INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Measurement of maximal inspiratory pressure (MIP) is a simple, quick and noninvasive clinical procedure for determining this index of inspiratory muscle strength both in healthy subjects and in patients with pulmonary or neuromuscular diseases (1). In the latter group, MIP is indicative of ventilatory capacity and the development of respiratory insufficiency (9), and is useful in assessing the degree of abnormality and in monitoring inspiratory muscle weakness in individual patients over time (3, 10). MIP is also helpful in evaluating the success of weaning patients from mechanical ventilators (4, 9), and in predicting the outcome of cardiac transplantation surgery in patients with chronic congestive heart failure (11).

We as well as others (7) have often observed the inability of normal subjects to reach the MIP reference values found in earlier studies (1, 6). Studies conducted to address this issue resulted in the publication of several reference equations (4, 7, 8). However, the screening procedures for good health, age range, and sample size have been inconsistent, not well defined, or inadequate. With the exception of the gender effect, there is little agreement about the effect of age, height, weight, and smoking history on MIP.

The objectives of this cross sectional study were to utilize a standardized electronic procedure for the measurement of MIP in a well-characterized group of men and women with a wide age range (20 to 90 yr), for the purpose of investigating the effect of age and for determining correlates, normal values, and reference equations for MIP.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Population

The Baltimore Longitudinal Study of Aging (BLSA) a long-term, multidisciplinary study of normal human aging, being conducted by the intramural research program of the National Institute on Aging since 1958 as an open-panel study that continuously recruits community-dwelling volunteers, mostly from the Washington-Baltimore area (12). The BLSA study participants, mostly white, well-educated, and generally in good health at the time of entry into the study, undergo 2 to 3 d of clinical and psychological measurements that include spirometry. They also fill out the American Thoracic Society (ATS) Division of Lung Diseases (DLD) pulmonary questionnaire (13). In 1990, MIP measurement was incorporated into the pulmonary function testing done in the BLSA in order to measure the effect of aging on inspiratory muscle strength. Between 1990 and 1993, MIP measurements were obtained from 435 men and 309 women, of whom 35 men and 14 women were excluded because of nonreproducible spirograms. Additionally, 19 men and eight women were excluded because of the unavailability of smoking histories and relevant anthropometric data, thereby resulting in a final study group that consisted of 381 men and 287 women with a total of 668 measurements (one measurement per subject). The subjects, ranging in age between 20 and 90 yr, with a mean age of 56 yr, were predominantly white (94%). A healthy subgroup of the 668 subjects was selected for the purpose of constructing reference equations. This was done by excluding 136 men and 57 women from the whole group because of a definite or possible diagnosis of coronary heart disease (angina pectoris, history of myocardial infarction, Q-wave abnormality, or ST depression on exercise testing). Another 82 men and 86 women were excluded because of pulmonary disease symptoms (asthma, bronchitis, emphysema, phlegm, shortness of breath, pulmonary congestion, bronchiectasis, interstitial/fibrotic lung disease, or tuberculosis), congestive heart failure, or renal failure. Additionally, 10 men and four women were excluded because of aortic valvular disease, neurologic deficit (stroke, parkinsonism or extrapyramidal disease, movement disorders, multiple sclerosis, hemiplegia, myoneural disorders, or muscular dystrophy). Another 10 men and 11 women were excluded for taking one or more of the following medications: systemic or inhaled glucocorticoids, mineralocorticoids, central nervous system (CNS) stimulants, barbiturates, or muscle relaxants. After excluding another four men and one woman for outlier MIP values or low FEV₁% predicted, the final, healthy BLSA subgroup consisted of 139 men and 128 women (Table 1). The subjects in this healthy subgroup ranged in age from 20 to 90 yr (mean: 50 yr) and were 95% white. Smoking histories were obtained from the subjects and classified as reported previously (14). Never-smokers were those who had not smoked more than five to 10 packs of cigarettes, 50 to 75 cigars, or three to five packages of pipe tobacco during their lifetime. Current cigarette smokers were those who smoked cigarettes every day or who had quit smoking less than 2 yr before the visit date. Former smokers were those who smoked cigarettes every day but had quit two or more years before the date of their visit. Occasional smokers were those who were currently smoking cigarettes less often than every day. Pipe/cigar smokers were those who were currently smoking pipes or cigars at the time of the visit. The majority of the subjects in the healthy subgroup were never-smokers (45% of the males and 57% of the females), followed by former smokers (35% of the males and 23% of the females), occasional smokers (9% of the males and 13% of the females), and current smokers (4% of the males and 7% of the females). The percentage of pipe/cigar smokers among the men was 8%, whereas there were no pipe or cigar smokers among the women. In the study, the effect of smoking was evaluated by alternately comparing never-smokers with ever-smokers, current smokers with subjects who were not currently smoking, and former smokers with those who were not. Anthropometric measurements were obtained at the time of the visit. Grip strength, one of the correlates of MIP (1, 4, 15), was unavailable in our data set. Summed forearm circumference correlated strongly with summed grip strength in male participants of the BLSA (16), and for this reason was used as a surrogate variable for grip strength. Forearm circumference, in centimeters, of each arm was measured at its widest point, with the arm relaxed and the hand open. Summed forearm circumference was obtained by adding the circumferences of the two arms.

MIP and Spirometric Measurements

MIP was measured with a solid-state pressure transducer (S&M Instruments Co., Doylestown, PA) interfaced with a computer, allowing visualization of real-time pressure-time measurements. Only one experienced technician performed the measurements. The pressure transducer calibration used for MIP measurement was checked weekly at zero and 160 cm H₂O against an anaeroid gauge that was incorporated into the equipment. If the calibration factor varied by more than 5% on repeated attempts, the equipment was recalibrated. The calibration values were stored in a log and reviewed weekly, along with a review of quality and reproducibility of the tracings. The maximum inspiratory effort was conducted by having the subject expire to a residual volume (RV) and then perform a maximum inspiratory maneuver. A cardboard mouthpiece with a hole 1 mm in diameter and 1 cm in length was used to reduce the effects of cheek muscles. An acceptable maneuver was defined as one that showed a 2-s plateau of inspiratory effort. A reproducible maneuver was defined as one that agreed with the largest acceptable maneuver to within 10%. The test session was conducted for a maximum of five efforts, in order to obtain three acceptable and two reproducible MIP maneuvers, and about 80% of the test sessions achieved the 10% reproducibility goal. MIP was measured as the most negative pressure attained after the first second of effort (4). For clarity of presentation, we report here the absolute value of MIP. Therefore, a larger value indicates better performance.

Spirometric measurements for the study were made as described earlier (14), using a Collins DS water-seal spirometer (Warren E. Collins, Braintree, MA) driven by a personal computer and controlled by Pulmo-Screen software with the National Heart, Lung, and Blood Institute (NHLBI)/Atherosclerosis Risk in Communities (ARIC) study modification (S&M Instrument Co.). The instrument met the accuracy criteria of the ATS (17) and, as reported earlier, reproducibility criteria were met when the second largest FEV₁ and FVC were within 5% of the largest values (14). FEV₁% predicted was calculated by dividing the observed FEV₁ by the corresponding predicted FEV₁ values. The latter were derived from the BLSA sex- and race-specific cross-sectional FEV₁ prediction equations (14).

Statistical Analyses

Cross-sectional analysis of the data was done with SAS software (SAS Institute, Cary, NC) (18). Stepwise multiple regression models were constructed, using MIP as the dependent variable. Independent candidate predictors were based on data from previous studies and on plausibility (1, 4, 8, 17). Candidate predictors included age, height, FVC, peak expiratory flow (PEF), weight, summed forearm circumference, race, presence of definite or possible coronary heart disease, smoking status, gender, and the interaction of age and gender. Even though the proportion of nonwhites in the study was small, we evaluated the effect of race on MIP. In some instances age squared (age²), and the coded variables for gender and coronary heart disease were also offered. For the purpose of deriving gender-specific reference equations, separate regression analyses were done for men and women, using age², age, height, and weight as the explanatory variables. Coefficients were considered significant at a value of p  0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

MIP results for the entire BLSA cohort (n = 668) were analyzed separately by gender. For ease of representation, and since the decline in MIP seems to be slower until middle age, age groups of unequal age range were selected. Table 2 shows mean MIP and FEV₁% predicted values by gender and age. In the youngest group (age < 39.9 yr), the mean MIP values for men and women were 117.6 cm H₂O and 79.5 cm H₂O, respectively. As it decreases with age, MIP reaches an average of 66.0 cm H₂O and 45.5 cm H₂O, respectively, for men and women in the oldest age group (age  75). The mean MIP values of the females in each of the age groups were approximately 70% of the values for the males. Moreover, MIP decline, in absolute terms, is smaller in women than in men in all age groups.

The significant predictors of MIP in the combined set of men and women in the BLSA were determined through stepwise linear regression. The full model for predicting MIP was the same as the one described earlier in the METHODS section. In the final model, weight and the coded variables for race, smoking, and presence of coronary heart disease were not significant, whereas age, gender, and the interaction of age and gender, in addition to height, FVC, PEF, and summed forearm circumference were significant (results not shown). This indicates that not only does gender affect MIP, but that aging affected MIP in men differently than in women in this set of subjects. When similar regression analysis was done by gender (Tables 3 and 4), age was found to be a negative predictor (age coefficient = 0.78), whereas PEF and summed forearm circumference were positive predictors for male subjects. For females, age and height were found to be negative MIP predictors (age coefficient = 0.36), whereas FVC, PEF, and weight were positive predictors. The term age² was not significant for either men or women. Moreover, race, smoking status, and presence of definite or possible coronary heart disease did not affect MIP in men or women in this group.

The healthy group (free of pulmonary, heart, and other diseases) was next analyzed as described previously. The mean MIP and FEV₁% predicted by gender and age are presented in Table 2. In general, the MIP values in the healthy subgroup were about the same as those of the whole BLSA group. In younger healthy subjects (< 55 yr) the average MIP in women was about 67% that of men. This value increased in older subjects to about 72%, probably because the mean age of women was lower than that of men. As in the entire study cohort, the decline in MIP, in absolute terms, was smaller in women than in men. Table 2 also shows that with two exceptions, the mean FEV₁% predicted levels in the healthy subset were slightly higher then those of the whole group of BLSA participants. The characteristics of this group of healthy BLSA subjects are shown in Table 1. The women in this group (mean age = 47.2 yr, mean MIP = 72.4 cm H₂O) were on average younger than the men (mean age = 51.8 yr, mean MIP = 101.2 cm H₂O). When compared with subjects in the whole group, the healthy men's and women's mean MIP values were slightly higher (8.6 cm H₂O and 5.4 cm H₂O, respectively), probably because they were on average about 6 yr younger.

Separate regression analyses for men and women were next conducted to determine the correlates of MIP among the healthy BLSA subgroup. Table 3 shows that among healthy males (n = 139), age was a significant negative predictor of MIP, whereas PEF and weight were positive predictors. Table 4 shows that in healthy females (n = 128), age and height were significant negative predictors of MIP, whereas PEF and weight were positive predictors. The variable summed forearm circumference was not offered in the regression model for females, since its correlation with weight was higher than that for males. Tables 3 and 4 also show that in the healthy subgroups of men and women, race and smoking status were not significant (this applies to all smoking comparisons described in METHODS). For this reason we had no race or smoking exclusions in this study. It is also noteworthy that the term age² was considered, but was not found to be significant either for men or for women.

Regression analyses of the data sets showed that age was a very strong negative predictor of MIP, with men having larger coefficients than women. PEF was a positive predictor of MIP, with larger coefficients for women than for men. Height was a negative predictor only for females, whereas weight, which was of borderline significance in the set of healthy males, was a very strong positive predictor of MIP in females, especially those in the healthy subgroup. Summed forearm circumference was significant in the analyses of both the pooled BLSA data set, and the male subset. In none of the analyses did we observe both, weight and summed forearm circumference to be simultaneously significant.

Reference equations were constructed separately by gender, using the two sets of healthy men and healthy women, respectively. The variables offered in the full model were age², age, weight, and height. The final models are presented in Table 5, which also shows the lower limit of the normal range (LLN). The latter was defined as the fifth percentile of negative residuals (actual  predicted values) of MIP (4).

Figures 1A and 1B compare the normative values derived from the two sets of healthy BLSA men and women with those derived from other cross-sectional studies. The BLSA values for the younger subjects are similar to those of Black and Hyatt (6). However, for the older subjects, the BLSA values fall between those of the other studies.

View larger version (18K): [in this window] [in a new window]   View larger version (20K): [in this window] [in a new window]   Figure 1.   Comparison of MIP reference equations for men (A) and for women (B). The mean and lower limit of the normal range (LLN) for maximal inspiratory pressure (MIP) in healthy BLSA men and women of average weight and height are compared with those for healthy men and women from studies by Black and Hyatt (6), Vincken and colleagues (7), Wilson and associates (8), and Enright and coworkers (4).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

MIP is an indicator of inspiratory muscle strength and a determinant of VC (19). Decline in inspiratory muscle strength, if severe, can lead to impaired airway clearance and inadequate ventilation (15). MIP is known to be decreased in pulmonary diseases such as COPD (20), degenerative neuromuscular diseases (21), congestive heart failure (11), and during long-term corticosteroid treatment (22). Given the value of MIP in assessing inspiratory muscle strength, and the range of normal reference values obtained because of the variety of procedures used and differences in population selection, we believe that it is important to establish reference equations for MIP for a well-defined, healthy population spanning a wide range of ages.

In this study we used standardized methods for spirometry and MIP-related measurements, resembling those of Enright and colleagues (4), in a well-defined group of male and female volunteers with a wide age range. From this group we selected a subgroup of healthy subjects who were free of pulmonary, coronary, or neuromuscular diseases and who were not taking any medication that might interfere with muscle strength. In this respect, it is important to emphasize that because of the extensive battery of tests given to BLSA participants, the healthy subgroup was very well defined and very healthy as compared with most other reference populations. The MIP values obtained are consistent with those obtained in other studies (7, 8) that reported lower MIP values than did Black and Hyatt (6) and Ringquist (1). Our results demonstrated a very strong gender effect, which is consistent with all previous findings reported in the literature (10). Separate analysis by gender showed that in all age groups, MIP in men was about 30% higher than in women. This is similar to results obtained by several investigators (7, 8, 10, 15, 23). Moreover, the coefficients of age were more negative for men than for women, indicating that the age-related decline in MIP in men was larger than that in women.

Our finding, from multiple linear regression analysis, that age is a significant negative predictor of MIP for both men and women is consistent with the results of Ringquist (1) and Vincken and colleagues (7), both of whom used study groups with wide age ranges, and with the results of Enright and coworkers (4, 23) in elderly subjects. However, our finding is contrary to results obtained by other investigators, such as Bruschi and associates (10) and Mcelvaney and colleagues (3), which indicated that age did not significantly affect MIP. Our results are also different from those of Wilson and associates (8) and Berry and associates (15), who found a negative relationship between age and MIP only in adult and elderly men, respectively, and from the results of Black and Hyatt, who reported a significant negative effect of age only in older women (6). The failure to show an effect of age on MIP in a group of young adults (19) is not surprising, and is consistent with previous findings that age-related decline in pulmonary function begins in the middle thirties (24). In addition to age, the other negative predictor of MIP in our multiple linear regression model was height. The latter was significant in the females but not in the males, a finding that is consistent with previously reported findings (1, 8, 19). This means that of two women who are of the same age and weight, the taller one would have a smaller MIP. We interpret this to mean that the shorter woman is the more muscular.

Our finding that weight is a significant positive predictor of MIP in healthy women and healthy men is consistent with that of Leech (19), but different from that of Ringquist (1). The positive effect of body weight on MIP may be due in part to the relationship between weight and the isometric length of different muscle groups (25), and to the fact that alterations in body weight have been shown to affect the diaphragm muscle mass (26). Schoenberg and associates (27) called the improvement in lung function that accompanies small weight increases, the "muscularity effect," and speculated that it is attributable to increased respiratory muscle force.

Ringquist used handgrip flexor strength in his linear regressions as a surrogate for overall isometric strength (1). Enright and coworkers (4) found a positive correlation between MIP and handgrip strength in the elderly. In the present study we found a high correlation between weight and forearm circumference, both of which were correlated separately with MIP. In the set of healthy males, weight was a better predictor of MIP than was forearm circumference, although the opposite was true for the larger BLSA male subset. PEF, known to be correlated with maximal respiratory pressure (19), was a significant positive predictor of MIP in all equations, whereas in the analyses of data for both healthy men and healthy women, FVC was not. Moreover, the results of our analyses showed that smoking status did not affect MIP. This was consistent with findings reported in the literature (7, 10), but contrary to those reported by Enright and coworkers (4). Race did not seem to affect MIP among either males or females, perhaps because the study population was overwhelmingly Caucasian. Even though the health exclusions used in this study were very rigorous, we did not find differences in MIP between the healthy subgroup and the overall group.

We compared our gender-specific reference equations with those from four study groups in which MIP was measured at lung volumes near RV (Table 6). With the exception of the study by Enright and coworkers (4) which had a much larger number of subjects (1,269 men and 1,602 women), the other studies had a few dozen subjects of each gender. In the studies by Black and Hyatt (6) and by Wilson and colleagues (8), the total numbers of subjects older than 55 yr were 50 and 26, respectively, whereas the study by Enright and coworkers (4) was restricted to subjects older than 65 yr. The instruments used by Enright and coworkers and Vincken and colleagues to measure MIP were comparable to ours, whereas Black and Hyatt and Wilson and associates used aneroid gauges. Our results, although close to those of Vincken and coworkers, and slightly higher than those of Enright and coworkers, are lower than those of Black and Hyatt or Wilson and colleagues (Table 6, Figures 1A and 1B).

In summary, we measured MIP in a healthy subgroup of BLSA subjects, using a standardized electronic procedure. In addition to showing a strong gender effect, MIP decreased with age in both men and women. The decline in both men was larger than that in women. The reference equations derived from this study are useful in assessing the strength of inspiratory muscles.