One-year Prevalence, Comorbidities, and Cost of Hospitalizations for Alpha-1 Antitrypsin Deficiency among Patients with Chronic Obstructive Pulmonary Disease in the United States

Objectives: Little is known about severe chronic obstructive pulmonary disease (COPD) exacerbations among patients with Alpha-1 Antitrypsin Deficiency (AATD). We assessed inpatients with AATD and COPD among a sample of COPD inpatients to ascertain demographic, clinical and economic differences in the course of disease and treatment. Methods: Using data from the 2009 Nationwide Inpatient Sample (NIS), we identified COPD (ICD-9-CM: 491.xx, 492.xx, or 496.xx) patients with AATD (273.4). We compared patient demographics and healthcare outcomes (eg, length of stay, inpatient death, type and number of procedures, and cost of care) between COPD patients with and without alpha-1 antitrypsin deficiency. Frequencies and percentages for patient demographics were compared using bivariate statistics (eg, chi-square test). Recognizing the non-parametric nature of length of stay and cost, we calculated median values and interquartile ranges for these variables for each group of patients. Finally, the risk of inpatient death was estimated using logistic regression. Results: Of 840 242 patients with COPD (10.8% of the NIS sample population), 0.08% (684) had a primary or secondary diagnosis code for AATD. COPD+AATD were younger (56 vs 70, p<0.0001) and as a result, less likely to be covered by Medicare (44% vs 62%, p<0.0001). AATD patients were also more likely to have comorbid non-alcoholic liver disease (7% vs 2%, p<0.0001), depression (17% vs 13%, p=0.0328), and pulmonary circulation disorders (7% vs 4%, p=0.0299). Patients with AATD had a 14% longer length of stay (IRR = 1.14, 95% CI 1.07, 1.21) and a mean cost of $1487 (p=0.0251) more than COPD inpatients without AATD. Conclusions: AATD is associated with increased mean length of stay and cost, as well as higher frequency of comorbid non-alcoholic liver disease, depression, and pulmonary circulation disorders. Future research should assess other differences between AATD and the general COPD population such as natural history of disease, treatment responsiveness and disease progression.


INTRODUCTION
Alpha-I-antitrypsin (AAT), the main protease inhibitor in human serum 1 , is produced in the liver and reaches the lungs through circulatory diffusion. 2In healthy individuals, AAT acts to defend against elastolytic enzymes in the lower airways 3 , which are a major contributing factor to emphysema. 4The release of AAT is up-regulated during inflammation, infection, cancer, and pregnancy. 5AAT deficiency (AATD) is caused by mutations of the SERPINA I gene encoding AAT-Z 6 , and presents with varying ranges of AAT serum concentrations, which has shown to be inversely related to the degree of disease severity.Over 100 alleles and more than 40 phenotypes have been identified, of which null-null and ZZ phenotypes have the lowest serum AAT concentrations, followed by SZ, SS, MZ, MS, and finally MM. 6 Resulting disease outcomes include pulmonary diseases, such as emphysema and bronchiectasis 7 , liver disease (eg, chronic hepatitis, cirrhosis and hepatoma) 8 , and, in rare cases, panniculitis. 9AATD is underrecognized and there are often long delays between symptom onset and diagnosis. 5,102][13] The first established the genetic predisposition in young adults for early-onset emphysema. 11The second established the preponderance of early-onset cirrhosis in children with marked decreases in alpha-1 globulin in serum protein electrophoresis patterns. 12ucasians in Europe and North America show the highest risk of mutation and, hence bear the predominant burden of AATD. 14An estimated 0.11% of Caucasian populations in the United States have AATD. 5,14An indirect epidemiologic study conducted in 2007 estimated US prevalence of AATD at 33,088 individuals.6][17] AATD is usually detected via genetic testing when patients present symptomatic liver and/or lung disease and report a familial history of AATD. 18This type of exhaustive enquiry is not always performed in all clinical settings and hence AATD continues to be under-recognized and its development is not well understood.Fewer than 10,000 individuals are currently receiving therapy for AAT augmentation, while prevalence estimates suggest this figure should be closer to 70 000 individuals in the United States, which lends support for the assertion that AATD is under-recognized by the medical community. 18ronic obstructive pulmonary disease (COPD) is a progressive, debilitating respiratory disease and includes emphysema and chronic bronchitis. 19The clinical characteristics of COPD comprise difficulty breathing, lung airflow obstructions, and nonproductive cough. 19Current prevalence estimates suggest that 6.3% of the US adults have COPD. 19COPD is influenced by both genetic and environmental factors 20 , suggesting AATD may contribute significantly to the pathogenesis of COPD in many patients.Previous studies have reported significant correlations between AATD and pulmonary function. 16,21Moreover, people with AATD are at an increased risk for developing early-onset COPD. 22 this study, we assessed the prevalence of AATD as well as evaluated the demographic, clinical characteristics and cost of hospitalization associated with identifiable and symptomatic AATD in a hospital discharge population with COPD.We also compared COPD patients with AATD to COPD without a recorded AATD diagnosis.

METHODS
We obtained and analyzed data from the 2009 Nationwide Inpatient Sample (NIS).The NIS contains 100% of the discharge records from 20% of community hospitals in the United States during 2009.The NIS includes information on each discharge, such as patient demographics, hospital characteristics, up to 25 diagnoses, up to 15 procedures, and total charges billed to individual patients.Using the cost-to-charge ratio provided by the NIS, it is possible to estimate the cost of care for each patient encounter.Discharge records in the NIS have been de-identified, so each encounter is treated as a unique event.
We defined our cohort as patients with COPD, limiting our definition of the disease to two conditions: chronic bronchitis and emphysema.The cohort was identified through International Classification of Diseases, Version 9, Clinical Modification (ICD-9-CM) diagnosis codes included in the discharge record.Within this cohort of patients, we identified patients that also had an AATD diagnosis code.Our classification scheme is depicted in Figure 1.(Appendix A includes a list of ICD-9-CM codes used in our classification scheme.)Since we sought to characterize the population of COPD patients with AATD diagnosis compared to those without a recorded AATD diagnosis, we described the two groups using basic descriptive statistics (eg frequency and proportions) and used bivariate statistics to ascertain differences.We also used multivariable regression models to better assess the incremental effect of AATD on healthcare outcomes while adjusting for patient characteristics, comorbidities, and healthcare use.Using the included sample weights, we calculated national hospitalization incidence estimates for both groups of patients.We compared patient demographics and healthcare outcomes (eg, length of stay, inpatient death, type and number of procedures, and cost of care) between COPD patients with and without AATD.Frequencies and percentages for patient demographics were compared using bivariate statistics (eg, chi-square test for binary and categorical variables, and t-test for continuous variables).As length of stay and cost tend not to be normally distributed, we calculated median values and interquartile ranges for these variables for each group of patients, comparing the values between the groups with the Wilcoxon-Mann-Whitney test.Multivariable models were used to assess the association between AATD and hospital outcomes including length of stay, risk of death and total inpatient cost while adjusting for confounding patient and hospital characteristics.We used a negative binomial regression model to assess length of stay, a logistic regression model to assess risk of inpatient death and a Gamma generalized linear model with log link to assess total inpatient cost.Cost was modeled with the assumption that our continuous outcome data would be non-negative and heteroskedastic.

RESULTS
Over 4.25 million COPD-related hospitalizations occurred in 2009.We estimated the national hospitalization incidence for COPD with AATD at 3426 patients (Table 1).COPD patients with AATD were roughly 14 years younger, on average, than COPD patients without AATD (55.79 vs 69.77 years, p<0.0001).Overall, approximately 4.0% of hospitalizations for COPD, regardless of AAT sufficiency, resulted in death.While fewer than ten COPD patients with AATD were hospitalized with a minor loss of function, 3.6% of COPD patients without a recorded AATD diagnosis entered the hospital with a minor loss of function (p<0.0001).COPD patients with AATD were more likely to be white than COPD patients without an AATD diagnosis (76.6% vs 69.7%, p<0.0001).Additionally, COPD patients with AATD were more likely to have private health insurance than COPD patients without an AATD diagnosis (23.1% vs 14.3%, p<0.0001).
Consistent with the distribution of inpatient deaths reported in Table 1, AATD was not associated with a statistically significant increase in the risk of inpatient death (Table 3) while adjusting for age, sex, region, hospital location (urban/rural), primary payer, loss of function, median household income and race.However, AATD was associated with a 14% increase in length of hospital stay (IRR: 1.14 [1.07-1.21])(Table 3) and $1487 greater cost (p=0.0251)after adjustment.

DISCUSSION
The current analysis compared demographic and clinical characteristics of a hospital discharge population comprising COPD patients with AATD versus patients with non-AATD associated COPD.Using data from the 2009 Nationwide Inpatient Sample, the results suggest important differences between these patient groups.
COPD inpatients with a diagnosis of AATD have been reported to be younger than COPD inpatients without evidence of AATD. 23,24In our sample, COPD patients with AATD were approximately 14 years younger than non-AATD CPOD patients.There are two possible explanations for this finding.First, patients with AATD may develop COPD at a much younger age.Previous reports have indicated that the dominant age group of AATD patients with COPD fall into a younger age group than AATD patients without COPD (49.1% aged 59 years or younger compared to 41.9% aged 60 years or older, respectively). 23Second, patients with AATD may experience more severe symptoms arising from COPD, therefore exacerbations are more likely to occur, prompting the patient to seek medical care.A registry of individuals with severe AATD from March 1989 to October 1992 indicated that AATD participants were more likely to have severe pulmonary function impairment for their age. 25Furthermore, an examination by Campos et al. found that patients with AATD, despite augmentation therapy (AT), experienced a high number of exacerbations with negative impact on quality of life and health resource utilization (85% of participants). 26However, without accurate AATD prevalence estimates, and therefore questionable classification of non-AATD associated COPD vs AATD associated COPD, it is difficult to make these claims.
COPD inpatients with a diagnosis of AATD in the current study presented similarly to those of younger, healthier COPD patients when compared to COPD inpatients without AATD.They were more likely to be of working age, to have private insurance and less comorbidities.Conversely, a study by McGrady et al .27found that AATD patients reported several comorbidities and with a higher frequency, although more AATD patients reported no comorbidities overall in comparison with COPD patients.However, McGrady's study noted possible recall bias or over/under-reporting, as it was a patient self-reported study. 27TD patients in the current study were more likely to have non-alcoholic liver-disease than their non-AATD counterparts.Research has shown an increase of liver disease in AATD in direct correlation with age and increased risk of cirrhosis and primary liver cancer in patients homozygous for AATD.29 It is well understood that liver manifestations primarily affect patients with the ZZ genotype (~30% adults, ~2.5% children), although MZ, SZ, and Z-null genotype patients are also at a (low) risk of developing liver disease.Across genotypes, approximately 15% of AATD patients present with liver dysfunction, which may include hepatocarcinoma, cirrhosis, cancer, hepatitis or liver failure.30,31

Table 3 .
Outcomes Associated with AATD