Highlights
-
•
Medication administration prior to in-hospital cardiac arrest remains understudied.
-
•
IHCA patients are more frequently introduced to new medication 24 h prior to IHCA.
-
•
Differences between IHCA cases and controls were seen in antithrombotics, diuretics, and opioids.
Keywords: In-hospital cardiac arrest, Medication, Prescription
Abstract
Background
Limited research has examined medication administration in patients with in-hospital cardiac arrest (IHCA). We aimed to identify and describe in-hospital medications administered 24 h prior to IHCA compared with matched controls.
Methods
This study was a registry-based, matched, case-control study including patients who were hospitalized in the Central Denmark Region from 2017 to 2022. Patients with IHCA were matched with hospitalized controls on age, sex, length, and period of admission. The index time was defined as the time of IHCA for cases, and a corresponding index time was assigned to matched controls. Medication administrations within 24 h prior to index time were identified and divided into 21 groups based on ATC codes, e.g. QTc-prolonging drugs. New medication administrations were defined as no prior redeemed prescription for the same ATC code within 2 years.
Results
A total of 1976 cases and 19,716 controls met the study criteria and were included in the study. Among cases, the most frequently administered medication groups were antithrombotic agents (43% of cases), antibiotics (42%), and diuretics (35%), whereas for controls the groups were antibiotics (33% of controls), antithrombotic agents (33%), and opioids (24%). Overall, cases were administered a median of 11 (IQR: 0–20) medications within 24 h prior to the index time as compared to 5 (IQR: 0–14) medications for controls.
Conclusion
Medications administered in-hospital within 24 h prior to IHCA differed from those in matched controls and were characterized by a higher proportion of newly initiated treatments. These findings may help form new hypotheses on strategies on prevention of IHCA.
Introduction
Approximately 2000 in-hospital cardiac arrests (IHCA) occur each year in Denmark, and around 300,000 suffer from IHCA each year in the United States.1, 2 Survival to hospital discharge remains at only 25–30%.1, 3, 4 According to international guidelines further research on IHCA is needed to develop preventative strategies and thereby reduce mortality.3, 5
Pharmacological causes, including medication-related toxicity, are widely recognized as potentially reversible causes of cardiac arrest.6, 7, 8 For example, several psychotropic medications, including tricyclic antidepressants and antipsychotics, are associated with QTc-prolongation, which can trigger ventricular arrhythmias and thereby sudden cardiac death.9, 10, 11 Further, opioids have been associated with an increased risk of IHCA in former studies, by increasing risk of respiratory depression which might lead to hypoxia, hypercapnia, and potentially cardiac arrest.12, 13 Furthermore, previous research has shown that clinical deterioration is associated with medication-related adverse events.14, 15 The majority of studies investigating medication use in cardiac arrest have focused on out-of-hospital cardiac arrest16, 17 or on the association between specific medications and IHCA, such as sedatives and opioids.12, 13, 18 These studies have been limited by confounding by indication and small study populations. As a result, only limited research has described the overall hospital medication administration for patients preceding IHCA.
Hospitalized patients are often treated with multiple medications simultaneously. Identifying and describing medications administered prior to cardiac arrest may help form new hypotheses on IHCA aetiologies and future strategies on prevention. The aim of this study was to identify and describe medications administered during hospitalization 24 h prior to IHCA among patients who experienced IHCA compared with matched controls.
Methods
Study design and data sources
The study was a registry-based matched case-control study using both national and regional Danish registries, including the Danish In-Hospital Cardiac Arrest Registry (DANARREST), the electronic medical records of the Central Denmark Region (MidtEPJ), the Danish National Prescription Registry, the National Patient Registry, and the Danish Civil Registration System. Data across all datasets were linked by the unique Civil Registration Number, assigned to all Danish residents. These data sources have previously been described in detail elsewhere,19, 20, 21, 22 but a brief overview is provided below. A previous study has been published using the same cohort and examined vital signs prior to IHCA.23
DANARREST is a national quality-control database containing data on IHCA in Denmark. The database contains data on cardiac arrest characteristics. From this registry, IHCA cases and the associated cardiac arrest characteristics were acquired.
MidtEPJ is a regional electronic medical record system, which contains documentation for all patients who have received outpatient or inpatient care at one of the hospitals in the Central Denmark Region. The Central Denmark Region consists of seven regional hospitals and one university hospital, offering healthcare to approximately 1.3 million people. Data from MidtEPJ are available for research purposes through the regions Business Intelligence Data Warehouse, and information for the study population’s in-hospital medication administration was collected through this.24
The Danish National Prescription Registry contains individual-level information on dispensing of medication across pharmacies. Therefore, it was possible to identify prescribed and purchased medication for the patients prior to hospital admission using this registry.
The Danish Civil Registration System provided information on sex, and dates of birth and death.
The National Patient Registry contains ICD-10 diagnostic codes, procedure codes, and operations codes, which was used to describe the study population in regards of comorbidities, former procedures and length of admission.22
Ethical approval is not required for register-based research in Denmark.
Study population
From DANARREST, information on adult IHCA patients (≥18 years) occurring in the Central Denmark Region in the period January 1st, 2017, to December 31st, 2022, was extracted. IHCA was defined as any patient needing chest compressions, defibrillation or both for any person within hospital walls. If a patient experienced more than one IHCA, only the first event was included. Patients were excluded if no rhythm analysis was performed, if data on rhythm analysis was missing, or if the first rhythm analysis showed a pulse-generating rhythm, as these patients were not considered to be in true cardiac arrest. Cases were excluded if they had IHCA in a psychiatric setting, as these patients represent a unique clinical population.
The index time for cases was defined as the time and date of IHCA. Each case was then matched with up to 10 controls using risk set sampling with replacement, meaning that controls could later be included as a case if they subsequently had a cardiac arrest. Matching was based on age (±5 years), sex, period of admission (±1 month), and length of hospital stay in days prior to the index time for the case. The latter matching criterion ensured that controls had been admitted at least the same number of days as the matched case’s length of admission to index time. For example, if a case experiences IHCA on day 2, a corresponding control must be admitted for at least 2 days. Patients with psychiatric admissions were excluded.
Medication administration
For the study population, we characterized whether patients received any in-hospital medication and identified the medication groups administered within 24 h prior to the index time. All information on medication administrations from the time of admission to the index time was obtained from MidtEPJ.
Each medication administered is registered in MidtEPJ with the corresponding Anatomical Therapeutic Chemical (ATC) Classification code.25, 26 The ATC code system is an international system recognized by the World Health Organization (WHO), which classifies medications with unique codes based on their therapeutic actions and active substance. Medications were pooled into 21 groups based on these ATC codes. The selected medication groups and the ATC search are presented in Supplemental Tables 1 and 2.
Statistical analysis
Continuous data were summarized using medians with interquartile ranges, and categorical data were summarized using counts and percentages.
The main analysis examined administration of the selected medication groups 24 h prior to the index time. The aggregate number and proportion of cases and controls receiving at least one medication from the predefined groups were reported. Further, we described patients in whom the administration of medication was part of a new prescription during their hospital stay rather than the continuation of previously prescribed medications. This was done by cross-referencing each ATC-code with redeemed prescriptions from the Danish National Prescription Registry. A medication was considered newly initiated if the patient had not redeemed a prescription within 2 years prior to admission date with the same ATC code, regardless of prior use of other medications within the same medication group.
Additional analyses were restricted to cases and controls who received at least one medication within 24 h prior to index time. In this subgroup, the proportion of patients receiving at least one medication from one of the predefined groups was reported. Moreover, time in hours from last administration to the index time and the total number of administrations within 24 h prior to the index time were reported for cases and controls. All the metrics above were also reported separately for cases according to whether the initial rhythm was shockable vs. non-shockable.
An additional analysis focused on the most frequently administered medications for the following groups: opioids, benzodiazepines (and benzodiazepine-related medication), and antipsychotics. The aim was to characterize medication patterns involving drugs known to increase the risk of adverse effects, such as QTc prolongation and respiratory depression.
All statistical analyses were descriptive and performed using R statistical software (R Foundation for Statistical Computing, Vienna, Austria), R version 4.4.3 (2025-02-28).
Results
Study population
A total of 1976 cases and 19,716 controls met the study criteria and were included in the final study population. A flowchart of patient inclusion is outlined in Fig. 1. Baseline population characteristics are presented in Table 1. The median age of the study population was 73 years (IQR: 65–80) and the population was predominantly male (65%). The largest differences in comorbidities between cases and controls were for ischemic heart disease (29% and 22%, respectively) and heart failure (20% and 11%, respectively).
Fig. 1.
Flowchart of patient inclusion.
Table 1.
Base population characteristics.
|
Cases N = 1976 |
Controls N = 19,716 |
|
|---|---|---|
| Demographics | ||
| Age | 73 (65–80) | 73 (65–80) |
| Male | 1276 (65%) | 12,731 (65%) |
| Female | 700 (35%) | 6985 (35%) |
| Comorbidities | ||
| Ischaemic heart disease | 575 (29%) | 4261 (22%) |
| Heart failure | 400 (20%) | 2227 (11%) |
| Hypertension | 710 (36%) | 6101 (31%) |
| Atrial fibrillation/flutter | 473 (24%) | 3882 (20%) |
| Other cardiac dysrhythmias | 270 (14%) | 1987 (10%) |
| Peripheral vascular disease | 226 (11%) | 1785 (9.1%) |
| Hypercholesterolemia | 221 (11%) | 2252 (11%) |
| Venous thromboembolism | 147 (7.4%) | 1385 (7.0%) |
| Pulmonary arterial hypertension | 36 (1.8%) | 194 (1.0%) |
| Cerebrovascular disease | 284 (14%) | 3460 (18%) |
| Other neurological diseases | 152 (7.7%) | 1846 (9.4%) |
| Chronic obstructive pulmonary disease | 330 (17%) | 2742 (14%) |
| Asthma | 98 (5.0%) | 842 (4.3%) |
| Diabetes Mellitus | 383 (19%) | 2886 (15%) |
| Obesity | 173 (8.8%) | 1608 (8.2%) |
| Liver disease | 139 (7.0%) | 1165 (5.9%) |
| Gastric ulcer | 103 (5.2%) | 705 (3.6%) |
| Pancreatitis | 46 (2.3%) | 467 (2.4%) |
| Renal disease | 309 (16%) | 2046 (10%) |
| Psychiatric disease | 371 (19%) | 3188 (16%) |
| Cardiac procedures prior to index admission | 730 (37%) | 4154 (21%) |
| Hospital admission characteristics | ||
| Any ICU admission between admission and index time | 451 (23%) | 1093 (5.5%) |
| Hours from admission to index time | 33 (4–109) | 32 (4–107) |
| Outcomes | ||
| 30-day survival | 544 (28%) | 18,241 (92.5%) |
| 180-day survival | 480 (24%) | 16,062 (81%) |
Results are reported as either counts with percentages or as median (quartiles).
Cardiac arrest characteristics for cases are presented in Table 2 overall and stratified by administration of any medication 24 h prior to index time. The proportion of patients with an initial shockable rhythm was 21%, and most patients suffered IHCA in a medical or surgical ward (46%).
Table 2.
Cardiac arrest characteristics.
|
All cases N = 1976 |
Cases administered any in-hospital medication within 24 h prior to IHCA N = 1323 |
Cases with no in-hospital medication administration within 24 h prior to IHCA N = 652 |
|
|---|---|---|---|
| Location of cardiac arrest | |||
| Hospital ward | 903 (46%) | 790 (60%) | 113 (17%) |
| Emergency department | 401 (20%) | 122 (9.2%) | 279 (43%) |
| Outpatient clinic | 22 (1.1%) | 9 (0.7%) | 13 (2.0%) |
| Cardiac Catheterization Laboratory | 123 (6.2%) | 38 (2.9%) | 85 (13%) |
| Operation theatre | 71 (3.6%) | 41 (3.1%) | 30 (4.6%) |
| Postoperative unit | 31 (1.6%) | 19 (1.4%) | 12 (1.8%) |
| Intensive care unit | 278 (14%) | 223 (17%) | 55 (8.4%) |
| Other | 147 (7.4%) | 81 (6.1%) | 65 (10.0%) |
| Cause of arrest | |||
| Non-cardiac | 1081 (56%) | 774 (60%) | 306 (47%) |
| Presumed cardiac | 862 (44%) | 521 (40%) | 341 (53%) |
| Intubated prior to cardiac arrest | |||
| Yes | 251 (13%) | 153 (12%) | 98 (15%) |
| No | 1717 (87%) | 1166 (88%) | 550 (85%) |
| Monitored cardiac arrest | |||
| Yes | 1046 (54%) | 618 (47%) | 428 (67%) |
| No | 909 (46%) | 693 (53%) | 215 (33%) |
| Initial recorded cardiac arrest rhythm | |||
| Ventricular fibrillation | 237 (12%) | 126 (9.5%) | 111 (17%) |
| Pulseless ventricular tachycardia | 167 (8.5%) | 106 (8.0%) | 61 (9.4%) |
| Pulseless electrical activity | 953 (48%) | 605 (46%) | 347 (53%) |
| Asystole | 619 (31%) | 486 (37%) | 133 (20%) |
| Witnessed cardiac arrest | |||
| Yes | 1552 (79%) | 956 (72%) | 595 (91%) |
| No | 422 (21%) | 366 (28%) | 56 (8.6%) |
| Reason for termination of CPR | |||
| Return of spontaneous circulation | 999 (51%) | 613 (47%) | 386 (59%) |
| Death | 913 (46%) | 677 (52%) | 235 (36%) |
| eCPR | 52 (2.6%) | 22 (1.7%) | 30 (4.6%) |
| Time of day | |||
| 07–15 | 1010 (51%) | 666 (50%) | 343 (53%) |
| 15–23 | 673 (34%) | 490 (37%) | 183 (28%) |
| 23–07 | 293 (15%) | 167 (13%) | 126 (19%) |
Results are reported as either counts with percentages or as median (quartiles).
Medication administration overall
Table 3 presents the distribution of medications administered 24 h prior to index time among cases and controls. For cases, the most frequently administered medication groups were antithrombotic agents (43%), antibiotics (42%), and diuretics (35%), whereas for controls it was antibiotics (33%), antithrombotic agents (33%), and opioids (24%).
Table 3.
Administration of medications 24 h prior to index time for cases and controls.
|
Cases N = 1976 |
Controls N = 19,716 |
|||
|---|---|---|---|---|
| Patients administered the medication – No. (%) | Patients administered the medication as a new administration – No. (%) | Patients administered the medication – No. (%) | Patients administered the medication as a new administration – No. (%) | |
| Opioids | 571 (29%) | 425 (22%) | 4794 (24%) | 2786 (14%) |
| Antidepressants | 182 (9.2%) | 8 (0.4%) | 1903 (9.7%) | 168 (0.9%) |
| Antipsychotics | 123 (6.2%) | 77 (3.9%) | 636 (3.2%) | 311 (1.6%) |
| Benzodiazepine and benzodiazepine-related medications | 262 (13%) | 181 (9%) | 1713 (8.7%) | 947 (4.8%) |
| β-blockers | 387 (20%) | 83 (4.2%) | 2950 (15%) | 368 (1.9%) |
| Ca2+ channel antagonists | 158 (8.0%) | 31 (1.6%) | 1755 (8.9%) | 309 (1.6%) |
| Antiarrhythmics (excluding β-blockers) | 202 (10%) | 119 (6%) | 904 (4.6%) | 365 (1.9%) |
| Antithrombotic | 850 (43%) | 648 (33%) | 6409 (33%) | 3883 (20%) |
| LMWH* | 486 (25%) | 484 (24%) | 2930 (15%) | 2874 (15%) |
| Antiplatelet agents | 350 (18%) | 166 (8.4%) | 2635 (13%) | 791 (4%) |
| Anticoagulants (excluding heparins) | 185 (9.4%) | 68 (3.4%) | 1663 (8.4%) | 421 (2.1%) |
| Insulin and analogs for injection | 228 (12%) | 151 (7.6%) | 1135 (5.8%) | 667 (3.4%) |
| Blood-glucose reducing medication excluding insulin | 92 (4.7%) | 10 (0.5%) | 915 (4.6%) | 58 (0.3%) |
| Diuretics | 699 (35%) | 310 (16%) | 4080 (21%) | 1082 (5.5%) |
| Glucocorticoids | 255 (13%) | 176 (8.9%) | 2266 (11%) | 1312 (6.7%) |
| Antibiotics | 833 (42%) | 800 (40%) | 6480 (33%) | 5932 (30%) |
| Antifungal | 135 (6.8%) | 124 (6.2%) | 922 (4.7%) | 767 (3.9%) |
| Chronic obstructive pulmonary disease medication | 357 (18%) | 279 (14%) | 1546 (8.0%) | 906 (4.6%) |
| NSAID | 56 (2.8%) | 40 (2%) | 865 (4.4%) | 610 (3.1%) |
| PPI** | 547 (28%) | 243 (12%) | 4057 (21%) | 1294 (6.6%) |
| QTc-prolonging medication | 354 (18%) | 146 (7.4%) | 2840 (14%) | 854 (4.3%) |
Results are reported as counts with percentages.
Low-molecular-weight heparin.
Proton pump inhibitor.
The medication group with most new medication administrations was antibiotics where 40% of cases versus 30% of controls received an antibiotic as a new medication administration. The largest absolute differences between cases and controls in new medication administrations were observed for antithrombotic agents, where 33% of cases and 20% of controls received an antithrombotic agent as a new medication administration.
Overall, cases were administered a median of 11 (IQR: 0–20) medications within 24 h prior to the index time as compared to 5 (IQR: 0–14) medications for controls. Similarly, cases were administered a median of 4 (IQR: 0–7) new distinct medications within 24 h prior to the index time as compared to 2 (IQR: 0–4) for controls.
Patients with at least one medication administration within 24 h prior to index time
1323 cases and 12,064 controls had any medication administered in the 24 h prior to their index time (see Fig. 1). Baseline characteristics, dividing cases and controls in those who received any medication 24 prior to index time and those who did not, are shown in Supplementary Table 3. The median time from admission to index time for cases who received any medication was 71 h (IQR: 31–161) and 1.7 h (IQR: 0.5–4.7) for cases who did not receive any medication with 24 h of the index time.
The proportion of patients with an initial shockable rhythm was 18% among patients receiving any medication within 24 h prior to IHCA and 26% among those who did not receive any medication. Among the patients who received any medication, 60% suffered IHCA in a general ward and 9.2% in the emergency department. In patients without medication administration within 24 h prior to IHCA, the primary locations of cardiac arrest were the emergency department (43%) and the general ward (17%).
Table 4 presents the medication groups for cases and controls in whom at least one medication was administered within 24 h of the index time. Here, the most frequently administered medication group for cases were antithrombotic agents (64%), whereas for controls it was antibiotics (54%), and the largest absolute differences in frequency between cases and controls were observed for diuretics (53% and 34%, respectively).
Table 4.
Administration of medications 24 h prior to index time for cases and controls in whom at least one medication was administered within 24 h of the index time.
|
Cases N = 1323 |
Controls N = 12,064 |
|||||||
|---|---|---|---|---|---|---|---|---|
| Patients administered the medication – No. (%) | Patients administered the medication as a new administration – No. (%) | Hours from last administration to index time | Number of administrations | Patients administered the medication – No. (%) | Patients administered the medication as a new administration – No. (%) | Hours from last administration to index time | Number of administrations | |
| Opioids | 571 (43%) | 425 (32%) | 4.8 (1.7–9.3) | 2 (1–4) | 4794 (40%) | 2786 (23%) | 7 (3–11.9) | 2 (1–4) |
| Antidepressants | 182 (14%) | 8 (0.6%) | 10.8 (5.4–11.4) | 1 (1) | 1903 (16%) | 168 (1.4%) | 11.5 (6.5–17.4) | 1 (1) |
| Antipsychotics | 123 (9.3%) | 77 (5.8%) | 7.6 (4.1–14.9) | 1 (1–2) | 636 (5.3%) | 311 (2.6%) | 10.3 (5–16.2) | 1 (1–2) |
| Benzodiazepine and benzodiazepine-related medications | 262 (20%) | 181 (14%) | 7.5 (3–13.1) | 1 (1–2) | 1713 (14%) | 947 (7.8%) | 11.4 (5.6–17.2) | 1 (1) |
| β-blockers | 387 (29%) | 83 (6.3%) | 9.1 (4.5–14.2) | 1 (1–2) | 2950 (24%) | 368 (3.1%) | 9.8 (4.9–15.7) | 1 (1–2) |
| Ca2+ channel antagonists | 158 (12%) | 31 (2.3%) | 11.9 (5.8–19.1) | 1 (1) | 1755 (15%) | 309 (2.6%) | 10.5 (4.8–16.9) | 1 (1) |
| Antiarrhythmics (excluding β-blockers) | 202 (15%) | 119 (9%) | 10.8 (4.4–16.2) | 1 (1–2) | 904 (7.5%) | 365 (3%) | 10.4 (5.2–16.4) | 1 (1) |
| Antithrombotic | 850 (64%) | 648 (50%) | 9.2 (4.3–15.4) | 1 (1–2) | 6409 (53%) | 3883 (32%) | 10.6 (5.5–16.4) | 1 (1–2) |
| LMWH* | 486 (37%) | 484 (37%) | 11.1 (5.5–16.7) | 1 (1) | 2930 (24%) | 2874 (24%) | 12.1 (6.7–17.5) | 1 (1) |
| Antiplatelet agents | 350 (26%) | 166 (12.5%) | 10 (4.7–17) | 1 (1–2) | 2635 (22%) | 791 (6.6%) | 10.9 (5.4–17.8) | 1 (1) |
| Anticoagulants (excluding heparins) | 185 (14%) | 68 (5.1%) | 10.9 (5.9–16.4) | 1 (1) | 1663 (14%) | 421 (3.5%) | 10.3 (5.4–15.6) | 1 (1–2) |
| Insulin and analogs for injection | 228 (17%) | 151 (11%) | 6.3 (2.9–11.9) | 2 (1–4) | 1135 (9.4%) | 667 (5.5%) | 7.4 (3.3–13.6) | 2 (1–4) |
| Blood-glucose reducing medication excluding insulin | 92 (7.0%) | 10 (0.8%) | 9.8 (4.5–15.2) | 2 (1–2) | 915 (7.6%) | 58 (0.5%) | 8.7 (4.2–14.5) | 2 (1–2) |
| Diuretics | 699 (53%) | 310 (23%) | 6.2 (2.5–11.9) | 2 (1–3) | 4080 (34%) | 1082 (9%) | 9.5 (4.4–15.5) | 1 (1–2) |
| Glucocorticoids | 255 (19%) | 176 (13%) | 8 (2.8–16.4) | 1 (1) | 2265 (19%) | 1312 (11%) | 11.2 (5.5–17.1) | 1 (1) |
| Antibiotics | 833 (63%) | 800 (60%) | 4.4 (2.1–7.2) | 3 (2–4) | 6480 (54%) | 5932 (49%) | 5.8 (2.9–10.4) | 3 (2–3) |
| Antifungal | 135 (10%) | 124 (9.4%) | 7.3 (3.0–14.2) | 2 (1–3) | 922 (7.6%) | 767 (6.4%) | 12.4 (3.8–12.4) | 2 (1–4) |
| Chronic obstructive pulmonary disease medication | 357 (27%) | 279 (21%) | 4.9 (2.1–10) | 3 (1–4) | 1546 (13%) | 906 (7.5%) | 7.1 (3.1–12.6) | 2 (1–4) |
| NSAID | 56 (4.2%) | 40 (3%) | 9.4 (5.2–18.1) | 1 (1) | 865 (7.2%) | 610 (5.1%) | 11.3 (7.3–18.3) | 1 (1) |
| PPI** | 547 (41%) | 243 (18%) | 10.3 (5.2–17.2) | 1 (1) | 4057 (34%) | 1294 (11%) | 10.7 (5.7–16.7) | 1 (1) |
| QTc-prolonging medication | 354 (27%) | 146 (11%) | 8.2 (3.9–13.8) | 1 (1–2) | 2840 (24%) | 854 (7%) | 10.1 (5.2–16.2) | 1 (1–2) |
Results are reported as counts with percentages or as median (quartiles).
Low-molecular-weight heparin.
Proton pump inhibitor.
Medication administration of antipsychotics, benzodiazepines, and opioids
Table 5 presents the most frequently administered medications within the antipsychotics, benzodiazepine and opioid groups for cases and controls in whom at least one medication was administered within 24 h of the index time as well as the number of patients administered the medication as a new administration.
Table 5.
New medications exposure for the most commonly administered antipsychotics, opioids, and benzodiazepines for cases and controls in whom at least one medication was administered within 24 h of the index time.
|
Cases N = 1323 |
Controls N = 12,064 |
|||
|---|---|---|---|---|
| Patients administered the medication – No. (%) | Number of Patients administered the medication as a new administration | Patients administered the medication – No. (%) | Patients administered the medication as a new administration – No. (%) | |
| Antipsychotics | 123 (9.3%) | 77 (5.8%) | 636 (5.3%) | 311 (2.6%) |
| Haloperidol | 21 (1.6%) | 20 (1.6%) | 96 (0.8%) | 73 (0.6%) |
| Droperidol | 15 (1.1%) | 15 (1.1%) | 94 (0.8%) | 94 (0.8%) |
| Olanzapine | 38 (2.9%) | 35 (2.6%) | 139 (1.2%) | 92 (0.8%) |
| Quetiapine | 26 (2.0%) | 0 (0%) | 194 (1.6%) | 30 (0.2%) |
| Risperidone | 24 (1.8%) | 13 (1%) | 82 (0.7%) | 39 (0.3%) |
| Opioids | 571 (43%) | 425 (32%) | 4794 (40%) | 2786 (23%) |
| Morphine | 366 (28%) | 267 (20%) | 3095 (26%) | 1684 (14%) |
| Oxycodone | 149 (11%) | 110 (8.3%) | 1235 (10%) | 797 (6.6%) |
| Tramadol | 64 (4.8%) | 25 (1.9%) | 500 (4.1%) | 155 (1.3%) |
| Fentanyl | 49 (3.7%) | 49 (3.7%) | 253 (2.1%) | 253 (2.1%) |
| Methadone | 16 (1.2%) | 10 (0.8%) | 197 (1.6%) | 83 (0.7%) |
| Benzodiazepine and benzodiazepine-related medications | 262 (20%) | 181 (14%) | 1713 (14%) | 947 (7.8%) |
| Zopiclone | 110 (8.3%) | 73 (5.5%) | 927 (7.7%) | 508 (4.2%) |
| Oxazepam | 26 (2.0%) | 12 (0.9%) | 233 (1.9%) | 73 (0.6%) |
| Midazolam | 50 (3.8%) | 50 (3.8%) | 227 (1.8%) | 207 (1.7%) |
| Zolpidem | 24 (1.8%) | 7 (0.5%) | 169 (1.4%) | 42 (0.3%) |
| Diazepam | 20 (1.5%) | 18 (1.4%) | 106 (0.9%) | 63 (0.5%) |
| Chlordiazepoxide | 39 (2.9%) | 27 (2%) | 88 (0.7%) | 51 (0.4%) |
| Clonazepam | 10 (0.8%) | 3 (0.2%) | 46 (0.4%) | 17 (0.1%) |
Results are reported as counts with percentages.
A total of 123 cases (9.3%) and 636 controls (5.3%) received antipsychotics. The antipsychotic with the largest observed absolute difference in new administration between cases and controls was olanzapine, which was a new administration of olanzapine among 2.6% of cases and 0.8% of controls.
Opioids were administered to 571 cases (43%) and 4794 controls (40%). Morphine showed the largest absolute difference in new administration between cases and controls (20% and 14%, respectively).
262 cases (20%) and 1713 controls (14%) were administered benzodiazepines. Midazolam showed the largest absolute difference in new administration between cases and controls (3.8% and 1.7%, respectively).
Discussion
Summary of major findings
In this matched case-control study, we described patterns of medication administration among IHCA patients compared to matched hospitalized controls. Focusing on medications initiated during hospitalization, IHCA patients were more frequently administered medications across multiple medication groups compared to controls, particularly antiarrhythmics, diuretics, and opioids. Antithrombotics were the most administered medication groups among cases, and antibiotics were the most common among controls.
Our findings suggest that IHCA patients are more frequently exposed to antithrombotic and antiarrhythmic therapies. This may reflect the underlying disease, as antithrombotic treatment is typically initiated for acute conditions such as pulmonary embolism or acute coronary syndrome,27 both of which are associated with an increased risk of cardiac arrest. In addition, antiarrhythmic drugs are also known to carry a proarrhythmic potential, which in some cases may contribute to life-threatening ventricular arrhythmias.28
Comparison with previous studies
In our study, both cases and controls were administered opioids frequently; cases had a higher proportion of newly initiated opioids, and opioids were administered closer to the index time compared to controls. Other studies have investigated the association between opioids, sedatives (including benzodiazepines) and the risk of IHCA, and found an increased risk of IHCA particularly when the two medication groups were combined.12, 13 These studies hypothesized that this could be explained by patients suffering IHCA secondary to opioid- and sedative-induced respiratory depression and hypoxia. However, these findings are limited by confounding by indication as the administration of opioids and sedatives may simply represent underlying severe illness making causal inference very difficult in these cases. In addition, no data were presented on the causes of IHCA in these patients, making interpretation of the results difficult.
A separate study investigated the association between QTc prolonging medications and IHCA. This study found an increased risk of IHCA in patients who were exposed to multiple QTc prolonging medications.18 This is in line with the known QTc-prolonging risk from these medications, and we also found IHCA cases to be more exposed to new administrations of QTc-prolonging medications. However, the study’s suggestion of an increased risk due to these medications is limited by a high risk of confounding by indication, as with the studies above.
Strengths and limitations
This study had several strengths. First, the individual-level linkage across Danish health registries provided data of high quality on medication administration and cardiac arrest characteristics with minimal missing data and loss to follow-up. Second, data from the National Prescription Registry enabled us to distinguish between medications that were simply continuations of already prescribed medications, which are of less interest from both a research and clinical perspective, and medications that were newly initiated during the index hospitalization. Third, we provided data from a matched control population in order to compare the medication administrations to a reference group. Fourth, we received data from the entire Central Denmark Region, including both regional and university hospital, increasing the external validity of our findings.
This study also has limitations. First, the observational design prevents establishing a causal link between medication administration and development of IHCA, which is why we chose a descriptive design. Second, we chose only to match on sex, age, admission length and time and not on, for example, admission type (elective vs emergency). This may explain the higher burden of comorbidity among cases compared to controls (Table 1); however, IHCA may occur in any hospitalized patient, and matching on admission type and diagnosis could have reduced the number of eligible controls for matching. Third, grouping by overall ATC code is a crude measure and does not provide information on specific medication indications. Lastly, this is a Danish study in a Danish setting, and our findings might therefore not be directly applicable to how medication is administered in other settings.
Areas for future research
This study utilized detailed data on medication administration for a large cohort of patients preceding IHCA. The main purpose of this study was to provide a detailed descriptive analysis of medications administered 24 h prior to IHCA during hospitalization. To our knowledge, no previous studies have described overall medication administration at this level of detail for IHCA patients compared to matched controls. Our findings suggest that IHCA does not occur in the presence of one single medication exposure but rather in the presence of multiple and often newly initiated medications. One explanation is that the medication administered may reflect the severity of illness and is the result of clinical deterioration and an escalation of treatment prior to cardiac arrest, for example antithrombotics administered to a patient experiencing acute coronary syndrome. A second explanation might be the effect of newly introduced medications in the context of polypharmacy, which may increase the risk of drug interactions, side effects, and physiological vulnerability in an already unstable patient. It is difficult to make a direct causal relationship between new medication administrations and IHCA in this, or any, observational study without risk of confounding, and therefore, we chose a descriptive approach to this investigation. We believe that our hypothesis-generating results have the potential to form the basis of new investigations into medication exposure for admitted patients and their association with the risk of IHCA. Future studies could investigate medication administration as one variable among several for inclusion in multivariable models or machine-learning algorithms with the potential for early prediction of IHCA.
Several patients did not receive any medication within 24 h of the index time. These patients had a very short interval between admission and the index time with a median of 2 h (IQR: 0.5–4.7 h) as compared to the cases and controls (Supplementary Table 3), who did receive medication, which had median times from admission to the index time of 73 h (IQR: 34–164 h). Moreover, of the IHCA patients who did not receive any medication within 24 h prior to IHCA, 43% suffered IHCA in the emergency department and 13% in the cardiac catheterization laboratory. For cases with any medication administration within 24 h prior to the index time, 9.2% suffered IHCA in the emergency department and 2.9% in the cardiac catheterization laboratory. These differences could suggest that IHCA patients constitute several phenotypes; some patients who suffer IHCA are critically ill at admission and suffer IHCA very early during their hospital stay, and other IHCA patients suffer IHCA after longer admissions. These findings are supported by other studies.29 These different IHCA phenotypes should be investigated further in future studies.
Conclusion
Patients with a IHCA are more often administered certain medications during their hospitalization compared to matched hospitalized controls. Differences were most pronounced for antithrombotics, diuretics, and opioids. Moreover, patients with IHCA are subject to a greater overall medication burden compared to controls. Further investigation is warranted to explore whether these differences reflect causal relationships.
Declaration of generative AI use
Generative AI was used to assist with coding.
Funding sources
This study was supported by the ZOLL Foundation and Laerdal Foundation.
CRediT authorship contribution statement
Stine H. Hansen: Writing – original draft, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Peter C. Lind: Writing – review & editing, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Frederik G. Hansen: Writing – review & editing, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Nikola Stankovic: Writing – review & editing, Validation, Supervision, Methodology, Data curation. Mathias J. Holmberg: Writing – review & editing, Validation, Supervision, Methodology, Data curation. Lars W. Andersen: Writing – review & editing, Validation, Supervision, Methodology, Data curation. Asger Granfeldt: Writing – review & editing, Supervision, Project administration, Methodology, Investigation, Funding acquisition, Conceptualization.
Declaration of competing interest
Asger Granfeldt and Lars W Andersen are members of the editorial board at Resuscitation Plus.
Acknowledgements
None
Footnotes
Supplementary material to this article can be found online at https://doi.org/10.1016/j.resplu.2026.101511.
Appendix A. Supplementary material
The following are the Supplementary material to this article:
ATC codes and medication administration divided in shockable vs. non-shockable rythm.
References
- 1.Andersen L.W., et al. In-hospital cardiac arrest: a review. JAMA. 2019;321(12):1200–1210. doi: 10.1001/jama.2019.1696. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Holmberg M.J., et al. Annual incidence of adult and pediatric in-hospital cardiac arrest in the United States. Circ Cardiovasc Qual Outcomes. 2019;12(7) [PMC free article] [PubMed] [Google Scholar]
- 3.Baldi E., et al. European Resuscitation Council guidelines 2025 epidemiology in resuscitation. Resuscitation. 2025;215(Suppl 1) doi: 10.1016/j.resuscitation.2025.110733. [DOI] [PubMed] [Google Scholar]
- 4.Andersen L.W., et al. Adult in-hospital cardiac arrest in Denmark. Resuscitation. 2019;140:31–36. doi: 10.1016/j.resuscitation.2019.04.046. [DOI] [PubMed] [Google Scholar]
- 5.Nallamothu B.K., et al. Ten steps toward improving in-hospital cardiac arrest quality of care and outcomes. Circ Cardiovasc Qual Outcomes. 2023;16(11) doi: 10.1161/CIRCOUTCOMES.123.010491. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Lind P.C., et al. Why do patients develop in-hospital cardiac arrest? A prospective clinical observational study (WHY-IHCA) Resuscitation. 2025;216 doi: 10.1016/j.resuscitation.2025.110782. [DOI] [PubMed] [Google Scholar]
- 7.Allencherril J., et al. Etiologies of In-hospital cardiac arrest: a systematic review and meta-analysis. Resuscitation. 2022;175:88–95. doi: 10.1016/j.resuscitation.2022.03.005. [DOI] [PubMed] [Google Scholar]
- 8.Reddy D.R.S., Hanmandlu A. The 4 H's and T's: how reliable is this mnemonic in classifying etiologies of in-hospital cardiac arrests? Resuscitation. 2022;175:3–5. doi: 10.1016/j.resuscitation.2022.03.035. [DOI] [PubMed] [Google Scholar]
- 9.Lægemiddelstyrelsen. Medikamentelt forlænget QT-interval; 2024 [2 February 2026]. Available from: https://pro.medicin.dk/laegemiddelgrupper/grupper/318345.
- 10.(AZCERT), ACfEaRoT. QT drugs lists; 2025 [2 February 2026]. Available from: https://crediblemeds.org/healthcare-providers/drug-lists/qt-drug-lists.
- 11.Stampe N.K., et al. Toxicological profile using mass spectrometry in sudden cardiac arrest survivors admitted to a tertiary centre. Resuscitation. 2024;198 doi: 10.1016/j.resuscitation.2024.110197. [DOI] [PubMed] [Google Scholar]
- 12.Overdyk F.J., et al. Association of opioids and sedatives with increased risk of in-hospital cardiopulmonary arrest from an administrative database. PLoS One. 2016;11(2) doi: 10.1371/journal.pone.0150214. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Lyons P.G., et al. Association between opioid and benzodiazepine use and clinical deterioration in ward patients. J Hosp Med. 2017;12(6):428–434. doi: 10.12788/jhm.2749. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Levkovich B.J., et al. Medication-related medical emergency team activations: a case review study of frequency and preventability. BMJ Qual Saf. 2023;32(4):214–224. doi: 10.1136/bmjqs-2021-014185. [DOI] [PubMed] [Google Scholar]
- 15.Levkovich B.J., et al. The frequency and nature of clinician identified medication-related rapid response system calls. Resuscitation. 2019;145:75–78. doi: 10.1016/j.resuscitation.2019.09.033. [DOI] [PubMed] [Google Scholar]
- 16.Cromhout P.F., et al. Exploring the use of psychotropic medication in cardiac patients with and without anxiety and its association with 1-year mortality. Eur J Cardiovasc Nurs. 2022;21(6):612–619. doi: 10.1093/eurjcn/zvab111. [DOI] [PubMed] [Google Scholar]
- 17.Lavonas E.J., et al. 2023 American Heart Association focused update on the management of patients with cardiac arrest or life-threatening toxicity due to poisoning: an update to the American Heart Association guidelines for cardiopulmonary resuscitation and emergency cardiovascular care. Circulation. 2023;148(16):e149–e184. doi: 10.1161/CIR.0000000000001161. [DOI] [PubMed] [Google Scholar]
- 18.De Bruin M.L., et al. In-hospital cardiac arrest is associated with use of non-antiarrhythmic QTc-prolonging drugs. Br J Clin Pharmacol. 2007;63(2):216–223. doi: 10.1111/j.1365-2125.2006.02722.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Andersen L.W., et al. The Danish in-hospital cardiac arrest registry (DANARREST) Clin Epidemiol. 2019;11:397–402. doi: 10.2147/CLEP.S201074. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Wallach Kildemoes H., Toft Sørensen H., Hallas J. The Danish national prescription registry. Scand J Public Health. 2011;39(7_suppl):38–41. doi: 10.1177/1403494810394717. [DOI] [PubMed] [Google Scholar]
- 21.Schmidt M., Pedersen L., Sørensen H.T. The Danish civil registration system as a tool in epidemiology. Eur J Epidemiol. 2014;29(8):541–549. doi: 10.1007/s10654-014-9930-3. [DOI] [PubMed] [Google Scholar]
- 22.Schmidt M., et al. The Danish national patient registry: a review of content, data quality, and research potential. Clin Epidemiol. 2015;7:449–490. doi: 10.2147/CLEP.S91125. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Hansen F.G., et al. Vital signs preceding in-hospital cardiac arrest: a matched case-control study. Resuscitation. 2025;215 doi: 10.1016/j.resuscitation.2025.110778. [DOI] [PubMed] [Google Scholar]
- 24.Description of the BI Data Warehouse for reasearch purposes in the Central Denmark Region; 2023 [cited 2026]. Available from: https://www.rm.dk/contentassets/7b6dcaf5c7554c139b4ec054822bc4cf/reference-til-forskning.pdf.
- 25.Agency DM. Oversigt ATC-kode; 2026 [cited 2026]. Available from: https://www.medicinpriser.dk/Default.aspx?id=65.
- 26.Organization WH. Anatomical Therapeutic Chemical (ATC) classification; 2026. Available from: https://www.who.int/tools/atc-ddd-toolkit/atc-classification.
- 27.Dreijer A.R., et al. Risk of bleeding in hospitalized patients on anticoagulant therapy: prevalence and potential risk factors. Eur J Intern Med. 2019;62:17–23. doi: 10.1016/j.ejim.2019.01.008. [DOI] [PubMed] [Google Scholar]
- 28.Fordyce C.B., et al. Prevention of complications in the cardiac intensive care unit: a scientific statement from the American Heart Association. Circulation. 2020;142(22):e379–e406. doi: 10.1161/CIR.0000000000000909. [DOI] [PubMed] [Google Scholar]
- 29.Lind P.C., et al. Blood laboratory analyses preceding in-hospital cardiac arrest: a matched case-control study. Acta Anaesthesiol Scand. 2024;68(8):1085–1093. doi: 10.1111/aas.14454. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
ATC codes and medication administration divided in shockable vs. non-shockable rythm.