Letter From the Editor

In their recent article [2], Sporrong and Styf recorded supraspinatus intramuscular pressure (IMP) and electromyographic (EMG) signal as functions of torque for concentric and eccentric isokinetic shoulder abduction in normal individuals. They concluded that there was a positive relationship for IMP, EMG, and torque for concentric muscle activation, but a negative relationship for eccentric muscle activation (Fig. I , [2]). The authors explained that the difference in responses between concentric and eccentric modes may have been due to “specific conditions affecting the anatomy of the supraspinatus muscle during eccentric activation”. In addition they stated, “we believe that agonistic muscles have different patterns of coactivation during eccentric and concentric muscle activity”. They did not comment, however, on the fact that their results differed substantially from previous work using similar techniques to measure IMP, EMG signal and torque [ 11. In contrast, Aratow et al. [l] noted that torque, EMG signal, and IMP showed similar relationships for concentric muscle activation and for eccentric muscle activation figures 3(A) and 3(B) are almost mirror images of each other-except that slightly higher IMP and torque was noted for the eccentric mode. I seriously doubt that the supraspinatus muscle is that different from the soleus muscle such that eccentric activation in the former would result in a negative relationship between EMG and IMP, while eccentric activation in the latter would result in a positive relationship between EMG and IMP. If these relationships were true, it would not bode well for the future of IMP recording as a reliable index of muscle force. However, I do have a possible explanation for the dissociation noted between EMG, torque, and IMP for eccentric activation of the supraspinatus muscle [2]. I believe an “x-axis” error may have been present during the collection of IMP data. The authors [2] used an isokinetic computerized ergometer (KinCom; Chattanooga Corporation, Chattanooga, TN, USA). This type of ergometer records shoulder torque and EMG data as functions of shoulder abduction angle. For concentric muscle activation, a tracing records data from left to right as the arm is abducted from 0” to 120”. For eccentric muscle activation, the tracing records data from right to lqft as the arm resists adduction from 120” to 0”. In contrast, a separate multi-channel ink recorder (Mingograft 82; Siemens Elema, Stockholm, Sweden), recorded IMP from left to right only. Thus if eccentric torque and EMG tracings recorded from right to left are to be integrated with IMP tracings recorded from left to right, then the torque and EMG tracings would need to be flipped 180” and read through the back side of the paper! Thus I ask Sporrong and Styf whether they remembered to do this, or whether they made an innocent mistake, but otherwise published and excellent and timely article? If the latter were true, I would be very interested in their further comments and discussion. Although an exact comparison would require that each individual IMP/torque data point be recalculated, note that if we flip the eccentric IMPItorque versus shoulder abduction curve 180” side-to-side, we see that it lies essentially parallel to the concentric curve (Fig. I , [2]). Thus if a simple x-axis error had been present and the correction were made, it would lead us to conclude that IMP, EMG signal, and torque demonstrate consistent relationships across muscles and contraction modes [ 1,2].