steel-Dwass tests were performed in four groups according to the presence or absence of delirium and antihypertensive medication. The four groups were as follows: A, AHT(+) and delirium(+); B, AHT(−) and delirium(+); C, AHT(+) and delirium(−); D, AHT(−) and delirium(−).
We used Fisher's exact test, Mann-Whitney U test, and logistic regression analysis. Missing values were handled using pairwise methods. All statistical analyses were performed using EZR 1.54 (Saitama Medical Center, Jichi Medical University, Saitama, Japan), a graphical user interface for R 4.0.2 (The R Foundation for Statistical Computing, Vienna, Austria) 22 . For the sensitivity analysis, the original cohort (n = 637) as a first sensitivity analysis and an ISS ≥ 16 cohort as a second sensitivity analysis that was also performed were matched and evaluated with the same outcomes. Statistical significance was set at P < 0.05 or based on the 95% CI.
In the first sensitivity analysis performed in the original cohort (n = 627), there was no significant difference in 28-day mortality between the two groups (3.6% vs. 4.2%; RR: 1.17, 95% CI: 0.54-2.53). Like the main analysis, the incidence of delirium was also lower in the AHT(+) group (24.1% vs. 14.3%; RR: 0.64, 95% CI: 0.45-0.91) than in the AHT(−) group. In the second sensitivity analysis, 187 patients with an ISS ≥ 16 were identified from the study population, and 55 were matched in both groups. Some variables (age, sex, comorbidities of dementia, and diabetes mellitus) were not balanced between the two groups even after matching. Like the main analysis, there was no significant difference in 28-day mortality between the two groups (10.9% vs. 10.9%; adjusted RR: 1.00, 95% CI: 0.34-2.9). The incidence of delirium was also lower in the AHT(+) group (46.9% vs. 19.1%; adjusted RR: 0.39, 95% CI: 0.20-0.77) than in the AHT(−) group (Table 3).

Discussion
To our knowledge, the relationship between the use of antihypertensive drugs and mortality or morbidity rates after an injury has not yet been investigated. Thus, this is the first study to show that antihypertensive drug usage may not affect mortality in older patients with trauma, although it may reduce the risk of delirium during hospitalization.
Previous studies on sepsis showed that using antihypertensive drugs reduced the risk of in-hospital mortality [17][18][19][20] . However, our results were inconsistent. One possible reason for this is the small number of outcomes; it might have reduced the effect size and curbed the required difference. The in-hospital mortality rate in our study population was approximately 5%; however, previous studies reported 30-to 90-day mortality rates of approximately 20-40%. In retrospect, a study of patients with a cluster of severe traumas (ISS ≥ 16) may be helpful, because the mortality of each group was 10.9% in this cohort. The difference in mortality among patients Table 1. Baseline characteristics of older trauma patients with or without antihypertensive drugs. Categorical variables are presented as counts and percentages (%), whereas numerical variables are presented as median and interquartile ranges. The balance between the two groups was evaluated using the SMD; an SMD > 0.1 was considered significant. *Others included crush, sports, animal-related, and abuse. AHT Antihypertensive drug, DNAR Do not attempt resuscitation, SMD Standardised mean difference. www.nature.com/scientificreports/ with an ISS ≥ 16 was also examined, despite the insufficient detection ability of the results owing to the small study population size. Alternatively, the degree of systemic inflammation may have been small, and the antiinflammatory effect of antihypertensive drugs may not have affected the mortality of the patients with trauma. Delirium is a clinically influential event that commonly occurs in hospitalized geriatric patients. No evidence is available to support the direct effects of antihypertensive drugs on the suppression of delirium risk because there has been no previous study on the direct relationship between the use of antihypertensive drugs and delirium 23 . Harrison et al. reported the difference in the rate of delirium depending on the type of antihypertensive drug. However, the comparison was not examined with patients who did not take antihypertensive drugs 23 . Zaal et al. investigated the risk factors for developing delirium in patients in the intensive care unit, and hypertension was mentioned as one of the factors. However, it was unclear whether patients had been taking antihypertensive drugs 24 .
One possibility of reducing the incidence of delirium is the anti-inflammatory effect of antihypertensive drugs 16 . Moreover, a systemic inflammatory response is induced after trauma 25 . Conversely, systemic inflammation and endothelial dysfunction are involved in delirium [26][27][28] . In our post-hoc analysis, the group with antihypertensive drugs did not suppress the increased inflammatory response after injury. Furthermore, the inflammatory response trend was significantly high in the group complicated with delirium. The fact that the Table 2. Characteristics of the study population after propensity score matching. Categorical variables are presented as counts and percentages (%), whereas numerical variables are presented as median and interquartile ranges. The balance between the two groups was evaluated using the SMD; an SMD > 0.1 was considered significant. *Others included crush, sports, animal-related, and abuse. AHT Antihypertensive drug, DNAR Do not attempt resuscitation, SMD Standardised mean difference.  AHT better AHT worse Figure 2. Primary and secondary outcome events. Relative risks of each outcome. Horizontal bars show 95% CIs. A thick vertical line represents a relative risk of 1.0, indicating no significant difference in outcome between AHT(−) and AHT(+) groups. † The results were statistically significant based on 95% CI, and a black circle plotted on the horizontal bar. Causes of death in the ATH(-) group were as follows: severe brain injury (n = 6), heart failure (n = 1), sepsis (n = 3), natural death (n = 1). Causes of death in the ATH(+) group were as follows: severe brain injury (n = 3), haemorrhage shock (n = 1), aortic dissection (n = 1), heart failure (n = 1), multiple organ failure (n = 1), sepsis (n = 2), malignant tumor (n = 2). AHT, antihypertensive drug; ARDS, acute respiratory distress syndrome; CI, confidence interval. www.nature.com/scientificreports/ values of the inflammatory response caused by trauma were the same between the AHT(+) and AHT(−) groups means that the protective effects of antihypertensive drugs on endothelial function have been speculated to suppress endothelial dysfunction, resulting in a reduced risk of delirium. One basis for this hypothesis is the incidence of VTE, which has also been reported to result from an acute inflammatory response and endothelial damage after trauma 29 . Although we did not find any statistically significant difference, the risk of VTE tended to be lower in the AHT(+) group. In this study, it was not possible to show the difference in the effects of delirium depending on the type or number of antihypertensive drugs administered. Studies on larger cohorts would be needed to better understand this. Consequent complications, such as delirium 30,31 , worsen the prognosis of geriatric patients with trauma 8 . Therefore, the prevention of delirium is important. Although the use of antihypertensive drugs before injury may have a preventive effect on delirium, whether the re-administration or the initiation of antihypertensive drugs after hospitalization exhibits a preventive effect on delirium remains unknown. In polytrauma cases, older patients may likely experience hemorrhagic shock during the acute phase or may experience sudden exacerbation because of infection or re-bleeding, even after they have overcome the acute phase. Furthermore, the use of as-needed antihypertensive medication is associated with increased mortality and prolonged hospital stay 32 . Thus, there is no clear recommendation regarding resuming the oral administration of antihypertensive drugs. Further studies should investigate the appropriate time to resume the use of antihypertensive drugs and the mechanism underlying the preventive effect on delirium.
This study had some limitations. First, this was a retrospective cohort study conducted at a single center. Thus, we could not eliminate bias in patient backgrounds. Additionally, it is difficult to show a causal relationship in a retrospective study owing to the difficulty in isolating confounders; therefore, only the statistical relationship of association could be stated. Second, we used propensity score matching, which could not evaluate the effect of unknown variables that might have led to biased results. Third, data on patient compliance until the day before the injury are unavailable due to the retrospective study design. If patients have histories of oral antihypertensive drug administration without actually taking them, the interpretation of the results may be altered. Fourth, we could not cover all types of antihypertensive drugs. We narrowed down the section to five types, ACE-inhibitors, ARBs, CCBs, β blockers, and Thiazides, because we referred to previous studies [17][18][19][20] . There are many types of drugs classified as antihypertensive drugs, and there are other types besides the five antihypertensive drugs, such as α-blockers. Fifth, we could not investigate the cause of hypertension since it was difficult to obtain information due to the study's retrospective design. Lastly, various antihypertensive drugs have been insufficiently The black thick horizontal bars are the mean value, and gray boxes are SDs. † The results were statistically significant based p-value < 0.05. We performed the Kruskal-Wallis test, a non-parametric analysis, because the homoscedasticity of the data rejected. (a) Significant differences were observed between groups A and C and groups A and D (Kruskal-Wallis test: p = 0.004). (b) Significant differences were observed between groups A and C, A and D, B and C, and B and D (Kruskal-Wallis test: p < 0.001). AHT, antihypertensive drug; CI, confidence interval; CRP, C-reactive protein; SD, standard deviation; WBC, white blood cell. www.nature.com/scientificreports/ investigated, and it is unclear which class of drugs had affected the results. Our cohort was too small to investigate the effects of various antihypertensive drugs individually.
In conclusion, the regular use of antihypertensive drugs did not affect outcomes in geriatric patients with trauma. However, the incidence of delirium might be reduced in patients with regular use of antihypertensive drugs. Our findings contribute to a growing body of evidence for improved care after admission for this vulnerable population and may provide new insights into medication management following admission.