Comorbidity and Mortality in COPD-Related Hospitalizations in the United States, 1979 to 2001^* FREE TO VIEW

COPD is the fourth-most-common cause of death in adults in the United States and is projected to be the third-most-common cause of death in both men and women by the year 2020.¹ COPD is also a leading cause of hospitalizations in adults in the United States, particularly in older populations.² Even though the prevalence of COPD has increased in the last 20 years,³ COPD is infrequently mentioned as a contributing or underlying cause of death even in patients with severe disease.⁴⁵ Although this has been attributed to an underreporting bias, it is also possible that COPD increases the risk of dying from other comorbid conditions. Since nonrespiratory diseases account for > 50% of the underlying causes of death in COPD,⁵ it is likely that the impact of COPD as a health burden is substantially underestimated. We hypothesized that patients discharged from the hospital with a diagnosis of COPD have a higher prevalence and in-hospital mortality from comorbidities when compared to patients discharged without a diagnosis of COPD. To determine this, we analyzed the prevalence and mortality from selected comorbidities in patients with and without a diagnosis of COPD using the National Hospital Discharge Survey (NHDS) [1979 to 2001].

The NHDS is a national survey that has been conducted continuously since 1965, and it provides data on inpatient utilization of non-Federal, short-stay hospitals in the United States. NHDS data are collected from a sample of inpatient records acquired from a national probability sample of hospitals. Because persons with multiple discharges during the year can be sampled more than once, the NHDS produces estimates for discharges, not persons. Only general hospitals, children’s general hospitals, or hospitals with an average length of stay < 30 days for all patients are eligible for inclusion in the survey. Federal, military, and Department of Veterans Affairs hospitals, as well as hospital units of institutions (such as prison hospitals), and hospitals with fewer than six beds staffed for patient use, are excluded. The NHDS data contain items that relate to the personal characteristics of the patient: birth date (converted to age), sex, race, ethnicity, marital status, zip code, and expected sources of payment. Administrative items such as admission and discharge dates, admission type and source, discharge status, and medical record number are also included. The medical information includes final medical diagnoses, procedures performed, and dates of surgery. A total of seven medical diagnoses per patient are obtained, and are coded according to the International Classification of Diseases, Ninth Revision (ICD-9).⁶⁷ The NHDS usually presents diagnoses and procedures in the order they are listed on the abstract form or obtained from abstract services. The first diagnosis listing usually represents the primary or main diagnosis for a particular hospital discharge, whereas the second to seventh listings represent the secondary or comorbid diagnosis (the order in which secondary diagnoses are listed is not determined by their degree of severity or clinical importance). In 1991, the ICD-9 codes for COPD were changed from 491.2 (obstructive chronic bronchitis: emphysematous, obstructive, and bronchitis with chronic airway obstruction, emphysema) to 491.20 (emphysema with chronic bronchitis without exacerbation) and 491.21 (COPD with acute exacerbation).

We used ICD-9 codes 490–492 and 496 to define COPD, and included COPD as either primary (first diagnosis listing) or secondary discharge diagnosis (from the second to the seventh diagnosis listing). We evaluated whether having a primary or secondary diagnosis of COPD in adults ≥ 25 years old was associated with increased prevalence and mortality from the following selected comorbidities: pneumonia (ICD-9 480–487.8); hypertension (ICD-9 401–406); diabetes (ICD-9 250); heart failure (ICD-9 428); ischemic heart disease (ICD-9 410–414); pulmonary vascular disease (ICD9 415–417, which includes acute cor pulmonale, pulmonary embolism, primary and secondary pulmonary hypertension, and cor pulmonale not otherwise specified); thoracic malignancies (ICD-9 160–165); and ventilatory failure (defined as acute respiratory failure [ICD-9 518.81]; pulmonary insufficiency following trauma or surgery [ICD-9 518.5]; asphyxia and respiratory arrest [ICD-9 799], respiratory distress [ICD-9 786.09]); acute renal failure (ICD-9 584); chronic renal failure (ICD-9 585–586); HIV (ICD-9 042); cerebrovascular accident (ICD-9 430- 438); and GI bleeding (ICD-9 578). To evaluate for any possible time-trend effects, we performed these analysis in five periods: 1979 to 1984, 1985 to 1988, 1989 to 1992, 1993 to 1996, and 1997 to 2001. Relative SEs were calculated using the weight for all-listed diagnosis provided in the 2000 NHDS data file documentation. We used the χ² test for differences between proportions and the Cochrane-Armitage test for trend using categorical data. For direct age adjustment using 2000 US population weights, a software program (PROC DESCRIPT in SUDAAN; Research Triangle Institute; Research Triangle Park, NC) was used. For analysis, we used statistical software (SAS version 8.0; SAS Institute; Cary, NC; and SUDAAN; Research Triangle Institute); p < 0.05 was considered statistically significant.

During 1979 to 2001, there were an estimated total of 47,404,700 hospital discharges with a diagnosis of COPD as either the primary or secondary discharge diagnosis (8.5% of all hospitalizations in adults > 25 years old), of which 37,540,374 discharges (79.2%) were listed with COPD as a secondary diagnosis and 9,864,278 discharges (20.8%) were listed with COPD as the primary discharge diagnosis. From 1979 to 2001, the yearly prevalence of hospital discharges with a diagnosis of COPD increased significantly (p for trend < 0.01), mainly due to the proportion of hospital discharges with COPD listed as a secondary diagnosis (mean increase per year, 0.3%; 95% confidence interval, 0.30 to 0.34), in comparison to discharges where COPD was listed as primary diagnosis (mean increase per year, 0.05%; 95% confidence interval, 0.02 to 0.07) [Fig 1] . The proportion of in-hospital mortality among discharges with and without COPD increased significantly with age (Table 1 ). The age-adjusted hospital prevalence of pneumonia, hypertension, heart failure, ischemic heart disease, pulmonary vascular disease, thoracic malignancies, and ventilatory failure were higher among hospital discharges with a diagnosis of COPD when compared to discharges without COPD (Fig 2 ). Further, a diagnosis of COPD was associated with higher age-adjusted in-hospital mortality for pneumonia, hypertension, heart failure, ventilatory failure, and thoracic malignancies when compared to hospital discharges with these comorbidities without a diagnosis of COPD (Fig 3 ). A discharge diagnosis of COPD was not associated with higher hospitalization prevalence or in-hospital mortality for acute and chronic renal failure, HIV, GI hemorrhage, and cerebrovascular disease (data not shown). Figure 4 shows the hospitalization prevalence and in-hospital mortality trends for ischemic heart disease, congestive heart failure, and pneumonia in patients discharged with and without COPD.

COPD is one of the leading causes for mortality and morbidity in the United States, yet little is known about the prevalence of comorbid conditions and specific causes of death in hospitalized patients with COPD.⁸ In this analysis, hospital discharges with primary or secondary COPD were also more frequently diagnosed with other comorbid conditions, including cardiac and pulmonary vascular disease, pneumonia, and thoracic malignancies. Also, the in-hospital mortality from congestive heart failure, hypertension, ischemic heart disease, and thoracic malignancies was higher among hospital discharges with any mention of COPD.

Progressive respiratory failure accounts for approximately one third of the COPD-related mortality; therefore, factors other than progression of lung disease must play a substantial role.⁹¹⁰ For example, a long-standing history of tobacco abuse in COPD patients may increase the risk for comorbidities such as cardiovascular disease and cancer. Some of the most common comorbid conditions that have been described in association with COPD include hypertension, diabetes, coronary artery disease,¹⁰¹¹ heart failure,¹² pulmonary infections, cancer, and pulmonary vascular disease.¹³ In a study¹⁴ of 312,664 decedents from England and Wales during 1993 to 1999, COPD was mentioned in 8.0% of all death certificates. In these death certificates, obstructive lung disease comprised 59.8% of the underlying cause of death. When obstructive lung disease was not the underlying cause of death, the most common causes of death were similar to those presented in our analysis (ischemic heart diseases, heart failure, malignant neoplasms of the lung, and bronchopneumonia).¹⁴

The number of preexisting comorbidities in patients with COPD has been associated with increased in-hospital mortality in a cross-sectional study¹⁵ of 71,130 patients admitted for COPD exacerbation. The in-hospital mortality from COPD exacerbations was 2.5%, which is considerably lower than the 5.9% in this study; this difference is probably explained by our broader inclusion of any mention of COPD in the discharge diagnoses. Comorbid conditions that have been associated with an increased mortality risk in COPD patients include chronic renal failure, cor pulmonale, and pulmonary vascular disease¹⁶; underlying heart diseases have not been consistently associated with a higher mortality risk.¹⁰ However, since COPD is frequently underreported, it is difficult to have an adequate estimate of how comorbid conditions influence COPD mortality or, inversely, how COPD affects the outcome of other diagnosis.⁴ For example, in a 22-year follow-up study of 5,542 adults in the first National Health and Nutrition Examination Survey, only 47.7% of patients with severe COPD at baseline had COPD listed in the death certificate, and only 23.1% had COPD listed as the underlying cause of death.⁵

The results from this study have several implications. First, this study shows that the burden of disease associated with COPD is largely underestimated, since having a diagnosis of COPD is associated with increase risk for hospitalization and in-hospital mortality from other common diagnoses. Second, the increase in the number of COPD-related hospital discharges since 1979 has been largely due to COPD discharges listed as secondary diagnosis relative to COPD discharges listed as primary diagnosis. We believe that any combination of the following factors could account for this finding: (1) outpatient treatment may reduce the incidence of COPD exacerbations requiring hospitalization¹⁷ (which are also more likely to be labeled as a primary discharge diagnosis); however, outpatient treatment may not reduce the effect that COPD has on the risk of being hospitalized for other comorbid conditions; (2) only the most severe cases of COPD are recognized during hospitalization,¹⁸ thereby underestimating the number of hospital discharges where moderate or mild COPD could also be labeled as primary diagnosis; and (3) even in patients with severe COPD, a large proportion of patients are admitted to the hospital for other comorbidities,¹⁰ and therefore COPD is labeled as a secondary diagnosis.

Third, after adjusting for age, patients with COPD appear to have an increasing trend in the hospitalization and mortality prevalence for pneumonia, congestive heart failure, and ischemic heart disease when compared to patients without COPD (Fig 4). This finding might have important implications for developing prevention strategies such as influenza and pneumonia vaccination, and long-term oxygen treatment.

Several limitations must be considered when interpreting the results from this study. Our study has several potential sources for exposure misclassification: first, the use of ICD-9 codes and the ability of physicians to adequately diagnose COPD in hospitalized patients. Since ICD-9 coding for COPD has a low degree of sensitivity (29 to 53%)¹⁹ and physicians often fail to recognize mild-to-moderate cases of COPD in hospitalized patients,¹⁸ this would result in a selection bias that would underestimate the true prevalence of COPD-related hospital discharges; however, this bias would not affect the relationship between COPD and comorbid conditions observed in our analysis since this is nondifferential with regards to the presence of comorbidities. The association between COPD must be taken with caution due to residual confounding from a lack of adjustment for other covariates (gender, disease severity, medications used, and occupational exposure, among others). Further studies are needed to characterize the association of COPD and how it relates to the prevalence and in-hospital mortality from other comorbidities (ie, degree of airway obstruction, smoking history, functional status, and body mass index). Third, our study cannot address whether or not there is a causal association between a diagnosis of COPD and the selected comorbidities used in the analysis. Fourth, no stratified analyses were done to determine the effect of COPD according to race, type of hospital, and insurance; therefore, these results may not be applicable across all ethnic groups, which differ in health-care access and socioeconomic status. Fifth, the increasing trends in the hospitalization prevalence and in-hospital mortality for ischemic heart disease, congestive heart failure, and pneumonia in patients with COPD should be taken with caution, since it is not possible to determine the extent to which other factors, such as changes in ICD coding patterns over time, could have affected these estimates.

As shown in this study, a discharge diagnosis of COPD is associated with higher prevalence and in-hospital mortality from selected comorbid conditions. Efforts toward earlier detection and treatment of COPD may result in decreasing the burden of disease.