276THE JOURNAL OF BONE AND JOINT SURGERY
M. G. Bowditch, BSc, FRCS, Orthopaedic Registrar
West Suffolk Hospital, Hardwick Lane, Bury St Edmunds, Suffolk IP33
P. Sanderson, FRCS Ed(Orth), Senior Orthopaedic Registrar
J. P. Livesey, FRCS, Senior House Officer
Northern General Hospital, Herries Road, Sheffield S5 7AU, UK.
Correspondence should be sent to Mr M. G. Bowditch at 41 Saffron Road,
Histon, Cambridge CB4 4LJ, UK.
©1996 British Editorial Society of Bone and Joint Surgery
THE SIGNIFICANCE OF AN ABSENT ANKLE REFLEX
M. G. BOWDITCH,P. SANDERSON,J. P. LIVESEY
From the Northern General Hospital, Sheffield, England
We assessed the prevalence of abnormal ankle reflexes
in 1074 adult patients attending orthopaedic clinics and
related it to age. Those with possible pathological causes
of reflex loss were excluded.
The absence of one or both reflexes was significantly
related to increasing age; all patients under 30 years
had both reflexes. Few had absent reflexes between 30
and 40 years, but over 40 years, the proportion with
both reflexes absent increased rapidly from 5% (40 to
50 years) to 80% (90 to 100 years). Unilateral absence
did not show the same pattern of increase being 3% to
5% at 40 to 60 years and 7% to 10% at over 60
Our results suggest that a significant number of
‘normal’ adults have unilateral absence of an ankle
reflex, but this finding is rare enough to be a definite
clinical sign, irrespective of age.
J Bone Joint Surg [Br] 1996;78-B:276-9.
Received 8 June 1995; Accepted after revision 4 October 1995
A patient with low back pain or sciatica and an absent
ankle reflex is often seen by orthopaedic and spinal sur-
geons. An absent ankle reflex has been described as a
definite neurological sign, possibly indicating nerve-root
compression (Maurice-Williams 1987; Frymoyer 1991;
Donaghy 1993; Dickson and Butt 1995).
For a clinical test to have useful significance, it must
have a high specificity and sensitivity with positive or
negative predictive values. Useful predictive values depend
strongly on prevalence (Altman 1991a). The predictive
value and the significance of an absent ankle reflex in
respect to nerve-root compression therefore depend on the
prevalence of absence in ‘normal patients’ from the same
population. Medical causes of absent ankle reflexes such as
vitamin deficiency, endocrine or inflammatory neuropathies
and vascular disorders (Maurice-Williams 1987) must be
The prevalance of absent ankle reflexes in the normal
adult ‘orthopaedic’ population is uncertain. Several authors
have studied elderly populations by clinical (Critchley
1931; Howell 1949; Smith 1956; Bryndum and Marquard-
sen 1964; Milne and Williamson 1972) and electrical meth-
ods (Bhatia and Irvine 1972); most of them report an
increase in prevalence of absent ankle jerks after the age of
70 years. There is some controversy over the pathological
significance of this finding (Ellenberg 1960; Bryndum and
Marquardsen 1964) but it has been attributed to a peri-
pheral neuropathy caused by either arteriosclerotic obstruc-
tion of the vasa vasorum or subclinical vitamin B defi-
ciency (Bryndum and Marquardsen 1964). These studies
have examined small select numbers of elderly patients
over the age of 60 years and have not tried to exclude
known medical causes.
We have studied the prevalence of absent ankle reflexes
in a ‘normal’ orthopaedic population.
PATIENTS AND METHODS
Between May and November 1994 we made a prospective
study of the presence or absence of ankle reflexes in 1074
adult patients attending the orthopaedic outpatient depart-
ment in two hospitals. Patients under 16 years of age and
those with a past or present history of spinal disease, low
back pain or sciatica, diabetes mellitus and neuropathic or
systemic medical disease (including rheumatoid arthritis)
which are associated with absent reflexes, were excluded.
There were 541 men and 533 women with an age range of
16 to 99 years.
The patients were examined by one of the authors using
a standard reflex hammer, in three positions:
a) sitting on a padded chair or couch with lower legs
b) kneeling on a padded chair with the feet over the edge
(Monrad-Krohn and Refsum 1964); and
c) lying on the bed either supine, with legs bent at the knee
and rotated outwards at the hip and the ankle in gentle
dorsiflexion (Monrad-Krohn and Refsum 1964), or in the
lateral position with some knee flexion and the ankle in a
neutral position (Bhatia and Irvine 1972).
To minimise interobserver error the examiner alone
decided between presence or absence of the reflex. A reflex
was recorded as present if elicited in any of the above
positions and absent if not obtained even after reinforce-
ment or attempts to relax the patient. The age and sex of the
patient were also recorded.
Fifty patients were examined separately by each of the
authors and the results compared. The kappa value (Altman
1991b) was then calculated to assess the interobserver
The results from all patients were considered in nine
groups, each representing a decade.
Statistical analysis. The percentage proportion (pre-
valence) and its 95% CI limits for each result of the reflex
examination namely ‘both present’, ‘both absent’ or ‘one
absent’ at each age group were calculated. The relationship
between the prevalence of absent (one or both) ankle
reflexes and increasing age was assessed using the chi-
277THE SIGNIFICANCE OF AN ABSENT ANKLE REFLEX
VOL. 78-B, NO. 2, MARCH 1996
Table I. The numbers of male and female patients and the
outcome of their ankle reflex examination in each age group
Both presentBoth absent
16 to 201919
21 to 3075580000
31 to 4063 451201
41 to 5071611322
51 to 6076666752
61 to 7063 55312567
71 to 804357 32367 11
81 to 901921 163337
91 to 1001211201
Table II. The prevalence (% proportion of total) of each ankle reflex outcome and the 95% CI for
both male and female patients in each age group
16 to 2038
21 to 30133
31 to 40112
93 to 99.8
41 to 50 140
90.5 to 98.1
51 to 60162
82.6 to 92.7
61 to 70187
56.2 to 70
71 to 80186
46.6 to 60.9
81 to 9099
30.8 to 50.1
91 to 10017
-0.5 to 35.7
-0.3 to 5.60.8 to 2.6
0.1 to 5.60.1 to 5.6
3.8 to 121.2 to 5.6
23 to 363.3 to 10
29.7 to 43.5 5.4 to 14
39.6 to 59.34.2 to 16
66.2 to 86.8-5.3 to 17.1
The prevalence of each ankle reflex outcome as a percentage within each age group, showing both the degree
and pattern of the changes with increasing age.
squared test for trends in contingency tables (Altman
1991c) and linear regression analysis.
To determine the age at which the most significant
changes in prevalence occurred, the results in every two
consecutive age groups (e.g. 51 to 60 and 61 to 70) were
compared. The difference in prevalence (%), the 95% CI
for this difference and the p value between the two consec-
utive groups, for both bilateral and unilateral absence, were
calculated by comparing proportions in contingency tables
and using Fisher’s exact test.
The results for gender in each age group were compared
separately using chi-squared tests.
We used the computer statistical software package
GraphPAD Software v 1.13 (Molfino, Toronto, Canada) and
a significance level of 5% on two-tailed tests throughout.
The interobserver error was very small with a kappa = 0.94
representing only one disagreement in the 50 patients
examined by two observers. Table I shows the number of
male and female patients and the outcome of the ankle
reflex examination in each age group.
Table II shows the percentage proportion (prevalence) of
each outcome and the 95% CI for all patients in each age
group. In all patients the prevalence of absent ankle reflexes
in one or both sides rises significantly with increasing age
(chi-squared trend = 41.7, correlation coefficient r = 0.88,
p < 0.001 and chi-squared trend = 76.1, r = 0.9, p < 0.0001,
The pattern and degree of this increase are different for
bilateral and unilateral absence, however, being much
greater for bilateral absence (Fig. 1). Table III shows the
difference in prevalence for both bilateral and unilateral
absence between consecutive age groups, with 95% CIs
and p values to show when the most significant increases or
changes occur. The prevalence of bilateral absence shows a
steady rise with two significant jumps, between the fifth
and sixth decades (23.8%, 95% CI 14.9% to 32.6%,
p < 0.00001) and between the seventh and eighth decades
(12.9%, 95% CI 0.9% to 25%, p < 0.05). Unilateral
absence increases steadily, but much less, with no sig-
nificant jumps between consecutive decades. We found no
significant difference between male and female patients in
any age group (Table IV).
It is generally accepted that there is loss of the ankle
reflexes with increasing age, but only the results for medi-
cal patients over the age of 60 years have been reported.
We studied loss of the ankle reflex over the full adult age
range in a population which represents those likely to be
seen by orthopaedic and spinal surgeons.
Over the age of 40 years, in either sex, the proportion of
278M. G. BOWDITCH, P. SANDERSON,J. P. LIVESEY
THE JOURNAL OF BONE AND JOINT SURGERY
Table III. The percentage difference in prevalence and its 95% CI between each two
consecutive age groups for bilateral and unilateral reflex absence
age groups (yr)% 95% CI
(16 to 20):(21 to 30)
(21 to 30):(31 to 40)
(31 to 40):(41 to 50)
(41 to 50):(51 to 60)
(51 to 60):(61 to 70)
14.9 to 32.6
(61 to 70):(71 to 80)
-2.9 to 16.2
(71 to 80):(81 to 90)
0.9 to 25
(81 to 90):(91 to 100)
<0.00001-2.3 to 7.50.36
0.19 02.9 to 8.30.36
0.04-6.8 to 7.70.99
Table IV. The level of sig-
nificance for comparison of the
reflex outcomes between male
and female patients in each age
Age groups (yr)
16 to 20
21 to 30-
31 to 40 0.3
41 to 50 0.5
51 to 600.55
61 to 70 0.83
71 to 80 0.78
81 to 90 0.3
91 to 100 0.4
Table V. The prevalence of ankle reflex outcome
and age (AROPA) index: a ‘working’ guide
Ankle reflex outcome prevalence
patients with absent ankle reflexes increases. The difference
between consecutive decades is small until that between the
fifth and sixth; at this level there is a significant increase,
principally due to a large rise in the proportion of patients
with bilateral absent reflexes (8% to 30%). This was not
seen for unilateral absence of reflexes. This increase at
about 60 years is similar to that reported in previous
studies, at about 70 years, and reduces the clinical value of
The prevalence of unilateral loss of the reflex has been
addressed by Bryndum and Marquardsen (1964) who report-
ed four cases in 100 patients over 65 years of age. In our
patients 1% to 10% of adults older than 40 years had
unilateral absence of an ankle reflex. Unilateral loss is there-
fore a more useful neurological sign and, where appropriate,
will require further investigation, irrespective of age.
The differences between bilateral absence and unilateral
absence suggest different aetiologies. Bilateral loss may
have a central or systemic cause, as suggested by Bryndum
and Marquardsen (1964), and unilateral loss may be due to
a peripheral or mechanical abnormality.
The predictive value of the absent ankle reflex for her-
niated lumbar disc is reported to be about 90% between 20
and 45 years of age and 60% over the age of 50 years
(Spangfort 1972); our findings support this view.
We present a simple ‘working’ guide to the prevalence of
ankle reflex outcomes with respect to the patient’s age
The authors wish to thank the Department of Statistics at the University of
Sheffield for their help and advice with this paper.
No benefits in any form have been received or will be received from a
commercial party related directly or indirectly to the subject of this
Altman DG. Practical statistics for medical research. London: Chapman
& Hall, 1991: a) Chapt. 14:411; b) Chapt. 14:404-7; c) Chapt.
Bhatia S, Irvine RE. Electrical recording of the ankle jerk in old age.
Gerontol Clin Basel 1972;15:357-60.
Bryndum B, Marquardsen J. The tendon reflexes in old age. Geront Clin
Critchley M. The neurology of old age. Lancet 1931;I:1221-30.
Dickson RA, Butt WP. Mini symposium: lumbar disc disease: 1). Clinical
and radiological assessment. Curr Orthop 1995;9:73-84.
Donaghy M. Disorders of peripheral nerves: lumbar disc lesions and
sciatica. In: Walton J, ed. Brain’s diseases of the nervous system. 10th
ed. Oxford: Oxford University Press, 1993:569-72.
Ellenberg M. The deep reflexes in old age. JAMA 1960;174:468-9.
Frymoyer JM. The adult spine: principles and practice. Vol. 1. New
York: Raven Press, 1991.
Howell TH. Senile deterioration of the central nervous system. Br Med J
Maurice-Williams RS. Disorders of the spinal nerve roots. In: Weatheral
DJ, et al, ed. Oxford textbook of medicine. 2nd ed. Oxford: Oxford
Medical Publications, 1987;21:109-15.
Milne JS, Williamson J. The ankle jerk in older people. Gerontol Clin
Monrad-Krohn GH, Refsum S. The clinical examination of the nervous
system. 12th ed. London: H. K. Lewis, 1964:189-90.
Smith RCF. The elicitation and incidence of neurological signs in the
aged. Practitioner 1956;177:59-66.
Spangfort EV. The lumbar disc herniation: a computer-aided analysis of
2504 operations. Acta Orthop Scand 1972:Suppl 142;1-95.
279THE SIGNIFICANCE OF AN ABSENT ANKLE REFLEX
VOL. 78-B, NO. 2, MARCH 1996