Peak flow meter is a sensitive test for assessing airflow limitation in patients with respiratory symptoms suggestive of asthma in primary health care in Brazil

Abstract

Background

The peak flow meter is considered a complementary tool to spirometry for monitoring obstructive lung diseases. In many countries, spirometry is not easily accessible for diagnosing asthma in primary care. In this context, there has been growing interest in evaluating the role of the peak expiratory flow meter (PEFM) for asthma management, particularly in primary care settings. This study aims to assess the usefulness of PEFM in detecting airflow limitation (AFL), as identified by conventional spirometry, in a primary health care (PHC) setting in Brazil.

Methods

This cross-sectional study was nested within a clinical trial. Participants aged 6 to 65, both sexes, with a history of respiratory symptoms suggestive of asthma and uncontrolled disease, from 28 PHC centers across all macroregions of Brazil were included. AFL was defined by spirometry according to international criteria and by PEFM when PEF was < 80% of the predicted value. The Global Lung Function Initiative (GLI) equations were used for parameterization. The accuracy of PEFM in detecting AFL, as identified by spirometry, was analyzed, and sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and area under the ROC curve (AUC) were reported.

Results

Among 298 participants, 107 (36%) had a confirmed asthma diagnosis. Spirometry identified AFL in 36% (n = 107) of participants, while PEFM detected AFL in 64.8% (n = 193). PEFM demonstrated an overall sensitivity of 82.4% (95% CI: 78.0%–86.7%) for detecting AFL compared with spirometry. Specificity was 44.4% (95% CI: 38.8%–50.0%), and AUC for the PEFM was 0.73. The Intraclass Correlation Coefficient (ICC) for absolute PEFM values was 0.01 (95% CI: -0.18 to 0.20, p = 0.473).

Conclusions

The PEFM demonstrates high sensitivity for AFL, but it should be used with caution as a stand-alone tool for asthma diagnosis. Given its low cost and easy access in PHC settings, PEFM seems to be a useful initial screening tool for AFL, especially in resource-limited settings.

Introduction

Asthma is a major noncommunicable disease (NCD) affecting both children and adults, and it is the most common chronic disease among children [1]. Asthma represents a serious global health problem, affecting approximately 300 million people around the world and causing around 1,000 deaths per day [2]. Most of these deaths are preventable and occur in low- and middle-income countries (LMICs) [3]. Some countries report very high rates (up to 90%) of uncontrolled asthma [4].

Spirometry, which measures forced expiratory volume in one second (FEV₁) and the FEV₁/forced vital capacity (FEV₁/FVC) ratio, is the gold standard for assessing lung function and diagnosing airflow limitation (AFL) in asthma [2]. However, spirometry is not always accessible in primary health care (PHC) settings, particularly in LMICs, due to its cost, the need for specialized equipment, and the requirement for trained personnel [3].

In the context of the growing global asthma burden, accurate and accessible diagnostic tools are essential for effective disease management, particularly in PHC settings. Recent studies have explored alternative diagnostic methods, such as the peak expiratory flow rate (PEFR). One study from South Kazakhstan found that PEFR monitoring is not inferior to spirometry for asthma screening, demonstrating high sensitivity and comparable overall accuracy [5]. Another study highlighted the effectiveness of combining PEFR measurements with symptom questionnaires and mini-spirometry to detect asthma and chronic obstructive pulmonary disease (COPD) in clinical practice, providing a viable alternative when high-quality spirometry is unavailable [6]. These findings highlight the potential of the peak expiratory flow meter (PEFM) as an accessible and cost-effective tool for asthma management in resource-limited contexts. However, to our knowledge, no previous study has assessed the usefulness of the PEFM in multiple primary health care centers, with the tests conducted entirely by PHC professionals under routine, real-world conditions. Our study contributes novel evidence to fill this gap, demonstrating its feasibility and scalability within everyday PHC practice.

Given the burden of asthma and the limitations of spirometry in PHC, there is an urgent need to evaluate alternative tools to detect AFL. PHC is the main entry point to Brazil’s public health system [7] and the level at which most asthma cases are initially managed. Therefore, it is essential to have diagnostic tools that are feasible, scalable, and suited to the realities of local services. The limited availability of spirometry at this level creates significant barriers to timely diagnosis and treatment. In this context, the PEFM stands out as a potentially valuable alternative, which is affordable, simple to implement, and capable of enhancing the quality of care from the first point of contact. Our study addresses this gap by assessing an alternative screening tool in a diverse, resource-limited setting. This study is the first of a two-phase project that evaluates, in the first place, the usefulness of the PEFM in detecting AFL compared with conventional spirometry in a real-world PHC context in Brazil. Phase 2 of this project will assess variable expiratory airflow, comparing PEFM with spirometry.

Methods

Study design and participants

This is a cross-sectional accuracy study nested within a pragmatic cluster randomized clinical trial [8], designed to evaluate the performance of the PEFM in detecting expiratory AFL, as defined by spirometry, among individuals with symptoms suggestive of asthma. During the first study visit, the collection of clinical data was followed by completing the asthma control test (ACT) and lung function tests (PEFM and spirometry). The research was conducted across 28 PHC centers across all macroregions of Brazil. These regions encompass diverse populations and varying levels of healthcare infrastructure, offering a comprehensive overview of asthma management in different primary care settings across Brazil.

Participants aged 6 to 65 years, of both sexes, presenting with symptoms suggestive of asthma and an ACT score < 20 (uncontrolled asthma) [9] were included. For participants aged 6–11 years, the Childhood Asthma Control Test (C-ACT) [10] was used, as per the instrument’s validation for this age group. These instruments were used to identify individuals with recent symptomatic activity. Patients were excluded if they had been hospitalized in the last 4 weeks; had lung diseases other than asthma, such as chronic obstructive pulmonary disease (COPD); had cognitive impairment or uncontrolled comorbidities, such as serious cardiovascular diseases and malignant disorders; had a significant smoking history (> 20 years of smoking or > 10 pack-years); or were pregnant, breastfeeding, or planning a pregnancy.

Procedures

Based on spirometric assessments, participants were classified as AFL-positive if they presented an FEV₁/FVC ratio < 0.80 (≥ 12 years) or < 0.90 (< 12 years), and an FEV₁ <80% of the predicted value [11]. Those who did not meet these criteria were classified as AFL-negative. These categories were adopted to reflect the presence or absence of expiratory AFL. Although the primary aim of this study was not to diagnose asthma, the AFL-positive group was characterized using clinical criteria consistent with asthma, such as symptom variability and a positive bronchodilator response. A positive bronchodilator response was defined as an increase of more than 10% in FEV₁ relative to the predicted value, as this criterion reduces confounding by lung size, sex, and age, and better reflects the physiological upper limit of normal response observed in healthy individuals [12]. This approach ensured the inclusion of participants with clinical and functional profiles typically observed in asthma, to evaluate the diagnostic performance of PEFM in detecting AFL. Complete sample characterization data were collected for the AFL-positive group. Participants classified as AFL-negative were considered a comparison group for performance analysis of the usefulness of the PEFM in detecting AFL compared with conventional spirometry. PEFM measurement was performed immediately before spirometry, and all tests were conducted by PHC professionals who had been previously trained by the main study team.

Lung function tests

Data were collected through spirometry and PEFM tests. Spirometry was performed using a Winspiro PRO spirometer (MIR - Medical International Research S.p.A., Viale Luigi Schiavonetti 270 00173, Roma, Italy) [13] following the guidelines published by the European Respiratory Society (ERS) and the American Thoracic Society (ATS) [14]. Acceptable and reproducible maximal flow-volume curves were obtained. Results were adjusted for height, age, ethnicity, and sex according to international reference values [15].

PEFM readings were taken using standardized peak flow meters, and participants received instructions on the proper use to minimize measurement variability. PEF was measured using a hollow cylinder ATS scale peak flow meter (Medicate^®), which ranged from 60 to 900 L/min. Peak flow measurements consisted of at least three acceptable blows after complete inhalation, with the highest two PEF readings repeatable within 40 L/min. The highest PEF was recorded as the patient’s PEF. Previously published reference equations were used to predict the PEF% [16]. The patients were classified as having expiratory AFL when the observed PEF value was less than 80% of the predicted value.

Statistical analysis

The power calculation for the sample (n = 298) of peak flow and spirometry tests focused on the diagnostic test’s sensitivity. The prevalence of expiratory AFL in the sample was 38.9%. We assumed an expected sensitivity of 80% and aimed to detect a minimum difference of 10% (delta = 0.10) from a lower reference value of 70%, with a significance level of 0.05. The R “pwr” package [17] was used to calculate the power of the sample, considering the test’s prevalence and characteristics. The calculation indicated a power of 86.3%, suggesting a high probability of detecting a significant difference in sensitivity if the true sensitivity is 80% and the minimum difference is 10%. This result indicates that the sample is sufficiently robust to reliably detect meaningful variations in the test’s sensitivity.

Statistical analysis was performed using the statistical software R (version 4.0.5, the R Foundation for Statistical Computing, Vienna, Austria; available from: https://www.R-project.org/) [18]. The statistical significance level was set as a p-value less than 0.05 for two-tailed tests. Continuous variables are presented as mean and standard deviation (SD) or median and interquartile range (IQR), as appropriate. Categorical variables are presented as absolute (n) and relative (%) frequencies. The usefulness of PEFM in detecting AFL, as identified by spirometry, was evaluated using several statistical measures. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV), and area under the ROC curve (AUC) were calculated to determine the diagnostic performance of PEFM. Agreement between PEFM and spirometry measurements was assessed using the Intraclass Correlation Coefficient (ICC) and the Bland-Altman method. The ICC was used to evaluate the consistency of measurements between both methods. The Bland-Altman analysis provided a graphical representation of the agreement, highlighting any systematic differences and the limits of agreement. Subgroup analyses were conducted to investigate variations in PEFM usefulness across age groups and eligibility status.

Results

A total of 298 participants completed all study procedures and were included in the analysis. Based on spirometry results, 107 individuals (36%) were classified as AFL-positive, while 191 (64%) did not meet AFL criteria and were classified as AFL-negative (Fig. 1). This classification guided all PEFM performance analyses aimed at detecting expiratory AFL. The characteristics of the participants are summarized in Table 1.

Fig. 1.

Protocol study flow-chart

Table 1.

Characteristics of the participants

Characteristics	Total AFL-positive group (n = 107)	Total AFL-negative group (n = 191)
Age (years), median (IQR)	29.78 (14.39; 47.45)	43.66 (26.26; 52.37)
Females, no. (%)	65 (60.7)	152 (79.5)
Body-mass index†, median (IQR)	24.9 (19.5; 31.8)	28.1 (23.7; 33.1)
Skin color — no. (%)
White	58 (54.2)	155 (81.1)
Black	6 (5.6)	5 (2.7)
Brown	43 (40.2)	31 (16.2)
Asthma severity* — no. (%)
Mild	17 (16.0)	-
Moderate	72 (68.0)	-
Severe	17 (16.0)	-
Comorbidities — no. (%)
Allergic rhinitis	46 (43.0)	-
Obesity	16 (15.0)	-
Gastroesophageal reflux	12 (11.2)	-
Others	33 (30.8)
Respiratory medication use — no. (%)
No medication	20 (18.6)	-
SABA as needed only	23 (21.4)	-
Any controller therapy	64 (60.0)	-
Lung function tests
PEF % predicted,	62 (50.5; 73.5)	77 (66; 89.5)
FEV1% predicted	59 (49; 72.5)	82.63 ± 18.6
FEV1/FVC %	75 (67.5; 78)	90 (84; 97)

Qualitative variables are expressed as absolute and relative frequencies, no. (%). Quantitative variables are presented as mean ± SD if normally distributed (as determined by the Shapiro-Wilk test; p > 0.05) or as median (IQR) otherwise

Abbreviations: SD Standard Deviation, IQR Interquartile Range, presented as (p25; p75), where p25 and p75 represent the 25th and 75th percentiles, respectively

†: The body-mass index is the weight in kilograms divided by the square of the height in meters

*: Analysis based on 106 participants; one missing data point

Sensitivity of PEFM in detecting AFL

AFL was identified in 36% (n = 107) of participants by spirometry. Using the PEFM, AFL was detected in 64.8% (n = 193) of the participants. PEFM demonstrated an overall sensitivity of 82.4% (95% CI: 78.0% to 86.7%) in detecting AFL compared to spirometry. Specificity was 44.4% (95% CI: 38.8% to 50.0%) and AUC for the PEFM was 0.73 (Fig. 2). The PPV was 43.5% (95% CI: 37.9% to 49.1%), and the NPV was 82.9% (95% CI: 78.6% to 87.1%).

Fig. 2.

Area under the ROC curve of PEFM for the detection of AFL (all patients)

Agreement between PEFM and spirometry

Agreement analyses were performed using the ICC and Bland-Altman plots. The ICC for absolute PEFM values was low, at 0.01 (95% CI: -0.18 to 0.20, p = 0.473), indicating poor agreement. Similarly, PEF% agreement with FEV₁% and FEV₁/FVC% yielded ICC values of 0.17 (95% CI: -0.02 to 0.35, p = 0.042) and 0.21 (95% CI: 0.02 to 0.38, p = 0.016), respectively. Bland-Altman plots demonstrated systematic bias, with wider limits of agreement, reinforcing the limited interchangeability between PEFM and spirometry (Fig. 3).

Fig. 3.

Bland-Altman plots of the differences between PEFM and spirometry measures

Subgroup analyses

In participants aged ≥ 18 years (n = 233), sensitivity reached 91.8% (95% CI: 88.3 to 95.4) and a specificity of 41.5% (95% CI: 35.1 to 47.8). The AUC for the PEFM was 0.79 (Fig. 4). Spearman correlation analyses in these participants demonstrated a moderate association between PFE% and FEV₁% (r = 0.513, p < 0.001) and between PFE% and FEV₁/FVC% (r = 0.515, p < 0.001), while FEV₁% and FEV₁/FVC% were strongly correlated (r = 0.643, p < 0.001). The ICC for absolute PEFM values was low, at 0.01 (95% CI: -0.23 to 0.24; p = 0.472), indicating poor agreement. Similarly, the agreement of PEF% with FEV₁% and FEV₁/FVC% yielded ICC values of 0.01 (95% CI: -0.23 to 0.24; p = 0.482) and 0.03 (95% CI: -0.2 to 0.27; p = 0.387), respectively. Additionally, Bland–Altman plots demonstrated wide limits of agreement, indicating limited interchangeability between the two methods (Fig. 5). In participants aged < 18 years (n = 65), sensitivity was 57.1% (95% CI: 45.1 to 69.2) and specificity was 56.8% (95% CI: 44.7 to 68.8). The AUC for the PEFM was 0.65. Spearman correlation analyses showed a weak, non-significant association between PEF% and FEV₁% (r = 0.166, p = 0.186) and a moderate correlation between PEF% and FEV₁/FVC% (r = 0.390, p = 0.001), while FEV₁% and FEV₁/FVC% were strongly correlated (r = 0.591, p < 0.001). The ICC for absolute PEFM values was 0.00 (95% CI: -0.32 to 0.32; p = 0.491), with similarly negligible agreement observed for PEF% with FEV₁% (ICC = 0.00, 95% CI: -0.32 to 0.32; p = 0.502) and PEF% with FEV₁/FVC% (ICC = 0.01, 95% CI: -0.32 to 0.32; p = 0.488). Bland–Altman plots also revealed wide limits of agreement in this subgroup.

Fig. 4.

Area under the ROC curve of PEFM for the detection of AFL (subgroup ≥18 y)

Fig. 5.

Bland-Altman plots of the differences between PEFM and spirometry measures (subgroup ≥18 y)

Exploratory correlation analysis between PEFM and spirometry

Spearman correlation analyses demonstrated that the absolute peak flow measurements were strongly correlated (r = 0.659, p < 0.001). The association between PEF% and FEV₁% was moderate (r = 0.431, p < 0.001), and a similar moderate correlation was observed between PEF% and FEV₁/FVC% (r = 0.462, p < 0.001). Moreover, FEV₁% and FEV₁/FVC% were strongly correlated (r = 0.633, p < 0.001).

Discussion

The present study evaluated the usefulness of the PEFM in detecting AFL compared to spirometry in a real-life setting of primary care in Brazil, which is the gold standard for assessing pulmonary function. Our findings indicate that while PEFM demonstrates high sensitivity, its specificity and overall agreement with spirometry remain limited. Importantly, the AUC results provide further insight, showing moderate overall discrimination and even better performance in adults, which underscores the practical value of the PEFM as a feasible tool to support the early identification of airflow limitation. However, the moderate correlation and low specificity suggest that PEFM should not be used as a standalone diagnostic method, but rather as part of a comprehensive asthma management strategy. These results highlight the potential of PEFM as a screening tool in PHC settings, particularly in resource-limited environments, where access to spirometry is restricted. In Brazil, restricted access is further compounded by the limited number of pulmonologists, who represent less than 1% of all medical specialists [19] and are unevenly distributed across the country. As they are typically responsible for interpreting spirometry, their scarcity contributes to diagnostic delays and the underuse of pulmonary function testing at the primary care level.

Our results are in agreement with previous studies. Thorat et al. [6] reported that a PEF cut-off of < 80% predicted yielded a sensitivity of 90% and a specificity of 50% for detecting AFL. Notably, when combined with a brief symptom questionnaire, the performance improved to a sensitivity of 84% and specificity of 93%. Similarly, Tian et al. [20] demonstrated that using a PEF < 80% predicted provided sensitivities ranging from 76.8% to 85.3% and specificities between 81.9% and 83.8% for detecting AFL in primary care settings. In addition, previous study [21] have also underscored that, while PEF measurements are useful for initial screening of obstructive airway diseases, their variability and poor agreement with spirometry limit their capacity to serve as definitive diagnostic tools, indicating that while PEF measurements can identify individuals at risk, they should ideally be complemented by additional lung function tests.

Our study demonstrated a moderate to strong correlation between PEFM and spirometric parameters (absolute PEF vs. FEV₁: r = 0.659, p < 0.001; PEF% vs. FEV₁%: r = 0.431, p < 0.001). This is consistent with the findings by Jackson and Hubbard [21], who also observed moderate correlations between PEF and spirometry values in a population-based study. However, despite this correlation, Bland-Altman plots revealed systematic bias and wide limits of agreement, highlighting poor concordance between the two methods. Critically, the high sensitivity of PEF for detecting AFL does not imply a numerical agreement with spirometry values; these metrics assess different properties of the test. This aligns with previous work by Aggarwal et al. [22] and Hegewald et al. [23], who emphasized that PEF measurements alone cannot reliably replace spirometry due to significant variability.

Subgroup analyses in our study revealed notable differences in diagnostic performance across age groups. In participants aged 18 years or older, accuracy reached 57.5%, with a high sensitivity of 91.9% and a specificity of 41.5%. These results suggest that PEFM may be more effective in detecting airflow limitation in adults than in children and adolescents. Tian et al. [20] similarly observed that PEF performed better in older populations, likely due to greater airflow limitation in adults, which PEF more readily detects. However, despite the high sensitivity, the poor specificity observed in our study raises concerns about false-positive diagnoses, further reinforcing the need for confirmatory spirometry. These findings indicate that while PEFM demonstrates high sensitivity and moderate correlation with spirometry, its low specificity and poor agreement suggest that it may not be a reliable standalone diagnostic tool for AFL. Nevertheless, given its accessibility in PHC settings, it can be useful as an initial screening tool, particularly for adult populations and in resource-limited environments.

Taken together, our findings suggest that although the PEFM should not replace spirometry, it may serve as a pragmatic tool to prioritise patients for referral to full pulmonary function testing, particularly in contexts where access is limited. This targeted approach reinforces the attributes of coordination and equity [24] in PHC by helping ensure that diagnostic resources are allocated to those most in need. Moreover, it may optimise system efficiency and enable more timely, evidence-based asthma management, ultimately strengthening the attributes of access and comprehensiveness at the frontline of Brazil’s public health system.

A key strength of our study is its multicenter design across diverse PHC settings, conducted entirely by primary care teams under real-world conditions. To our knowledge, no previous study has evaluated the usefulness of the PEFM in this way, providing novel, practice-relevant evidence to support its feasibility and scalability in routine PHC. However, our study has important limitations. First, the observed low specificity of the PEFM can result in an increased rate of false positives, which could lead to unnecessary referrals or anxiety if used in isolation. Second, this phase of the study did not assess bronchodilator response, which is a key criterion for diagnosing asthma and distinguishing it from other causes of fixed airflow limitation. We acknowledge that the comparative analysis of variable expiratory airflow using the two methods is currently the most important criterion for diagnosing asthma. However, the availability of a very simple method for assessing AFL (PEFM) makes it a feasible tool for screening for airway obstruction in low-resource settings, and our Phase 2 study will address this question. To strengthen our hypothesis, we are currently analyzing, in a larger sample of the present population, the role of PEFM in assessing the variable expiratory airflow (positive bronchodilator responsiveness) compared with spirometry. Third, a notable finding from our subgroup analysis was the differential performance of PEFM across age groups. The test demonstrated high sensitivity in adults, but lower sensitivity in children and adolescents. This indicates that PEFM’s overall screening accuracy can be influenced by the adult population in our sample.

Future studies should focus on multicenter studies with larger, more diverse populations to validate and refine optimal PEF cut-off values. In addition, the potential role of mini-spirometers, despite some limitations in accuracy, may offer additional functional insights and warrants further investigation. Such studies could develop combined algorithms that integrate PEF measurements with clinical parameters and symptom assessments to improve overall accuracy and precision. In addition, our pragmatic design prioritized real-world performance over analysis of confounders, such as demographic (e.g., gender) or contextual (e.g., region, socioeconomic status) variables, which should be explored in future implementation studies.

As a low-cost, easy-to-use device, the PEFM offers a feasible, scalable solution for detecting AFL and supporting initial asthma screening at the primary care level, particularly in LMICs. While its utility as a screening tool appears promising in adult primary care, it may be limited in pediatric settings, underscoring the need for age-specific validation. Our results suggest that, where applicable, PEFM facilitates screening and may improve teams’ responsiveness in initial management, contributing to more efficient allocation of diagnostic resources and reduced health costs in the context of the high burden of asthma.

Finally, in the real-world PHC setting, the PEFM stands out as a sensitive screening tool for patients with respiratory symptoms suggestive of asthma, with a low risk of false-negative results. A normal reading makes the presence of significant airflow limitation unlikely, providing greater confidence in the initial exclusion of the diagnosis. Conversely, abnormal results should be interpreted as a warning sign rather than diagnostic confirmation, underscoring the need for further investigation, preferably through spirometry. Thus, the peak flow meter plays a strategic role in structuring asthma care within PHC by supporting early clinical decision-making, optimizing referrals, and expanding access to functional respiratory assessment in resource-limited settings.

Authors’ contributions

All authors contributed to the statistical analysis plan and revised this manuscript: APTL, FF, BSR, GLA, IRF, LSC, LDS, MSC, TRC, MPSPT, AAC, MHJ, and PMP. Additionally, PMP conceived the study, and FF and MHJ contributed to its design. All authors read and approved the final manuscript.

Funding

This study was funded by the Brazilian Ministry of Health through the Brazilian Unified Health System Institutional Development Programme (PROADI-SUS). The funding body has had no role in the design of the study; the collection, analysis, or interpretation of data; or the writing of the manuscript. Its role is limited to providing supporting information for site recruitment.

Data availability

The datasets generated and/or analyzed during the current study are not publicly available. However, they may be available from the corresponding author on reasonable request and with authorization of the Brazilian Ministry of Health.

Declarations

Ethics approval and consent to participate

The CuidAR study complies with Brazilian National Health Council Resolution No. 466/2012 on human subject research, the Declaration of Helsinki, the Document of the Americas, and ICH/GCP E6 (R2) 2016. The study was approved by the Institutional Review Board (IRB) of the coordinating center (HMV; approval number 7.016.676, CAEE: 61054222.3.0000.5330). A waiver of informed consent was requested from the IRB (or IRB equivalent) of each participating center.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.