Psychological Profile and Ventilatory Response to ... - ATS Journals

2 downloads 0 Views 151KB Size Report
and cognitive abnormality model of hypochondriasis: amplification and physiological reactivity in women. Psychosom. Med. 58:423–431. 22. Brack, T., A. Jubran ...
Psychological Profile and Ventilatory Response to Inspiratory Resistive Loading MARC H. LAVIETES, CARLOS W. SANCHEZ, LANA A. TIERSKY, NEIL S. CHERNIACK, and BENJAMIN H. NATELSON Departments of Medicine (Pulmonary) and Neurosciences and the Chronic Fatigue Syndrome Center, University of Medicine and Dentistry of New Jersey, New Jersey Medical School, Newark, New Jersey

The purpose of this study was to explore the contribution of psychological state to both the ventilatory response and the intensity of dyspnea experienced after the addition of small inspiratory loads to breathing. We hypothesized that patients with either a specific psychiatric diagnosis or a specific psychological trait will associate a greater degree of dyspnea with a loaded breathing task than will control subjects. To insure the inclusion of persons with relevant psychological profiles, we recruited both subjects enrolled in the Chronic Fatigue Center and normal control subjects. In all, 52 subjects inspired first through a small (1.34 cm H2O/L/s) and second through a moderate (3.54 cm H2O/L/s) inspiratory resistive load (IRL). Ventilation was monitored throughout the 5-min sessions. Dyspnea was quantified with the Borg scale at specified times during the protocol. Standard psychological tests were administered. We found that subjects could be divided into two groups. One, the “responders,” reported Borg scores higher than those of the second, or “nonresponder” group, at all times during the protocol. By contrast, there was no difference between groups with respect to ventilation. Responders had higher scores on tests of depression (the Center for Epidemiological Study depression scale) than did nonresponders. We conclude that the variability observed in subjective responses to IRL is explained, in part, by differences in psychological state. Lavietes MH, Sanchez CW, Tiersky LA, Cherniack NS, Natelson BH. Psychological profile and ventilatory response to inspiratory resistive loading. AM J RESPIR CRIT CARE MED 2000;161:737–744.

Physicians are frequently confronted with patients for whom the distinction between a subtle presentation of an organic illness and an exaggerated verbal response to a minor somatic stimulus is unclear. Somatization is said to occur when affective or other benign impulses associated with psychological distress or with normal physiologic function are misinterpreted as symptoms of physical disease (1). Symptom amplification differs from somatization only in that the former implies the exaggeration (rather than the absence) of any underlying pathophysiological basis for a subjective complaint. Moreover, neuroticism and/or psychological distress are associated with excessive complaining or symptom amplification (2). The study of these conditions is important because somaticizers and symptom amplifiers frequently seek medical care. They are not easily recognized by physicians. As a result, they are subjected to extensive testing, all with added costs and risks. Loaded breathing (breathing through a circuit designed to impede inspiration, expiration, or both) may be a helpful tool to study symptom amplification. Most people, when confronted with a moderate inspiratory resistive load (IRL), ad(Received in original form October 20, 1998 and in revised form July 20, 1999 ) Supported in part by Center Grant No. U01A1-32247 from the National Institutes of Health. Correspondence and requests for reprints should be addressed to Marc H. Lavietes, M.D., University Hospital #I354, 100 Bergen Street, Newark, NJ 071032406. Am J Respir Crit Care Med Vol 161. pp 737–744, 2000 Internet address: www.atsjournals.org

just to the load and maintain a constant ventilation. By contrast, the degree to which subjects report dyspnea associated with a given inspiratory load can be highly variable. Any differences in the perception of dyspnea associated with loading between subjects or groups, therefore, would reflect differences in their subjective responses to the load. A subject who associates a high grade of dyspnea with the introduction of a small inspiratory load to the airway could be considered a “symptom amplifier,” at least with respect to the symptom of dyspnea. Although the ventilatory response to loads is governed primarily by respiratory system mechanics and by reflexes, behavioral or cognitive factors may play a role in this response as well (3, 4). We have studied the subjective and objective responses of persons free from organic disease to loading in two groups: (1) subjects with well-defined chronic fatigue syndrome (CFS); (2) age- and sex-matched control subjects. We chose subjects with CFS because somatization is thought to explain, in part, the symptoms of CFS (5, 6). To assess subjective responses, we used the Borg scale to obtain personal estimates of the magnitude of dyspnea experienced during loading. Measurements of ventilation, e.g., tidal volume, breathing frequency, and minute ventilation, served as measures of the objective responses to loading. Each subject received an extensive psychological assessment. The goal of this study was twofold: first, to identify subjects with an exaggerated subjective response to loading, and, second, to explore the relationships between emotional or psychological state, symptom amplification, and the ventilatory responses to loading in our subjects. We hypothesized

738

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

that quantitative assessment of dyspnea associated with IRL may differ between individual subjects; these differences may be explained, in part, by differences in psychological state or trait.

METHODS Subjects Subjects were recruited from the Chronic Fatigue Syndrome Center, Newark, New Jersey. Fifty-two subjects, 29 with CFS and 23 without CFS matched for age, sex, and socioeconomic background, were entered into the study. CFS subjects met established criteria for this diagnosis (7). Normal subjects free from psychiatric, medical, or surgical disease were recruited by advertisements directed to the general public as non-CFS control subjects and were paid for their time. All gave informed consent.

Methods Screening pulmonary function testing, which included VC and FEV1 (as %VC), was performed with a standard testing system (DSII plus; Warren E. Collins, Braintree, MA).

Loaded Breathing Subjects sat in a comfortable chair, arms supported by arm rests. They inspired from a mouthpiece, wearing noseclips, from an open circuit. End-tidal CO2 was sampled continuously from the mouthpiece (Beckman LB-2 infrared analyzer; Beckman Instruments, Fullerton, CA). Throughout the loaded breathing trials, inspiratory flow was measured by a pneumotachygraph (Warren E. Collins) attached to a pressure transducer and amplifier (Validyne, Northridge, CA). The analog output from the transducer was digitized at 20 Hz (Coulbourn “Lablinc,” Allentown, PA) and processed by computer (Northgate 386). Commercially available programs (Codas; Asystant) were used to display the flow signal on the oscilloscope, integrate flow, and calculate respiratory volumes and frequencies.

Subjective Assessment of Dyspnea The modified Borg scale was used to assess dyspnea in a quantitative manner throughout the loaded breathing trials (8). This scale combines properties of both categorical and interval scales and is thus an ideal tool for the quantitative assessment of dyspnea. Subjects were asked to scale “discomfort with breathing” or “shortness of breath.” Care was taken to avoid the term “effort” when instructing subjects (9). Subjects were asked to think of zero as “no discomfort whatsoever” and 10 as “the most discomfort or shortness of breath you have ever associated with breathing”; subjects were to grade their sense of discomfort at various intervals throughout the protocol.

Psychological Assessment Subjects underwent a clinical psychiatric interview using the computerized Quick Diagnostic Interview schedule (Q-DIS) (10). Both a standardized measure of depressive symptoms (the Center for Epidemiological Study depression scale, or CESD) and the Neuroticism subscale of the NEO Personality Inventory-Revised (NEO PI-R)

Protocol Subjects first underwent screening spirometry and psychological assessment. Each was then asked to assess his or her degree of dyspnea using the Borg scale before using the mouthpiece. Subjects next placed the mouthpieces in their mouths and breathed quietly through them until airflow and end-tidal PCO2 tracings were stable on the oscilloscope. This occurred within 10 min. Subjects again assessed their degree of dyspnea by the Borg scale. A small resistive load (1.34 cm H2O/L/s measured at 1.0 L/s) was then surreptitiously added to the inspiratory side of the breathing circuit. After 30 s, dyspnea was again rated by the Borg scale. For the first 27 subjects, the loaded breathing period was extended for 5 min. After the fifth minute, each subject gave a final estimate of his or her degree of dyspnea. After a brief rest period, these same subjects repeated the entire protocol; this time, a larger load (3.54 cm H2O/L/s measured at 1.0 L/s) was used. The loaded breathing segment of the protocol was carried out in a “single blind” fashion in that the experimenters performing the tests and analyzing the data did not know which subjects belonged to the CFS and which to the non-CFS groups.

Data Collection and Analysis · For all subjects, ventilation (V E) was calculated from the last three breaths prior to loading and again from the first breath after the introduction of the load. End-tidal PCO2 was measured simultaneously; Borg scores were taken before mouthpiece breathing, during the mouthpiece breathing, but before loading, and 30 s after introduction of the load. For the 27 subjects who participated in the entire protocol, addi· tional measurements of V E were made at the following intervals: 20, 40, 60, and 90 s, and 2, 3, 4, and 5 min. In addition, for each of these time periods the following data were computed: tidal volume (VT); inspiratory time (TI); inspiratory flow (VT/TI); and breathing frequency (f). Dyspnea was evaluated at the fifth minute. Data analysis was performed with a standard statistical package (SAS; Cary, NC). For the repeated measures analyses, a two-way ANOVA with repeated measures model was used. Standard t tests or nonparametric tests were used when appropriate. Grouping of Data for Analysis We first compared data from CFS with that from non-CFS subjects. We next identified from among the 52 subjects, 23 who reported a higher Borg score within 30 s after introduction of the small load; these subjects were defined as “responders.” The other 29, who did not report an increased Borg score after introduction of the small load were called “nonresponders.” Of the 27 subjects who participated in the extended loading protocol, 10 were responders and 17 were nonresponders. Because CFS subjects were distributed evenly between the groups, the data analysis consisted of comparisons between responder (R) and nonresponder (NR) groups rather than CFS and non-CFS subjects. Although sex differences may play a role in determining differences in responses to inspiratory loading, CFS subjects are predominately female (13). Of the 52 subjects studied, there were seven men. Five were identified as CFS (versus two as non-CFS); similarly, five

TABLE 1

Responder Nonresponder

n (M/F )

Age (yr)

FVC (% pred )

FEV1/FVC (%)

Height (cm)

Weight (kg)

23 (5/18) 29 (2/27)

37 ⫾ 9 39 ⫾ 9

94 ⫾ 12 93 ⫾ 11

82 ⫾ 5 82 ⫾ 6

168 ⫾ 10 164 ⫾ 9

72 ⫾ 21 69 ⫾ 20

* Values are means ⫾ SD.

2000

were administered (11, 12). Subjects were also asked whether or not they had ever been given a standard psychiatric diagnosis as defined by the DSM-III.

GROUP ANTHROPOMETRIC AND LUNG FUNCTION DATA Group

VOL 161

739

Lavietes, Sanchez, Tiersky, et al.: Psychological Profile and Loading

were responders (versus two nonresponders). Therefore, we made no effort to analyze male and female subjects separately.

RESULTS CFS versus Non-CFS Subjects

For CFS subjects, the mean Borg score prior to introduction of the small load was 1.24 Borg scale units; for the non-CFS, 1.04. For both of the groups, the median was 0.5; the lower and higher quartiles were 0 and 2, respectively. Immediately after loading, the mean Borg for the CFS group increased to 1.71; for the non-CFS group it was 1.67. For both groups, the median Borg response after loading was 1.0; quartiles were 0 and 3. There were no statistical differences between data for both groups prior to loading. Neither group achieved a statistically significant increase in the Borg response after the load. · VE for the CFS group prior to loading, 12.9 ⫾ 0.65 (SEM) did not differ from that of the non-CFS group, 11.4 ⫾ 3.9. No change in ventilation was seen in either group immediately after loading. By contrast, differences were found between groups with respect to their psychological profile. CFS subjects scored 76 ⫾ 33 (SD) on the neuroticism scale; non-CFS subjects, 54 ⫾ 19 (t ⫽ 2.83; p ⫽ 0.007). Similarly, CFS subjects scored higher on the CESD depression scale (22 ⫾ 14 SD) than did non-CFS subjects (5 ⫾ 4; t ⫽ 6.35, p ⭐ 0.001). Responder versus Nonresponder Subjects

Anthropometric and lung function data for these groups appear in Table 1. There were no differences between groups with respect to age, height, weight, or lung function. · Pooled data for VE, Borg dyspnea rating, and PCO2 taken both before introduction of the small load (but while breathing via the mouthpiece) and again immediately after loading · appear in Figure 1. VE prior to loading was 11.9 ⫾ 0.68 (SEM) L/min for the 23 responder (R) group subjects and 12.1 ⫾ 0.77 for the 29 nonresponder (NR) group subjects. Although ventilation increases, on average, immediately after loading in R but not in NR group subjects, the difference between the groups is not significant (p ⫽ 0.09; Wilcoxon’s rank sum test). The Borg score before loading was 1.39 ⫾ 0.30 for the R group and 0.97 ⫾ 0.25 for the NR group. Immediately after loading, the Borg score increased in R, but not in NR, subjects. This was expected since the assignment of subjects to either group was made on the basis of whether the subjects increased (R) or did not increase (NR) their estimates of their degree of dyspnea immediately after loading. PCO2 was 35 ⫾ 0.9 for R subjects before loading and 35 ⫾ 0.8 for NR subjects before loading. PCO2 did not differ between groups prior to loading; nor did it change with the introduction of the load. Data for Borg scale responses without and again with the mouthpiece (but prior to loading) appear in Table 2. In this table, the Borg scale of numerical responses appears in the lefthand column. Data for the Borg responses of NR subjects

Figure 1. Ventilation (upper panel ), Borg scale response (middle panel ), and end-tidal PCO2 (lower panel ) before and after the addition of a small resistive load (1.34 cm H2O/L/s measured at 1.0 L/s) to the inspiratory side of an open breathing circuit. All measurements were made while subjects breathed through a mouthpiece

(MPC). Ventilation was computed by extrapolation of the tidal volume and breath duration of the first breath immediately after loading to 1 min. End-tidal PCO2 was that of the first breath after loading. The Borg scale response was elicited 30 s after institution of the load. Data are shown as a mean and standard deviation. Responders ⫽ subjects whose Borg scale score increased after introduction of the 1.34 cm H2O load. Nonresponders ⫽ those whose Borg scale scores did not increase.

740

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

VOL 161

2000

TABLE 2 BORG SCALE WITH AND WITHOUT MOUTHPIECE Borg Scale

NR w/o

NR w

R w/o

Rw

10 · · 5 4 3 2 1 0.5 0

— — — — 0 1 0 1 2 25

— — — — 3 1 3 3 4 15

— — — — 0 3 1 0 0 19

— — — — 2 5 2 3 5 6

Total

29

29

23

23

Definition of abbreviations: NR ⫽ nonresponders; R ⫽ responders; w/o ⫽ without mouthpiece; w ⫽ with mouthpiece.

breathing without (w/o) the mouthpiece appear in the second column from the left. Note that 25 of 29 subjects reported zero, or no dyspnea, with breathing unencumbered by the mouthpiece, whereas only one of 29 described moderate dyspnea (a Borg score of 3). Similarly, the column labeled R w/o shows that 19 of 23 R group subjects reported zero Borg scores during nonmouthpiece breathing, whereas three responded with Borg scores of 3. There was no difference between these two groups. Comparison of R group subjects’ estimates of their dyspnea during unloaded mouthpiece breathing (column R w) with their dyspnea prior to introduction of the mouthpiece (R w/o) shows an increase in the subjective estimate of dyspnea induced by the mouthpiece (p ⫽ 0.001, Wilcoxon’s rank sum test). The mouthpiece induced an increase in the dyspnea reported by the NR group as well (p ⫽ 0.0037). · VE and Borg scores for both the 10 R group and 17 NR group subjects who· participated in the extended protocol appear in Figure 2. VE prior to introduction of the small load was 13.3 ⫾ 3.6 (SD) L/min in the R group and 12.5 ⫾ 4.6 in the NR group. Two-way ANOVA with repeated measures showed neither group (F ⫽ 2.28; p ⫽ 0.11), time (F ⫽ 1.12), nor interactive (F ⫽ 1.04) effects. The Borg scale score for R group subjects breathing with no mouthpiece was 0.6 ⫾ 1.3; for the NR group it was 0.06 ⫾ 0.2. In Figure 2, the Borg score is shown to have increased progressively for the R but not for the NR group. Formal analysis showed group (F ⫽ 9.25; p ⫽ 0.006), time (F ⫽ 28.39; p ⫽ 0.0001), and interactive (F ⫽ 8.38; p ⫽ 0.0001) effects. This analysis thus identified differences between the means of all data collected from R group subjects when compared with NR subjects (group effect) as well as between the means for all data (R and NR subjects combined) when compared at different time intervals (time effect). The relationship of R subjects with time differed from that of NR subjects with time as well (interactive effect). · VE and Borg scores for nine of the 10 R and 13 of the 17 NR group subjects breathing for 5 min on the larger load appear in Figure 3. Analysis of data from serial measurements of · VE shows a time effect (F ⫽ 4.66; p ⫽ 0.0001) but no group (F ⫽ 0.79) or interactive (F ⫽ 0.69) effects when the larger rather than the smaller load was used. By contrast, analysis of serial Borg scale responses for the two groups during the trial using the larger load continued to show group (F ⫽ 8.02; p ⫽ 0.01), time (F ⫽ 43.6; p ⫽ 0.0001), and interactive (F ⫽ 2.65; p ⫽ 0.056) effects. Perusal of individual data showed some overlap between groups, however. For example, Borg scores given by the nine R group subjects 30 s after the introduction of the large load ranged from 3 to 9; for the 13 NR subjects it

Figure 2. Ventilation (upper panel) and Borg score response (lower panel), both measured at discrete intervals over a 5-min period after introduction of the 1.34 cm H2O load. Ventilation was measured four times during the first minute after loading, during the first breath, and again after 20, 40, and 60 s. Data are given as mean ⫾ SD.

ranged from zero to 5. Further inspection of Figures 2 and 3 show that although the R group reported more dyspnea than did the NR group at all times throughout both trials, it is also true that both R and NR group subjects reported more dyspnea breathing through the larger (as compared with the smaller) load. For example, the mean Borg score for R group subjects after breathing for 5 min on the large load was 4.5 ⫾ 2.4; after 5 min on the small load it was 3.2 ⫾ 1.7. This difference was statistically significant at the 0.006 significance level (paired t ⫽ 3.21). Similarly, for the NR group, the mean of Borg scores after 5 min breathing through the large load was greater than the mean Borg score, taken at the corresponding time, while subjects breathed through the smaller load (paired t ⫽ 3.14; p ⫽ 0.004). Spirograms illustrating the breathing patterns for R and NR subjects before, during the first breath, after loading and in the fifth minute after loading appear for the small (Figure 4) and the large (Figure 5) loads. Analysis showed that inspiratory flow increased with time when subjects breathed

741

Lavietes, Sanchez, Tiersky, et al.: Psychological Profile and Loading

Figure 4. Spirograms depicting the ventilatory pattern while breathing through the mouthpiece before loading (control), during the first breath after loading, and after 5 min of breathing from the 1.34 cm H2O load.

Figure 3. Ventilation and Borg scale responses after introduction of a 3.54 cm H2O load. The format of this illustration is identical to that of Figure 2.

through the smaller load (F ⫽ 2.24; p ⫽ 0.021). Inspiratory flow, inspiratory time, and tidal volume all increased with respect to time when subjects breathed through the larger load. No group or interactive effects were seen, however. Data Pertaining to Psychological State

Complete data for psychological testing was available for 46 subjects. We found a correlation between the Borg score selected by each subject to characterize his sense of dyspnea during mouthpiece breathing but before introduction of the load and his score on the neuroticism measure (p ⫽ 0.029; Spearman’s rank correlation test). Thus, the addition of a mouthpiece to the breathing cirucuit is, in itself, a small stressor. As shown in Figure 6, a higher prevalence of psychopathology was found in the R group. Specifically, 10 of 22, or 45%, of the R group subjects met the DSM III-R diagnostic criteria for a psychiatric disorder, whereas only four of 24, or 17%, of the nonresponder subjects met such criteria. Finally, among the 46 subjects tested, R group subjects had higher depression scores than did NR subjects (t ⫽ 2.04; p ⫽ 0.049). This was so because the CESD score for R group CFS subjects (30 ⫾ 14 SD) was greater than the CESD score for the NR CFS subjects (15 ⫾ 10; t ⫽ 2.65; p ⬍ 0.02). By contrast,

there was no difference between the CESD scores of the nonCFS R and NR groups (Figure 7).

DISCUSSION It is well known that healthy persons vary in the degree of dyspnea they experience with added inspiratory loads. This study has shown first that, if subjects free from lung disease are presented with small, barely perceptible, inspiratory loads, some, whom we have labeled “responders,” will report increased dyspnea in response to the load, whereas others, the “nonresponders,” will not. Some NR subjects will not maintain their ventilation. Second, when a larger, more easily perceived, load is introduced, these same responders on the smaller load will report more dyspnea when compared with the same nonresponder subjects. With the larger load however, both R and NR subjects maintain ventilation. The fact that the same subjects (R group) report more dyspnea with both loads attests to the reproducibility of these findings. Furthermore, study of the spirograms shown in Figures 4 and 5 reveals that, at least for the larger load, the pattern of ventilation was similar for the two groups during the 5-min trial. Therefore, any differences in the subjective assessment of dyspnea during that trial could not be attributed to any differences in the ventilatory patterns between the R and NR groups. The Chronic Fatigue Syndrome and Inspiratory Resistive Loading

We selected patients with CFS because many believe CFS to be a psychosomatic disorder (14). If so, we would have expected similar ventilatory responses to IRL between groups

742

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

VOL 161

2000

Figure 7. CESD scores for subjects divided into four groups. Units (vertical scale) are those of the CESD depression scale. Data appear as mean ⫾ standard deviation. CESD score for the R/CFS⫹ group was greater than that for the NR/CFS⫹ group (t ⫽ 2.65; p ⬍ 0.02).

Figure 5. Spirograms depicting the ventilatory pattern before and after the introduction of the 3.54 cm H2O load. The format of this illustration is identical to that of Figure 4.

along with exaggerated subjective responses among the CFS subjects. CFS was initially thought to be an atypical manifestation of a hyperventilation syndrome. Lung function and CO2 stimulation testing have shown no apparent differences in respiratory regulation between CFS and matched control subjects, however (15, 16). The fact that CFS subjects are randomly distributed between the R and the NR groups in this study provides further evidence that CFS subjects do not exhibit any subtle abnormalities of respiratory regulation.

Figure 6. The numbers of subjects with (dark areas) and without (hatched areas) psychiatric diagnoses in the two groups.

Critique of the Methodology

We chose the resistive type of load because it roughly simulates airway resistive changes caused by bronchoconstriction. Effort could have become a confounding variable in this protocol, however, because a subject—by varying inspiratory flow and time—could conceivably vary the peak mouth or esophageal pressure associated with each breath. Ideally, the load used in this study should require a constant effort on the part of all subjects. It is unlikely, however, that differences in effort could explain the differences in dyspnea reported by the two groups. Airway pressure developed during inspiration, a major determinant of dyspnea, is a function of both resistance and flow (17). Because all subjects were free from airway obstruction, it is unlikely that any differences in airway resistance occurred between groups. The similarity in flow (VT/TI) seen in the spirograms of both groups throughout the study would also make differences in effort between subjects and/or groups unlikely. We chose a small load (1.34 cm H2O/L/s) for data collection. Weber’s law of perception describes the threshold level of an added load that can be readily perceived by a subject. In general, both normal and asthmatic subjects are likely to recognize an added resistive load if that external added load is at least 30% of the magnitude of the subject’s intrinsic load (18, 19). Airway resistance in humans ranges between 0.5 and 2.0 cm H2O/L/s. Because our subjects all had normal spirometry, it is likely that their airway resistance fell within the normal range. We reasoned, therefore, that since our added small load was close to the normal intrinsic airway resistance, attentiveness to the load and/or mental state might be critical factors that would determine which subjects would identify and experience dyspnea when confronted with this small added load. Two further points require mention. First, scaling of perception of a load is not identical to scaling of the discomfort

Lavietes, Sanchez, Tiersky, et al.: Psychological Profile and Loading

743

associated with that load. This study has considered only the scaling of discomfort associated with resistive loading. Subjects did not scale either the size of the load or the effort required for breathing through the load. Second, although a learning effect has been associated with repeated presentations of large threshold loads, we did not consider the possibility of a learning effect in the design of this study (20).

sation response (23). Studies of load detection in both heartlung transplant recipients and patients with spinal cord lesions suggest that neither airway nor chest wall afferents play an important role in the detection of an IRL (24, 25). Airway afferents may, however, affect the determination of the magnitude of dyspnea a person associates with loading (26, 27). Cognitive factors could modify a subject’s load response in either of two ways. First, subjects with the appropriate psychological makeup may symptom amplify, or exhibit an exaggerated subjective response to a load. The tendency to report physical symptoms in general is related to a subject’s psychological state. The symptom amplifier tends to be anxious, selfconscious, and have low self-esteem. Furthermore, a person who is likely to report one specific symptom is likely to report others as well (28–30). Second, subjects with certain psychological characteristics, e.g., depression, may demonstrate a diminished physiological, or ventilatory, response to a load when compared with those of an appropriate control group. Two observations give support to the first hypothesis, that psychological state or trait may play a role in the subjective response to an IRL. First, symptom amplifiers, or the “responder” group subjects, demonstrated more depressive symptoms and were more likely to meet the diagnostic criteria for a DSM III-R psychiatric diagnosis than did “nonresponder” group subjects. Second, the subset of R group patients who had CFS scored highly on the CESD depression scale. A CESD score of ⭓ 27 is thought to identify subjects with depression (31). By contrast, the trait neuroticism was not associated with R group subjects. This is shown by two observations. One, neuroticism scores differed greatly between CFS and non-CFS subjects, whereas dyspnea scores did not. Two, there was no difference between the neuroticism scores of the R and NR groups. These data do not support the alternate hypothesis that subjects with a heightened subjective response to loading maintain ventilation, whereas those with lesser subjective responses to loading do not. Some published studies in which subjects were asked to identify loads (rather than report the degree of dyspnea associated with the loads as in this study) do tend to support this second hypothesis. In one study, anxious, dependent subjects, selected from a group of normal hospital workers, required higher inspiratory resistances for load recognition than did adaptive or rigidly independent subjects (3). In another study, subjects with generalized anxiety or panic disorder were unable to grade the magnitudes of a series of inspiratory resistances (4). Whether subjects could both fail to identify a load but experience dyspnea when confronted with the same load is unclear.

Subjective Responses to Loading

This study has clearly identified a subset of subjects who, when compared with the remainder of the study group, consistently report more dyspnea when presented with repeated inspiratory loads. The designation of responders versus nonresponders was made early in the protocol on the basis of each subject’s first response, when presented with the Borg scale, to the small load. The final determination of dyspnea while breathing through the small load, as well as all responses while breathing through the larger load, was made after subjects were assigned to either group. Because we could expect only a small response to the small inspiratory load, we therefore designated any increase of Borg score as a positive response. This study design precluded the possibility of randomization of the sequence of small and large loads. The fact that the subjective responses of our two groups were clearly separated while breathing through the larger load, even while their ventilatory patterns were closely matched, suggests that the introduction of an easily perceived load may be an appropriate tool for the general study of symptom amplification. Many stimuli that have been used previously to study symptom amplification elicit both physiologic and perceptual or sensory responses. Submersion of an extremity in ice water, for example, elicits both a physiological response (e.g., tachycardia and/or heightened postural muscle activity) and a perceptual response, pain (21). Possibly, the cognitive appraisal of the stress associated with the ice-water submersion reflects the physiological responses to the ice-water. On the other hand, differences in subjective responses to ice-water submersion between responders and nonresponders may merely reflect differences in the subjective perception of pain by symptom-amplifiers. By contrast, since all subjects in this study maintained a constant ventilation when challenged by a 3.54 cm H2O/L/s resistive load, any differences in Borg score responses between groups could not be explained by differences in the physiological responses to the load. Ventilatory Responses to Loading

Perusal of Figures 2 and 3 shows that R group subjects were more likely than NR subjects to maintain a constant ventilation while breathing for 5 min on the small load. By contrast, all subjects maintained ventilation constant during the trial involving the larger load. These observations are compatible with the notion that variability in the breath-by-breath computation of minute ventilation or its components, tidal volume and frequency, is diminished as the magnitude of an applied inspiratory load is increased. Studies in which elastic loads have been presented to normal subjects have demonstrated this phenomenon (22). The loads in that study were of greater magnitude (9 and 18 cm H2O/L) than those presented in this study and were, therefore, more readily recognized by the subjects. Psychological State and Load Compensation

In conscious humans, the ventilatory response to an added external inspiratory resistive load (IRL) depends upon cortical, intrinsic muscle, and reflex factors. Increased diaphragm tension induced by an inspiratory load initiates the load compen-

Mouthpiece Breathing

A mouthpiece is known to stimulate ventilation, in part because it represents added dead space but also because it stimulates oral respiratory afferents as well (32, 33). This study has shown that many subjects report an increased sense of dyspnea when breathing via a mouthpiece (when compared with unencumbered breathing). Furthermore, a high Borg score with mouthpiece breathing before loading was associated with a high neuroticism score. Neuroticism is a trait that contrasts adjustment or emotional stability with maladjustment. Subjects who score high on a neuroticism scale, that is, subjects who experience a wide range of negative affects such as fear, anger and/or guilt, associate dyspnea with small tasks related to respiration such as placing one’s mouth on a mouthpiece. A mouthpiece is clearly a confounding factor in studies that evaluate dyspnea. In this study the two control periods, one with-

744

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

out, the other with, the mouthpiece allowed us to separate the effect of the mouthpiece from the effect of the load upon Borg score responses. We conclude the degree to which people associate dyspnea with a given task differs. Indices that describe and quantify emotional state may explain, in part, such intersubject differences. The subjective experience during loaded breathing is, in part, independent of the ventilatory response to the load. Acknowledgment : The writers thank Dr. Matthew Marin for his assistance with our illustrations, Mr. Martin Feuerman for help with the statistical analysis, and Mr. Waldo Duran for technical assistance.

References 1. Ciccone, D. S., N. Just, and E. B. Bandilla. 1996. Non-organic symptom reporting in patients with chronic non-malignant pain. Pain 68:329– 341. 2. Costa, P. T., Jr., and R. R. McCrae. 1980. Somatic complaints in males as a function of age and neuroticism: a longitudinal analysis. J. Behav. Med. 3:245–255. 3. Hudgel, D. W., D. M. Cooperson, and R. A. Kinsman. 1982. Recognition of added resistive loads in asthma. Am. Rev. Respir. Dis. 126:121–125. 4. Tiller, J., M. Pain, and N. Biddle. 1987. Anxiety disorder and perception of inspiratory resistive loads. Chest 91:547–551. 5. Johnson, S. K., J. DeLuca, and B. H. Natelson. 1996. Assessing somatization disorder in the chronic fatigue syndrome. Psychosom. Med. 58:50– 57. 6. Sharpe, M. 1997. Cognitive behavior therapy for functional somatic complaints. Psychosomatics 38:356–362. 7. Schluederberg, A., S. E. Straus, P. Peterson, S. Blumenthal, A. L. Komaroff, S. B. Spring, A. Landay, and D. Buchwald. 1992. Chronic fatigue syndrome research: definition and medical outcome assessment. Ann. Intern. Med. 117:325–331. 8. Borg, G. A. V. 1982. Psychophysical bases of perceived exertion. Med. Sci. Sports Exercise 14:377–381. 9. Demediuk, B. H., H. Manning, J. Lilly, V. Fencl, S. E. Weinberger, J. W. Weiss, and R. M. Schwartzstein. 1992. Dissociation between dyspnea and respiratory effort. Am. Rev. Respir. Dis. 146:1222–1225. 10. Marcus, S., L. N. Robbins, and K. Bucholz. 1990. Quick diagnostic interview schedule 3R, Version 1. Washington University School of Medicine, St. Louis, MO. 11. Radloff, L. S. 1977. The CES-D: a self-report depression scale for research in the general population. Appl. Psych. Meas. 1:385–401. 12. Costa, P. T., and R. R. McCrae. 1992. Revised NEO Personality Inventory (NEO-PI-R): Professional Manual. Psychological Assessment Resources, Odessa, FL. 13. Popovic, R. M., and D. P. White. 1998. Upper airway muscle activity in normal women: influence of hormonal status. J. Appl. Physiol. 84:1055– 1062. 14. Kruesi, M. J., J. Dale, and S. E. Straus. 1989. Psychiatric diagnoses in patients who have chronic fatigue syndrome. J. Clin. Psychiatry 50:53–56.

VOL 161

2000

15. Rosen, S. D., J. C. King, J. B. Wilkinson, and P. G. F. Nixon. 1990. Is chronic fatigue syndrome synonymous with effort syndrome? J. R. Soc. Med. 83:761–764. 16. Lavietes, M. H., B. H. Natelson, D. L. Cordero, S. P. Ellis, and W. N. Tapp. 1996. Does the stressed patient with chronic fatigue syndrome hyperventilate? Int. J. Behav. Med. 3:70–83. 17. Bradley, T. D., D. A. Chartrand, J. W. Fitting, K. J. Killian, and A. Grassino. 1986. The relation of inspiratory effort sensation to fatiguing patterns of the diaphragm. Am. Rev. Respir. Dis. 134:1119–1124. 18. Wiley, R. L., and F. W. Zechman, Jr. 1966. Perception of added airflow resistance in humans. Respir. Physiol. 2:73–87. 19. Rubinfeld, A. R., and M. C. F. Pain. 1977. Conscious perception of bronchospasm as a protective phenomenon in asthma. Chest 72:154–158. 20. Eastwood, P. R., D. R. Hillman, A. R. Morton, and K. E. Finucane. 1998. The effects of learning on the ventilatory responses to inspiratory threshold loading. Am. J. Respir. Crit. Care Med. 158:1190–1196. 21. Gramling, S. E., E. P. Clawson, and M. K. McDonald. 1996. Perceptual and cognitive abnormality model of hypochondriasis: amplification and physiological reactivity in women. Psychosom. Med. 58:423–431. 22. Brack, T., A. Jubran, and M. J. Tobin. 1997. Effect of elastic loading on variational activity of breathing. Am. J. Respir. Crit. Care Med. 155: 1341–1348. 23. Zechman, F. W., S. R. Muza, P. W. Davenport, R. L. Wiley, and R. Shelton. 1985. Relationship of transdiaphragmatic pressure and latencies for detecting added inspiratory loads. J. Appl. Physiol. 58:236–243. 24. Tapper, D. P., S. R. Duncan, S. Kraft, F. T. Kagawa, S. Marshall, and J. Theodore. 1992. Detection of inspiratory resistive loads by heart-lung transplant recipients. Am. Rev. Respir. Dis. 145:458–460. 25. Axen, K. 1982. Ventilatory responses to mechanical loads in cervical cord-injured humans. J. Appl. Physiol. 52:748–756. 26. Peiffer, C., D. Silbert, J. Cerrina, F. L. Ladurie, P. Dartevelle, A. Chapelier, and P. Herve. 1996. Respiratory sensation related to resistive loads in lung transplant recipients. Am. J. Respir. Crit. Care Med. 154: 924–930. 27. DiMarco, A. F., D. A. Wolfson, S. B. Gottfried, and M. D. Altose. 1982. Sensation of inspired volume in normal subjects and quadriplegic patients. J. Appl. Physiol. 53:1481–1486. 28. Schwartz, G. E., R. J. Davidson, and D. J. Goleman. 1978. Patterning of cognitive and somatic processes in the self-generation of anxiety: effects of mediation versus exercise. Psychosom. Med. 40:321–328. 29. Taylor, J. 1953. A personality scale of manifest anxiety. J. Abnormal Soc. Psychol. 48:285–290. 30. Lipman, R., K. Rickels, L. Covi, L. Derogatis, and E. Uhlenhuth. 1969. Factors of symptom distress. Arch. Gen. Psychiatry 21:328–338. 31. Schulberg, H. C., M. Saul, M. McClelland, M. Ganguli, W. Christy, and R. Frank. 1985. Assessing depression in primary medical and psychiatric disorders. Arch. Gen. Psychiatry 42:1164–1170. 32. Gilbert, R., J. H. Auchincloss, Jr., J. Brodsky, and W. Boden. 1972. Changes in tidal volume, frequency, and ventilation induced by their measurement. J. Appl. Physiol. 33:252–254. 33. Freedman, A. R., B. T. Mangura, and M. H. Lavietes. 1983. Minute ventilation in asthma. Am. Rev. Respir. Dis. 128:800–805.