J Oral Maxillofac Surg

62:592-600, 2004

Effect of Exposed Inferior Alveolar

Neurovascular Bundle During Surgical
Removal of Impacted Lower Third Molars
Andrew Ban Guan Tay, BDS, MDS, FDS RCSEd, FAMS,*
and Wee Ser Go, BDS, MDS, FDS RCS, FAMS†

Purpose: In this prospective study, we sought to determine the incidence of inferior alveolar nerve
(IAN) paresthesia in patients with an exposed IAN bundle seen intraoperatively.
Patients and Methods: We included consecutive patients undergoing third molar surgery in whom an
exposed IAN bundle was seen in the third molar socket intraoperatively. Data recorded from patients
included radiographic findings, variations in surgical method, intraoperative findings, complications, and
postoperative sequelae. Patients were reviewed 1 week after surgery and evaluated subjectively and
objectively to determine the incidence of paresthesia when the IAN bundle was exposed. Patients with
paresthesia were followed for 2 years or until it resolved.
Results: An exposed IAN bundle was seen in 192 operation sites in 170 patients over a 5-year period,
of which 166 patients with 187 operation sites were included in this study. Thirty-eight operation sites
(20.3%) showed paresthesia at 1 week after surgery (95% confidence interval, 14.5% to 26.1%); 20 sites
(15.0%) showed abnormal objective assessment results. By 3 months from surgery, 57.9% of nerves had
recovered to normal sensation, 65.8% of nerves recovered by 6 months, and 71.1% of nerves recovered
by 1 year. Eight patients were lost to follow-up. Logistic regression showed that the operator, male
gender, older age, and root curvature were significant risk factors for paresthesia.
Conclusions: Sighting an exposed intact IAN bundle during third molar surgery indicates its intimate
relationship with the third molar and carries a 20% risk of paresthesia, with a 70% chance of recovery by
1 year from surgery.
© 2004 American Association of Oral and Maxillofacial Surgeons
J Oral Maxillofac Surg 62:592-600, 2004

The close proximity of the inferior alveolar nerve
(IAN) to the roots of the impacted lower third molar
is well known. Therefore, the possibility of injury to
the IAN resulting in paresthesia in the course of sur-
gical removal of impacted lower third molars has
been widely demonstrated. The reported incidence of
IAN injury after surgical removal of impacted lower
third molars ranges from 0.4% to 8.4%.1-25
The prediction of close proximity of the IAN to the
roots of the mandibular third molar relies mainly on
Received from the Department of Oral and Maxillofacial Surgery,
National Dental Centre, Singapore.
*Senior Registrar.
†Visiting Consultant.
Address correspondence and reprint requests to Dr Tay: Depart-
ment of Oral and Maxillofacial Surgery, National Dental Centre, 5,
Second Hospital Avenue, Singapore 168938; e-mail: tinyknots@
hotmail.com
© 2004 American Association of Oral and Maxillofacial Surgeons
0278-2391/04/6205-0078$30.00/0
doi:10.1016/j.joms.2003.08.033
radiographic analysis. The main drawback of radiog-
raphy lies in the fact that it provides a 2-dimensional
image of the relationship of the IAN to the lower third
molar. Nonetheless, Rood and Nooraldeen Shehab
(1990) reported that 3 of 7 radiologic signs indicate a
higher risk of IAN injury, implying close proximity of
the IAN to the lower third molar.26 The 3 radiologic
signs were diversion of the mandibular canal, darken-
ing of the root, and interruption of the white line.
Besides radiographic prediction, clinical features that
may suggest a higher risk of IAN injury are third
molars fully impacted in bone, horizontal impac-
tions,6,18 and older patients.21,23,34 The experience of
the operator,25 duration of surgery,24 and use of mul-
tiple sectioning or deep drilling with a bur23 have also
been reported to have a relationship with the inci-
dence of IAN injury.
Intraoperative events that indicate or suggest close
proximity of the IAN with the lower third molar
include observation of the neurovascular bundle in
the socket3,6 and excessive hemorrhage from the
socket.6 In 1960, Howe and Poyton3 used visual iden-

592
TAY AND GO
tification of the neurovascular bundle in the lower
third molar socket as the criterion for a “true relation-
ship” of the third molar with the IAN. Third molars
that appeared to be in an intimate relationship with
the mandibular canal on intraoral radiography were
described as being in “apparent relationship.” In
1980, Kipp et al6 reported in a retrospective study
that both the clinical observation of the neurovascular
bundle during surgery and excessive hemorrhage
from the socket correlated with a higher incidence of
paresthesia of the IAN postoperatively.
The aim of this prospective study was to determine
the incidence of IAN paresthesia in patients with an
exposed inferior alveolar neurovascular bundle
(IANB) found intraoperatively and to determine if
such a finding of an exposed IANB intraoperatively is
predictive of paresthesia of the IAN.
Patients and Methods
In a prospective case series study, consecutive pa-
tients who underwent surgical excision of mandibular
third molars and in whom an exposed IAN bundle
was seen in the third molar socket intraoperatively
were included in the study. Cases included those
operated on while under local or general anesthesia
by surgeons from the Department of Oral and Maxil-
lofacial Surgery in the Government Dental Clinic/
National Dental Centre (GDC/NDC) in Singapore.
A common surgical procedure was used by sur-
geons in this department. After local anesthesia was
administered by IAN block and buccal infiltration, a
buccal envelope mucoperiosteal flap was raised, bone
around the third molar crown was removed with a
bur to expose it, the crown of the tooth was divided
with a bur as appropriate, and all fragments of the
tooth were elevated out. Any remnant follicle was
removed with a curette and the socket was irrigated
with saline. The socket was then briefly explored for
fragments or exposed tissue with light and suction.
Hemostasis was achieved if necessary with gauze
packing or with hemostatic material such as Surgicel
(Ethicon, Neuchatel, Switzerland) if necessary before
closure of the wound with sutures. Care was taken
not to compress an exposed neurovascular bundle if
it was seen. There usually was no further effort made
to explore exposed tissue through further bone re-
moval around the tissue, except for removal of the
interradicular septum if required for visual access.
The use of corticosteroids to reduce postoperative
swelling was left to the discretion of the operating
surgeon.
The criteria for identifying an exposed IANB were
as follows:
593
● A tubular, pale or whitish structure in the third
molar socket
● Structure oriented in a mesiodistal direction
● Structure at a level in the socket consistent with
the radiographic appearance
Exposed tissue in the lingual wall that was not
tubular or had no distinctly mesiodistal orientation
would more likely be lingual tissue exposed through
a lingual plate perforation and therefore was not in-
cluded. Either periapical or panoramic radiographs
were used to check the level of the IAN canal against
the third molar root.
A proforma was used to document the intraopera-
tive findings and the postoperative sequelae. Data
documented included whether the surgeon thought
the IAN was manipulated or damaged in any way,
whether the third molar was divided, the method of
tooth division, whether roots were fractured or other
complications occurred, and the use of hemostatic
material and corticosteroids.
Each patient who was noted to have an exposed
IANB was reviewed at 1 week after the surgery and
assessed subjectively for the sensation of the lower lip
and chin by stroking the lower lip and chin with a
finger and asking if the sensation felt was normal or
abnormal. If the patient reported feeling abnormal
sensation, objective assessment was then performed
with: light touch with cotton wisp, direction sense by
stroking lightly with a probe, 2-point discrimination
with a Disk-criminator (Smith & Nephew Royan Inc,
Germantown, WI), pinprick with a needle, and vital-
ity of a sound second or first premolar with an elec-
tropulp vitality tester. For light touch, direction sense,
and pinprick, the responses recorded were normal if
the patient was able to feel the test stimulus and
abnormal if the patient was unable to feel the stimu-
lus. For 2-point discrimination, a normal result was
recorded if the discrimination distance was less than
14 mm and an abnormal result was recorded if the
discrimination distance was 14 mm or more. Results
from both affected and unaffected sides were re-
corded for comparison with unaffected sides used as
controls. Patients with sensory alteration were fol-
lowed up monthly for at least 1 year and then every 2
to 3 months for an additional year or until sensation
returned to normal. The majority of the objective
neurosensory testing was performed by the authors.
An attempt to minimize the number of patients lost to
follow-up by calling patients who had missed fol-
low-up visits for new appointments or for telephone
interviews.
Radiographs were retrieved and evaluated at the
end of the study. Impaction of the third molars was
categorized by depth (superficial, moderate, and
deep) and type of impaction (vertical, mesioangular,
594 EXPOSED IAN BUNDLE DURING THIRD MOLAR SURGERY

Table 1. PATIENT CHARACTERISTICS (N ⴝ 166)

because the clinical records were lost and could not
be found during the analysis phase of this study.
Characteristic Of the remaining 166 patients, 21 patients (12.7%)
had bilateral IAN bundles exposed during lower third
Gender, n (%) molar surgery. The patient characteristics (gender,
Male 76 (45.8)
Female 90 (54.2) age, race, third molar side) are shown in Table 1. The
Age (yr) gender distribution was nearly equal, with more fe-
Mean 25.8 males than males (mean age, 25.8 years [SD, 7.9
SD 7.9 years]; median age, 23 years; age range, 12 to 78
Median 23 years).
Range 12 to 78
Race, n (%) The features of third molars are shown in Table 2.
Chinese 133 (80.1) Of the 187 operation sites, 46.5% were on the right
Malay 18 (10.8) side and 53.5% were on the left side. The majority of
Indian 12 (7.2) third molars had 2 roots (70.6%) and were of moder-
Others 3 (1.8) ate depth of impaction (65.2%). Most of third molars
Side of third molar
Left only 79 (47.6) had mesioangular (48.1%) or horizontal (44.4%) im-
Right only 66 (39.8) pactions. Nearly a quarter of third molars had groov-
Bilateral 21 (12.7) ing of the root by the IAN (23.0%) or were curved
Total 166 (100.0) (23.0%).
The features of third molar surgeries are illustrated
in Table 3. There were more surgical procedures
horizontal, and distoangular). The criteria for deter- performed under general anesthesia than under local
mining third molar depth used the relationship of the anesthesia; one procedure was carried out under local
deepest part of the third molar crown to the adjacent anesthesia with intravenous sedation. The great ma-
second molar: superficial (at or above the second jority of third molars were removed with division or
molar cervicoenamel junction [CEJ]), moderate (be- sectioning (92.5%). Less than a tenth of the third
low the second molar CEJ and above the midheight
point of the second molar distal root), and deep (be-
low the midheight point of the second molar distal Table 2. THIRD MOLAR FEATURES (N ⴝ 187)
root). The type of impaction of third molars was
determined by using the relationship of the long axis Characteristic
of the third molar to the occlusal plane of the adjacent Side of third molar, n (%)
molars: horizontal (⬍30°), mesioangular (30° to 70°), Left 100 (53.5)
vertical (70° to 110°), and distoangular (⬎110°). Right 87 (46.5)
No. of roots, n (%)
0 (incomplete root) 7 (3.7)
Results 1 35 (18.7)
2 132 (70.6)
An exposed IAN bundle was seen in the third molar 3 11 (5.9)
socket in a total of 192 operation sites in 170 patients 4 2 (1.1)
over a 5-year period from June 1996 to May 2001. Of Depth of impaction, n (%)
Superficial 10 (5.3)
these, 4 patients (5 operated sides) were excluded Moderate 122 (65.2)
from the study for the following reasons: in 2 patients, Deep 54 (28.9)
there was a variation in the surgical technique that Missing data 1 (0.5)
would have had a possible effect on the surgical Type of impaction, n (%)
outcome, and in 2 patients (3 sides), the records Vertical 11 (5.9)
Mesioangular 90 (48.1)
contained an error or were lost. The lingual split Horizontal 83 (44.4)
technique was used in 1 operation instead of the Distoangular 2 (1.1)
prescribed buccal approach, with paresthesia of the Missing data 1 (0.5)
IAN occurring postoperatively. One patient had a cyst Root grooved by inferior alveolar nerve, n (%)
surrounding the third molar root that was peeled off Yes 43 (23.0)
No 139 (74.3)
the exposed IAN bundle with no postoperative par- Missing data 5 (2.7)
esthesia. There was an unsolvable error in recording Root curved, n (%)
the side in which the nerve bundle was seen in one Yes 43 (23.0)
patient who underwent bilateral third molar removal No 140 (74.9)
and developed unilateral paresthesia. Another patient Missing data 4 (2.1)
Total 187 (100.0)
with bilateral IAN bundle exposure was excluded
TAY AND GO 595

Table 3. THIRD MOLAR SURGERY FEATURES

The gender distribution of patients with paresthesia
(N ⴝ 187) was nearly equal but with more males (54.1%) than
females (45.9%). The median age was 27 years, and
Characteristic the age range was 20 to 78 years. The racial distribu-
Anesthesia tion of patients with paresthesia was similar to that of
General 103 (55.1) all patients with an exposed nerve.
Local (1 case with intravenous sedation) 84 (44.9) The results of the objective neurosensory assess-
Tooth division ment using light touch, direction sense, 2-point dis-
Yes 173 (92.5)
crimination, and pinprick are shown in Table 4. Of
No 14 (7.5)
Root fractured the 38 operation sites that had paresthesia after third
Yes 17 (9.1) molar surgery, 10 sites showed normal results for all 4
No 166 (88.8) tests of light touch, direction sense, 2-point discrimi-
Missing data 4 (2.1) nation, and pinprick, whereas 6 sites showed abnor-
Complication intraoperative
Yes 25 (13.4) mal results for all 4 tests. Vitality testing of the canine
No 158 (84.5) or premolar was found to be of no clinical use in
Missing data 4 (2.1) evaluating injury-induced sensory alteration of the
Hemostatic material IAN, as there was no difference in the vitality testing
Yes 34 (18.2)
results between sites with paresthesia and those with-
No 153 (81.8)
Steroids used out.
Yes 78 (41.7) Eight patients were lost to follow-up before com-
No 109 (58.3) plete neurosensory recovery at various points of time
Inferior alveolar nerve manipulated after surgery. Three patients did not return for fol-
Yes 6 (3.2)
No 180 (96.3) low-up within the first month after surgery, whereas 5
Missing data 1 (0.5) patients with paresthesia at their last visit were lost to
Inferior alveolar nerve damaged follow-up only 1 year after surgery. The patients lost
Yes 0 (0.0) to follow-up were predominantly male (75%), older
No 186 (99.5)
(median age, 31.5 years; age range, 21 to 78 years),
Missing data 1 (0.5)
Total 187 (100) and Chinese (87.5%).
Recovery of cases with paresthesia over time is
shown in Table 5 and illustrated in Figure 1 as the line
molar surgeries were complicated by a fractured root with circle markers; the upper line with diamond
(9.1%). A slightly higher percentage of surgeries had markers shows the trend of recovery assuming that
other complications (13.4%), mostly bleeding in the patients who were lost to follow-up had persistent
operation site. Hemostatic material (Surgicel) was paresthesia.
used in a higher proportion of third molar operation The proportion of patients who still had numbness
sites (18.2%). Steroids were used in 41.7% of surger- at 1 year after surgery was 5.9% of all operations
ies; except in 1 patient who received oral dexameth- where the IAN bundle was exposed, or 28.9% of
asone, all other patients were given intravenous dexa- operations that resulted in paresthesia. Assuming a
methasone (usually 8 mg) intraoperatively. worst case scenario, that the 3 patients who were lost
Of all of the 187 operation sites in which an ex- to follow-up before reaching 1 year had persistent
posed IAN neurovascular bundle was noted, 38 sites paresthesia at 1 year after surgery, the highest possi-
(20.3%; 95% confidence interval [CI], 14.5% to 26.1%) ble percentage of patients with persistent paresthesia
had sensory alteration at 1-week review. Only 1 pa- at 1 year was 7.5% of all operations and 36.8% of
tient with IAN bundle exposure bilaterally had pares- procedures resulting in neurosensory deficits. The
thesia on both sides. These 37 patients were followed proportion of patients with persistent paresthesia at 2
up as far as possible until the paresthesia resolved. years after surgery was 2.1% of all operations or 10.5%

Table 4. OBJECTIVE SENSORY TESTING RESULTS AT 1 WEEK AFTER SURGERY

Test Light Touch Direction Sense Two-Point Discrimination Pinprick

Abnormal 11 (28.9%) 13 (34.2%) 17 (44.7%) 16 (42.1%)

Normal 25 (65.8%) 23 (60.5%) 18 (47.4%) 20 (52.6%)
No record 2 (5.3%) 2 (5.3%) 3 (7.9%) 2 (5.3%)
Total 38 (100%) 38 (100%) 38 (100%) 38 (100%)
596 EXPOSED IAN BUNDLE DURING THIRD MOLAR SURGERY

Table 5. RELATIONSHIP OF NUMBER OF SITES WITH PARESTHESIA TO TIME AFTER SURGERY

Time After Surgery

1 week 1 month 3 months 6 months 1 year 2 years

No. of sites with paresthesia 38 30 16 13 11 4

Percent of 187 operations 20.3 16.0 8.6 7.0 5.9 2.1
Percent of 38 operations with paresthesia 100 78.9 42.1 34.2 28.9 10.5

of operations that resulted in paresthesia. Assuming
that the 8 patients lost to follow-up had paresthesia
up to 2 years postoperatively, the highest possible
incidence of persistent paresthesia at 2 years after
surgery was 6.4% of all operations and 31.6% of pro-
cedures resulting in paresthesia.
Two patients who had no paresthesia did not have
their sensory status checked at 1-week postoperative
review, and an additional 8 patients with paresthesia
at 1 week postoperatively failed to return for follow-
up. These 10 patients were contacted by telephone
later to ascertain the status of their paresthesia. They
could not be assessed with objective neurosensory
testing, and the information of the time of recovery in
these patients was obtained retrospectively. The
records of these 10 patients (10 sides) were censored
and the data set was reanalyzed. The incidence of
paresthesia at 1 week after third molar surgery in the
censored data set was 16.9% (95% CI, 11.7% to
23.3%); this was not significantly different from the
incidence derived from the uncensored data set.
There was no significant difference between the un-
censored and censored populations with regard to
gender, age, and race. The proportion of patients who
recovered sensation at various time points in the
censored group was found to be similar to that of the
uncensored cases. We therefore used the data from
the uncensored cases in the analysis for risk factors of
paresthesia.
The proportion of procedures in which the sur-
geon thought that the IAN bundle had been manipu-
lated was 3.2% (6 of 186) of all operations with re-
corded data (Table 3). Half of the procedures (3 of 6)
in which the nerve bundle was manipulated resulted
in paresthesia. There was a higher percentage of ma-
nipulation of the IAN bundle in the group with post-
operative paresthesia (7.9%) compared with the
group without postoperative paresthesia (2.0%).
However, if the 6 patients in whom the nerve bundle
had been manipulated are excluded, the overall pro-
portion of procedures that resulted in paresthesia was
not significantly affected: 19.4% (35 of 180) compared
with 20.3% (38 of 187). There were no cases in the
study where the surgeon recorded having directly
damaged the IAN bundle.
The outcome of paresthesia stratified according to
the grade of seniority (Table 6) showed that the inci-
dence of paresthesia was lowest among the registrar/
senior registrar grade (intermediate seniority) and
highest among the consultants (greatest seniority).
The incidence of paresthesia varied among the vari-
ous individual operators and ranged from 9.2% to
31.3%. The surgeon who had contributed the largest
number of cases (40.6%) to this study also had the
lowest incidence of paresthesia (9.2%). Three other
surgeons contributed 34.8% of cases (32, 22, and 11
cases, respectively) to the study. The individual inci-
dence of paresthesia for each of these 3 surgeons
ranged from 27.3% to 31.3%. An additional 11 sur-
geons who contributed cases each provided small
numbers (⬍10 each).

Table 6. RELATIONSHIP OF PARESTHESIA TO

SENIORITY OF SURGEON

Registrar/Senior
Grade Consultant Registrar Resident

No. with
paresthesia/
total 21/74 13/97 4/16
Percent with
paresthesia 28.4 13.4 25.0
FIGURE 1. Recovery of sensory alteration over time.
TAY AND GO
Logistic regression was used to check the correla-
tion between various aspects of the third molar sur-
gery with the incidence of paresthesia. The main
effects of the following variables were considered in
the regression model: patients’ gender, age, and race;
tooth side; root number; depth of impaction; type of
impaction; surgery under local anesthesia or general
anesthesia; tooth divided or not; root grooved by IAN;
root curved or not; root fractured in surgery; compli-
cation in surgery; hemostatic material used or not;
steroids used or not; and whether manipulation of
IAN bundle occurred. The outcomes of cases oper-
ated on by the busiest surgeon compared with other
surgeons were also included in the logistic regression
analysis. Eight sites with missing values were ex-
cluded from the analysis. There were a total of 179
sites included in the logistic regression analysis. None
of the main effects were found to be significant in the
full model (significant level at .05), except the main
effect of the operator (P ⫽ .0063; patients treated by
the busiest surgeon had less risk of paresthesia than
those treated by other operators, odds ratio [OR] ⫽
0.198; 95% confidence interval, 0.058 to 0.600).
Using backward selection with removal level at 0.1,
we found that the main effects of gender (P ⫽ 0.0337,
females had less risk of paresthesia than males, OR ⫽
0.410; 95% CI, 0.180 to 0.934), age (P ⫽ 0.0281, an
increase of age in 1 year increased the odds of pares-
thesia by 6.9%, OR ⫽ 1.069; 95% CI, 1.007 to 1.135),
operator (P ⫽ 0.0021, patients treated by the busiest
surgeon had less risk of paresthesia than those treated
by other doctors, OR ⫽ 0.229; 95% CI, 0.089 to
0.586), and root curvature (P ⫽ 0.0332, patients who
have curved roots had more risk of paresthesia than
those who do not, OR ⫽ 2.537; 95% CI, 1.077 to
5.977) were significant.
Discussion
The published incidence of IAN injury after surgical
removal of impacted lower third molars ranges from
0.4% to 8.4%.1-25 However, some of the often quoted
studies are not necessarily comparable for several
reasons: some studies are retrospective or are of un-
certain prospectivity,1,2,4-6,10,11,13,14 2 studies were tri-
als and not observational studies,16,20 1 was an age-
controlled study,7 and another was a questionnaire
survey.14 In some studies, the incidence of nerve
injury was presented on a per-patient basis rather
than on a per-operation site basis13,14,16 or did not
specify the time period from surgery after which
paresthesia was noted.7,11,13,14 In 1 study, the inci-
dence of nerve injury included both IAN and lingual
nerve.22 Studies using the lingual split technique4,9,10
or modifications of it8 were excluded to allow better
comparison with this present study. Inclusion of only
597
prospective, large-sample (ⱖ300 third molar surgery
sites) observational studies that did not use the lingual
split technique, that focused on the incidence of IAN
injury in third molar surgery, and that published inci-
dences of IAN injury per operation site at around 1
week after surgery with sensory testing to confirm
sensory alteration resulted in 5 studies with an inci-
dence range of 1.3% to 3.6%.15,19,23–25 The incidence
of sensory disturbance when the IAN bundle is seen
exposed during lower third molar surgery (20.3%;
95% CI, 14.8% to 26.8%) is significantly higher than
the incidence of IAN sensory alteration with third molar
surgery in the literature, which is expected intuitively.
This result is still significantly higher compared with the
highest reported incidence of IAN sensory alteration
with third molar surgery in literature (8.4%).20
In this study, the criteria for sensory alteration are
patient-reported paresthesia in response to stroking
of the lower lip and chin, with objective sensory
assessment using light touch, direction sense, 2-point
discrimination, and pinprick. Vitality testing of the
ipsilateral canine or premolar was not found to be of
clinical use. Objective sensory assessment in 28 sites
produced an abnormal result in at least 1 of the 4
tests, with 6 sites showing abnormalities in all 4 tests.
In 10 sites, the 4 tests produced normal results. This
indicates that the objective testing methods as they
were used were not able to identify less severe sen-
sory alterations. Of these 10 sites with normal objec-
tive test results, 1 site continued to show paresthesia
up to 2 years. The incidence of sensory alteration after
third molar surgery where an IAN bundle is exposed
and confirmed by objective neurosensory assessment
is 15.0%, which is higher than the reported incidence
of sensory alteration after third molar surgery in the
literature.
Howe and Poyton3 reported that of 1,355 mandib-
ular third molars in 1,046 patients seen over a 2-year
period, 101 third molars (7.5%) were found to be in
“true relationship” with the IANB; that is, the IANB
was visually identified in the socket during surgery.
An “apparent relationship” was found in 837 teeth
(61.8%) where the lower third molars were in appar-
ent superimposition with the IAN on radiography.
The incidence of IAN paresthesia in the 101 third
molars in “true relationship” with the IANB was 35.6%
(36 of 101). This was significantly higher than the
incidence of 2.7% (34 of 1,254) of the other third
molars that were not in “true relationship” with the
IANB. Kipp et al6 found in their retrospective study
that in 8 cases (0.6%) of a total of 1,377 extractions,
the mandibular canal was observed intraoperatively.
Of these 8 cases, 2 (25%) resulted in paresthesia
postoperatively. Gülicher and Gerlach23 reported that
of 1,106 mandibular third molars, 39 operated sides
showed paresthesia 1 week after surgery. In their
598
study, the nerve canal was “opened” in 132 sites, of
which 22 sites (16.7%) resulted in paresthesia, com-
pared with 17 of 974 sites (1.7%) that did not involve
the nerve canal (␹2 test, P ⬍ .001). Valmaseda-Castel-
lón et al24 also reported that exposure of the IAN
increased the risk of nerve damage. The incidence of
IAN paresthesia at 1 week after surgery, during which
the IAN bundle was seen exposed, is 20.3% in our
study and is significantly lower than Howe and Poy-
ton’s result3 (␹2 test, P ⫽ .005) but closer to the other
studies.6,23
Possible mechanisms of nerve injury in patients
who sustained sensory deficits after third molar sur-
gery with observed, intact IAN bundles include com-
pression injury or crush injury. The process of nerve
regeneration after compression or less severe crush
injuries usually requires several weeks to 6 months.27
If there is no sensory recovery during this time, the
loss of continuity in the nerve trunk should be ex-
pected. Peripheral nerve injuries have been classified
in various ways,28 with the best-known systems pro-
posed by Seddon29,30 and Sunderland.29,31,32 Both clas-
sification systems are based on the severity of nerve
injury. The Sunderland classification expands the Sed-
don classification of neuropraxia, axonotmesis, and
neurotmesis into 5° of nerve injury in increasing order
of severity.
Because the IAN bundles were seen to be clinically
intact, the initial assumption was that the IAN injuries
in this study at worst would be Sunderland third-
degree injuries. The sites with altered sensation that
recovered within 3 months (57.9%) were most likely
Sunderland first- and second-degree nerve injuries,
and the majority of sites with altered sensation at 6
months (34.2%) were probably Sunderland third-de-
gree nerve injuries. Collateral reinnervation from ad-
jacent nerves may account for some instances of early
sensory recovery.27 However, the persistence of sen-
sory alteration in 28.9% of sites at 1 year suggests the
presence of Sunderland fourth-degree injury in some
instances, with intraneural fascicular disruption
and/or scarring. It should be remembered that the
time of recovery of nerve injury in the IAN has not
been correlated in humans with the Sunderland or
Seddon classifications. From a clinical standpoint, the
finding of an exposed intact IAN bundle during third
molar surgery suggests that there is a 1:5 chance of
developing paresthesia; patients with postoperative
paresthesia have a 1:4 chance of it persisting for 1
year or more.
In this study, the number of sites or nerves with
paresthesia reduced markedly by nearly 60% after 3
months; approximately 65% of sites recovered by 6
months postoperatively, and about 70% of sites re-
gained sensation at 1 year after surgery. From 1 to 2
years after surgery, there was a second smaller de-
EXPOSED IAN BUNDLE DURING THIRD MOLAR SURGERY
crease of about 18% of sites (lower line with circle
markers in Figure 1). However, if it is assumed that all
patients who were lost to follow-up had persistent
paresthesia (upper line with diamond markers), the
trend of recovery plateaus at 1 year after surgery. The
true trend would lie somewhere in between these 2
lines. Other studies have noted the majority of IAN
injuries recover within the first 4 to 6 months. Howe
and Poyton3 reported that 90% (63 of 70) patients
with “true relationship” of the IAN to the third molar
recovered within 4 months, with 2.9% (2 of 70) per-
sisting at 1 year. Rood9 noted that the majority of 105
IAN injuries after lower third molar surgery with the
lingual split technique recovered by 4 months after
surgery and that 4.8% (5 of 105) continued to have
sensory deficit up to 1 year. Ferdousi and MacGre-
gor,33 reporting on 56 injuries to nerves in the trigem-
inal system from various causes, noted that 60% of
patients recovered after 6 months from injury. Robin-
son34 reported that 53.8% of 13 patients with com-
pression nerve injuries of the IAN after various types
of oral surgery recovered by 4 months after surgery,
with the rest having sensory deficit persisting for
more than 1 year. In our series, 28.9% (11 of 38) of
sites with paresthesia (5.9% of all operations) had
persistent paresthesia for 1 year compared with 2.9%
and 4.8% reported by Howe and Poyton3 and Rood,9
respectively, but closer to 23.0% (12 of 52) in Car-
michael and McGowan’s report18 and 26.7% (4 of 15)
in the study by Valmaseda-Castellón et al.24 It should
be noted that objective neurosensory assessment was
performed only in the study by Valmaseda-Castellón
et al.24 Robinson34 reported a higher 46.2% (6 of 13)
of patients with abnormal sensory function lasting
more than 1 year. There appears to be some possibil-
ity of further recovery after 1 year, in at least 5% of
sites.
Statistical analysis showed that male gender, older
age, root curvature, and the operator correlated sig-
nificantly with an increased risk of IAN sensory deficit
in third molar surgeries where the IANB was clinically
observed. A higher incidence of paresthesia was
noted in males in another article21 and also associated
with older patients in other studies.7,21,23,24 A few
other reports did not find similar associations with
gender23,24 or age.22,25 Deep or fully impacted third
molars were associated with greater risk of IAN par-
esthesia,18,23 as was radiographic indications of close
proximity between the third molar and IAN.23,24 The
use of general anesthetic was significantly associated
with a higher risk of IAN injury in 1 article22 but not
others21,24 nor in this present study. It is clinically
significant that the use of corticosteroids was not
associated with a lower incidence of IAN sensory
deficit. The correlation of root curvature with in-
creased risk of IAN sensory deficit is plausible but is
TAY AND GO
difficult to fully substantiate, as there are no repro-
ducible and reliable criteria that different examiners
could use to diagnose a curved root consistently. The
influence of individual surgeons on the outcome of
IAN numbness was noted in a few reports,23,24 as was
the experience of the operator.25 This is not surpris-
ing because the outcome of surgery is to a great
extent dependent on the skill and technique of the
operator. The most consistent predictors of nerve
injury risk in third molar surgery appear to be older
patients and features of close proximity between the
third molar and IAN, specifically radiographic signs
and deep impactions.
Limitations in this study are the disproportionate
contribution of cases from various surgeons and the
difference in the incidence of IAN paresthesia be-
tween individual surgeons. Cases in this study were
contributed by staff of all grades, from residents to
consultants in the department. However, a significant
proportion of cases was contributed to this study by
one surgeon with a number of surgeons contributing
a variable number of cases each. Two other surgeons
in the department who had a significant patient load
did not record any cases of exposed IAN in the study.
This strongly suggests that there was a degree of
selection bias. This may be related to the absence of
an impartial examiner to evaluate all third molar op-
eration sites during the period that this study was
performed. The notably different rates of IAN pares-
thesia between surgeons (who contributed suffi-
ciently large numbers of cases), ranging from 9.2% to
31.3%, also suggests either a selection bias or con-
founding related to expertise. It is difficult to deter-
mine the extent to which this bias or confounding
affects the overall incidence of IAN sensory alteration.
Identification of the neurovascular bundle may
sometimes be difficult because of inadequate access
to visualize the bundle (eg, in a lingually tilted third
molar) or hemorrhage obscuring the bundle. Misiden-
tification of other soft tissue as the IAN bundle is also
possible, such as lingual soft tissue seen in a lingual
wall perforation, but the criteria for identification of
the IANB should have minimized this. Tomography
can be used postoperatively to determine if a sighted
tubular structure in the socket during surgery is in-
deed the IANB but may not affect subsequent man-
agement.
Another limitation of this study was that the exam-
iner evaluating sensory alteration was not blinded to
the side where the exposed IAN was seen. In addi-
tion, the objective neurosensory testing format used
in this study did not provide a qualitative grading of
sensory impairment that correlated reliably with the
severity of injury. It would have been useful other-
wise to monitor the trend of sensory recovery quali-
599
tatively and to estimate the need for surgical interven-
tion of nerve injury.
The visual sighting of an intact IANB during third
molar surgery indicates that the third molar is in
intimate relationship with the IAN and carries an
about 20% risk of subsequent paresthesia, with a
range of approximately 15% to 25% risk (95% CI). The
incidence of sensory alteration after exposure of the
intact IANB, confirmed by an abnormal neurosensory
test result, was 15%. About 60% of cases may be
expected to recover after 3 months, around 65% re-
cover within 6 months, and approximately 70% re-
cover by 1 year after surgery. Sensory alteration that
fails to resolve after 1 year is more likely to persist,
although gradual recovery is still possible.
Acknowledgments
The authors would like to thank Shen Liang, Biostatistician,
National Medical Research Council, Clinical Trials and Epidemiol-
ogy Research Unit, for performing the statistical analysis; Dr Goh
Bee Tin, Department of Oral and Maxillofacial Surgery, National
Dental Centre, for her invaluable assistance in data compilation;
and Prof John R. Zuniga, Department of Oral and Maxillofacial
Surgery, University of North Carolina, for his very helpful com-
ments and suggestions in the writing of this paper. The authors also
gratefully acknowledge the contribution and support of all the
surgeons and nurses in the Department of Oral and Maxillofacial
Surgery, National Dental Centre, toward this study.
References
1. Robinson TO: Paresthesia: A postoperative complication fol-
lowing removal of the mandibular third molar. Am J Orthod
26:278, 1940
2. Frank VH: Paresthesia: Evaluation of 16 cases. J Oral Surg 17:27,
1959
3. Howe GL, Poyton HG: Prevention of damage to the inferior
dental nerve during the extraction of mandibular third molars.
Br Dent J 109:355, 1960
4. Rud J: The split-bone technique for removal of impacted man-
dibular third molars. J Oral Surg 28:416, 1970
5. Van Gool AV, Ten Bosch JJ, Boering G: Clinical consequences
of complaints and complications after removal of the mandib-
ular third molars. Int J Oral Surg 6:29, 1977
6. Kipp DP, Goldstein BH, Weiss WW: Dysesthesia after mandib-
ular third molar surgery: A retrospective study and analysis of
1,377 surgical procedures. J Am Dent Assoc 100:185, 1980
7. Bruce RA, Frederickson GC, Small GS: Age of patients and
morbidity associated with mandibular third molar surgery.
J Am Dent Assoc 101:240, 1980
8. Hochwald DA, Davis WH, Martinoff J: Modified disto-lingual
splitting technique for the removal of impacted mandibular
third molars. Incid Pathol 56:9, 1983
9. Rood JP: Degrees of injury to the inferior alveolar nerve sus-
tained during the removal of impacted mandibular third molars
by the lingual split technique. Br J Oral Surg 21:103, 1983
10. Rud J: Re-evaluation of the lingual split-bone technique for
removal of impacted mandibular third molars. J Oral Maxillofac
Surg 42:114, 1984
11. Goldberg MH, Nemarich AN, Marcoll WP: Complications after
mandibular third molar surgery: A statistical analysis of 500
consecutive procedures in private practice. J Am Dent Assoc
111:277, 1985
12. Osborn TP, Frederickson G, Small IA, et al: A prospective study
of complications related to mandibular third molar surgery.
J Oral Maxillofac Surg 43:767, 1985
600
13. Sisk AL, Hammer WB, Shelton DW, et al: Complications follow-
ing removal of impacted third molars. The role of the experi-
ence of the surgeon. J Oral Maxillofac Surg 44:855, 1986
14. Alling CC: Dysesthesia of the lingual and inferior alveolar
nerves following third molar surgery. J Oral Maxillofac Surg
44:454, 1986
15. Wofford DT, Miller RI: Prospective study of dysesthesia follow-
ing odontectomy of impacted mandibular third molars. J Oral
Maxillofac Surg 45:15, 1987
16. Von Arx DP, Simpson MT: The effect of dexamethasone on
neuropraxia following third molar surgery. Br J Oral Maxillofac
Surg 27:477, 1989
17. Rood JP: Permanent damage to inferior alveolar and lingual
nerves during the removal of impacted mandibular third mo-
lars. Comparison of two methods of bone removal. Br Dent J
172:108, 1992
18. Carmichael FA, McGowan DA: Incidence of nerve damage
following third molar removal. A West Scotland Oral Surgery
Research Group study. Br J Oral Maxillofac Surg 30:78, 1992
19. Schultze-Mosgau S, Reich RH: Assessment of inferior alveolar and
lingual nerve disturbances after dentoalveolar surgery, and of
recovery of sensitivity. Int J Oral Maxillofac Surg 22:149, 1993
20. Lopes V, Mumenya R, Feinmann C, et al: Third molar surgery: An
audit of the indications for surgery, post-operative complaints and
patient satisfaction. Br J Oral Maxillofac Surg 33:33, 1995
21. Black CG: Sensory impairment following lower third molar sur-
gery: A prospective study in New Zealand. N Z Dent J 93:68, 1997
22. Brann CR, Brickley MR, Shepherd JP: Factors influencing nerve
damage during lower third molar surgery. Br Dent J 186:514,
1999
EXPOSED IAN BUNDLE DURING THIRD MOLAR SURGERY
23. Gülicher D, Gerlach KL: Sensory impairment of the lingual and
inferior alveolar nerves following removal of impacted mandib-
ular third molars. Int J Oral Maxillofac Surg 30:306, 2001
24. Valmaseda-Castellón E, Berini-Aytes L, Gay-Escoda C: Inferior
alveolar nerve damage after lower third molar surgical extrac-
tion: A prospective study of 1117 surgical extractions. Oral
Surg Oral Med Oral Pathol Oral Radiol Endod 92:377, 2001
25. Bataineh AB: Sensory nerve impairment following mandibular
third molar surgery. J Oral Maxillofac Surg 59:1012, 2001
26. Rood JP, Nooraldeen Shehab BAA: The radiological prediction
of inferior alveolar nerve injury during third molar surgery. Br J
Oral Maxillofac Surg 28:20, 1990
27. Zuniga JR: Normal response to nerve injury. Oral Maxillofac
Surg Clin North Am 4:323, 1992
28. LaBanc JP: Classification of nerve injuries. Oral Maxillofac Surg
Clin North Am 4:285, 1992
29. Mackinnon SE, Dellon AL: Surgery of the Peripheral Nerve.
New York, NY, Thieme Medical Publishers, 1988
30. Seddon HJ: Three types of nerve injury. Brain 66:237, 1943
31. Sunderland S: A classification of peripheral nerve injuries pro-
duced by loss of function. Brain 74:491, 1951
32. Sunderland S: Nerve Injuries and Their Repair. A Critical Ap-
praisal. New York, NY, Churchill Livingstone, 1991, p 221
33. Ferdousi AM, MacGregor AJ: The response of the peripheral
branches of the trigeminal nerve to trauma. Int J Oral Surg
14:41, 1985
34. Robinson PP: Observations on the recovery of sensation fol-
lowing inferior alveolar nerve injuries. Br J Oral Maxillofac Surg
26:177, 1988