🧬 biology

Resistance Training Improves Integrated Health-Related Fitness in University Students: A Phenotype-Based Longitudinal Analysis

This longitudinal study utilized a hierarchical principal component analysis and K-means clustering to identify three distinct, stable, yet modifiable integrated health-related fitness phenotypes among university students, demonstrating that a 16-week resistance training intervention significantly facilitated upward transitions in fitness profiles for a subset of participants.

Original authors: Sung-Wook Chun, Soo-Min Ha, YuKun Han, JeongHoon Park, Sangho Lee, Na-Young An

Published 2026-08-07

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Sung-Wook Chun, Soo-Min Ha, YuKun Han, JeongHoon Park, Sangho Lee, Na-Young An

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ⚕️ This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer

Imagine your health as a complex video game character. In the old days, coaches and doctors would check your stats one by one: "How strong are your arms?" "How fast can you run?" "How much do you weigh?" They treated these as separate scoreboards. But in the real world, your body doesn't work in isolation; your muscles, your fat, and your heart all talk to each other. Scientists call this "health-related fitness," a big umbrella term that covers everything from how much muscle you have to how well your heart pumps blood. The tricky part is that these stats often get analyzed separately, like looking at a puzzle piece by piece instead of seeing the whole picture. Researchers wanted to know: Can we combine all these different stats into one single "health score" that tells us who is truly fit, not just in one area, but across the board? And if we give these students a workout, does their whole "character" level up, or just one specific stat?

This study, conducted with university students, decided to build a new kind of health report card. Instead of listing ten different numbers, the researchers used a clever mathematical trick to blend body composition (like muscle and fat) and physical fitness (like strength and running) into a single, integrated "Health Score." They then grouped the students into three teams: Low, Moderate, and High fitness phenotypes. Think of these not as fixed labels, but as different "character builds" in a game. The "Low" build had less muscle and more fat with lower strength, while the "High" build was the opposite. The researchers then watched what happened over 16 weeks. One group of students did a supervised weight training class, while the other group took a health management class without heavy lifting.

Here is what they found. First, the new "Health Score" worked perfectly. It successfully sorted the 259 students into three distinct groups that made sense: the Low group had the lowest muscle and strength, the High group had the most, and the Moderate group was right in the middle. The math was solid, with the groups staying stable even when the researchers tested the data over and over.

When it came to the 16-week workout, the results were a mix of stability and subtle change. Most students, about 89.1%, stayed in the same "character build" they started with. They didn't jump from the "Low" team to the "High" team. However, the students in the resistance training group did show a real, measurable improvement in their overall Health Score compared to the control group. It's like they gained experience points and leveled up their stats, even if they didn't change their entire character class. About 11.4% of the weight-training group managed to "level up" their phenotype, moving from a lower fitness group to a higher one, while only 5.5% of the non-training group did the same. Very few people (only 1.2%) got worse.

The study suggests that while it takes a lot of effort to completely change your overall fitness "type" in just 16 weeks, resistance training definitely improves the underlying numbers that make up that type. The researchers are careful to say this is a promising new way to look at fitness, but it's still an early-stage tool that needs to be tested on more people before it becomes the standard rulebook for everyone. They also noted that their method used a specific type of body scanner (bioelectrical impedance) and that the groups had different numbers of men and women, which might influence the results. But the main takeaway is clear: looking at fitness as a single, combined story gives a clearer picture than looking at the chapters separately, and lifting weights helps write a better story, even if the chapter title doesn't change immediately.

Technical Summary: Resistance Training Improves Integrated Health-Related Fitness in University Students: A Phenotype-Based Longitudinal Analysis

Problem Statement
Health-related fitness (HRF) encompasses both physical fitness and body composition, yet these domains are frequently analyzed in isolation or summarized via simple composite scores. While such methods describe group-level differences, they fail to capture how fitness and body composition interact within specific individuals over time. Existing multivariate approaches often treat fitness and body composition as separate profile domains or rely on cross-sectional designs. Consequently, there is a methodological gap in classifying university students into ordered, multidimensional HRF phenotypes that integrate field-based fitness measures with bioelectrical impedance analysis (BIA)-derived body composition, and in tracking longitudinal transitions within these integrated profiles.

Methodology
This study employed a secondary analysis of data from a 16-week university-based exercise intervention involving 259 participants (186 in a resistance training group, 73 in a control group). The study utilized a hierarchical Principal Component Analysis (PCA) framework followed by K-means clustering to develop an integrated HRF phenotype model.

Data Reduction:
- Fitness Domain: Four field-based tests (grip strength, modified sit-up, 20-m shuttle run, sit-and-reach) were reduced to a single fitness-domain component (PC1) via PCA, explaining 52.67% of the variance.
- Body Composition Domain: Three BIA-derived variables (body fat mass, body fat percentage, skeletal muscle mass) were reduced to a body composition-domain component (PC1), explaining 68.58% of the variance. Note: Body fat variables were inverted prior to standardization to ensure higher scores reflected favorable profiles.
- Integration: The two domain-specific PC1 scores were standardized and entered into a final PCA to generate a single "Integrated HRF Score," explaining 84.02% of the total variance.
Phenotype Classification:
- K-means clustering ( $K=3$ ) was applied to the baseline Integrated HRF Scores to classify participants into three ordered phenotypes: Low ( $n=90$ ), Moderate ( $n=99$ ), and High ( $n=70$ ).
- Cluster validity was assessed using internal indices (silhouette coefficient, Calinski–Harabasz, Davies–Bouldin), and stability was confirmed via 500 iterations of repeated subsampling (mean Adjusted Rand Index = 0.917).
Longitudinal Analysis:
- Follow-up assessments at 16 weeks were processed using the baseline PCA loading matrices and standardization parameters to ensure comparability.
- Phenotype transitions were categorized as Improved (movement to a higher phenotype), Stable (remaining in the same phenotype), or Declined (movement to a lower phenotype).
- Statistical analysis included linear mixed-effects models to assess group-by-time interactions and Bowker's test for symmetry to evaluate transition asymmetry.

Key Results

Phenotype Distinctiveness: The three identified phenotypes showed clear, graded differences across all input variables. The High phenotype exhibited superior muscular strength, endurance, cardiorespiratory fitness, and skeletal muscle mass, alongside lower adiposity, compared to the Low phenotype. Notably, Body Mass Index (BMI) did not differ significantly across phenotypes, whereas body fat percentage and skeletal muscle mass did, highlighting the limitations of BMI in distinguishing fitness profiles.
Intervention Effects: Resistance training significantly improved the continuous Integrated HRF Score compared to the control group (Group $\times$ Time interaction: $\beta = 0.139$ , 95% CI 0.047–0.230, $p=0.003$ ).
Phenotype Transitions: Over the 16-week period, the majority of participants (89.1%) remained in their baseline phenotype category. However, a subset showed upward transitions: 9.7% of the total sample improved to a higher phenotype, while only 1.2% declined.
Asymmetry: Bowker's test indicated significant asymmetry in transitions ( $\chi^2 = 17.37$ , $p < 0.001$ ), with a directional bias toward upward movement rather than downward decline.
Stability: The clustering solution demonstrated high robustness, with 99.1% of participants in the resistance training group and 93.2% in the control group maintaining their phenotype status, suggesting that while individual variables may change, the overall multidimensional profile is relatively stable over short-to-moderate periods.

Significance and Claims
The authors claim that this study successfully developed a hierarchical PCA-based framework that integrates field-based fitness and body composition into a unified, interpretable phenotype model. The primary significance lies in shifting the analytical focus from isolated component changes to integrated multidimensional profiles.

The paper posits that:

Integrated Assessment: Combining fitness and body composition via hierarchical PCA allows for the identification of ordered HRF profiles (Low, Moderate, High) that are more informative than isolated metrics or simple BMI categorization.
Complementary Perspectives: Continuous changes in the Integrated HRF Score (detected by linear mixed models) and categorical phenotype transitions (detected by clustering) offer complementary views of exercise adaptation. The study demonstrates that significant physiological improvements can occur even when an individual does not cross the threshold to a new phenotype category.
Precision Exercise: This framework provides a foundation for "precision exercise assessment," potentially helping to distinguish students with unfavorable multidimensional profiles who may require targeted interventions, even if their BMI or single fitness test scores appear normal.

The authors maintain a modest stance, noting that the framework is an early-stage assessment tool requiring external validation in diverse populations (e.g., clinical, athletic, older adults) and that the 16-week duration may have been insufficient for many participants to achieve the magnitude of change required for phenotype-level transitions. They emphasize that the study describes a profiling method rather than a predictive model for phenotype change.

Technical Summary: Resistance Training Improves Integrated Health-Related Fitness in University Students: A Phenotype-Based Longitudinal Analysis

More like this