Journal of Manipulative and Physiological Therapeutics
Volume 22 • Number 9 • November/December 1999
0161-4754/99/$8.00 + 0

76/1/102747 © 1999 JMPT

628

INTRODUCTION

Donelson et al

1

compared McKenzie

assessment with discography and diagnostic
imaging (magnetic resonance imaging or
computed tomography) in the detection of
discogenic lower back pain and annular
incompetence. They concluded that in a
chronic, out-of-work, worker’s compensa-
tion-litigation patient population, McKenzie
assessment reliably differentiated discogenic and
nondiscogenic pain. They further concluded that this
noninvasive, low-tech, relatively inexpensive clinical assess-
ment with repeated end-range lumbar test movements pro-
vided more relevant information than noninvasive imaging
studies.

This commentary briefly reviews the Donelson study and

provides statistics to appraise the validity of McKenzie
assessment as a diagnostic test.

REVIEW

Donelson et al,

1

on the basis of a dynamic internal disc

model, hypothesized that change in location of lumbar and
referred pain during McKenzie testing would reliably distin-
guish discogenic from nondiscogenic lower back pain and
competent from incompetent annuli in symptomatic discs.

In McKenzie assessment repeated lumbar end-range

motions are used to provoke change in pain patterns. Change
in location of back pain and leg pain is used to assist in clas-
sifying patients. Pain that moves proximally toward the
spine is termed centralization. Pain that moves distally away
from the spine is termed peripheralization, and pain that
does not change location in response to movement is termed
no change.

Donelson et al classified patients with chronic lower back

pain who were referred for discography into 3 groups on the
basis of centralization, peripheralization, or no change in
pain during McKenzie assessment. These groups were then
compared with blinded discography (for discogenic pain)
and diagnostic imaging (for annular incompetency).

The patient population consisted of 63 patients (41 men

and 22 women; average age, 39.6 years) with lower back pain
symptoms present for more than 3 months (median duration,
15.3 ± 12.2 months). The majority of patients had pain below
the knee, and none had neurologic deficits. The majority of
patients (69.9%) were not working because of their pain. In-
surance compensation was divided among worker’s compen-
sation (42.8%), medical-legal compensation (30.1%), and

private self-pay (27.1%). No other characteris-

tics of the study population were noted.

Data were evaluated for significant differ-

ences (P < .05) with descriptive statistics by

using

χ

2

analysis for categorical variables

and t or z tests for continuous variables. Dur-

ing McKenzie assessment, referred pain cen-

tralized in 31 (49.2%) patients. Twenty-three

(74%) of these had positive discography results,

of which 21 (91%) had intact annuli. Sixteen (25%)

patients experienced pain peripheralization. Eleven

(69%) of these subjects had positive discograms, and 6 (54%)
of this subgroup had intact annuli. Distal pain was unchanged
in 16 (25%) patients, two (12.5%) of whom had positive
discograms, and all of whom had intact annuli (100%).

There were no significant differences between the 3 pain

response groups in regard to sex, age, duration of symptoms,
or insurance type. The ability of McKenzie assessment to dis-
tinguish between a positive and negative discography result
on the basis of these pain responses alone was highly signif-
icant (P < .001). The incidence of discs with a competent an-
nulus that occurred in centralizers was significantly greater
than that found in peripheralizers (P < .042). Donelson et al
concluded that McKenzie assessment reliably differentiated
discogenic and nondiscogenic pain and was helpful in distin-
guishing a competent from an incompetent annulus.

STATISTICAL COMMENTARY

Standard statistics establishing the validity of McKenzie

assessment as a diagnostic test were lacking in the original
Donelson study.

1

Sensitivity, specificity, and likelihood ratio

were calculated herein (Table 1) by using Bayesian probabil-
ity analysis of the data summary table from the Donelson
study (p 1118).

Sensitivity is the true-positive rate, a measure of how well

the test detects positive cases. Specificity, the true-negative
rate, measures how well it detects negative cases. Positive
and negative predictive values represent the probability of a
disorder being present or absent if a test result is positive or
negative, respectively. The likelihood ratio summarizes how
the test discriminates between true-positive and false-posi-
tive results.

2

If any change in pain pattern occurs, whether centraliza-

tion or peripheralization, McKenzie assessment has a high
sensitivity: 94% for detecting discogenic pain and 100% for
detecting an incompetent annulus. If peripheralization of
pain occurs, McKenzie testing has a moderate specificity:

COMMENTARY

The Diagnostic Utility of McKenzie Clinical Assessment for Lower Back Pain

Paul M. Delaney, PhD, DC,

a

and Michael J. Hubka, DC

a

Journal of Manipulative and Physiological Therapeutics

Volume 22 • Number 9 • November/December 1999

Commentary

• Delaney and Hubka

629

82% for detecting discogenic pain and 86% for detecting an
incompetent annulus. The likelihood ratio for McKenzie
testing was not impressive, with the highest (3.58) being for
peripheralization of pain (Table 1).

These statistics differ from those of Bogduk and Lord,

3

whose commentary on the Donelson study misinterpreted
the data. This altered their statistical analysis and subse-
quent inferences about McKenzie testing for discogenic
pain. On page 246, in their “Results” section, they state that
“[d]istal pain was unchanged in 16 subjects (25%), only 2 of
whom (12.5%) had intact anuli.” In fact, all 16 patients had a
competent annulus, but 2 of the 16 had positive discography
results (Donelson R, personal communication).

The most informative McKenzie result regarding disco-

genic pain is no change in pain during assessment. Peripher-
alization is the most informative test result regarding annular
incompetency. These pain-pattern classifications minimized
false-positive and false-negative results compared with other
pain-pattern groups for those diagnoses (Table 1).

A useful perspective is provided by comparing McKenzie

assessment to physical examination procedures published in
the Agency for Health Care Policy and Research guidelines
for acute lower back problems in adults (p 18, Table 2).

4

These guidelines cited sensitivity and specificity for 9 phys-
ical examination procedures used to diagnose lumbar disc
herniation in patients with sciatica on the basis of surgical
case series from multiple literature sources. The highest sen-
sitivity calculated for McKenzie testing (Table 1) exceeded
that of the 9 diagnostic procedures noted above, but its high-
est specificity was exceeded by 3 tests (Table 2).

McKenzie assessment meets many of the evidence-based

criteria for diagnostic test validity and clinical utility
described by Sackett et al.

2

However, on the basis of the sta-

tistics discussed above, McKenzie assessment does not reli-
ably distinguish discogenic from nondiscogenic pain.
Although the Donelson study was large enough to demon-
strate statistical significance, more research with larger sam-
ple sizes taken from general clinical populations is needed
to verify the validity of McKenzie assessment. The issue of
its intertester reliability must also be clarified.

1,5-8

CONCLUSION

Changes in pain patterns during McKenzie assessment

were informative but not definitive (ie, high sensitivity but
only low-to-moderate specificity) for detecting discogenic
pain. The validity and intertester reliability of McKenzie
assessment need further studies, with larger sample sizes
from general clinical populations, to determine whether it
can reliably detect discogenic pain and annular incompe-
tence.

Paul M. Delaney, PhD, DC

Associate Clinical Professor

LACC’s Glendale Chiropractic Clinic

1425 E Colorado St

Glendale, CA 91205

Michael J. Hubka, DC

Associate Clinical Professor

LACC’s Thie Chiropractic Clinic

1192 N Lake Ave

Pasadena, CA 91104

Table 1.

Contingency tables for the validity of McKenzie assessment in the diagnosis of discogenic pain and annular incompetence

Positive

Negative

Yes

No

Sensitivity

Specificity

LR

predictive value

predictive value

Discogenic pain

Centralizers

23

8

0.64

0.70

2.16

0.74

0.59

All others

13

19

Peripheralizers

11

5

0.31

0.82

1.65

0.69

0.47

All others

25

22

Change in pain

34

13

0.94

0.52

1.96

0.72

0.88

No change in pain

2

14

Incompetent annulus

Centralizers

10

21

0.50

0.51

1.02

0.32

0.69

All others

10

22

Peripheralizers

10

6

0.5

0.86

3.58

0.63

0.79

All others

10

37

Change in pain

20

27

1

0.37

1.59

0.43

1.00

No change in pain

0

16

Based on data from Donelson et al.

1

LR, Likelihood ratio.

Table 2.

Comparison of McKenzie assessment to diagnostic tests

for lower back pain in the Agency for Health Care Policy and
Research guidelines

Diagnostic test

Sensitivity Specificity

1. Ipsilateral SLR

0.80

0.40

2. Crossed SLR

0.25

0.90

3. Ankle dorsiflexion weakness

0.35

0.70

4. Great toe extensor weakness

0.50

0.70

5. Impaired ankle reflex

0.50

0.60

6. Sensory loss

0.50

0.50

7. Patellar reflex

0.50

NA

8. Ankle plantarflexion weakness

0.60

0.95

9. Quadriceps weakness

<0.01

0.99

10. McKenzie assessment for discogenic pain

0.94

0.82

11. McKenzie assessment for incompetent

1.00

0.86

annulus

Data for tests 1 to 9 were derived from Bigos et al (Table 2, p 18).

4

Data

for tests 10 and 11 were derived from Donelson et al.

1

In tests 10 and 11,

sensitivity is for change in pain during McKenzie assessment, and speci-
ficity is for peripheralization of pain during McKenzie assessment.

SLR, Straight leg raise; NA, not applicable.

ACKNOWLEDGMENTS

We thank Ms Nehmat Saab and the staff of the Learning

Resource Center at the Los Angeles College of Chiropractic
for their help accessing the literature.

REFERENCES

1. Donelson R,Aprill C, Medcalf R, Grant W. A prospective study of

centralization of lumbar and referred pain. A predictor of sympto-
matic discs and anular competence. Spine 1997;22:1115-22.

2. Sackett D, Richardson S, Rosenberg W, Haynes B. Evidence-

based medicine: how to practice and teach EBM. Oxford:
Churchill-Livingstone Publishers; 1997. p. 118.

3. Bogduk N, Lord SM. A prospective study of centralization of lum-

bar and referred pain: a predictor of symptomatic discs and anular
competence. Commentary. Pain Med J Club J 1997;3:246-8.

4. Bigos S, Bowyer O, Braen G, Brown K, Deyo R, Haldeman S,

et al. Acute low back problems in adults. Clinical practice

guideline, quick reference guide number 14. Rockville (MD):
US Department of Health and Human Services, Public Health
Service, Agency for Health Care Policy and Research; 1994
Dec. AHCPR Publication No. 95-0643.

5. Donahue MS, Riddle JL, Sullivan MS. Intertester reliability of

a modified version of McKenzie’s lateral shift assessments
obtained on patients with low back pain. Phys Ther 1996;
76:706-26.

6. Kilby J, Stigant M, Roberts A. The reliability of back pain as-

sessment by physiotherapists, using a “McKenzie algorithm.”
Physiotherapy 1990;76:9.

7. Riddle DL, Rothstein JM. Intertester reliability of McKenzie’s

classifications of the syndrome types present in patients with
low back pain. Spine 1993;18:1333-44.

8. Roach KE, Brown MD, Albin RD, Delaney KG, Lipprandt HM,

Rangelli D. The sensitivity and specificity of pain response to
activity and position in categorizing patients with low back pain.
Phys Ther 1997;77:730-8.

Journal of Manipulative and Physiological Therapeutics
Volume 22 • Number 9 • November/December 1999

Commentary

• Delaney and Hubka

630