INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: obstructive lung diseases; exacerbations; computed tomography; health status

Patients with ₁-antitrypsin deficiency (₁-ATD) of the PiZ phenotype are at increased risk of developing emphysema, particularly if they smoke (1). Observational data suggest that this process may be slowed by intravenous augmentation therapy with a purified preparation of human ₁-antitrypsin (2), possibly by an effect on exacerbations (3, 4). However, the efficacy of this intervention has never been proven in a randomized placebo-controlled study. This is primarily because power calculations using FEV₁ as the main outcome indicated that such a trial would be logistically and economically prohibitive (5). However, quantitative computed tomography scanning may be a more sensitive marker of progression of emphysema in these patients and thus may facilitate the investigation of an effective treatment (6).

Whereas prevention of physiological and radiological deterioration is an important therapeutic goal, patients are most interested in symptomatic benefits. Patients with ₁-ATD have extensive physiological abnormalities that relate to their health status (7). However, there are no data published currently regarding health status decline and factors, other than smoking, that influence this process in ₁-ATD.

The aim of the current study was to document the natural history of lung disease and its consequences in a group of patients with emphysema and ₁-ATD who never received augmentation therapy. In particular, we monitored prospectively the decline not only in FEV₁ but also in other measures of lung function, high-resolution computed tomography (HRCT), and health status with a view to assessing their suitability as evaluative instruments in future trials. Finally, we examined a number of factors including exacerbation frequency that may influence the rate of decline in these patients.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

The ₁-AT level and phenotype were confirmed by immunoassay and isoelectric focusing, respectively, in a central U.S. laboratory (Heredilab, Salt Lake City, UT) using a dried finger prick blood spot. At the time of analysis, 45 patients with the PiZ phenotype and airflow obstruction (prebronchodilator FEV₁ < 80% predicted and FEV₁/FVC ratio of less than 0.7) had completed three annual assessments over 24 mo. Two of these were excluded due to comorbid disease, one with liver cirrhosis who underwent liver transplantation during follow up and a second with fibrotic lung disease. Eleven subjects with emphysema and airflow obstruction had withdrawn (n = 8) or died (n = 3) during the study period, having completed only one (n = 5) or two (n = 6) assessments.

Clinical History

A full clinical history was obtained with particular attention to the presence of chronic sputum expectoration (8) and the frequency of acute exacerbations. These were defined as clear episodes characterized by at least two of the following criteria: new or increased sputum volume, increased sputum purulence, and increased breathlessness that persisted for more than 48 h (9). Exacerbation data were collected at the 6-mo assessment and supported where possible by clinical notes from contact episodes between assessments.

All subjects gave written informed consent to the study, which was approved by the University of Birmingham Hospital NHS Trust Research Ethics Committee.

Lung Function Testing

All subjects performed dynamic spirometry before and after dual bronchodilatation with nebulized ₂-agonist and ipratropium bromide as described previously (7). Lung volumes were measured by helium dilution (Morgan Medical, Kent UK) and gas transfer (diffusing capacity of the lung for CO [DL_CO]) by the single breath carbon monoxide method and corrected for effective alveolar volume (DL_CO/VA). All tests were performed to British Thoracic Society/Association of Respiratory Technicians and Physiologists (BTS/ARTP) guidelines (10). In addition, an arterialized earlobe capillary blood sample was obtained to estimate arterial Pa_O2 (11).

Computed Tomography

The HRCT scanning protocol has been described in detail in a previous publication (7). Briefly, for baseline scans, 1-mm-thick slices were obtained at 10-mm intervals at full inspiration and 30-mm intervals at full expiration. The inspiratory scans were examined for the macroscopic changes of emphysema and bronchiectasis (12). The CT image consists of pixels, which represent the density contained within the corresponding 1-mm-thick volume of lung (voxel). Density mask analysis using a threshold of -910 HU was performed on single slices through the upper (at the level of the aortic arch) and lower (at the level of the inferior pulmonary vein) zones in order to quantify the extent of emphysematous tissue (13). The results were expressed as the voxel index (VI), that is, the number of low-density voxels (910 HU and below) expressed as a percentage of the total number of voxels representing lung tissue. For follow up scans, 30-mm increments were taken for both inspiratory and expiratory phases and all scans were usually performed within 24 h of lung function testing and always within 3 wk.

Reproducibility of the HRCT measurements was determined by assessing 10 patients with chronic obstructive pulmonary disease (COPD) not related to ₁-ATD (mean: age 68 [SD] = 7.5], FEV₁ 0.84 L [SD = 0.30], FEV₁ % predicted 36 [SD = 12]) on three separate occasions over a 1-mo period.

Health Status

Disease-specific health status was assessed using the St. George's Respiratory Questionnaire (SGRQ) (14, 15) and generic health status using the Short-form 36 (SF36) (16) as described previously (7). Each of the domains was scored from 0 to 100 with a high score indicating worse impairment for the SGRQ and the reverse for the SF36.

Statistical Analysis

Data were analyzed using a statistical software package (SPSS version 10.0.5). Longitudinal changes were assessed using the Repeated Measures option from the General Linear Modelling menu. Sensitivity to change for each of the measures was determined by dividing the mean decline above that expected for age (10) by the standard error of the actual decline (6). No significant deterioration in lung density is expected in normal subjects over 2 yr, and thus, for this parameter, any change seen was assumed to be related to disease progression alone (17).

To determine factors that influence disease progression, Pearson's correlation coefficients were used to identify significant bivariate relationships. Along with categorical variables (e.g., age and sex), these data were then entered as independent variables into stepwise multiple regression analysis with the change in lung function, HRCT, or health status examined as the dependent variable. Significance was accepted at the 5% level.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Baseline

Of the 43 PiZ subjects with airflow obstruction, 32 were male (74%) and 38 were current or ex-smokers (88%). Twenty-three patients (53%) described chronic sputum expectoration that fulfilled the MRC criteria for the diagnosis of chronic bronchitis (8).

The baseline lung function and HRCT data are shown in Table 1. Although a wide range of impairment was observed, the mean values indicate moderate to severe airflow obstruction, gas trapping, and reduced gas transfer consistent with pulmonary emphysema. The mean ± SD improvement in FEV₁ after nebulized ₂-agonist and ipratropium bromide was 282 ± 164 ml or an improvement of 8.7 ± 4.4 in the value expressed as a percentage of the predicted value.

Visual assessment of the HRCT scans confirmed the presence of emphysema in all cases, affecting the lower zones predominantly. This was confirmed by the greater voxel indices for the lower zones for both inspiratory and expiratory phase scans (p < 0.001, Table 1). In addition, seven subjects (16%) had HRCT evidence of bronchiectasis, although this was invariably cylindrical and limited in distribution.

Health status scores indicated marked impairment and disability at the start of the study for both the disease-specific and generic questionnaires (SGRQ total score: mean 58.2 ± [SD] 18; normal range 5-7; SF36 Physical Functioning: mean 35 ± 28, mean for U.K. population aged 55-64 = 76 [15]).

The subjects who died or withdrew were at baseline (p > 0.05) of a similar age to those who completed three assessments but, as a group, had a lower FEV₁ (29% predicted) and DL_CO/VA (49% predicted) (p < 0.05 for both comparisons).

HRCT Reproducibility

The mean voxel indices for all four scan measurements did not change over the 1-mo period (Figure 1). The average coefficients of variation for each patient ranged from 4.6% (SD ± 3.4) for the upper zone inspiratory scans to 9.3% (SD ± 3.2) for the lower zone expiratory scans reflecting the mean values for those scans. Internal consistency as assessed by Cronbach's alpha coefficient was high, ranging from 0.98 for the lower zone scans to 0.99 for the upper zone scans (inspiratory and expiratory).

View larger version (43K): [in this window] [in a new window]   Figure 1.   Short-term reproducibility of HRCT voxel index in 10 patients with usual COPD. The mean and standard error bars are displayed for the first (white), second ( gray), and third (black) scans performed over a 2-wk period. No significant difference was seen for the group data for any scan region analyzed at any two points.

Longitudinal Change

Airflow obstruction, vital capacity, gas trapping, and gas transfer all deteriorated during the study period, indicating physiological deterioration and, in particular, the FEV₁ showed a mean annual decline of 67 ml (SE ± 15 ml). During this time, the HRCT voxel indices also increased, although this achieved significance only for the upper zone scans. The mean values at each time point for the lung function and HRCT parameters are shown in Table 2.

The SGRQ total score did not decline significantly over the 24 mo (mean at baseline = 58.2 ± 18; 24 mo = 60.3 ± 15.7, p = 0.45). However, there was a consistent and significant (p = 0.001) decline of > 4 units/yr in the Activity domain of the SGRQ (Figure 2). The overall score did not reflect this change due to an improvement in the mean ± SE Symptoms score, which decreased from 73.7 ± 18.3 at baseline to 66.8 ± 19.3 at 24 mo (p = 0.02). Within the components of the Symptom score, a significant decrease in the reporting of wheeze appeared to account for this improvement (p = 0.03), with no change in any of the other parameters.

View larger version (44K): [in this window] [in a new window]   Figure 2.   Change in health status over 2 yr. The mean and standard error bars for the St. George's Respiratory Questionnaire (SGRQ) Activity domain (black) and the SF36 Physical Functioning domain (gray) are displayed. The increasing SGRQ score (p < 0.001) and the decreasing SF36 score (p = 0.01) indicate worsening health status related to physical activity in the 43 PiZ subjects.

Although there is no summary score for the SF36, there was a significant decrease in the mean ± SE score for the Physical Functioning domain from 34.9 ± 4.3 at baseline to 26.2 ± 3.6 at 24 mo (p = 0.01), consistent with the changes in SGRQ Activity (Figure 2). There was no change in the remainder of the SF36 domains with the exception of Role Emotional, which showed an improvement with the mean score increasing from 65.1 ± 41.8 to 82.2 ± 35.1 over the 2 yr (p = 0.03).

Comparison of each measurement to determine its sensitivity to change (see METHODS) indicated that the upper zone voxel index was the most sensitive followed by gas transfer (DL_CO). The data for these and other variables are summarized in Table 3.

Predictors of Decline

None of the categorical variables, including age and sex, predicted the rate of decline of FEV₁. However, there was a relationship between FEV₁ decline and the baseline (postbronchodilator) FEV₁ with the greatest change occurring in those subjects with the least initial impairment (r = 0.56, p < 0.001). In addition, FEV₁ also declined more rapidly in those subjects with greatest bronchodilator reversibility expressed as a percentage of the predicted value (r = 0.52, p < 0.001). Although bronchodilator reversibility was related to baseline FEV₁ (r = 0.49, p < 0.001), multiple regression analysis revealed that both baseline FEV₁ and bronchodilator reversibility predicted FEV₁ rate of decline independently (baseline FEV₁: r² = 0.29, baseline FEV₁ and bronchodilator reversibility combined: r² = 0.36).

Regular use of inhaled corticosteroids did not influence the decline in FEV₁ or the number of exacerbations reported, although there were too few patients (n = 8) not receiving such medication to form firm conclusions. Similarly, smoking status did not affect decline, although only three patients continued to smoke during the study and these three admitted to an average of less than five cigarettes per day.

Neither the presence of chronic bronchitis nor the frequency of exacerbations showed any relationship to FEV₁ decline in the group as a whole. However, exacerbation frequency was related to the decline in FEV₁ in patients with a baseline postbronchodilator FEV₁ > 35% predicted (n = 26, r = 0.38, p = 0.026). In addition, the number of exacerbations during the study period for the group as a whole did relate to decline in vital capacity (VC) (r = 0.50, p < 0.001, Figure 3) and DL_CO (r = 0.31, p = 0.02). Stepwise multiple regression analysis confirmed that both the number of exacerbations and the presence of chronic bronchitis were independent predictors of the decline in VC (combined adjusted r² = 0.29, p < 0.05).

View larger version (16K): [in this window] [in a new window]   Figure 3.   Relationship between change in vital capacity and exacerbation frequency. The number of patients (n), Pearson correlation coefficient (r) level of significance (p), and linear trendline are displayed.

The decline in health status as assessed by the Activity domain of the SGRQ was related to the decline in FEV₁ (r = 0.45, p = 0.002), VC (r = 0.35, p = 0.01), and DL_CO (r = 0.50, p < 0.001). When these three variables were entered as dependent factors in stepwise multiple regression analysis with SGRQ Activity decline as the dependent factor, DL_CO remained the only independent predictor (adjusted r² = 0.23).

There was no difference in the rate of decline of FEV₁, DL_CO/VA, or HRCT voxel index in the patients who completed the study and the six subjects who completed two assessments before withdrawal (p > 0.05, data not shown).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The current study describes in detail the natural history of lung disease in ₁-antitrypsin-deficient (PiZ) patients attending a single center. Over a 2-yr period, this relatively homogeneous group with established, mainly lower zone pulmonary emphysema demonstrated a significant deterioration in measurements of airflow obstruction, gas trapping, and gas transfer. This was accompanied by an increase in the extent of emphysema seen on HRCT with the clearest deterioration occurring in the upper lung fields. Furthermore, these physiological and radiological changes were accompanied by a statistically and clinically important deterioration in health status related to physical activity.

The mean rate of decline in FEV₁ for the group as a whole was 67 ml/yr, which is similar to that found by other workers in subjects with ₁-ATD (2, 6). However, the median decline was somewhat lower (45 ml/yr) reflecting the fact that function declined more slowly in the relatively large number of patients who had already developed severe COPD and more rapidly in the smaller number of patients with better lung function. This was confirmed by the correlation between the initial FEV₁ and its subsequent decline with the more severely affected patients showing the least change. The subjects with a baseline postbronchodilator FEV₁ of > 35% predicted had a decline of 101 ml/yr (SE ± 20 ml/yr). This did not appear to be due to a "survivor effect" as the limited data on those who withdrew after 12 mo revealed a similar rate compared with those who completed 24 mo of the study. Previous studies have shown that the annual decline in ex-smokers with nondeficient COPD who had an average baseline FEV₁ 75% predicted was 34 ml/yr (18). The current study therefore confirms that ₁-ATD predisposes to a more rapid average decline in FEV₁ even in ex-smokers.

FEV₁ decline was faster in those subjects with greatest bronchodilator reversibility as assessed by the improvement expressed as a percent of the predicted normal values. This change was independent of the baseline FEV₁ and is consistent with data from the U.S. ₁-ATD Registry (2) and the association noted between bronchial hyperreactivity and FEV₁ decline in COPD without ₁-ATD (19). A possible explanation for this observation is that bronchodilator reversibility and possibly hyperreactivity are associated with increased airway inflammation that may independently lead to more rapid development of obstructive changes. Prospective studies including lung biopsies and bronchial hyperreactivity testing will be necessary, however, to resolve this issue.

Recurrent exacerbations are episodes that also increase the inflammatory burden in the airways (20). In the current study, the group as a whole demonstrated no relationship between exacerbation frequency and FEV₁ decline. However, when subjects with mild to moderate airflow obstruction at baseline (i.e., those patients demonstrating the most rapid decline in FEV₁) were examined separately, a significant correlation between exacerbation frequency and FEV₁ decline was found. In addition, the finding that deterioration in VC and DL_CO was also related to the number of exacerbations supports the hypothesis that the increased inflammatory burden associated with exacerbations in the airways is associated with greater lung damage. This observation suggests that treatment of exacerbations should be a key aim in the management of ₁-ATD. Indeed, retrospective analysis indicates that augmentation therapy is associated with a reduction in exacerbations (3) and a lesser decline in FEV₁ for patients with moderate airflow obstruction (2). It remains possible that the putative beneficial effect of augmentation therapy on FEV₁ decline therefore reflects (at least in part) amelioration of exacerbations. Clearly, this concept is worthy of further prospective study.

Overall, health status as determined by the SGRQ Total score did not change over the study period despite a decline in lung function. However, closer examination of each of the domains showed opposing effects. Symptoms improved, which may relate to beneficial effects of joining the program and would be supported by the improvement in the role emotional domain of the SF36. In particular, the Symptom score improvement was accounted for by a reduction in reported wheezing that may reflect optimization of bronchodilator therapy following the initial assessment. Nevertheless, the decline of more than four points in health status as assessed by the SGRQ Activity domain is considered to represent a clinically important deterioration (14, 15). Such a change occurred annually in the current study and was related to the change in FEV₁ and gas transfer, although not to the progression of emphysema on HRCT. However, the lower number of CT scans available for analysis may have been a factor in the failure to demonstrate an association. Overall, however, the current findings support the hypothesis that preservation of lung function will be reflected in symptomatic benefit and should be a feature of successful intervention therapy.

Upper zone inspiratory HRCT analysis was the measure most sensitive to disease progression. This finding is in agreement with the study by Dirksen and coworkers (6), although in the current study the superiority of HRCT was less clear, with the physiological measures (and DL_CO in particular) also proving relatively sensitive to change. The CT protocol employed in the current study has practical advantages over those employed in previous studies (21, 22) and is therefore a more realistic tool for sequential monitoring. In addition, in ₁-ATD where emphysema is relatively homogeneous in distribution our limited slice approach has been shown to relate to lung function, exercise capacity, and health status (7, 23) and is supported by previous work suggesting little loss of sensitivity using a single HRCT slice 5 cm below the carina (21).

Although lower zone emphysema predominates in the early stages in ₁-antitrypsin deficiency, the greater deterioration of the upper zone HRCT seen here suggests that the active disease process eventually switches to the more normal areas of the lungs as pulmonary ventilation and perfusion changes when the lower zones are destroyed. This is consistent with cross-sectional data from our center that demonstrates a curvilinear relationship between upper and lower zone voxel indices (24). These findings also have implications for CT scanning used to monitor disease progression and suggests that assessment of the upper zones should be included.

In summary, patients with airflow obstruction and ₁-antitrypsin deficiency demonstrate loss of several measures of lung function over a 2-yr period, accompanied by the development of more extensive emphysema (particularly in the upper zones). The deterioration of lung function is related independently to baseline FEV₁, bronchodilator reversibility, and the number of exacerbations that occur. These changes were accompanied by a statistically and clinically significant deterioration in health status. Although there was a significant deterioration in the FEV₁ during the study, HRCT voxel index and DL_CO were more sensitive to change and should therefore be considered as good indicators of progression and should be included as outcome measures in clinical trials in ₁-ATD.