PEDIATRICS Vol. 105 No. 2 February 2000, pp. 354-358

Can Peak Expiratory Flow Predict Airflow Obstruction in Children With Asthma?

Nemr Eid, MD^*, Ben Yandell, PhD, Laura Howell, MSN, CPNP^*, Martha Eddy, MSN, CPNP^*, and Shahid Sheikh, MD^*

From the ^* Division of Pediatric Pulmonary Medicine, Kosair Children's Hospital/University of Louisville; and the  Department of Clinical Information Management, Alliant Health System, Louisville, Kentucky.

	ABSTRACT

Top Abstract Methods Results Discussion References

Study Objectives.  A recent trend in the treatment of asthma has been the widespread, independent use of peak expiratory flow (PEF). We examined whether PEF monitoring creates inaccuracies in assessment of children with moderate to severe asthma.

Methods.  We compared the negative predictive value of PEF in relation to the forced expiratory volume in 1 second (FEV₁), and to the forced expiratory flow between 25% and 75% of the vital capacity (FEF_25-75%) at different levels of air trapping as determined by the residual volume over total lung capacity ratio (RV/TLC).

Results.  The study included 244 patients, ages 4 to 18 years with all classes of asthma severity, with FEV₁ ranging from 28% to 134% of predicted value. We analyzed 367 sets of pulmonary function tests performed throughout a 3-year period. Thirty percent of patients with a normal PEF value had an abnormal FEV₁ or FEF_25-75%. As air trapping increased, the ability of a normal PEF to predict normal FEV₁ and FEF_25-75% readings fell from 83% to 53%. The negative predictive value was significantly lower for patients with RV/TLC ratio >30 compared with patients with RV/TLC <30.

Conclusions.  The results of this study suggest that it might be possible to identify children for whom the PEF is likely to give false-negative results. As air trapping increases, it causes the PEF to give misleading reassurance of normal pulmonary function. Furthermore, poor predictiveness of PEF is obtained when values 80% of predicted for age are considered normal.  Key words:  asthma, pulmonary function, air trapping, peak expiratory flow.

National and global asthma guidelines recommend objective measures of lung function such as spirometry and peak expiratory flow (PEF) to assess the severity of asthma and to evaluate the long-term response to therapy.^1,2 Because of its simplicity and easy availability, daily home PEF monitoring has been advocated for patients with moderate to severe asthma, to help them control their symptoms and to alert them of impending asthma exacerbations.^1,2 The correlation of home PEF monitoring with other objective measures of lung function such as the forced expiratory volume in 1 second (FEV₁), and the forced expiratory flow between 25% and 75% of the vital capacity (FEF_25-75%) has been the subject of ongoing debate. The advisability of home PEF monitoring, especially for children with asthma, has been challenged by many asthma experts.^3-13

PEF is the maximum flow rate generated at any time during a forceful exhalation. It usually occurs in the first 150 milliseconds of a forced expiratory maneuver.¹⁴ It reflects and measures only the rate of flow from the large airways, and is affected by the few hundred milliliters of expired air starting from full inflation of the lungs.¹⁴ The PEF is also affected by the strength of the thoracic and abdominal muscles, and the degree of muscular effort generated by the patient. PEF can be measured either by a pneumotachometer, which uses a transducer to convert flow to electric output during spirometry, or by simple potable flow gauge devices called peak flow meters. Even the PEF measured by spirometry is far less reproducible and seems to have a wider range of normal values than other parameters such as FEV₁ and FVC. Therefore, to be most meaningful, the PEF measured by portable devices should be correlated with the PEF measured by spirometry. The FEV₁ reflects large and medium-size airway function and is usually measured at lower lung volumes than that of the PEF. The FEF_25-75% reflects the function of small airways. These last 2 parameters are often referred to as effort-independent, because they occur late in the forced vital capacity maneuver, where airflow is mainly dependent on the elastic recoil pressure of the lung.

Objective measures of airflow, together with clinical manifestations, have been used successfully to classify the severity of asthma into 4 stages: intermittent, mild persistent, moderate persistent, and severe persistent.² Moderate-to-severe persistent asthma is characterized by airflow limitations: decreased FEV₁, FEF_25-75%, and PEF; and increased lung volumes: residual volume (RV), functional residual capacity (FRC), and total lung capacity (TLC).^15,16 Hyperinflation and air trapping are partly caused by partial obstruction of the airway, predominantly during expiration, and by decreased elastic properties of the lung.¹⁶ Hyperinflation is also affected by airway closure because of shortening of airway smooth muscles, and tonic activity of the respiratory muscles leading to inspiratory position of the chest wall.¹⁷ Children with chronic severe asthma who have hyperinflation and air trapping, have an increased incidence of irreversible lung damage, life-threatening exacerbations, and a higher incidence of steroid requirement.^1518-20 These patients are vulnerable to frequent asthma attacks and, therefore, in theory, are the best candidates for daily home PEF monitoring. However, if PEF measurements underestimate the degree of airflow obstruction, the measurements may lead to false reassurance and to catastrophic outcome in such patients.²¹

We hypothesized that in children with moderate-to-severe asthma characterized by air trapping (increased RV/TLC ratio), the PEF will underestimate the degree of airway obstruction, because the patient with air trapping may be able to generate a quick burst of air at the beginning of a forced exhalation, sufficient to give a normal PEF reading. We studied the negative predictive value (NPV) of the PEF in relation to FEV₁ and FEF_25-75% at different levels of air trapping, and we examined whether PEF monitoring creates serious inaccuracies in the assessment of an identified clinical population with asthma.

	METHODS

Top Abstract Methods Results Discussion References

We compared the NPV of PEF to FEV₁ and FEF_25-75% at different levels of RV/TLC in 244 patients. The mean age was 10.2 ± 3.6 years (range, 4-18 years). Patients were free of any symptoms at the time of the pulmonary function tests (PFTs). The study included 367 sets of PFTs throughout a 3-year period. There were 214 PFTs performed on outpatients for routine asthma monitoring, and 153 on inpatients just before hospital discharge when patients were thought to be free of any symptoms. All PFTs were performed at a children's hospital laboratory by the same registered respiratory therapist. Forced expiratory flows were measured by a Collins DS water-seal spirometer (Collins, Braintree, MA), according to American Thoracic Society standards.²² At least 3 forced vital capacity maneuvers were performed, and the best sum test for FEV₁ and FVC was selected by the computer. Two FVC values had to be within 5% of each other. For PEF and FEV_1, values of 80% of predicted or greater were classified as normal (ie, a negative test for poor pulmonary function) and values <80% of predicted were classified as abnormal (positive test). For FEF_25-75% values <65% of predicted were classified as abnormal.

Lung volumes were measured by a Collins BP-plethysmograph (Braintree, MA). Panting breathing movements against the shutter were made at a rate of 1 cycle/second. The cheeks and chins of each patient were firmly supported with both hands during the procedure. Multiple FRC determinations were made until three measurements were within 5% of each other. All measurements were reported with race adjustments according to American Thoracic Society guidelines.²² When a patient had >1 set of PFTs, we used the most recent set for the analysis with the following exception. Because we expected air trapping, as measured by RV/TLC ratio, to affect the predictive value of PEF readings, we wanted the most accurate and stable measure possible of RV/TLC. Accordingly, we used the mean RV/TLC value for patients with multiple readings. As it turns out, this precaution was unnecessary: mean RV/TLC reading correlated 0.94 with the patient's first RV/TLC and 0.96 with the most recent reading. Absence of air trapping was defined as RV/TLC ratio <25; RV/TLC values of 25 to 30 were considered mild air trapping, and values >30 were considered moderate to severe.

This study focuses on the NPV of PEF. The NPV looks only at patients with negative (ie, normal) results on a test and determines what percentage of those patients are, in fact, disease-free (as determined by some other method). It is the percentage of negative tests that indicates true negatives. The NPV was chosen because it looks at the use of a test from the clinician's perspective. A test may be a relatively good test in the sense of being sensitive and specific, yet still present a problem in interpretation for the clinician if its predictive value is low in the patient population. A test with a low NPV gives the clinician false reassurance about a patient's status. Unlike other measures of test validity such as sensitivity and specificity, the predictive value of a test is affected by the prevalence of disease in a given population. The NPV drops as the disease prevalence increases. The more prevalent the disease, the less believable a normal test result becomes. Even a test that is valid in a general population may not be valid in a sicker population.²³

	RESULTS

Top Abstract Methods Results Discussion References

Table 1 shows descriptive statistics for the sample, and a breakdown of the variables by asthma severity, as assessed by the patient's most recent FEV₁. As the minimum and maximum values demonstrate, the study included patients with all classes of asthma severity. Most recent readings ranged from 28% to 134%.

The PEF, FEV₁, and FEF_25-75% correlated fairly well with each other, with correlations ranging from 0.59 to 0.73 (data not shown). These 3 variables were inversely related to air trapping (RV/TLC): as air trapping increased, patients had significantly poorer readings on all 3 measures. White children showed a slight tendency to have better pulmonary function, with statistically significant differences for FEF_25-75% and RV/TLC.

Table 2 shows typical validity statistics assessing the ability of PEF to predict FEV₁ and FEF_25-75%. The top section of Table 2 analyzes all patients, and shows that PEF is a reasonable predictor of pulmonary function as measured by FEV₁ and FEF_25-75%. The findings in the "either" row of Table 2 means that PEF will detect most patients with abnormal pulmonary function (sensitivity 76%), and will classify most well patients correctly (specificity 77%)_. When PEF suggests that a patient is having airflow obstruction, the patient is having airflow obstruction (positive predictive value 81%), and the numeric value of PEF is a decent linear predictor of the numeric value of FEV₁ (reasonably strong r value).

The bottom 3 sections of Table 2 subdivide the patients by their degree of air trapping, and Fig 1 shows this finding graphically in 115 patients who have normal PEF values. These drops in NPV are statistically significant by 2-tailed Fisher's exact test (P = .02 for FEV₁ and P = .008 for FEF_25-75% using RV/TLC levels of 30 or more as the cutoff for moderate to severe air trapping).

View larger version (27K): [in this window] [in a new window]   Fig. 1.   Graph of negative predictive value of PEF at different residual volume/total lung capacity ratios. All patients had normal PEF readings.

The same decrease in the believability of a normal PEF value occurs if we classify patients by level of FEV₁ or FEF_25-75% rather than RV/TLC ratio. For example, only 50% of patients with normal PEF also had a normal FEF_25-75% if their FEV₁ was <90% of predicted (NPV of 50; 20 of 40 patients), whereas 88% of patients with a normal PEF also had a normal FEF_25-75% if their FEV₁ was 90% of predicted or greater (NPV of 88; 66 of 75 patients). Using FEF_25-75%, 75% of patients with a normal PEF also had a normal FEV₁ if their FEF _25-75% was 75% of predicted (NPV of 75; 30 of 40 patients), whereas 95% of patients with a normal PEF also had a normal FEV₁ if their FEF_25-75% was 75% of predicted or greater (NPV of 95; 71 of 75 patients). In other words, RV/TLC ratio, FEV_1, and FEF_25-75% are surrogates for poor lung function and seem to predict unsatisfactory NPVs for PEF. Stated differently, the more abnormal readings a patient has, the less trustworthy a normal reading becomes.

	DISCUSSION

Top Abstract Methods Results Discussion References

Multiple studies have shown that PEF can be a poor predictor of airway obstruction and have questioned the correlation between PEF, FEV₁, and FEF_25-75% in children with asthma.^3-13

Ferguson³ reported normal PEF values in asymptomatic children documented to have persistent airway obstruction by FEV₁ and FEF_25-75%. Indeed during the periods in which symptoms were absent and PEF was normal, 76% of these periods were associated with a decrease in FEF_25-75%.³ Similar findings were reported in 65 asymptomatic children with asthma seen at a general pediatric and pediatric pulmonary clinic.⁴ Although the PEF was normal, 17% of the children had FEV_{1 25-75%} <65% of the predicted values.⁴ Klein et al,⁵ found 18% of children with moderate to severe asthma attending a summer camp to have normal PEF but abnormal FEF_25-75%. The FEF_25-75% reduction was associated with significant clinical asthma even in the presence of a normal PEF.⁵ Similarly, in our study, it was typically the FEF_25-75% that indicated poor pulmonary function, but unlike the patients of Klein and colleagues,⁵ our patients and those of Ferguson³ and Bye and colleagues⁴ were asymptomatic at the time the tests were performed. Uwyyed et al⁶ found a poor correlation between home PEF and spirometry for all classes of asthma severity. Similar findings were reported by Sly et al,⁷ in 12 boys with asthma attending boarding school. Less than half of the clinically important reduction of lung function detected by spirometry were detected by 4 different types of home peak flow devices.⁷

Daily monitoring of PEF has not been shown to improve asthma illness score when compared with symptoms-only recognition plan.^10,11 In general clinical practice, rigid adherence to long-term daily PEF monitoring has been found to produce no changes in outcome.¹² Even twice daily PEF monitoring has been found to be of limited value for diagnostic purposes or for assessing the degree of asthma severity.¹³ When comparing asthma intervention which includes self-management principles and skills, education about medication use, and symptoms recognition, with or without PEF monitoring, two randomized, controlled studies have found significant improvement in lung function, symptoms, and medication use.^24-25 However, the relative importance of PEF monitoring to the overall effectiveness of the asthma intervention could not be determined from these studies. Only 1 study found that a peak flow-based action plan is effective in reducing emergency medical visits for asthma when compared with symptoms-based plan or to no action plan.²⁶ It should be mentioned however that those studies were done in adults, and there are no similar pediatric trials to date.

In our study, we found that PEF correlates well with FEV₁ in children with known asthma. Children with poor PEF readings are more likely than children with better PEF readings to have airway obstruction, and this pattern holds even in children with severe asthma. PEF seems to have shortcomings with this sample of patients when it comes to the NPV. Our primary concern is the believability of a normal test result (NPV). The lower the NPV is, the less believable a test becomes. For example, the NPV of 70 means that in fully 30% of the cases in which PEF was normal, either FEV₁ or FEF_25-75% was abnormal. Incidentally, these findings are not owing to having several almost well patients, because 17% of patients with normal PEF had either FEV₁ or FEF_25-75% values <60% of predicted, and 9% had values <50% of predicted (data not shown). Even after raising the PEF normal value cutoff to 90% of predicted, 25% of patients still had an abnormal FEV₁ or FEF_25-75%; raising the PEF cutoff would of course also increase the percentage of false-positives. The bottom 3 sections of Table 2 show that normal PEF results are fairly believable in patients with little air trapping, but become less believable as the degree of air trapping increases. Indeed, among children with increased air trapping (RV/TLC ratio >30), 47% of those with a normal PEF actually had obstructed airways as measured by FEV₁ or FEF_25-75% (NPV 53). The question is, then, how much faith should we place in normal PEF readings? And how can the clinician predict when PEF is likely to miss poor pulmonary function?

The results of this study suggest that it may be possible to identify children for whom the PEF is likely to give false-negative results. The degree of air trapping correlates well with disease severity. As air trapping increases, it causes the PEF to give misleading reassurance about pulmonary function, because the patient with air trapping is often able to produce a peak burst of exhaled air, sufficient to give a normal PEF value. However, as the expiration progresses, airway obstruction as measured by FEV₁ and FEF_25-75% becomes evident. A similar theory has been proposed by Sly,²⁷ in which the PEF is normal yet severe airway obstruction exists. He theorized that in such patients the transient " supramaximal" flow is because of explosive decompression of the gas in the central airway producing very high flow that cannot be sustained by the airway during forced expiration.²⁷

National and global asthma guidelines suggest the use of home PEF monitoring in patients with moderate to severe asthmaexactly the population that is less likely to benefit from such measurements according to our findings. Given the shortcomings of home peak flow meters, as shown above, and given the fact that even PEF measured by spirometry underestimates the degree of airway obstruction in moderate to severe asthmatics, routine use of home peak flow meters should be reviewed, and widespread independent measurements of PEF should be used with caution. An alternative is active in-clinic monitoring, with comprehensive tests of airflow obstruction, in children with moderate to severe asthma. In the future, hand-held spirometers may offer a viable alternative to the current peak flow meters.

Our work should not, however, be taken as an argument to stop using PEF home monitoring in the treatment of children with asthma for 2 reasons. First, PEF is a good test by several standards: it is sensitive, specific, and has good positive predictive value. It also has reasonable NPV for patients with little air trapping or with normal FEV₁ readings. Second, this study looks only at a point-in-time PEF, and does not examine the validity of using changes in the trend of PEF as a predictor of pulmonary function. It is rather interesting to note that in our study the poor NPV of PEF was obtained when values 80% of predicted for age were taken as normal. Perhaps, the newly revised asthma guidelines,²⁸ which recommend following the trend of the patient's personal best PEF determined throughout a 2- to 3-week period (not just the value predicted for age), may prove more beneficial in assessing early signs of airway obstruction. Nevertheless, long-term studies on outcome measures are sorely needed in this respect.

What our study does show is that clinicians caring for asthmatic children must be suspicious of normal PEF readings in patients with moderate to severe asthma. They should be aware that in this population, a normal PEF reading cannot be taken as reassurance that the patient has normal pulmonary function. Furthermore, the value of spirometry and lung volume measurement in detecting chronic airflow limitation even in asymptomatic children should not be underestimated. It is well known that persistent airway obstruction is present in asymptomatic children with asthma.^3,4 Chronic unrecognized airway obstruction can lead to progressive worsening asthma, hyperinflation, and chronic obstructive pulmonary disease later in life.^18-20 Our results confirm these observations in a rather dramatic way because all our patients were asymptomatic, but 43% of them still had FEV₁ <80% of predicted (Table 1). This finding is rather alarming as children with moderate to severe persistent asthma adapt to their chronic airflow limitation and have poor perception of their disease severity. Indeed, 1 study estimates that one third to one half of patients with asthma are unable to estimate airway obstruction by symptoms alone.¹⁴ Perhaps, more emphasis should be placed on the utility of monitoring spirometry and lung volumes in the overall treatment of children with asthma.

	ACKNOWLEDGMENTS

We thank Tom Spalding, RRT, CPFT, for performing all pulmonary function tests, and Bobbi Calloway for secretarial help.

	FOOTNOTES

Received for publication Mar 12, 1999; accepted Jun 10, 1999.

Presented in part at the Sixty-fourth Annual International Scientific Assembly, XIX World Congress on Diseases of the Chest, American College of Chest Physicians; November 10, 1998; Toronto, Canada.

Reprint requests to (N.E.) University of Louisville, Department of Pediatrics, Pediatric Pulmonary Medicine, 571 S Floyd St, Suite 414, Louisville, KY 40202. E-mail: nseid{at}louisville.edu

	ABBREVIATIONS

PEF, peak expiratory flow; FEV₁, forced expiratory volume in 1 second; FEF_25-75%, forced expiratory flow between 25% and 75% of the vital capacity; RV, residual volume; FRC, functional residual capacity; TLC, total lung capacity; NPV, negative predictive value; PFT, pulmonary function test.

	REFERENCES

Top Abstract Methods Results Discussion References

1. Expert Panel Report: Guidelines for the Diagnosis and Management of Asthma. Bethesda, MD: National Asthma Education Program, Office of Prevention, Education, and Control, National Heart, Lung and Blood Institute, National Institutes of Health; 1991. Publication No. 91-3042
2. Global Initiative for Asthma. Bethesda, MD: National Institutes of Health; 1995. Publication No. 95-3659
3. Ferguson AC Persisting airway obstruction in asymptomatic children with asthma with normal peak expiratory flow rates. J Allergy Clin Immunol. 1988; 82:19-22 [CrossRef][Medline]
4. Bye MR, Kerstein D, Barsh E The importance of spirometry in the assessment of childhood asthma. Am J Dis Child. 1992; 146:977-978 [Abstract]
5. Klein RB, Fritz GK, Yeung A, McQuaid EL, Mansell A Spirometric patterns in childhood asthma: peak flow compared with other indices. Pediatr Pulmonol. 1995; 20:372-379 [Medline]
6. Uwyyed K, Springer C, Avital A, Bar-Yishay E, Godfrey S Home recording of PEF in young asthmatics: does it contribute to management? Eur Respir J. 1996; 9:872-879 [Abstract]
7. Sly PD, Cahill P, Willet K, Burton P Accuracy of mini peak flow meters in indicating changes in lung function in children with asthma. Br Med J. 1994; 308:572-574 [Abstract/Free Full Text]
8. Clough JB, Sly PD Association between lower respiratory tract symptoms and falls in peak expiratory flow in children. Eur Respir J. 1995; 8:718-722 [Abstract]
9. Malo J-L, L'Archevêque RT, Trudeau RT, d'Aquino C, Cartier A Should we monitor peak expiratory flow rates or record symptoms with a simple diary in the management of asthma? J Allergy Clin Immunol. 1993; 91:702-709 [CrossRef][Medline]
10. Charlton I, Charlton G, Broomfield J, Mullee MA. Evaluation of peak flow and symptoms only self-management plans for control of asthma in general practice. Br Med J. 1990;1355-1359
11. Hammond N, Abdalla M, Beattie JAG, Effectiveness of routine self monitoring of peak flow in patients with asthma. Br Med J. 1994; 308:564-567 [Abstract/Free Full Text]
12. Jones KP, Mullee MA, Middleton M, Chapman E, Holgate ST, British Thoracic Society Research Committee Peak flow based asthma self-management: a randomized controlled study in general practice. Thorax. 1995; 50:851-857 [Abstract]
13. Linna OVE Twice-daily peak expiratory flow rate monitoring for the assessment of childhood asthma. Allergy Proc. 1993; 14:33-36 [Medline]
14. Li TCJ Home peak expiratory flow rate monitoring in patients with asthma. Mayo Clin Proc. 1995; 70:649-656 [Medline]
15. Liu AH, Brugman SM, Schaeffer EB, Irving CG Reduced lung elasticity may characterize children with severe asthma. Am Rev Respir Dis. 1990; 141:A906
16. Finucane KE, Colebath HJH Elastic behavior of the lung in patients with airway obstruction. J Appl Physiol. 1969; 26:330-338 [Free Full Text]
17. Muller N, Bryan AC, Zamel N Tonic inspiratory muscle activity as a cause of hyperinflation in asthma. J Appl Physiol. 1981; 50:279-282 [Abstract/Free Full Text]
18. Ostergaard PA Non-IgE-mediated asthma in children. Acta Paediatr Scand. 1985; 74:713-719 [Medline]
19. Kraemer R, Meister B, Schaad UB, Rossi E Reversibility of lung function abnormalities in children with perennial asthma. J Pediatr. 1983; 102:347-350 [Medline]
20. Greenough A, Everett L, Pool J, Price JF A 2-year longitudinal study of lung hyperinflation in young asthmatics. Respir Med. 1991; 85:379-382 [Medline]
21. Saetta M, Thienne G, Crescioli S, Fabbri LM Fatal asthma in a young patient with severe bronchial hyperresponsiveness but stable peak flow records. Eur Respir J. 1989; 2:1008-1012 [Abstract]
22. American Thoracic Society Medical Section of the American Lung Association. Standardization of spirometry 1994 update. Am R Respir Crit Care Med. 1994; 152:1107-1136
23. Lilienfield, DE, Stolley PD. Demographic studies: morbidity statistics. In: Foundations of Epidemiology. New York, NY: Oxford University Press; 1994:123-129
24. Woolcock AJ, Yan K, Salome CM Effects of therapy on bronchial hyperresponsiveness in the long-term management of asthma. Clin Allergy. 1988; 18:165-176 [CrossRef][Medline]
25. Ignacio-Garcia JM, Gonzales-Santos P Asthma self-management education program by home monitoring of peak expiratory flow. Am J Respir Crit Care Med. 1995; 151:353-359 [Abstract]
26. Cowie RL, Revitt SG, Underwood MF, Field SK The effects of peak flow-based action plan in the prevention of exacerbations of asthma. Chest. 1997; 112:1534-1538 [Abstract/Free Full Text]
27. Sly PD Peak expiratory flow monitoring in pediatric asthma: is there a role? J Asthma. 1996; 33:277-287 [Medline]
28. Highlights of the Expert Panel Report 2: Guidelines for the Diagnosis and Management of Asthma. Bethesda, MD: National Asthma Education and Prevention Program, National Institutes of Health, National Heart, Lung and Blood Institute; 1997. Publication No. 97-4051A