Isolated Reduction in Single-Breath Diffusing Capacity in the Evaluation of Exertional Dyspnea

Isolated Reduction in Single-Breath Diffusing Capacity in the Evaluation of Exertional Dyspnea* Zab Mohsenifar, M.D., F.C.C.P.; jack Collier; M.D.; MichaelJ Belman, M.D., F.C.C.P.; and Spencer K. Koerner; M.D., F.C.C.P.

Study SeHinga and lnteroentioru: Sixty individuals complaining oC dyspnea on exertion, but with normal spirometry and lung volumes and normal chest roentgenograms were reviewed for this study. These individuals were selected from a large group of outpatients (55! individuals over a seven-year period) who were referred to our laboratory for exercise testing to determine the cause of their exertional dyspnea. They were grouped aeconling to the single-breath diffusing capacity, with those less than 70 percent of predicted in the low Dco group (group I) and those greater than 70 percent of predicted in the normal Dco group (group !.). Both study groups underwent an incremental exercise test. Beaulta: Twenty-three individuals had a Dco less than 70 percent of predicted. During exercise, seven of these (30 percent) had an abnormal Pa01 and &ve had an abnormal P(A-a)01 • Thirty-seven people had a normal Dco. Thirtysix of these (97 percent) had a normal Pa01 and P(A-a)01 during exercise. Overall, eight individuals had an abnormal

The single-breath diffusing capacity for carbon monixide is considered a direct assessment of gas transfer across the lung. 1 A reduced Dco may be a physiologic manifestation of many pulmonary disorders. These include obstructive lung diseases, pulmonary vasoocclusive diseases, pulmonary vasculitis, and interstitial lung diseases. In addition, the Dco has been utilized to evaluate the severity of disease, and with serial testing, it has been employed as a marker of disease progression. 2 •3 Over the years, we have seen a large number of men and women with complaints of dyspnea on exertion. We saw 60 individuals with dyspnea on exertion who had normal spirometry, lung volumes, and normal diagnostic evaluations. Some of these individuals had a reduced Dco as their only abnormality and some had a normal Dco. The purpose of this study was to investigate the significance of reduced Dco with regard to the impact that it had on gas exchange at rest and during exercise.

Pa01 or P(A-a)O. during exercise; seven of these had an abnormally low Dco at rest. Concluaion3: Based on this selected group of a subpopulation, we conclude that the Dco is an important determinant of the diagnostic approach to a patient with dyspnea who is otherwise normal. If all pulmonary functions, including Dco are normal, an exercise study will fail to reveal abnormal Pa01 or P(A-a)01 in 97 percent of the cases. However, a low Dco has a poor predictive value with respect to abnormal gas exchange during exercise. Therefore, when investigating exertional dyspnea, based on this selected subpopulation, if the spirometry, lung volumes, and Dco are normal, one may forego additional invasive gas exchange evaluation. However, an abnormal Dco warrants further physiologic testing. (Cheat 199!; 101:965-69)

I

Dco=single breath diffusing capacity; VoNT=dead space/

_ tidal volume ratio; VE =minute ventilation

I _

MATERIAL AND METHODS

Subjects

From 1979 to 1986, 552 outpatients with dyspnea on exertion were referred to our laboratory for exercise testing to determine the etiology of their dyspnea. Sixty of these individuals who had normal spirometry, lung volumes, hematocrit values, and normal chest roentgenograms are the subject of this retrospective study. These individuals had undergone studies by their primary physicians to rule out cardiac diseases, including occult ischemic heart disease, and pulmonary diseases, including ventilation perfusion scans for pulmonary embolism when indicated and serologic tests for collagen vascular diseases. Patients with a history of cough, bronchial asthma, a bronchodilator response, history of any neuromuscular diseases, anemia, and/or obesity were excluded. Twentythree of these 60 individuals (18 men, 5 women) had a reduced Dco, and the remaining 37 (32 men, 5 women) had normal Dco values. The individuals were grouped according to the Dco, based on Sue et al, • with those less than 70 percent of predicted in the low Dco group (group 1), and those greater than 70 percent of predicted in the normal Dco group (group 2). These individuals had mild exertional dyspnea which is defined as dyspnea upon moderate or heavy exercise. Five individuals in gn>Up 1 and seven in group 2 had smoking histories of less than four pack-years; the remainder never smoked. Methods

*From the Division of Pulmonary Medicine, Department of Medicine, UCLA School of MediciDe, Cedars-Sinai Medical Center, Los Angeles. Manuscript received June 17; revision accepted August 2. Reprint requuts: Dr. Moluenifor, Dioiaion Of ftllmonary Medicine, Ctidars-SffllJf Medkal Center; lDs Angelea 90038

Slow vital capacity, FVC, and FEV, were measured in triplicate using a 12-L dry rolling seal spirometer. Predicted values for FVC and FEV, were calculated using the data of Schmidt et al. • Lung volumes were measured using the helium dilution technique,• and predicted values were calculated using the data of Goldman and CHEST I 101 I 4 I APRIL, 1992

-

Becklake,7 and Boren et al.• Arterial blood gases were measured using a blood gas analyzer; the P(A-a)02 was calculated using the equation: PaO

•

Paco. =PiO - - •

R

where Pa02 =alveolar oxyen tension, Pi02 =inspiratory oxygen tension, PaC02 =arterial carbon dioxide tension, and R =respiratory exchange quotient. The single-breath diffusing capacity was measured using a modified CPI with manual valve (CardioPulmonary Instruments) in the sitting position in duplicate, by the singlebreath method described by Ogilvie et al. • The highest of the two acceptable maneuvers which agreed within 5 percent were reported. linearity of the CO analyzer was checked weekly. Breathholding times during Dco determinations were close to 10 s and were measured by the method of Jones and Mead.•• Smokers were told to refrain from smoking 6 h prior to testing, and carboxyhemoglobin levels were measured in these patients. Predicted values for Dco were calculated using the data of Miller et al. 11 The measurement of lung function was performed with standard protocols, and the American Thoracic Society guidelines were used to determine acceptability. 12 All individuals underwent an incremental symptom-limited exercise testing on an electrically braked bicycle. After 2 to 5 min at rest and 2 to 4 min of pedaling without added resistance, the work rate was increased by a fixed amount between 10 to 30 W every minute until the subject no longer wished to continue. With the individual seated on a cycle ergometer, expiratory flow was measured by a Fleisch No. 3 pneumotachograph after passing through a two-way valve and integrated electrically by a respiratory integrator to obtain minute ventilation. The respired gas was then passed through a mixing chamber and was sampled by a mass spectrometer for a determination of oxygen and carbon dioxide concentrations, from which Vo2 , Vco2 , and the respiratory exchange ratio (R) were calculated. Dead space/tidal volume ratio was calculated using the modified Bohr equation. 13 During the final minute of exercise at each workload, exhaled gas was collected and arterial blood gases were simultaneously withdrawn for the determination of minute ventilation Vo2 and Vo/VT. Arterial oxygen tension during exercise was considered abnormal if it was below 80 mm Hg, P(A-a)O. was considered abnormal if it was greater than 35 mm Hg and Vo!VT during exercise was deemed abnormal if it was greater than 0.30.•• The results are expressed as means± SO. Analysis of variance was used to test for specific intergroup differences. For all comparisons, p
Table 2-Gaa &change Data at Rest and at ~ Variable

Group 1

Group2

No. oflndividuals Pa01 , mmHg Rest P(A-a)01 , mm Hg Rest Pa02 , mmHg Exercise P(A-a)02 , mm Hg Exercise RR, min Exercise VE,L Exercise Vo1 ,ml Exercise Vco1 , ml Exercise Vo!VT Exercise Heart rate, beats/min Rest Heart rate, beats/min Exercise

23 83±12

37 84±9

19±ll

15±7*

87±11

94±7*

28±13

20±6*

33±7

34±10

50±14

53±16

1210±308

1382±420*

1292±370

1495±492*

0.30±0.08

0.24±0.07*

79±12

84±14

131 ±21

139±25

*RR, respiratory rate during exercise; p70 percent predicted.

REsuLTS Anthropometric characteristics and results of spirometry, lung volumes, and single-breath Dco at rest for both groups are shown in Table 1. At rest, group 1 was similar to group 2 with respect to age, weight, FEV1, FVC, and TLC. The mean Dco in group 1 was

Table 1-Characteristica ofPatientB According to Singk-BretJth Diffusing Capacity Variable

No. of Individuals Age, yr Height, em Weight, kg Vital capacity, L (% predicted) FEV,,L (% predicted) FEV,IFVC,% MVV, L (% predicted) Total lung capacity, L (% predicted) Diffusing capacity, mVminlmm Hg (% predicted) Hematocrit, %

Group 1

Group2

23 37 51±14 52±12 171±10 171±6 76±10 78±9 4.03±0.8 4.05±0.6 (98±13) (102± 13) 3.13±0.8 3.17±0.5 (97±12) (103± ll) 79±6 79±6 137±41 136±28 (101 ± 17) (96±21) 5.80±0.8 5.90±0.8 (104± 14) (102± ll) 17.2±3.0 24.0±4.0 (59±9) (86±12) 45±3 45±3

*Group 1, Dco>70% predicted; group 2, Dco>70% predicted. MVV = Maximum voluntary ventilation

&erciae

•

110

~

101

~

100

-

• •• • •• • •• • ••• ~-· •• • ••• ••• • •• ••• ••• • • • • • ee e • • • •

•-

.

~ E 10 ~ E

~

• 70

••

•

I

~

80 75

..

~

-

I

30

•• • I

40

•• •

I

I

50 60 70 80 10 100 110 120 DLCO % PREDICTED

FIGURE 1. Comparison of diffusing capacity for carbon monoxide at rest (Dco, percent predicted) with arterial Po1 (PaOJ at peak exercise. Evaluallon of Exertlonal Dyapnea (~

et 81)

.10-

••• •

.40.35

•

•

f

~

• •

'

• •

•

• •

.41

•• •• • • • • •••• •• • •• ••

~.30 .25-

•

•

.20

.15-

•

.10I

30

40

50

80

70

80

10 100 110 120

•

• • • •• • •••• •• • • • • •• h •• ••••• • •••• • • ••• • •• ••••• • •• • • • •• I

30 40

DLCO " PREDICTED

I

50

I

80

70

80

I

•

•

• I

I

I

90 100 110 120

DLCO" PREDICTED

FIGURE

2. Comparison of diffusing capacity for carbon monoxide at rest (Dco, percent predicted) with alveolar-arterial gradient P(A-a)02 at peak exercise.

3. Comparison of diffusing capacity for carbon monoxide at rest (DGo, percent predicted) with deadspace tidal volume ratio (VDIVT) during exercise.

59± 9 percent of predicted, and in group 2, it was 86 ± 12 percent of predicted. Gas exchange data at rest and during exercise are shown in Table 2. Our results demonstrate a significant difference in Vo2 at peak exercise, exercise Pa02 , rest and exercise P(A-a)02 , and exercise VnNT between group 1 and group 2. In general, Vo2 was reduced in both groups. 14 However, both groups had significant cardiac and ventilatory reserves at maximum exercise, implying that the exercise study was symptom-limited and submaximal. As can be seen from Figure 1, seven of the individuals in the low Dco group had arterial hypoxemia during exercise while only one individual in the normal Dco group had exercise hypoxemia. Similarly, the P(A-a)02 at peak exercise as shown in Figure 2 was abnormally high in five individuals in the low Dco group and only in one individual in the normal Dco group. Figure 3 shows that an abnormal VnNT during exercise was observed in 9 of 23 individuals in group 1 and 10 of37 individuals in group 2. Sensitivity, specificity, positive predictive value, and negative predictive value oflow Dco in predicting an abnormal P(A-a)0 2 gradient, Pa02 , and VDNT during exercise are shown in Table 3.

Dco is normal, in the same clinical setting, gas exchange is normal both at rest and during exercise, obviating the need for invasive gas exchange analysis during exercise in most of such individuals. Since the classic studies of Krogh, 15 single-breath diffusing capacity has been used to detect abnormalities of gas transfer across the alveolar capillary membrane. This measurement has been used in physiologic studies of various lung diseases. A low Dco is often present in patients with emphysema, interstitial lung diseases, pulmonary vasoocclusive disease, and pulmonary vasculitis. 3 The 1986 American Thoracic Society statement on the evaluation of impairment/ disability secondary to occupational lung diseases has included Dco as a part of the initial evluation in order to rate the degree of impairment. 16 Many authors have addressed the predictive value of abnormal Dco at rest. For example, Sue et al4 Table 3-Ability of Single Bf'86th Di.lfuaing Capocity for Carbon MORO%ide to Predict an Abnormallb01 , P(A-a)01 and VD!VT During Exercile

DISCUSSION

In the setting of normal pulmonary function in individuals with dyspnea on exertion, an isolated reduction in Dco is a poor predictor of gas exchange abnormalities during exercise. However, when the

FIGURE

Sensitivity % Specificity % Positive predictive value% Negative predictive value%

Pa02 (Less Than 80mmHg)

P(A-a)02 (Greater Than 35mmHg)

VDIVT (Greater Than 0.30)

87

83 67 22

47

69 30 97

97

73

66

39

CHEST I 101 I 4 I APRIL, 1992

887

investigated the significance of Dco measurement among shipyard workers with an asbestos exposure history of 23 to 29 years. They found that six of the eight patients (75 percent) with low Dco (less than 70 percent of predicted) and normal FEV 1/FVC ratio had abnormal exercise variables. In contrast, in our study where we excluded any patients with cardiac or pulmonary diseases, only 7 of 23 individuals (30 percent) with low Dco had an abnormal gas exchange measurement. In the study conducted by Sue et al, 4 26 percent of patients with a normal Dco and a normal FEV1/FVC had an abnormal gas exchange factor. Conversely, in our study, only one of 37 individuals with a normal Dco and a normal FEV 1/FVC ratio had an abnormal Pa0 2 or P(A-a)02 gradjent during exercise. However, an abnormal VoNT was seen in 10 of 37 individuals. The reason for the discrepancy between our study and Sue et al4 most probably is explained by the difference in our patient populations. We were careful to exclude any patient with any pulmonary, neuromuscular disease, or cardiac disease, and they studied a cohort of workers exposed to heavy amounts of asbestos. Owens et al, 17 in a study of patients with chronic obstructive pulmonary disease, concluded that a diffusing capacity above 55 percent of predicted was 100 percent specific in excluding arterial desaturation. On the other hand, 19 of 28 patients (68 percent) with a diffusing capacity of 55 percent of predicted or less developed arterial desaturation with exercise. This study highlights the significance of a low Dco in predicting abnormalities during exercise. Their study, however, was conducted in patients with previously documented pulmonary disease, unlike our investigation. Other studies have also used the single-breath diffusing capacity for carbon monoxide to predict an abnormal gas exchange during exercise in workers exposed to asbestos and silica and patients with interstitial lung disease. 1s-20 To our knowledge, our study is the first to examine the significance of low Dco in individuals with dyspnea who are otherwise normal and have had no evident cardiopulmonary or neuromuscular diseases. This study does not provide an answer to the ultimate cause of dyspnea on exertion in our study population, and we have no follow-up information at this time. The presence of occult interstitial or vasoocclusive lung diseases cannot be ruled out based on the available data. However, our study provides insight into the importance oflow Dco with respect to gas exchange findings during exercise. It should be noted that Dco measurement is dependent on the pulmonary capillary blood volume and the alveolar capillary membrane of the lung. The accurate determination of this pulmonary function requires attention to various technical considerations. The accuracy of the breathholding test in measuring 968

the true overall pulmonary diffusion has been questioned. Although available evidence suggests that it is reasonably precise in normal individuals, any inaccuracy in this measurement might increase the variability of the test beyond its true physiologic limits. In addition, single-breath Dco has been found to be dependent upon alveolar oxygen tension, hemoglobin concentration, and the degree of lung inflation at the time of breathholding. 21 •22 These factors cannot be completely controlled in the clinical application of this test; however, in our study, all of our subjects had normal lung volumes and normal hemoglobins. The use of a particular predicted value for singlebreath diffusing capacity for carbon monoxide may affect any discussion regarding the importance of the diffusing capacity. We used the predicted equation of Miller et al 11 for nonsmokers. The values in this predicted equation are slightly higher than the values ofCotes, 23 and slightly lower than the values of Crapo and Morris. 24 If we used Crapo and Morris's predicted equation, the number of individuals with a diffusing capacity less than 70 percent of predicted would increase by six individuals, increasing our sensitivity and negative predictive value to 100 percent, with a corresponding decrease in specificity and positive predictive value. If we used the predictive equation of Cotes, 23 our results would remain unchanged and the sensitivity, specificity, and predictive values would not change. There were a number of individuals in our study with an abnormal Dco performed at rest who had normal Pa02 and P(A-a)02 during exercise. We propose the following morphologic model to explain these different responses. In these individuals, the pulmonary capillaries were probably still largely intact, and therefore, capable of being recruited during exercise, but relatively underperfused at rest, resulting in a low Dco. We also saw individuals who had an abnormal Dco and normal VoNT during exercise. It might be argued that a normal VoNT in these individuals during exercise could have occurred as a consequence of a marked increase in tidal volume in the presence of a large fixed anatomic dead space. Mohsenifar et al• and Overland et al116 have previously reported patients with collagen vascular diseases and drug abusers with a moderate reduction in Dco who had a normal VoNT at rest, as well as during exercise. In the present study, there were ten individuals who had a normal Dco and an abnormal VoNT during exercise. Abnormal VoNT during exercise cannot be ascribed to low tidal volume since the mean tidal volume in both groups exceeded 1,350 ml. 27 It is possible that in these individuals, the increase in pulmonary perfusion during exercise occurred nonuniformly, creating high ventilation/perfusion regions, which would result in a high VoNT. Evaluallon of Exertionll Dyapnaa (~eta/)

In summary, single-breath Dco was helpful in assessing individuals with complaints of exertional dyspnea. Diffusing capacity for carbon monoxide of greater than 70 percent predicted was associated with a normal Pa02 and a normal alveolar-arterial difference in Pa02 in 97 percent of the cases. However, an abnormal diffusing capacity for carbon monoxide had a very poor predictive value, associated with an abnormal Pa02 and alveolar-arterial gradient during exercise in only 30 percent of the cases. Overall, seven of eight individuals with an abnormal Pa02 during exercise had a low Dco. Based on the data from our subpopulation, we conclude that a normal singlebreath Dco is a good predictor of a normal Pa02 and P(A-a)02 , such that most subjects with a normal Dco will have a normal Pa02 during exercise. Conversely, an abnormal Dco is a poor predictor of abnormal gas exchange during exercise. Therefore, in the assessment of dyspnea, based on this selected subpopulation, if the spirometry, lung volumes, and single-breath Dco are normal, one may forego additional invasive gas exchange evaluation. However, an abnormal Dco in a individual with complaints of dyspnea warrants additional physiologic investigation. Follow-up studies in individuals with isolated low Dco and exertional dyspnea will elucidate true structure-function correlation in such individuals. ACKNOWLEDGMENT: The authors wish to thank Professor Janet ElashoiJ for statistical comments, and Debra Craig for secretarial assistance.

REFERENCES 1 Beck K, Hyatt R, Staats B, Enright P, Hodarte J. Carbon monoxide diiJusing capacity of the lungs determined by singlebreath and steady-state exercise methods. Mayo Clio Proc 1989; 64:51-59 2 Crystal R, Fulmer J, Roberts W, Moss M, Line B, Reynolds H. Idiopathic pulmonary fibrosis: clinical, histologic, radiographic, physiologic, scintigraphic, cytologic and biochemical aspects. Ann Intern Med 1976; 85:769-88 3 Burrows B, Kasik J, Niden A, Barclay W. Clinical usefulness of the single-breath pulmonary diiJusing capacity test. Am Rev Respir Dis 1961; 84:789-806 4 Sue D, Oren A, Hansen J, Wasserman K. Diffusing capacity as a predictor of gas exchange during exercise. N Engl J Med 1987; 316:1301-06 5 Schmidt C, Dickman M, Gardner R, Brough F. Spirometric standards for healthy elderly men and women. Am Rev Respir Dis 1973; 108:933-39 6 McMichael J. A rapid method of determining lung capacity. Clio Sci 1939; 4:167-73 7 Goldman H, Becklake M. Respiratory function tests: normal values at median altitude and the prediction of normal results. Am Rev Tuberc Pulmon Dis 1959; 79:457-67 8 Boren H, Kory R, Syner J. The Veterans administration-army cooperative study of pulmonary function: II. The lung volume and its subdivisions in normal man. Am J Med 1966; 41:96-114

9 Ogilvie C, Forster R, Blakemore W, Morton J. A standardized breathbolding technique for the clinical measurement of the diiJusing capacity of the lung for carbon monoxide. J Clio Invest 1957; 36:1-17 10 Jones R, Mead F. A theoretical and experimental analysis of anomalies in the estimation of pulmonary diffusing capacity by the single-breath method. QJ Exp Physiol1961; 46:131-43 11 Miller A, Thornton J, Warshaw R, Anderson H, 1ierstein A, SelikoiJ I. Single-breath diftUsing capacity in a representative sample of the population of Michigan, a large industrial state: predicted values, lower limits of normal, and frequencies of abnormality by smoking history. Am Rev Respir Dis 1983; 127: 9:10-71 12 American Thoracic Society: Epidemiology standardization project. Am Rev Respir Dis 1978; 118:1-20 13 Comroe J, Forster R, Dubois A, Brisco W, Carlsen E. Physiologic dead space, in the lung. Chicago: Year Book Medical Publishers, 1962 14 Hansen J, Sue D, Wasserman K. Predicted values for clinical exercise testing. Am Rev Respir Dis 1984; 129(suppl):49-55 15 Krogh M. The diiJusion of gases through the lungs of man. J Physiol1915; 36:1-17 16 Evaluation of impairment/disability secondary to respiratory disorders. Am Rev Respir Dis 1986; 133:1205-09 17 Owens G, &gers R, Pennock B, Levin D. The diffusing capacity as a predictor of arterial oxygen desaturation during exercise in patients with chronic obstructive pulmonary disease. N Engl J Med 1984; 310:1218-21 18 Ward J, Harper P, Popovich J. Diffusing capacity predicts arterial oxygen desaturation in patients with lung diseases of multiple etiologies. Chest 1986; 89:514 19 Garcia JGN, Griftith DE, Williams JS, Blevins Jw. Kronenberg RS. Reduced diffusing capacity as an isolated finding in asbestos and silica exposed workers. Chest 1990; 98:105-11 20 Schwartz D, Galvin J, Dayton C, Standford W, Merchant J, Hunningbake G. Determinants of restrictive lung function in asbestos-induced pleural fibrosis. J Appl Physiol1990; 68:193237 21 FreyTM, Crapo RO, Jensen RL, Kanner RE, Kass JE, Castriotta RJ, et al. Adjustment of Dco for varying COHb, and alveolar P01 using a theoretical adjustment equation. Respir Physiol 1990; 81:303-12 22 American Thoracic Society Statement. Single-breath carbon monoxide diiJusing capacity (transfer factor): recommendation for a standard technique. Am Rev Respir Dis 1987; 136:1299'Jm 23 Cotes J. Lung function: assessment and application in medicine. 3rd ed. Oxford: Blackwell, 1975 24 Crapo R, Morris A. Standardization single-breath normal values for carbon monoxide diiJusing capacity. Am Rev Respir Dis 1981; 123:185-89 25 Mobsenifar Z, 'Illsbkin D, Levy S, Bjerke R, Clements P, Furst D. Lack of sensitivity of measurements of VDIVT at rest and during exercise in detection of hemodynamically significant pulmonary vascular abnormalities in collagen vascular disease. Am Rev Respir Dis 1981; 123:508-12 26 Overland ES, Nolan AJ, Hopewell PC. Alterations of pulmonary function in intravenous drug abusers, prevalence, severity and characterization of gas exchange abnormalities. Am J Med 1980; 68:231-37 27 Mobsenifar Z, Brown UV, Koerner SK. EIJect of breathing pattern on deadspace ventilation, VDIVT during exercise. Respiration 1985; 47:232-36

CHEST I 101 I 4 I APRIL, 1992

881