Costs and health-related quality of life in Alpha-1-Antitrypsin Deficient COPD patients

Data was obtained from the baseline examination of the German multicentre COPD cohort COSYCONET (German COPD and Systemic Consequences – Comorbidities Network). The study included 2,741 subjects with a clinical diagnosis of COPD. Recruitment took place in 31 study centres between September 2010 and December 2013. Participants had to be ≥40 years of age and stable. Detailed information on recruitment and exclusion criteria can be found in Karch et al. [16]. COPD was defined according to the GOLD criteria after performing standardized post-bronchodilator spirometry [17]. For GOLD staging, reference values according to the Global Lung Initiative (GLI) were used [18]. 450 subjects with a Tiffeneau index above 70% at baseline examination were excluded from this analysis (Fig. 1).

AATD patients were identified by using self-reports on AATD and its genotype, AT from the medication list and the measured AAT serum levels. Participants with the heterozygous genotype PiMZ were classified as COPD patients without AATD. If a person answered that they did not know his or her genotype and was not receiving AT, an AAT serum level cut-off-point of <50 mg/dl was selected to distinguish COPD patients with and without AATD. With n = 131 patients (6.0%), including 106 patients with AT and 25 without AT, AATD was overrepresented in COSYCONET (Fig. 1).

Healthcare utilisation was assessed using standardised questionnaires and interviews. The calculation of costs has been described previously [14]. In brief, outpatient physician visits (last 3 months, subdivided by 13 specialists), and physiotherapist visits, rehabilitation and days spent in hospital (past 12 months) were considered for the calculation of direct healthcare costs. These quantities were multiplied by the German unit costs (Additional file 1) of the year 2012 [19]. Furthermore, currently used prescribed pharmaceuticals were considered. Utilization of pharmaceuticals was estimated based on patients’ information on the drug code, defined daily doses [20] and German pharmacy retail prices of the year 2012 [21]. For the calculation of costs per year all costs were extrapolated to 12 months. In case of augmented AATD patients, the estimated mean annual costs of AT, €72,255 [21], were subtracted from the medication costs. Indirect costs as a consequence of production losses [22] included the temporary inability to work and premature retirement for participants aged < 65 years. For cost calculations the human capital approach was applied that accounts full labour costs for all sick days (€177 per missed day in 2012) [22]. Premature retirement (<65 years) was valued with mean annual labour costs of €37,126 per year [23]. As there is an ongoing discussion about the appropriate valuation of indirect costs [22], the friction cost approach was additionally applied. This means, that missed working days were calculated at only €142 and only for the first 50 days of work absenteeism, and costs for premature retirement were omitted [24].

Participants completed the Saint-George’s-Respiratory-Questionnaire (SGRQ) [25] and the COPD Assessment Test (CAT) [26]. The scoring of the SGRQ ranges from 0 to 100 whereas CAT ranges from 0 to 40 with higher scores indicating worse HRQL [25, 26]. The generic EuroQol 5 dimensions (EQ-5D-3 L) questionnaire including Visual Analogue Scale (VAS) was used in addition as an established German EQ-5D-3 L utility tariff [27]. This has been demonstrated as suitable for cost-utility analyses [28]. In both parts of the EQ-5D-3 L, higher values are associated with a better HRQL.

Participants were asked: “Has a physician ever diagnosed you with one of the following diseases?”. Thereby information on 33 pre-defined comorbidities was assessed. Comorbidities were added up to a comorbidity count. Such a count has been shown to perform well in quantifying comorbidity in COPD patients [29].

For comparing characteristics between AATD patients and COPD control subjects, or between the two subgroups of AATD patients, t-tests for continuous, Chi²-tests for categorical variables, and Wilcoxon signed-rank tests for ordinal variables were applied.

The association of AATD and AT with direct and indirect costs was examined via generalised linear regression modelling (GLM). The covariates GOLD grade 1–4, age, sex, school education, smoking status, BMI and comorbidity count were added to the model. As cost data was highly right-skewed, which means that costs only assume values ≥0 and more individuals have costs below the mean than above, a gamma model with log-link was assumed. In the case of indirect costs only participants aged younger than 65 years were included. Participants with zero-costs were assigned €1 for regression modeling. Resulting cost estimates can be directly interpreted as factors. Adjusted means in direct and indirect costs were estimated as recycled predictions from 1000 bootstrap replications [30]. GLM with normal distribution was performed for HRQL scores. Therefore, estimates for HRQL represent additional units on the HRQL scale.

Statistical analyses were performed using SAS software (SAS Institute Inc., Cary, NC, USA, version 9.3). For significance testing, a p-value of 0.05 or less was required. Graphics were edited in R [31] using the package ggplot2 [32].

To consider potential residual confounding due to the categorization of lung function in GOLD grades 1 to 4, GLM was also performed using the exact predicted percentage of Forced Expiratory Volume in 1 second (FEV₁ %pred.) instead of GOLD grades [18].