Sepsis, representing a complex immune response to infections and progressing to an unregulated condition, is a major cause of mortality in critical care worldwide (1–3). Hospital mortality increases substantially when sepsis progresses to septic shock. Because of gradually aging populations and regularly renewed guidelines, the reported incidence of sepsis has shown an increasing trajectory (3), and sepsis outcomes might be improved through prior use of certain common medications, such as lipid-lowering agents, especially in patients with multiple comorbidities who are prone to infectious diseases (4).
The hypertension management guidelines of the Journal of the American Heart Association recommend angiotensin-converting enzyme inhibitors (ACEIs) and angiotensin receptor blockers (ARBs) as the first-line treatments for hypertensive patients with concomitant chronic kidney disease (CKD) or diabetes mellitus (DM) (5). Furthermore, as these CKD or diabetic patients are relatively immunocompromised, they are much more prone to severe infections and therefore have increased hospital mortality and morbidity (6, 7).
In addition to calcium channel blockers (CCBs), ACEIs and ARBs are also frequently prescribed to treat hypertension. ACEIs and ARBs inhibit the synthesis of angiotensin II and block the effects of angiotensin II, respectively, inducing vasodilatation and decreasing the glomerular filtration rate (GFR). Angiotensin II thus plays an important role in the maintenance of GFR, especially in hypovolemia or hypotension (8). An increasing number of studies have debated the impact of ACEIs and ARBs on the development of sepsis and related hospital outcomes (8–12).
In the present study, we used the unique database of a sepsis cohort that included 223,560 patients with a first admission for sepsis from the National Health Insurance Research Database (NHIRD) of Taiwan. We aimed to examine the impact of preadmission antihypertensive drug use on the hospital mortality of sepsis and septic shock, and compare the various effects of all types of antihypertensive drugs, especially ACEIs and ARBs.
Data sources and study participants
We conducted this retrospective observational study using the unique database of a sepsis cohort between 1999 and 2013 from the NHIRD of Taiwan. The National Health Insurance Program was launched by the National Health Insurance Administration (NHIA) in 1995 in Taiwan and currently provides coverage for more than 23.03 million residents (>99% of the entire population). The NHIA releases deidentified patient information and claims data to the National Health Research Institute for the NHIRD. The confidentiality and credibility of these data are strictly maintained in accordance with the NHIRD regulations, with documented high quality in previous studies (13–16).
We defined the admission date of the first hospitalization for sepsis as the index date. Comorbidities were also defined using the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM); diagnoses made within a 1-year period before the index date were considered the underlying comorbidities of a patient. The patients were classified as using certain drugs if they took them for more than 1 week within a 3-month period before the index date.
The database in the NHIRD contains deidentified basic demographic information, disease diagnoses, prescriptions, procedures, and examinations for each enrollee. Each enrollee in the NHIRD is coded with an encrypted identifier, which can be used to link each patient's complete medical data (17). Because of the above advantage, we can track medications and comorbidities, among others, in hospitalization or outpatient visit records before or after the index hospitalization for sepsis.
However, because of the limitation of the database, we can trace each patient back to only the beginning of the database, that is, 1999. The first hospitalization for sepsis since 1999 was therefore defined as the first episode. Because patients become increasingly weaker with each repeated sepsis attack, only the first episode of sepsis for each enrollee was included in further analysis to avoid inclusion of a second or subsequent episode.
Credibility of the disease coding of sepsis
Our study and comparison cohorts were retrieved between 1999 and 2013. During the study period, the definitions of sepsis and septic shock were based on the systemic inflammatory response syndrome “SIRS” criteria. Furthermore, the national health insurance system in Taiwan is a single-payer design and supervised by the National Health Institute, which regularly monitors the accuracy of coding. Throughout the study period, the ICD-9-CM was used. The accuracy of sepsis coding (ICD-9-CM: 038.9) was also validated in a high-quality journal (14), and many important studies on sepsis were published using the same claims database (14, 18–22). The infection sites of sepsis origin were also retrieved using the diagnosis codes of ICD-9-CM (Supplement Table 1, http://links.lww.com/SHK/A890). We classified the infection site into the following categories: central nervous system, respiratory system, cardiovascular system, gastrointestinal/biliary tract system, genitourinary system, soft tissue/musculoskeletal system, device related, and unclassified.
Validation for septic shock
The identification of patients with septic shock was validated by analysis of randomly selected samples (100 patients) from the claims database of Taichung Veterans General Hospital (a 1,520-bed tertiary referral medical center in central Taiwan) for 1999 to 2013. The content of this claims database is used for reimbursement and is similar to that of the NHIRD. An emergency physician and an intensivist (M-SH and S-YH) independently reviewed the clinical data and the medical records for all selected samples. All instances of disagreement were resolved by consensus. We defined a septic shock patient as someone who had shock induced by a certain infection that necessitated vasopressor agent support even after adequate fluid resuscitation. To define “how sepsis patients receive adequate fluid resuscitation,” we referred to “nursing records” to determine whether fluid resuscitation was administered with the goal of restoring sepsis-induced tissue hypoperfusion or hypotension before the administration of vasopressors. The initial fluid resuscitation should meet the guideline of 20 mL/kg of crystalloid or albumin equivalent, except for contraindications such as acute respiratory distress syndrome. We could therefore define a patient with septic shock by following the above criterion. Ultimately, a patient with the main discharge diagnosis of sepsis plus the claims of a certain vasopressor agent (including dopamine and norepinephrine) was defined as admitted with the diagnosis of septic shock.
To assess interobserver agreement on the identification of septic shock, we calculated the Cohen kappa coefficient (κ) to express probability beyond chance. During the period from January 1, 1999, through December 31, 2013, a total of 10,734 patients at Taichung Veterans General Hospital had claims data indicating septic shock according to the above definition; 100 of these patients were randomly selected for validation. Among these 100 patients, the diagnosis of septic shock was confirmed for 97 and unconfirmed for 3 (sensitivity, 97.0%). Interobserver agreement was excellent (κ=0.887). The validation process was approved by the Institutional Review Board of Taichung Veterans General Hospital (CE18102A).
Antihypertensive drug use
The unique sepsis cohort included 223,560 patients who had at least one admission for sepsis from 1999 to 2013. For patients with repeated admissions for sepsis, only data for the first admission were used. Antihypertensive drug users were defined as using certain antihypertensive drugs for more than 1 week within a 3-month period before the index date. Antihypertensive drug nonusers were retrieved from the same database and included patients who did not take any antihypertensive drugs within 3 months before the index date and were frequency matched with the antihypertensive drug users by age, sex, and index year.
In Taiwan, antihypertensive drugs are not available over-the-counter, and a physician's decision to prescribe antihypertensive drugs should not only follow the treatment guidelines for specific diseases, but also the payment regulations by the NHIA. If prescriptions are against the rules, the NHIA may not only refuse to pay the medical fee, but may also punish the physicians with a maximal 100-fold rebound (because the national health insurance program is a single-payer, compulsory insurance coverage policy in Taiwan and the NHIA has the full authority to control all medical facilities and healthcare professionals). Therefore, selection bias in this study was minimal.
Healthy user bias
The “healthy user bias” has frequently been discussed in this type of study, that is, “certain drugs in sepsis outcomes,” and describes a type of sampling bias. Specifically, antihypertensive drug users may have more health-seeking behaviors (these patients are more likely to see a doctor on a regular basis, exercise, stop smoking, receive vaccines, and adhere to treatments) that help them stay healthier than antihypertensive drug nonusers.
In fact, information on health-seeking behavior is difficult to obtain and adjust in the large claims database and therefore may “overexaggerate” the therapeutic effect of a certain drug or procedure in observational studies (23, 24). Because the health-seeking tendency correlates strongly with socioeconomic status and urbanization level, we used the above two variables as surrogates in the regression model to adjust for “health user bias” (25, 26).
In this study, the comparison cohort (antihypertensive drug nonusers) was composed of patients with hypertension without the need for corresponding treatment (13,162 patients, 39.63%) and patients without hypertension (20,051 patients, 60.37%).
Although information on individual income could not be directly obtained in the NHIRD, the insurance premium paid was a useful surrogate for household income level, and this method has widely been adopted in NHIRD-associated studies. Specifically, within the national health insurance system of Taiwan, the fee for an individual's insurance premium is proportional to his or her work salary (27).
The algorithm used for participant selection for the study (antihypertensive drug users) and comparison (antihypertensive drug nonusers) cohorts is shown in Supplement Figure 1, http://links.lww.com/SHK/A891.
As the NHIRD contains deidentified secondary data for research, our study was exempted from the requirement of informed consent from participants. This study was approved by the Institutional Review Board of China Medical University (CMUH104-REC2-115).
A frequency matching method was used to retrieve the study and comparison cohorts and match participants in each cohort with a 1:1 ratio according to age, sex, and index year (defined as the calendar year of the index date). Differences in demographic characteristics, baseline comorbidities (including hyperlipidemia, congestive heart failure [CHF], chronic obstructive pulmonary disease [COPD], chronic liver disease [CLD], CKD, ischemic heart disease [IHD], and cancer), medications (including aspirin, clopidogrel, statins, steroids, and immunosuppressants), infection sites of sepsis origin, and hospital outcomes (length of hospital stay, intensive care unit [ICU] admission, septic shock, and 28-day and total hospital mortality) were examined by using the chi-square test and two-sample t test.
Odds ratios (ORs) with 95% confidence intervals (95% CIs) for total hospital mortality were calculated for each variable in the logistic regression model. Adjusted ORs were further obtained after adjustment for possible confounders, including age, sex, insurance premium, urbanization level, and comorbidities in multivariate analysis. Stratification analysis was conducted according to age, sex, and septic shock status during hospitalization.
A joint effect analysis was performed to analyze the synergistic effect of different types of antihypertensive drugs, including ACEIs, ARBs, beta-blockers, CCBs, and diuretics, on the outcome of total hospital mortality for sepsis. Because there were 5 categories of antihypertensive drugs, 25 different regimens of drug combinations were evaluated using nonuse of antihypertensive drugs as a reference. The adjusted OR for each regimen was calculated by logistic regression after adjusting for age, sex, insurance premium, urbanization level, and comorbidities.
The statistical analyses were performed using the SAS 9.4 statistical package (SAS Institute Inc., Cary, NC). A P value of 0.05 was set as indicating significance.
Initially, 223,560 sepsis patients were included, with 186,307 (84.82%) antihypertensive drug users and 33,346 (15.18%) antihypertensive drug nonusers. After frequency matching for age, sex, and the index year (the calendar year), 33,213 antihypertensive drug users with sepsis and an equal number of antihypertensive drug nonusers with sepsis were included for further analysis.
Before and after matching, a greater proportion of the antihypertensive drug users with sepsis than the antihypertensive drug nonusers with sepsis had comorbidities, such as hyperlipidemia, CHF, COPD, CLD, CKD, IHD, and cancer (all P < 0.0001), and a greater proportion of the former than the latter accepted medications, including aspirin, clopidogrel, statins, steroids, and immunosuppressants (all P < 0.0001) (Table 1). Furthermore, compared with the antihypertensive drug nonusers with sepsis, a greater proportion of antihypertensive drug users with sepsis were diagnosed with septic shock (4.36% vs. 2.31%, P < 0.0001) during the index hospitalization and had higher 28-day (30.77% vs. 24.01%, P < 0.0001) and total hospital (38.42% ss. 24.57%, P < 0.0001) mortality rates.
In the logistic regression model, after adjusting for age, sex, insurance premium, urbanization level, and baseline comorbidities, preadmission antihypertensive drug use was associated with an increased adjusted OR for total hospital mortality (adjusted OR = 1.80 [95% CI, 1.74–1.87], P < 0.0001) compared with preadmission antihypertensive drug nonuse (Table 2). Surprisingly, preadmission antihypertensive drug use was associated with a decreased adjusted OR for total hospital mortality in septic shock (adjusted OR = 0.66 [95% CI, 0.55–0.80], P < 0.0001) (Table 3).
Among sepsis patients, compared with preadmission antihypertensive drug nonuse subjects, both males and females and those in different age subgroups with preadmission antihypertensive drug use had increased adjusted ORs for total hospital mortality (Supplement Figure 2, http://links.lww.com/SHK/A892).
In the stratification analysis performed according to the type of antihypertensive drugs, only ACEIs and ARBs exhibited a decreased adjusted OR for total hospital mortality under both sepsis and septic shock conditions (i.e., ACEI users compared with ACEI nonusers; ARB users compared with ARB nonusers) (ACEI use, adjusted OR for total hospital mortality = 0.93 [95% CI, 0.88–0.98], P = 0.0085 in sepsis and adjusted OR for total hospital mortality = 0.85 [95% CI, 0.81–0.90], P < 0.0001 in septic shock) (ARB use, adjusted OR for total hospital mortality = 0.85 [95% CI, 0.81–0.90), P < 0.0001 in sepsis and adjusted OR for total hospital mortality = 0.62 [95% CI, 0.49–0.79], P < 0.0001 in septic shock) (Table 3).
We further conducted a joint effect analysis to evaluate the synergistic effects of different combinations of antihypertensive drugs on total hospital mortality. The results showed that the use of ACEIs was associated with a decreased adjusted OR for total hospital mortality, except when ACEIs were combined with diuretics (Fig. 1); ACEI use alone had a decreased adjusted OR for total hospital mortality (adjusted OR = 0.67, 95% CI, 0.51–0.88, P = 0.0043). Similar results were also observed for ARB use; ARB use alone had a decreased adjusted OR for total hospital mortality (adjusted OR = 0.62, 95% CI, 0.45–0.85, P = 0.0029). CCB use alone and beta-blocker use alone were also associated with decreased ORs for total hospital mortality but without significance (adjusted OR = 0.94 [95% CI, 0.84–1.04], P = 0.2181 and adjusted OR = 0.89 [95% CI, 0.76–1.04], P = 0.1484, respectively).
In this study, we demonstrated that preadmission antihypertensive drug use was associated with an increased risk of total hospital mortality, but that the risk decreased under septic shock conditions. Furthermore, preadmission ACEI or ARB use was associated with a decreased risk of total hospital mortality under both sepsis and septic shock conditions. These results support not only findings from preclinical studies regarding the physiologically protective effects of ACEIs and ARBs in sepsis, but also the suggestion that ACEIs and ARBs may be good alternatives for antihypertensive management during sepsis, for which there has been only inconclusive data.
In an animal study by Laesser et al. in 2004, pretreatment with candesartan increased cardiac output, portal venous blood flow, arterial standard base excess, and mixed venous oxygen saturation during Escherichia coli infections (9). In 2007, Lund et al. demonstrated that enalapril and L-158809 (a type of ARB) reduce superoxide levels and further enhance relaxation of the aorta by a mechanism related to the actions of acetylcholine. This finding suggested that blocking the renin-angiotensin system (RAS) by ACEIs or ARBs relieves oxidative stress and endothelial dysfunction after endotoxin injection (11).
Regarding human studies, in 2007, Mortensen et al. conducted a retrospective national cohort study and found that preadmission ARB use was significantly associated with decreased 30-day hospital mortality in sepsis patients (10). Later, in 2010, Doerschug et al. evaluated ICU patients and demonstrated the RAS to be activated in clinically severe sepsis and correlate with acute organ injury and mortality. Despite macrovascular fluid resuscitation, systemic RAS activation persisted in many sepsis patients; therefore, the microvascular response to ischemia was continuously impaired, further inducing vital organ dysfunction (12).
In a case–control study conducted by Dial et al. in 2014, ARB use in hypertensive patients was not associated with an increased risk of sepsis compared with the use of other classes of antihypertensive drugs. In addition, ARB use was not associated with an increased risk of renal failure or death during sepsis (8). Despite growing evidence, further studies are still necessary to confirm the benefit of ACEIs and ARBs in sepsis patients.
In this study, we conducted subgroup analysis according to septic shock status and observed that in septic shock patients, most antihypertensive drugs were associated with better hospital outcomes regarding survival. Because data regarding when ACEI or ARB or other antihypertensive drugs were held and restarted were unavailable, we speculate that this phenomenon may be explained by a rebounding effect of discontinuing antihypertensive drugs during shock. However, the protective effect of ACEIs and ARBs under even nonseptic shock conditions may be explained by other complex mechanisms, such as relieving oxidative stress and decreasing endothelial dysfunction.
In 2017, Wiewel et al. demonstrated in a prospective study that prior use of CCBs was associated with decreased mortality in patients with sepsis following ICU admission (28). In the same year, Lee et al. also reported in a nationwide cohort study that preadmission use of CCBs was associated with improved outcomes in sepsis patients (20). In the retrospective study by Zheng et al. (2017), preadmission use of CCBs was associated with reduced risks of developing respiratory failure, bacteremia, and severe sepsis in patients admitted with pneumonia but was not associated with total hospital mortality (29). In our study, preadmission CCB use was beneficial to septic shock patients with regard to total hospital mortality but not to the entire sepsis cohort (all sepsis patients, adjusted OR = 1.21 [95% CI, 1.17–1.26, P < 0.0001], septic shock patients, adjusted OR = 0.64 [95% CI, 0.53–0.77, P < 0.0001]). Furthermore, in joint effect analysis of the entire sepsis population, the adjusted OR of CCB use alone (not combined with any other antihypertensive drug) for total hospital mortality was 0.94 (95% CI, 0.84–1.04, P = 0.2181). In conclusion, despite growing evidence from animal and clinical studies, the benefit of CCBs under sepsis conditions remains controversial (30–34). Further clinical studies with a stratification of sepsis severity may help to resolve this issue.
Circulatory shock is a medical emergency that requires rapid interventions (35, 36), and the Federal Drug Administration has approved the use of angiotensin II agents (novel, noncatecholamine agents) as auxiliary therapeutic vasopressor agents for septic shock patients (37). In 2018, Wakefield et al. showed that the use of angiotensin II agents in vasodilatory shock patients was able to reduce the background catecholamine dose and facilitate achievement of a blood pressure goal (35). Moreover, recent advancements have shown that the use of angiotensin II agents in septic shock patients may represent an alternative treatment strategy, especially under vasodilatory conditions that do not respond to high-dose vasopressors (37, 38). We speculate that ACEI and ARB use is associated with a rebounding reaction. That is, discontinuation of preadmission ACEI and ARB use during septic shock induces a rebounding of angiotensin activity that was previously inhibited by ACEIs and ARBs and therefore increases the survival of patients with septic shock. Nonetheless, further clinical and basic studies are still needed to clarify the positive and negative effects of angiotensin II use in septic shock patients.
Originally, “healthy user bias” referred to comparison of patients (the study cohort) receiving a certain medication for “primary” prevention and the other patients (the comparison cohort) not receiving this medication. The frequently used example is hormone replacement therapy (HRT) in postmenopausal women to reduce the risk of cardiovascular disease in 1985 (27). Although several observational studies supported this finding at that time, it was abandoned after subsequent randomized control trials. The diverse results were caused by unmeasurable variables, that is, “healthy user bias,” whereby the women who took HRT were more likely to engage in healthier behaviors, such as a healthy diet, abstinence from alcohol, regular exercise, and maintenance of an adequate body weight, compared with the non-HRT users.
Another popular example concerns the use of statins and the reduced risk of hip fracture, Alzheimer's disease, and sepsis (39–42). The following description is cited from Majumdar et al. in 2006. “Specifically, preventive therapy, such as use of statins, is more likely to be prescribed to relatively healthy or health-seeking patients (healthy users) … That is not to say that healthy users do not have comorbidities. For example, consider two 70-year-old men discharged from the hospital after myocardial infarction. The first patient sees his doctor regularly, stops smoking, loses weight, starts exercising, receives his immunizations, and, after making a request for statins, adheres to treatment. The second patient does not see a doctor, continues to smoke, does not change his diet or lifestyle, and, after discharge, does not even fill his statin prescription. Administrative databases would consider both to be equivalent 70-year-old male survivors of myocardial infarction, but it is apparent that the first patient is a healthy user of statins, whereas the other patient is not. Irrespective of statin use, it might be surmised that the first patient would have better health-related outcomes (42).” Based on the above descriptions, we know that the comparison cohort is not restricted to patients with or without a certain disease but should be relatively comparable between study and comparison cohorts, which is also the reason we conducted frequency matching and further adjusted the urbanization level and individual income (23).
In a meta-analysis by Psaty et al. published in JAMA in 1997, the authors conclude that until large long-term clinical trials of CCBs and ACEIs are conducted, the available evidence recommends diuretics and beta-blockers at low doses as first-line agents (43). In 2003, Psaty et al. reported that low-dose diuretics were the most effective treatment for preventing cardiovascular comorbidity and mortality (44), and in 2007, the European Society of Cardiology/European Society of Hypertension (ESC/ESH) guidelines continued to recommend that thiazide diuretics should be considered for the initiation and maintenance of antihypertensive treatment (45). Later, in Diabetes Care by Ehud Grossman et al. (2011), thiazide-type diuretics were considered in the debate regarding diuretics for hypertension treatment to be at least as effective as CCBs, ACEIs, and beta-blockers in reducing cardiovascular events. In addition, thiazide diuretics were found to be particularly effective in preventing stroke and heart failure in hypertensive patients, especially in elderly and very elderly patients (46). The British Hypertension Society also recommends that diuretics and CCBs should be first-line drugs in hypertensive patients aged at least 55 years, whereas ACEIs (or ARBs) should be first-line drugs in hypertensive patients younger than 55 years of age (47).
As our sepsis study cohort was retrieved between 1999 and 2013, common use of thiazide or thiazide-like diuretics (alone or with other antihypertensive drugs) in hypertensive patients was not unexpected. Furthermore, the mean age of the patients with diuretic use alone was 58.71 ± 17.70 years. Diuretics may therefore be prescribed to antihypertensive patients more often.
Regardless, some unmeasured associated variables in the group of patients with diuretic use (alone or combined with other antihypertensive drugs) may still exist. Because another important effect of diuretics in clinical practice is treatment of CHF, we conducted sensitivity analysis excluding all sepsis patients with CHF in both the study and comparison cohorts. However, the adjusted OR of diuretics use for total hospital mortality remained increased (adjusted OR=2.62, 95% CI, 2.51–2.74). Further study is needed to clarify the mechanism involved.
This study has several strengths. First, the study was conducted using a unique cohort including 223,560 sepsis patients within a 15-year period, and the patients had varying degrees of disease severity from sepsis to septic shock. Second, to explore the future use of NHIRD in critical care, we conducted a validation study, enrolling 100 patients admitted for septic shock, using the hospital claims database. The study results supported the definition of “the septic shock condition” by concomitantly using the main discharge diagnosis code of sepsis (038.9) plus claims data for vasopressor agents (dopamine and norepinephrine). Third, this study will help clinicians to better understand the use and mechanism of newly approved “angiotensin II agents.”
The limitations of this study include the following. (1) There was a lack of laboratory data, such as inflammatory markers, which is inevitable in large epidemiologic studies. (2) The definition of sepsis in this study relied on the sepsis-2 definition with the basis of “SIRS” rather than the sepsis-3 definition of “Sequential Organ Failure Assessment, SOFA.” However, this critical change in the new sepsis definition should not substantially interfere with the study result if the entire population was retrieved using the same definition. Cortes-Puch et al. in 2016 commented in the American Journal of Respiratory and Critical Care(45) “Moreover, these previous definitions and the SIRS criteria have been widely adopted for use at the bedside and for hospital- and statewide quality improvement initiatives worldwide. Numerous controlled trials have relied on them, and this scientific database should not be discarded until unequivocal evidence indicates that superior diagnostic criteria exist.” (3) Despite the regular quality check and internal control for disease coding by the national health institute, missing or even wrong disease coding is inevitable. Therefore, further prospective clinical studies are needed before completely adopting the result of this study.
In this study, we demonstrated that preadmission antihypertensive drug use was associated with a decreased risk of total hospital mortality in septic shock but not under less severe nonshock conditions. The validation process in this study helped define septic shock in a large epidemiologic database without laboratory data. Furthermore, preadmission ACEI or ARB use was associated with decreased total hospital mortality, regardless of sepsis or septic shock conditions. The use of ACEIs or ARBs in hypertensive patients with a high risk of sepsis may be regarded as a first choice for antihypertensive medications. The study also provides a reference for newly approved angiotensin II drugs in septic shock patients.
The authors thank the Biostatistics Task Force of Taichung Veterans General Hospital, Taichung, Taiwan, Republic of China, for their assistance and advice with the statistical analyses, and also thank Dr. Sung-Yuan, Hu for his help in the validation process.
1. Vincent JL, Marshall JC, Namendys-Silva SA, Francois B, Martin-Loeches I, Lipman J, Reinhart K, Antonelli M, Pickkers P, Njimi H, et al. Assessment of the worldwide burden of critical illness: the intensive care over nations (ICON) audit. Lancet Respir Med
2 (5):380–386, 2014.
2. Fleischmann C, Scherag A, Adhikari NK, Hartog CS, Tsaganos T, Schlattmann P, Angus DC, Reinhart K. Assessment of global incidence and mortality of hospital-treated sepsis
. Current estimates and limitations. Am J Respir Crit Care Med
193 (3):259–272, 2016.
3. Singer M, Deutschman CS, Seymour CW, Shankar-Hari M, Annane D, Bauer M, Bellomo R, Bernard GR, Chiche JD, Coopersmith CM, et al. The Third International Consensus Definitions for Sepsis
and Septic Shock
315 (8):801–810, 2016.
4. Lai CC, Lee MT, Lee SH, Hsu WT, Chang SS, Chen SC, Lee CC. Statin treatment is associated with a decreased risk of active tuberculosis: an analysis of a nationally representative cohort. Thorax
71 (7):646–651, 2016.
5. Kovell LC, Ahmed HM, Misra S, Whelton SP, Prokopowicz GP, Blumenthal RS, McEvoy JW. US hypertension management guidelines: a review of the recent past and recommendations for the future. J Am Heart Assoc
2015; 4 (12): pii: e002315.
6. Mansur A, Mulwande E, Steinau M, Bergmann I, Popov AF, Ghadimi M, Beissbarth T, Bauer M, Hinz J. Chronic kidney disease
is associated with a higher 90-day mortality than other chronic medical conditions in patients with sepsis
. Sci Rep
7. Bomberg H, Kubulus C, List F, Albert N, Schmitt K, Graber S, Kessler P, Steinfeldt T, Standl T, Gottschalk A, et al. Diabetes: a risk factor for catheter-associated infections. Reg Anesth Pain Med
40 (1):16–21, 2015.
8. Dial S, Nessim SJ, Kezouh A, Benisty J, Suissa S. Antihypertensive agents acting on the renin-angiotensin system
and the risk of sepsis
. Br J Clin Pharmaco
78 (5):1151–1158, 2014.
9. Laesser M, Oi Y, Ewert S, Fandriks L, Aneman A. The angiotensin II receptor blocker candesartan improves survival and mesenteric perfusion in an acute porcine endotoxin model. Acta Anaesthesiol Scand
48 (2):198–204, 2004.
10. Mortensen EM, Restrepo MI, Copeland LA, Pugh JA, Anzueto A, Cornell JE, Pugh MJ. Impact of previous statin and angiotensin II receptor blocker use on mortality in patients hospitalized with sepsis
27 (12):1619–1626, 2007.
11. Lund DD, Brooks RM, Faraci FM, Heistad DD. Role of angiotensin II in endothelial dysfunction induced by lipopolysaccharide in mice. Am J Physiol Heart Circ Physiol
293 (6):H3726–H3731, 2007.
12. Doerschug KC, Delsing AS, Schmidt GA, Ashare A. Renin-angiotensin system
activation correlates with microvascular dysfunction in a prospective cohort study of clinical sepsis
. Crit Care
14 (1):R24, 2010.
13. Wu CY, Chen YJ, Ho HJ, Hsu YC, Kuo KN, Wu MS, Lin JT. Association between nucleoside analogues and risk of hepatitis B virus-related hepatocellular carcinoma recurrence following liver resection. JAMA
308 (18):1906–1914, 2012.
14. Chao PW, Shih CJ, Lee YJ, Tseng CM, Kuo SC, Shih YN, Chou KT, Tarng DC, Li SY, Ou SM, et al. Association of postdischarge rehabilitation with mortality in intensive care unit
survivors of sepsis
. Am J Respir Crit Care Med
190 (9):1003–1011, 2014.
15. Chang SL, Huang YL, Lee MC, Hu S, Hsiao YC, Chang SW, Chang CJ, Chen PC. Association of varicose veins with incident venous thromboembolism and peripheral artery disease. JAMA
319 (8):807–817, 2018.
16. Chang SH, Wen MS, Kuo CF. Drug interactions with non-vitamin K oral anticoagulants-reply. JAMA
319 (8):830–831, 2018.
17. Lin LY, Warren-Gash C, Smeeth L, Chen PC. Data resource profile: the National Health Insurance Research Database
). Epidemiol Health
18. Shen HN, Lu CL, Yang HH. Epidemiologic trend of severe sepsis
in Taiwan from 1997 through 2006. Chest
138 (2):298–304, 2010.
19. Ou SM, Chu H, Chao PW, Lee YJ, Kuo SC, Chen TJ, Tseng CM, Shih CJ, Chen YT. Long-term mortality and major adverse cardiovascular events in sepsis
survivors. A nationwide population-based study. Am J Respir Crit Care Med
194 (2):209–217, 2016.
20. Lee CC, Lee MG, Lee WC, Lai CC, Chao CC, Hsu WH, Chang SS, Lee M. Preadmission use of calcium channel blocking agents is associated with improved outcomes in patients with sepsis
: a population-based propensity score-matched cohort study. Crit Care Med
45 (9):1500–1508, 2017.
21. Lee CC, Lee MG, Hsu TC, Porta L, Chang SS, Yo CH, Tsai KC, Lee M. A population-based cohort study on the drug-specific effect of statins on sepsis
153 (4):805–815, 2018.
22. Ou SY, Chu H, Chao PW, Ou SM, Lee YJ, Kuo SC, Li SY, Shih CJ, Chen YT. Effect of the use of low and high potency statins and sepsis
outcomes. Intensive Care Med
40 (10):1509–1517, 2014.
23. Shrank WH, Patrick AR, Brookhart MA. Healthy user and related biases in observational studies of preventive interventions: a primer for physicians. J Gen Intern Med
26 (5):546–550, 2011.
24. Brookhart MA, Patrick AR, Dormuth C, Avorn J, Shrank W, Cadarette SM, Solomon DH. Adherence to lipid-lowering therapy and the use of preventive health services: an investigation of the healthy user effect. Am J Epidemiol
166 (3):348–354, 2007.
25. Humphrey LL, Chan BK, Sox HC. Postmenopausal hormone replacement therapy
and the primary prevention of cardiovascular disease. Ann Intern Med
137 (4):273–284, 2002.
26. Kinjo M, Chia-Cheng Lai E, Korhonen MJ, McGill RL, Setoguchi S. Potential contribution of lifestyle and socioeconomic factors to healthy user bias in antihypertensives and lipid-lowering drugs. Open Heart
4 (1):e000417, 2017.
27. Stampfer MJ, Willett WC, Colditz GA, Rosner B, Speizer FE, Hennekens CH. A prospective study of postmenopausal estrogen therapy and coronary heart disease. N Engl J Med
313 (17):1044–1049, 1985.
28. Wiewel MA, van Vught LA, Scicluna BP, Hoogendijk AJ, Frencken JF, Zwinderman AH, Horn J, Cremer OL, Bonten MJ, Schultz MJ, et al. Prior use of calcium channel blockers
is associated with decreased mortality in critically ill patients with sepsis
: a prospective observational study. Crit Care Med
45 (3):454–463, 2017.
29. Zheng L, Hunter K, Gaughan J, Poddar S. Preadmission use of calcium channel blockers
and outcomes after hospitalization with pneumonia: a retrospective propensity-matched cohort study. Am J Ther
24 (1):e30–e38, 2017.
30. Li G, Qi XP, Wu XY, Liu FK, Xu Z, Chen C, Yang XD, Sun Z, Li JS. Verapamil modulates LPS-induced cytokine production via inhibition of NF-kappa B activation in the liver. Inflamm Res
55 (3):108–113, 2006.
31. Sirmagul B, Kilic FS, Tunc O, Yildirim E, Erol K. Effects of verapamil and nifedipine on different parameters in lipopolysaccharide-induced septic shock
. Heart Vessels
21 (3):162–168, 2006.
32. Li XQ, Cao W, Li T, Zeng AG, Hao LL, Zhang XN, Mei QB. Amlodipine inhibits TNF-alpha production and attenuates cardiac dysfunction induced by lipopolysaccharide involving PI3K/Akt pathway. Int Immunopharmacol
9 (9):1032–1041, 2009.
33. Wyska E. Pretreatment with R(+)-verapamil significantly reduces mortality and cytokine expression in murine model of septic shock
. Int Immunopharmacol
9 (4):478–490, 2009.
34. Xu H, Garver H, Fernandes R, Galligan JJ, Fink GD. Altered L-type Ca2+ channel activity contributes to exacerbated hypoperfusion and mortality in smooth muscle cell BK channel-deficient septic mice. Am J Physiol Regul Integr Comp Physiol
307 (2):R138–R148, 2014.
35. Wakefield BJ, Sacha GL, Khanna AK. Vasodilatory shock in the ICU
and the role of angiotensin II. Curr Opin Crit Care
24 (4):277–285, 2018.
36. Bauer SR, Sacha GL, Lam SW. Safe use of vasopressin and angiotensin II for patients with circulatory shock. Pharmacotherapy
38 (8):851–861, 2018.
37. Khanna A, English SW, Wang XS, Ham K, Tumlin J, Szerlip H, Busse LW, Altaweel L, Albertson TE, Mackey C, et al. Angiotensin II for the treatment of vasodilatory shock. N Engl J Med
377 (5):419–430, 2017.
38. Chawla LS, Russell JA, Bagshaw SM, Shaw AD, Goldstein SL, Fink MP, Tidmarsh GF. Angiotensin II for the Treatment of High-Output Shock 3 (ATHOS-3): protocol for a phase III, double-blind, randomised controlled trial. Crit Care Resusc
19 (1):43–49, 2017.
39. Ray WA, Daugherty JR, Griffin MR. Lipid-lowering agents and the risk of hip fracture in a Medicaid population. Inj Prev
8 (4):276–279, 2002.
40. Esposito K, Capuano A, Sportiello L, Giustina A, Giugliano D. Should we abandon statins in the prevention of bone fractures? Endocrine
44 (2):326–333, 2013.
41. Haley RW, Dietschy JM. Is there a connection between the concentration of cholesterol circulating in plasma and the rate of neuritic plaque formation in Alzheimer disease? Arch Neurol
57 (10):1410–1412, 2000.
42. Majumdar SR, McAlister FA, Eurich DT, Padwal RS, Marrie TJ. Statins and outcomes in patients admitted to hospital with community acquired pneumonia: population based prospective cohort study. BMJ
333 (7576):999, 2006.
43. Psaty BM, Smith NL, Siscovick DS, Koepsell TD, Weiss NS, Heckbert SR, Lemaitre RN, Wagner EH, Furberg CD. Health outcomes associated with antihypertensive therapies used as first-line agents. A systematic review and meta-analysis. JAMA
277 (9):739–745, 1997.
44. Psaty BM, Lumley T, Furberg CD, Schellenbaum G, Pahor M, Alderman MH, Weiss NS. Health outcomes associated with various antihypertensive therapies used as first-line agents: a network meta-analysis. JAMA
289 (19):2534–2544, 2003.
45. Mancia G, De Backer G, Dominiczak A, Cifkova R, Fagard R, Germano G, Grassi G, Heagerty AM, Kjeldsen SE, Laurent S, et al. 2007 Guidelines for the Management of Arterial Hypertension: the Task Force for the Management of Arterial Hypertension of the European Society of Hypertension (ESH) and of the European Society of Cardiology (ESC). J Hypertens
25 (6):1105–1187, 2007.
46. Grossman E, Verdecchia P, Shamiss A, Angeli F, Reboldi G. Diuretic treatment of hypertension. Diabetes Care
34: (Suppl. 2): S313–S319, 2011.
47. Hypertension: management of hypertension in adults in primary care. Available at: https://www.nice.org.uk/guidance/CG34
. Accessed April 17, 2019.