Reference values for peak expiratory flow in Indian adult population using a European Union scale peak flow meterRR Kodgule1, V Singh2, R Dhar3, BG Saicharan4, SJ Madas1, JA Gogtay5, SS Salvi1, PA Koul6
1 Department of Academic Research, Chest Research Foundation, Pune, Maharashtra, India
2 Department of Research, Asthma Bhawan, Jaipur, Rajasthan, India
3 Research Department, National Asthma, Allergy, and Bronchitis Institute (NAABI), Kolkata, West Bengal, India
4 Research Department, Lung Care Center, Hyderabad, Andhra Pradesh, India
5 Respiratory Department, Cipla Pharmaceuticals Limited, Mumbai, Maharashtra, India
6 Department of Internal and Pulmonary Medicine, Sheri-Kashmir Institute of Medical Sciences, Srinagar, Jammu and Kashmir, India
Correspondence Address: Source of Support: The study was supported by grant from Cipla Limited,, Conflict of Interest: Author Gogtay JA is an employee of Cipla limited that
provided an unrestricted grant for the conduct of the study. DOI: 10.4103/0022-3859.132311
Source of Support: The study was supported by grant from Cipla Limited,, Conflict of Interest: Author Gogtay JA is an employee of Cipla limited that provided an unrestricted grant for the conduct of the study.
Background: Reference values of peak expiratory flow (PEF) in Indian adults have to date been derived locally, using an old (Wright) scale peak flow meter. There are thus no reliable reference values for PEF for Indians and this formed the aim of the study. Materials and Methods: A European Union (EU) scale peak flow meter (PFM) was used for the study. A respiratory health and demographic questionnaire was administered to 1000 male and female adults from randomly selected locations in the country in this multi centric study. The locations represented different geographic, ethnic, and socioeconomic backgrounds. Patients were stratified according to height and age. The PEF values were measured using the Breathometer™ (Cipla Ltd., India) with EU scale. Reference equations were derived from multiple regression analysis. Results: A total of 3608 participants were excluded. In 80% of the remaining 6138 healthy adults (M: 3720; F: 2418), the predicted regression equations were derived. Gender, age, and height were the significant determinants of PEF. The equations in L/minute are: Females: PEF = -1.454 (Age) + 2.368 (Height) Males: PEF = -1.807 (Age) + 3.206 (Height). The derived equation was validated by comparing the predicted PEF values with the measured values in the remaining sample of 20% (Mean ΔPEF: M = 1.85 L/minute, CI = -2.76, 6.47; F = 1.64, CI = -2.89, 6.18). An Indian adult with average height and age was found to have approximately 30% lower PEF compared to the corresponding European adult using the Nunn and Gregg equation. Conclusion: We derived reference values of PEF for Indian adults using a validated EU scale peak flow meter.
Keywords: Adult, asthma, clinical respiratory medicine, COPD, peak expiratory flow rate, peak flow meter, reference values, respiratory function tests
Spirometry and peak expiratory flow (PEF) measurements are the most widely used tests to assess airflow limitation. Although spirometry is a more repeatable and objective measure of airflow limitation, the peak flow meter (PFM) is a handy, easy-to-use tool, which can be an important aid in both the diagnosis and monitoring of asthma. It measures PEF rate, which is the maximum airflow achieved during a forceful expiratory maneuver, beginning with the lungs fully inflated. It can be used to confirm the diagnosis of asthma (by measuring the bronchodilator reversibility and/or diurnal peak flow variability over a period of two weeks), to grade asthma severity, and also for monitoring the response to pharmacotherapy. During acute asthma exacerbations, the PEF measurement serves as an important guide for disease management as spirometry may be contraindicated. The Global Initiative for Asthma (GINA) guidelines, for asthma management, advocate classification of the severity of asthma exacerbations, based on the percent predicted values of PEF and also recommend its use for guiding asthma management.  Regular monitoring of PEF is especially useful among those who are 'poor perceivers' of their symptoms and among those who have severe asthma.  Early diagnosis by screening is crucial for preventive intervention and arresting disease progression during the management of chronic obstructive pulmonary disease (COPD), as therapeutic options have limited effectiveness. More recently, the peak flow meter has been shown to be a cost-effective screening tool for COPD  and has been recommended by the International COPD Coalition for screening high-risk groups for COPD. 
The Mini Wright's PFM is an old device, that has a linear scale with equidistant readings, and has been used to record PEF values for almost three decades. This device was subsequently found to over-read by about 70 L/minute in the middle of the flow range and under-read by about 50 L/minute in the higher flow range. , Therefore, in 2004, the scale of this device was replaced with a new 'European Union (EU) scale' [Figure 1]. This scale has been shown to be more accurate than the old scale, and as a result, all cylindrical peak flow meters are now recommended to have the EU scale, to ensure accurate readings. ,
The PEF values obtained in a patient from a peak flow meter need to be compared with values from a normal, healthy population (reference values). These reference values need to be derived locally as they are influenced by ethnicity, race, climatic conditions, and nutritional status. Although guidelines recommend comparing the measured PEF values with the patient's personal best, this approach has been found to be practically inaccurate and unreliable. A large American study reported that only 29% of asthma patients admitted to the Emergency Department knew their personal best PEF and out of them 45% had a measured PEF greater than the reported personal best, during the follow-up. Reliance only on 'the personal best' has the potential to lead to an inappropriate discharge from the Emergency Department.  In the past, reference equations for PEF for the Indian population have been derived by investigators from different cities and towns, including, Chennai,  Chandigarh,  North Arcot,  and Lucknow.  However, the results of these studies are limited by (a) The study population being limited only to a local region, (b) skewed age and height distributions, and (c) use of the Mini Wright's scale device to derive the predicted equations. Large differences noted between the predicted PEF values derived from these different studies, has compounded the problem. The present study thus aimed to derive reference values for PEF for an adult Indian population, using an EU scale peak flow meter and attempting to eliminate inadequacies of previous Indian studies.
The study was given the acronym 'PERFORM' which meant 'Peak Expiratory Flow Rate among Indian adults' and was primarily coordinated by the Chest Research Foundation, Pune, India. The study was approved by the Independent Ethics Committee of the Chest Research Foundation and written informed consent was obtained from all participants.
Five centers (Srinagar, Jaipur, Kolkata, Pune, and Hyderabad), which represented different ethnic races, cultures, geographies, and dietary habits, from north, west, east, central, and south zones of India, respectively, were selected as study locations. Qualified and trained investigators (who in turn trained four field workers) The PEF values were obtained using the Breathometer ® (Cipla Ltd., India) peak flow meter. Their quality of data acquisition was then examined in a pilot study of 20 subjects after which the actual study was initiated.
Subjects aged >18 years, who provided written informed consent and were able to perform satisfactory blows on peak flow meter were included in the study. Subjects with respiratory symptoms and/or history of significant exposure to risk factors for obstructive airways disease were excluded from the study.
Sample size and sampling
A sample size of between 200-500 participants has been suggested previously to be appropriate for multiple regression.  We chose a value of 500 males and 500 females to account for drop outs and for separate regression equations for each gender. Assuming that only 40% of the study subjects would be healthy for derivation of the reference equations (+ 15% withdrawals), based on our experience from earlier studies, , we estimated that a sample size of 1000 male and 1000 female subjects per center (i.e., a total of 10,000 subjects from India) for a reliable estimate. The study subjects were recruited from randomly selected market places, colleges, universities, government offices, and religious places (convenience sampling).
Instruments for assessment
A validated health questionnaire designed by the Chest research foundation to distinguish healthy from unhealthy participants was used. The Breathometer ® (EU Scale Peak Flow Meter, Cipla Limited, India) was used for PEF measurement. The Breathometer conformed to the EN 13826:2003 standards for manufacturing, and calibration as recommended by the European Union. The English version of the questionnaire was translated and back translated into the local languages (Hindi, Marathi, and Bengali) before it was used at all the centers.
Equipment calibration and accuracy
Each device was found accurate after being tested by the manufacturer in the laboratory, by using a calibrating device that could generate accurate flows between 0-900 L/minute, with a 30 milliseconds rise time and 10 milliseconds dwell time, with an abrupt fall of flow after reaching the PEF. Each breathometer was tested at low flows, mid flows, and high flows between 0-800 L/minute both under conditions of constant flow and with flow patterns generated by the calibrating device. In an earlier study, we validated the accuracy of the Breathometer ® device by comparing it with a standard EU scale peak flow meter (Wright's EU scale peak flow meter) (Abstract No 1310, European Respiratory Society Congress, 2008). We also demonstrated the robustness of the device, by showing that the values obtained by one device were similar to those obtained by the other devices. In accordance with the international guidelines,  the subjects performed at least three acceptable blows into the peak flow meter until the two highest PEF values were reproducible within 40 L/minute. The function and accuracy of the peak flow meter was checked daily by measuring the PEF of two field workers at each center (none of them suffered from asthma or other airway or lung disease) and ensuring that the readings were within 20 L/minute of the initial reading.  The same peak flow meter was used every day by the field worker and was replaced by another peak flow meter only if the daily calibration failed.
PEF - case definition
An acceptable PEF value was defined as one that was produced with the hardest blow, after taking a maximum deep inhalation, as interpreted by the field worker on visual appearance. The highest of the three acceptable readings was recorded as the PEF of the individual.
Variables such as age, height, and gender, which are known to affect the lung function, were used to stratify the study population. Based on the age-wise percent population of India,  the estimated sample of 10,000 was stratified into age groups of five years, starting from age 18 years up to 75 years as follows: 18-24 (1000), 25-29 (750), 30-34 (700), 35-39 (600), 40-44 (500), 45-49 (400), 50-54 (350), 55-59 (250), 60-64 (200), 65-69 (150), 70-75 (100). Similarly, the height was also stratified into groups of 5 cm, starting from 150 cm up to 180 cm (the last group showing a spread of 6 cm).
Data management and analysis
The data obtained was centrally entered at the co-coordinating center using the double data entry method and discrepant and incomplete data were either rectified or excluded. After cleaning, the data was locked and subsequently analyzed. Center specific analysis was also done. The demographic and health-related data were expressed descriptively. Eighty percent of the population was randomly selected from each age-and-height group, for deriving the regression equation for the respective populations. As height and age are known to be contributory factors for PEF, the regression equation was expressed as
PEF = β0 + β1 * Age + β2 * Height
β0 , β1 , and β2 in this equation are the regression coefficients. Separate equations were derived for males and females. The derived equation was then tested on the remaining 20% of the population for validation. The equation was considered valid if the difference between the predicted PEF value derived from the equation and the actual measured PEF value in the remaining 20% of the study population was not statistically significant (P value >0.05). The contributions of the independent variables, age and height, to the variability in the PEF values (r 2 ) were also estimated.
The equations for each center were compared with each other and with the median equation by comparing models of height and age using the paired t-test in SPSS software version 11.5. The significance levels of these differences were expressed as actual P values.
A total of 9746 subjects (males: 5284; females: 4462) participated in the study, out of which 3608 (37.02%) subjects were excluded as unhealthy [Table 1]. The main reasons for exclusion were presence of respiratory symptoms (wheezing in last 12 months-16.5%, cough affected by weather-13.8%, frequent wheezing-9.6%) and significant exposure to chulla smoke (7.7%).
Data from the remaining 6138 (Males: 3720; Females: 2418) healthy subjects was used to derive center-specific and national predicted equations [Table 1]. Gender, height, and age were the most important determinants of the predicted PEF. In both the genders and across all the centers, the PEF values decreased with increasing age and increased with increasing height [Figure 2]. Using age and height as regression coefficients, separate linear regression equations were derived for males and females, using SPSS software Version 11.5 software. We randomly selected 80% (2976 males and 1934 females) of the total 3720 males and 2418 females each, and from each age and height bracket, to derive the median reference equation for the entire Indian population. The regression equations obtained for predicting PEF were as follows:
PEF in L/minute (male) = 69.259 - 2.290 (age) + 2.888 (height) (r 2 = 0.27)
PEF in L/minute (female) = 168.551 - 1.776 (age) + 1.354 (height) (r 2 = 0.24)
Age and height together explained only 27 and 24% of the variance for PEF values for males and females respectively. The residual analysis against age showed that the error variance increased with increasing age. Hence, we divided both sides of the earlier models by age, based on the model used by Prasad et al.  and considered the following model for both the genders:
PEF / age = β0 / age + β1 + β2 *height / age
In the stepwise regression analysis we found only height/age as a significant predictor of PEF/age, while the independent variable 1/age did not contribute significantly. Omission of 1/age did not change the r 2 value, and hence, we omitted it from the equation. We multiplied both the sides by age and obtained the following final equations:
Final median national equations
PEF in L/min (male) = -1.807 * age + 3.206 * height (r 2 : 0.892)
PEF in L/min (female) = -1.454 * age + 2.368 * height (r 2 : 0.882)
The highest PEF values were found between the age groups of 18 and 24 years among both males and females, after which it reduced by 2.3 L/minute and 1.8 L/minute per year, on an average, in males and females, respectively. The PEF increased by about 2.9 L/minute and 1.3 L/minute for every centimeter of increase in height, in males and females, respectively. These equations were tested with the remaining 20% of the population (744 male and 484 female subjects) using the paired-t test. The measured PEF values were not different from the predicted values both in males (Δ = 1.85L/minute, P value >0.05) and females (Δ = 1.64 L/minute, P value >0.05). This, therefore, helped us in validating the equation internally.
Similarly reference equations were derived for each center [Table 2] and compared with each other by models for height and age, using dummy variables. Statistically significant differences (P < 0.01) were observed within the reference equations for all the centers, as shown in [Table 3]. The mean difference in PEF values between these centers ranged from 6 L/minute to a maximum of 65 L/minute (Jaipur females versus Kolkata females). The equations of each center were compared with the median equation, by comparing the regression coefficients. We found statistically significant differences (P < 0.01) between the median equation and equations for all the centers in both males and females [Table 4]. These differences ranged from 2 L/minute to a maximum of 35 L/minute (Jaipur versus median equation).
The predicted PEF of an arbitrary male and female of age 35 years and heights of 170 cm and 160 cm, respectively, was calculated using the derived equation for Indian population, the Nunn and Gregg equation ,, for Europeans, and NHANES III equation for American Caucasians.  The predicted values of PEF according to the Indian equation were around 30 and 20% lower compared to the Nunn and Gregg equation ,, for Europeans and NHANES III equation for American Caucasians,  respectively [Figure 3] and [Table 5].
We belive this is the first study in the country conducted using the new EU scale peak flow meter, to derive reference values of peak expiratory flow in an adult population. Like some of the similar previous studies , and unlike others, , we recruited subjects from the community by selecting subjects from the representative locations. However, unlike all the similar previous studies ,,, our study was conducted at five different locations in India thus making it both representative and all encompassing.
We gender, height, and age to be the largest determinants of variability in PEF, which was consistent with the previous observations. ,,, We also found that females on an average had 30% lower PEF values than males. We did not capture the effect of weight, as its role in determining PEF has been debated, because of two opposing effects: Increased weight due to muscular development increases PEF and increased weight due to obesity decreases PEF. 
As expected, we found wide differences between the predicted values obtained in this study and those obtained by the previous Indian studies, before and after correcting the scale change.  One of the reasons could be that the distribution of the age and height in our study was wider than in the previous Indian studies. Interestingly, the peak PEF was achieved at an earlier age (18-24 years) in the present study, as compared to the previous studies.
Compared to the Nunn and Gregg equation ,, for the Europeans and the NHANES III equation for the American Caucasians,  our PEF values were roughly around 30 and 20% lower, respectively. The most common practice in India is to use 90% of the European Community for Coal and Steel (ECCS) equation, derived for the European population, as the reference values for the spirometry parameters. However, our study suggests that the lung function values of the Indians are around 70% compared to those of the Europeans. This argues against the current practice of applying a correction factor of 90%, which is both arbitrary and has been shown to have poor agreement with the Indian equations.  As, there is confusion regarding the predicted values for lung function in India, the reference values for PEF derived by us are probably the only reliable reference values that can be used for assessment of lung function in the adult Indian population, in the current scenario. Use of the western predicted values is likely to misclassify the severity of obstructive airway disease toward increased severity, leading to overtreatment. Furthermore, lower PEF values suggest that a large number of borderline subjects will lie in the middle of the PEF scale, an area where the old scale Peak Flow Meter over-reads. Hence, use of the old scale Peak Flow Meter has the potential for under-diagnosis in such patients. Clearly there is an urgent need to confirm the low lung function in the Indian population by deriving nationwide predicted values for spirometry.
Consistent with the previous observations, we found that the South Indians have lower PEF values compared to North Indians. , However, this is the first study that has compared the PEF values in both halves of the country using a uniform and objective methodology, and at the same time. We also found significant differences between the five equations. These regional differences in the PEF values could be because of inherent genetic, , climatic, and nutritional differences between these populations, and may vouch for using regional equations in clinical practice. We also found statistically significant differences between the median equation and equations for each center, with the differences ranging between 2 L/minute to a maximum of 35 L/minute (Jaipur versus median equation). A clinically-meaningful difference in the PEF values has been suggested to be 60 L/minute or more by the official European Respiratory Society (ERS) statement.  These differences, being much less than 60 L/minute, may be considered as clinically insignificant, suggesting that a median equation may provide the reference values for PEF, for the entire adult Indian population. Considering the importance of the simplicity of the equation for its acceptance in clinical practice,  we recommend using the single median reference equation for the entire country.
Our study has some inherent weaknesses. We omitted 37.02% of the total participants as 'not healthy', thus binging to fore the generalizability of the reference equation. This omission was done in order to prevent even a minor contribution from any disease or abnormality to the PEF values.
We classified the status of subjects as 'healthy' based on a questionnaire which may not have had adequate sensitivity. Variation in the data collection activities at various centers could have contributed to the differences in PEF at various centers. However, as we trained the field workers centrally in data collection and used the same brand of peak flow meter, chances of differences in the PEF values introduced by the variation in data collection are likely to be minimal. Measuring the PEF requires one good forceful blow. Upon training and demonstration, this maneuver is relatively easy to perform even in children above the age of five years. In addition, we obtained three acceptable blows from the subjects with a repeatability of 40 L/minute between the highest two blows and only the highest value was recorded. Yet another limitation of our study is the inclusion of English or Hindi speaking population from Hyderabad,. Hyderabad as a city may not be truly representative of South India. Selection of the study population from the randomly selected market places, colleges, universities, government offices, and religious places again can introduce a minimal sampling bias. An important limitation of our study could be the lack of capturing, more objectively, the important predictors of lung health like the socioeconomic status, nutritional status, and the extent of exposure to environmental pollution.
No criteria have been defined for assessing the quality of the measured values of PEF. The lack of use of objective quality criteria is an important factor that reduces the reliability of PEF values; both while deriving the reference values as well while using them in the clinics. In conclusion, in this nationwide, multicenter study we have,
The authors acknowledge the sincere efforts of Dr Umar Hafiz Khan, Sheri-Kashmir Institute of Medical Sciences, Srinagar, Dr Subhasis Mukherjee, NAABI, Kolkata, and Dr Udaiveer Singh, Asthma Bhawan, Jaipur, in data collection, during their role as investigators in the study. They also acknowledge Cipla Pharmaceuticals Limited, Mumbai, India, for providing unrestricted grants for the conduct of the study.
[Figure 1], [Figure 2], [Figure 3]
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5]