Documente Academic
Documente Profesional
Documente Cultură
ORIGINAL ARTICLE
Abstract
Objective: To assess the diagnostic accuracy of the surface electromyography (sEMG) parameters associated with referred anterior knee pain in
diagnosing patellofemoral pain syndrome (PFPS).
Design: Sensitivity and specificity analysis.
Setting: Physical rehabilitation center and laboratory of biomechanics and motor control.
Participants: Pain-free subjects (nZ29) and participants with PFPS (nZ22) selected by convenience.
Interventions: Not applicable.
Main Outcome Measure: The diagnostic accuracy was calculated for sEMG parameters reliability, precision, and ability to differentiate
participants with and without PFPS. The selected sEMG parameter associated with anterior knee pain was considered as an index test and was
compared with the reference standard for the diagnosis of PFPS. Intraclass correlation coefficient, SEM, independent t tests, sensitivity,
specificity, negative and positive likelihood ratios, and negative and positive predictive values were used for the statistical analysis.
Results: The medium-frequency band (B2) parameter was reliable (intraclass correlation coefficientZ.80e.90), precise (SEMZ2.71e3.87
normalized unit), and able to differentiate participants with and without PFPS (P<.05). The association of B2 with anterior knee pain showed
positive diagnostic accuracy values (specificity, .87; sensitivity, .70; negative likelihood ratio, .33; positive likelihood ratio, 5.63; negative predictive value, .72; and positive predictive value, .86).
Conclusions: The results provide evidence to support the use of EMG signals (B2 e frequency band of 45e96Hz) of the vastus lateralis and
vastus medialis muscles with referred anterior knee pain in the diagnosis of PFPS.
Archives of Physical Medicine and Rehabilitation 2014;95:1521-6
2014 by the American Congress of Rehabilitation Medicine
0003-9993/14/$36 - see front matter 2014 by the American Congress of Rehabilitation Medicine
http://dx.doi.org/10.1016/j.apmr.2014.03.028
1522
D. Ferrari et al
Methods
Participants
Index test
List of abbreviations:
ICC
LRL
LRD
NPV
PFPS
PPV
PSD
sEMG
VAS
VL
VM
Instrumentation
The experimental design included a staircase of 7 steps, each
18cm high and 28cm deep, and a walkway in front of and behind
the staircase.
The sEMG signals were obtained using a conditioner module
(ADS1000-AC1160a) that used a fourth-order, zero-lag, band-pass
Butterworth digital filter with cutoff frequencies of 20 and 500Hz
and an amplifier with a gain of 50. The sample frequency used was
4000Hz. The preamplifier circuit on the electrode cable had a gain
of 20, a common mode rejection ratio of >80dB, and an impedance of 1012U. To collect the VM and the VL sEMG data, 2 pairs
of bipolar surface-capture Ag/AgCl electrodes (Medi-Traceb) with
diameters of 10mm were positioned on the VM and VL muscles at
www.archives-pmr.org
1523
Table 1
Variable
Control Group
PFPS Group
Age (y)
Height (m)
Weight (kg)
N
20.422.33
1.640.05
57.948.51
29
22.652.08
1.640.06
61.7910.65
22
.001*
.814
.200
an interelectrode distance of 20mm. The positions of the electrodes were determined according to the location of the motor
point. The electrodes were positioned 2cm below the motor point
in the direction of the muscle belly.18 An electrostimulation device
(Nemesysc) was used to find the motor point. To reduce the
measurement variability that might be caused by the positions of
the electrodes, a template was developed that used the anatomic
references of the patella and the hip of the participant.
Processing data
The analyzed sEMG signals were referenced by the vertical
component of the ground reaction force measured by the force
plate in the fourth step; that is, the beginning and end values of all
the sEMG signals considered for analysis were determined in
accordance with the vertical component of the ground reaction
force measurement in the fourth step.
All the analyses were performed in MATLAB. The power
spectrum density (PSD) of the filtered sEMG time series was
calculated using the fast Fourier transform.19 The median frequency of the PSD was calculated as the frequency at which the
integral of the left side of the spectrum was equal to that of the
right side.20
The intensity of the PSD was normalized using the following
steps: (1) calculation of the spectral distribution function, which is
the cumulative sum of the power spectrum divided by its
maximum value and multiplied by 100, and (2) calculation of the
derived spectral distribution function to obtain a PSD with intensity values normalized between 0 and 100. From the normalized PSD, the mean intensity was calculated for each of the 3
frequency bands considered for analysis: low (15e45Hz), medium
(45e96Hz), and high (96e400Hz).21
Experimental protocol
For the reliability analysis, the trials were performed in the same
manner on 2 separate days, with an interval of 2 to 7 days between
Table 2
Statistical analyses
The descriptive values (means SDs) were obtained using SPSS
software version 18.d The data were analyzed for normality of
distribution using the Shapiro-Wilk W test. Independent t tests
were used to quantify the differences between the groups. For a
relative measure of reliability, the intraclass correlation coefficient
(ICC)(2, k) model was used.22 ICC values in the .00 to .25 range
indicated little, if any, reliability; the .26 to .49 range indicated
poor reliability; the .50 to .69 range indicated moderate reliability;
the .70 to .89 range indicated high reliability; and the .90 to 1.0
range indicated very high reliability.23,24 The SEM was used to
express the reliability in absolute values,22 indicating the precision
of the measurement.25 A lower SEM indicated better reliability of
the measurement.24
After calculating the reliability of the sEMG parameters, the
diagnostic accuracy was determined. The term accuracy refers to
the amount of agreement between the information from the
reference standard and that of the index test.17 To examine the
diagnostic accuracy value of the index test, the sensitivity, specificity, positive and negative predictive values (PPVs and NPVs),
and positive and negative likelihood ratios (LR and LR) were
calculated.26,27
Determination of the true positive, false positive, true negative,
and false negative by the index test followed the criteria listed in
appendix 1. To determine the cutoff of the sEMG signal for
diagnosis, several confidence intervals (70%, 80%, 90%, and
95%) of the control group were considered and the sensitivity
and specificity of each interval was calculated. The confidence
interval selected was the one that showed the best sensitivity
and specificity.
Results
From May 2011 through June 2012, 78 participants were recruited
using the inclusion criteria, and after diagnosis, 51 participants
Mean, SD, ICC, and SEM values of the sEMG frequency domain parameters in the Control Group and the PFPS Group for the VM muscle
PFPS Group
Control Group
Day 1
Day 2
Day 1
Day 2
Parameter
Mean SD
Mean SD
ICC
SEM
Mean SD
Mean SD
ICC
SEM
MF (Hz)
B1 (nu)
B2 (nu)
B3 (nu)
57.4412.42
47.84*12.13
32.90*8.42
1.761.26
55.898.14
49.8911.47
32.848.03
1.590.87
0.82
0.85
0.90
0.63
5.8
5.83
3.52
0.79
52.718.0
55.68*10.56
28.21*5.0
1.470.78
51.377.44
56.08.52
27.744.67
1.430.61
0.83
0.87
0.81
0.91
4.05
4.53
2.71
0.28
Abbreviations: B1, low-frequency band; B2, medium-frequency band; B3, high-frequency band; MF, median frequency; nu, normalized unit.
* A significant difference (P<.05) between groups only for day 1.
www.archives-pmr.org
1524
Table 3
D. Ferrari et al
Mean, SD, ICC, and SEM values of the sEMG frequency domain parameters in the Control Group and the PFPS Group for the VL muscle
PFPS Group
Day 1
Control Group
Day 2
Day 1
Day 2
Parameter
Mean SD
Mean SD
ICC
SEM
Mean SD
Mean SD
ICC
SEM
MF (Hz)
B1 (nu)
B2 (nu)
B3 (nu)
56.7712.95
48.1811.18
31.90*7.28
1.811.34
56.1210.51
49.2511.42
31.407.99
1.811.07
0.89
0.73
0.85
0.88
5.25
7.43
3.87
0.54
53.7711.87
54.9612.98
25.88*4.93
1.861.14
55.5712.43
52.6812.64
26.294.75
2.061.16
0.85
0.84
0.80
0.91
6.16
6.58
2.81
0.45
Abbreviations: B1, low-frequency band; B2, medium- frequency band; B3, high-frequency band; MF, median frequency; nu, normalized unit.
* A significant difference (P<.05) between groups only for day 1.
were selected and divided into the control group and PFPS group.
The sample loss was due to not following the diagnostic and
exclusion criteria used in this study. Table 1 presents the
descriptive data of the sample.
The data reliability of the electromyography parameters is
presented in tables 2 and 3. The results showed a moderate to very
high reliability associated with a low variability in the parameters
of domain frequency of the VM and VL muscles for both groups.
To compare the groups, the mean of day 1 was used, and the
differences were significant. The low-frequency band from the
VM muscle was able to differentiate the participants with and
without PFPS. The medium-frequency bands from both muscles
were reliable and able to differentiate between the control group
and the PFPS group (see tables 2 and 3).
The outcomes of the diagnostic accuracy of the index test refer
to the medium-frequency band because this parameter from both
muscles was able to differentiate the groups. The cutoff value
determined corresponds to a confidence interval of 95%. For the
VM muscle, the interval was between 26.72 and 31.26, and for the
VL muscle, the interval was between 24.75 and 29.01. The participants who reported values within these ranges were considered
pain-free, and the participants with values outside these ranges
were considered to have PFPS. The sensitivity, specificity, PPV
and NPV, and LR and LR results are presented in table 4.
Discussion
To our knowledge, this is the first study to explore the diagnostic
accuracy of the association of clinical and biomechanical parameters, represented by the electromyography variable, for
diagnosing PFPS. Our results revealed positive diagnostic accuracy values and suggest the use of the medium-frequency band
associated with anterior knee pain for the diagnosis of PFPS.
Reference
Standard
Positive
Negative
Total
Positive
Negative
Total
19 (true positive)
8 (false negative)
27
3 (false positive)
21 (true negative)
24
22
29
51
Study limitations
The use of the index test presented in this study showed positive
diagnostic accuracy values. This study was a preliminary investigation, and its applicability requires caution. This study has
some limitations that need to be considered for future studies, such
as a larger sample size and no restriction of gender. Although the
incorporation of part of the standard reference in the index test
might be considered bias, we are in search of the best method of
diagnosis and believe that the classical clinical questions cannot
be excluded from the diagnostic process.
Conclusions
The findings of this study revealed a positive outcome of the
diagnostic accuracy of the sEMG parameters associated with
referred anterior knee pain of a level of at least 2cm on a VAS and
pain in functional activity. The results of this study provide
diagnostic evidence for the use of this screening test in the diagnosis of PFPS.
Suppliers
a. Lynx, Rua Doutor Jose Elias, 358, Sao Paulo - SP, 05083-030,
Brazil.
b. Kendall, 15 Hampshire St, Mansfield, MA 02048.
c. Quark Medical, Rua do Rosario, 1519 - Centro, Piracicaba - SP,
13400-186, Brazil.
d. SPSS Inc, 233 S Wacker Dr, 11th Fl, Chicago, IL 60606.
Keywords
Anterior knee pain syndrome; Electromyography; Rehabilitation;
Sensitivity; Specificity
Corresponding author
Fabio Mcolis de Azevedo, PhD, Rua Roberto Simonsen, 305,
Presidente Prudente, SP 19060-900, Brazil. E-mail address:
micolis@fct.unesp.br.
www.archives-pmr.org
1525
False positive
True negative
False negative
Anterior knee pain level of at least 2cm on a 10cm VAS in the last month
Anterior knee pain during at least 2 of the
following activities: prolonged sitting, climbing
stairs, squatting, running, kneeling, and hopping
and jumping
VM and/or VL altered in electromyography analyses
Anterior knee pain level of at least 2cm on a 10cm VAS in the last month
Anterior knee pain during at least 2 of the
following activities: prolonged sitting, climbing
stairs, squatting, running, kneeling, and hopping
and jumping
VM and/or VL normal in electromyography analyses
Not have pain in knee
VM and/or VL normal in electromyography analyses
Not have pain in the knee
VM and/or VL altered in electromyography analyses
References
1. Fagan V, Delahunt E. Patellofemoral pain syndrome: a review on the
associated neuromuscular deficits and current treatment options. Br J
Sports Med 2008;42:789-95.
2. Ng GYF, Zhang AQ, Li CK. Biofeedback exercise improved the EMG
activity ratio of the medial and lateral vasti muscles in subjects with
patellofemoral pain syndrome. J Electromyogr Kinesiol 2008;18:128-33.
3. Lin F, Wilson NA, Makhsous M, et al. In vivo patellar tracking
induced by individual quadriceps components in individuals with
patellofemoral pain. J Biomech 2010;43:235-41.
4. Powers CM, Bolgla LA, Callaghan MJ, Collins N, Sheehan FT.
Patellofemoral pain: proximal, distal, and local factors, 2nd
International Research Retreat. J Orthop Sports Phys Ther 2012;42:
A1-54.
5. Pappas E, Wong-Tom WM. Prospective predictors of patellofemoral
pain syndrome: a systematic review with meta-analysis. Sports Health
2012;4:115-20.
6. Thomas MJ, Wood L, Selfe J, Peat G. Anterior knee pain in younger
adults as a precursor to subsequent patellofemoral osteoarthritis: a
systematic review. BMC Musculoskelet Disord 2010;11:201.
7. Nunes GS, Stapait EL, Kirsten MH, de Noronha M, Santos GM.
Clinical test for diagnosis of patellofemoral pain syndrome: systematic
review with meta-analysis. Phys Ther Sport 2013;14:54-9.
8. Cook C, Mabry L, Reiman MP, Hegedus EJ. Best tests/clinical findings for screening and diagnosis of patellofemoral pain syndrome: a
systematic review. Physiotherapy 2012;98:93-100.
9. Lankhorst NE, Bierma-Zeinstra SM, van Middelkoop M. Factors
associated with patellofemoral pain syndrome: a systematic review. Br
J Sports Med 2012;47:193-206.
10. Von Tscharner V, Valderrabano V. Classification of multi muscle
activation patterns of osteoarthritis patients during level walking. J
Electromyogr Kinesiol 2010;20:676-83.
11. Bolgla LA, Malone TR, Umberger BR, Uhl TL. Reliability of electromyographic methods used for assessing hip and knee neuromuscular activity in females diagnosed with patellofemoral pain
syndrome. J Electromyogr Kinesiol 2010;20:142-7.
12. Rathleff MS, Samani A, Olesen JL, et al. Neuromuscular activity and
knee kinematics in adolescents with patellofemoral pain. Med Sci
Sports Exerc 2013;45:1730-9.
1526
13. Farina D, Merletti R, Enoka RM. The extraction of neural strategies
from the surface EMG. J Appl Physiol (1985) 2004;96:1486-95.
14. Nijs J, Van Geel C, Van der auwera C, Van de Velde B. Diagnostic
value of five clinical tests in patellofemoral pain syndrome. Man Ther
2006;11:69-77.
15. Van Tiggelen D, Cowan S, Coorevits P, Duvigneaud N, Witvrouw E.
Delayed vastus medialis obliquus to vastus lateralis onset timing
contributes to the development of patellofemoral pain in previously
healthy men: a prospective study. Am J Sports Med 2009;37:
1099-105.
16. Uliam Kuriki H, Mcolis de Azevedo F, de Faria Negrao Filho R,
Alves N. Comparison of different analysis techniques for the determination of muscle onset in individuals with patellofemoral pain
syndrome. J Electromyogr Kinesiol 2011;21:982-7.
17. Bossuyt PM, Reitsma JB, Bruns DE, et al. Towards complete and
accurate reporting of studies of diagnostic accuracy: the STARD
initiative. Fam Pract 2004;21:4-10.
18. Hermens HJ. Development of recommendations for SEMG sensors
and sensor placement procedures. J Electromyogr Kinesiol 2000;10:
361-74.
19. Smoliga JM, Myers JB, Redfern MS, Lephart SM. Reliability and
precision of EMG in leg, torso, and arm muscles during running. J
Electromyogr Kinesiol 2010;20:e1-9.
20. Solomonow M, Baten C, Smit JOS, Baratta R, Hermens H,
Ambrosia RD. Electromyogram power spectra frequencies associated
with motor unit recruitment strategies. J Appl Physiol (1985) 1990;68:
1177-85.
21. Ferrari D, Kuriki HU, Alves N, De Azevedo FM. Characterization of
the patellofemoral pain syndrome by frequency parameters of EMG
signal. J Biomech 2012;45:S390.
22. Weir JP. The intraclass correlation coefficient and the SEM. J Strength
Cond Res 2005;19:231-40.
D. Ferrari et al
23. Mathur S, Eng JJ, MacIntyre DL. Reliability of surface EMG during
sustained contractions of the quadriceps. J Electromyogr Kinesiol
2005;15:102-10.
24. Kellis E, Katis A. Reliability of EMG power-spectrum and amplitude
of the semitendinosus and biceps femoris muscles during ramp isometric contractions. J Electromyogr Kinesiol 2008;18:351-8.
25. Denegar CR, Ball DW. Assessing reliability and precision of measurement: an introduction to intraclass correlation and standard error
of measurement. J Sport Rehabil 1993;2:35-42.
26. Fritz JM, Wainner RS. Examining diagnostic tests: an evidence-based
perspective. Phys Ther 2001;81:1546-64.
27. Jaeschke R, Guyatt GH, Sackett DL. Users guides to the medical
literature. III. How to use an article about a diagnostic test. B. What
are the results and will they help me in caring for my patients? The
Evidence-Based Medicine Working Group. J Am Med Assoc 1994;
271:703-7.
28. Deeks JJ. Systematic reviews of evaluations of diagnostic and
screening tests. BMJ 2001;323:157-62.
29. Cook C, Hegedus E, Hawkins R, Scovell F, Wyland D. Diagnostic
accuracy and association to disability of clinical test findings associated with patellofemoral pain syndrome. Physiother Can 2010;62:
17-24.
30. Sweitzer BA, Cook C, Steadman JR, Hawkins RJ, Wyland DJ. The
inter-rater reliability and diagnostic accuracy of patellar mobility tests
in patients with anterior knee pain. Phys Sportsmed 2010;38:90-6.
31. Cavazzuti L, Merlo A, Orlandi F, Campanini I. Delayed onset of
electromyographic activity of vastus medialis obliquus relative to
vastus lateralis in subjects with patellofemoral pain syndrome. Gait
Posture 2010;32:290-5.
32. Wong Y. Recording the vastii muscle onset timing as a diagnostic
parameter for patellofemoral pain syndrome: fact or fad? Phys Ther
Sport 2009;10:71-4.
www.archives-pmr.org