Comments
Abstract
Rationale: There is continuing debate about whether to define airflow obstruction by a post-bronchodilator ratio of forced expiratory volume in 1 second (FEV1) and forced vital capacity (FVC) below 0.70, or by ratio values falling below the age-dependent lower limit of normal (LLN) derived from general population data.
Objectives: To determine whether using the LLN criterion affects the classification and outcomes of patients previously defined as having chronic obstructive pulmonary disease by the fixed FEV1/FVC ratio.
Methods: We applied the LLN definition to pooled data from the Tiotropium Safety and Performance in Respimat study that used the fixed FEV1/FVC ratio for the clinical diagnosis of chronic obstructive pulmonary disease.
Results: A total of 17,072 patients were analyzed; of these, 1,807 (10.6%) patients had a ratio greater than or equal to LLN. Patients with a ratio greater than or equal to LLN had similar risks of death from any cause and fatal major adverse cardiovascular (CV) event as those below LLN. Patients with a ratio below LLN had a significantly lower risk of major adverse CV events (hazard ratio = 0.69; 95% confidence interval [CI] = 0.55–0.86; P = 0.001), and had significantly greater risks of moderate to severe exacerbation (rate ratio = 1.48; 95% CI = 1.36–1.61; P < 0.0001) and severe exacerbation (rate ratio = 2.01; 95% CI = 1.68–2.40; P < 0.0001) when compared with patients greater than or equal to LLN. Study outcomes by treatment arm (5 μg tiotropium Respimat vs. 18 μg HandiHaler) were comparable.
Conclusions: Using the LLN to define airflow obstruction would have excluded patients in the Tiotropium Safety and Performance in Respimat study with a higher risk of nonfatal major adverse CV events and a lower risk of exacerbation; study outcomes by treatment arm (2.5 μg/5 μg tiotropium Respimat vs. 18 μg HandiHaler) remained similar.
Clinical trial registered with www.clinicaltrials.gov (NCT01126437).
International guidelines for chronic obstructive pulmonary disease (COPD) define the presence of airflow obstruction spirometrically as an forced expiratory volume in 1 second/forced vital capacity (FEV1/FVC) ratio below 0.70 (sometimes described as 70%), as this is a simple and objective way to identify persistent airflow limitation in clinical practice (1–4). However, the FEV1/FVC ratio decreases with age, leading to overdiagnosis of COPD in the elderly and potential misdiagnosis of adults below age 55 years (particularly those with mild disease and few symptoms) (1, 5–8).
One way to address these age-related problems is to use a statistically defined, age-specific lower limit of normal (LLN) of the FEV1/FVC ratio as the threshold to define airflow limitation. LLN is defined as the lower fifth percentile of the frequency distribution for a reference “healthy” population (equivalent in sex, height, ethnicity, and age), and its use may reduce the risk of misclassification (5, 7, 8). However, there is still debate about which criterion is most appropriate in clinical COPD diagnosis, in part due to the lack of a reference standard or comparative data (1, 5). The correlation of fixed ratio versus LLN with clinical outcomes has been reported in a number of studies (5, 9–13), with unclear results about which criterion might be better. One pragmatic approach has been to confine the diagnosis of COPD to patients with a ratio less than 0.70 and an absolute FEV1 less than 80% predicted (14). Nonetheless, advocates of the LLN approach have pointed out the potential risks of inappropriate diagnosis and the impact this might have on the results of clinical trials (15, 16).
The TIOSPIR (Tiotropium Safety and Performance in Respimat) trial was a prospective evaluation of the safety and exacerbation reduction efficacy of tiotropium Respimat versus tiotropium HandiHaler (17, 18). As TIOSPIR is the largest (N = 17,135) long-term, randomized international (50 countries) study of patients with COPD performed to date, we explored the clinical implications of the LLN criteria in a large population of patients with a clinical diagnosis of COPD over an extended period. Specifically, we examined how many patients would be reclassified as “normal” when we applied the LLN threshold (19), what the characteristics and clinical course of these patients were, and whether this reclassification affected the result of the TIOSPIR trial.
TIOSPIR was a large (N = 17,135), long-term (2–3 yr), randomized, double-blind, double-dummy, parallel-group, active-controlled, event-driven trial, which has been described in detail previously (17, 18).
The primary safety endpoint was time to all-cause death. The primary efficacy endpoint was time to first COPD exacerbation. Secondary endpoints included number of COPD exacerbations, time to first moderate/severe exacerbation (moderate exacerbations required a prescription for antibiotics, systemic corticosteroids, or both [with no hospitalization]; severe exacerbations required hospitalization [due to COPD]), time to and number of severe exacerbations, and time to major adverse cardiovascular (CV) events (MACE; comprising stroke, transient ischemic attack [TIA], myocardial infarction [MI], sudden death, cardiac death, sudden cardiac death, or fatal event in the system organ classes [SOCs] for cardiac and vascular disorders). SOCs were defined according to the Medical Dictionary for Regulatory Activities standardizing attributions of adverse events (AEs) globally. The cause of death was adjudicated by an independent mortality adjudication committee.
Patients were randomized to once-daily 2.5 μg tiotropium Respimat (two puffs of 1.25 μg) or 5 μg (two puffs of 2.5 μg), or once-daily 18 μg tiotropium HandiHaler. The study did not include a washout period from COPD treatments before randomization. All patients (including those with premature discontinuation) were followed for vital status until end of study. However, data on exacerbations were only available for patients who did not withdraw from follow-up. The trial was performed in accordance with the Declaration of Helsinki; the study protocol and procedures were approved by the relevant institutional review boards and ethics committees. All patients provided written informed consent.
TIOSPIR enrolled patients with a clinical diagnosis of COPD according to accepted criteria (1), aged 40 years or greater, with a post-bronchodilator FEV1 of 70% predicted or less, FEV1/FVC of 0.70 or less, and 10 pack-years or greater of smoking history. Except for those participants in the lung function substudy (20), spirometry assessment was based on routinely acquired data within 6 months of study entry. Usual background treatment for COPD (except for other inhaled anticholinergics) was permitted.
The LLN of FEV1/FVC was calculated using equations defined by the European Respiratory Society’s Global Lung Function Initiative (GLI) in 2012 (19) (with and without adjustment for ethnicities other than white) or the European Community for Steel and Coal (ECSC) in 1993 (21); these equations also accommodate age, height, and sex. Patients were allocated to subgroups according to FEV1/FVC ratio less than LLN or greater than or equal to LLN (GLI or ECSC criteria). As primary endpoints did not differ between 2.5/5 μg tiotropium Respimat and 18 μg HandiHaler groups (18), data were pooled for most analyses.
For analysis of risk of death (time to all-cause death, including fatal MACE), events occurring during treatment and vital status follow-up (follow-up for deaths even if the patient discontinued early) were considered, whereas, for other analyses (MACE and exacerbations), only events within the on-treatment period were evaluated.
Hazard ratios (HR) and 95% confidence intervals (CIs) were calculated using a Cox proportional hazards regression model (without covariate adjustment) and demonstrated via Kaplan–Meier plots. Incidence rate ratios and 95% CIs were used to compare incidence rates. Negative binomial regression models were used to examine annual exacerbation rates. HRs adjusted for covariates (patient demographics and baseline characteristics) were calculated using stepwise regression (please refer to the online supplement for further details.).
Boehringer Ingelheim, the study sponsor, participated in the design of the study, the collection and analysis of the data, and the preparation of this manuscript.
Of 17,116 patients treated in TIOSPIR, 17,072 were classified by the LLN criteria (71.5% male, mean age of 65.0 yr), of whom 271 patients (1.6%) had an FEV1 between 70% and 80% predicted, and 64 (0.4%) had an FEV1 greater than or equal to 80% predicted (Table 1); 44 were excluded from the analysis (1 elderly patient aged >95 yr and 43 patients who had missing FEV1/FVC ratio data). As the baseline characteristics of patients categorized by LLN were similar when using either the GLI or ECSC equations (see Table E1 in the online supplement), the GLI equations for white individuals were selected to further determine the effects of the reclassification on the clinical outcomes of CV safety and mortality, and the impact on exacerbations.
| Baseline Characteristics | LLN Classification Group | Total (N = 17,072) | |
|---|---|---|---|
| FEV1/FVC < LLN Predicted n (%) = 15,265 (89.4) | FEV1/FVC ≥ LLN Predicted n (%) = 1,807 (10.6) | ||
| Duration of follow-up, mean (SD), d | 838.2 (145.1) | 837.7 (144.3) | 838.2 (145.0) |
| Age, mean (SD), yr | 64.6 (9.1) | 68.5 (8.4) | 65.0 (9.1) |
| Male, n (%) | 10,870 (71.2) | 1,334 (73.8) | 12,204 (71.5) |
| Race, n (%) | |||
| White | 12,394 (81.2) | 1,532 (84.8) | 13,926 (81.6) |
| Black | 233 (1.5) | 23 (1.3) | 256 (1.5) |
| Asian | 2,246 (14.7) | 177 (9.8) | 2,423 (14.2) |
| BMI, mean (SD), kg/m2 | 25.9 (5.6) | 28.7 (5.9) | 26.2 (5.7) |
| BMI class, n (%) | |||
| Underweight, <18.5, kg/m2 | 1,053 (6.9) | 44 (2.4) | 1,097 (6.4) |
| Normal, 18.5 to <25, kg/m2 | 6,220 (40.7) | 447 (24.7) | 6,667 (39.1) |
| Overweight, 25 to <30, kg/m2 | 4,850 (31.8) | 647 (35.8) | 5,497 (32.2) |
| Obese, ≥30, kg/m2 | 3,142 (20.6) | 669 (37.0) | 3,811 (22.3) |
| mMRC breathlessness score, n (%) | |||
| <2 | 6,837 (44.8) | 930 (51.6) | 7,767 (45.6) |
| ≥2 | 8,411 (55.2) | 873 (48.4) | 9,284 (54.4) |
| Sputum-producing cough, n (%) | 9,739 (63.8) | 1,122 (62.2) | 10,861 (63.7) |
| FEV1 % predicted, classes, n (%) | |||
| FEV1 ≥80% | 37 (0.2) | 27 (1.5) | 64 (0.4) |
| GOLD stage II: FEV1 50 to <80%, n (%) | 5,853 (38.3) | 1,358 (75.2) | 7,211 (42.2) |
| FEV1 70 to <80% | 177 (1.2) | 94 (5.2) | 271 (1.6) |
| FEV1 50 to <70% | 5,676 (37.2) | 1,264 (70.0) | 6,940 (40.7) |
| FEV1 30 to <50% | 7,050 (46.2) | 387 (21.4) | 7,437 (43.6) |
| FEV1 <30% | 2,325 (15.2) | 35 (1.9) | 2,360 (13.8) |
| FEV1 % predicted, mean (SD) | 45.1 (13.4) | 56.9 (11.2) | 46.3 (13.7) |
| FVC % predicted, mean (SD) | 73.4 (18.1) | 64.7 (12.8) | 72.5 (17.8) |
| FEV1/FVC ratio, mean (SD) | 0.48 (0.10) | 0.67 (0.04) | 0.50 (0.11) |
| FVC % predicted, classes, n (%) | |||
| ≥80% | 5,306 (34.8) | 134 (7.4) | 5,440 (31.9) |
| 50% to <80% | 8,486 (55.6) | 1,437 (79.5) | 9,923 (58.1) |
| 30% to <50% | 1,385 (9.1) | 211 (11.7) | 1,596 (9.3) |
| <30% | 88 (0.6) | 25 (1.4) | 113 (0.7) |
Based on the GLI (white) equation, 10.6% (n = 1,807) of patients had an FEV1/FVC greater than or equal to LLN (Table 1). For most patients who would have been excluded had the LLN been used as the spirometric inclusion criterion rather than the fixed threshold of 0.70, the FEV1/FVC was close to the LLN threshold (Figure 1A). The cumulative distribution (Figure 1B) shows that the majority of patients had an FEV1 % predicted close to 70%.
Figure 1. (A) Cumulative distribution and number of patients by difference of forced expiratory volume in 1 second/forced vital capacity (FEV1/FVC) % from the lower limit of normal (LLN). (B) Cumulative distribution and number of patients by FEV1 % predicted, for patients who would have been excluded from TIOSPIR (Tiotropium Safety and Performance in Respimat) if the FEV1/FVC greater than or equal to LLN criteria were used. Global Lung Function Initiative 2012 criteria were used to calculate FEV1 % predicted.
[More] [Minimize]Patients with FEV1/FVC greater than or equal to LLN were older and had a higher body mass index (BMI) at baseline than those with FEV1/FVC less than LLN (Table 1). Post-bronchodilator screening at baseline showed that patients with FEV1/FVC greater than or equal to LLN also had a relatively greater reduction in FVC than patients with FEV1/FVC less than LLN, and the majority (>75%) of patients with FEV1/FVC greater than or equal to LLN were classified as GOLD (Global Initiative for Chronic Obstructive Lung Disease) stage II (FEV1 = 50 to <80%; Table 1; Table E1).
Patients with FEV1/FVC greater than or equal to LLN were more likely to report a history of CV disorders (including MI, ischemic heart disease/coronary artery disease, cardiac arrhythmia, stroke, and TIA) and use more CV medication and less pulmonary medication than those with FEV1/FVC less than LLN at baseline (Table 2; Table E1).
| Baseline Characteristics | LLN Classification Group | Total (N = 17,072) | |
|---|---|---|---|
| FEV1/FVC < LLN Predicted n (%) = 15,265 (89.4) | FEV1/FVC ≥ LLN Predicted n (%) = 1,807 (10.6) | ||
| No. of exacerbations in previous year, n (%) | |||
| 0 | 7,820 (51.2) | 979 (54.2) | 8,799 (51.5) |
| 1 | 4,357 (28.5) | 512 (28.3) | 4,869 (28.5) |
| ≥2 | 3,082 (20.2) | 313 (17.3) | 3,395 (19.9) |
| Cardiac history, n (%)* | 3,864 (25.3) | 596 (33.0) | 4,460 (26.1) |
| MI | 875 (5.7) | 148 (8.2) | 1,023 (6.0) |
| IHD/CAD | 2,203 (14.4) | 386 (21.4) | 2,589 (15.2) |
| Cardiac arrhythmia | 1,596 (10.5) | 226 (12.5) | 1,822 (10.7) |
| History of, n (%) | |||
| Stroke | 319 (2.1) | 68 (3.8) | 387 (2.3) |
| TIA | 207 (1.4) | 38 (2.1) | 245 (1.4) |
| Pulmonary medication, n (%) | 13,905 (91.1) | 1,556 (86.1) | 15,461 (90.6) |
| ICS (not LABA) | 1,097 (7.2) | 132 (7.3) | 1,229 (7.2) |
| LABA (not ICS) | 1,473 (9.6) | 226 (12.5) | 1,699 (10.0) |
| Both ICS and LABA | 8,098 (53.0) | 756 (41.8) | 8,854 (51.9) |
| Neither ICS nor LABA | 4,597 (30.1) | 693 (38.4) | 5,290 (31.0) |
| CV medication, n (%) | 7,558 (49.5) | 1,174 (65.0) | 8,732 (51.1) |
Patients with FEV1/FVC greater than or equal to LLN had a risk of all-cause death (HR = 1.06; 95% CI = 0.90–1.26) and fatal MACE (HR = 1.06; 95% CI = 0.75–1.49), which was similar to that in patients with FEV1/FVC less than LLN (Figures 2A and 2B and Table 3). Mortality risk was not affected by LLN classification in any treatment arm (5 μg Respimat or 18 μg HandiHaler; Table E2).
Figure 2. (A) Risk of death, (B) fatal major adverse cardiovascular event (MACE), (C) MACE, (D) moderate to severe exacerbation, and (E) severe exacerbations (hospitalization), by the lower limit of normal (LLN) forced expiratory volume in 1 second/forced vital capacity (FEV1/FVC) (Global Lung Function Initiative [white]) classification. *Vital status follow-up. †On-treatment analysis: exacerbations, from randomization to drug stop date + 1 day; MACE, from randomization to drug stop date + 30 days. Comparisons are for risk, measured as time to first event (≥LLN vs. <LLN). MACE included stroke, transient ischemic attack, myocardial infarction, sudden death, cardiac death, sudden cardiac death, or fatal event in the system organ classes for cardiac and vascular disorders. Severe exacerbations were defined as those requiring hospitalization. CI = confidence interval; HR = hazard ratio; pred = predicted.
[More] [Minimize]| FEV1/FVC <LLN (n = 15,265) | FEV1/FVC ≥LLN (n = 1,807) | HR or IRR (95% CI) ≥LLN vs. <LLN | |
|---|---|---|---|
| Patients who died, n (%)* | |||
| Causes of death, n (rate)† | 1,152 (7.5) | 145 (8.0) | IRR, 1.06 (0.90–1.26) |
| Cardiac | 58 (0.2) | 8 (0.2) | IRR, 1.17 (0.56–2.44) |
| Gastrointestinal | 26 (0.1) | 6 (0.1) | IRR, 1.95 (0.80–4.74) |
| General | 277 (0.8) | 41 (1.0) | IRR, 1.25 (0.90–1.74) |
| Hepatobiliary | 4 (0.0) | 2 (0.0) | IRR, 4.23 (0.77–23.08) |
| Infections | 86 (0.2) | 16 (0.4) | IRR, 1.57 (0.92–2.68) |
| Injury | 34 (0.1) | 4 (0.1) | IRR, 0.99 (0.35–2.80) |
| Neoplasms | 262 (0.7) | 42 (1.0) | IRR, 1.36 (0.98–1.88) |
| Nervous | 39 (0.1) | 3 (0.1) | IRR, 0.65 (0.20–2.10) |
| Respiratory | 348 (1.0) | 19 (0.5) | IRR, 0.46 (0.29–0.73) |
| Vascular | 10 (0.0) | 3 (0.1) | IRR, 2.54 (0.70–9.21) |
| Patients with MACE | |||
| Fatal MACE, n (%)* | 295 (1.9) | 37 (2.0) | HR = 1.06 (0.75–1.49) |
| MACE, n (%)‡ | 554 (3.6) | 92 (5.1) | HR = 1.45 (1.16–1.81) |
| Stroke | 134 (0.9) | 30 (1.7) | HR = 1.94 (1.31–2.89) |
| TIA | 58 (0.4) | 17 (0.9) | HR = 2.55 (1.49–4.38) |
| MI | 172 (1.1) | 22 (1.2) | HR = 1.11 (0.71–1.72) |
Subgroup analyses indicated that patients with GOLD stage II and patients considered overweight (BMI = 25 to <30 kg/m2) with FEV1/FVC less than LLN had a significantly reduced risk of all-cause mortality compared with patients with FEV1/FVC greater than or equal to LLN (HR = 0.61, 95% CI = 0.49–0.76; and HR = 0.69; 95% CI = 0.51–0.94, respectively). Mortality risk did not differ in any other subgroup, and no significant interaction with respect to increased mortality risk was observed in any subgroup when analyzed by treatment and LLN classification.
Patients with FEV1/FVC greater than or equal to LLN had, statistically, a significantly greater risk of MACE compared with patients with FEV1/FVC less than LLN (HR = 1.45; 95% CI = 1.16–1.81; P = 0.0010). The risk of (time to) MACE in patients with FEV1/FVC less than LLN versus patients with FEV1/FVC greater than or equal to LLN is shown in Figure 2C. The results did not change when HRs were adjusted for demographics and baseline characteristics (Table E3). For individual MACE components, a statistically significantly greater risk of stroke and TIA was observed among patients with FEV1/FVC greater than or equal to LLN (Table 3).
The classification of patients by the LLN criterion did not change the outcomes of the study by treatment arm, with no difference between patients on 5 μg tiotropium Respimat and 18 μg HandiHaler. Indeed, the treatment differences (Respimat vs. HandiHaler) in patients with FEV1/FVC greater than or equal to LLN were similar to those in the FEV1/FVC less than LLN group, and to those from the overall study (Table E2) (18).
When comparing patients with FEV1/FVC greater than or equal to LLN to patients with FEV1/FVC less than LLN, data showed that those with FEV1/FVC greater than or equal to LLN had significantly lower risks of moderate to severe COPD exacerbation (adjusted mean rate of events [per patient-year] 0.40 vs. 0.60 [Table E4]; rate ratio = 0.68; 95% CI = 0.62–0.73; P < 0.0001) and severe exacerbation (adjusted mean rate of events [per patient-year] 0.06 vs. 0.13 [Table E4]; rate ratio = 0.50; 95% CI = 0.42–0.59; P < 0.0001). The risks of (time to) moderate to severe COPD exacerbation and severe exacerbation are presented in Figures 2D and 2E, respectively. Adjusted HRs and rate ratios showed similar findings (Tables E3 and E4).
Risks of (time to) first moderate to severe COPD exacerbation and of first severe COPD exacerbation were similar for each tiotropium treatment arm within both LLN classifications (Table E2).
The incidence of AEs and serious AEs was similar between patients with FEV1/FVC less than LLN (65.9% and 33.0%, respectively) and patients with FEV1/FVC greater than or equal to LLN (62.4% and 31.4%, respectively) (Table E5).
In this post hoc analysis, changing the definition of airflow obstruction from a fixed FEV1/FVC threshold of 0.70 to a definition based on the LLN of the FEV1/FVC would have excluded 1,807 study participants. This discordant group, in which there was no agreement between the two definitions of airflow obstruction, was at greater risk of nonfatal MACE and experienced fewer exacerbations, but this exclusion did not change the outcome of the trial.
Even in this population of clinically diagnosed patients with COPD recruited to have a post-bronchodilator FEV1 less than or equal to 70% predicted, 10.6% did not meet the LLN criteria for obstruction. This discordant group was smaller than groups seen in reports derived from population-based studies (7, 22, 23), in which the severity of baseline FEV1 impairment was less. We saw no significant differences between the different lung function equations that we used to generate the LLN data, possibly reflecting similarities in lung function across regions in a population defined as having symptomatic COPD.
The GLI equations are widely available and have been used in previous studies (15, 24), so we adopted them for our primary comparison. Discordant patients were some 4 years older, on average, than those meeting both definitions, but were younger than people studied previously (11, 23). The BMI of discordant patients was higher and their FVC, however expressed, was lower, in part due to the higher incidence of obesity. The relationship between obesity and lung function is complex (25), but an increased closing volume is common in this setting, and leads to greater loss of FVC and a relatively higher FEV1/FVC ratio than might otherwise be the case. The ratio of our discordant patients was close to the threshold value of 0.7, and, in many cases, lay close to the threshold value for LLN. Thus, small differences in test performance or between-day variations in lung function could have influenced whether a diagnosis of airflow obstruction was made.
At baseline, the discordant patients reported symptoms, such as cough productive of sputum and breathlessness, as often as those for whom there was diagnostic agreement (Table 1). Although not directly comparable, our patients were more symptomatic than those discordant patients identified in older population-based studies (22). Misdiagnosis might explain this symptomatology, and there was a small excess of cerebrovascular and cardiac diagnoses in the discordant group, but, overall, the difference from patients with definite COPD was surprisingly small. Both the reported history of prior exacerbations and the use of respiratory medication before study entry were similar between groups and higher than that in other discordant populations.
The observed clinical outcomes showed both similarities to and differences from the overall TIOSPIR population. The all-cause mortality did not differ between the groups, despite the better baseline lung function of the discordant patients. A previous study by Mannino and Diaz-Guzman (9) showed that, compared with people with normal lung function, mortality was increased in the obstructed and restricted subjects classified as normal using the LLN. Likewise, another study, by Bhatt and colleagues (26), showed that a fixed-ratio group had more exacerbations than smoking control subjects, and that the use of LLN would fail to identify a number of patients with significant pathology and respiratory morbidity. As has been noted elsewhere, the rate of cardiac events was higher and the rate of exacerbations lower among discordant patients (12, 13), although this does not seem to be true in the elderly population (11, 22). Whether this difference in clinical outcomes relates to the diagnosis of COPD in these patients cannot be addressed by our data. Clearly, the study entry criteria have an important influence on the subsequent outcomes. However, irrespective of LLN classification, patients who received Respimat or HandiHaler had similar risks of death, fatal MACE, MACE, and exacerbation within each category, and their inclusion in this trial did not alter the conclusions of TIOSPIR.
Our analysis has both strengths and limitations. We believe that this is the largest clinical trial to date in which the effects of diagnostic reclassification based on the LLN have been assessed. Follow-up for mortality outcomes was complete, and the causes of death were independently adjudicated. Study entry was based on post-bronchodilator spirometry, unlike many of the studies in which the impact of the LLN has been assessed. It is recognized that in borderline cases, such as those described here, use of a bronchodilator will change the classification based on a fixed ratio and LLN threshold (27), although we do not believe that this would be a significant issue in our patients whose FEV1 was less than or equal to 70% predicted.
The cut-off of FEV1 less than or equal to 70% predicted most likely increased specificity; hence, we may have excluded a number of patients who might be positive by the fixed criterion with FEV1 % predicted between 70% and 80%. However, although the fixed-ratio approach may work in conjunction with a decreased FEV1, it might be less useful in patients with COPD with a post-bronchodilator FEV1 greater than 80%, exhibiting normal lung function. In our study, only a small number of patients (1.6%) fell within this range, and only 0.4% had FEV1 greater than or equal to 80% predicted (Table 1). We also acknowledge that, by the nature of our inclusion criteria, we had no LLN-positive, but fixed-negative participants in the study.
The majority of the spirometric measurements were obtained in routine laboratories, and this makes our results generalizable beyond the rather specific field of clinical trials. Crucially, study entry depended on the clinician’s clinical diagnosis of COPD in a population with a known risk factor (tobacco smoking), which was supported by spirometry rather than relying on the identification of airflow obstruction in a representative population sample.
We did not exclude patients on the basis of cardiac disease, which not only increases the generalizability of the data, but also helps to explain the similarity of the concordant and discordant groups in this respect. We focused on the GLI (white) equation to define LLN; however, baseline characteristics using other equations (GLI [region] and ECSC) showed only slight differences between criteria (Table E1). Differences between this and the other equations are known to be small (6). Adjustment for ethnicity affects FEV1 and FVC, but FEV1/FVC ratio is virtually independent of race (19). Our study was generally representative of the patients recruited into clinical trials, but we cannot extrapolate these findings to much older people with milder disease, who would more closely correspond to those identified from population studies of the prevalence of airflow obstruction, nor can we establish whether younger patients not meeting the fixed-ratio criterion would benefit from long-acting bronchodilator treatment.
In conclusion, this post hoc analysis of the TIOSPIR study provides reassuring data showing that the great majority of patients studied met the diagnostic criteria for COPD no matter how specified. Many of those who did not meet the LLN criterion might have done so had they been retested, given the known between-test variation of spirometry. Potentially, the most important difference was the presence of respiratory symptoms as a prerequisite for the diagnosis, although restricting recruitment to patients with FEV1 less than or equal to 70% predicted may also have played a role. Even though exacerbations were less frequent in the discordant group, they were not absent and were likely to contribute to impaired health status (28).
There is renewed interest in symptomatic individuals at risk of COPD who demonstrate structural abnormalities in their lungs before airflow obstruction is seen, whatever the diagnostic thresholds selected (29). Future studies will need to explore how defining groups by symptoms and risk factors relates to outcomes in discordant patients of any age. For now, we are reassured that, in trials like ours, outcomes are unlikely to have been materially affected by the diagnostic recruitment strategy used.
The authors thank the other Tiotropium Safety and Performance in Respimat Publication Steering Committee members (Professors Antonio Anzueto, Ronald Dahl, and Daniel Dusser) for their input into the analyses. Writing and editorial support was provided by Sarah J. Petit, Yogeeta Surinder Kumar, and Carol Richter of PAREXEL, and was funded by Boehringer Ingelheim.
| 1 . | Global Initiative for Chronic Obstructive Lung Disease. Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease. GOLD Web site; 2016 Dec [accessed 2016 Dec 17]. Available from: http://goldcopd.org/gold-2017-global-strategy-diagnosis-management-prevention-copd/. |
| 2 . | O’Donnell DE, Hernandez P, Kaplan A, Aaron S, Bourbeau J, Marciniuk D, et al. Canadian Thoracic Society recommendations for management of chronic obstructive pulmonary disease—2008 —highlights for primary care. Can Respir J 2008;15(suppl A):1A–8A. |
| 3 . | Guidelines and Protocols Advisory Committee. Chronic obstructive pulmonary disease (COPD). British Columbia Ministry of Health Service; 2011 Jan [accessed 2017 Dec 6]. Available from: https://www2.gov.bc.ca/gov/content/health/practitioner-professional-resources/bc-guidelines/copd. |
| 4 . | Qaseem A, Wilt TJ, Weinberger SE, Hanania NA, Criner G, van der Molen T, et al.; American College of Physicians; American College of Chest Physicians; American Thoracic Society; European Respiratory Society. Diagnosis and management of stable chronic obstructive pulmonary disease: a clinical practice guideline update from the American College of Physicians, American College of Chest Physicians, American Thoracic Society, and European Respiratory Society. Ann Intern Med 2011;155:179–191. |
| 5 . | Cerveri I, Corsico AG, Accordini S, Niniano R, Ansaldo E, Antó JM, et al. Underestimation of airflow obstruction among young adults using FEV1/FVC <70% as a fixed cut-off: a longitudinal evaluation of clinical and functional outcomes. Thorax 2008;63:1040–1045. |
| 6 . | Quanjer PH, Brazzale DJ, Boros PW, Pretto JJ. Implications of adopting the Global Lungs Initiative 2012 all-age reference equations for spirometry. Eur Respir J 2013;42:1046–1054. |
| 7 . | Swanney MP, Ruppel G, Enright PL, Pedersen OF, Crapo RO, Miller MR, et al. Using the lower limit of normal for the FEV1/FVC ratio reduces the misclassification of airway obstruction. Thorax 2008;63:1046–1051. |
| 8 . | Miller MR, Levy ML. Chronic obstructive pulmonary disease: missed diagnosis versus misdiagnosis. BMJ 2015;351:h3021. |
| 9 . | Mannino DM, Diaz-Guzman E. Interpreting lung function data using 80% predicted and fixed thresholds identifies patients at increased risk of mortality. Chest 2012;141:73–80. |
| 10 . | Murphy DE, Panos RJ. Diagnosis of COPD and clinical course in patients with unrecognized airflow limitation. Int J Chron Obstruct Pulmon Dis 2013;8:199–208. |
| 11 . | Turkeshi E, Vaes B, Andreeva E, Matheï C, Adriaensen W, Van Pottelbergh G, et al. Airflow limitation by the Global Lungs Initiative equations in a cohort of very old adults. Eur Respir J 2015;46:123–132. |
| 12 . | van Dijk W, Tan W, Li P, Guo B, Li S, Benedetti A, et al.; CanCOLD Study Group. Clinical relevance of fixed ratio vs lower limit of normal of FEV1/FVC in COPD: patient-reported outcomes from the CanCOLD cohort. Ann Fam Med 2015;13:41–48. |
| 13 . | Wollmer P, Engström G. Fixed ratio or lower limit of normal as cut-off value for FEV1/VC: an outcome study. Respir Med 2013;107:1460–1462. |
| 14 . | Celli BR, Halbert RJ, Isonaka S, Schau B. Population impact of different definitions of airway obstruction. Eur Respir J 2003;22:268–273. |
| 15 . | Pellegrino R, Viegi G, Brusasco V, Crapo RO, Burgos F, Casaburi R, et al. Interpretative strategies for lung function tests. Eur Respir J 2005;26:948–968. |
| 16 . | Quanjer PH, Ruppel G, Brusasco V, Pérez-Padilla R, Fragoso CA, Culver BH, et al. COPD (confusion over proper diagnosis) in the zone of maximum uncertainty. Eur Respir J 2015;46:1523–1524. |
| 17 . | Wise RA, Anzueto A, Calverley P, Dahl R, Dusser D, Pledger G, et al. The Tiotropium Safety and Performance in Respimat trial (TIOSPIR), a large scale, randomized, controlled, parallel-group trial-design and rationale. Respir Res 2013;14:40. |
| 18 . | Wise RA, Anzueto A, Cotton D, Dahl R, Devins T, Disse B, et al.; TIOSPIR Investigators. Tiotropium Respimat inhaler and the risk of death in COPD. N Engl J Med 2013;369:1491–1501. |
| 19 . | Quanjer PH, Stanojevic S, Cole TJ, Baur X, Hall GL, Culver BH, et al.; ERS Global Lung Function Initiative. Multi-ethnic reference values for spirometry for the 3–95-yr age range: the global lung function 2012 equations. Eur Respir J 2012;40:1324–1343. |
| 20 . | Anzueto A, Wise R, Calverley P, Dusser D, Tang W, Metzdorf N, et al. The Tiotropium Safety and Performance in Respimat® (TIOSPIR®) trial: spirometry outcomes. Respir Res 2015;16:107. |
| 21 . | Quanjer PH, Tammeling GJ, Cotes JE, Pedersen OF, Peslin R, Yernault JC. Lung volumes and forced ventilatory flows. Eur Respir J 1993;6:5–40. |
| 22 . | Vaz Fragoso CA, Concato J, McAvay G, Van Ness PH, Rochester CL, Yaggi HK, et al. The ratio of FEV1 to FVC as a basis for establishing chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2010;181:446–451. |
| 23 . | Vaz Fragoso CA, Concato J, McAvay G, Van Ness PH, Gill TM. Respiratory impairment and COPD hospitalisation in older persons: a competing risk analysis. Eur Respir J 2012;40:37–44. |
| 24 . | Vaz Fragoso CA, McAvay G, Van Ness PH, Casaburi R, Jensen RL, MacIntyre N, et al. Phenotype of normal spirometry in an aging population. Am J Respir Crit Care Med 2015;192:817–825. |
| 25 . | O’Donnell DE, Ciavaglia CE, Neder JA. When obesity and chronic obstructive pulmonary disease collide: physiological and clinical consequences. Ann Am Thorac Soc 2014;11:635–644. |
| 26 . | Bhatt SP, Sieren JC, Dransfield MT, Washko GR, Newell JD Jr, Stinson DS, et al.; COPDGene Investigators. Comparison of spirometric thresholds in diagnosing smoking-related airflow obstruction. Thorax 2014;69:409–414. |
| 27 . | Pérez-Padilla R, Hallal PC, Vázquez-García JC, Muiño A, Máquez M, López MV, et al.; PLATINO group. Impact of bronchodilator use on the prevalence of COPD in population-based samples. COPD 2007;4:113–120. |
| 28 . | Labonté LE, Tan WC, Li PZ, Mancino P, Aaron SD, Benedetti A, et al.; Canadian Respiratory Research Network; CanCOLD Collaborative Research Group. Undiagnosed COPD contributes to the burden of health care utilization: data from the CanCOLD Study. Am J Respir Crit Care Med 2016;194:285–298. |
| 29 . | Woodruff PG, Barr RG, Bleecker E, Christenson SA, Couper D, Curtis JL, et al.; SPIROMICS Research Group. Clinical significance of symptoms in smokers with preserved pulmonary function. N Engl J Med 2016;374:1811–1821. |
Supported by Boehringer Ingelheim, which funded the Tiotropium Safety and Performance in Respimat study; Boehringer Ingelheim participated in the design of the study, the collection and analysis of the data, and the preparation of the manuscript.
Author Contributions: P.M.A.C., A.M., A.F., N.M., and R.A.W. had full access to the data, and contributed to the conception and design of this study and to the analysis and interpretation of the data; they were involved in drafting the manuscript or revising it critically for important intellectual content, have read and approved the final manuscript, and assume full responsibility for the integrity of the submission as a whole.
This article has an online supplement, which is accessible from this issue’s table of contents at www.atsjournals.org.
Author disclosures are available with the text of this article at www.atsjournals.org.
