The process of screening asymptomatic adults for chronic obstructive pulmonary disease (COPD) can either involve a targeted screening approach (with questionnaires [key question 2]) or no risk stratification whereby asymptomatic unselected adults go directly to pulmonary function screening tests (key question 3). The dashed line indicates an established association between an intermediate outcome and a health outcome.
aUsing prescreening questionnaires.
Diagnostic criterion for chronic obstructive pulmonary disease is a postbronchodilator ratio of forced expiratory volume in 1 second (FEV1) to forced vital capacity (FVC) of less than 0.70. KQ indicates key question.
aDetails about reasons for exclusion are as follows. Relevance: study aim not relevant. Setting: study was not conducted in a country relevant to US practice. Population: study was not conducted in asymptomatic adults 40 years and older. Quality: study did not meet criteria for fair or good quality (ie, it was poor quality). Study design: study did not use an included design. Intervention: study used an excluded intervention or screening approach. Outcomes: study did not have relevant outcomes or had incomplete outcomes. Non-English: study was published in a non-English language.
eMethods. Systematic review literature search strategies
eTable 1. Quality assessment criteria
eTable 2. Scoring details for externally validated prescreening questionnaires
Guirguis-Blake JM, Senger CA, Webber EM, Mularski RA, Whitlock EP. Screening for Chronic Obstructive Pulmonary DiseaseEvidence Report and Systematic Review for the US Preventive Services Task Force. JAMA. 2016;315(13):1378-1393. doi:10.1001/jama.2016.2654
Chronic obstructive pulmonary disease (COPD) is the third leading cause of death in the United States.
To systematically review literature on the accuracy of screening questionnaires and office-based screening pulmonary function testing and the efficacy and harms of treatment of screen-detected COPD.
MEDLINE, PubMed, and the Cochrane Central Register of Controlled Trials for relevant English-language studies published through January 2015.
Two reviewers independently screened abstracts and studies. The search yielded 13 141 unique citations; 465 full-text articles were reviewed, and 33 studies met the inclusion criteria.
Data Extraction and Synthesis
Two reviewers rated the quality of each study using USPSTF criteria.
Main Outcomes and Measures
Diagnostic accuracy (sensitivity, specificity, positive predictive value [PPV], and negative predictive value [NPV]; treatment efficacy (COPD exacerbations, all-cause mortality, quality of life, and dyspnea); and treatment harms.
All screening questionnaires were based on symptoms as well as risk factors such as age and smoking history. The COPD Diagnostic Questionnaire was the most extensively studied (5 studies, n = 3048), with moderate overall performance for COPD detection: area under the receiver operating characteristic curve (AUC), 0.65 to 0.72; sensitivity, 80% to 93%; and specificity, 24% to 49%, at a threshold of greater than 16.5. Positive predictive value and NPV ranged from 17% to 45% and 76% to 98%, respectively. For pulmonary function–based screening tools, FEV1/FEV6 was the best studied (3 studies, n = 1587), with AUC ranging from 0.84 to 0.85. Sensitivity ranged from 51% to 80%. Specificity (range, 90%-95%) and PPV (range, 63%-75%) appeared better than questionnaires. There was not strong evidence to support that screening and supplying smokers with spirometry results improves smoking cessation rates. Treatment trials were unavailable for screen-detected patients. Trials that reported outcomes in patients with mild to moderate COPD included 2 trials of long-acting β-agonists (LABAs) (n = 3174), 1 RCT of LABAs and inhaled corticosteroids (ICS) (n = 1097), 5 RCTs of the long-acting muscarinic antagonist tiotropium (n = 4592), and 6 RCTs of ICS (n = 3983). They suggested no benefit in all-cause mortality, but a decrease in annual rates of exacerbations with pharmacologic treatments. Few trials reported harms for any individual drug class. Adverse effects were generally mild (eg, dry mouth and cough).
Conclusions and Relevance
There was no direct evidence available to determine the benefits and harms of screening asymptomatic adults for COPD using questionnaires or office-based screening pulmonary function testing or to determine the benefits of treatment in screen-detected populations. Indirect evidence suggests that the COPD Diagnostic Questionnaire has moderate overall performance for COPD detection. Among patients with mild to moderate COPD, the benefit of pharmacotherapy for reducing exacerbations was modest.
Quiz Ref IDChronic lower respiratory disease, composed chiefly of chronic obstructive pulmonary disease (COPD), was the third leading cause of death in the United States in 2013.1,2 Cigarette smoke exposure, either directly or indirectly, has been highly correlated with the development of COPD and COPD mortality.3- 7 In theory, primary care physicians can identify undetected COPD by screening relatively unselected, asymptomatic individuals or by targeting a high-risk asymptomatic population using screening spirometry, followed by confirmatory diagnostic spirometry in primary care or pulmonary specialty clinics.7,8 Current clinical practice guidelines recommend against screening for COPD in asymptomatic patients; however, many professional organizations recommend case-finding among patients presenting with respiratory symptoms associated with the disease, such as dyspnea, chronic cough, or sputum production.7- 9
Quiz Ref IDIn 2008, the US Preventive Services Task Force (USPSTF) recommended against screening asymptomatic adults for COPD using spirometry (grade D).10 The USPSTF concluded that this method had no net benefit and had large associated opportunity costs. The aim of this systematic review is to update the evidence on the benefits and harms of screening for COPD using questionnaires and spirometry, including the diagnostic accuracy of primary care–feasible screening instruments; the effect of spirometric screening on uptake of targeted preventive services; and the effectiveness, benefits, and harms of treating screen-detected patients (generally those with mild to moderate COPD) since the last recommendation.
To conduct this review, an analytic framework was developed with 8 key questions (KQs) (Figure 1) that examined the effect of screening asymptomatic adults 40 years and older for COPD on health outcomes (KQ1); the accuracy and harms of screening questionnaires and pulmonary function tests (KQs 2-4); the effectiveness and harms of COPD screening on the uptake of targeted preventive services (KQs 5 and 6); and the effectiveness and harms of treatment of asymptomatic mild to moderate COPD (KQs 7 and 8). Detailed methods and results are available in the full evidence report (http://www.uspreventiveservicestaskforce.org/Page/Document/final-evidence-review143/chronic-obstructive-pulmonary-disease-screening).11 The analytic framework, review questions, and methods for locating and qualifying evidence reflect public input after posting on the USPSTF website.
Searches included MEDLINE, PubMed, the Cumulative Index to Nursing and Allied Health Literature, and the Cochrane Central Register of Controlled Trials from January 2000 to January 2015, supplemented by checking reference lists from relevant systematic reviews. Evaluating the effect of a COPD diagnosis on pneumococcal and influenza immunization rates was a new question to this review; therefore, databases were searched from inception through January 2015. Since January 2015, we have continued to conduct ongoing surveillance through article alerts and targeted searches of high-impact journals to identify major studies published in the interim that may affect the conclusions or understanding of the evidence and therefore the related USPSTF recommendation. The last surveillance was conducted on January 22, 2016, and identified no new studies.
Two reviewers independently reviewed 13 141 unique citations and 465 full-text articles against a priori inclusion criteria (Figure 2 and eMethods in the Supplement). For KQs 1 through 6, we initially considered studies including asymptomatic adults 40 years and older (limited to current smokers for KQ5a). For questions 7 and 8, we restricted the population further to include only asymptomatic adults 40 years and older who were also diagnosed with mild COPD (forced expiratory volume in 1 second [FEV1] ≥80% normal) to moderate COPD (FEV1 50%-79% normal) or a mean population FEV1 greater than or equal to 60% predicted to approximate a population of mild to moderate COPD. Asymptomatic patients were defined as those in 1 of the following states: free of the disease; the disease is present, but the patient has physical symptoms that are undetected by the patient or the clinician; or the patient has nonspecific symptoms that have gone unrecognized as being related to COPD. For KQs 2 and 4, we analyzed COPD prescreening questionnaires feasible in primary care with published studies describing their original development, internal validation, and external validation; results are reported only for COPD screening questionnaires with external validation, which is the minimal requirement for consideration in clinical practice.12,13 For KQ2, the initial search was for risk factor–only based screening questionnaires, which would capture an asymptomatic population. However, because none were identified, risk factor– and symptom-based prescreening questionnaires were included. For KQ3, we examined primary care–feasible screening pulmonary function tests (eg, handheld devices or prebronchodilator testing requiring minimal personnel training).
For the treatment questions, the search included treatment efficacy literature for the following COPD drug classes or combinations of any of the following: long-acting β-agonists (LABAs), long-acting anticholinergics, and inhaled corticosteroids (ICS). Because there were no trials in screen-detected or asymptomatic populations, the included population was expanded to those diagnosed with mild to moderate disease because observational studies show that 84% to 95% of screen-detected patients are expected to have mild to moderate COPD.14- 17
For KQs 1, 5, and 7, the study design was limited to randomized clinical trials (RCTs). For KQs 2 and 3, designs were limited to diagnostic accuracy studies (including cross-sectional and cohort studies) with a reference standard COPD definition of a postbronchodilator ratio of FEV1 to forced vital capacity (FVC) of less than 0.70.7 For KQs 4 and 6, RCTs, large screening registry or database observational studies, and cohort studies were considered. When evaluating harms associated with the treatment of COPD (KQ8), the data were limited to what was reported in the efficacy trials included in KQ7, large screening registries, and systematic reviews, supplemented with information reported by the US Food and Drug Administration.
One reviewer extracted study-level data into standardized evidence tables; a second checked for accuracy. Articles meeting inclusion criteria were critically appraised by 2 independent reviewers using predefined criteria18- 21 with disagreements resolved by a third investigator. Included studies were limited to those published in English that were rated as good or fair quality using USPSTF quality rating standards.18 (Details are available in eTables 1 and 2 in the Supplement.)
Data from the included studies were qualitatively examined to identify a range of results. Given the clinical heterogeneity of studies, meta-analyses were not conducted for any of the questions in this review.
For studies of diagnostic accuracy, 2 × 2 tables were constructed from data reported in the primary studies. When 95% CIs were not reported for diagnostic accuracy estimates, these intervals were calculated using Jeffrey confidence intervals (Stata version 13.1). For diagnostic accuracy studies, in addition to the standard test performance characteristics (area under the receiver operating characteristic [ROC] curve, sensitivity, specificity, positive predictive value [PPV], negative predictive value [NPV]), we calculated the following outcomes: COPD prevalence in the population, percentage of patients screening positive, false-positive rate, and the percentage of missed cases.
Thirty-three studies (48 articles) met the inclusion criteria for this systematic review (Figure 2). This article provides a summary of results that supported the USPSTF recommendation process.
KQ1: Does screening asymptomatic adults 40 years and older for COPD with prebronchodilator screening spirometry improve health-related quality of life or reduce morbidity or mortality?
There was no direct evidence comparing the effectiveness of COPD screening and no screening on patient health outcomes.
KQ2: Can high-risk asymptomatic adults who are more likely to test positive on screening for COPD be reliably identified using prescreening questionnaires?
KQ4: What are the adverse effects of screening for COPD using prescreening questionnaires?
No relevant studies of COPD screening questionnaires in asymptomatic populations were identified that were based solely on risk factors. Three externally validated prescreening questionnaires were identified that assessed risk factors and respiratory symptoms to select high-risk patients for screening spirometry: the COPD Diagnostic Questionnaire (CDQ), the Lung Function Questionnaire (LFQ), and the COPD Population Screener (COPD-PS). The predictive accuracy of these questionnaires in all included studies was measured against the postbronchodilator FEV1/FVC reference standard according to the Global Initiative for Chronic Obstructive Lung Disease COPD definition and American Thoracic Society and European Respiratory Society quality standards.7,22,23
The CDQ is an externally validated, 8-item, self-administered, symptom- and risk factor–based COPD prescreening questionnaire used to select high-risk patients for screening spirometry, which assigns scores for established risk factors including age, pack-years of smoking, and body mass index (BMI) as well as symptoms and allergy history (Table 1).14,17,24- 28 Possible scores range from 0 to 38 (higher scores associated with higher COPD risk), with highest scores attributed to older age (score 10 for ≥70 years), greater pack-years (score 7 for ≥50 pack-years), and lower BMI (score 5 for BMI). Two cut points (16.5 and 19.5) have been proposed to select patients for screening spirometry based on ROC curves from the original development study.30 The original development and internal validation was performed in a primary care–based US and UK cross-sectional study of 818 past and current smokers 40 years and older.31
The CDQ has been externally validated in 5 fair- to good-quality diagnostic accuracy studies mainly focusing on primary care European and Australian populations.14,17,24- 26 The study populations varied; 3 studies recruited solely current or ever-smokers from primary care, the general population, or both,17,24,26 and 2 studies recruited patients from primary care clinics without regard to smoking history.14,25 Chronic obstructive pulmonary disease was diagnosed by spirometry in 10.3% to 41.1% of participants in each of the 4 studies that reported this outcome, with the highest prevalence (41.1%) being reported in a study that required participants to be current smokers with at least a 10 pack-year history and have at least 1 respiratory symptom; these participants were essentially prescreened, thereby selecting for those most likely to have COPD.26 Prevalence of COPD in the studies recruiting ever-smokers ranged from 13.1% to 27.9%,17,24 and 1 general population study with more than half nonsmoking participants had an overall COPD prevalence of 10.3%, which was higher (17.2%) among ever-smokers.14 The majority of patients found to have COPD were identified as having mild or moderate disease (83.8%-94.7%).
Most external validation studies reported that a CDQ score greater than 16.5 had a sensitivity ranging from 80% to 91% and specificity ranging from 24% to 49% for identifying those who test positive using spirometric confirmation for COPD (Table 1). Choosing a higher cut point (19.5) reduced sensitivity and NPV but increased specificity and PPV. The proportion of cases missed by the CDQ (false-negative rate) varied widely, from 9.0% to 37.0%, and was lowest when using the most sensitive screening threshold (see full evidence report11). For the threshold of less than 16.5 for screen negatives, and limiting to studies in which fewer than 20% of spirometry tests were invalid or incomplete, the proportion of missed spirometry-diagnosed COPD was around 10%. In these studies, increasing the screening threshold to less than 19.5 increased the missed COPD cases to 27.9% to 34.2% in best estimates.
Simple tables were constructed to compare screening test performance using the CDQ across a range of populations, using the mean sensitivity and specificity of applicable studies. Table 2 shows the trade-offs with missed cases and false-positive tests in populations with varying COPD prevalence at the 2 cut points (16.5 and 19.5).
The LFQ and COPD-PS are both 5-item self-administered risk factor– and symptom-based questionnaires (Table 1). The LFQ assigns scores to age; smoking history (pack-years, never/current/former smoker); and presence of wheezing, dyspnea, and mucous productive cough. Possible scores range from 5 to 25, with lower scores associated with higher COPD risk. A threshold of 18 or less has been proposed as a cut point for COPD risk warranting pulmonary function diagnostic workup. Although the LFQ was specifically developed in the National Health and Nutrition Examination Survey population with chronic bronchitis and studied in US primary care practices,32,33 data for this questionnaire were limited to a single validation study.27 This external validation study, however, had quality concerns (31% of spirometry was invalid or incomplete) and relatively poorer test performance than the CDQ (lower sensitivity, specificity, PPV, and NPV) when used in similar populations. In addition, we could not assess the harms of screening (ie, rate of false positives or proportion of missed cases) using the LFQ because only a subset of screen-negative patients were selected for spirometry.
The COPD-PS assigns scores for age, smoking history, dyspnea, sputum production, and dyspnea-related functional limitations. Possible scores range from 0 to 10, with higher scores being associated with a higher risk of COPD. A threshold of 5 to 6 or more has been proposed as a cut point for COPD risk warranting pulmonary function test workup. Although the COPD-PS was derived in an enriched sample of US pulmonary and primary care clinics,34 its external validation in a single Japanese population-based study makes conclusions regarding generalizability of accuracy results limited.28 The COPD-PS has recently been applied in a multisite, US-based primary care, pragmatic COPD screening trial (n = 8770); however, this trial did not include the reference standard of spirometry for accuracy estimation.35
KQ3: What is the test performance of screening pulmonary function tests in predicting diagnosis of COPD based on confirmation using postbronchodilator spirometry to identify fixed airflow obstruction in asymptomatic adults?
KQ4: What are the adverse effects of screening for COPD using screening pulmonary function tests?
One good-quality and 4 fair-quality diagnostic accuracy studies were identified that evaluated 2 different handheld pulmonary function screening tests against a postbronchodilator FEV1/FVC reference standard: FEV1/FEV614,17,36 (delivered either before or after the bronchodilator) (Table 3) and peak expiratory flow (PEF).37,38 The included populations varied in their selectivity in terms of age, smoking status, symptomatology, and exclusion of preexisting COPD.
Three studies (1 good-quality and 2 fair-quality) reported the screening test performance of FEV1/FEV6 (n = 1587).14,17,36 Two studies assessing prebronchodilator FEV1/FEV6 among ever-smokers found similar sensitivities (51% and 53%) and specificities (90% and 93%) (Table 3).17,36 The third study14 used postbronchodilator FEV1/FEV6 for screening; sensitivity was much higher (80%) than the 2 prebronchodilator studies,17,36 and specificity was as good or better (95%).14 However, the sample was enriched with current smokers, which would increase the predictive value. In a subsample limited to ever-smokers, postbronchodilator screening appeared similar to screening test performance in the entire population, suggesting that postbronchodilator FEV1/FEV6 performs better than prebronchodilator testing.
Harms of screening pulmonary function testing included false positives and false negatives (missed cases). False-negative rates (proportion of total diagnoses missed) ranged from 14% to 49%, depending on the threshold used (see full evidence report11). The FEV1/FEV6 index test threshold of less than 0.70 showed the lowest rate of false negatives (19.8%) seen after postbronchodilator index testing.14 Using a prebronchodilator cutoff of FEV1/FEV6 less than 0.70, the missed cases in 2 of the trials approached 50% (see full evidence report11).17,36 False-positive rates for the less than 0.70 threshold ranged from 5.2% to 10.5%,14,17,36 with the lowest rate seen in the single study using postbronchodilator testing.14
Mean sensitivity and specificity were used to construct simple tables to compare screening test performance using the prebronchodilator and postbronchodilator FEV1/FEV6 across a range of populations. Table 2 shows the trade-offs with missed cases and false-positive tests in populations with varying COPD prevalence.
Two fair-quality studies of PEF evaluated the largest number of patients (n = 23 098).37,38 However, these were based on large population-based studies whose primary aims were to describe the prevalence of COPD internationally. Because the studies did not exclude persons with preexisting COPD and also included several developing countries with environmental exposures that would be not be considered generalizable to the United States, their results are less applicable to screening.
KQ5: Does identifying asymptomatic adults with fixed airflow obstruction through screening improve the delivery and uptake of targeted preventive services?
KQ5a: Does screening for COPD increase smoking cessation rates among asymptomatic adults compared with usual care?
KQ5b: Does screening for COPD increase relevant immunization rates among asymptomatic adults compared with usual care?
KQ6: What are the adverse effects of COPD screening, including the effect of targeted preventive services in this population?
Five fair-quality studies (n = 1694) were identified that addressed the incremental value of adding spirometry to existing smoking cessation counseling interventions to improve smoking cessation rates (Table 4).39- 43 There was not strong evidence to support the premise that supplying smokers with spirometry results improves smoking cessation rates. No trials randomized patients without known COPD diagnoses to screening spirometry vs no spirometry in order to estimate the independent value of screening spirometry. Instead, in all studies control groups received almost the same smoking cessation support as the spirometry group; studies varied in whether the control group received spirometry testing39,41 or not40,42,43 and in whether smoking cessation support was tailored based on spirometry or other medical examination findings. Of the 5 included RCTs, a single fair-quality trial giving patients their “lung age” reported a statistically significant difference in biochemically validated abstinence in the intervention group compared with the control group rates after 12 months of follow-up (n = 561; 13.6% vs 6.4%; validated quit rate difference, 7.2% [95% CI, 2.2%-12.1%]; number needed to treat [NNT], 14); however, both the control and treatment groups underwent spirometry, so the trial actually tested the method of communicating spirometry results rather than the value of the screening spirometry itself.41 Most of the other trials either reported higher abstinence rates that were not statistically significant or no difference in the intervention group compared with control (Table 4).
There was little evidence examining the potential negative effect of COPD screening on targeted preventive services; little can be concluded from the included qualitative study reporting that some smokers felt that confrontation with spirometry would interfere with personal choice.39
No trials were identified reporting the effect of COPD screening on recommended immunization uptake rates.
KQ7: Does treatment for asymptomatic adults identified with mild to moderate COPD through screening improve health-related quality of life or reduce morbidity or mortality?
KQ8: What are the adverse effects of COPD treatments in this population?
Twenty studies of 14 distinct trials were identified for the 3 included drug classes and 1 combination treatment: 2 trials of LABAs,44,45 1 RCT of LABAs plus ICS,45 5 RCTs of the long-acting muscarinic antagonist (LAMA) tiotropium,44,46- 48 and 6 RCTs of ICS45,49- 53 (Table 5).54,55
No studies were found in patients with screen-detected COPD and relatively few in patients with mild COPD (FEV1 ≥80% predicted). Most of the subanalyses of patients with mild to moderate COPD in treatment trials included populations at the more severe end of moderate COPD.
One post hoc subanalysis of a large 4-group RCT and 1 post hoc pooled subanalysis from 3 other RCTs (n = 3174) were identified (Table 5).44,45 Based on 1 subanalysis reporting each outcome (all-cause mortality, exacerbations, health-related quality of life [HrQOL], and dyspnea scores), LABAs appeared to reduce exacerbations and dyspnea scores; results were mixed for HrQOL, and no trials reported exercise capacity. Results for harms were rarely reported, with few differences between treated and untreated groups for a variety of individual adverse events; however, lower rates of study withdrawal and pneumonia were reported in 1 trial in patients treated with salmeterol (Table 6).44- 46,50,51,53- 55
One trial of patients with mild to moderate COPD, 2 post hoc subanalyses of RCTs, and 2 RCTs of patients with a mean FEV1 greater than or equal to 60% (n = 3983) were identified (Table 5). Despite the rarely reported outcomes and limitations, overall results seemed to indicate a reduction in COPD exacerbations with ICS; however, exacerbations were variably defined, and therefore annual rates of exacerbations varied widely. Results from the 1 trial in patients with mild to moderate COPD (EUROSCOP; n = 1175) showed a statistically significant difference in exacerbation rates, but because the annual rates of exacerbations were very low (<0.1 exacerbations/y) in patients with milder COPD severities, the absolute difference was very small (absolute difference of 0.02 exacerbations/y).49 Data were not sufficiently reported to make firm conclusions about the effect of ICS treatment on dyspnea or HrQOL.
Six RCTs reported treatment harms associated with ICS among patients with mild to moderate COPD (Table 6).45,49- 53 Overall, withdrawal rates between treatment groups were similar in the 4 trials that reported these data.45,49- 51 Results of the composite outcome of any adverse event or serious adverse events were mixed but generally showed few differences between treated and untreated groups. Data on pneumonia, bone mineral density, and fractures were sparse and mixed. One post hoc subanalysis reported more ischemic cardiac events among those in the placebo group (3.0% vs 5.3%; P = .048), although these results should be interpreted with caution due to study methods (see full evidence report11).54
The 1 included post hoc subanalysis among patients with moderate COPD (n = 1097) suggested a possible all-cause mortality benefit that was not seen in the main trial across all stages of COPD (Table 5).45 In addition, data showed a statistically significant, but probably not clinically meaningful, improvement in HrQOL and a reduction in exacerbations; however, more evidence is required to make firm conclusions.
Two treatment effectiveness RCTs provided data on harms associated with treating patients with mild to moderate COPD with the combination of LABAs and ICS (Table 6).45,51 Withdrawal rates appeared to be mixed, with the subanalysis of the TORCH trial reporting lower rates of withdrawal among patients treated with salmeterol and fluticasone than those treated with placebo,45 and another trial reporting similar rates of withdrawal between treatment groups.51 Only the subanalysis of the TORCH trial reported on the incidence of composite or individual adverse events. It showed similar rates between treated and control groups, except perhaps a higher risk for pneumonia with treatment, in contrast to findings for LABAs in the same study.45 Paucity of data made robust conclusions challenging.
A single trial of the LAMA tiotropium in patients naive to maintenance treatment with moderate COPD (n = 457)55 and 4 subgroup analyses examining those with mild to moderate COPD were identified (Table 5).44,46- 48 Results were mixed for the effect of tiotropium on exacerbations and HrQOL, although the majority of the evidence suggested a beneficial effect on both outcomes. The trial of treatment-naive patients with moderate COPD most approximated a screen-detected population and showed a statistically significant reduction in exacerbations and statistically significant, but probably not clinically meaningful, difference in work productivity scores.55
Two treatment effectiveness RCTs46,55 and 1 post hoc analysis of pooled study data44 provided few data on harms associated with tiotropium (Table 6). One trial46 reported very similar withdrawal rates with and without tiotropium, with approximately half of these withdrawals attributed to adverse events in both groups; 1 study reported higher rates of any adverse event in the tiotropium group compared with control (67% v 55.9%, no statistical testing), but 1 study reported no difference in serious events.44,55
Overall, the treatment literature was largely based on patients with COPD on the more severe range of moderate, so applicability to a screen-detected asymptomatic population may not be appropriate. Furthermore, there were a number of limitations in the included subgroup analyses for all classes of medications, such as (1) the primary trials were powered for the entire population, not the subgroup; (2) analyses were mostly post hoc; and (3) interaction testing and adjustment for confounders were rarely performed. The inconsistency in reported outcomes across the studies further limited the strength of available evidence. The most inconsistency was seen in the definitions of exacerbations used in the studies. Most studies defined an exacerbation as requiring treatment with an antibiotic or systemic corticosteroid; however, other studies included patient-reported increase in symptoms. Fewer than 5 trials reported harms for any individual medication class in patients with mild to moderate COPD, limiting the ability to make firm conclusions regarding the risk of treating patients with early disease.
An overall summary of the evidence is presented in Table 7. No population-based trials were identified that compared screening with no screening to determine whether primary care screening for COPD improves health outcomes. Quiz Ref IDThrough the indirect evidence pathway, we found that the CDQ appears to be the strongest risk factor– and symptom-based questionnaire, demonstrating reasonable sensitivity and specificity for scores greater than 16.5, based on 5 external validation studies. However, no treatment trials in screen-detected populations were identified; treatment trials and subanalyses in populations with symptomatic mild to moderate COPD supported a modest reduction in exacerbation frequency. Strong evidence was not found to support screening as a means to improve smoking cessation rates or other preventive services.
Quiz Ref IDThe value assigned to screening for COPD depends not only on the accuracy of screening tools; it requires judgments about the evidence for the benefit of earlier identification balanced against the harms of missed cases, false-positive diagnostic workups, and treatment harms. Earlier identification could lead to net screening benefits if there was evidence of beneficial outcomes derived from downstream treatment in early-stage asymptomatic disease or improvements in the uptake of preventive interventions. However, the treatment literature was largely limited to subgroup analyses, almost exclusively among individuals with moderate COPD and primarily the severe end of moderate (eg, FEV1% predicted of approximately 60% in many studies). Even among these groups, the benefits on exacerbations and dyspnea scores with early treatment in patients with moderate COPD are not strongly established, and the clinical significance of the observed reduction may be limited. Absolute treatment benefit estimates would be expected to be lower in screen-detected populations with mostly mild disease compared with the populations in the trials available for this systematic review. Epidemiologic studies report that patients with mild to moderate COPD have an average of less than 1 exacerbation per year.7 A systematic review of RCTs and cohort studies reported an annual event-based exacerbation frequency (defined as physician office visits, antibiotic use, steroid use, or hospitalizations) of 0.82 (95% CI, 0.46-1.49) for mild disease and 1.17 (95% CI, 0.93-1.50) for moderate disease.56 Patients with screen-detected COPD might be expected to have even fewer exacerbations, which would render the absolute benefit as at best modest. Limited data on harms reported in the treatment effectiveness trials suggest that there are no substantial serious adverse effects for most medications. However, some concerns remain about ICS-containing medications increasing the incidence of pneumonia in patients with more severe COPD and effects on bone mineral density and fracture risk.57,58 Data were too limited to make firm conclusions regarding this potential harm in the included trials for screen-detected individuals with mild to moderate COPD.
Arguments have been made that the high prevalence of undiagnosed COPD (10%-20%),16 as well as clinical COPD misdiagnoses in smokers who are found to have alternate treatable diagnoses (eg, congestive heart failure) could be considered as potential benefits with few screening-related harms, because spirometry is a simple, noninvasive test.59,60 In contrast, however, are concerns about the patient-focused benefits of population screening efforts in largely asymptomatic patients, particularly in light of little evidence on treatment benefit in mild disease, opportunity costs associated with screening, and high monthly costs of these inhaled medications.61- 63
A large potential benefit from COPD screening would be increasing smoking cessation rates, because smoking cessation is the only proven beneficial treatment for reducing progression in mild to moderate COPD.64 Smoking cessation counseling and pharmacotherapy are effective in patients with COPD,65- 67 even though there is some evidence that smokers with COPD differ in their motivation to quit compared with smokers without COPD.68- 71 The lung age trial by Parkes et al41 was the only trial reporting a statistically significant absolute increase in biochemically confirmed cessation rates (7%) when screening results reported lung age to participants (NNT = 14). Because both groups received spirometry and counseling, the communication of lung damage might be the critical component in counseling. However, these positive results have not been replicated in other trials.40 On the other hand, we did not identify literature to support the premise that false reassurance in those with normal spirometry may dampen motivation to quit. No completed or pending trials reporting the effect of COPD screening on recommended immunization uptake rates were identified.
A challenging issue when considering screening for COPD is the requirement for an asymptomatic population. All questionnaires included symptoms as part of their scoring, and the rationale for screening has largely been a case-finding one (ie, there is substantial undiagnosed disease seen in primary care). As per USPSTF scope, the systematic review was focused on asymptomatic individuals. However, the distinction between patients who are symptomatic and those who are undetected or who present with nonspecific symptoms was difficult to determine from available clinical research. Because there were few RCTs or screening studies, and the ones identified were clinically and methodologically diverse, we were limited to qualitative analysis. Our a priori methods specified patient-focused outcomes and did not include changes in FEV1 because it is unclear how changes in FEV1 correlate with changes in patient-oriented health outcomes such as exacerbation rates. In addition, we relied on harms data as reported in the effectiveness RCTs and thus may not have captured the full range of potential adverse effects or their population-based incidence. It is unlikely, however, that observational studies in screen-detected populations applicable to US-based primary care are readily available given current practice.
Quiz Ref IDThere was no direct evidence available to determine the benefits and harms of screening asymptomatic adults for COPD using questionnaires or office-based screening pulmonary function testing or to determine the benefits of treatment in screen-detected populations. Indirect evidence suggests that the CDQ has moderate overall performance for COPD detection. Among patients with mild to moderate COPD, the benefit of pharmacotherapy for reducing exacerbations was modest.
Corresponding Author: Janelle M. Guirguis-Blake, MD, Kaiser Permanente Research Affiliates Evidence-based Practice Center and Department of Family Medicine, University of Washington, 521 Martin Luther King Jr Way, Tacoma, WA 98405 (email@example.com).
Author Contributions: Dr Guirguis-Blake had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: All authors.
Acquisition, analysis, or interpretation of data: Guirguis-Blake, Senger, Webber, Whitlock.
Drafting of the manuscript: Guirguis-Blake, Senger, Webber, Whitlock.
Critical revision of the manuscript for important intellectual content: Guirguis-Blake, Senger, Webber, Mularski.
Statistical analysis: Senger, Webber.
Obtained funding: Whitlock.
Administrative, technical, or material support: Senger, Webber, Mularski.
Study supervision: Guirguis-Blake, Senger.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.
Funding/Support: This research was funded by the Agency for Healthcare Research and Quality (AHRQ) under a contract to support the US Preventive Services Task Force (USPSTF).
Role of the Funder/Sponsor: This review was conducted by the Kaiser Permanente Research Affiliates Evidence-based Practice Center under contract to AHRQ. Investigators worked with USPSTF members and AHRQ staff to develop the scope, analytic framework, and key questions for this review. AHRQ had no role in study selection, quality assessment, or synthesis. AHRQ staff provided project oversight, reviewed the report to ensure that the analysis met methodological standards, and distributed the draft for peer review. Otherwise, AHRQ had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: We gratefully acknowledge the following individuals for their contributions to this project: the AHRQ staff; the US Preventive Services Task Force; and Evidence-based Practice Center staff members Smyth Lai, MLS; Kevin Lutz, MFA; and Keshia Bigler, BS. USPSTF members and peer reviewers did not receive financial compensation for their contributions.
Additional Information: A draft version of this evidence report underwent external peer review from 3 content experts (Daniel Kotz, PhD, Maastricht University; Anthony Stanley, MD, University of South Wales; and Barbara Yawn, MD, University of Minnesota). Comments were presented to the USPSTF during its deliberation of the evidence and were considered in preparing the final evidence review.
Editorial Disclaimer: This evidence report is presented as a document in support of the accompanying USPSTF Recommendation Statement. It did not undergo additional peer review after submission to JAMA.