- Email: [email protected]
Markers of disease severity in chronic obstructive pulmonary disease
Pulmonary Pharmacology & Therapeutics 19 (2006) 189–199 www.elsevier.com/locate/ypupt
Markers of disease severity in chronic obstructive pulmonary disease Luigi G. Franciosia,*, Clive P. Pageb, Bartolome R. Cellic, Mario Cazzolab,d, Michael J. Walkere, Meindert Danhofa, Klaus F. Rabef, Oscar E. Della Pasquaa,g a
Gorlaeus Laboratories, Leiden/Amsterdam Center for Drug Research, Leiden University, Einsteinweg 55, 2300 RA, Leiden, The Netherlands b Sackler Institute of Pulmonary Pharmacology, King’s College, London, UK c St Elizabeth’s Hospital, Tufts University, Boston, MA, USA d Department of Respiratory Medicine, A. Cardarelli Hospital, Naples, Italy e Department of Pharmacology and Therapeutics, University of British Columbia, Vancouver, Canada f Department of Pulmonology, Leiden University Medical Center, Leiden, The Netherlands g Clinical Pharmacology and Discovery Medicine, GlaxoSmithKline, Greenford, UK Received 1 February 2005; revised 11 May 2005; accepted 12 May 2005
Abstract Background and objectives: Diagnosis and assessment of treatment effect in chronic obstructive pulmonary disease (COPD) have relied primarily on the examination of a complex set of symptoms and the use of spirometry. However, these methods require long periods of assessment to determine whether patients show clinically relevant improvements after intervention. We therefore wanted to determine how existing clinical and laboratory measures change with COPD severity and identify disease markers that can serve as better endpoints for diagnosis and assessment of COPD progression and treatment effect. Methods: Using standard COPD keywords and terms, we searched PubMed, ISI Web of Science, and Cochrane Review databases for retrospective and prospective clinical studies published since 1966. We identified 652 studies (nZ146,255) from 1978 to September 2003 based on the availability of spirometric and demographic data, investigation of possible markers, absence of acute exacerbations and comorbidities, and the withdrawal of standard COPD medication. Central tendencies and dispersions of subject baseline measures were collected according to study sample size, smoking status, and mild, moderate and severe COPD stages. A fixed effect meta-analysis was then conducted on each measure at various disease stages. Results: Arterial oxygen tension, sputum neutrophils and IL-8, and serum TNF-a and C-Reactive Protein showed a trend toward separation between COPD stages. Other measures such as pack-years and St George’s Respiratory Questionnaire only distinguished between disease and disease-free states. Conclusions: We observed little separation between disease stages for many measures used in COPD diagnosis and clinical trials. This demonstrates the poor sensitivity of such endpoints to define a patient’s clinical status and to quantify treatment effect. Therefore, we recommend that longitudinal studies and disease modelling be the primary methods for assessing whether potential markers of disease progression can be used for COPD diagnosis and clinical trials. q 2005 Elsevier Ltd. All rights reserved. Keywords: COPD; Disease progression; Biological marker; Mathematical model
1. Introduction Chronic obstructive pulmonary disease (COPD) is a respiratory syndrome associated with a progressive, non-
* Corresponding author. Tel.: C44 20 8966 5765; fax: C44 20 8966 2123. E-mail address: [email protected] (L.G. Franciosi).
1094-5539/$ - see front matter q 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.pupt.2005.05.001
reversible limitation to airflow and abnormal inflammatory responses involving the small airways [1]. Current diagnosis of COPD involves an assessment of smoking and occupational history as well as recording of symptoms such as cough, sputum and dyspnea. However, due to poor public awareness of this disease and the overlap of symptoms with other comorbidities, many patients are not diagnosed until later when the disease has made a significant impact on their quality of life [2,3]. For many years, spirometry has been the only means of confirming and monitoring airflow obstruction. Since
190
L.G. Franciosi et al. / Pulmonary Pharmacology & Therapeutics 19 (2006) 189–199
diagnosis and the assessment of treatment effect in COPD have been based on measures that require both significant vigilance and long-term assessment, the American Thoracic Society (ATS) and the European Respiratory Society (ERS) have jointly published new guidelines to help general practitioners and pulmonary specialists better identify and manage COPD patients as well as highlight areas in medical care, where more research is needed [1]. One such area is the improvement of standards for classifying and staging COPD. Current disease staging systems are mainly based on the Forced Expired Volume in one second (FEV1), the most commonly used spirometric measure for diagnosis and evaluation of treatment effect in COPD [4–6]. Its utility has been demonstrated in a number of major studies conducted over the past 30 years; these include Fletcher and Peto’s landmark study showing a progressive decline in FEV1 over a life time [7]; and the US Lung Health Study demonstrating a return of an accelerated FEV1 decline to normal with smoking cessation [8]. More recent studies on the long-term effect of corticosteroids have shown small changes in the rate of FEV1 decline and short-term symptomatic relief [9– 13]. All these findings have eluded to the possibility that changes in FEV1 may be the best variable to reflect changes in the underlying disease mechanisms. However, the longterm monitoring of FEV1 changes and symptom-related measures requires a great deal of effort and resources, particularly, when assessing treatment benefit [2]. It has been suggested that the discovery of new markers of disease severity would not only enhance our understanding of COPD’s natural history and pathogenesis but also offer new endpoints that can be used along side standard measures in COPD diagnosis and clinical trials [1,2]. Presently, the medical literature contains numerous crosssectional studies that have tried to correlate clinical variables, counts of cell types and concentrations of inflammatory mediators to FEV1 [14–30]. Examples include body mass index (BMI) [31], neutrophil cell counts [32,33], and tumour necrosis factor (TNF-a) [23,34,35]. Despite the increasing medical burden of COPD, longitudinal studies investigating such potential markers of disease progression are often scarce or currently ongoing [2]. The elucidation of a marker’s potential to discriminate between different disease stages would be useful in establishing appropriate clinical endpoints, particularly for the assessment of treatment effect [2]. Therefore, we reviewed the medical literature using a commonly cited disease staging system to determine the statistical distributions of clinical and laboratory measures obtained from patients with stable COPD.
2. Methods 2.1. Study identification and selection Our search strategy involved the use of PubMed, ISI Web of Science, and Cochrane Review databases and keywords
obtained from US National Library of Medicine, major guidance documents and reference books on stable COPD. Additional details about our search can be found in Appendix A. Study selection was based on the availability of FEV1 and other spirometric, demographic, and possible marker data, the absence of exacerbations and life— threatening co-morbidities, and the withdrawal from COPD medication such as oral corticosteroids for an extended period of time prior to study entry and baseline measurements. Studies that primarily investigated potential markers in COPD patients but did not record FEV1 and other spirometry measures were excluded. 2.2. Objective of the literature review Our objective was to determine the characteristics of published study populations and to understand how baseline variables and potential markers changed with COPD stages. Study subjects included those with mild, moderate and severe COPD that were assessed according to the 1995 ATS COPD staging criteria [6] as well as healthy subjects that were matched for age, sex or height. If baseline information of a specific biochemical measure was unavailable for the healthy subject group of COPD studies, data was retrieved from asthma studies that had healthy subjects of similar characteristics. 2.3. Data abstraction From each study, we abstracted any spirometric, demographic, clinical, cytological, or biochemical variable that was considered standard in the stable COPD literature or had been suggested as a potential marker of COPD severity (Table 1). Other data retrieved concerned the study question, experimental design, sponsorship, and the presence or absence of individual subject data. We first sorted the data according to the four FEV1 postbronchodilator categories of the 1995 ATS disease staging system (Table 2) [6]. This system was selected in lieu of past and present disease staging systems because of its scientific evidence [36–46] and its common use in the design and eligibility requirements of past COPD studies. The boundaries of its FEV1 categories are similar to those found in the newly adopted ATS/ERS 2004 disease staging system (Fig. 1). (Each FEV1 category has been reclassified to a more severe level of disease and the previous ‘Severe’ stage is now spit into ‘Severe’ and ‘Very Severe’. These changes like those seen with past disease staging systems were made mainly for reasons related to clinical management rather than pathophysiological findings [47–49]). The data was then further organized according to sample size as well as smoking status, i.e. non-, ex- and current smoker. Cytological and biochemical data were also classified according to specimen type.
L.G. Franciosi et al. / Pulmonary Pharmacology & Therapeutics 19 (2006) 189–199
191
Table 1 Demographics, spirometry and possible markers of stable COPD Demographics Spirometry Clinical variables
Age; gender; height; weight; pack-years; disease years FEV1 (litres and % predicted); FVC (litres and % predicted); FEV1/FVC ratio Arterial oxygen tension; arterial carbon dioxide tension; b2-agonist reversibility; methacholine, histamine or AMP challenge; body mass index; baseline dyspnea index; 6 minute walking distance; St George’s respiratory questionnaire
Cytological variables Local (Sputum, BAL, Biopsy) Systemic (Plasma or Serum) Biochemical variables Local (Sputum, BAL, Biopsy) Systemic (Plasma or Serum) Exhaled breath
Neutrophils; macrophages; eosinophils; CD8C lymphocytes Neutrophils; macrophages; eosinophils; CD8C lymphocytes Interleukin-6; interleukin-8; TNF-a; fibrinogen; C-reactive protein Interleukin-6; interleukin-8; TNF-a; fibrinogen; C-reactive protein Nitric oxide; carbon monoxide
FEV1, forced expired volume in one second; FVC, forced vital capacity; BAL, broncho-alveolar lavage; TNF-a, tumour necrosis factor a.
2.4. Statistical methods The typical study measures of central tendency and dispersions (i.e. means, medians, standard errors, standard deviations, 95% confidence interval, and inter-quartile ranges) were tabulated and calculated using Microsoft Excel. A fixed effect meta-analyses was used to summarize point estimates, 95% confidence intervals, and study data distributions (i.e. 2SD) [50]. A two-tailed Z-test was then performed whenever possible to statistically assess differences between healthy subjects and a disease stage. The a-level was adjusted for multiple testing according to the Bonferroni correction procedure. In the event a specific disease stage had a large number of studies with only median data available, the normal distribution of the data was assumed and medians were treated as means. During the analyses, data showing non-normality was not logtransformed because individual data sets were not available to us.
3. Results 3.1. Description of studies We identified 652 suitable studies published between 1978 and mid September 2003. These are available at http:// www.pharmacology.leidenuniv.nl/ under the link that
concerns the COPD project summary and reference lists. The total number of subjects in these studies was 146,255, of which 46% were healthy subjects, 26% mild, 16% moderate, and 13% severe COPD patients. 3.2. Clinical variables and relationship with COPD stage The meta-analyses of typical study demographics are provided in Table 3. We found that the pack-years in mild, moderate and severe stages were similar, indicating that analysis of study subjects according to smoking status was not necessary for our analysis. We also found that twothirds of studies provided gender information; studies with healthy subjects had about an equivalent number of males and females whilst the number of males increased and the number of females decreased with advancing COPD. As expected, the point estimates and 95% confidence intervals of FEV1 % predicted (Fig. 2A) showed a clear separation between the healthy state and the different COPD stages from the studies sampled. This reflects the fact that the ATS’ COPD stages are based on determinations of FEV1 % predicted. Among the possible markers of COPD, only the point estimates of arterial oxygen tension (PaO2) demonstrated the same profile as FEV1 (Fig. 2B). Other variables showed either no difference between healthy and mild stages, as with arterial carbon dioxide tension (PaCO2), BMI and the six minute walking distance (6MWD) (Fig. 2C), or showed a binary pattern in that
Table 2 Disease staging systems-past and present Staging system ERS (1995) [4] ATS (1995) [6] BTS (1997) [5] GOLD (2001) [49] GOLD (2003) [48] ATS/ERS (2004) [47]
Disease severity listed according to FEV1 (% predicted) criterion At risk
Mild (%)
Moderate (%)
Severe (%)
Very severe (%)
N.Ap. N.Ap. N.Ap. Normal Normal R80
R70 R50 60–79 R80 R80 R80
50–69 35–49 40–59 30–79 50–79 50–79
!50 !35 !40 !30 or !50CCRF 30–49 30–49
N.Ap. N.Ap. N.Ap. N.Ap. !30 or !50CCRF !30
ERS, European respiratory society; ATS, American thoracic society; BTS, British thoracic society; GOLD, global initiative for chronic obstructive lung disease; CRF, chronic respiratory failure; and N.Ap., not applicable.
192
L.G. Franciosi et al. / Pulmonary Pharmacology & Therapeutics 19 (2006) 189–199
3.3. Cytological variables
Fig. 1. Comparison of ATS 1995 and ATS/ERS 2004 disease staging systems.
statistically significant differences were only seen between the presence of disease irrespective of stage and the healthy state. This was the case for baseline dyspnea index (P! 0.001), beta2 agonist reversibility (P!0.001), and the St George’s Respiratory Questionnaire (SGRQ) (Fig. 2D, P! 0.001). A shortage of studies or appropriate data across disease stages for pharmacological challenge tests (i.e. methacholine, histamine and AMP) did not allow us to evaluate the statistical distributions of these variables.
The two main cell types sampled in sputum—neutrophils and macrophages—showed a significant difference between the healthy state and each of the stages of COPD, while the eosinophils did not. Significant overlap was seen in distributions (2SD) for mild, moderate and severe stages for each cell type. However, the point estimates obtained from the sputum neutrophil studies demonstrated an upward trend from the healthy to severe stages of COPD (Fig. 3A). A meta-analysis of cell types in bronchoalveolar lavage (BAL) and lung biopsies was not possible due to the limited amount of studies and data available (Table 4). 3.4. Biochemical variables Research on biochemical markers in plasma or serum consisted of IL-6, IL-8, TNF-a and CRP. CRP serum concentrations showed no statistical differences between the healthy and each of the COPD stages, but we observed an upward trend with increasing disease severity
Table 3 Typical subject demographics Variable/COPD stagea
Total studiesb
Total subjectsc
Point estimate (95% CI)
Study data distributions (2SD)
Significance testd Z value
P value
Age (years) Healthy Mild Moderate Severe Height (cm) Healthy Mild Moderate Severe Weight (kg) Healthy Mild Moderate Severe Pack-yearse Healthy Mild Moderate Severe Disease years Mild Moderate Severe
628 194 218 229 207 164 39 36 71 67 159 37 32 66 73 204 67 107 67 30 27 11 12 4
143,756 66,312 36,411 22,846 18,187 18,439 6761 3873 6536 1269 12,921 3586 3716 4192 1427 63,785 27,223 23,943 11,861 758 8839 1467 7179 193
– 46.3 (45.6–47.1) 61.2 (60.4–62.1) 65.0 (64.3–65.8) 64.0 (63.3–64.7) – 170.3 (168.6–172.1) 169.3 (167.7–170.8) 167.4 (166.3–168.5) 167.5 (166.3–168.8) – 74.7 (71.7–77.7) 71.2 (67.8–74.5) 68.0 (65.7–70.4) 62.3 (60.6–64.0) – 17.3 (15.8–18.8) 40.5 (38.2–42.9) 39.0 (35.7–42.3) 39.6 (35.1–44.2)
– 36.5–56.2 47.6–74.8 52.0–78.1 53.7–74.3 – 158.2–172.1 159.2–170.8 157.3–168.5 157.5–168.8 – 53.4–96.0 50.5–91.8 46.7–89.4 48.5–76.1 – 7.2–27.5 17.4–63.6 11.0–67.0 17.6–61.7
5.0 (3.6–6.4) 8.2 (5.5–11.0) 10.0 (7.0–12.9)
– – –
– – 26.0 35.0 33.9 – – 0.87 2.77 2.54 – – 1.53 3.39 7.01 – – 16.2 11.7 9.11 – – – –
– – !0.001 !0.001 !0.001 – – 0.38 (NS) 0.0056 0.011 – – 0.13 (NS) !0.001 !0.001 – – !0.001 !0.001 !0.001 – – – –
a
ATS COPD Stages: mild COPD is FEV1R50% predicted; Moderate COPD is FEV1 between 35 and 49% predicted; Severe COPD is FEV1!35% predicted. b Bold numbers indicate total studies (without duplicates) for the specific variable of interest. c Bold numbers indicate total subjects for all disease stages with respect to the specific variable of interest. d Mild, moderate and severe COPD stages were each compared to the healthy stage; a-level adjusted according to the Bonferroni Correction to account for multiple testing; NS, non-significant difference. e Most studies determined pack years from the number of packs per day multiplied by number of years smoked; one pack contains 20 cigarettes.
L.G. Franciosi et al. / Pulmonary Pharmacology & Therapeutics 19 (2006) 189–199
193
Fig. 2. Fixed effect meta-analysis results of selected clinical variables. For each clinical variable, the point estimates (point), 95% confidence intervals (box), and two standard deviations (bars) are presented for mild (FEV1R50% predicted), moderate (FEV1 35–49%), and severe COPD (FEV1!35%). ((A) the two standard deviations of FEV1 % predicted have been omitted for theoretical reasons). N, signifies the total studies, n, the total subjects for each stage and asterisks P!0.001.
(Fig. 3B). For serum TNF-a, there were no statistical differences between the healthy subjects and any of the COPD stages despite a trend for increases with increasing severity of the stages (Fig. 3C). In the case of IL6, IL-8 and fibrinogen, there was insufficient data to perform any meta-analysis (Table 4). There was far less studies in which possible biochemical markers of COPD were measured in healthy subjects, or in patients at various stages of the disease, that were suitable for analysis. We only found enough studies investigating IL-8 in which
sputum values for healthy subjects were significantly different from mild and moderate COPD patients (Fig. 3D). There was insufficient data for the metaanalysis of sputum TNF-a, IL-6, fibrinogen, and CRP. The number of studies that assessed the biochemical markers IL-6, IL-8, TNF-a, fibrinogen and CRP in BAL was also scarce. From the data available, only BAL IL-8 was found to show a statistical difference between the healthy and mild stages (P!0.001). From the evaluation of nitric oxide and carbon monoxide levels in exhaled
194
L.G. Franciosi et al. / Pulmonary Pharmacology & Therapeutics 19 (2006) 189–199
Fig. 3. Fixed effect meta-analysis results of selected biochemical variables. For each biochemical variable, the point estimates (point), 95% confidence intervals (box), and two standard deviations (bars) have been presented for mild (FEV1R50% predicted), moderate (FEV1 35–49%), and severe COPD (FEV1!35%). N, indicates the total studies; n, the total subjects for each stage, and asterisks P!0.001.
breath, the meta-analyses revealed no obvious relationship with COPD severity (Table 4). 4. Discussion The diagnosis and determination of treatment benefit for any disease should be based on sensitive clinical and laboratory measures that reflect differences in disease severity. In particular, each measure should show a clear separation of their statistical distributions across disease
stages. This has been previously illustrated in HIV infection, where CD4 counts and mRNA viral load are correlated with clinical status, antiviral treatment effect, and risk of AIDS [51]. Despite the ongoing research on biomarkers of COPD, little effort has been made in assessing whether such correlations exist. Moreover, the selection of endpoints for the assessment of clinical status and treatment effect have not been based on any formal validation. This is striking since the identification and validation of endpoints that reflect changes in disease status are requirements for both
L.G. Franciosi et al. / Pulmonary Pharmacology & Therapeutics 19 (2006) 189–199
195
Table 4 Variables demonstrating little relationship with COPD severity or have insufficient data for meta-analysis Variable
Total studies
Total subjects
PaCO2 (mmHg) BMI (kg/m2) b2 Agonist reversibility (% FEV1 predicted) Sputum eosinophils (% differential count) 6MWD (meters) Sputum macrophages (% differential count) SGRQ (Total/100) BAL neutrophils (!103/mL) Sputum IL-8 (nanograms/mL) BAL macrophages (!103/mL) Baseline dyspnea index (Score/12) Exhaled nitric oxide (parts per billion) BAL Eosinophils (!103/mL) Biopsy neutrophils (cells/mm2) Biopsy CD8C lymphocytes (cells/mm2) Biopsy eosinophils (cells/mm2) Histamine challenge (mg/mL) Biopsy macrophages (cells/mm2) BAL IL-8 (pg/mL) Methacholine challenge (mg/mL) Exhaled carbon monoxide (parts per million) AMP Challenge (mg/mL) Sputum TNF-a (ng/mL) Blood IL-6 (pg/mL) Blood fibrinogen (g/L) Blood IL-8 (pg/mL) Sputum fibrinogen (mg/L) Sputum IL-6 (pg/mL) BAL TNF-a (pg/mL) Blood eosinophils (!109/mL) BAL IL-6 (pg/mL) Blood CD8C cells (% lymphocytes) Biopsy IL-8 (IL-8C mRNA/mm2) Biopsy TNF-a (TNF-aC mRNA/mm2) Blood neutrophils (!109/mL) Sputum CD8C lymphocytes (% lymphocytes) BAL CD8C Lymphocytes Blood macrophages Sputum C-reactive protein BAL Fibrinogen BAL C-Reactive protein Biopsy IL-6 Biopsy fibrinogen Biopsy C-reactive protein
230 111 78 55 54 53 33 28 26 25 23 22 17 17 14 14 13 13 11 8 7 7 6 5 4 4 3 3 3 2 2 1 1 1 1 1 N.Ap. N.Ap. N.Ap. N.Av. N.Av. N.Av. N.Av. N.Av.
20,977 54,463 23,621 1621 3181 1550 8328 925 864 829 5750 915 522 415 376 306 695 255 397 144 591 228 172 195 6758 106 68 188 103 83 67 66 57 57 51 13 N.Ap. N.Ap. N.Ap. N.Av. N.Av. N.Av. N.Av. N.Av.
medical practice and drug development [52]. Hence, we conducted this study to determine such changes in variables and suggested markers of COPD with respect to clinical status. Given the limited amount of published longitudinal studies and availability of individual time courses, the information derived from the cross-sectional studies that we sampled primarily reflects overall inter-individual variability. Nevertheless, pooling of this baseline data provides insight into which variables may be potential markers of disease severity according to FEV1 status. We determined that PaO2 was the only clinical variable unrelated to spirometry that correlated with the FEV1 staging criteria (Fig. 2B). Other clinical variables such as
BMI, 6MWD (Fig. 2C) and SGRQ (Fig. 2D) displayed overlap in effect sizes, confidence intervals and data distributions. Normally, these variables were only able to distinguish between healthy subjects and patients. With respect to the cell types, only sputum neutrophils exhibited a trend toward separation between COPD stages (Fig. 3A). We were not able to assess the statistical distributions of many cells types in other body specimens due to a shortage of data in the moderate and severe stages. When we reviewed the biochemical variables, we found a separation in point estimates in serum TNF-a and CRP as well as in sputum IL-8 (Fig. 3). However, we were faced with a shortage of data in the later stages of the disease.
196
L.G. Franciosi et al. / Pulmonary Pharmacology & Therapeutics 19 (2006) 189–199
Another aspect we had to consider was the specificity of these cytological and biochemical variables. The few inflammatory cells and mediators that were affected across COPD stages may not bare a causal relationship with the severity of the disease itself. Rather, they may only reflect local and systemic inflammatory conditions and may not be specific to COPD. In addition, the dynamics of inflammatory cell migration was not considered in the sampled matrices (i.e. blood, tissue biopsy, BAL, and sputum). The correspondence between local and systemic measurements of migrating inflammatory cells such as lymphocytes was often unclear. These measurements were usually obtained at different times from different individuals rather than concurrently within each individual. Hence, a parallel sampling scheme for local and systemic matrices would be required for characterising the inflammatory component of this disease. An interesting finding is that pack-years did not change with increasing disease severity. This observation is in agreement with previous studies that 10–15% of smokers get COPD [53,54]. A recent study has implicated differences in activation of bronchial epithelial cells between smokers with COPD and smokers without airway obstruction in releasing neutrophils into the airways [55]. These preliminary findings have a number of limitations. We considered the characteristics of the COPD population mainly from cross-sectional studies. The studies included in this analysis were selected based on a set of criteria (outlined in the Study Identification and Selection section) that should minimize heterogeneity across the populations of interest and eliminate potential bias during comparisons. We realize, however, that protocol violations may have occurred that were not described in the original publications that we reviewed. During the data abstraction process, we noticed that many studies did not explicitly state whether FEV1 was a pre- or post-bronchodilator value. We assumed that investigators were following international guidelines and that they used a post-bronchodilator value for study eligibility. Another issue related to FEV1 is the consistency of its measurement [17,56]. In addition, environmental and biological factors may have contributed to the variation in FEV1 data [36]. With respect to study subjects, the possibility exists that a given subject may have been enrolled in more than one COPD study. This implies that the subject numbers that support this review may be smaller than indicated. Another consideration is that although subjects may be off their medication for study eligibility, it is reasonable to assume that patients might have been on some medication (e.g. long-term oxygen therapy, bronchodilators and oral corticosteroids) that could affect FEV1 status at the time of baseline measurements. We intend to further explore the contribution of confounding factors and covariates such as therapy and age to explain the profiles of FEV1 and other potential markers. As well, the value of multidimensional scales (e.g. the BODE index that combines body-mass index, airflow obstruction, dyspnea,
and exercise capacity [57]) will also be examined for disease classification. However, individual data sets are required for these purposes. Our group is currently performing this work. The use of a disease staging system for the basis for an exploratory analysis of clinical, cytological and biochemical variables implies that its categories have been rigorously evaluated using criterion-based validity methods [58]. However, we realized that past and present systems have FEV1 categories which were mainly inferred from clinical management rather than scientific studies [36,48,49]. In that sense, there is growing awareness that a mechanism-based approach should be used to define and classify disease markers and staging systems [52,59,60]. Disease markers can characterise a process on the causal pathway between disease and treatment response [52]. A mechanism-based classification has been proposed which divides biological markers into six categories from genotype to clinical scales [59,60]. Thus far, it appears that only physiological measures and clinical scales have been included (Types 4 and 6 biomarkers, respectively) in COPD trials. There has been no systematic evaluation of genotype/phenotype (Type 0), target expression and activation (Type 3), or disease process (Type 5) other than the rate and frequency of COPD exacerbations. The process of selecting and validating a marker involves evaluating the effect of disease on the marker as well as its sensitivity to treatment effect. This can be achieved by mathematical modelling of the time course of disease and its variability. A rank of sensitive markers can then be used to study treatment response in patient populations or in individual patients. In addition, such markers should be specific in differentiating disease from the healthy state. These requirements are essential to minimise false positives and false negatives in diagnosis and treatment. In chronic diseases, clinical use of a biomarker requires two additional aspects to be taken into account, namely the rate and magnitude of change that ultimately contribute to the estimation of group size and statistical power of a study. If endpoints do not change over time, how do we define treatment efficacy? We need markers that are sensitive to disease severity. Despite the separation of point estimates for the potential markers that we identified in this investigation, their statistical distributions often demonstrate considerable overlap when their data have been ranked accordingly. For example, when the data retrieved from the cross-sectional studies of sputum neutrophils were ranked from lowest to highest available percentages, the ATS’ FEV1-based disease stages begin to overlap with increasing disease severity (Fig. 4). This highlights the need to look beyond FEV1 as the major endpoint for classifying COPD severity and to explore these elements in longitudinal studies. An established approach to assessing disease severity and progression as well as predicting treatment effect is mechanism-based mathematical modelling. Such models
L.G. Franciosi et al. / Pulmonary Pharmacology & Therapeutics 19 (2006) 189–199
197
Fig. 4. Ranked sputum neutrophil data demonstrating overlap of the ATS’ FEV1-based disease stages. Presented are mean or median percentage of sputum neutrophils in a percent differential count that have been ranked from the lowest to highest value obtained from the sampled literature. The blue bars represent studies with control healthy subjects; the green bars are studies with mild COPD patients (FEV1R50% predicted); the yellow bars are studies with moderate COPD patients (FEV1 35–49%); and the red bars are studies with severe COPD patients (FEV1!35%). N, indicates the total sum of subjects obtained from all available studies sampled.
can be useful in a number of ways, particularly to support the selection of dosing regimen as well as study design and treatment duration [51,61]. A major advantage of modelling is the ability to estimate the placebo effect and separate it from the pharmacological response, which eliminates the enormous bias of the ‘last observation carried forward (LOCF)’ procedure common in longitudinal studies. One example of this approach is pharmacokinetic-pharmacodynamic modelling in which mechanism-based parameters are used as fixed effects to explain the time course of disease or treatment effect. Moreover, it incorporates random effects that describe intra-individual and inter-individual differences and their change over time. Similarly, changes in disease severity can be encompassed in one or more states using Markov models. With this approach, one can predict the number of individuals in each disease state at each specified observation time and their rate of transition to subsequent states. Given the characteristics of COPD, we feel that the current research process is not robust enough to elucidate possible biomarkers or endpoints for treatment response. Pharmacostatistical modelling is a powerful tool to characterise the time course of disease and treatment effect in both populations and individual patients as well as estimates of uncertainty. This is paramount to our current efforts to identify biomarkers in COPD. In summary, the statistical distributions obtained from our cross-sectional studies revealed that many clinical, cytological and biochemical variables in COPD showed little or no relationship with FEV1-based disease staging criteria. From the available studies, where more than spirometry was measured, we did identify five possible markers, i.e. PaO2, sputum neutrophils and IL-8, and systemic TNF-a and CRP, that showed a separation
between disease stages. In addition, this review provides insight about the areas, where more data should be acquired to conclude about the relevance of the measures investigated thus far. Longitudinal studies are required to determine how these possible markers reflect the time course of disease within a single individual, but the implementation of such studies may have practical, ethical and financial consequences. Therefore, we are undertaking an integrated mechanism-based modelling approach that offers a better solution for improving the diagnosis as well as the assessment of treatment effect in patients with COPD.
Acknowledgements We like to thank Sandra van den Berg-van Tol, secretary at the LACDR Division of Pharmacology for her assistance in the preparation of the website containing the study references for this work. We also like to thank senior statisticians Nigel Dallow and Andrew M. Wright from GlaxoSmithKline UK for their statistical advice and guidance during the meta-analyses. The current research is part of L. Franciosi’s post-doctoral fellowship on COPD disease progression funded by GlaxoSmithKline UK.
Appendix A. Search strategy We performed a search of the PubMed, Medline, ISI Web of Science, and Cochrane Review databases up to mid September 2003. All language types were included. The principal search for the disease’s name were: chronic obstructive pulmonary disease; COPD; chronic obstructive
198
L.G. Franciosi et al. / Pulmonary Pharmacology & Therapeutics 19 (2006) 189–199
airway disease; COAD; chronic airflow obstruction; CAO; chronic obstructive lung disease; COLD; chronic bronchitis; CB; bronchitis; blue bloaters; pulmonary emphysema; pink puffers; and small airways disease. These terms were then combined with terms designating a classification or a subpopulation such as control; healthy; mild; moderate; severe; very severe; ATS (American Thoracic Society); BTS (British Thoracic Society); ERS (European Respiratory Society); GOLD; disease, stages, states; clinical; human; current smoker, never smoker, non-smoker, ex-smoker; and stable. Major COPD symptoms like chronic cough, sputum, dyspnea (or dyspnoea) as well as drug classes such as beta agonists, corticosteroids, anticholinergics, and phosphodiesterase four inhibitors were also included. The resulting citation list was then refined to include retrospective and prospective clinical studies, but not abstracts, reviews, meta-analyses, case reports, letters, editorials or duplicate publications. Then, common variables and suggested markers of clinical, cytological or biochemical origin obtained from a preliminary assessment of large studies, review articles, and guidance documents were searched within this list. These variables were further searched in combination with their respective measurement technique or specimen type. Keywords included spirometry, biopsy, bronchial; bronchoalveolar lavage; BAL; sputum, induction; exhaled breath; and systemic, plasma, serum. The resulting list contained 934 citations of interest. From the title and abstract of each these citations, we assessed the availability of variable information and spirometry data for subsequent disease classification. Studies that investigated COPD patients with respect to pharmacoeconomics, respiratory failure, lung transplant surgery, or other advanced life–threatening co-morbidities such as terminal cancer were excluded. Original articles were then retrieved in electronic or hard copy formats and reviewed to determine if each clinical study had a clear description of the subject eligibility criteria, smoking status, sample sizes and the presence of spirometry and anthropometric data at baseline or study entry. The final database contained 652 studies that were eligible for descriptive analysis.
[5]
[6]
[7] [8]
[9]
[10]
[11]
[12]
[13]
[14] [15]
[16] [17]
[18] [19]
[20]
References [1] American Thoracic Society/European Respiratory Society. Definition, Diagnosis and Staging of COPD. American Thoracic Society Website 2004 May 24 [cited 2004 May 24]; Available from: URL: http://www.thoracic.org/COPD/1/definitions.asp. [2] Croxton TL, Weinmann GG, Senior RM, Wise RA, Crapo JD, Buist AS. Clinical research in chronic obstructive pulmonary disease: needs and opportunities. Am J Respir Crit Care Med 2003;167(8): 1142–9. [3] Calverley PM, Walker P. Chronic obstructive pulmonary disease. Lancet 2003;362(9389):1053–61. [4] Siafakas NM, Vermeire P, Pride NB, Paoletti P, Gibson J, Howard P, et al. Optimal assessment and management of chronic obstructive
[21]
[22]
[23]
pulmonary disease (COPD). The European respiratory society task force. Eur Respir J 1995;8(8):1398–420. The COPD Guidelines Group of the Standards of Care Committee of the BTS. BTS guidelines for the management of chronic obstructive pulmonary disease. Thorax 1997;52(5):S1–28 American Thoracic Society. Standards for the diagnosis and care of patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1995;152(5 Pt 2):S77–121 Fletcher C, Peto R. The natural history of chronic airflow obstruction. Br Med J 1977;1(6077):1645–8. Anthonisen NR, Connett JE, Murray RP. Smoking and lung function of Lung Health Study participants after 11 years. Am J Respir Crit Care Med 2002;166(5):675–9. Burge PS, Calverley PM, Jones PW, Spencer S, Anderson JA, Maslen TK. Randomised, double blind, placebo controlled study of fluticasone propionate in patients with moderate to severe chronic obstructive pulmonary disease: the ISOLDE trial. BMJ 2000; 320(7245):1297–303. Pauwels RA, Lofdahl CG, Laitinen LA, Schouten JP, Postma DS, Pride NB. Long-term treatment with inhaled budesonide in persons with mild chronic obstructive pulmonary disease who continue smoking. European respiratory society study on chronic obstructive pulmonary disease. N Engl J Med 1999;340(25):1948–53. Vestbo J, Sorensen T, Lange P, Brix A, Torre P, Viskum K. Longterm effect of inhaled budesonide in mild and moderate chronic obstructive pulmonary disease: a randomised controlled trial. Lancet 1999;353(9167):1819–23. Lung Health Study Research Group. Effect of inhaled triamcinolone on the decline in pulmonary function in chronic obstructive pulmonary disease. N Engl J Med 2000;343(26):1902–9. Weir DC, Bale GA, Bright P, Sherwood BP. A double-blind placebocontrolled study of the effect of inhaled beclomethasone dipropionate for 2 years in patients with nonasthmatic chronic obstructive pulmonary disease. Clin Exp Allergy 1999;29(2):125–8. Barnes PJ. Mechanisms in COPD: differences from asthma. Chest 2000;117(2):10S–114. Kessler R, Faller M, Fourgaut G, Mennecier B, Weitzenblum E. Predictive factors of hospitalization for acute exacerbation in a series of 64 patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1999;159(1):158–64. Weiss ST, DeMeo DL, Postma DS. COPD: problems in diagnosis and measurement. Eur Respir J Suppl 2003;41:4s–112. Crapo RO, Jensen RL, Hargreave FE. Airway inflammation in COPD: physiological outcome measures and induced sputum. Eur Respir J Suppl 2003;41:19s–228. Wouters EF, Creutzberg EC, Schols AM. Systemic effects in COPD. Chest 2002;121(5):127S–130. Agusti AG, Noguera A, Sauleda J, Sala E, Pons J, Busquets X. Systemic effects of chronic obstructive pulmonary disease. Eur Respir J 2003;21(2):347–60. Tashkin DP, Altose MD, Connett JE, Kanner RE, Lee WW, Wise RA. Methacholine reactivity predicts changes in lung function over time in smokers with early chronic obstructive pulmonary disease. The Lung Health Study research group. Am J Respir Crit Care Med 1996;153(6 Pt 1):1802–11. Scanlon PD, Connett JE, Waller LA, Altose MD, Bailey WC, Buist AS. Smoking cessation and lung function in mild-to-moderate chronic obstructive pulmonary disease. The Lung Health Study. Am J Respir Crit Care Med 2000;161(2 Pt 1):381–90. Rutgers SR, Timens W, Kauffman HF, Postma DS. Markers of active airway inflammation and remodelling in chronic obstructive pulmonary disease. Clin Exp Allergy 2001;31(2):193–205. Yasuda N, Gotoh K, Minatoguchi S, Asano K, Nishigaki K, Nomura M, et al. An increase of soluble Fas, an inhibitor of apoptosis, associated with progression of COPD. Respir Med 1998; 92(8):993–9.
L.G. Franciosi et al. / Pulmonary Pharmacology & Therapeutics 19 (2006) 189–199 [24] Barnes PJ. Chronic obstructive pulmonary disease. N Engl J Med 2000;343(4):269–80. [25] Walter RE, Beiser A, Givelber RJ, O’Connor GT, Gottlieb DJ. Association between glycemic state and lung function: the Framingham heart study. Am J Respir Crit Care Med 2003;167(6):911–6. [26] Dahl M, Tybjaerg-Hansen A, Vestbo J, Lange P, Nordestgaard BG. Elevated plasma fibrinogen associated with reduced pulmonary function and increased risk of chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2001;164(6):1008–11. [27] Dev D, Wallace E, Sankaran R, Cunniffe J, Govan JR, Wathen CG, et al. Value of C-reactive protein measurements in exacerbations of chronic obstructive pulmonary disease. Respir Med 1998;92(4): 664–7. [28] Soler N, Ewig S, Torres A, Filella X, Gonzalez J, Zaubet A. Airway inflammation and bronchial microbial patterns in patients with stable chronic obstructive pulmonary disease. Eur Respir J 1999;14(5): 1015–22. [29] Zhu J, Majumdar S, Qiu Y, Ansari T, Oliva A, Kips JC, et al. Interleukin-4 and interleukin-5 gene expression and inflammation in the mucus-secreting glands and subepithelial tissue of smokers with chronic bronchitis. Lack of relationship with CD8(C) cells. Am J Respir Crit Care Med 2001;164(12):2220–8. [30] Bhowmik A, Seemungal TA, Sapsford RJ, Wedzicha JA. Relation of sputum inflammatory markers to symptoms and lung function changes in COPD exacerbations. Thorax 2000;55(2):114–20. [31] Landbo C, Prescott E, Lange P, Vestbo J, Almdal TP. Prognostic value of nutritional status in chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1999;160(6):1856–61. [32] Stockley RA. Neutrophils and the pathogenesis of COPD. Chest 2002; 121(5):151S–15. [33] Beeh KM, Beier J, Kornmann O, Mander A, Buhl R. Long-term repeatability of induced sputum cells and inflammatory markers in stable, moderately severe COPD. Chest 2003;123(3):778–83. [34] Keatings VM, Collins PD, Scott DM, Barnes PJ. Differences in interleukin-8 and tumor necrosis factor-alpha in induced sputum from patients with chronic obstructive pulmonary disease or asthma. Am J Respir Crit Care Med 1996;153(2):530–4. [35] Sakao S, Tatsumi K, Igari H, Watanabe R, Shino Y, Shirasawa H, et al. Association of tumor necrosis factor-alpha gene promoter polymorphism with low attenuation areas on high-resolution CT in patients with COPD. Chest 2002;122(2):416–20. [36] American Thoracic Society. Lung function testing: selection of reference values and interpretative strategies. Am Rev Respir Dis 1991;144(5):1202–18. [37] Burrows B, Lebowitz MD, Camilli AE, Knudson RJ. Longitudinal changes in forced expiratory volume in one second in adults. Methodologic considerations and findings in healthy nonsmokers. Am Rev Respir Dis 1986;133(6):974–80. [38] Knudson RJ, Lebowitz MD, Holberg CJ, Burrows B. Changes in the normal maximal expiratory flow-volume curve with growth and aging. Am Rev Respir Dis 1983;127(6):725–34. [39] Crapo RO, Morris AH, Gardner RM. Reference spirometric values using techniques and equipment that meet ATS recommendations. Am Rev Respir Dis 1981;123(6):659–64. [40] Paoletti P, Pistelli G, Fazzi P, Viegi G, Di PF, Giuliano G, et al. Reference values for vital capacity and flow-volume curves from a general population study. Bull Eur Physiopathol Respir 1986;22(5): 451–9. [41] Lebowitz MD, Quackenboss J, Camilli AE, Bronnimann D, Holberg CJ, Boyer B. The epidemiological importance of intraindividual changes in objective pulmonary responses. Eur J Epidemiol 1987;3(4):390–8.
199
[42] Cochrane GM, Prieto F, Clark TJ. Intrasubject variability of maximal expiratory flow volume curve. Thorax 1977;32(2):171–6. [43] Miller MR, Pincock AC. Predicted values: how should we use them? Thorax 1988;43(4):265–7. [44] Miller AJ. Pulmonary function tests in clinical and occupational lung disease. Orlando: Grune and Stratton; 1986. [45] Oldham PD. Percent of predicted as the limit of normal in pulmonary function testing: a statistically valid approach. Thorax 1979;34(4): 569–70. [46] Sobol BJ. Assessment of ventilatory abnormality in the asymptomatic subject: an exercise in futility. Thorax 1966;21(5):445–9. [47] Celli BR, MacNee W. Standards for the diagnosis and treatment of patients with COPD: a summary of the ATS/ERS position paper. Eur Respir J 2004;23(6):932–46. [48] Fabbri LM, Hurd SS. Global strategy for the diagnosis, management and prevention of COPD: update. Eur Respir J 2003;22(1):1–2. [49] Pauwels RA, Buist AS, Calverley PM, Jenkins CR, Hurd SS. Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease. NHLBI/WHO Global initiative for chronic obstructive lung disease (GOLD) workshop summary. Am J Respir Crit Care Med 2001;163(5):1256–76. [50] Hedges LV, Vevea JL. Fixed- and random-effects models in metaanalysis. Psychol Methods 1998;3(4):486–504. [51] De Gruttola VG, Clax P, DeMets DL, Downing GJ, Ellenberg SS, Friedman L, et al. Considerations in the evaluation of surrogate endpoints in clinical trials. summary of a national institutes of health workshop. Control Clin Trials 2001;22(5):485–502. [52] Frank R, Hargreaves R. Clinical biomarkers in drug discovery and development. Nat Rev Drug Discov 2003;2(7):566–80. [53] Saetta M, Turato G, Maestrelli P, Mapp CE, Fabbri LM. Cellular and structural bases of chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2001;163(6):1304–9. [54] Sherrill DL, Lebowitz MD, Burrows B. Epidemiology of chronic obstructive pulmonary disease. Clin Chest Med 1990;11(3): 375–87. [55] Schulz C, Kratzel K, Wolf K, Schroll S, Kohler M, Pfeifer M. Activation of bronchial epithelial cells in smokers without airway obstruction and patients with COPD. Chest 2004;125(5): 1706–13. [56] Eaton T, Withy S, Garrett JE, Mercer J, Whitlock RM, Rea HH. Spirometry in primary care practice: the importance of quality assurance and the impact of spirometry workshops. Chest 1999; 116(2):416–23. [57] Celli BR, Cote CG, Marin JM, Casanova C, Montes de OM, Mendez RA, et al. The body-mass index, airflow obstruction, dyspnea, and exercise capacity index in chronic obstructive pulmonary disease. N Engl J Med 2004;350(10):1005–12. [58] Colburn WA. Optimizing the use of biomarkers, surrogate endpoints, and clinical endpoints for more efficient drug development. J Clin Pharmacol 2000;40(12 Pt 2):1419–27. [59] Aarons L, Karlsson MO, Mentre F, Rombout F, Steimer JL, van Peer A. Role of modelling and simulation in Phase I drug development. Eur J Pharm Sci 2001;13(2):115–22. [60] Danhof M. Markers of Pharmacological and Toxicological Action. Haemostas Foundation Leiden Workshop 2001 April 19 [cited 2004 May 24]; Available from: URL: http://www.haemostas.nl/files/ abstract23danhof.doc. [61] Borg S, Ericsson A, Wedzicha J, Gulsvik A, Lundback B, Donaldson GC, et al. A computer simulation model of the natural history and economic impact of chronic obstructive pulmonary disease. Value Health 2004;7(2):153–67.
Markers of disease severity in chronic obstructive pulmonary disease Luigi G. Franciosia,*, Clive P. Pageb, Bartolome R. Cellic, Mario Cazzolab,d, Michael J. Walkere, Meindert Danhofa, Klaus F. Rabef, Oscar E. Della Pasquaa,g a
Gorlaeus Laboratories, Leiden/Amsterdam Center for Drug Research, Leiden University, Einsteinweg 55, 2300 RA, Leiden, The Netherlands b Sackler Institute of Pulmonary Pharmacology, King’s College, London, UK c St Elizabeth’s Hospital, Tufts University, Boston, MA, USA d Department of Respiratory Medicine, A. Cardarelli Hospital, Naples, Italy e Department of Pharmacology and Therapeutics, University of British Columbia, Vancouver, Canada f Department of Pulmonology, Leiden University Medical Center, Leiden, The Netherlands g Clinical Pharmacology and Discovery Medicine, GlaxoSmithKline, Greenford, UK Received 1 February 2005; revised 11 May 2005; accepted 12 May 2005
Abstract Background and objectives: Diagnosis and assessment of treatment effect in chronic obstructive pulmonary disease (COPD) have relied primarily on the examination of a complex set of symptoms and the use of spirometry. However, these methods require long periods of assessment to determine whether patients show clinically relevant improvements after intervention. We therefore wanted to determine how existing clinical and laboratory measures change with COPD severity and identify disease markers that can serve as better endpoints for diagnosis and assessment of COPD progression and treatment effect. Methods: Using standard COPD keywords and terms, we searched PubMed, ISI Web of Science, and Cochrane Review databases for retrospective and prospective clinical studies published since 1966. We identified 652 studies (nZ146,255) from 1978 to September 2003 based on the availability of spirometric and demographic data, investigation of possible markers, absence of acute exacerbations and comorbidities, and the withdrawal of standard COPD medication. Central tendencies and dispersions of subject baseline measures were collected according to study sample size, smoking status, and mild, moderate and severe COPD stages. A fixed effect meta-analysis was then conducted on each measure at various disease stages. Results: Arterial oxygen tension, sputum neutrophils and IL-8, and serum TNF-a and C-Reactive Protein showed a trend toward separation between COPD stages. Other measures such as pack-years and St George’s Respiratory Questionnaire only distinguished between disease and disease-free states. Conclusions: We observed little separation between disease stages for many measures used in COPD diagnosis and clinical trials. This demonstrates the poor sensitivity of such endpoints to define a patient’s clinical status and to quantify treatment effect. Therefore, we recommend that longitudinal studies and disease modelling be the primary methods for assessing whether potential markers of disease progression can be used for COPD diagnosis and clinical trials. q 2005 Elsevier Ltd. All rights reserved. Keywords: COPD; Disease progression; Biological marker; Mathematical model
1. Introduction Chronic obstructive pulmonary disease (COPD) is a respiratory syndrome associated with a progressive, non-
* Corresponding author. Tel.: C44 20 8966 5765; fax: C44 20 8966 2123. E-mail address: [email protected] (L.G. Franciosi).
1094-5539/$ - see front matter q 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.pupt.2005.05.001
reversible limitation to airflow and abnormal inflammatory responses involving the small airways [1]. Current diagnosis of COPD involves an assessment of smoking and occupational history as well as recording of symptoms such as cough, sputum and dyspnea. However, due to poor public awareness of this disease and the overlap of symptoms with other comorbidities, many patients are not diagnosed until later when the disease has made a significant impact on their quality of life [2,3]. For many years, spirometry has been the only means of confirming and monitoring airflow obstruction. Since
190
L.G. Franciosi et al. / Pulmonary Pharmacology & Therapeutics 19 (2006) 189–199
diagnosis and the assessment of treatment effect in COPD have been based on measures that require both significant vigilance and long-term assessment, the American Thoracic Society (ATS) and the European Respiratory Society (ERS) have jointly published new guidelines to help general practitioners and pulmonary specialists better identify and manage COPD patients as well as highlight areas in medical care, where more research is needed [1]. One such area is the improvement of standards for classifying and staging COPD. Current disease staging systems are mainly based on the Forced Expired Volume in one second (FEV1), the most commonly used spirometric measure for diagnosis and evaluation of treatment effect in COPD [4–6]. Its utility has been demonstrated in a number of major studies conducted over the past 30 years; these include Fletcher and Peto’s landmark study showing a progressive decline in FEV1 over a life time [7]; and the US Lung Health Study demonstrating a return of an accelerated FEV1 decline to normal with smoking cessation [8]. More recent studies on the long-term effect of corticosteroids have shown small changes in the rate of FEV1 decline and short-term symptomatic relief [9– 13]. All these findings have eluded to the possibility that changes in FEV1 may be the best variable to reflect changes in the underlying disease mechanisms. However, the longterm monitoring of FEV1 changes and symptom-related measures requires a great deal of effort and resources, particularly, when assessing treatment benefit [2]. It has been suggested that the discovery of new markers of disease severity would not only enhance our understanding of COPD’s natural history and pathogenesis but also offer new endpoints that can be used along side standard measures in COPD diagnosis and clinical trials [1,2]. Presently, the medical literature contains numerous crosssectional studies that have tried to correlate clinical variables, counts of cell types and concentrations of inflammatory mediators to FEV1 [14–30]. Examples include body mass index (BMI) [31], neutrophil cell counts [32,33], and tumour necrosis factor (TNF-a) [23,34,35]. Despite the increasing medical burden of COPD, longitudinal studies investigating such potential markers of disease progression are often scarce or currently ongoing [2]. The elucidation of a marker’s potential to discriminate between different disease stages would be useful in establishing appropriate clinical endpoints, particularly for the assessment of treatment effect [2]. Therefore, we reviewed the medical literature using a commonly cited disease staging system to determine the statistical distributions of clinical and laboratory measures obtained from patients with stable COPD.
2. Methods 2.1. Study identification and selection Our search strategy involved the use of PubMed, ISI Web of Science, and Cochrane Review databases and keywords
obtained from US National Library of Medicine, major guidance documents and reference books on stable COPD. Additional details about our search can be found in Appendix A. Study selection was based on the availability of FEV1 and other spirometric, demographic, and possible marker data, the absence of exacerbations and life— threatening co-morbidities, and the withdrawal from COPD medication such as oral corticosteroids for an extended period of time prior to study entry and baseline measurements. Studies that primarily investigated potential markers in COPD patients but did not record FEV1 and other spirometry measures were excluded. 2.2. Objective of the literature review Our objective was to determine the characteristics of published study populations and to understand how baseline variables and potential markers changed with COPD stages. Study subjects included those with mild, moderate and severe COPD that were assessed according to the 1995 ATS COPD staging criteria [6] as well as healthy subjects that were matched for age, sex or height. If baseline information of a specific biochemical measure was unavailable for the healthy subject group of COPD studies, data was retrieved from asthma studies that had healthy subjects of similar characteristics. 2.3. Data abstraction From each study, we abstracted any spirometric, demographic, clinical, cytological, or biochemical variable that was considered standard in the stable COPD literature or had been suggested as a potential marker of COPD severity (Table 1). Other data retrieved concerned the study question, experimental design, sponsorship, and the presence or absence of individual subject data. We first sorted the data according to the four FEV1 postbronchodilator categories of the 1995 ATS disease staging system (Table 2) [6]. This system was selected in lieu of past and present disease staging systems because of its scientific evidence [36–46] and its common use in the design and eligibility requirements of past COPD studies. The boundaries of its FEV1 categories are similar to those found in the newly adopted ATS/ERS 2004 disease staging system (Fig. 1). (Each FEV1 category has been reclassified to a more severe level of disease and the previous ‘Severe’ stage is now spit into ‘Severe’ and ‘Very Severe’. These changes like those seen with past disease staging systems were made mainly for reasons related to clinical management rather than pathophysiological findings [47–49]). The data was then further organized according to sample size as well as smoking status, i.e. non-, ex- and current smoker. Cytological and biochemical data were also classified according to specimen type.
L.G. Franciosi et al. / Pulmonary Pharmacology & Therapeutics 19 (2006) 189–199
191
Table 1 Demographics, spirometry and possible markers of stable COPD Demographics Spirometry Clinical variables
Age; gender; height; weight; pack-years; disease years FEV1 (litres and % predicted); FVC (litres and % predicted); FEV1/FVC ratio Arterial oxygen tension; arterial carbon dioxide tension; b2-agonist reversibility; methacholine, histamine or AMP challenge; body mass index; baseline dyspnea index; 6 minute walking distance; St George’s respiratory questionnaire
Cytological variables Local (Sputum, BAL, Biopsy) Systemic (Plasma or Serum) Biochemical variables Local (Sputum, BAL, Biopsy) Systemic (Plasma or Serum) Exhaled breath
Neutrophils; macrophages; eosinophils; CD8C lymphocytes Neutrophils; macrophages; eosinophils; CD8C lymphocytes Interleukin-6; interleukin-8; TNF-a; fibrinogen; C-reactive protein Interleukin-6; interleukin-8; TNF-a; fibrinogen; C-reactive protein Nitric oxide; carbon monoxide
FEV1, forced expired volume in one second; FVC, forced vital capacity; BAL, broncho-alveolar lavage; TNF-a, tumour necrosis factor a.
2.4. Statistical methods The typical study measures of central tendency and dispersions (i.e. means, medians, standard errors, standard deviations, 95% confidence interval, and inter-quartile ranges) were tabulated and calculated using Microsoft Excel. A fixed effect meta-analyses was used to summarize point estimates, 95% confidence intervals, and study data distributions (i.e. 2SD) [50]. A two-tailed Z-test was then performed whenever possible to statistically assess differences between healthy subjects and a disease stage. The a-level was adjusted for multiple testing according to the Bonferroni correction procedure. In the event a specific disease stage had a large number of studies with only median data available, the normal distribution of the data was assumed and medians were treated as means. During the analyses, data showing non-normality was not logtransformed because individual data sets were not available to us.
3. Results 3.1. Description of studies We identified 652 suitable studies published between 1978 and mid September 2003. These are available at http:// www.pharmacology.leidenuniv.nl/ under the link that
concerns the COPD project summary and reference lists. The total number of subjects in these studies was 146,255, of which 46% were healthy subjects, 26% mild, 16% moderate, and 13% severe COPD patients. 3.2. Clinical variables and relationship with COPD stage The meta-analyses of typical study demographics are provided in Table 3. We found that the pack-years in mild, moderate and severe stages were similar, indicating that analysis of study subjects according to smoking status was not necessary for our analysis. We also found that twothirds of studies provided gender information; studies with healthy subjects had about an equivalent number of males and females whilst the number of males increased and the number of females decreased with advancing COPD. As expected, the point estimates and 95% confidence intervals of FEV1 % predicted (Fig. 2A) showed a clear separation between the healthy state and the different COPD stages from the studies sampled. This reflects the fact that the ATS’ COPD stages are based on determinations of FEV1 % predicted. Among the possible markers of COPD, only the point estimates of arterial oxygen tension (PaO2) demonstrated the same profile as FEV1 (Fig. 2B). Other variables showed either no difference between healthy and mild stages, as with arterial carbon dioxide tension (PaCO2), BMI and the six minute walking distance (6MWD) (Fig. 2C), or showed a binary pattern in that
Table 2 Disease staging systems-past and present Staging system ERS (1995) [4] ATS (1995) [6] BTS (1997) [5] GOLD (2001) [49] GOLD (2003) [48] ATS/ERS (2004) [47]
Disease severity listed according to FEV1 (% predicted) criterion At risk
Mild (%)
Moderate (%)
Severe (%)
Very severe (%)
N.Ap. N.Ap. N.Ap. Normal Normal R80
R70 R50 60–79 R80 R80 R80
50–69 35–49 40–59 30–79 50–79 50–79
!50 !35 !40 !30 or !50CCRF 30–49 30–49
N.Ap. N.Ap. N.Ap. N.Ap. !30 or !50CCRF !30
ERS, European respiratory society; ATS, American thoracic society; BTS, British thoracic society; GOLD, global initiative for chronic obstructive lung disease; CRF, chronic respiratory failure; and N.Ap., not applicable.
192
L.G. Franciosi et al. / Pulmonary Pharmacology & Therapeutics 19 (2006) 189–199
3.3. Cytological variables
Fig. 1. Comparison of ATS 1995 and ATS/ERS 2004 disease staging systems.
statistically significant differences were only seen between the presence of disease irrespective of stage and the healthy state. This was the case for baseline dyspnea index (P! 0.001), beta2 agonist reversibility (P!0.001), and the St George’s Respiratory Questionnaire (SGRQ) (Fig. 2D, P! 0.001). A shortage of studies or appropriate data across disease stages for pharmacological challenge tests (i.e. methacholine, histamine and AMP) did not allow us to evaluate the statistical distributions of these variables.
The two main cell types sampled in sputum—neutrophils and macrophages—showed a significant difference between the healthy state and each of the stages of COPD, while the eosinophils did not. Significant overlap was seen in distributions (2SD) for mild, moderate and severe stages for each cell type. However, the point estimates obtained from the sputum neutrophil studies demonstrated an upward trend from the healthy to severe stages of COPD (Fig. 3A). A meta-analysis of cell types in bronchoalveolar lavage (BAL) and lung biopsies was not possible due to the limited amount of studies and data available (Table 4). 3.4. Biochemical variables Research on biochemical markers in plasma or serum consisted of IL-6, IL-8, TNF-a and CRP. CRP serum concentrations showed no statistical differences between the healthy and each of the COPD stages, but we observed an upward trend with increasing disease severity
Table 3 Typical subject demographics Variable/COPD stagea
Total studiesb
Total subjectsc
Point estimate (95% CI)
Study data distributions (2SD)
Significance testd Z value
P value
Age (years) Healthy Mild Moderate Severe Height (cm) Healthy Mild Moderate Severe Weight (kg) Healthy Mild Moderate Severe Pack-yearse Healthy Mild Moderate Severe Disease years Mild Moderate Severe
628 194 218 229 207 164 39 36 71 67 159 37 32 66 73 204 67 107 67 30 27 11 12 4
143,756 66,312 36,411 22,846 18,187 18,439 6761 3873 6536 1269 12,921 3586 3716 4192 1427 63,785 27,223 23,943 11,861 758 8839 1467 7179 193
– 46.3 (45.6–47.1) 61.2 (60.4–62.1) 65.0 (64.3–65.8) 64.0 (63.3–64.7) – 170.3 (168.6–172.1) 169.3 (167.7–170.8) 167.4 (166.3–168.5) 167.5 (166.3–168.8) – 74.7 (71.7–77.7) 71.2 (67.8–74.5) 68.0 (65.7–70.4) 62.3 (60.6–64.0) – 17.3 (15.8–18.8) 40.5 (38.2–42.9) 39.0 (35.7–42.3) 39.6 (35.1–44.2)
– 36.5–56.2 47.6–74.8 52.0–78.1 53.7–74.3 – 158.2–172.1 159.2–170.8 157.3–168.5 157.5–168.8 – 53.4–96.0 50.5–91.8 46.7–89.4 48.5–76.1 – 7.2–27.5 17.4–63.6 11.0–67.0 17.6–61.7
5.0 (3.6–6.4) 8.2 (5.5–11.0) 10.0 (7.0–12.9)
– – –
– – 26.0 35.0 33.9 – – 0.87 2.77 2.54 – – 1.53 3.39 7.01 – – 16.2 11.7 9.11 – – – –
– – !0.001 !0.001 !0.001 – – 0.38 (NS) 0.0056 0.011 – – 0.13 (NS) !0.001 !0.001 – – !0.001 !0.001 !0.001 – – – –
a
ATS COPD Stages: mild COPD is FEV1R50% predicted; Moderate COPD is FEV1 between 35 and 49% predicted; Severe COPD is FEV1!35% predicted. b Bold numbers indicate total studies (without duplicates) for the specific variable of interest. c Bold numbers indicate total subjects for all disease stages with respect to the specific variable of interest. d Mild, moderate and severe COPD stages were each compared to the healthy stage; a-level adjusted according to the Bonferroni Correction to account for multiple testing; NS, non-significant difference. e Most studies determined pack years from the number of packs per day multiplied by number of years smoked; one pack contains 20 cigarettes.
L.G. Franciosi et al. / Pulmonary Pharmacology & Therapeutics 19 (2006) 189–199
193
Fig. 2. Fixed effect meta-analysis results of selected clinical variables. For each clinical variable, the point estimates (point), 95% confidence intervals (box), and two standard deviations (bars) are presented for mild (FEV1R50% predicted), moderate (FEV1 35–49%), and severe COPD (FEV1!35%). ((A) the two standard deviations of FEV1 % predicted have been omitted for theoretical reasons). N, signifies the total studies, n, the total subjects for each stage and asterisks P!0.001.
(Fig. 3B). For serum TNF-a, there were no statistical differences between the healthy subjects and any of the COPD stages despite a trend for increases with increasing severity of the stages (Fig. 3C). In the case of IL6, IL-8 and fibrinogen, there was insufficient data to perform any meta-analysis (Table 4). There was far less studies in which possible biochemical markers of COPD were measured in healthy subjects, or in patients at various stages of the disease, that were suitable for analysis. We only found enough studies investigating IL-8 in which
sputum values for healthy subjects were significantly different from mild and moderate COPD patients (Fig. 3D). There was insufficient data for the metaanalysis of sputum TNF-a, IL-6, fibrinogen, and CRP. The number of studies that assessed the biochemical markers IL-6, IL-8, TNF-a, fibrinogen and CRP in BAL was also scarce. From the data available, only BAL IL-8 was found to show a statistical difference between the healthy and mild stages (P!0.001). From the evaluation of nitric oxide and carbon monoxide levels in exhaled
194
L.G. Franciosi et al. / Pulmonary Pharmacology & Therapeutics 19 (2006) 189–199
Fig. 3. Fixed effect meta-analysis results of selected biochemical variables. For each biochemical variable, the point estimates (point), 95% confidence intervals (box), and two standard deviations (bars) have been presented for mild (FEV1R50% predicted), moderate (FEV1 35–49%), and severe COPD (FEV1!35%). N, indicates the total studies; n, the total subjects for each stage, and asterisks P!0.001.
breath, the meta-analyses revealed no obvious relationship with COPD severity (Table 4). 4. Discussion The diagnosis and determination of treatment benefit for any disease should be based on sensitive clinical and laboratory measures that reflect differences in disease severity. In particular, each measure should show a clear separation of their statistical distributions across disease
stages. This has been previously illustrated in HIV infection, where CD4 counts and mRNA viral load are correlated with clinical status, antiviral treatment effect, and risk of AIDS [51]. Despite the ongoing research on biomarkers of COPD, little effort has been made in assessing whether such correlations exist. Moreover, the selection of endpoints for the assessment of clinical status and treatment effect have not been based on any formal validation. This is striking since the identification and validation of endpoints that reflect changes in disease status are requirements for both
L.G. Franciosi et al. / Pulmonary Pharmacology & Therapeutics 19 (2006) 189–199
195
Table 4 Variables demonstrating little relationship with COPD severity or have insufficient data for meta-analysis Variable
Total studies
Total subjects
PaCO2 (mmHg) BMI (kg/m2) b2 Agonist reversibility (% FEV1 predicted) Sputum eosinophils (% differential count) 6MWD (meters) Sputum macrophages (% differential count) SGRQ (Total/100) BAL neutrophils (!103/mL) Sputum IL-8 (nanograms/mL) BAL macrophages (!103/mL) Baseline dyspnea index (Score/12) Exhaled nitric oxide (parts per billion) BAL Eosinophils (!103/mL) Biopsy neutrophils (cells/mm2) Biopsy CD8C lymphocytes (cells/mm2) Biopsy eosinophils (cells/mm2) Histamine challenge (mg/mL) Biopsy macrophages (cells/mm2) BAL IL-8 (pg/mL) Methacholine challenge (mg/mL) Exhaled carbon monoxide (parts per million) AMP Challenge (mg/mL) Sputum TNF-a (ng/mL) Blood IL-6 (pg/mL) Blood fibrinogen (g/L) Blood IL-8 (pg/mL) Sputum fibrinogen (mg/L) Sputum IL-6 (pg/mL) BAL TNF-a (pg/mL) Blood eosinophils (!109/mL) BAL IL-6 (pg/mL) Blood CD8C cells (% lymphocytes) Biopsy IL-8 (IL-8C mRNA/mm2) Biopsy TNF-a (TNF-aC mRNA/mm2) Blood neutrophils (!109/mL) Sputum CD8C lymphocytes (% lymphocytes) BAL CD8C Lymphocytes Blood macrophages Sputum C-reactive protein BAL Fibrinogen BAL C-Reactive protein Biopsy IL-6 Biopsy fibrinogen Biopsy C-reactive protein
230 111 78 55 54 53 33 28 26 25 23 22 17 17 14 14 13 13 11 8 7 7 6 5 4 4 3 3 3 2 2 1 1 1 1 1 N.Ap. N.Ap. N.Ap. N.Av. N.Av. N.Av. N.Av. N.Av.
20,977 54,463 23,621 1621 3181 1550 8328 925 864 829 5750 915 522 415 376 306 695 255 397 144 591 228 172 195 6758 106 68 188 103 83 67 66 57 57 51 13 N.Ap. N.Ap. N.Ap. N.Av. N.Av. N.Av. N.Av. N.Av.
medical practice and drug development [52]. Hence, we conducted this study to determine such changes in variables and suggested markers of COPD with respect to clinical status. Given the limited amount of published longitudinal studies and availability of individual time courses, the information derived from the cross-sectional studies that we sampled primarily reflects overall inter-individual variability. Nevertheless, pooling of this baseline data provides insight into which variables may be potential markers of disease severity according to FEV1 status. We determined that PaO2 was the only clinical variable unrelated to spirometry that correlated with the FEV1 staging criteria (Fig. 2B). Other clinical variables such as
BMI, 6MWD (Fig. 2C) and SGRQ (Fig. 2D) displayed overlap in effect sizes, confidence intervals and data distributions. Normally, these variables were only able to distinguish between healthy subjects and patients. With respect to the cell types, only sputum neutrophils exhibited a trend toward separation between COPD stages (Fig. 3A). We were not able to assess the statistical distributions of many cells types in other body specimens due to a shortage of data in the moderate and severe stages. When we reviewed the biochemical variables, we found a separation in point estimates in serum TNF-a and CRP as well as in sputum IL-8 (Fig. 3). However, we were faced with a shortage of data in the later stages of the disease.
196
L.G. Franciosi et al. / Pulmonary Pharmacology & Therapeutics 19 (2006) 189–199
Another aspect we had to consider was the specificity of these cytological and biochemical variables. The few inflammatory cells and mediators that were affected across COPD stages may not bare a causal relationship with the severity of the disease itself. Rather, they may only reflect local and systemic inflammatory conditions and may not be specific to COPD. In addition, the dynamics of inflammatory cell migration was not considered in the sampled matrices (i.e. blood, tissue biopsy, BAL, and sputum). The correspondence between local and systemic measurements of migrating inflammatory cells such as lymphocytes was often unclear. These measurements were usually obtained at different times from different individuals rather than concurrently within each individual. Hence, a parallel sampling scheme for local and systemic matrices would be required for characterising the inflammatory component of this disease. An interesting finding is that pack-years did not change with increasing disease severity. This observation is in agreement with previous studies that 10–15% of smokers get COPD [53,54]. A recent study has implicated differences in activation of bronchial epithelial cells between smokers with COPD and smokers without airway obstruction in releasing neutrophils into the airways [55]. These preliminary findings have a number of limitations. We considered the characteristics of the COPD population mainly from cross-sectional studies. The studies included in this analysis were selected based on a set of criteria (outlined in the Study Identification and Selection section) that should minimize heterogeneity across the populations of interest and eliminate potential bias during comparisons. We realize, however, that protocol violations may have occurred that were not described in the original publications that we reviewed. During the data abstraction process, we noticed that many studies did not explicitly state whether FEV1 was a pre- or post-bronchodilator value. We assumed that investigators were following international guidelines and that they used a post-bronchodilator value for study eligibility. Another issue related to FEV1 is the consistency of its measurement [17,56]. In addition, environmental and biological factors may have contributed to the variation in FEV1 data [36]. With respect to study subjects, the possibility exists that a given subject may have been enrolled in more than one COPD study. This implies that the subject numbers that support this review may be smaller than indicated. Another consideration is that although subjects may be off their medication for study eligibility, it is reasonable to assume that patients might have been on some medication (e.g. long-term oxygen therapy, bronchodilators and oral corticosteroids) that could affect FEV1 status at the time of baseline measurements. We intend to further explore the contribution of confounding factors and covariates such as therapy and age to explain the profiles of FEV1 and other potential markers. As well, the value of multidimensional scales (e.g. the BODE index that combines body-mass index, airflow obstruction, dyspnea,
and exercise capacity [57]) will also be examined for disease classification. However, individual data sets are required for these purposes. Our group is currently performing this work. The use of a disease staging system for the basis for an exploratory analysis of clinical, cytological and biochemical variables implies that its categories have been rigorously evaluated using criterion-based validity methods [58]. However, we realized that past and present systems have FEV1 categories which were mainly inferred from clinical management rather than scientific studies [36,48,49]. In that sense, there is growing awareness that a mechanism-based approach should be used to define and classify disease markers and staging systems [52,59,60]. Disease markers can characterise a process on the causal pathway between disease and treatment response [52]. A mechanism-based classification has been proposed which divides biological markers into six categories from genotype to clinical scales [59,60]. Thus far, it appears that only physiological measures and clinical scales have been included (Types 4 and 6 biomarkers, respectively) in COPD trials. There has been no systematic evaluation of genotype/phenotype (Type 0), target expression and activation (Type 3), or disease process (Type 5) other than the rate and frequency of COPD exacerbations. The process of selecting and validating a marker involves evaluating the effect of disease on the marker as well as its sensitivity to treatment effect. This can be achieved by mathematical modelling of the time course of disease and its variability. A rank of sensitive markers can then be used to study treatment response in patient populations or in individual patients. In addition, such markers should be specific in differentiating disease from the healthy state. These requirements are essential to minimise false positives and false negatives in diagnosis and treatment. In chronic diseases, clinical use of a biomarker requires two additional aspects to be taken into account, namely the rate and magnitude of change that ultimately contribute to the estimation of group size and statistical power of a study. If endpoints do not change over time, how do we define treatment efficacy? We need markers that are sensitive to disease severity. Despite the separation of point estimates for the potential markers that we identified in this investigation, their statistical distributions often demonstrate considerable overlap when their data have been ranked accordingly. For example, when the data retrieved from the cross-sectional studies of sputum neutrophils were ranked from lowest to highest available percentages, the ATS’ FEV1-based disease stages begin to overlap with increasing disease severity (Fig. 4). This highlights the need to look beyond FEV1 as the major endpoint for classifying COPD severity and to explore these elements in longitudinal studies. An established approach to assessing disease severity and progression as well as predicting treatment effect is mechanism-based mathematical modelling. Such models
L.G. Franciosi et al. / Pulmonary Pharmacology & Therapeutics 19 (2006) 189–199
197
Fig. 4. Ranked sputum neutrophil data demonstrating overlap of the ATS’ FEV1-based disease stages. Presented are mean or median percentage of sputum neutrophils in a percent differential count that have been ranked from the lowest to highest value obtained from the sampled literature. The blue bars represent studies with control healthy subjects; the green bars are studies with mild COPD patients (FEV1R50% predicted); the yellow bars are studies with moderate COPD patients (FEV1 35–49%); and the red bars are studies with severe COPD patients (FEV1!35%). N, indicates the total sum of subjects obtained from all available studies sampled.
can be useful in a number of ways, particularly to support the selection of dosing regimen as well as study design and treatment duration [51,61]. A major advantage of modelling is the ability to estimate the placebo effect and separate it from the pharmacological response, which eliminates the enormous bias of the ‘last observation carried forward (LOCF)’ procedure common in longitudinal studies. One example of this approach is pharmacokinetic-pharmacodynamic modelling in which mechanism-based parameters are used as fixed effects to explain the time course of disease or treatment effect. Moreover, it incorporates random effects that describe intra-individual and inter-individual differences and their change over time. Similarly, changes in disease severity can be encompassed in one or more states using Markov models. With this approach, one can predict the number of individuals in each disease state at each specified observation time and their rate of transition to subsequent states. Given the characteristics of COPD, we feel that the current research process is not robust enough to elucidate possible biomarkers or endpoints for treatment response. Pharmacostatistical modelling is a powerful tool to characterise the time course of disease and treatment effect in both populations and individual patients as well as estimates of uncertainty. This is paramount to our current efforts to identify biomarkers in COPD. In summary, the statistical distributions obtained from our cross-sectional studies revealed that many clinical, cytological and biochemical variables in COPD showed little or no relationship with FEV1-based disease staging criteria. From the available studies, where more than spirometry was measured, we did identify five possible markers, i.e. PaO2, sputum neutrophils and IL-8, and systemic TNF-a and CRP, that showed a separation
between disease stages. In addition, this review provides insight about the areas, where more data should be acquired to conclude about the relevance of the measures investigated thus far. Longitudinal studies are required to determine how these possible markers reflect the time course of disease within a single individual, but the implementation of such studies may have practical, ethical and financial consequences. Therefore, we are undertaking an integrated mechanism-based modelling approach that offers a better solution for improving the diagnosis as well as the assessment of treatment effect in patients with COPD.
Acknowledgements We like to thank Sandra van den Berg-van Tol, secretary at the LACDR Division of Pharmacology for her assistance in the preparation of the website containing the study references for this work. We also like to thank senior statisticians Nigel Dallow and Andrew M. Wright from GlaxoSmithKline UK for their statistical advice and guidance during the meta-analyses. The current research is part of L. Franciosi’s post-doctoral fellowship on COPD disease progression funded by GlaxoSmithKline UK.
Appendix A. Search strategy We performed a search of the PubMed, Medline, ISI Web of Science, and Cochrane Review databases up to mid September 2003. All language types were included. The principal search for the disease’s name were: chronic obstructive pulmonary disease; COPD; chronic obstructive
198
L.G. Franciosi et al. / Pulmonary Pharmacology & Therapeutics 19 (2006) 189–199
airway disease; COAD; chronic airflow obstruction; CAO; chronic obstructive lung disease; COLD; chronic bronchitis; CB; bronchitis; blue bloaters; pulmonary emphysema; pink puffers; and small airways disease. These terms were then combined with terms designating a classification or a subpopulation such as control; healthy; mild; moderate; severe; very severe; ATS (American Thoracic Society); BTS (British Thoracic Society); ERS (European Respiratory Society); GOLD; disease, stages, states; clinical; human; current smoker, never smoker, non-smoker, ex-smoker; and stable. Major COPD symptoms like chronic cough, sputum, dyspnea (or dyspnoea) as well as drug classes such as beta agonists, corticosteroids, anticholinergics, and phosphodiesterase four inhibitors were also included. The resulting citation list was then refined to include retrospective and prospective clinical studies, but not abstracts, reviews, meta-analyses, case reports, letters, editorials or duplicate publications. Then, common variables and suggested markers of clinical, cytological or biochemical origin obtained from a preliminary assessment of large studies, review articles, and guidance documents were searched within this list. These variables were further searched in combination with their respective measurement technique or specimen type. Keywords included spirometry, biopsy, bronchial; bronchoalveolar lavage; BAL; sputum, induction; exhaled breath; and systemic, plasma, serum. The resulting list contained 934 citations of interest. From the title and abstract of each these citations, we assessed the availability of variable information and spirometry data for subsequent disease classification. Studies that investigated COPD patients with respect to pharmacoeconomics, respiratory failure, lung transplant surgery, or other advanced life–threatening co-morbidities such as terminal cancer were excluded. Original articles were then retrieved in electronic or hard copy formats and reviewed to determine if each clinical study had a clear description of the subject eligibility criteria, smoking status, sample sizes and the presence of spirometry and anthropometric data at baseline or study entry. The final database contained 652 studies that were eligible for descriptive analysis.
[5]
[6]
[7] [8]
[9]
[10]
[11]
[12]
[13]
[14] [15]
[16] [17]
[18] [19]
[20]
References [1] American Thoracic Society/European Respiratory Society. Definition, Diagnosis and Staging of COPD. American Thoracic Society Website 2004 May 24 [cited 2004 May 24]; Available from: URL: http://www.thoracic.org/COPD/1/definitions.asp. [2] Croxton TL, Weinmann GG, Senior RM, Wise RA, Crapo JD, Buist AS. Clinical research in chronic obstructive pulmonary disease: needs and opportunities. Am J Respir Crit Care Med 2003;167(8): 1142–9. [3] Calverley PM, Walker P. Chronic obstructive pulmonary disease. Lancet 2003;362(9389):1053–61. [4] Siafakas NM, Vermeire P, Pride NB, Paoletti P, Gibson J, Howard P, et al. Optimal assessment and management of chronic obstructive
[21]
[22]
[23]
pulmonary disease (COPD). The European respiratory society task force. Eur Respir J 1995;8(8):1398–420. The COPD Guidelines Group of the Standards of Care Committee of the BTS. BTS guidelines for the management of chronic obstructive pulmonary disease. Thorax 1997;52(5):S1–28 American Thoracic Society. Standards for the diagnosis and care of patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1995;152(5 Pt 2):S77–121 Fletcher C, Peto R. The natural history of chronic airflow obstruction. Br Med J 1977;1(6077):1645–8. Anthonisen NR, Connett JE, Murray RP. Smoking and lung function of Lung Health Study participants after 11 years. Am J Respir Crit Care Med 2002;166(5):675–9. Burge PS, Calverley PM, Jones PW, Spencer S, Anderson JA, Maslen TK. Randomised, double blind, placebo controlled study of fluticasone propionate in patients with moderate to severe chronic obstructive pulmonary disease: the ISOLDE trial. BMJ 2000; 320(7245):1297–303. Pauwels RA, Lofdahl CG, Laitinen LA, Schouten JP, Postma DS, Pride NB. Long-term treatment with inhaled budesonide in persons with mild chronic obstructive pulmonary disease who continue smoking. European respiratory society study on chronic obstructive pulmonary disease. N Engl J Med 1999;340(25):1948–53. Vestbo J, Sorensen T, Lange P, Brix A, Torre P, Viskum K. Longterm effect of inhaled budesonide in mild and moderate chronic obstructive pulmonary disease: a randomised controlled trial. Lancet 1999;353(9167):1819–23. Lung Health Study Research Group. Effect of inhaled triamcinolone on the decline in pulmonary function in chronic obstructive pulmonary disease. N Engl J Med 2000;343(26):1902–9. Weir DC, Bale GA, Bright P, Sherwood BP. A double-blind placebocontrolled study of the effect of inhaled beclomethasone dipropionate for 2 years in patients with nonasthmatic chronic obstructive pulmonary disease. Clin Exp Allergy 1999;29(2):125–8. Barnes PJ. Mechanisms in COPD: differences from asthma. Chest 2000;117(2):10S–114. Kessler R, Faller M, Fourgaut G, Mennecier B, Weitzenblum E. Predictive factors of hospitalization for acute exacerbation in a series of 64 patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1999;159(1):158–64. Weiss ST, DeMeo DL, Postma DS. COPD: problems in diagnosis and measurement. Eur Respir J Suppl 2003;41:4s–112. Crapo RO, Jensen RL, Hargreave FE. Airway inflammation in COPD: physiological outcome measures and induced sputum. Eur Respir J Suppl 2003;41:19s–228. Wouters EF, Creutzberg EC, Schols AM. Systemic effects in COPD. Chest 2002;121(5):127S–130. Agusti AG, Noguera A, Sauleda J, Sala E, Pons J, Busquets X. Systemic effects of chronic obstructive pulmonary disease. Eur Respir J 2003;21(2):347–60. Tashkin DP, Altose MD, Connett JE, Kanner RE, Lee WW, Wise RA. Methacholine reactivity predicts changes in lung function over time in smokers with early chronic obstructive pulmonary disease. The Lung Health Study research group. Am J Respir Crit Care Med 1996;153(6 Pt 1):1802–11. Scanlon PD, Connett JE, Waller LA, Altose MD, Bailey WC, Buist AS. Smoking cessation and lung function in mild-to-moderate chronic obstructive pulmonary disease. The Lung Health Study. Am J Respir Crit Care Med 2000;161(2 Pt 1):381–90. Rutgers SR, Timens W, Kauffman HF, Postma DS. Markers of active airway inflammation and remodelling in chronic obstructive pulmonary disease. Clin Exp Allergy 2001;31(2):193–205. Yasuda N, Gotoh K, Minatoguchi S, Asano K, Nishigaki K, Nomura M, et al. An increase of soluble Fas, an inhibitor of apoptosis, associated with progression of COPD. Respir Med 1998; 92(8):993–9.
L.G. Franciosi et al. / Pulmonary Pharmacology & Therapeutics 19 (2006) 189–199 [24] Barnes PJ. Chronic obstructive pulmonary disease. N Engl J Med 2000;343(4):269–80. [25] Walter RE, Beiser A, Givelber RJ, O’Connor GT, Gottlieb DJ. Association between glycemic state and lung function: the Framingham heart study. Am J Respir Crit Care Med 2003;167(6):911–6. [26] Dahl M, Tybjaerg-Hansen A, Vestbo J, Lange P, Nordestgaard BG. Elevated plasma fibrinogen associated with reduced pulmonary function and increased risk of chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2001;164(6):1008–11. [27] Dev D, Wallace E, Sankaran R, Cunniffe J, Govan JR, Wathen CG, et al. Value of C-reactive protein measurements in exacerbations of chronic obstructive pulmonary disease. Respir Med 1998;92(4): 664–7. [28] Soler N, Ewig S, Torres A, Filella X, Gonzalez J, Zaubet A. Airway inflammation and bronchial microbial patterns in patients with stable chronic obstructive pulmonary disease. Eur Respir J 1999;14(5): 1015–22. [29] Zhu J, Majumdar S, Qiu Y, Ansari T, Oliva A, Kips JC, et al. Interleukin-4 and interleukin-5 gene expression and inflammation in the mucus-secreting glands and subepithelial tissue of smokers with chronic bronchitis. Lack of relationship with CD8(C) cells. Am J Respir Crit Care Med 2001;164(12):2220–8. [30] Bhowmik A, Seemungal TA, Sapsford RJ, Wedzicha JA. Relation of sputum inflammatory markers to symptoms and lung function changes in COPD exacerbations. Thorax 2000;55(2):114–20. [31] Landbo C, Prescott E, Lange P, Vestbo J, Almdal TP. Prognostic value of nutritional status in chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1999;160(6):1856–61. [32] Stockley RA. Neutrophils and the pathogenesis of COPD. Chest 2002; 121(5):151S–15. [33] Beeh KM, Beier J, Kornmann O, Mander A, Buhl R. Long-term repeatability of induced sputum cells and inflammatory markers in stable, moderately severe COPD. Chest 2003;123(3):778–83. [34] Keatings VM, Collins PD, Scott DM, Barnes PJ. Differences in interleukin-8 and tumor necrosis factor-alpha in induced sputum from patients with chronic obstructive pulmonary disease or asthma. Am J Respir Crit Care Med 1996;153(2):530–4. [35] Sakao S, Tatsumi K, Igari H, Watanabe R, Shino Y, Shirasawa H, et al. Association of tumor necrosis factor-alpha gene promoter polymorphism with low attenuation areas on high-resolution CT in patients with COPD. Chest 2002;122(2):416–20. [36] American Thoracic Society. Lung function testing: selection of reference values and interpretative strategies. Am Rev Respir Dis 1991;144(5):1202–18. [37] Burrows B, Lebowitz MD, Camilli AE, Knudson RJ. Longitudinal changes in forced expiratory volume in one second in adults. Methodologic considerations and findings in healthy nonsmokers. Am Rev Respir Dis 1986;133(6):974–80. [38] Knudson RJ, Lebowitz MD, Holberg CJ, Burrows B. Changes in the normal maximal expiratory flow-volume curve with growth and aging. Am Rev Respir Dis 1983;127(6):725–34. [39] Crapo RO, Morris AH, Gardner RM. Reference spirometric values using techniques and equipment that meet ATS recommendations. Am Rev Respir Dis 1981;123(6):659–64. [40] Paoletti P, Pistelli G, Fazzi P, Viegi G, Di PF, Giuliano G, et al. Reference values for vital capacity and flow-volume curves from a general population study. Bull Eur Physiopathol Respir 1986;22(5): 451–9. [41] Lebowitz MD, Quackenboss J, Camilli AE, Bronnimann D, Holberg CJ, Boyer B. The epidemiological importance of intraindividual changes in objective pulmonary responses. Eur J Epidemiol 1987;3(4):390–8.
199
[42] Cochrane GM, Prieto F, Clark TJ. Intrasubject variability of maximal expiratory flow volume curve. Thorax 1977;32(2):171–6. [43] Miller MR, Pincock AC. Predicted values: how should we use them? Thorax 1988;43(4):265–7. [44] Miller AJ. Pulmonary function tests in clinical and occupational lung disease. Orlando: Grune and Stratton; 1986. [45] Oldham PD. Percent of predicted as the limit of normal in pulmonary function testing: a statistically valid approach. Thorax 1979;34(4): 569–70. [46] Sobol BJ. Assessment of ventilatory abnormality in the asymptomatic subject: an exercise in futility. Thorax 1966;21(5):445–9. [47] Celli BR, MacNee W. Standards for the diagnosis and treatment of patients with COPD: a summary of the ATS/ERS position paper. Eur Respir J 2004;23(6):932–46. [48] Fabbri LM, Hurd SS. Global strategy for the diagnosis, management and prevention of COPD: update. Eur Respir J 2003;22(1):1–2. [49] Pauwels RA, Buist AS, Calverley PM, Jenkins CR, Hurd SS. Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease. NHLBI/WHO Global initiative for chronic obstructive lung disease (GOLD) workshop summary. Am J Respir Crit Care Med 2001;163(5):1256–76. [50] Hedges LV, Vevea JL. Fixed- and random-effects models in metaanalysis. Psychol Methods 1998;3(4):486–504. [51] De Gruttola VG, Clax P, DeMets DL, Downing GJ, Ellenberg SS, Friedman L, et al. Considerations in the evaluation of surrogate endpoints in clinical trials. summary of a national institutes of health workshop. Control Clin Trials 2001;22(5):485–502. [52] Frank R, Hargreaves R. Clinical biomarkers in drug discovery and development. Nat Rev Drug Discov 2003;2(7):566–80. [53] Saetta M, Turato G, Maestrelli P, Mapp CE, Fabbri LM. Cellular and structural bases of chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2001;163(6):1304–9. [54] Sherrill DL, Lebowitz MD, Burrows B. Epidemiology of chronic obstructive pulmonary disease. Clin Chest Med 1990;11(3): 375–87. [55] Schulz C, Kratzel K, Wolf K, Schroll S, Kohler M, Pfeifer M. Activation of bronchial epithelial cells in smokers without airway obstruction and patients with COPD. Chest 2004;125(5): 1706–13. [56] Eaton T, Withy S, Garrett JE, Mercer J, Whitlock RM, Rea HH. Spirometry in primary care practice: the importance of quality assurance and the impact of spirometry workshops. Chest 1999; 116(2):416–23. [57] Celli BR, Cote CG, Marin JM, Casanova C, Montes de OM, Mendez RA, et al. The body-mass index, airflow obstruction, dyspnea, and exercise capacity index in chronic obstructive pulmonary disease. N Engl J Med 2004;350(10):1005–12. [58] Colburn WA. Optimizing the use of biomarkers, surrogate endpoints, and clinical endpoints for more efficient drug development. J Clin Pharmacol 2000;40(12 Pt 2):1419–27. [59] Aarons L, Karlsson MO, Mentre F, Rombout F, Steimer JL, van Peer A. Role of modelling and simulation in Phase I drug development. Eur J Pharm Sci 2001;13(2):115–22. [60] Danhof M. Markers of Pharmacological and Toxicological Action. Haemostas Foundation Leiden Workshop 2001 April 19 [cited 2004 May 24]; Available from: URL: http://www.haemostas.nl/files/ abstract23danhof.doc. [61] Borg S, Ericsson A, Wedzicha J, Gulsvik A, Lundback B, Donaldson GC, et al. A computer simulation model of the natural history and economic impact of chronic obstructive pulmonary disease. Value Health 2004;7(2):153–67.