SPINE Volume 31, Number 18, pp E621E627

2006, Lippincott Williams & Wilkins, Inc.

The Treatment of Chronic Coccydynia With

Intrarectal Manipulation
A Randomized Controlled Study
Jean-Yves Maigne, MD,* Gilles Chatellier, MD, Michel Le Faou, MD,*
and Marie Archambeau, MD*

Study Design. Randomized open study.

Objective. To evaluate the efficacy of intrarectal manual treatment of chronic coccydynia; and to determine the
factors predictive of a good outcome.
Summary of Background Data. In 2 open uncontrolled
studies, the success rate of intrarectal manipulation of the
coccyx was around 25%.
Methods. Patients were randomized into 2 groups of
51 patients each: 1 group received three sessions of coccygeal manipulation, and the other low-power external
physiotherapy. The manual treatment was guided by the
findings on stress radiographs. Patients were assessed, at
1 month and 6 months, using a VAS and (modified)
McGill Pain, Paris (functional coccydynia impact), and
(modified) Dallas Pain questionnaires.
Results. At baseline, the 2 groups were similar regarding
all parameters. At 1 month, all the median VAS and questionnaire values were modified by 34.7%, 36.0%, 20.0%,
and 33.8%, respectively, in the manipulation group, versus 19.1%, 7.7%, 20.0%, and 15.7%, respectively, in the
control (physiotherapy) group (P 0.09 [borderline], 0.03,
0.02, and 0.02, respectively). Good results were twice as
frequent in the manipulation group compared with the control group, at 1 month (36% vs. 20%, P 0.075) and at 6
months (22% vs. 12%, P 0.18). The main predictors of a
good outcome were stable coccyx, shorter duration, traumatic etiology, and lower score in the affective parts of the
McGill and Dallas questionnaires.
Conclusions. We found a mild effectiveness of intrarectal manipulation in chronic coccydynia.
Key words: manipulation, spinal manipulative therapy,
massage, coccyx, coccydynia, randomized controlled trial.
Spine 2006;31:E621E627

Intrarectal manipulation was the first published modality for the treatment of coccydynia. The technique was
mentioned in 1634, by Ambroise Pare (cited by Sugar1).
The object was to reduce a presumed fracture. In the
20th century, coccygeal pain was successively attributed
to a painful spasm of the intrapelvic muscles,2 loss of
From the *Department of Physical Medicine, Hopital Hotel-Dieu,
Paris, France; Unite dEpidemiologie Clinique, Paris, France; Faculte
de Medecine Rene Descartes Paris 5, Paris, France; and Hopital Europeen Georges Pompidou, Paris, France.
Acknowledgment date: February 14, 2006. Acceptance date: March
30, 2006.
The manuscript submitted does not contain information about medical
device(s)/drug(s).
No funds were received in support of this work. No benefits in any
form have been or will be received from a commercial party related
directly or indirectly to the subject of this manuscript.
Address correspondence and reprint requests to Jean-Yves Maigne,
MD, Department of Physical Medicine, Hotel-Dieu University Hospital, Place du Parvis de Notre-Dame, F-75181 Paris Cedex 04, France;
E-mail: jy.maigne@htd.aphp.fr

coccygeal mobility,3 or a sprain.4 Three main intrarectal

techniques were devised to address these pathogenic
mechanisms: Thiele described levator ani, coccygeus,
and piriformis massage2; Mennell proposed repeated
joint mobilization with circumduction of the coccyx,3
whereas Maigne practiced mobilization of the coccyx in
extension, with energetic levator ani stretching.5,6
More recently, we described a technique using dynamic views (stress radiographs).7 This technique clearly
reveals the mobility of the coccyx as the subject passes
from the standing to the sitting position. Mobility in
coccydynia patients was compared with that in healthy
controls. Normal movement was found to be usually
into flexion and amounted to between 0 and 25. A
pathologic exaggeration of this movement is termed hypermobility (found in 25% of the coccydynia patients).
Less often, normal movement was into extension, of between 0 and 20. A pathologic exaggeration of this
movement results in posterior luxation (20% of the coccydynia patients), which occurs only in the sitting position. These two lesions are bracketed together under the
term instability of the coccyx.7 A third type of lesion is
a coccygeal spicule (found in 15% of the coccydynia
patients). This abnormal bony spur at the distal end of
the coccyx may irritate the subcutaneous tissues if the
coccyx is stiff.8 The causative role of these lesions in
coccydynia was confirmed, in earlier studies, using intradiscal anesthetic blocks or injection of anesthetic around
the spicule.7,8
The rationale for the treatment of coccydynia with
manipulation is twofold: The lesions described above
may cause or maintain a painful tension of the intrapelvic muscles, which may aggravate the pain.9 Manual
techniques are thought to relax the muscles.9,10 Manipulation may also make a stiff coccyx slightly more mobile, which would reduce the potential for subcutaneous
tissue irritation by the coccyx. These techniques were
previously assessed in two open studies. Wray et al (in a
nonrandomized study) showed that patients who, before
manipulation, had received a local corticosteroid injection into the soft tissues around the sides and tip of the
coccyx, had a 25% better 3-month success rate than did
patients managed with local injections only (85% vs.
60%).11 In that study, the success criterion was freedom
from symptoms at the 3-month review. Our randomized
study compared the Thiele and the Maigne techniques,
using a third, personal technique (mild stretching of the
levator muscles without mobilizing the coccyx itself).9
E621

E622 Spine Volume 31 Number 18 2006

The overall success rate was 25.7% at 6 months and

24.3% at 2 years. Success was defined as 60% pain
relief on a visual analogue scale (VAS). Outcomes varied
slightly as a function of the cause of the coccydynia (assessed on stress radiographs) and the technique used.
Levator massage and mild stretching of the levators had
the greatest effectiveness, while the Maigne technique
worked best in patients with normal coccygeal mobility.
The two studies did not use controls; and although the
success rates were similar (25%), a simple placebo effect
could not be ruled out. However, a controlled study had
to be undertaken. We present the results of an open randomized trial comparing the effectiveness of intrarectal
manipulation with that of a control treatment, in patients with chronic coccygodynia.
Methods
Participants. Between October 1999 and April 2003, 286 new
patients with coccydynia were examined by the senior author
(J.-Y.M.). Coccydynia was defined as pain in or around the
coccyx, without any significant radiation, which is present
mainly in the sitting position or when passing from the sitting
to the standing position.7

Sample Size Estimation. It was calculated that, in order to

achieve 80% confidence in detecting a difference between a placebo (expected recovery rate, 10%) and an active treatment (expected recovery rate, 35%), the sample size required was 102
patients.
Inclusion Criteria. The following inclusion criteria were
used: 1) chronic (2 months) coccydynia (since, in our experience, many cases of acute coccydynia will remit spontaneously
in under 2 months); 2) age over 25 and under 70 years (the
lower age limit reflects our policy of not performing intrarectal
manipulation in subjects under 25 years of age); 3) no ongoing
antidepressant treatment; 4) no medicolegal litigation or occupational-accident etiology; 5) good knowledge of the questionnaire language (French); 6) no prior manual treatment; 7) consent to the study (and, in particular, to repeated intrarectal
treatments); and 8) ability to return for further visits as required by the treatment and the study protocol.
Baseline Assessment. The parameters recorded were patient
age, gender, body mass index (BMI), duration of the complaints, and traumatic or nontraumatic etiology (coccydynia
was defined as traumatic if there had been less than 1 month
between a traumatic event or childbirth and the onset of the
complaint).8 In patients randomized to manipulation, the pelvic muscle tone was assessed rectally, before any manipulation,
using a previously described technique.9 Muscle tone was
graded subjectively as normal or diminished.
Stress radiographs were obtained. In light of these stress radiographs, patients were first classified as having a stable coccyx
(with normal mobility of the coccyx in flexion or extension), or an
unstable coccyx (hypermobility or luxation). Hypermobility has
been defined as a pathologic exaggeration of the movement of
flexion and luxation as a pathologic alteration of the normal pattern of movement of extension.8 This classification was the most
important parameter for the current study.

Next, the following three angles, previously defined in the

literature,8,12 were also recorded: angle of coccygeal incidence,
angle of sagittal pelvic rotation, and sacrococcygeal angle.
These angles are related with each other and with the direction
of the coccygeal motion when sitting.
Pain intensity was assessed using a VAS and questionnaires.
The 100-mm VAS was used to establish the mean pain intensity
in the sitting position (If you are sitting normally, without
trying to avoid putting weight on the coccyx, how bad, on
average, is your pain?), or the mean pain intensity when passing from the sitting to the standing position, if the sitting position as such was not painful. The maximum score was 100
points (1 point for every millimeter on the scale).
The impact of the pain was assessed using three questionnaires: the (modified) McGill Pain Questionnaire, the Paris
(functional coccydynia impact) Questionnaire, and the (modified) Dallas Pain Questionnaire. The modified (abridged)
McGill Pain Questionnaire comprising 15 items (1 8 sensory; 9 15 affective) has been validated in a French-language
version.13 The maximum score was 60 points. The Frenchlanguage version of the Dallas Pain Questionnaire has also
been validated.14 This instrument is a low back pain questionnaire; for purposes of this study, only the nine sections relevant
to coccygeal pain (Sections VIIIXVI; each with a maximum
score of 10 points) were used. The section concerning pain
during car travels was counted twice, to produce a maximum
total score of 100, since experience has shown pain during car
travel to be the most sensitive and representative feature of
coccygeal pain. The Paris Questionnaire was designed in our
department specifically for this study and has therefore not
been submitted to formal validation. It covers functional aspects only and has a maximum number of 100 points (see
Appendix). Thus, the total number of points that could be
obtained (called the individual global score) was 360, with
higher scores reflecting a greater impact of the coccydynia.

Treatments. The active treatment comprised three 5-minute

sessions of intrarectal manipulation, over a period of 10 days
(manipulation group). It was delivered by the senior author
(J.-Y.M., a physician who has been practicing manual medicine
for 25 years). Based on the study by Maigne et al,9 the maneuver used was a composite rather than a single one. As a general
rule, coccyges that were nonmobile or which displaced in flexion were treated with strong levator muscle stretching and mobilization of the coccyx in extension; while only slight levator
stretching, without coccygeal mobilization, was used in coccyges that displaced in extension (with or without posterior luxation). In all cases, levator muscles that were tender to internal
palpation would be massaged longer. The intensity of the maneuvers was gradually increased during the three sessions. At
the start of each session, only slight stretching, preceded by
isometric external-sphincter contraction (muscle energy technique), would be used, to avoid all pain.
The active treatment was compared against external shortwave (magnetic field) physiotherapy applied to the sacrum at
the lowest possible power; this treatment was delivered in three
sessions over a period of 10 days (control group). This modality
was tantamount to a placebo.
Patients in both groups were informed that the treatment
was designed to reduce the internal muscle spasm. If improvement was not obtained until after the second or the third session, a fourth session could be arranged. At the end of the

Treatment of Chronic Coccydynia Maigne et al E623

Table 1. Baseline Data at Enrollment

Gender (female/male)
Age (yr)
Duration of complaints (mo)*
Etiologic trauma (%)
BMI (kg/m2)
Stable coccyx (%)
Unstable coccyx (%)
VAS 100 mm (maximum 100 points)*
MPQ (maximum 60 points)*
Paris (maximum 100 points)*
DPQ (modified) (maximum 100
points)*
Individual global score (maximum
360 points)

Manipulation
(n 51)

Controls
(n 51)

46/5
45.2 (11.5)
15 (2120)
18 (35.3)
24.4 (4.2)
26 (51)
25 (49)
62 (2595)
18 (452)
50 (2080)
37 (1161)

46/5
44.6 (13)
11 (260)
22 (43.1)
24.5 (4.7)
22 (43)
29 (57)
68 (2295)
19.5 (552)
50 (3090)
41.5 (1085)

168 48

186 41.2

BMI body mass index; VAS visual analog scale; MPQ (modified) McGill
Pain Questionnaire; Paris Paris Questionnaire; DPQ (modified) Dallas Pain
Questionnaire.
Data are mean (SD) or % (n).
*Data are median (range).

treatment, the patients were given the questionnaires and the

VAS for the subsequent assessments.

One-Month and Six-Month Follow-up. One month after

the end of the treatment, the patients filled in and returned the
first set of their questionnaires and VAS. Patients were routinely contacted by telephone by one of the authors (M.A.),
who would see them in the outpatient clinic if need be. If the
patient felt that the outcome was unsatisfactory, he or she
would be removed from the study, and the treatment would be
regarded as a failure. If the patient was satisfied with the outcome, he or she would wait, without any other treatment, for 5
months. At the end of this period (i.e., 6 months from the end of
treatment), he or she would fill in the second set of questionnaires and VAS, and was then telephoned by the assessor.
Outcome Measures. The primary outcome measure was the
score obtained in the questionnaires and the VAS. The secondary outcome measure was the good or poor nature of the
final outcome, which was determined from the individual
global score (maximum score, 360 points) at 0, 1, and 6
months. It was arbitrarily decided that patients who had met
the two conditions of1 50% decrease in the individual global
score at 1 month, and2 60% decrease in that score at 6
months, would be considered to have had a good outcome.
Anything else would be considered a poor outcome, in par-

ticular if the improvement was not observed until after the

1-month milestone. Thus, four subgroups (good and poor outcomes, respectively, following manipulation; and good and
poor outcomes, respectively, following the control treatment)
were established and compared. This subgroup analysis is presented without any statistical test, to avoid erroneous conclusions resulting from inflation of the alpha risk.

Statistical Methods and Analysis. The investigation was

designed as an open randomized study. Randomization was by
sealed envelopes opened in the presence of the patient. The
randomization list was balanced at every 10th patient. The
study had been approved by the local Ethics Committee, and
each patients written informed consent was obtained.
The statistical analysis of the outcome measures was performed using the customary tests: 2 test (for categorical variables) and t test or Mann-Whitney test, as appropriate. Differences with a P value of less than 0.05 were considered
statistically significant. Statview (version 5.0, SAS Institute
Inc.) and Stata (StataCorp LP 7, 2005) software was used.
Results are given as numbers (percent), mean values (1 SD), or
median (range) values for variables that did not have a gaussian
distribution.

Results
Patient Baseline Characteristics
A total of 102 patients were enrolled. Two were lost to
follow-up (1 after the first manipulation session; the other,
a Control group patient, after the 1-month follow-up visit,
which had shown 50% improvement). Thus, in each
group, 50 patients were available for review. At enrollment,
the 2 groups were well matched for the parameters studied
(Table 1).
One-Month Results
At 1 month, all the VAS and questionnaire results differed statistically between the 2 groups, with the manipulation group showing the better outcomes (Table 2).
The mean individual global score had improved by 26%
in the manipulation group, as against 14% in the control
group. Equal to or greater than 50% improvement in the
individual global score at 1 month was seen in 36% of
the manipulation group patients and in 20% of the control group patients.

Table 2. VAS and Questionnaire Results at 1 Month

Instrument (P value)
VAS (P 0.09)
Group
Manipulation (n 50)
Decrease M0M1 (%)
Control (n 50)
Decrease M0M1 (%)

M0

M1

62 (2595)
40.5 (0100)
34.7%
68 (2295)
55 (0100)
19.1%

MPQ (P 0.03)
M0

M1

18 (452)
11.5 (041)
36%
19.5 (552)
18 (154)
7.7%

Paris (P 0.02)
M0

M1

50 (2080)
40 (090)
20%
50 (3090)
60 (0100)
20%

DPQ (P 0.02)
M0

M1

37 (1161)
24.5 (071)
33.8%
41.5 (1085)
35 (087)
15.7%

VAS visual analog scale; MPQ (modified) McGill Pain Questionnaire; Paris Paris Questionnaire; DPQ (modified) Dallas Pain Questionnaire.
Data are median (range). Data were compared using a Mann-Whitney test.

E624 Spine Volume 31 Number 18 2006

Six-Month Results
The patients who at 1 month had had a good outcome
(50% improvement in the individual global score) continued in the study out to 6 months. In the manipulation
group, 34 patients were observed up for that period; in
the control group, there were 24. At 6 months, there
were 11 good outcomes (50% improvement in the individual global score at 1 month and 60% at 6 months)
in the manipulation group (22% of the patients); in the
control group, there were six good outcomes (12% of the
patients), a 10% absolute rate of improvement (95%
confidence interval, 6.5% to 22.2%). If a more stringent definition of a good outcome (e.g., 50% improvement at 1 month and 90% at 6 months) had been applied, the respective numbers and percentages would
have been 4 (8%) and 2 (4%), a 4% absolute rate of
improvement (95% confidence interval, 7.1% to
11.6%).

dence, sacrococcygeal, and sagittal pelvic rotation) showed

that coccyges that responded well were slightly straighter
than those that showed a poor response (Table 3).

Factors Associated With a Good Outcome

Subgroup analysis was performed to establish the factors
associated with a good outcome in the 2 groups.

Treatment Results
Despite its limitations (see below), the current study
broke new ground by comparing the outcomes, in
chronic coccydynia, of intrarectal coccygeal manipulation versus those obtained with a noneffective control
treatment. Under the study conditions, manipulation
was more effective than the control treatment. Improvement was definite using questionnaire results (the primary outcome measure), but only modest using the rate
of good outcomes (the secondary outcome measure). The
6-month rate of good results was 22%, which compares
favorably with the outcomes of 2 earlier, uncontrolled
studies, but meant that the efficacy was modest. The effectiveness observed appears to be less than that of intradiscal injections,15 or of coccygectomy, which is, however, performed in unstable coccyges only. 16 It is
conceivable that the various treatment methods are specific to different types of coccydynia, with manipulation
particularly suited to patients who are not relieved by
intradiscal injections (i.e., patients who do not have intradiscal inflammation). However, this hypothesis re-

Clinical Factors
A short time interval between the onset of the complaints
and initiation of treatment, and etiologic trauma, were
predictive of a good outcome following manipulation.
These factors were not found in the patients with a good
outcome following the control treatment, which suggests
that the improvement observed in the manipulation
group was attributable to the manipulation. Normal (undiminished) pelvic muscle tone was predictive of a good
outcome at 1 month, but not at 6 months (Table 3).
Radiographic Factors
Patients with a stable coccyx had a 30.8% chance of obtaining relief from manipulation, as against a 13.6%
chance of relief from the control treatment. On the other
hand, unstable coccyges did not respond any better to manipulation than they did to the control treatment. The angles measured on the stress radiographs (coccygeal inci-

Pain Intensity and Pain Impact

The intensity and nature of the pain, and the impact of
the pain on traveling in a car, did not have any predictive
value (Table 4).
Psychosocial Factors
Some of these factors had a good predictive value: manipulation worked best in patients with the lowest scores
in the McGill Questionnaire (in particular, for the affective items) and in the Depression and Interpersonal Relationships sections of the Dallas Pain Questionnaire
(Table 4). Interestingly, the same pattern was observed in
the Control group.
Discussion

Table 3. Association of Some Clinical and Radiographic Factors With Outcome in the 2 Groups

Clinical factors
Median time to treatment (mo)
Etiologic trauma (%)
Pain on getting up from sitting (%)
Diminished pelvic muscle tone (%)
BMI (mean)
Radiographic factors
Stable coccyx
Coccygeal incidence angle ()
Pelvic sagittal rotation angle ()
Sacrococcygeal angle ()

Patients With a
God Outcome*
Following Manipulation
(n 11)

Patients With a
Poor Outcome*
Following Manipulation
(n 39)

Patients With a
Good Outcome*
Following the
Control Treatment
(n 7)

Patients With a
Poor Outcome*
Following the
Control Treatment
(n 43)

5
55
73
45
23.4

18
28
64
44
24.6

19.8
0
43
NA
23.4

16.2
46.5
58
NA
24.5

8 (73%)
38.8
35.7
147

18 (46.2%)
22.9
39.8
138.7

3 (43%)
20
33
136

19 (44.2%)
24
37
130

NA not applicable (not measured in this group); BMI body mass index.
*A good outcome was defined as at least 50% improvement at 1 month, and at least 60% improvement at 6 months.

Treatment of Chronic Coccydynia Maigne et al E625

Table 4. Association of Baseline Pain Intensity and Pain Impact, and Psychosocial Factors, With Outcome
in the 2 Groups
Patients With
a Good Outcome*
Following Manipulation
(n 11)
Pain intensity and pain impact
VAS (median)
Paris Questionnaire (median)
DPQ Section IX: How much
does pain interfere with
traveling in a car?
DPQ Section X: How much
does pain interfere with
your job? (median)
Psychosocial factors
MPQ affective items (max 28
points) (median)
DPQ Section VIII: How much
does pain interfere with
your social life? (median)
DPQ Section XI: How much
control do you feel you
have over demands made
on you? (median)
DPQ Section XII: How much
control do you feel you
have over your emotions?
(median)
DPQ Section XIII: How
depressed have you been
since the onset of pain?
(median)
DPQ Section XIV: How much
do you think your pain
has changed your
relationship with others?
(median)
DPQ Section XV: How much
support do you need from
others to help you?
(median)
DPQ Section XVI: How much
do you think others
express irritation,
frustration, or anger
toward you because of
your pain? (median)

Patients With a
Poor Outcome
Following Manipulation
(n 39)

Patients With a
Good Outcome
Following the
Control Treatment
(n 7)

Patients With a
Poor Outcome
Following the
Control Treatment
(n 43)

50
50
60

62
50
70

66
50
60

71
50
65

50

50

70

55

12

12

43

54

60

50

30

30

25

40

18

24

31.5

40

10

23

23

35

16.5

10

14

8.5

8.5

10

4.5

8.5

10

VAS visual analog scale; DPQ (modified) Dallas Pain Questionnaire; MPQ (modified) McGill Pain Questionnaire; max maximum score (Unless otherwise
indicated, the maximum score that could be obtained in any one test was 100 points).
*A good outcome was defined as at least 50% improvement at 1 month, and at least 60% improvement at 6 months.

mains to be tested. It is also possible that the efficacy of

manipulation could be increased if more sessions were
provided. Thiele used up to 11 treatments and recommended up to 6 treatments.2 In light of our results, it
would be desirable to add 2 or 3 manipulation sessions
after the first month in an attempt to preserve the beneficial effects obtained during the first month. It should be
noted that there were no adverse events.
Limitations of the Study
To our knowledge, this is the first randomized controlled
study published on this topic. However, our study still
had several limitations.
First, it was, by definition, an open study. Therefore,
the validity of questionnaires is a critical aspect. Among

the 3 questionnaires used in the study, only the Paris

questionnaire had not been formally validated. However, its results were seen to vary in the same way as those
of the Dallas and the McGill questionnaires. These 2
questionnaires are commonly used in pain studies. They
assess the sensory and affective components of the pain
(McGill) and the extent to which chronic pain affects
daily and work activities, leisure activities, anxietydepression, and social interest (Dallas). Whereas the
McGill questionnaire may be used for any pain of vertebral origin, the Dallas questionnaire had to be modified,
by the elimination of the items relating to low back pain,
to make it suitable for use in patients with chronic coccydynia. The Paris Questionnaire was specifically de-

E626 Spine Volume 31 Number 18 2006

signed for coccydynia. It covers, specifically, the functional aspects of the pain, which are not assessed by the
other two questionnaires. It explores the pain when sitting (and differs, in this respect, from the VAS, which
only measures the intensity of the pain), when passing
from sitting to standing (very specific of coccydynia with
radiologic findings8), when walking and at night (pain on
these occasions is frequent in severe coccydynia), and
when traveling (the most sensitive detector of coccygeal
pain). This set of three questionnaires offers a correct
representation of the pain impact in chronic coccydynia.
With the 4 scales (VAS and the questionnaires), we obtained differences between the 2 groups (P ranging from
0.09 to 0.02). This fact is explained by the interrelationship between the scales. However, it adds value to our
study: the difference due to the intervention was greater
than the variability due to random error.
Second, the control treatment (lowest-power short
waves, applied to the sacrum) was not strictly speaking a
placebo. However, the control groups success rate at 6
months was only 12%, which suggests that the effectiveness of the control treatment may be considered to have
been very slight to nonexistent.
Finally, the patients were not followed up beyond 6
months; however, in a pilot study, good results had been
seen to be maintained in the period between 6 months
and 24 months.9
Subgroup Comparison
These analyses were performed post hoc and should
therefore be interpreted with caution. Some factors were
predictive of a good outcome of manipulation. Manipulation patients with a stable coccyx did better than those
with an unstable coccyx; whereas in the control group,
stability was not a factor of good prognosis. This would
appear to derive from the mechanism of action of manipulation. The technique can make the coccyx more
supple and, in particular, relieve the painful tension of
the pelvic muscles. We think that for this relief to be
durable, the causative coccygeal disc lesion must be minimal or healing. In that case, the painful muscle tension
may be assumed to be purely reflex and to lend itself to
treatment with manipulation.9
Timing
The procedure should be performed as early as possible:
In patients with coccydynia of more than 1 years standing, the results of manipulation were markedly poorer.
In cases with etiologic trauma, the rate of good results
was twice that in nontrauma-related cases. These 2 observations suggest that a minimal coccygeal disc lesion
(especially if caused by trauma) will heal spontaneously.
If there is still pain after 1 year, the cause should be
sought elsewhere. The culprits may, for instance, be intradiscal inflammation or psychosocial factors. The current study demonstrated the role of psychosocial factors
and appears, indeed, to be the first study to show the
importance of these factors in chronic coccydynia.

Mechanisms of Action of Coccygeal Manipulation

Our results suggest that manipulation works only if the
coccyx is stable. With an unstable coccyx, the pain is
likely to be due to the abnormality (hypermobility or
luxation) of the affected coccygeal motion segment, and
manipulation will not relieve the pain. With a stable coccyx, the pain may (in some cases) be due to painful tension of the pericoccygeal muscles (the levator ani and the
coccygeus), and manipulation may afford relief. With a
stiff coccyx, manipulation may render the coccyx more
supple and thus ensure that the subcutaneous tissues are
not subjected to excessive bony pressure when the subject is in the sitting position. To test this hypothesis,
postmanipulation dynamic radiographs would have
been required.
Conclusion
In light of the current study, coccygeal manipulation may
be recommended in the treatment of chronic coccydynia.
However, it should be borne in mind that its effectiveness is
mild. Manipulation appears to work best where the coccydynia is posttraumatic, of recent onset, and not associated
with coccygeal instability or adverse psychosocial factors.
Key Points
In chronic coccydynia, intrarectal manual treatment was better than the control treatment (22%
vs. 12%, at 6 months).
The effectiveness observed was mild.
Manipulation was more effective in posttraumatic coccydynia of recent onset and not associated with instability of the coccyx.
Psychosocial factors had an adverse effect on the
effectiveness of the treatment.

References
1. Sugar O. Coccyx, the bone named for a bird. Spine 1995;20:379 83.
2. Thiele GH. Coccydynia and pain in the superior gluteal region and down the
back of the thigh: causation by tonic spasm of the levator ani, coccygeus and
piriformis muscles and relief by massage of these muscles. JAMA 1937;109:
12715.
3. Mennell JB. The Science and Art of Joint Manipulation. London: Churchill,
1949.
4. Howorth B. The painful coccyx. Clin Orthop 1959;14:145 61.
5. Maigne R. Les manipulations vertebrales, 3rd ed. Paris: Expansion Scientifique Francaise, 1961:180.
6. Maigne R. Diagnosis and Treatment of Pain of Vertebral Origin: A Manual
Medicine Approach. Baltimore: Williams & Wilkins, 1996:339 40.
7. Maigne JY, Guedj S, Straus C. Idiopathic coccygodynia: lateral roentgenograms in the sitting position and coccygeal discography. Spine 1994;19:
930 4.
8. Maigne JY, Doursounian L, Chatellier G. Causes and mechanisms of common coccydynia: role of body mass index and coccygeal trauma. Spine
2000;25:30729.
9. Maigne JY, Chatellier G. Comparison of three manual coccydynia treatments: a pilot study. Spine 2001;26:E479 83; discussion E484.
10. Maigne JY, Vautravers P. Mechanism of action of spinal manipulative
therapy. Joint Bone Spine 2003;70:336 41.
11. Wray CC, Easom S, Hoskinson J. Coccydynia: aetiology, treatment. J Bone
Joint Surg Br 1991;73:335 8.
12. Maigne JY, Tamalet B. Standardized radiologic protocol for the study of
common coccygodynia and characteristics of the lesions observed in the

Treatment of Chronic Coccydynia Maigne et al E627

13.
14.

15.
16.

sitting position: clinical elements differentiating luxation, hypermobility,

and normal mobility. Spine 1996;21:2588 93.
Boureau F, Luu M, Doubre`re JF. Comparative study of the validity of four
French McGill Pain Questionnaire (MPQ) versions. Pain 1992;50:59 65.
Marty M, Blotman F, Avouac B, et al. Validation of the French version of the
Dallas Pain Questionnaire in chronic low back pain patients. Rev Rhum
Engl Ed 1998;65:126 34.
Rouhier B. Resultats des injections coccygiennes dans le traitement de la
coccygodynie chronique [Dissertation]. Paris: University of Paris V, 2003.
Maigne JY, Lagauche D, Doursounian L. Instability of the coccyx in coccydynia. J Bone Joint Surg Br 2000;82:1038 41.

Appendix
Annex 1. The Paris (functional coccydynia impact)
Questionnaire
Circle the number which best describes your response.
To ensure that your questionnaire will count, please answer all 5 questions.
1) When I am sitting:
I have no pain (0)
I have slight discomfort (1)

2)

3)

4)

5)

I can only sit in certain positions (2)

Sitting is virtually impossible for me (3)
When standing up from sitting:
I do not feel any worse than when I am sitting (0)
Standing up from sitting is painful
Standing up from sitting is extremely painful (2)
Standing still or walking:
I have no pain (0)
I am in pain (1)
Traveling (by car, train, bus, coach):
Traveling is not uncomfortable (0)
Traveling is slightly uncomfortable (1)
Traveling is extremely uncomfortable (2)
At night:
I have no pain during the night (0)
I sometimes have pain during the night (1)
I regularly get woken by pain during the night (2)
Total 10 final score (out of maximum of 100
points)