Secular Trends in the Physical Fitness of United States Army Recruits on Entry to Service, 1975–2013

Abstract

Knapik, JJ, Sharp, MA, and Steelman, RA. Secular trends in the physical fitness of United States Army recruits on entry to service, 1975–2013. J Strength Cond Res 31(7): 2030–2052, 2017—A systematic literature search was conducted to identify and analyze articles that reported on physical fitness of new US Army recruits. The National Library of Medicine's PubMed and the Defense Technical Information Center were searched using the keywords (military personnel OR trainee OR recruit OR soldier) AND (physical fitness OR strength OR endurance OR flexibility OR balance OR coordination OR muscle contraction OR running OR exercise OR physical conditioning). Reference lists of obtained articles and contact with authors enhanced the search. Studies were selected if they involved recruits in Basic Combat Training or One-Station Unit Training, provided a quantitative assessment of at least one fitness measure, and the fitness measure(s) were obtained early in training. Average values for each fitness measure were obtained, plotted by the year of data collection, and fitted to linear regression models (fitness measure × year). Fifty-three articles met the review criteria. Regression analysis indicated little temporal change in height, but body weight, body mass index, body fat, and fat-free mass increased over time. Limited V̇o₂max data suggested no temporal change in male recruits, but those in female recruits V̇o₂max seem to have slightly improved. Apparently contradicting the V̇o₂max findings, performance on endurance runs (1- and 2-mile) declined, possibly because of the increase in body weight. Muscular endurance (push-ups, sit-ups) demonstrated little systematic change over time. Limited but multiple measures of muscular strength suggest a temporal increase in strength. Specific components of US Army recruit fitness seem to have changed over time.

Introduction

Physical fitness is widely acknowledged as important for health and optimal physical performance. Lower physical fitness is a risk factor for cardiovascular disease (^{10,37,157,170}), certain types of cancers (^116,145), type 2 diabetes (^30,73), hypertension (^29,40), stroke (⁶⁵), and all-cause mortality (^10,117). In the military services, higher levels of physical fitness are important not only for health, but also for military-specific task performance, injury prevention, and reduced attrition from military service. In operational and experimental research, it has been demonstrated that fitter soldiers exhibit higher levels of performance on actual or simulated military tasks (^{88,150,159,162}). Numerous studies have now shown that the performance on military-specific tasks like load carriage, repetitive lifting, obstacle courses, casualty drags, and other militarily relevant tasks are improved by higher levels of physical fitness (^{61,63,83,88,112,159}). Individuals entering the military with low levels of fitness have been shown to be more susceptible to injury (^74,99,103) and more likely to be discharged early in their military service (^86,179).

Temporal trends in the physical fitness in army recruits are broadly representative of changes in the American population because US service members are drawn from a large cross-section of the United States. US Army recruits represent a unique age group, generally 18–35 years, although skewed to the younger ages. Many national samples have examined high-school aged individuals (15–18 year olds) (^68,180), and data on older individuals are only available in limited, nonrepresentative studies (^120,148,189). Furthermore, large-scale sampling of the fitness of US children and youth largely ended in the 1980s (¹⁴²), and there are no other nationally representative fitness trends that have been published since that time. In 2012, the CDC began conducting youth fitness measures (¹⁶), but this is a recent development, and although some of these data have been reported (³⁹), temporal changes are not available at this time. Thus, the data collected on recruits involve a broad national sample, a unique age group, and a time period in which other US national data are not available.

This article will first define physical fitness, then systematically review the literature on the physical fitness of US Army recruits on entry to service. We previously examined secular changes in the physical fitness of recruits from 1975 to 2003 (¹⁰¹), and this paper expands on this work by including the extensive and more recently published data.

Physical Fitness Definitions and Components

Physical fitness has been defined in a number of ways (^{17,22,55,58,67,128,141,182,188}), but a simple definition that incorporates many aspects of the others is that physical fitness is “a set of attributes that allows the performance of physical activity” (¹²⁸). The attributes or components of physical fitness have been determined based on factor analytic studies that provide construct validity for the fitness concept (^48,101). In these studies, individuals were administered a broad array of physical performance tests for which quantitative measures could be obtained. Correlational and factor analytic techniques were used to assemble the tests into groups having commonality. After a long series of studies, a number of components were identified, although different authors categorized these components somewhat differently because of the types of tests involved in the different studies (^{25,48–50,64,135,197}). For example, a factor related to cardiovascular endurance could not emerge if there was no long-term endurance activity, which was the case in studies before 1972. Table 1 provides the components of physical fitness and their definitions assembled from a number of sources (^{22,25,48,58,64,141,188}). Physiological studies assisted in identifying the fitness components by linking the components found in the factor analysis studies to particular biological characteristics. These included the energy systems recruited to fuel-specific activities (e.g., aerobic, anaerobic), muscle fiber types, the neuromuscular control necessary to accomplish specific tasks, and the physical principles involved in specific activities (^{56,66,128,182,190}).

Table 1.: Components of physical fitness.*

Some investigators have separated the components of physical fitness into those that are health-related and those that are performance-related as shown in Table 1 (^25,55,141). Health-related components are those that have been specifically linked to health. For example, a number of investigations indicate that low levels of aerobic fitness are associated with higher risk of cardiovascular disease and mortality and certain types of cancers and cancer mortality (^{37,154,156,167}). Performance-related components are those related to skilled activity in sports, exercise, or occupational tasks. There is considerable overlap between health-related and performance-related fitness as shown in Table 1. Some authors have combined balance, speed, agility, and coordination into a single component called “mobility,” which is characterized as the functional application of strength and endurance to produce purposeful and effective body motion (⁵).

Body composition, body weight, and body mass index ([BMI], defined as weight/height²) are often considered components of fitness because of their relation to health and their interaction with the other fitness factors. Overweight and obesity are associated with an increased risk of cardiovascular disease, type 2 diabetes, musculoskeletal disorders, and certain types of cancers (³²). Muscle mass is highly correlated with absolute strength (^71,72,127), power production (⁶⁰), cardiorespiratory endurance (¹⁸³), and the performance of many physical tasks (^59,183). Individuals with more fat tend to have more difficulty performing certain tasks, especially those requiring weight bearing activity and cardiorespiratory endurance (^31,183). Body weight is generally negatively associated with lower-body power tests like the broad jump, vertical jump, and short sprints (^21,24,62,149) and positively associated with upper-body power tasks (^62,129).

Methods

For this investigation, the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were followed (¹³³). The authors followed a review protocol that is described below.

Information Sources and Search

National Library of Medicine's PubMed and the Defense Technical Information Center (DTIC) were searched to find studies on the physical fitness of US Army recruits on entry to Basic Combat Training (BCT) or One-Station Unit Training (OSUT). In the DTIC search, only unclassified articles with full text and unlimited public access (i.e., approved for public release, distribution unlimited) were examined. Keywords used in the search included military personnel OR trainee OR recruit OR Soldier AND physical fitness OR strength OR endurance OR flexibility OR balance OR coordination OR muscle contraction OR running OR exercise OR physical conditioning. The reference lists of obtained articles were searched for other articles not found in the retrieval services. The files of a senior researcher with experience in studying recruit fitness were also examined. Eleven authors were contacted personally to clarify data collection methods or to obtain original data. No limitations were placed on the dates of the searches, and the final search was completed in May 2016.

Study Selection and Data Collection Process

Articles were selected for the review if (a) they were written in English, (b) they involved recruits in BCT or OSUT, (c) they provided a quantitative assessment of at least one physical fitness measure, and (d) the physical fitness measures were obtained early in training, generally within the first week, before any substantial physical training had occurred. Publication titles were first examined, and abstracts were reviewed if the article seemed to involve military personnel and physical fitness. The full text of the article was retrieved if the abstract suggested that recruits were involved in the study and fitness measures obtained. If the authors did not explicitly report on a specific physical fitness measure but data were available within the article to calculate the measure, it was included. This happened primarily in the cases of BMI and fat-free mass (FFM). If height and weight were provided, then BMI was estimated as follows: weight/(height × height) (kg·m⁻²). If body weight and body fat were provided, FFM was estimated as follows: body weight-body fat (kilogram). If weight provided as kilogram and body fat as a percent, FFM was estimated as follows: weight − ([body fat × 0.01] × weight).

Not included in the review were studies which involved (a) military personnel other than new recruits, (b) studies that did not contain quantitative data, (c) studies that combined male and female data (because of the large differences in fitness between the sexes), and (d) studies that combined data of recruits plus soldiers who were not in basic training. There were studies that involved a single data collection period for which the data were used in multiple publications. Authors were contacted to identify these investigations, and studies with the largest number of subjects were used in the analyses. In several cases, authors provided their original data that contained the largest number of cases, and these data were used in the analyses. These latter cases were footnoted in the tables.

Data Analyses

Where only 2 years of data were available on a particular measure, the data were placed in a table and the differences in the 2 points in time examined. Where data were available for at least 3 years, the mean values were plotted. Where data from a single data collection period overlapped 2 years (e.g., 2005 and 2006), the year with the greater number of days was selected for the plot. Mean data were placed into a database, and the Statistical Package for the Social Sciences (SPSS, Version 21) were used to apply a variety of curve fitting techniques to examine trends over time (years). Linear, polynomial, exponential, logarithmic, and power fits were applied to the mean data, but with few exceptions, a linear fit was able to account for almost as much variance as alternate fits; alternatives to linear modeling generally added little explanatory information. Linear correlation coefficients (fitness measures vs. years) were calculated and p values and standard errors of estimate (SEE) were determined for these coefficients and shown on graphs. In the regression analysis, years were indicated by sequential numbers (1, 2, 3 etc.) rather than the actual year, so that the intercept indicated the estimated fitness value at the earliest point at which the data were collected.

Results

Figure 1 shows the publications included and excluded at each stage of the literature search. The PubMed search yielded 4,619 titles and the DTIC search 3,146. After reading titles and selected abstracts, 215 full-text articles were obtained. Fifty-three articles in 24 independent data collection periods fully met the review criteria. These studies had been performed between 1975 and 2013, a 38-year span. Table 2 shows that these studies separated into those that examined (a) physical characteristics (weight, height, and BMI) and body composition (fat and FFM) and (b) other fitness variables including cardiorespiratory endurance, muscle strength, muscular endurance, and power. In some studies, the number of recruits tested differed for some variables, so a range is shown. All studies used convenience samples of recruits rather than random samples. The average ages of the recruits in the studies ranged from 19 to 23 years.

Figure 1.: Publications included and excluded at each stage of literature review.

Table 2.: Studies measuring physical fitness of Army recruits on entry to service.*

Table 2-A.: Studies measuring physical fitness of Army recruits on entry to service.*

Physical Characteristics and Body Composition

Data on physical characteristics and body composition were available over 38 years from 1975 to 2013. Height was generally measured with a statiometer and weight with a digital or beam balance scale, with one study reporting self-reported heights and weights (⁹⁴). Most studies provided body fat and FFM estimated from skinfolds (^{2,4,14,74,75,78,80,85,99,100,102,105,109–111,119,122,124,130,131,140,143,161,169,193}), but a few investigations reported data only from dual X-ray absorptiometry (^53,187) or circumferential techniques (^54,146). Where data on multiple body composition techniques were provided, skinfold estimates were selected because most studies used this method.

Figure 2 shows that men and women demonstrated little change in height over the 38-year survey period. Figure 3 shows a trend of increasing body weight over time for both men and women: linear regression indicated an estimated 71.5 kg in 1975 and 80.0 kg in 2013 for men (12% difference); for women, these values were 58.5 kg and 63.7 kg (9% difference). Figure 4 shows a trend of increasing BMI over time for both men and women: linear regression indicated an estimated 23.3 kg·m⁻² in 1975 and 25.6 kg·m⁻² in 2013 for men (10% difference); for women, these values were 22.0 and 23.9 (9% difference). Figure 5 shows that there was little difference over time when body fat was expressed as a percentage of the body weight. By contrast, Figure 6 shows a trend of increasing body fat mass (kg) over time for both men and women: linear regression equations indicated an estimated 11.8 kg in 1975 and 14.0 kg in 2013 for the men (19% difference); for women, these values were 15.4 kg and 18.4 kg (19% difference). Figure 7 shows a trend of increasing fat-free mass over time for both men and women: linear regression equations indicated an estimated 59.0 kg in 1975 and 69.4 kg in 2013 for the men (18% difference); for women, these values were 42.1 kg and 46.5 kg (10% difference).

Figure 2.: Trends in height of US Army recruits (1975–2013).

Figure 3.: Trends in body weight of US Army recruits (1975–2013).

Figure 4.: Trends in body mass index of US Army recruits (1975–2013).

Figure 5.: Trends in percent body fat of US Army recruits (1975–2013).

Figure 6.: Trends in body fat (kg) of US Army recruits (1975–2013).

Figure 7.: Trends in fat-free mass of US Army recruits (1975–2013).

Two studies measured the bone mineral density of female recruits with DEXA, and there was no difference in mean values between the studies. A 1993 report on 158 female recruits found bone mineral density averaged (±SD) 1.204 ± 0.079 gm·cm⁻² (¹⁸⁷). A 1998 report on 168 female recruits found an average (±SD) bone mineral density of 1.204 ± 0.086 gm·cm⁻² (⁹⁹).

Cardiorespiratory Endurance

Figure 8 shows studies (^143,161,184) that examined V̇o₂max of recruits on 3 occasions between 1975 and 1998 (23 years). V̇o₂max was obtained from direct measures of oxygen consumption using a graded, uphill running treadmill protocol. The correlations between the V̇o₂max values and the years were not statistically significant, but there were only 3 data points and thus low statistical power. For men, the high correlation coefficient (i.e., good linear fit of points) and relatively flat regression line (small regression slope) suggested little differences in the V̇o₂max of the men among the 3 periods. The linear regression equation indicated that the 1975 value differed by less than 1% from the 1998 value. Women also had a high correlation coefficient, but in contrast to the men, they demonstrated a slight increase in V̇o₂max over the same time period. The linear regression equation indicated that, compared with the 1975 value, the 1978 value was 3% lower, but the 1998 value was 6% higher.

Figure 8.: Trends in V̇o ₂max of US Army recruits (1975–1998).

Figures 9 and 10 show data on 2- and 1-mile run times, respectively. The runs were administered by drill sergeants as part of the Army Physical Fitness Test (APFT) (⁵), and run times were obtained from the records of the training companies. Recruits wore a numbered vest or carried a numbered plaque for identification. Drill sergeants lined up the recruits at a starting point, started the run, and recorded the time it took for each recruit to complete the distance. Figure 9 shows 2-mile run times collected between 1998 to 2008 (21-year period), and the graph suggests a decline in performance (i.e., longer run times) for both men and women: linear regression indicated an estimated 14.8 minutes in 1987 and 17.0 minutes in 2008 (15% difference); for women, these values were 19.5 and 21.3 minutes, respectively (9% difference). One-mile run times from 3 studies conducted between 1984 and 2003 are shown are shown in Figure 10. Although the regression was not statistically significant, both the male and female data suggest a trend of decreasing performance over time in consonance with the 2-mile run data.

Figure 9.: Trends in 2-mile run times of US Army recruits (1998–2007 for men; 1988 to 2008 for women).

Figure 10.: Trends in 1-mile run times of US Army recruits (1984–2003).

Muscular Endurance

Studies that examined push-up (PU) and sit-up (SU) performance of new army recruits from 1984 to 2008 (24-year period) are shown in Figure 11 (PU) and Figure 12 (SU). Tests were administered by drill sergeants as part of the APFT (⁵), and the data were obtained from the records of the training companies. Drill sergeants monitored individual performances and were very familiar with administration of the well standardized events. For PUs, the recruit was required to lower the body in a generally straight line to a point where his or her upper arms were parallel to the ground, and then return to the starting point with elbows fully extended. For the SU, the recruit's knees were bent at a 90° angle, fingers were interlocked behind the head, and a second person held the participant's ankles to keep his or her heels firmly on the ground. The recruit raised his or her upper body to a vertical position so that the base of the neck was anterior to the base of the spine and then returned to the starting position. Drill sergeants recorded the number of PUs and SUs successfully completed within 2 minutes. The 2 tests were administered together, with the PUs performed first and the SUs second with at least 10 minutes of rest between tests.

Figure 11.: Trends in push-up performance of US Army recruits (1984–2007 for men; 1984 to 2008 for women).

Figure 12.: Trends in sit-up performance of US Army recruits (1984–2007 for men; 1984 to 2008 for women).

Figure 11 shows that there was little difference in PU performances over the 24 years. For the SUs shown in Figure 12, slope was negative, suggesting a decline in performance, but there was considerable variability in performance, and the linear regression fits were not significant.

Muscle Strength

Four measures of strength were obtained on recruits over periods separated by 15–20 years (^{27,99,105,161,169,187}). These measures included isometric upper-body strength, isometric lower-body strength, isometric upright pull strength, and incremental dynamic lift (IDL) strength. For a particular strength measure, the same testing devices and methodology was used. Figure 13A shows the device used to measure isometric upper-body and lower-body strength. For upper-body strength, the recruit was securely seated in a chair with a belt around his or her waist. With both hands, the recruit grasped a piece of aluminum tubing near the head such that the recruit's elbow was at a 90° angle with the upper arm and parallel to the floor. The recruit pulled downward on the bar exerting as much force as possible, and the isometric force was measured on a transducer (⁹⁸). For lower-body strength, the recruit was securely seated with feet resting on a bar and knees at a 90° angle (similar to a leg press). Hands were placed on circular bars just below the seat. The recruit exerted as much force as possible on the leg bar, and the isometric force was measured (⁹⁸). For the upright pull (Figure 13B), the recruit was standing and grasped with an alternating grip an aluminum bar 38 cm above the ground. The recruit bent his or her knees, kept the back straight and head up. The recruit pulled up vertically on the bar exerting as much force as possible, and the isometric force was measured with a transducer (¹⁰⁶). The IDL device (Figures 13C, D) consisted of an adjustable stack of weights (weight carriage) attached to 2 handles located 20 cm from the ground. The weight carriage allowed for the progressive stacking of weight to determine maximal lifting strength. The recruit grasped both handles of the device with feet shoulder distance apart, knees bent, back straight, and head up (Figure 13C). The recruit lifted the handles and weight carriage to a height of 152 cm (Figure 13D). An initial load of 18 kg was progressively increased until the maximal load the recruit could lift was determined using a 1-repetition maximum procedure (¹⁶⁹).

Figure 13.: Devices used to measure muscle Strength. A) isometric upper and lower body strength device; (B) isometric upright pull strength; (C) incremental dynamic lift, starting Position; (D) incremental dynamic lift, final position.

Table 3 shows studies (^105,161,169) that examined the 3 measures of isometric strength in new army recruits. Male upper-body and lower-body strength were, respectively, 15 and 11% higher in men in 1998 compared with 1978; for women, the values were 18 and 4% higher, respectively (20 years difference). Upright pull strength was 5% higher in men and 4% higher in women in 1998 compared with 1983 (15 years difference).

Table 3.: Strength and power measures of New US Army recruits (values are mean ± SD).*

Incremental dynamic lift data are plotted in Figure 14. The regression was not statistically significant because of the low statistical power (only 3 data points), but the high correlation coefficient suggested a strong trend toward increasing strength values over the time period. Linear regression indicated that male IDL strength was 22% higher in 1998 compared with 1983; female IDL strength was 40% higher in the same time period.

Figure 14.: Trends in incremental dynamic lifting performance of US Army recruits (1983–1998).

Leg Power

Two studies reported on female recruit vertical jump performance as shown in Table 3. The 1998 value (⁹⁹) using a Vertec apparatus was 13% higher than the 1993 value using a wall-and-chalk method (¹⁸⁷).

Discussion

This study examined temporal trends in US Army recruit physical fitness by reviewing and analyzing quantitative fitness measures available in the literature and extends previous findings (¹⁰¹) by adding 34 studies published since 2003. Our earlier analysis (¹⁰¹) suggested a slight but significant increase in height, but that was not substantiated when 6 more recent studies were added to the analysis. Interestingly, body weight and BMI have continued to increase at rates very similar to those noted earlier (¹⁰¹). The results regarding body fat (%) were not clear in the previous studies (¹⁰¹), but the current analysis indicated that although there was only a small difference in body fat over time when expressed as a proportion of the total body weight (i.e., % body fat), when expressed in absolute terms (i.e., kilogram), body fat tended to increase over time. The trend in fat free mass continued in men at a rate similar to that reported previously (¹⁰¹), but the rate for women was greater than that was indicated earlier (¹⁰¹). Importantly, our data suggest that the secular increase in body weight was not due solely to an increase in body fat, but also due to an increase in fat-free mass. There were no new studies examining V̇o₂max, and the limited V̇o₂max studies suggested no change in this measure among male recruits in the 23-year period between 1975 and 1998; however, female recruit V̇o₂max may have improved slightly over the same period. Adding additional studies on endurance runs for time (1- and 2-mile runs) continued to apparently contradict the V̇o₂max findings by suggesting performance declines (slower run times) for both men and women. Adding new studies on muscular endurance continued to suggest little systematic change over time. Limited but multiple measures of muscular strength suggested a temporal increase in strength over time in consonance with the temporal increase in fat-free mass.

A number of civilian studies have examined secular changes using BMI as a marker of overweight and obesity (^{41,45,107,113,152}). It should be pointed out that there are both advantages and disadvantages with the use of this index. The BMI removes the dependency of weight on height (⁸¹). It is easy to obtain measure, and large publicly available databases (e.g., Behavioral Risk Factor Surveillance Survey, Youth Risk Behavior Survey, and National Health and Nutrition Examination Survey) can be used to describe populations and trends, as many investigations have done (^41,45). The correlation between body fat and BMI is about 0.7 in both civilian samples (^46,81,153) and in new army recruits (⁸⁵). However, a correlation of 0.7 indicates that only about 1/2 of the variance in BMI is in common with body fat. Furthermore, BMI does not reflect the change in body composition (increase in fat and reduction in fat-free mass) that occurs naturally with aging, even within the age range of recruits in basic training (¹⁴⁷). There is also evidence that BMI may be associated with different proportions of body fat in different racial groups (³³), and that leg length and body build can affect the association between fat and BMI (¹¹⁵). An individual can have a high BMI not only because of higher body fat but also because of higher muscle and bone mass because the index does not distinguish between the 2.

There is historical evidence that BMIs have been increasing since at least the late 19th century in the United States (^52,107). Data obtained from large national US samples (i.e., National Health and Nutrition Examination Survey and National Health Examination Survey) did not begin until about 1960 and suggested little change in BMI from about 1960 to 1980, but after this, BMIs have increased over years (^42,47,118). The average increase in BMI from about 1980 to the early 2000s has been estimated at 11% (25.5–28.3 kg·m⁻²) for men and 13% (25.2–28.5 kg·m⁻²) for women (¹⁹²). Both the historical data and the post-1960 national data consistently show that the most rapid increases were among those who were the most overweight and obese, with those individuals becoming more overweight and obese (^42,47,107). The most recent NHANES data suggested a leveling off in the prevalence of obesity after the 2000s (^41,44). Recruits' BMI data reported here only involved mean values spanning the period from 1975 to 2013 (38 years) but showed a trend of increasing BMIs with similar rates of increase among men and women. Similar secular increases in weight and/or BMI have been reported in the US population and that of other countries (^{3,69,70,126,155,163,176,177,181,186,194,196}).

The weight for height standards for entry into the army have changed over time and could have affected the secular trends. A sample of the changing weight for height standards converted to BMIs are shown in Tables 4 and 5 for men and women, respectively. If a recruit exceeded the weight for height standard, then body fat could be estimated. Army regulations before 1991 stated that “body composition measurements may be used as the final determinate…” but it was not until 1991 that actual body fat standards were published as shown in Table 6, and these changed little over time. For both men and women, there was no change in the minimum BMI for entry to service until 2008, when it increased for most heights. After October 1991, the maximum allowable BMI for men was actually lower than previously allowed and did not change after this time. Before October 1991, maximal allowable BMIs ranged from 30.8 to 31.9 kg·m⁻², and after this time 27.1–28.0 kg·m⁻². Despite this more rigorous standard, men's body weight and BMI rose over time, as shown in Figures 3 and 4. For women, the maximum allowable BMI increased in each of the 3 periods. Before September 1991, the maximal allowable BMI ranged from 22.1 to 25.4 kg·m⁻²; from October 1991 to January 2008, 23.4–25.0 kg·m⁻²; and after January 2008, 25.5–26.4 kg·m⁻². These changes may have affected the women's average BMI, but women's BMI actually rose at a slightly slower rate than the men's (Figure 4).

Table 4.: Changes in minimum and maximal BMI for men to enter the US Army at various periods.*

Table 5.: Changes in minimum and maximal BMI for women to enter the US Army at various periods.*

Table 6.: Maximal allowable body fat if recruit exceeds weight for height tables.*

In accounting for the increase in BMI and body weight, it is useful to remember that body weight is regulated by a balance between energy output (activity) and energy intake (food). If activity output is greater than food input, body weight will decrease; if food intake is greater than activity output, weight will increase (¹⁶⁵). The civilian increase in body weight and BMI since 1980 has been ascribed to both a reduction in physical activity and an increase in food intake (^{42,43,137,152}). It should be noted that analyses of studies examining temporal trends in physical activity are complicated by differences in measurement techniques, questions asked on questionnaires, and data collection procedures. One study compared adult (≥16 or ≥18 years of age) data from 3 national survey systems including the Behavioral Risk Factor Surveillance System (BRFSS), National Health and Nutrition Examination Survey (NHANES), and National Health Interview Survey (NHIS) using standard definitions for “active” and “inactive.” The BRFSS indicated a slight increase in the prevalence of active individuals and a slight decline in inactive individuals from 2001 to 2007. The NHANES indicated a similar trend from 1999 to 2006. The NHIS indicated no significant trend from 1998 to 2007. Data from the nationally representative Youth Risk Behavior Surveillance Survey (YRBSS) are collected on high-school students (9th to 12th grade) every 2 years and showed only small and inconsistent changes in various measures of physical activity over time from 1991 (the first year of the survey) to 2015 (^1,121,195) for 18-year-old youth (prime US Army recruiting age). The NHANES began using accelerometers to objectively measure physical activity in 2005, and one study examining these data found little difference in the physical activity of adolescents in the limited timeframe from 2003 to 2006 (⁵⁷). The YRBSS indicated no change in the proportion of high-school students participating in physical education classes or on sports teams from 1991 to 2015 (¹⁷⁸). Earlier surveys and other data obtained as far back as 1965 suggest a small increase in the prevalence of adult leisure-time physical activity from 1965 to 1982 (¹⁶⁶), but occupational activity has decreased (²³). Other indicators of adolescent physical activity such as active commuting to school, high-school sport participation, outdoor play, and sedentary behaviors suggested that physical activity in this group has declined over time (¹²). Thus, there are contradictory data on temporal changes in physical activity in both adolescents and adults.

On the other hand, there is firmer evidence that energy intake has increased. Although complicated by differences in data collection methods over time, national data from NHANES indicate that energy intake increased in the 4 decades from the early 1970s to the 2000s. The increase in energy intake came primarily from carbohydrate sources and was found in most BMI groups (normal, overweight, and obese) (^11,191,192). More recent NHANES data extending to 2010 suggested that energy intake peaked in the 2003 to 2004 period, then decreased slightly in subsequent years, which was accounted for by small decreases in carbohydrates and fat intake (⁵¹).

These data on body composition suggested that the increase in BMI (and body weight because there was little change in height) was due to an increase in both body fat and fat-free mass. Interestingly, when fat was expressed as a proportion of the total body weight (% body fat), there was little change over time, but when expressed in absolute terms (kilogram), body fat did increase over time. Olds (¹³⁸) reviewed 154 studies conducted from 1951 through 2004 that had provided skinfold data to estimate body fat. He found that the average increase in body fat was 0.8 kg/decade, whereas the increase in fat-free mass was 0.6 kg/decade. In this study, body fat increased at a rate of 0.8 kg/decade for men and 0.6 kg/decade for women, whereas fat-free mass increased at rates of 2.8 kg/decade for men and 1.2 kg/decade for women. Both investigations support the contention that the secular increase in body weight is accounted for by increases in both fat and fat-free mass.

Eisenmann and Malina (³⁸) reviewed studies on peak V̇o₂max of younger US men and women during the 20th century. They found no change in the peak V̇o₂max of 15- to 19-year-old males from the mid-1930s to the 1990s but an increase in the V̇o₂max of 15- to 19-year-old females from the 1960s to the 1990s. The males and females in the Eisenmann and Malina (³⁸) study included 18- to 19-year-olds who were old enough to enter the military, and their data are broadly in consonance with the very limited V̇o₂max data involving new army recruits (^143,161,184). The increase in V̇o₂max of the women could possibly be associated with the increased participation of women in high school and college sports initiated in the US by Title IX of the Education Amendments Act of 1972. This act stated that no person in the US could be excluded from participation in federally funded education programs or activities (including sports) on the basis of sex (^35,171).

There are a few studies of secular trends in the cardiorespiratory fitness in older US individuals (^120,148,189), but these involve unique populations. Only one investigation has included a nationally representative sample, and this investigation did not examine secular trends (¹⁸⁵). Cross-sectional, longitudinal studies conducted at Cooper Clinic involved individuals who could afford an exclusive preventive medicine examination in Dallas, TX. In one of these studies, 20- to 34-year-old men who were tested between 1980 and 1989 had a higher estimated V̇o₂max (treadmill time, Balke protocol) than those tested in 1970–1979 (41–45 ml·kg⁻¹·min⁻¹), but those attending the clinic in later decades (up to 2009) had modestly lower V̇o₂max values compared with the 1980–1989 group (45 vs. 43 ml·kg⁻¹·min⁻¹) (¹⁸⁹). Women attending the clinic were not separated into age groups, but those with an average age of about 40 years showed an increase in estimated V̇o₂max from 1970 through to 1989 (31–36 ml·kg⁻¹·min⁻¹) and little change after this (¹²⁰). Another cross-sectional study examined students at a Seventh-Day Adventist college (average age 22 years) in each year from 1996 to 2008. They found a secular decline in estimated V̇o₂max (bicycle ergometer, Astrand protocol) that amounted to yearly linear declines of 0.8 and 0.4 ml·kg⁻¹·min⁻¹ for men and women, respectively (¹⁴⁸). In a nationally representative sample of Canadians, there was a decline in estimated V̇o₂max (step test) from 1981 and 2007–2009 (27-year period) (²⁸). Except for the female data from Cooper Clinic, these investigations are generally not in consonance with the V̇o₂max data in this study, but it should be noted that V̇o₂max was measured directly in the basic training studies, whereas in the other investigations (^{28,120,148,189}), it was estimated.

Several civilian studies have suggested a secular decline in the performance of endurance running tasks like aerobic shuttle runs and long distance runs for time (e.g., 1- and 1.5-miles) among both US males and females of high-school age (^{114,123,173,175,180}). This has been reported not only in the United States but also more recently in other countries (^{34,36,69,174,176,181}), including Finnish conscripts (¹⁵⁵), over periods ranging from 7 to 40 years. In agreement, we found a secular decline in both 1- and 2-mile run performance of recruits. There were 11 studies involving men over a 21-year period (1987–2007) and 10 studies involving women over a similar period (1988–2008). The linear trend showing a decline in running performance was supported by relatively high and statistically significant correlation coefficients between run times and years. Data on the 1-mile run times was more sparse (only 3 studies), but the correlations were high, and the trend was in the same direction as the 2-mile run data.

The endurance running data seem to contradict the studies on V̇o₂max because these are both measures of cardiorespiratory fitness. However, a distinction must be made between direct physiological tests like V̇o₂max and performance tests involving endurance runs for time. Endurance runs are used as surrogate measures of aerobic capacity because they have a large aerobic component (^82,172). Nonetheless, running tests can be influenced by factors not directly related to physiological capability like experience (practice, pacing ability, ability to tolerate running discomfort), motivation, and environmental factors (instructions, terrain, weather). One factor to explore in the apparent discrepancy between V̇o₂max and running performance is the secular increase in body weight (Figure 3). The body weight increase appeared to be due to increases in both fat mass and fat-free mass (Figures 6, 7). Increased fat mass would be expected to decrease running performance since weight added to the body increases energy cost (^31,77,79) without adding to the oxidative capacity of the body. On the other hand, additional fat-free mass can increase V̇o₂max in proportion to its contribution as oxidative muscle tissue involved in exercise (¹⁸³). One study that matched younger individuals for age, gender, BMI, and triceps skinfold found overall performance declines in 1.6-km runs and 20-m shuttle run tests amounting to 29–61%, suggesting that while markers of fatness might account for some of the decline, other factors (some of which are mentioned above) are also involved (¹³⁹).

No data was found on secular changes in the muscle strength of US youth or young adults in the civilian literature. Among investigations examining this in other countries, results are conflicting with some (^34,134,186) but not all (^163,177) studies generally showing higher strength values in later years compared to earlier years. There were only a few studies on strength in Army recruits and these used a variety of strength measures (^{27,105,160,161,169}) so caution must be exercised in interpreting this limited data. Nonetheless, the results were consistent in showing that 3 measures of static strength and one measure of dynamic strength (IDL) were all higher in 1998 than in 1978 or 1983.

The suggestion of increasing recruit strength over time is supported by self-reported physical activity and increases in fat-free mass over the same time period. In the YRBSS, a greater proportion of high school students reported participation in strength training activities over the period 1991 to 2015 (¹⁷⁸). Fat-free mass was also linearly increasing over time among recruits as shown here (Figure 7). About 1/2 of fat-free mass is muscle mass (¹²⁸) and the cross-sectional amount of muscle mass is highly correlated with muscle strength (^71,72,127).

US civilian data on secular changes in muscular endurance was limited to the PU performance of boys, flexed-arm hang performance of girls, and SU performance of both genders (^26,180). In general, there appeared to be a small upward trend such that the PU performance of 16- to 17-year old males and flexed-arm hang performance of 16- to 17-year old females increased slightly from 1965 to 1985 (^26,180). SU performance of 10- to 17-year males and females increased 9% from 1980 to 1989 (¹⁸⁰). Studies in other countries have been mixed, but with few exceptions (⁷⁰) have generally shown declines in various measures of muscular endurance over time (^126,181,186).

In the recruit data there were a relatively large number of studies involving PU and SU performance over a period of 24 years from 1984 to 2008 (^{2,14,18–20,27,74–76,84,87,89,90,95–97,99,100,102,104,146,151,161,164}). The PU scores demonstrated little systematic change over the years and the correlation coefficient describing the linear regression fit was low. Although there was a trend of declining scores in the sit-up data there was considerable variation and the trend was not significant. These data suggest that there has been little change in muscular endurance of recruits in the time period examined.

One potential confounder for APFT events (PUs, SUs, and two-mile run) is the fact that the standards for the test have changed over time (⁹¹). However, recruits are not required to achieve the standard on entry to service, only near the end of training. The purpose of the initial APFT is generally to provide the recruits and trainers a “baseline” level of fitness for physical training. Thus, it is unlikely that the change in standards over time effected the initial measurements analyzed here.

One measure of female leg power (vertical jump) was higher in 1998 than in 1993. Caution is suggested in interpretation because of the short period of time (5 years) and the fact that measurement methods were slightly different in the 2 investigations. There was little change in the standing long jump performance of youth between 1965 and the 1980's (¹²³). In early periods (1958–1975) there appears to be a secular increase in performance on this test (^68,123), presumably due to national efforts to improve testing results in schools (¹²³). Studies in other countries have reported mixed results with secular performance declines or little change in standing broad jump performance among Lithuanian and Finnish males and females, (^70,181) but an increase in performance on the vertical jump among Swedish males (¹⁸⁶).

There are several limitations to the analyses reported here. Limitations with regard to BMI and APFT measures (PUs, SUs, and 2-mile runs) have been discussed above. Studies on which this review is based have involved convenience samples and not random samples of recruits. Nonetheless, BCT and OSUT training sites draw recruits from across the entire US and are likely representative of the US population at any particular time. With one exception (¹⁶¹), data were not collected to specifically examine longitudinal trends. For some fitness measures the number of studies and subjects was low with the possibility that the sample was not representative of the year indicated but all available data was used in the analyses for completeness and to follow the ad hoc research plan. Methods for measuring body composition differed between studies, although when multiple measures were provided skinfolds were selected for analysis since most investigations used this method. Nonetheless, skinfold sites and equations used to estimate body composition differed, adding variability to the body composition values. On some variables like muscle strength and V̇o₂max there was very limited data and further studies could clarify the trends reported here.

Practical Applications

This analysis indicated that secular changes in the physical fitness of recruits differed depending on the component of fitness measured. Efforts should be made to more systematically track fitness in the US to compliment the efforts made here. Current work by the CDC to conduct the National Youth Fitness Survey is promising in this regard. Fitness surveys have been administered in the past (^26,68,180) and organizations like the American Alliance for Health, Physical Education, Recreation and Dance have historically provided national fitness test batteries (^15,26,68) but there was little in the way of a systematic effort to develop a nation-wide database. For tracking self-reported height and weight, a publicly available database called the BRFSS does exist (https://apps.nccd.gov/brfss/trends) and the NHANES has provided periodic measurement of height and weight (^43,47,137). In the US Army, measures of height, weight, PU performance, SU performance, and 2-mile run times are routinely collected during recruit training and twice a year in operational Army units (¹⁴⁴) but the data are not placed into an overarching database. There are efforts in progress to change this test (⁹¹). Nonetheless, a comprehensive fitness surveillance system that provides overlapping fitness measures from civilian and military sources would allow systematic tracking of fitness trends, periodic assessments, and comparisons between populations.