Rowing, a primarily endurance and weight-supported type of sport, is associated with a lower rate of injuries compared to other sports (9). It has been reported that injuries in rowing occur as a result of the intensive training regime that the rowers follow (12), whereas others blame the rowing action or technique (1). Rowing technique requires high levels of dedication in every stroke. The stroke begins with flexed legs, forward lean, and extended arms and ends with extended legs, backward lean, and flexed arms, thus involving virtually every muscle group. This recent description of a new rowing technique, which emphasizes the drive off the legs, raises the incidence of low back pain (18).
There are 2 types of rowing, known as sweep and scull (Figure 1). Scullers have control of 2 oars, whereas sweep rowers are responsible for 1 oar. Sweep rowers are further defined depending on which side they row. Port-side rowers have their oars to their right (port side of the boat), whereas starboard-side rowers have their oars to their left (starboard side of the boat) (14). Sweep rowing involves the oarsmen loading the back in a rotated and flexed position. Repetition of an asymmetry activity may lead to the development of muscle asymmetry injury (11) and reinjury (15). Hides et al. (10) noted right to left differences in terms of muscle cross-sectional area in subjects with and without low back pain, suggesting that muscle asymmetry may be important in the development of low back pain.
Additionally, muscular tightness, which restricts the range of motion, is also believed to predispose the muscle to injury and to impair performance (20). Harvey (8) suggested that adequate flexibility prevents overuse injuries and undue stress on the body andhelps achieves a good technique. Moreover, adequate hip flexibility allows the thighs to approximate the trunk in the catch position. Previous study investigating the effects of shortened hamstring length on the upright and toe touching position found that decreased hamstring length led to a decrease in lumbar flexion and an increase in thoracic flexion (5).
Previous studies have investigated the relationship between imbalance in muscle strength and the occurrence of injuries, including scullers and sweep rowers. Karlson (14) identified asymmetry in the isokinetic strength of the quadriceps muscle group in oarsmen. Koutedakis et al. (16) noted a low hamstring to quadriceps strength ratio, suggesting weakness of the hamstring muscle groups in rowers with low back pain. They suggested that this abnormal hamstring/quadriceps ratio might interfere with the lumbo-pelvic rhythm, leading to increased stress on the spine.
The causes of these imbalances are not known but could be to because of the nature of sweep rowing. It remains unclear, however, if uninjured and sweep rowers present strength or mobility asymmetries in the muscles of lower extremities. There is concern that the incidence of low back pain is rising, particularly in stroke side compared to bow side rowers. Therefore, the present study was designed to investigate lower extremity asymmetries in isokinetic strength and joint mobility according to sweep side in experienced oarsmen.
Experimental Approach to the Problem
To examine lower extremities asymmetries according to rowing side, isokinetic strength and joint mobility in the muscles of lower extremities between ports and starboards were assessed using a 2 × 2 (side by leg) analysis of variance (ANOVA) model for repeated measures. In particular, both left and right peak torque of extensors and flexors at 60 and 180°·s−1 isokinetic speed, and joint mobility for knee, hip, and ankle, were measured. All measurements were performed at the end of the preparatory training period.
Twenty six well-trained Greek male rowers volunteered to participate in this study. The rowers were divided into 2 groups as follows: twelve straboard (S; n = 12, age 18.6 ± 0.5 years, height 181.6 ± 1.4 cm, weight 77.6 ± 2.5 kg, training age 5.5 ± 0.5 years) and 14 port (P; n = 14, age 20.6 ± 1.3 years, height 182.2 ± 1.8 cm, weight 78.8 ± 1.6 kg, training age 6.1 ± 0.9 years). All rowers were training as port or starboard and had been doing so for the last 5.55 ± 0.52 and 6.09 ± 0.95 years, respectively (Table 1).
Most rowers were members of the Greek rowing national team, some of these had been national champions, and others were placed second or third in the Greek national championship. Both groups were following the training program under guidelines of Greek Nation Rowing Federation including indoor and outdoor rowing, running and 3-up to 5-training sessions of weight training. The participants did not change their rowing side when lifting weights, and they did not insist in nonoarside extremities in weight room or with flexibility training. Measurements took place at the end of preparatory training period (i.e., in February). The rowers were familiarized with all laboratory procedures.
All participants were asked to maintain their current diet and refrain from any physical activity within 24 hours of testing and to abstain from eating for 3 hours before testing. On test days, the rowers were asked not to drink coffee or other caffeine-containing beverages. Each rower was tested at the same time of the day between tests. Each subject volunteered to participate in the study after being fully informed of the nature, risks, and benefits of participation in the investigation and signed an informed consent form. All participants were nonsmokers and did not use any nutritional supplements or medications known to affect their performance. They reported no musculoskeletal injuries of the lower limbs. All experimental procedures were performed in accordance with the policy statement of the American College of Sports Medicine on research with human subjects as published in Medicine & Science in Sports & Exercise and were approved by the Institutional Review Board for use of Human subjects.
Initially, rowers reported to the laboratory on 2 separate occasions. Intervals of at least 48 hours were intermitted among sessions. Session 1 included assessment of anthropometric measurements and isokinetic strength measurements. During the second session, joint mobility was measured. All tests were conducted in an air-conditioned laboratory at sea level (mean room temperature 25 ± 2°C, relative humidity 45 ± 5% and barometric pressure 760 ± 5 mm Hg).
Isokinetic Strength Measurements (Session 1)
Subjects' body mass was measured to the nearest 0.1 kg with subjects lightly dressed and barefooted. Standing height was measured to the nearest 0.5 cm and body surface area (m2) was calculated (Seca 220e, Hamburg, Germany).
Maximum isokinetic strength was recorded as torque of the quadriceps and hamstrings muscles at angular velocities of 60 and 180°·s−1. Peak torque was measured using a speed-controlled isokinetic dynamometer (Cybex II, Lumex Inc., Ronkonkoma 11779, New York, NY, USA).
Before testing, each subject underwent a standardized warm-up on a Monark cycle ergometer for 5 minutes with no resistance at 60 rpm before all strength assessments. This exercise was followed by 10 minutes of static stretching of the hamstrings and quadriceps. For each angular velocity, peak isokinetic torque was recorded simultaneously, and the torque generated by the limb weight and the dynamometer arm was extracted from the obtained data.
The subjects were seated on the chair of the dynamometer, with stabilization straps at the trunk, thigh, and tibia to prevent extraneous joint movement. The knee to be tested was positioned at 90°·s−1 of flexion (0°·s−1 fully extended knee) to align the axis of the dynamometer lever arm with the distal point of the lateral femoral condyle. Subjects were instructed to kick the leg as hard and as fast as they could through a complete range of motion. Verbal encouragement was given during every trial.
Three repetitions were carried out at each angular velocity, and the highest peak torque value was used for the analyses. A 30-second rest period was taken between each trial and a 60-second rest period between each velocity measurement.
Joint Mobility Measurements (Session 2)
The flexibility of the participant's hip joint (in flexion) was measured using a goniometer (Myrin Flexometer, Medical Research Ltd, Leeds, United Kingdom). This instrument is a circular scale with a weighted pointer controlled by gravity attached to the center. The flexometer was strapped to the segment being tested.
The same pair of testers, a physiotherapist, and an orthopedic surgeon performed all measurements. Both had previous experience in measuring joint motion. They followed written instructions defining positions of the extremity and all aligment of the goniometer. The terminology used was that recommended by the American Academy of Orthopedic Surgeons. Environmental influences were standardized by the measuring at the same time of the day and at a room temperature of 20°C. No warm-up exercises were done.
Each session repeated the entire measuring procedure including positioning of the oarsman marking of anatomic land marks and positioning of measuring devices. Each trial repeated the range of movement without changing the positioning of the oarsman or the measuring devices and while using the same anatomic land markings. One trial was performed on both right and left sides in each session.
With the subject supine on the bench, the flexometer was strapped to the lateral side of the thigh 5 cm above the patella and was adjusted to zero. Velcro bands immobilized the pelvis and opposite leg. Examiner 1 then slowly raised the subject's leg keeping both hands on the knee. Examiner 2 read the result when examiner 1 felt the knee flex.
The subject lay prone on a bench while the flexometer was strapped to the leg above the lateral malleolus and adjusted to zero. Examiner 1 passively flexed the knee, and examiner 2 read the result at the movement hip flexion started.
Ankle Dorsiflexion, Knee Flexed
Standing on the floor with the foot of the leg to be tested on a bench, the subject leaned forward to produce a maximal ankle dorsiflexion with the heel in contact with the bench and knee maximally flexed. The examiner checked with 1 hand that the subject's heel was in contact with the bench.
The distribution of all depended variables was examined by the Shapiro-Wilk test and was found not to differ significantly from normal. Two-way repeated measures ANOVA (side × leg) was used for the analysis of rowing side and leg's peak torque. To determine the meaningfulness of the effect of rowing side on each dependent variable, effect sizes were calculated using the following equation:
(peak torque and mobility). Effect sizes of 0.2, 0.6, 1.2, 2.0, and 4.0 were considered to be small, moderate, large, very large, and nearly perfect, respectively, according to a modified Cohen's scale (2) (http://newstats.org). For comparison, the values of the original Cohen's scale are 0.2 for small, 0.5 for moderate, and 0.8 for large effects (2). Data are presented as mean ± SE. The test-retest reliability of peak torque and mobility was determined by performing the infraclass reliability test. The level of statistical significance was set at α = 0.05.
Descriptive statistic in port and starboard for all variables is presented in Table 2 (mean ± SD). The 95% confidence interval for means with lower and upper bounds are presented in Figures 2 and 3 for joint mobility and peak torque, respectively.
Quadriceps Peak Torque at 60 and 180°·s−1
The 2-way interaction for side × leg (p < 0.001) was significant at 60°·s−1. Pairwise analysis showed that ports exhibited a significantly higher quadriceps peak torque compared to starboards in the right leg (p = 0.001, 95% CI 34.43-104.59) and that starboards exhibited a higher quadriceps peak torque in the left leg compared to the right leg (p = 0.044, 95% CI −73.84 to 1.01) (Table 3). The effect size for the quadriceps peak torque at 60°·s−1 between right and left legs for ports and starboards in the actual value was 0.94.
The 2-way interaction for side × leg (p < 0.001) was significant at 180°. Pairwise analysis showed that ports exhibited a significantly higher quadriceps peak torque compared to starboards in the right leg (p = 0.004, 95% CI 13.59-61.75) otherwise starboards exhibited a higher quadriceps peak torque in the left leg compared to in the right leg (p = 0.020, 95% CI −43.77 to 4.13) (Table 3). The effect size for quadriceps peak torque at 180°·s−1 between the right and left legs for ports and starboards in the actual value was 0.94.
Hamstrings Peak Torque at 60 and 180°·s−1
The 2-way interaction for side × leg (p < 0.001) was significant at 60°·s−1. Pairwise analysis showed that ports exhibited a significantly higher hamstrings peak torque compared to starboards in the right leg (p = 0.05, 95% CI 0.53-37.30). No significant difference was observed for starboards in the left hamstrings peak torque compared to in the right (p > 0.05, 95% CI −21.65 to 12.62) (Table 3). The effect size for hamstrings peak torque at 60°·s−1 between the right and left legs for ports and starboards in the actual value was 0.61.
The 2-way interaction for side × leg (p < 0.001) was significant at 180°·s−1. Pairwise analysis showed that ports exhibited a significantly higher hamstrings peak torque compared to starboards in the right leg (p = 0.004, 95% CI 7.95-35.87). No significant difference was seen for starboards in the left hamstrings peak torque compared to the right (p > 0.05, 95% CI −13.06 to 13.49) (Table 3). The effect size for hamstrings peak torque at 180°·s−1 between right and left legs for ports and starboards in the actual value was 0.84.
The 2-way interaction for side × leg (p < 0.001) was significant for hip mobility. Pairwise analysis showed that ports exhibited a significantly higher hip mobility compared to starboards in the right leg (p = 0.011, 95% CI 1.60-11.28) and starboards exhibited a higher hip mobility in the left leg compared to in the right leg (p = 0.001, 95% CI −14.55 to 4.33) (Table 3). The effect size for hip mobility between the right leg and the left leg for ports and starboards in the actual value was 0.94.
The 2-way interaction for side × leg was not significant for knee mobility. Pairwise analysis showed no significant difference between ports and starboards in both the right (p > 0.05, 95% CI −16.28 to −3.70) and left (p > 0.05, 95% −13.23 to 3.05) knee mobilities (Table 3). The effect size for hip mobility between the right and left legs for ports and starboards in the actual value was 0.52.
The 2-way interaction for side × leg (p < 0.001) was significant for ankle mobility. Pairwise analysis showed no significant difference between ports and starboards in the right leg (p > 0.05, 95% CI −0.23-6.93). Otherwise starboards exhibited higher ankle mobility in the left leg compared to in the right leg (p = 0.001, 95% CI −13.15 to −4.72) (Table 3). The effect size for ankle mobility between the right and left legs for ports and starboards in the actual value was 0.87.
The aim of this study was to determine the effects of sweep rowing on isokinetic strength and joint mobility of lower extremities in oarside rowers. Our findings demonstrate a bilateral strength difference in favor of the oarside. Port-side rowers exhibit a significantly higher peak torque in oarside-right quadriceps and hamstrings at 60 and 180°·s−1 velocities compared to the left side. In a respective manner, starboard-side rowers showed a higher peak torque in left quadriceps and hamstrings at 60 and 180°·s−1 compared to the right side rowers. Furthermore, right hip mobility was significantly higher in port compared to starboard, whereas left hip and ankle mobilities were lower in port compared to starboard.
Because of the highly competitive nature of rowing and the associated oarside training, it is reasonable to expect asymmetry and imbalances in the leading muscle groups. A strength discrepancy of 10 ± 15% or more between the 2 sides is considered to represent a significant asymmetry (3,6,13). Parkin et al. (19) reported that stroke side had no influence on leg strength although influence between the left and the right erector spinae muscles was significantly related to rowing side. It is difficult to compare the results of Parkin et al. (19), with those of the present study, because in our study, port and starboard were distinguished. In a previous work by Kramer et al. (17), a tendency for light-weight sweep rowers to develop greater strength in the oarside knee extensors at velocities of 160 and 200°·s−1 was reported. Our data are in agreement with these findings
Early studies have shown that a bilateral leg strength difference of 10% or greater may be a factor contributing to injury (7,21). Knapik et al. (15) suggested that players with a strength imbalance of >15% were 2.6 times more likely to suffer injury in the weaker leg. Fowler and Reilly (4) reported a 20% difference in bilateral muscle strength in professional soccer players prone to injury. In our study, all participants had severe (up to 12%) asymmetries in lower extremities in either quadriceps or hamstrings. This fact may increase the incidence of injury (7).
The findings of this study also showed significant bilateral imbalance in flexibility, that is, hip and ankle flexibility was higher in row-side compared to that in non-row side. Muscular tightness, which restricts the range of motion, is also believed to predispose the muscle to injury and to impair performance in sports where flexibility is important (20). In particular, rowers require adequate flexibility to prevent overuse injuries and undue stress on the body and to achieve a good technique. Kramer et al. (17) reported that hip flexibility allows the thighs to approximate the trunk in the catch position.
More recently, another study investigating the effects of shortened length on the upright and toe touching position found that decreased hamstring length led to a decrease in lumbar flexion and an increase in thoracic flexion (5). These findings suggest that decreased hamstring length may lead to overuse injury with the repetitive rowing action. Previous data of Gajdosik et al. (5) are not comparable with that of our study because in the former study measurements were performed in the standing position replicating the forward reach and catch phase of rowing. However, in support of our findings, Harvey (8) concluded that rowers had more flexibility in the quadriceps but less flexibility in the illiopsas. This might be a possible explanation for any potential difference in knee mobility between port and starboard.
We concluded that sweep rowing produces asymmetries in isokinetic strength and increase flexibility in lower extremities depending on rowing side. Further investigations are required to examine additionally the rowing side effects on cross-sectional area and any possible correlation between side asymmetry and low back pain.
Based on the findings of the present study, it is clear that sweep rowing produces a unique lateral dominance in lower extremities. Sweep rower presents a higher peak torque in oarside knee flexors and extensors than that in nonoarside-right side for port and left side for starboard. Furthermore, joint mobility is quite higher in oarside compare to that in nonoarside in either port or starboard. All the abovementioned asymmetries may predispose sweep rowers to injury. In particular, the limitation in isokinetic asymmetries and mobility of flexor muscles dramatically increases the risk for low back pain in rowers. Therefore, coaches when designing the strengthening and stretching training programs should bear in mind the fact that they may have to compensate for any difference between side and nonoarside. Thus, prophylactic strengthening and stretching programs should aim at correcting strength and mobility imbalance and thereby reducing the occurrence of injuries.
The authors would like to thank Dr. Vassilio Karagianni, Research Associate, Department of Fundamental Dental Sciences, School of Dentistry, Aristotle University of Thessaloniki, for his statistical support.
1. Christiansen, E and Kanstrup, IL. Increased risk of stress fractures of the ribs in elite rowers. Scand J Med Sci Sports
7: 49-52, 1997.
2. Cohen, J. Statistical Power Analysis for the Behavioral Sciences
. Hillsdale, NJ: Lawrence Erlbaum, 1998.
3. Elliot, J. Assessing muscle strength isokinetically. J Am Med Assoc
240: 2408-2410, 1978.
4. Fowler, NE and Reilly, T. Assessment of muscle strength asymmetry in soccer players. In: Contemporary Ergonomics
. Lovesey, EJ, ed. London, United Kingdom: Taylor and Francis, 1993. pp. 327-332.
5. Gajdosik, R, Albert, C, and Mitman, J. Influence of hamstring length on the standing position and flexion range of motion of the pelvic angle, lumbar angle and thoracic angle. J Ortho Sports Phys Ther
20: 213-219, 1994.
6. Gleim, G, Nicholas, J, and Webb, J. Isokinetic evaluation following leg injuries. Phys Sports Med
6: 75-82, 1978.
7. Grace, TG, Sweetser, ER, and Nelson, MA. Isokinetic muscle imbalance and knee-joint injuries. J Bone Joint Surg
66: 734-739, 1984.
8. Harvey, D. Assessment of flexibility of elite athletes using the modified Thomas test. Br J Sports Med
32: 68-70, 1998.
9. Hickey, GJ, Fricker, PA, and McDonald, WA. Injuries to elite rowers over a 10 year period. Med Sci Sports Exerc
10. Hides, JA, Stokes, MJ, Saide, M, Jull, GA, and Cooper, DH. Evidence of lumbar multifidus muscle wasting ipsilateral to symptoms in patients with acute/subacute low back pain. Spine
19: 165-172, 1994.
11. Hodges, PW and Richardson, CA. Inneficient muscular stabilization of the lumbar spine associated with low back pain. Spine
21: 2640-2650, 1996.
12. Jones, BH, Cowan, DN, and Knapic, JJ. Exercise, training and injuries. Sports Med
18: 202-214, 1994.
13. Kannus, P. Isokinetic evaluation of muscular performance: Implications for muscle testing and rehabilitation. Int J Sports Med
15: S11-S18, 1994.
14. Karlson, AK. Rib stress fractures in elite rowers: A case series and proposed mechanism. Am J Sports Med
26: 516-519, 1998.
15. Knapik, JJ, Bauman, CL, Jones, BH, Harris, JM, and Vaugh, L. Pre-season strength and flexibility imbalances associated with athletic injuries in female collegiate athletes. Am J Sports Med
19: 76-81, 1990.
16. Koutedakis, Y, Frischknecht, R, and Murthy, M. Knee flexion to extension peak torque ratios and low back injuries in highly active individuals. Int J Sports Med
18: 290-295, 1997.
17. Kramer, J, Leger, A, and Morrow, A. Oarside and non-oarside knee extensor measures and their relationship to rowing ergometer performance. J Orth Sport Phys Ther
14: 213-219, 1991.
18. McGragor, AH, Anderton, L, and Gedroyc, WMW. The trunk muscles of elite oarsmen. Br J Sports Med
36: 214-217, 2002.
19. Parkin, S, Nowicky, AV, Rutherford, OM, and McGregor, AH. Do oarsmen have asymmetries in the strength of their back and leg muscles? J Sports Sci
19: 521-526, 2001.
20. Rahnama, N, Reilly, T, Lees, A, and Graham-Smith, P. A comparison of musculoskeletal function in elite and sub-elite English soccer players. In: Kinanthropometry VII
. Reilly, T and Marfell-Jones, M, eds. London, United Kingdom: Routledge, 2003. pp. 151-164.
21. Wyatt, MP and Edwards, AM. Comparison of quadriceps and hamstring torque values during isokinetic exercise. J Orthopedic Sports Phys Ther
3: 48-56, 1991.