Hospital discharge diagnoses of ventricular arrhythmias and cardiac arrest were useful for epidemiologic research

doi:10.1016/j.jclinepi.2005.04.009

Journal of Clinical Epidemiology

Volume 58, Issue 12, December 2005, Pages 1325-1329

https://doi.org/10.1016/j.jclinepi.2005.04.009 Get rights and content

Abstract

Objective

We investigated the validity of hospital discharge diagnosis regarding ventricular arrhythmias and cardiac arrest.

Methods

We identified patients whose record in the PHARMO record linkage system database showed a code for ventricular or unspecified cardiac arrhythmias according to codes of the International Classification of Diseases, 9th revision, clinical modification (ICD-9-CM). The validity of ICD codes for ventricular arrhythmias and cardiac arrest (427.1, 427.4, 427.41, 427.42, 427.5, 427.69) and ICD codes for unspecified cardiac arrhythmias (427.2, 427.60, 427.8, 427.89, 427.9) was ascertained through manual review of hospital clinical records. The positive predictive value (PPV) was calculated, and differences between characteristics of true and false positives were evaluated.

Results

The PPV of ICD codes for ventricular arrhythmias and cardiac arrest was 82% (95% confidence interval CI = 72–92). True positive results were associated with male gender (P = .09) and younger age (P = .05). Of the unspecified cardiac arrhythmias 10% (95% CI = 2–18) were identified as ventricular arrhythmias or cardiac arrest.

Conclusion

Hospitalizations for ventricular cardiac arrhythmias and cardiac arrest (coded according to ICD-9-CM as paroxysmal ventricular tachycardia, ventricular fibrillation, ventricular flutter, ventricular premature beats, or cardiac arrest) have a high PPV and are useful for selecting events in epidemiological studies on drug-induced arrhythmias.

Introduction

Adverse drug reactions that require hospitalization fulfill the criteria for serious according to the ICH Harmonised Tripartite Guideline [1]. Hospitalizations are widely used to define outcomes in pharmacoepidemiologic database studies on side effects of drugs [2], [3], [4], [5], [6]. The most commonly used coding system to categorize hospital discharge diagnosis is the International Classification of Diseases (ICD) [7]. Classification in a computerized medical database can be subject to errors, however, ranging from incomplete reporting by a physician to typing errors by a coding clerk [8]. Several researchers have assessed the validity of hospital discharge records for various diseases, comparing the data with either original medical records [9], [10], [11], [12] or computerized information from the clinical chemical laboratory [13].

Drugs that prolong cardiac repolarization, manifested as a prolonged QTc interval on the surface electrocardiogram (ECG), may induce torsade de pointes ventricular tachycardia, which is generally preceded by ventricular premature beats [14]. Torsade de pointes may develop into ventricular fibrillation, which leads to cardiac arrest and requires electrocardioversion to restore normal rhythm [15]. Hospitalizations for ventricular arrhythmias and cardiac arrest may function as an epidemiologic study endpoint of interest, when assessing the risk for drug-induced arrhythmias among patients taking QTc-prolonging drugs.

We investigated the positive predictive value (PPV) of hospital discharge diagnosis of ventricular cardiac arrhythmias and cardiac arrest, classified according to ICD codes matching this definition. In addition, ICD codes for unspecified cardiac arrhythmias were scrutinized for ventricular cardiac arrhythmias and cardiac arrest as well.

Section snippets

Setting

Data were obtained from the PHARMO record linkage system, which contains drug dispensing records from community pharmacies and linked hospital discharge records of a defined population of 330,000 residents of eight medium-sized Dutch cities. In the Netherlands, as in many other countries, every hospital must to collect diagnostic data on all admissions. Discharge diagnoses, coded according to the International Classification of Diseases, 9th revision, clinical modification (ICD-9-CM), are

Results

During the study period, a total number of 111 patients were admitted to one of the four hospitals for ventricular or unspecified cardiac arrhythmias for the first time. All medical records could be retrieved and included 61 records of patients with an ICD code for ventricular arrhythmias or cardiac arrest and 50 records of patients with an ICD code for unspecified cardiac arrhythmias (Table 1). Overall, 55 of the 111 selected discharge diagnoses were validated as being either ventricular

Discussion

In this study, we found a positive predictive value of 82% for ventricular arrhythmias and cardiac arrest of hospitalizations specifically coded as such according to the ICD-9-CM classification. The PPV did not depend on hospital, nor on year of hospitalization, and therefore extrapolation of the results to other hospitals and other years may be justified. Other diagnoses most likely to include some of the misclassified ventricular cardiac arrhythmias and cardiac arrest cases (ICD codes for

Acknowledgments

This research was funded by the Utrecht Institute for Pharmaceutical Sciences, and an unrestricted grant from the Dutch Medicines Evaluation Board. None of the authors has any financial conflicts of interest. We gratefully thank the PHARMO Institute and the hospitals that participated in this project.

References (23)

N.S. Rawson et al.
Hospitalizations for aplastic anemia and agranulocytosis in Saskatchewan: incidence and associations with antecedent prescription drug use
J Clin Epidemiol
(1998)
C. Cattaruzzi et al.
Positive predictive value of ICD-9th codes for upper gastrointestinal bleeding and perforation in the Sistema Informativo Sanitario Regionale database
J Clin Epidemiol
(1999)
A. Chio et al.
Piemonte and Valle d'Aosta Register for ALS. Validity of hospital morbidity records for amyotrophic lateral sclerosis: a population-based study
J Clin Epidemiol
(2002)
Y. Kiyota et al.
Accuracy of Medicare claims-based diagnosis of acute myocardial infarction: estimating positive predictive value on the basis of review of hospital records
Am Heart J
(2004)
K.L. Movig et al.
Validity of hospital discharge International Classification of Diseases (ICD) codes for identifying patients with hyponatremia
J Clin Epidemiol
(2003)
E.H. Locati et al.
Spontaneous sequences of onset of torsade de pointes in patients with acquired prolonged repolarization: quantitative analysis of Holter recordings
J Am Coll Cardiol
(1995)
J.P. Hanrahan et al.
Terfenadine-associated ventricular arrhythmias and QTc interval prolongation: a retrospective cohort comparison with other antihistamines among members of a health maintenance organization
Ann Epidemiol
(1995)
J.A. Staffa et al.
Risk of selected serious cardiac events among new users of antihistamines
Clin Ther
(1995)
A.M. Walker et al.
The risk of serious cardiac arrhythmias among cisapride users in the United Kingdom and Canada
Am J Med
(1999)
Committee for Proprietary Medicinal Products (CPMP)
Note for guidance on definitions and standards for expedited reporting (CPMP/ICH/3945/03)
(2003)

J.G. Donahue et al.

Inhaled steroids and the risk of hospitalization for asthma

JAMA

(1997)

Cited by (52)

Clinical and Economic Burden of Out-of-Hospital Cardiac Arrest in US Commercial Insurance Population (2014 to 2019)
2022, American Journal of Cardiology
Little is known about the economic burden incurred by out-of-hospital cardiac arrest (OHCA) in the US commercial insurance setting. We used IBM MarketScan Commercial Claims and Encounters Database (January 2014 to March 2019) to identify patients hospitalized with OHCA based on the International Classification of Diseases codes. Patients who survived the initial OHCA episode were stratified by prognosis based on discharge setting and classified into mild (discharged home), moderate (skilled nursing facility), severe (inpatient rehabilitation or long-term hospital), and very severe (hospice) prognosis groups, respectively. Patients were followed up for 12 months after discharge for health care resource utilization and medical costs, which were inflated to year 2020. Overall, 23,512 patients with OHCA hospitalization were identified, of whom 14,667 were <65 years and 60.5% were men. The incidence of OHCA per 100,000 was steady in patients <65 years over the years (17.9 in 2014; 17.5 in 2018) but among those ≥65 years, decreased from 139.7 in 2014 to 111.1 in 2018. Total medical costs 12 months after discharge generally increased with severity of prognosis, with an average for the mild, moderate, and severe prognosis group, respectively, estimated to be $52,746, $100,394, and $130,530 among patients <65 years, and $63,194, $65,794, and $70,973 among those ≥65 years. Costs were lower for those with very severe prognosis ($7,102 for <65 years; $2,553 for ≥65 years), possibly due to high mortality. In conclusion, OHCA continues to pose a substantial clinical and economic burden on patients and the US health care system, which increases with the severity of disease prognosis.
Association of Physical Activity With Primary Cardiac Arrest Risk in the General Population: A Nationwide Cohort Study of the Dose-Response Relationship
2022, Mayo Clinic Proceedings
Citation Excerpt :
To minimize the strength of this problem, we applied the definition that was validated in 2 independent tertiary hospitals, and the diagnostic reliability of the cohort data was deemed high. The validity of the identification of cardiac arrest using ICD codes has been established in several studies, showing reliability and high positive predictive value.33-35 Third, we were unable to examine the levels of occupational physical activity or sedentary time in the current study.
To quantify the dose-response relationship between moderate to vigorous physical activity and primary cardiac arrest (PCA).
There were 504,840 participants older than 18 years who underwent the Korean National Health Screening Program, including a self-administered questionnaire for physical activity from January 1, 2009, through December 31, 2014. Physical activity levels were converted into metabolic equivalent tasks (METs) per week and categorized to correspond with multiples of public health recommendations. We evaluated the quantitative and categorical dose-response relationship between physical activity and PCA.
A curvilinear dose-response relationship between physical activity and PCA was observed; the benefits started at two-thirds (5 MET-hour/week) of the United States and World Health Organization guidelines–recommended minimum (7.5 MET-hour/week) and continued to 5 times (40 MET-hour/week) the recommended minimum (P nonlinearity <.001). The largest benefit was noted at a level of 2 to 3 times the recommended minimum (hazard ratio, 0.6; 95% CI, 0.4 to 0.8). In addition, there was no evidence of an increased PCA risk at a level more than 5 times the recommended minimum (hazard ratio, 0.7; 95% CI, 0.5 to 1.1). These associations were consistent regardless of age, sex, body mass index, comorbid conditions, and estimated 10-year risk for cardiovascular disease.
The beneficial effect of physical activity on PCA started at two-thirds of the recommended minimum and continued to 5 times the recommended minimum. No excess risk for PCA was present among individuals with activity levels more than 5 times the recommended minimum regardless of cardiovascular disease or lifestyle risk factor presence.
Sudden cardiac arrest in epilepsy patients undergoing continuous video electroencephalogram monitoring: The national inpatient sample
2020, Epilepsy Research
To examine the relationship between epilepsy and sudden cardiac arrest (SCA) and identify clinical and healthcare system related predictors of SCA in patients with a discharge diagnosis of epilepsy undergoing continuous video EEG (cVEEG) monitoring.
The national inpatient sample was used as data source to identify adults (18 years and older) with a primary discharge diagnosis of epilepsy who were at some point during their hospitalization on cVEEG monitoring. We applied a logistic regression model to identify independent patient-related and hospital/healthcare system-related factors associated with SCA.
A total of weighted 10,059 (0.71 %) patients with epilepsy on cVEEG had a secondary discharge diagnosis of SCA. The main independent factors associated with SCA were the presence of any of the following secondary discharge diagnoses: paroxysmal arrhythmia (OR: 2.29, 95 %CI: 1.96–2.66), myocardial infarction (OR: 3.78, 95 %CI: 2.83–5.05), congestive heart failure (OR: 2.27, 95 %CI: 1.93–2.62), and anoxic brain injury (OR: 57.6, 95 %CI: 50.83–67.27). There was no association between refractory epilepsy and SCA (OR: 0.99, 95 %CI: 0.51–1.93).
SCA is a rare event occurring in < 1% of patients with epilepsy undergoing cVEEG monitoring in the United States. Key independent contributors to occurrence of SCA are presence of select cardiovascular conditions and anoxic brain injury.
Contemporary impacts of a cancer diagnosis on survival following in-hospital cardiac arrest
2019, Resuscitation
The objective of this study was to determine whether survival and post-arrest procedural utilization following in-hospital cardiac arrest (IHCA) differ in patients with and without comorbid cancer.
We retrospectively reviewed all adult (age ≥18 years old) hospital admissions complicated by IHCA from 2003 to 2014 using the National Inpatient Sample (NIS) dataset. Utilizing propensity score matching using age, gender, race, insurance, all hospital level variables, HCUP mortality score, diabetes, hypertension and cardiopulmonary resuscitation use, rates of survival to hospital discharge and post-arrest procedural utilization were compared.
From 2003 to 2014, there were a total of 1,893,768 hospitalizations complicated by IHCA, of which 112,926 occurred in patients with history of cancer. In a propensity matched cohort from 2012 to 2014, those with cancer were less likely to survive the hospitalization (31% vs. 46%, p < 0.0001). Following an IHCA, rates of procedural utilization in patients with cancer were significantly less when compared to those without a concurrent malignancy: coronary angiography (4.0% vs. 13.0%), percutaneous coronary intervention (2.2% and 8.0%), targeted temperature management (0.8% vs. 6.0%); p < 0.0001 for all comparisons. This patient population was less likely to have acute coronary syndrome (12.6% vs. 27.0%) or congestive heart failure (24.5% vs. 38.2%); p < 0.0001 for both comparisons. Survival improved in both groups over the study period (p < 0.0001).
Patients with a history of cancer who sustain IHCA are less likely to receive post-arrest procedures and survive to hospital discharge. Given the expected rise in the rates of cancer survivorship, these findings highlight the need for broader application of potentially life-saving interventions to lower risk cancer patients who have sustained a cardiac arrest.
The risk of cardiac events in patients who received concomitant levofloxacin and amiodarone
2019, International Journal of Infectious Diseases
Levofloxacin and amiodarone are both known to prolong the QT interval. This study was conducted to estimate the risk of cardiac events in patients receiving concomitant levofloxacin and amiodarone.
The study included patients who were admitted to a large academic community medical center from 1/2012 to 12/2015 and received both levofloxacin and amiodarone at some point during their hospitalization. Patients received concomitant or non-concomitant levofloxacin and amiodarone during hospitalization. The primary outcome was the occurrence of cardiac events during therapy. The secondary outcome was the proportion of patients with an electrocardiogram performed before and after initiation of therapy. Odds ratios for cardiac events were calculated using a multivariable logistic regression model with and without adjusting for the study variables. The concomitant group was further evaluated for predictors of the primary outcome using multivariable logistic regression.
This study included 240 patients, 164 (68.3%) of whom received concomitant levofloxacin and amiodarone. Concomitant medication therapy was associated with a greater than six-fold increased risk of cardiac events after adjusting for the study variables (Odds Ratio = 6.20; 95% Confidence Interval = 1.34–28.62).
Patients receiving concomitant amiodarone and levofloxacin experienced a five-fold increase in cardiac events compared to patients given either medication alone.
Long-term risk of seizures among cardiac arrest survivors
2018, Resuscitation
The long-term risk of seizures in cardiac arrest survivors is not established. We hypothesized that survivors of cardiac arrest face an increased long-term risk for seizures.
We performed a retrospective cohort study using 2008–2015 claims data from a nationally representative 5% sample of Medicare beneficiaries ≥66 years-old. Our exposure of interest was a hospital diagnosis code of cardiac arrest, defined by previously validated ICD-9-CM codes. Since we were interested in long-term risk, we excluded patients with a history of seizure, and those who were diagnosed with a seizure during hospitalization or died during the index hospitalization. Our outcome was a diagnosis of seizure. Survival statistics were used to calculate seizure incidence and Cox proportional hazards models were used to determine the association between cardiac arrest and long-term seizures after adjustment for demographics and Charlson comorbidities.
Among 1,764,508 beneficiaries with a mean 4.5 years of follow-up, we identified 57,437 patients with cardiac arrest who survived to discharge without a seizure. The annual incidence of seizures was 1.26% (95% confidence interval [CI], 1.20–1.33%) compared to 0.61% (95% CI, 0.61–0.62%) in other Medicare patients. In unadjusted analysis, cardiac arrest was associated with an increased risk of post-discharge seizures (hazard ratio [HR], 1.8; 95% CI, 1.7–1.9), but the association was lost after adjustment for demographics and comorbidities (HR 0.9; 95% CI, 0.9–1.0; P = 0.12).
The long-term risk of seizures was not elevated in patients with cardiac arrest who survived to hospital discharge without a seizure.

View all citing articles on Scopus

View full text

Brief ReportHospital discharge diagnoses of ventricular arrhythmias and cardiac arrest were useful for epidemiologic research

Abstract

Objective

Methods

Results

Conclusion

Introduction

Section snippets

Setting

Results

Discussion

Acknowledgments

J Clin Epidemiol

J Clin Epidemiol

J Clin Epidemiol

Am Heart J

J Clin Epidemiol

J Am Coll Cardiol

Ann Epidemiol

Clin Ther

Am J Med

Note for guidance on definitions and standards for expedited reporting (CPMP/ICH/3945/03)

Inhaled steroids and the risk of hospitalization for asthma

JAMA

Brief Report
Hospital discharge diagnoses of ventricular arrhythmias and cardiac arrest were useful for epidemiologic research