Abstract
Maximum inspiratory pressure (MIP), an indicator of inspiratory muscle strength, is reported on 13,005 African-American and white participants from the Atherosclerosis Risk in Communities Study. Sex-specific associations between MIP and age, anthropometric measures, physical activity, health status, smoking status, and education level are presented. In this cohort of subjects 47 to 68 yr of age, MIP decreased 0.93 cm H2O (p ⩽ 0.0001) in women and 1.2 cm H2O (p ⩽ 0.0001) in men for each 1-yr increase in age. In both sexes, after adjusting for age, the associations between MIP and current body weight, weight at 25 yr of age, waist girth, body mass index, and calf girth were statistically significant and nonlinear (convex quadratic). MIP was positively associated with standing height in both sexes after controlling for age (p ⩽ 0.03). The age-adjusted association between MIP and sitting height was nonlinear in men and linear in women. MIP was also associated (positively) with education level, health status, physical activity, and (negatively) cigarette smoking. The study was not designed to evaluate associations with race, but data patterns suggest the absence of a race effect on MIP.
Maximum inspiratory pressure (MIP), a nonspecific indicator of inspiratory muscle strength, is measured in clinical settings to facilitate the diagnosis and management of patients with primary or secondary inspiratory muscle strength abnormalities (1). The method suggested by Black and Hyatt (2) continues to serve as the standard for determining MIP (3). Enright and colleagues (4) have since provided standard values for older adults based on 2,871 healthy participants 65 yr of age and older in the Cardiovascular Health Study. Analyzing the full cohort of 4,443 participants, they also reported positive associations between MIP and male sex, FVC, handgrip strength, and higher lean body mass, and negative associations with age, current smoking, poor health status, and waist girth.
The objective of including the measurement of MIP in the Atherosclerosis Risk in Communities Study (ARIC) was to assess whether respiratory muscle weakness has a role in the association of impaired spirometry with cardiovascular mortality. The current report is based on MIP determinations in more than 13,000 men and women 47 to 68 yr of age examined in ARIC. This is believed to be the largest study of MIP to date. The purpose of this report is to describe the association between MIP and demographic and anthropometric indices in this biethnic community sample of middle-aged adults.
ARIC is a prospective epidemiologic study of new and established risk factors for atherosclerosis and community trends in coronary heart disease. The baseline cohort of 15,792 men and women 45 to 64 yr of age was selected from four U.S. communities: Forsyth County, NC; Jackson, MS; selected suburbs of Minneapolis, MN; and Washington County, MD. After a home interview to assess demographic and health history, the baseline examination was conducted between November 1986 and March 1990; a second clinic examination was performed between February 1990 and March 1993; ARIC is now in its fourth examination cycle. Annual telephone interviews have been conducted since the baseline examination. Protocols for each examination were approved by appropriate institutional review boards, and informed written consent was obtained from each participant at each clinic examination. Details of the ARIC study design are published elsewhere (5).
MIP was measured at the second clinic examination, as were most of the data included in this report. However, a few demographic and other variables were collected only at the preceding annual telephone interview (health status), the first examination (physical activity, standing height, sitting height, and calf girth), or the initial recruitment home interview (education level).
MIP was measured at residual volume after spirometric testing. Methodology was standardized across the four field centers, and pulmonary technicians were trained and certified annually. Solid-state analog maximum respiratory pressure transducers were assembled with aneroid pressure gauges by the S&M Instrument Company (Doylestown, PA). Each system was connected to an IBM PC/XT by an analog-to-digital interface so that real time pressure-by-time curves could be displayed on the computer monitor. A 1-mm leak was utilized to minimize pressure generated by oral muscles.
Participants were instructed to apply noseclips, slowly exhale completely until the lungs were empty, insert a disposable cardboard mouthpiece between the teeth, sealing the lips around it, and then inhale with as much force as possible. The technique was described to the participants as “like trying to suck a thick chocolate malt through a narrow straw.” All MIP measurements were taken with the participant in a seated position. In addition to coaching by the technician, participants observed the computer monitor displaying their time-pressure curve as feedback and to provide incentive for improvement. At least three, and as many as five, MIP trials were completed by each participant with the goal of obtaining acceptable (test lasting at least 2 s) and reproducible (second largest pressure within 10% of the largest pressure) results.
Anthropometric measures were determined by trained, certified technicians following a detailed, standardized protocol. All anthropometric measurements were made on fasting (8 to 12 h) participants wearing light-weight, nonconstricting underwear and hospital scrub suits. Standing height was determined with the participant's head positioned in the Frankfort horizontal plane. Sitting height was measured by having the participant sit on a stool approximately 32 inches high in a standardized position. Both participant and stool heights were recorded; actual sitting height was calculated as the seated participant's height minus the stool height. All height measurements were recorded to the centimeter, rounding down. Body weight was assessed on a balance scale, measured to the pound, rounding down. Body mass index (BMI) was calculated as weight (kg) ÷ [height (m)]2. With the participant seated and leg hanging freely, calf girth was recorded as the maximum circumference over the calf muscle, preferably on the right side. Waist girth was determined at the level of the umbilicus. Girth measurements were recorded to the centimeter, rounding down. Elbow breadth, utilized as an indicator of frame size, was measured at the epicondyles of the humerus while the elbow was flexed at 90 degrees. The average of two triceps skinfold measures, taken at the midpoint between the acromion process and olecranon at the midline of the back of the right arm, was utilized in analyses. Skinfolds and elbow breadth measurements were recorded to the millimeter, rounding down. Detailed descriptions of the anthropometric procedures are provided in the ARIC Cohort Component Procedures manuals (6, 7) for the baseline and follow-up examinations.
Trained and certified interviewers collected information on smoking, health status, weight at 25 yr of age, physical activity, and education. During the annual telephone contact prior to the second examination, participants were asked to categorize their health, relative to others their own age, as excellent, good, fair, or poor. The physical activity questionnaire was based on an instrument developed by Baecke and colleagues (8). A highly reliable index of physical activity (sports or exercise during leisure time), ranging from 1 (low) to 5 (high), was calculated based on frequency and intensity of reported activities (9). For data analysis, this index was categorized into low (1 to 1.9), medium (2 to 2.9), and high (3 to 5) levels of physical activity. Education level was analyzed as basic (0 to 11 yr), intermediate (high school graduate or GED through vocational school), or advanced (1 yr or more of college).
SAS System software (10) was utilized for all statistical analyses. General linear modeling techniques (11) were used; all models fit were sex-specific and adjusted for age by analysis of covariance. In addition, sex-specific stepwise regression with forward selection (p ⩽ 0.15 to enter the model; p ⩽ 0.05 to remain in the model) was utilized to identify the strongest set of independent predictors of MIP. The effects of center, race, education, smoking status, physical activity, and health status were modeled using indicator variables. Former and never cigarette smokers who reported current use of pipes or cigars were excluded from the analysis of smoking. Bonferroni corrections for multiple comparisons were applied to p values for comparisons of age-adjusted mean MIP between levels of selected covariates. To reduce the influence of outlying values in the tails of the distributions, values more extreme than the 1st and 99th sex-specific percentiles were excluded from models and graphs for sitting and standing heights, calf and waist girths, current weight, weight at 25 yr of age, BMI, and triceps skinfolds. Although MIP is a negative pressure, it is reported ignoring the sign as is customary in the literature.
Participants were classified into one of five possible center and race groups: Forsyth blacks, Forsyth whites, Jackson blacks, Minneapolis whites, and Washington County whites. Individual center and race effects are confounded since three of the centers comprised participants of a single race group.
Because of the large number of observations, it was not practical to present individual data points on the figures. To facilitate assessment of goodness of fit, figures present predicted values from selected age-adjusted linear models overlaid with observed values representing the unadjusted mean MIP value for observations falling into equal-width categories. Because of the sparsity of data in the tails of the distributions, outermost categories were combined with the adjacent category; for example, women weighing 103 to 105 lbs were included with those weighing 106 to 110 lbs, the adjacent full-width category.
This cross-sectional analysis was based on participants who completed the second examination (n = 14,348; 91% of the baseline cohort). Excluded from the analysis data set were participants missing all pulmonary function data (n = 345), race groups other than African-American or white (n = 42), nonwhites at Minneapolis and Washington County field centers (n = 49), participants outside of the age range 47 to 68 yr at the time of their second examination (n = 14), and those for whom MIP was missing (n = 392). Initial analyses suggested a consistent bias in the MIP determinations by one technician. The mean MIP for the 501 participants tested by this technician was 71 in comparison with means of 82 to 103 for 27 other technicians. Mean MIP measured by this technician remained outlying after adjusting for sex, race, education level, and income level. After excluding the participants tested by this technician, data for 13,005 participants remained (82% of the original cohort).
In comparison to participants in the second examination who were excluded, participants included in the present analysis were slightly younger (1.1 yr), heavier in weight (2.2 pounds), and with better FVC (3.7 versus 3.4 L) and FEV1 (2.7 versus 2.5 L). A greater proportion of African-Americans were excluded from the analysis primarily because of failure to complete the pulmonary or MIP procedures. Participants included in the analysis tended to have more years of education (79% with intermediate or advanced education versus 73%) and a better self-reported health status (35% excellent versus 28%), and they were less likely to be current cigarette smokers (22 versus 26%).
Fifty-five percent (n = 7,197) of the participants were female. This study included 3,095 African-American participants from two field centers; 2,817 (91%) were from the Jackson center, which enrolled only African-American participants, and the remainder were from the Forsyth County center. Approximately 21% of the cohort reported an education level less than high school graduate. Although a similar percentage of the male (24%) and female (21%) participants were current cigarette smokers, 51% of the women never smoked, whereas only 27% of the men were never smokers. Only 16% of the cohort reported their health as “fair” or “poor” relative to others of their own age. Within sex, a greater proportion of women were categorized into low (32%) and a smaller proportion into high (23%) levels of physical activity than were men (22% low; 36% high).
The average age of this cohort of 5,808 men and 7,197 women was 57 yr. Unadjusted mean MIP was 106 cm H2O (SD = 28.4) in men and 79 cm H2O (SD = 24.8) in women. Pearson's coefficients for correlation between MIP and age were statistically significant in women (r = −0.21) and in men (r = −0.24). Across the age range of this cohort, MIP decreased 0.93 cm H2O (p ⩽ 0.0001) in men for each 1-yr increase in age (Figure 1a).
Fig. 1. Observed and predicted maximum inspiratory pressure by age and anthropometric measures in men (open stars) and women (open circles).
[More] [Minimize]Differences in the mean, age-adjusted MIP levels were observed by study population and ethnicity in both sexes (Table 1). In women, Minneapolis and Washington County whites and Jackson African-Americans were similar, with significantly higher average MIP than Forsyth County whites and Forsyth County African-Americans. In men, the mean MIP for Minneapolis and Washington County whites were similar, but significantly higher than the other center and race groups. The center and race effect remained notable in both sexes after adjusting for both age and education level.
| Center and Race Variable | n (%) | Unadjusted | Age-adjusted | |||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Mean | SD | Mean | 95% CI | |||||||
| Men | ||||||||||
| Forsyth County, Black | 106 (1.8) | 95 | 36.5 | 95.3 | 90.1–100.5 | |||||
| Jackson, Black | 1,027 (20.9) | 102 | 29.7 | 100.1 | 98.5–101.8 | |||||
| Forsyth County, White | 1,214 (17.7) | 102 | 28.0 | 102.2 | 100.7–103.7 | |||||
| Minneapolis, White | 1,795 (30.9) | 110 | 26.1 | 110.0 | 108.8–111.3 | |||||
| Washington County, White | 1,666 (28.7) | 108 | 28.9 | 109.1 | 107.8–110.5 | |||||
| Women | ||||||||||
| Forsyth County, Black | 172 (2.4) | 70 | 26.0 | 70.6 | 67.0–74.2 | |||||
| Jackson, Black | 1,790 (24.9) | 81 | 25.4 | 80.8 | 79.7–81.9 | |||||
| Forsyth County, White | 1,366 (19.0) | 73 | 23.3 | 73.5 | 72.2–74.7 | |||||
| Minneapolis, White | 1,951 (27.1) | 80 | 23.9 | 79.4 | 78.4–80.5 | |||||
| Washington County, White | 1,918 (26.6) | 81 | 25.2 | 81.8 | 80.7–82.8 | |||||
| After applying a Bonferroni correction for multiple comparisons, p values for testing the equality of age-adjusted mean MIP were: | ||||||||||
| Center and Race Comparison | p Values | |||||||||
| Male | Female | |||||||||
| Forsyth Black versus | Jackson Black | 0.8222 | 0.0001 | |||||||
| Forsyth White | 0.1249 | 1.0000 | ||||||||
| Minneapolis White | 0.0001 | 0.0001 | ||||||||
| Washington White | 0.0001 | 0.0001 | ||||||||
| Jackson Black versus | Forsyth White | 0.7459 | 0.0001 | |||||||
| Minneapolis White | 0.0001 | 0.8202 | ||||||||
| Washington White | 0.0001 | 1.0000 | ||||||||
| Forsyth White versus | Minneapolis White | 0.0001 | 0.0001 | |||||||
| Washington White | 0.0001 | 0.0001 | ||||||||
| Minneapolis White versus | Washington White | 1.0000 | 0.0264 | |||||||
Average MIP increased monotonically with increasing levels of education (Table 2). Differences in mean MIP between levels of education were attenuated after adjustment for age, remaining statistically significant only in men.
| Unadjusted | Age-adjusted | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Variable | n (%) | Mean | SD | Mean | 95% CI | |||||
| Men | ||||||||||
| Education level | ||||||||||
| Basic | 1,239 (21.4) | 98 | 29.6 | 99.6 | 98.1–101.2 | |||||
| Intermediate | 2,127 (36.7) | 107 | 27.5 | 106.1 | 104.9–107.2 | |||||
| Advanced | 2,431 (41.9) | 110 | 27.7 | 109.4 | 108.3–110.5 | |||||
| Smoking | ||||||||||
| Current | 1,383 (23.8) | 100 | 28.2 | 98.8 | 97.3–100.2 | |||||
| Former | 2,881 (49.6) | 107 | 27.9 | 108.5 | 107.4–109.5 | |||||
| Never | 1,540 (26.5) | 109 | 28.7 | 109.0 | 107.6–110.4 | |||||
| Physical activity | ||||||||||
| Low | 1,298 (22.4) | 101 | 29.4 | 101.0 | 99.6–102.5 | |||||
| Medium | 2,407 (41.6) | 105 | 27.9 | 104.8 | 103.7–105.9 | |||||
| High | 2,083 (36.0) | 111 | 27.8 | 110.8 | 109.6–112.0 | |||||
| Health status | ||||||||||
| Excellent | 2,149 (37.0) | 111 | 26.8 | 110.6 | 109.5–111.8 | |||||
| Good | 2,795 (48.2) | 105 | 28.5 | 105.4 | 104.4–106.4 | |||||
| Fair | 715 (12.3) | 97 | 29.2 | 98.0 | 96.0–100.0 | |||||
| Poor | 142 (2.4) | 91 | 28.9 | 92.3 | 87.8–96.7 | |||||
| Women | ||||||||||
| Education level | ||||||||||
| Basic | 1,545 (21.5) | 76 | 25.2 | 77.8 | 76.5–79.0 | |||||
| Intermediate | 3,275 (45.6) | 79 | 24.6 | 79.2 | 78.4–80.1 | |||||
| Advanced | 2,368 (32.9) | 81 | 24.6 | 79.6 | 78.6–80.6 | |||||
| Smoking | ||||||||||
| Current | 1,478 (20.5) | 75 | 24.4 | 74.5 | 73.2–75.7 | |||||
| Former | 2,059 (28.6) | 81 | 25.3 | 80.4 | 79.4–81.5 | |||||
| Never | 3,657 (50.8) | 80 | 24.6 | 80.1 | 79.4–80.9 | |||||
| Physical activity | ||||||||||
| Low | 2,286 (31.9) | 78 | 24.9 | 77.6 | 76.6–78.6 | |||||
| Medium | 3,235 (45.1) | 79 | 24.8 | 79.3 | 78.5–80.1 | |||||
| High | 1,655 (23.1) | 80 | 24.4 | 80.5 | 79.4–81.7 | |||||
| Health status | ||||||||||
| Excellent | 2,348 (32.7) | 81 | 24.7 | 80.8 | 79.8–81.7 | |||||
| Good | 3,585 (49.9) | 79 | 24.4 | 78.9 | 78.1–79.7 | |||||
| Fair | 1,061 (14.8) | 76 | 25.2 | 76.9 | 75.4–78.3 | |||||
| Poor | 197 (2.7) | 73 | 28.7 | 73.4 | 70.0–76.7 | |||||
Unadjusted mean MIP was 7 to 9 cm H2O lower (p ⩽ 0.0001) in male current cigarette smokers and 5 cm H2O lower (p ⩽ 0.0001) in female current cigarette smokers than in former or never smokers (Table 2). After adjustment for age, lower MIP in current smokers persisted, and the differential increased with increasing age. The decline in MIP associated with cigarette smoking was also evident in analyses using cigarettes smoked per day. After controlling for age, each cigarette smoked per day was associated with a MIP lower on average by 0.3 cm H2O (p ⩽ 0.0001) in men and in women. When this analysis was restricted to current smokers only, a one cigarette per day increment was associated with a reduction in MIP of 0.23 cm H2O in women (p ⩽ 0.0001), but it was not significant in men.
In men, the crude mean MIP was significantly lower in those who reported less exercise during leisure time than in those with greater physical activity. Although a similar pattern existed in women, the magnitude of the difference in MIP with increasing physical activity level was smaller.
Self-reported health status in comparison with other participants of the same age was positively associated with MIP in both men and women, independent of age. A monotonic association was observed such that mean MIP was greatest in participants reporting excellent health status and lowest in those reporting poor health status. The magnitude of the association was greater in men than in women.
Sex-specific mean values of selected anthropometric characteristics and their correlation with MIP are presented in Table 3.
| Men (n = 5,808) | Women (n = 7,197 ) | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Characteristic | Mean | SD | Correlation with MIP (Pearson's r ) | Mean | SD | Correlation with MIP (Pearson's r ) | ||||||
| Height, cm | 176 | 6.6 | 0.07‡ | 162 | 6.0 | 0.05‡ | ||||||
| Sitting height, cm | 92 | 3.7 | 0.14‡ | 86 | 3.5 | 0.09‡ | ||||||
| Weight, lb | 190 | 32.1 | 0.11‡ | 164 | 37.0 | 0.14‡ | ||||||
| Weight at age 25, lb | 166 | 26.4 | 0.11‡ | 128 | 21.4 | 0.10‡ | ||||||
| Current weight/weight age 25 ratio | 1.16 | 0.170 | 0.00 NS | 1.28 | 0.236 | 0.09‡ | ||||||
| Body mass index | 27.8 | 4.3 | 0.08‡ | 28.2 | 6.1 | 0.13‡ | ||||||
| Waist, cm | 100 | 11.2 | 0.02 NS | 96 | 16.1 | 0.06‡ | ||||||
| Hip, cm | 104 | 8.4 | 0.07‡ | 107 | 12.0 | 0.11‡ | ||||||
| Waist/hip ratio | 0.97 | 0.054 | −0.05† | 0.89 | 0.084 | −0.02 NS | ||||||
| Calf, cm | 38.1 | 3.1 | 0.13‡ | 37.0 | 3.9 | 0.13‡ | ||||||
| Elbow, mm | 73.9 | 4.1 | 0.05‡ | 64.9 | 4.8 | 0.08‡ | ||||||
| Subscapular, mm | 20.2 | 8.3 | 0.03* | 23.4 | 11.0 | 0.12‡ | ||||||
| Triceps, mm | 16.8 | 7.1 | 0.02 NS | 27.4 | 8.0 | 0.14‡ | ||||||
Although Pearson's correlations between MIP and anthropometric measures were generally statistically significant but modest in magnitude (r < 0.15), most anthropometric measures were highly correlated with each other. Pearson's correlation coefficients of 0.80 or greater were evident in both sexes between weight and waist girth, hip girth, and BMI, and between BMI and waist and hip girths. Elbow breadth, included in some models to adjust for frame size, and sex- specific correlation coefficients less than 0.60 with other anthropometric measurements. The range of observed values for some anthropometry variables was substantially wider in women than in men.
As illustrated in Figure 1b, the relationship between MIP and body weight was nonlinear. A quadratic regression model fit the observed data reasonably well, but it underestimated mean MIP somewhat for the highest weights in men. Because identifying the best predictive model was not the focus of this investigation, higher level polynomials were not pursued. However, piecewise regression (12) was performed, but it did not provide a better fit. Parameter estimates for selected regression models, including those depicted graphically in the panels of Figure 1, are presented in the Appendix.
Self-reported weight at 25 yr of age was also evaluated for its association with MIP. Again, the relationship between MIP and body weight was nonlinear (Figure 1c).
As suggested by Figure 1d, the association between MIP and waist girth after adjusting for age was reasonably well characterized by a quadratic model in both sexes. Waist remained a statistically significant independent predictor of MIP after additionally controlling for frame size based on elbow breadth (data not shown).
Figure 1e illustrates the quadratic association between BMI and MIP in men and women. Thus, all three indicators of obesity (weight, waist girth, and BMI) demonstrated the same general shape of the association with MIP, that is, quadratic, with a positive linear coefficient and a negative quadratic term. However, although the degree of curvilinearity was similar in men and women for body weight, a greater degree of curvature for BMI and waist girth associations was evident in men than in women. To investigate whether the curvilinear associations of MIP with BMI, weight, and waist arose from confounding by center or race, each nonlinear association was evaluated in sex, race, and center-specific models. The quadratic associations were consistent and statistically significant within center and race-specific models for men, except for Forsyth County African-Americans for which the quadratic terms were not statistically significant. In women, the quadratic terms were nonsignificant in models for Forsyth County and Washington County whites, and borderline significant for the waist model in Jackson African-Americans (p ⩽ 0.07).
For women, MIP was linearly related to sitting height. On average, for each centimeter increase, MIP increased 0.34 cm H2O. The relationship between MIP and sitting height in men was nonlinear and better depicted as a quadratic function, as shown in Figure 1f.
By comparison, the association between MIP and standing height was linear in both sexes. Age-adjusted MIP increased 0.11 cm H2O in women and 0.17 cm H2O in men for each centimeter increase in standing height (Figure 1g).
The association between MIP and calf girth appeared nonlinear in both sexes and was fitted with a quadratic model as shown in Figure 1h.
In women, the association between triceps skinfold and MIP was linear, with MIP increasing 0.38 cm H2O for each millimeter increase in triceps skinfold. In men, linearity checks indicated that both quadratic and cubic skinfold terms were highly statistically significant (p < 0.0001); however, the magnitude of the association was minimal, reflected in a relatively horizontal slope. The simple linear model (without a quadratic term) was not statistically significant for triceps skinfold measurements in men (data not shown).
In order to identify the strongest independent predictor variables of MIP, a stepwise regression with forward selection was performed evaluating age, education level, smoking status, physical activity, current health status, center and race, and linear and quadratic terms for waist, height, weight, BMI, calf girth, sitting height, and weight at 25 yr of age. In women, age, education, smoking, physical activity, health status, center and race, standing height, weight at 25 yr of age, current weight and weight2, calf and calf2, waist2, and sitting height2 were all statistically significant independent predictors of MIP. BMI, which had statistically significant age-adjusted univariate associations with MIP, was not statistically significant when considered in the larger model for women. In men, age, education, smoking, physical activity, health status, center and race, weight, calf, waist2, and BMI and BMI2 were all statistically significant independent predictors of MIP. Standing height, sitting height, and weight at 25 yr of age that had statistically significant age-adjusted univariate associations with MIP were not statistically significant when considered in the larger model for men.
The ARIC Study provides the largest population-based sample on which MIP has been measured to date. Because of the large sample size, there is sufficient power to identify statistically significant associations that are not necessarily of a clinically meaningful magnitude. Within sex, statistically significant associations between MIP and age, education, center and race, smoking, health status, body weight, weight at 25 yr of age, waist girth, BMI, sitting height, standing height, and calf girth were demonstrated. Adjustment for age, and allowing for nonlinearity when appropriate, explained relatively little of the overall variance in MIP within sex (see model R2 in Appendix). Sex-specific model R2 values were lower in this study than those reported for comparable models in other studies (4, 13, 14). The nonlinearity of the association between MIP and certain measures of body size explained, in part, the modest magnitude of Pearson's correlation coefficients observed between these variables and MIP.
The observed mean MIP values of 106 in men and 79 in women were slightly higher than normal values reported by Black and Hyatt (2) for this age group and, as expected, notably higher than those reported by Enright and colleagues (4) for an older cohort.
The magnitude of the decline in average MIP with age demonstrated in ARIC after adjusting for weight (see Appendix) was similar to that reported by Enright and colleagues (4) in reference equations based on a healthy subgroup of the Cardiovascular Health Study cohort 65 to 85 yr of age. Black and Hyatt (2) found a much smaller association with age in their study, which remained statistically significant only in women when the analyses were performed separately by age group (20 to 54 yr, 55 yr and older). However, their study was limited by a relatively small sample size, which included only 10 men and 10 women in each age decade stratum.
Because of the study design, race associations cannot be thoroughly evaluated in ARIC independent of field center. However, given the observed distributions of MIP by center, race, and sex groups, a consistent pattern of mean levels of MIP by race was not evident.
Previous reports have examined the linear association between MIP and adult body size with inconsistent results. Whereas Vincken and colleagues (13) found that height, weight, and relative weight did not contribute to the explanation of variability in MIP beyond that explained by age and sex, Enright and colleagues (4) reported that height, weight, and waist were significantly related to MIP. Wilson and colleagues (14) reported height as a significant predictor of MIP in women, but not in men. The ARIC study has demonstrated that the association between MIP and several measures of ponderosity is nonlinear.
Standing height showed a direct (positive) and linear relationship with MIP in both sexes. This is likely to reflect an association between stature and inspiratory muscle strength. Also, pulmonary function studies have found that lung volumes are directly related to height (15).
Sitting height, as an indicator of chest cavity length, was expected to be associated with MIP, and to be a stronger predictor than overall body height. The apparent sex difference in the nature of the association between sitting height and MIP was not anticipated and warrants further consideration. MIP increased linearly with increasing sitting height in women, whereas in men the association was quadratic, with a notably steeper slope over most of the range, reaching a plateau at approximately 90 cm sitting height.
The associations between current weight and MIP and between self-reported weight at 25 yr of age and MIP were quadratic in both men and women. In both sexes MIP was positively associated with current weight through approximately 212 pounds, after which the association became negative.
A lower mean level of MIP at higher values of body weight, waist girth, and BMI is consistent with the reported effects of obesity on the respiratory system. Severe obesity leads to inspiratory muscles working inefficiently (16, 17) and a reduction in chest wall compliance (18-20).
Calf girth was expected to be positively associated with MIP, consistent with the report by Enright and colleagues (4) of an association between MIP and handgrip strength. Because there is no biologic reason to expect MIP to decline at the highest levels of muscularity, obesity may limit the validity of the use of calf girth as an indicator of muscularity. Instead, a threshold for maximal MIP and possible confounding by obesity are more plausible explanations for the observed quadratic relationship with calf girth.
This study confirmed findings by Enright and colleagues (4) indicating that current smokers have lower MIP values than former or never smokers. By contrast, other investigators (13, 21) have reported no association between smoking status and MIP. Further research in this area is needed to explain the contradictory reports.
The association between MIP and self-reported health status in this cohort 47 to 68 yr of age was similar to that demonstrated in an older cohort (4), in that both studies report monotonically increasing MIP with better health status. It is not certain whether symptomatic aspects of inspiratory muscle weakness itself is perceived by the participant as “unhealthy” and affects the self-evaluation, or whether reduced MIP is associated with (either causing or resulting from) other health problems that sway the evaluation.
A limitation of the present study is that the MIP data were cross-sectional and not all covariates of interest were measured concurrently. An implicit assumption was that these variables (e.g., sitting and standing height, calf girth) did not change markedly over the approximate 3-yr period between their measurement and the MIP determination. Another limitation of this study is that little information is available on measurement error for MIP. Although standardized training and protocols were used to collect study data, and pulmonary technicians were certified annually, there was evidence of differences in mean MIP by center and technician.
Most MIP determinations met acceptability (98%) and reproducibility (85%) criteria, but this does not guarantee achieving maximal effort (22). Considering that average MIP values in the ARIC cohort were similar to or of larger magnitude than most other published studies of MIP, we believe ARIC has performed MIP in a manner comparable with established laboratory procedures.
On the basis of this unprecedented opportunity to investigate MIP in a large, population-based sample of African-Americans and whites 47 to 68 yr of age, this report has documented sociodemographic and anthropometric correlates of MIP in considerable detail. With the provisions identified in this section, the findings are applicable to middle-aged men and women in typical community settings in four geographic regions of the United States.
| Age: | ||
| Females (0.045): | 131.94439 − 0.93490 (age) | |
| Males (0.059): | 175.22283 − 1.20493 (age) | |
| Center and Race: | ||
| Females | ||
| (0.062): | 133.46648 − 0.91382 (age) − 11.17614 (Forsyth Blacks) − 8.31467 (Forsyth Whites) − 0.95865 (Jackson Blacks) − 2.32738 (Minneapolis Whites) + 0.00000 (Washington Whites) | |
| (0.064): | 132.34045 − 0.87351 (age) − 11.56260 (Forsyth Blacks) − 8.71588 (Forsyth Whites) − 0.55102 (Jackson Blacks) − 2.93007 (Minneapolis Whites) + 0.000000 (Washington Whites) − 3.36422 (Basic Education) − 0.63413 (Intermediate Education) + 0.00000 (Advanced Education) | |
| Males | ||
| (0.082): | 181.90268 − 1.26867 (age) − 13.84170 (Forsyth Blacks) − 6.94357 (Forsyth Whites) − 9.00650 (Jackson Blacks) + 0.88148 (Minneapolis Whites) + 0.00000 (Washington Whites) | |
| (0.096): | 169.46647 − 0.98204 (age) + 51.75368 (Forsyth Blacks) + 5.02701 (Forsyth Whites) + 14.16408 (Jackson Blacks) + 8.08147 (Minneapolis Whites) + 0.00000 (Washington Whites) − 1.14431 (Forsyth Blacks × age) − 0.23137 (Forsyth Whites × age) −0.39638 (Jackson Blacks × age) − 0.15926 (Minneapolis Whites × age) + 0.00000 (Washington Whites × age) − 8.84663 (Basic Education) − 3.83245 (Intermediate Education) + 0.00000 (Advanced Education) | |
| Education Level: | ||
| Females (0.046): | 131.22031 − 0.91210 (age) − 1.85859 (Basic Education) − 0.37025 (Intermediate Education) + 0.00000 (Advanced Education) | |
| Males (0.075): | 172.73352 − 1.10391 (age) − 9.77081 (Basic Education) − 3.33274 (Intermediate Education) + 0.00000 (Advanced Education) | |
| Weight: | ||
| Females | ||
| (0.063): | 114.60755 − 0.90113 (age) + 0.09565 (weight) | |
| (0.069): | 76.18978 − 0.91114 (age) + 0.55863 (weight) − 0.00132 (weight)2 | |
| Males | ||
| (0.063): | 157.04599 − 1.14195 (age) + 0.07849 (weight) | |
| (0.072): | 70.00270 − 1.13767 (age) + 0.98201 (weight) − 0.00230 (weight)2 | |
| Weight at age 25: | ||
| Females (0.056): | 70.53938 − 0.88006 (age) + 0.74836 (weight at age 25) − 0.00222 (weight at age 25)2 | |
| Males (0.066): | 111.61329 − 1.16261 (age) + 0.64663 (weight at age 25) − 0.00163 (weight at age 25)2 | |
| Waist: | ||
| Females | ||
| (0.055): | 79.09995 − 0.98936 (age) + 1.01600 (waist) − 0.00439 (waist)2 | |
| (0.058): | 62.41431 − 0.98440 (age) + 0.97255 (waist) − 0.00438 (waist)2 + 0.31510 (elbow) | |
| Males | ||
| (0.068): | −6.58730 − 1.21592 (age) + 3.56787 (waist) − 0.01723 (waist)2 | |
| (0.070): | −32.12062 − 1.22838 (age) + 3.50182 (waist) − 0.01723 (waist)2 + 0.44551 (elbow) | |
| Body Mass Index: | ||
| Females (0.070): | 75.05811 − 0.95387 (age) + 3.45108 (BMI) − 0.04724 (BMI)2 | |
| Males (0.077): | 27.17294 − 1.17291 (age) + 9.84372 (BMI) − 0.16135 (BMI)2 | |
| Sitting Height: | ||
| Females (0.047): | 100.23339 − 0.89475 (age) + 0.34379 (sitting height) | |
| Males (0.071): | −673.17799 − 1.12520 (age) + 17.46312 (sitting height) − 0.09001 (sitting height)2 | |
| Standing Height: | ||
| Females (0.046): | 113.01471 − 0.92054 (age) + 0.11187 (standing height) | |
| Males (0.059): | 143.73857 − 1.17019 (age) + 0.16745 (standing height) | |
| Calf Girth: | ||
| Females | ||
| (0.062): | −35.69052 − 0.89375 (age) + 8.04646 (calf) − 0.09564 (calf)2 | |
| (0.063): | −40.62758 − 0.90319 (age) + 7.84659 (calf) − 0.09440 (calf)2 + 0.17158 (elbow) | |
| Males | ||
| (0.073): | −137.93695 − 1.13179 (age) + 15.28046 (calf) − 0.18699 (calf)2 | |
| (0.074): | −150.69750 − 1.14655 (age) + 15.33549 (calf) − 0.18926 (calf)2 + 0.20057 (elbow) | |
| Triceps Skinfolds: | ||
| Females (0.058): | 119.33430 − 0.89411 (age) + 0.37987 (triceps) | |
| Males (0.063): | 152.67959 − 1.20601 (age) + 3.75535 (triceps) − 0.18326 (triceps)2 + 0.00272 (triceps)3 | |
| Smoking: | ||
| Females (0.054): | 134.31265 − 0.95743 (age) − 5.66866 (current smk) + 0.26542 (former smk) + 0.00000 (never smk) | |
| Males (0.080): | 180.99010 − 1.26018 (age) − 9.96256 (current smk) − 0.47300 (former smk) + 0.00000 (never smk) | |
| Physical Activity: | ||
| Females (0.047): | 132.74762 − 0.94489 (age) + 1.23173 (high activity) − 1.66123 (low activity) + 0.00000 (medium activity) | |
| Males (0.077): | 175.16271 − 1.22614 (age) + 5.96943 (high activity) − 3.78691 (low activity) + 0.00000 (medium activity) | |
| Health Status: | ||
| Females (0.049): | 125.12048 − 0.91478 (age) + 7.40182 (excellent) + 5.54921 (good) + 3.40959 (fair) + 0.00000 (poor) | |
| Males (0.084): | 157.11429 − 1.13076 (age) + 18.36347 (excellent) + 13.12007 (good) + 5.75494 (fair) + 0.00000 (poor) | |
| Stepwise models: | ||
| Females (0.105): | 84.0756 − 0.7369 (age) − 1.6291 (basic education) − 3.2516 (current smk) − 3.1118 (low activity) − 1.6440 (medium activity) − 8.5939 (poor health) − 5.9122 (fair health) −2.7894 (good health) −10.4829 (Forsyth Blacks) − 7.6007 (Forsyth Whites) − 3.8903 (Minneapolis Whites) − 0.0016 (waist) −0.5030 (standing height) + 0.5153 (weight) − 0.0007 (weight2) + 3.1050 (calf) − 0.0466 (calf2) + 0.0510 (weight at age 25) + 0.0028 (sitting height2) | |
| Males (0.149): | 55.4328 − 0.9841 (age) − 4.0934 (basic education) − 1.8522 (intermediate education) − 4.2764 (current smk) − 4.2335 (low activity) − 2.9495 (medium activity) − 13.8032 (poor health) − 7.1393 (fair health) − 3.2239 (good health) − 13.2479 (Forsyth Blacks) − 7.4260 (Forsyth Whites) − 10.1264 (Jackson Blacks) − 3.5681 (Minneapolis Whites) − 0.0055 (waist) + 0.2703 (weight) + 8.7973 (BMI) − 0.1246 (BMI2) − 0.5694 (calf) |
The writers thank Wanda Alexander, Karen Birkholz, Dot Buckingham, Kelli Collins, Delilah Cook, Bob Ellison, Julie Fleshman, Amy Haire, Doris Harper, Patricia Hawbaker, Agnes Hayes, Joel Hill, Kathleen Hunt, Mary Hurt, Jane Johnson, Phyllis Johnson, Kim Jones, Marilyn Knowles, Penny Lowery, Patricia Martin, Catherine Paton, Pam Pfile, Hogan Pham, Delilah Posey, Charlie Rhodes, Dawn Scott, Selma Soyal, Valarie Stinson, Teri Trevino, Climmon Walker, Lily Wang, Laurie Wormuth, Virginia Wyum, Kiduk Yang, and Ding-yi Zhao.
Supported by Contracts N01-HC-55015, N01-HC-55016, N01-HC-55018, N01-HC-55019, N01-HC-55020, N01-HC-55021, and N01-HC-55022 from the National Heart, Lung, and Blood Institute.
| 1. | Clausen, J. L. 1982. Maximal inspiratory and expiratory pressures. In Pulmonary Function Testing: Guidelines and Controversies. Academic Press, New York. |
| 2. | Black L. F., Hyatt R. E.Maximal respiratory pressures: normal values and relationship to age and sex. Am. Rev. Respir. Dis.991969696702 |
| 3. | Rochester D. F.Tests of respiratory muscle function. Clin. Chest Med.91988249261 |
| 4. | Enright P. L., Kronmal R. A., Manolio T. A., Schenker M. B., Hyatt R. E.Respiratory muscle strength in the elderly: correlates and reference values. Cardiovascular Health Study Research Group. Am. J. Respir. Crit. Care Med.1491994430438 |
| 5. | ARIC InvestigatorsThe Atherosclerosis Risk in Communities (ARIC) Study: design and objectives. A.J.E.1291989687702 |
| 6. | ARIC Investigators. 1988. Atherosclerosis Risk in Communities Study Protocol Manual 2: Cohort Component Procedures, Version 2.0. Collaborative Studies Coordinating Center, Chapel Hill, NC. |
| 7. | ARIC Investigators. 1990. Atherosclerosis Risk in Communities Study Protocol Manual 2: Cohort Component Procedures for the Second Examination, Version 3.0. Collaborative Studies Coordinating Center, Chapel Hill, NC. |
| 8. | Baecke J. A. H., Burema J., Frijters J. E. R.A short questionnaire for the measurement of habitual physical activity in epidemiologic studies. Am. J. Clin. Nutr.361982939942 |
| 9. | Richardson M. T., Ainsworth B. E., Wu H.-C., Jacobs D. R. J., Leon A. S.Ability of the Atherosclerosis Risk in Communities (ARIC)/Baecke questionnaire to assess leisure-time physical activity. Int. J. Epidemiol.241995685693 |
| 10. | SAS Institute Inc. 1989. SAS/STAT User's Guide, Version 6, 4th ed, Vol. 2. SAS Institute Inc., Cary, NC. |
| 11. | Kleinbaum, D. G., L. L. Kupper, and K. E. Muller. 1988. Applied Regression Analysis and Other Multivariable Methods, 2nd ed. PWS-Kent Publishing Company, Boston. |
| 12. | Neter, J., W. Wasserman, and M. H. Kutner. 1985. Applied Linear Statistical Models: Regression, Analysis of Variance, and Experimental Designs, 2nd ed. Richard D. Irwin, Inc., Homewood, IL. |
| 13. | Vincken W., Ghezzo H., Cosio M. G.Maximal static respiratory pressures in adults: normal values and their relationship to determinants of respiratory function. Bull. Eur. Physiopathol. Respir.231987435439 |
| 14. | Wilson S. H., Cooke N. T., Edwards R. H. T., Spiro S. G.Predicted normal values of maximal respiratory pressures in Caucasian adults and children. Thorax391984535538 |
| 15. | O'Brien R. J., Drizd T. A.Roentgenographic determination of total lung capacity: normal values from a national population survey. Am. Rev. Respir. Dis.1281983949952 |
| 16. | Ray C. S., Sue D. Y., Bray G., Hansen J. E., Wasserman K.Effects of obesity on respiratory function. Am. Rev. Respir. Dis.1281983501506 |
| 17. | Sharp J. T., Barrocas M., Chokroverty S.The cardiorespiratory effects of obesity. Clin. Chest Med.11980103118 |
| 18. | Sharp J. T., Henry J. P., Sweany S. K., Meadows W. R., Pietras R. J.Effects of mass loading the respiratory system in man. J. Appl. Physiol.191964959966 |
| 19. | Naimark A., Cherniack R. M.Compliance of the respiratory system and its components in health and obesity. J. Appl. Physiol.151960377382 |
| 20. | Luce J. M.Respiratory complications of obesity. Chest781980626631 |
| 21. | Karvonen J., Saarelainen S., Nieminen M. M.Measurement of respiratory muscle forces based on maximal inspiratory and expiratory pressures. Respiration6119942831 |
| 22. | Aldrich T. K., Spiro P.Maximal inspiratory pressure: does resproducibility indicate full effort? Thorax5019954043 |
