Assessment of Isokinetic Knee and Shoulder Strength in Elite Youth Boxers: Correlation with Punch Impact Force

This study aimed to assess the isokinetic strength performance of the knee and shoulder in amateur boxer, and further to explore the contribution of knee and shoulder peak torque to different punches. Boxers (6 females and 8 males) were measured the dynamic concentric torque of knee exion(F) and extension(E), shoulder external rotation (ER) and internal rotation (IR) at 60°/S, 180°/S. We tested the punching force of three techniques by Kistler Instruments. At 60°/S, F/E ratios of the dominant (D) knee of female boxers were lower than those of non-dominant (ND) side (p=0.03). At two angular velocities, all boxers demonstrated higher ER/IR ratios in ND (p=0.011 for males, p=0.025 for females). Bilateral difference of knee exor in females was greater (p=0.038) than that in males. The maximum correlation was found at 180°/S between punching forces and the joints strength. ANOVA analyses revealed that athletes with greater punching force tend to have higher knee strength. Although correlations between shoulders strength and punching forces were strong, they did not discriminate between athletes who punched with higher punching force. In order to effectively transfer muscle strength to punching force, rapid completion of action should be highlighted in strength training, especially lower-limb strength.


Introduction
Amateur boxing is an Olympic sport. To be the winner, boxers must deliver clear and powerful punches to score or knock down opponents. There are three punching techniques: straight, hook and uppercut, impact force of which is one of the vital performance indicators and the key factor to the victory of the competition 1 , 2. It was reported that the impact force was more elevated in elite boxers than in mediocre boxers and higher in the mediocre boxers than in novices 3. These studies showed that it was paramount for boxers to enhance their punching force, which was in uenced by the muscle strength of upper and lower limbs. Therefore, it is necessary to explore the contribution of strength of knee and shoulder to punching force 4 , 5.
Isokinetic strength test, known as "gold standard" method, is a reliable and safe approach to measure the strength of muscles 6, to optimize the athletic performance and to prevent sports injury. It was reported that once the bilateral or ipsilateral limbs strength asymmetry exceeded a certain scope, not only the athletic performance would be negatively in uenced but also the athletes7, especially the young athletes whose skeleton and muscle systems are nevertheless immature, tended to suffer from muscle strain, anterior cruciate ligament (ACL)sprains and rupture 8 , 9.Many studies have examined the isokinetic strength characteristics of athletes in diverse sports 10-12. However, few studies have engaged in characteristics of isokinetic strength in boxers. Only one has assessed isokinetic strength characteristics of shoulder's rotators muscles of amateur boxer athletes by far 13, yet there is no report about the lower limbs' isokinetic strength characteristics. It's necessary to complement in this area of research.
During the last decade, there has been growing recognition of the signi cant association between muscle strength and sport performance 14. In fact, the correlation of muscle strength with sport performance has been studied in many sports, such as football 15, basketball 16, handball 11, water-polo 17. However, it is known that there is no report about the correlation of boxers' upper and lower limbs peak muscle torque with punching force. Only one study described the relationship between punching force and muscular strength by countermovement bench throw and jump18. Previous study lacked the speci c quantitative value on muscular strength. The aim of the present study, therefore, was to (1) analyze isokinetic strength characteristics of knee and shoulder of boxers (2) examine the relationship between the peak torque of knees and shoulders and the punching force of three techniques.

Participants
Fourteen (six females and eight males) young amateur boxers participated in the present study, with the average age 17.86±0.9 years, height 171.21±4.05cm, weight of males 67.43±4.42kg, weight of females 60.35±2.97kg. All the players self-reported the right limb as their dominant limb. All participants were professional-level athletes and they have participated in national-level youth competitions. None of them have injured within six weeks before the test. At the beginning of the study, after they were informed the bene ts and potential risks, they gave informed consent to participate. The study was approved by ethics committee of Shanghai University of Sport (102772021RT029) and in accordance with the Helsinki declaration.
Isokinetic Testing-knee Isokinetic concentric torque of knee extensors and exors was measured by an isokinetic dynamometer IsoMed2000 (Germany) at 60°/s and 180°/s. According to the instruction manual for the IsoMed 2000 Systems, the participant was seated in appropriate and stable position using xed straps to avoid compensatory trunk movement. The rotation axis of the machine and the knee joint were aligned, and the gravity compensation were performed. Before isokinetic test, each player performed a 15-20 minutes warm-up including stretching quadriceps, hamstrings, adductors, iliopsoas and gluteus. During continuous (bidirectional) knee exion-extension movements the concentric isokinetic torque of the quadriceps and hamstrings was recorded at the angular velocities of 60º/s and 180º/s, through a knee moving range from 0° ( exed) to 90° (full extension). Five maximal exion and extension were performed at the two different angular velocities with 1-minute recovery between series. During the test, the participants must develop the maximum force at the same time during the two phases of extension and exion. The test began with the dominant limb (D), then non-dominant limb (ND).

Isokinetic Testing-shoulder
Isokinetic concentric torque of shoulder external rotation (ER) and internal rotation (IR) was measured by an isokinetic dynamometer IsoMed 2000 (Germany) at 60°/s and 180°/s. participants were seated with 45° of shoulder abduction in the scapular plane, which was regarded as reliable for external rotation and internal rotation strength assessment. The trunk and the hip were kept stable by two straps cross-placed in front of the chest and xed to the back of the chair. The limb measurements began with the D side, then ND. The length of the lever arm was adjusted for each participant, and each limb was weighed and used for gravity compensation. The maximal voluntary contraction of external rotation and internal rotation in angle of 60°/s and 180°/s was repeated ve times with 1 minute of recovery between series.

Maximum punching forces
The equipment employed to assess the punching force was a force plate (Kistler Instruments, Winterthur, Switzerland) xed on the tripod, with a built-in signal ampli er, and the sampling frequency sets at 1000Hz. Athletes were required to punch the target area in full force and then return to the starting position and keep stable and static. Each of three techniques of punches Jab, Cross, Lead hook, Rear hook, Lead uppercut, Rear uppercut were performed ve times with a 15-second resting interval after each punch. Every technique of punches has a 1-minute resting interval. Due to different arm lengths, every participant was allowed to choose their own punching distance for their best performance. Verbal motivation was provided to each boxer to inspire the maximal punching force.
Peak torque of concentric contraction of knee and shoulder was expressed in absolute N and relative to body weight N/kg. The exor to extensor(F/E) peak torque ratio was equal to dividing knee F concentric peak torque by knee E concentric peak torque. The external rotation to internal rotation (ER/IR) peak torque ratio was equal to dividing shoulder ER concentric peak torque by shoulder IR concentric peak torque. The BD% was the percentage difference between the peak torque of the D and ND, as follows: BD% = [(Higher PT -Lower PT)/ Higher PT] × 100. Calculations were performed in Microsoft Excel.

Statistical analysis
All statistical analyses were performed using PRISM (GraphPad Software, Inc. Version 9.0.2 for Windows) and IBM SPSS Statistics (version 23). Descriptive statistics (means and standard deviations) were performed to summarize all data. The Shapiro-Wilk test for data normality was applied. Pearson or Spearman correlation analysis was employed to determine the relationship between maximal punching forces of three techniques of punches and shoulder IR and knee extensor peak torque of the participants.
To exclude the impact of weight, all the comparisons mentioned in the article were relative quantities. As the qualitative evaluation of the degree of relationship, r=0 was considered null, between 0 and 0.3 weak, 0.3 and 0.6 moderate, 0.6 and 0.9 strong, 0.9 and 1 very strong, and 1 full19. Differences between genders and bilateral limbs were analyzed by unpaired two-tailed t-test with Welch correction when normality or equal variance of samples failed. When equal or normality was satis ed, the difference was analyzed using unpaired two-tailed t-test. * P value less than 0.05 was considered signi cant; ** P less than 0.01 was considered extremely signi cant.
The participants were divided into three groups according to their maximal punching forces; lower-level group (LG), medium-level group (MG), and higher-level group (HG), respectively. The rst group contained conducted using Fisher's least signi cant difference (LSD) test. For all statistical tests, signi cant wans P 0.05. The Levene method was used to test the homogeneity of variance. Effect sizes (ES) were calculates using cohen's d: 0.25 is trivial, 0.25-0.50 is small, 0.50-1.0 is moderate, and .1.0 is large20.

Results
The knee relative exors to extensors peak torque (F/E) ratio was illustrated in Table 1and Figure 1. At 60°/S and 180°/S, F/E ratios ranged respectively 0.61-0.63 for males and 0.58-0.63 for females. At 60°/S, The F/E ratios of D knee in female were lower than those of ND side (p=0.03) and lower than that in males (p=0.006). The ratio of relative shoulder ER/IR peak torque was shown in Table 2 and Figure 1. At 60°/S and 180°/S, ER/IR ratios ranged respectively 0.65-0.67 for males and 0.66-0.70 for females. At 60°/S, ND showed greater (p=0.011 for males, p=0.025 for females) ER/IR ratio than D regardless of sex. At 180°/S, ER/IR ratio in the ND showed gender differences that the ER/IR ratios from female were signi cantly higher (p=0.021) than those from the male. BD% values from extensor and exor, IR and ER at two angular velocity were presented in Figure 2. The female athletes demonstrated greater BD% in exor when compared to males 9.45±1.72VS7.61±1.25%;p=0.038 at 180°/S. The absolute and relative values of punching force in Jab, Cross, LH, RH, LU, RU were shown in Table 3. Figure 3 shows the relationship between relative punching forces (RPF) and relative peak torque (RPT). The RPF of Jab had the greatest association with the RPT of D knee at 180°and indicated regular degree r=0.536, p=0.031 . Likewise, correlations between Cross RPF and the RPT of D knee at 180°were greatest, belonging to strong relevance r=0.603, p=0.007 . The RPF of LH and LU exhibited moderate levels of associated with the RPT of ND shoulder at 180° (r=0.535, p=0.035 and r=0.572, p=0.029, respectively). The RPF of RH and RU was greatly associated with the RPT of D shoulder at 180°, which lay in strong and moderate correlation (r=0. 663, p=0.006 and r=0.574, p=0.027, respectively).

Discussion
The study focused on the Isokinetic strength pro le of knee and shoulder young boxers. Based on this, we explored the relationship between peak muscle torques and punching force, which provide advice for coaches and athletics. The study found that:1) at 180°/S knee F/E ratio and shoulder ER/IR ratio were higher than those values at 60°/S, and the gender difference appeared in F/E ratio of D at 60°/S and ER/IR of ND at 180°/S, 2) with the increase in angular velocity, the correlation coe cient between the punching force of three punching techniques and the RPT of knee extensor and shoulder IR was on a rise.
The F/E ratios of boxers increased with the increase in angular velocity. These results were in accordance with outcomes obtained in other sports 11 , 21. This may be due to the fact that exor was characterized by relatively faster muscle bers in comparison with extensor22. In low velocity movement, F/E ratio of D and ND limbs in female showed dramatic difference. This nding was contrary to previous studies on female basketball players which have suggested that there was no signi cant difference in D and ND 16.
The difference between the results of studies may derive from the different modes of power generation in lower limbs. Boxing players prefer to rstly use the D limb to push off the ground, while basketball players jump with simultaneous two limbs pedal and stretch. Unfortunately, since there is no report about isokinetic strength in lower limbs of boxers, it is impossible compare the F/E ratio of knee among boxers.
It was reported that the normal value of knee F/E ratio was higher than 0.6 23, and if it is lower than 0.6,imbalance between the F and E will lower the stability and increase the risk of injury 17 times 24. In this study, F/E ratios in the D limb of female boxers were lower than 0.6 at 60°/S, and lower than those in males. These results seem to be consistent with other research which found that female players were more vulnerable to injury. It was reported that female athletes tended to suffer anterior cruciate ligament (ACL) injure 25.One reason for this may be weaker co-contraction of the exor during knee extension, and stronger F may reduce overload imposed on the ACL 26. To increase the stability of knee and lower the risk of injury, it is necessary to correct the muscle imbalance of female boxer.
In the present study, ER/IR ratio exhibited lower values in the D regardless of gender, and ER/IR increased as the angular velocity did, which was consistent with previous research on boxer 13. Unfortunately, only male players have been included in previous study. After the female players were involved in present study, identical trends were found at two angular velocities. However, only at 180°/S, ER/IR ratios indicated striking sex difference. A possible explanation for this might be that as the angular velocity increased, IR muscles were capable of keeping stronger in males. For this reason, males ER/IR ratios were lower. In present study, only at 60°/S, ER/IR ratios showed signi cant difference within bilateral limb, while previous studies revealed this difference at both 60°/S and 180°/S 27. This inconsistency may be due to the difference in frequency of using two sides of the upper limbs. Previous study was based on volleyball spikers so that the use frequency of D limbs is higher than in ND limbs. But boxers showed similar frequency in use of D and ND limbs.
The value of ER/IR can be used to assess the shoulder dynamic stability. It was reported that ER/IR ratio ranged 0.66-0.75 could prevent the shoulder injuries 28. In our study, the ratio scope was 0.62-0.71. But there was no need to excessively worry about this ratio. Although boxing belongs to combat sports, shoulder strains/sprains happened in boxers was no more, or even lower than that in other sports, while cerebral concussion was the one to be prevented and avoided in boxing 29 , 30.
Female boxers demonstrated higher BD% in shoulder and knee than males. Nevertheless, only the BD% in knee exor at 180°/S showed signi cant difference. Differences were not signi cant at other angular velocity and movement. In general, a 10-15% BD may be applied as a standard to stop athlete from exercise unless the balance recovers to normal range 31 , 32. Athletes with a BD% over 15% were 2.6 times more likely to get injured in weak limbs 33. In our research, only a few athletes demonstrated BD% higher than 10%, but lower than 15%. It is to be emphasized that the BD% will increase as the training experience does 34. Therefore, it is necessary to pay sustained attention to this indicator in order to prevent sport injury and training effects.
This study is the rst to describe the relationship between the RPT of knee and shoulder at two angular velocities and the RPF of three punching techniques. It was found that correlation coe cients were higher at the 180°/S than at 60°/S. These relationships may partly be explained by the fact that the punching power is generated by fast muscle bers, which also determine the power generated by high speed movement 35. As a result, punching force and the strength produced at 180°/S had stronger correlation. Many papers have also studied that the correlation between the explosive strength performance and peak torque in Isokinetic strength testing became greater as the angular velocity increased 16 , 36. In order to improve the punching force, plyometric training and complex training should be given more emphasis in training.
Jab, cross punching force was more positively correlated with the knee extensors peak torque, compared with shoulder IR. The explanation might be that because of the long distance to the target, boxers can fully extend their knee, and the extent of extension has strong correlation with power output 37 , 38. The extension of lower limbs would deliver power to upper limbs which accomplish the punching 39-41. we explained that the increase in muscle power of lower limbs would enhance the punching force from another viewpoint. The correlation coe cient between LH, LU, RH, RU and shoulder was greater, which may be related to other factors that in uence the punching force, including extent of extension of knee and body rotation. Because hooks and uppercuts have shorter distance to the target, knee cannot extend fully. However, the contribution of knee strength is still large. This can be observed from three groups according to the punching force. HG showed greater knee strength compared with LG and MG. There results suggested the importance of lower-limb strength on punching force and it may be a key factor to generate higher punching force. The present study rst described the Isokinetic strength characteristics of knee and shoulder of boxers, and based on which the correlation between the strength of knee and shoulder and the force of three punches was determined. Former studies have reported the relationship between imbalance of knee strength and injury. In addition, athlete should emphasis speed of action in strength training, converting muscle power into punching force, especially lower-limb strength. We were aware of two major limitations in this study. First, the sample size was small. Second, athletes were free to choose their own punching time and distance in the impact force test, which was di cult to represent the real performance in practical competition. It was the limitation in the existing literature on punching performance of combat sport.

Conclusion
We suggested developing the strength of exor to increase F/E ratios and to decrease BD% of knee, especially for females. RPF has stronger correlation with the RPT of knee and shoulder at 180°/S than at 60°/S. Therefore, athletics should emphasize the rapid completion of the movement in training exercise(e.g. plyometric training), which is favor to the transfer from muscle strength to punching force. Low-limbs strength should be paid more attention, because boxers with greater knee strength can produce higher punching force. Future studies about boxers need to focus more on punching performance in the formal competition, since knocking out the opponent is the ultimate goal of boxing.

Declarations Data availability
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request. Figure 1 Scatter plot of knee F/E ratios shoulder ER/IR ratios of D and ND limbs from boxing athletes at 60°/S and 180°/S. *Signi cant difference between D and ND, # signi cant difference between male and female.

Figure 2
Box and whisker presenting 5-95 percentile of bilateral difference in peak torque of F and E of the knee and in peak torque of ER and IR of the shoulder at 60°/S and 180°/S *higher than males.

Figure 3
The analysis of correlation plot between RPT and RPF at two angular velocities. Red: maximum correlation *p 0.05 **p 0.01