Severe adverse reactions to dental local anaesthetics: prolonged mandibular and lingual nerve anaesthesia


  • PJ Sambrook,

    1. Oral and Maxillofacial Surgery Unit, The University of Adelaide, South Australia.
    2. RACDS representative, Editorial Advisory Panel, Australian Prescriber.
    Search for more papers by this author
  • AN Goss

    1. Oral and Maxillofacial Surgery Unit, The University of Adelaide, South Australia.
    2. Member, Dental Therapeutics Committee, Australian Dental Association Inc.
    Search for more papers by this author

Professor Alastair Goss
Oral and Maxillofacial Surgery Unit
Faculty of Health Sciences
The University of Adelaide
Adelaide SA 5005


Background:  Prolonged anaesthesia may occur following dental local anaesthetic blocks. This paper reviews the possible mechanisms of injury. Direct injury to the nerve by the needle, although commonly thought to be the mechanism, is unlikely. It is much more likely that the injury is from neurotoxicity and/or interference with the vascularization of the nerve.

Methods:  Estimation of the frequency of injury was complicated by the fact that although local anaesthetics are prescription-only (S4) drugs, they are supplied without prescription by dental supply houses. Unlike all other S4 drugs, there is no statutory requirement to record supply. The pharmaceutical and supply houses relied on that and ‘commercial confidentiality’ to not supply information.

Results:  An informed estimate of 1 in 27 415 was made but this figure has wide confidence limits. Management of cases of prolonged anaesthesia following local anaesthetic injection is discussed.

Conclusions:  Patients who suffer this uncommon complication suffer considerable distress and feel injured, so care must be exhibited in their management. Specialist referral is recommended.

Abbreviations and acronyms:

local anaesthetic


oral and maxillofacial surgeon


Injection of local anaesthetic (LA) solution is the most common application of a prescription or S4 medication in dentistry. Its use is so ubiquitous that many dentists do not consider it to be restricted medication but a material to be used like a restorative material, or a topical over-the-counter medication. This view is reinforced in that dentists do not prescribe local anaesthetics like they would write a prescription for an antibiotic or analgesic. Local anaesthetic cartridges are distributed to Australian dentists, along with other stores, by dental supply houses. There are several different types of LA agent with multiple brand names (Table 1).

Table 1.   Dental local anaesthetic agents used in Australia (all S4 drugs)
LignocainePlain or with adrenalineTrade names: lignocaine, neurocaine, xylocaine, lignospan
PrilocainePlain or with felypressin or adrenalineTrade name: citanest
ArticainePlain or with adrenalineTrade names: bucanest, deltazine, septanest
MepivocaineWith adrenalineTrade names: scandonest, septodont
BupivacainePlain or with adrenalineTrade name: Marcain

The aim of LA injection is to temporarily reduce the pain and other sensations associated with dental treatment. Usually, it is a safe and effective procedure and sensation returns to normal in a short period.1 However, there may be systemic adverse reactions ranging from briefly feeling unwell to total collapse, which is reviewed separately.2 A different group of adverse reactions to LA injection is when the loss of sensation is prolonged. This may be temporary, where after a few days, weeks or months, sensation returns, or it may be permanent. This mainly involves lingual or mandibular nerves, or both. Prolonged anaesthesia following infiltration of the maxillary nerves is rare and will not be discussed further in this paper.

An understanding of the anatomy and physiology of the mandibular and lingual nerves is essential to understanding LA-related prolonged anaesthesia (Fig 1).3 Microscopically, each nerve fibre or axon with its myelinated sheath is covered by an endoneural layer. This connective tissue cover contains a basal laminar, collagenous fibres and endoneural capillaries. There are 7000 to 12 000 axons in each mandibular or lingual nerve.4 The individual nerve fibres are grouped into fascicles which are surrounded by a perinuerium. This connective tissue layer supports, protects and provides nutrition to the fascicle. The mandibular nerve has a minimum of three fascicles at the lingualar, increasing to usually seven and sometimes more, within the mandibular canal.5 The lingual nerve may only have one fascicle at the lingualar, this increases up to seven at the third molar region.5 The individual fascicles are surrounded by a fibrous perinuerium which protects the nerve against functional and compressive forces. Surrounding the epineurium is a loose areolar connective tissue layer, the mesoneurium, which contains a segmental blood and lymphatic supply. The vessels penetrate into the epineurial nerve sheath to maintain nutrition and excretion of the nerve. Thus, the mandibular and lingual nerves are complex, well protected and independently well nourished.

Figure 1.

 Structure and vasculature of a nerve. Note the complex intraneuronal blood supply which includes lymphatic drainage. (Drawing by Dr C Swann, PhD. Medical illustrator, The University of Adelaide, based on the drawing in Lundburg.3)

Anatomically, the lingual nerve is at greater risk than the mandibular nerve from a block injection. In the standard, mouth-open position for a mandibular block, the lingual nerve is stretched quite tightly and is just a few millimetres beneath the mucosal surface. This makes it less likely to deflect during an injection.6 For this reason, lingual nerve injury is more common than mandibular nerve injury when caused by LA injections.6 Nerves, although resistant to injury, can be injured and these were classified by Sunderland over half a century ago (Table 2).7 It must be noted that this staging system was based on a histological study involving rabbits. Thus, apart from Stage 5, where there is complete separation of the whole nerve, the damage is internal so cannot be seen from direct inspection of the nerve. Seddon developed a clinical classification which can be related to the degree of nerve injury and also relates to the response to injury (Table 2).8 Neither of these classifications can be applied prospectively to individual human nerve injuries. Assessment of nerve injuries are essentially subjective, and made from the description of the sensations the patient experiences over time. The definitions of the various descriptions are presented in Table 3. It is important that these definitions are accurately and thoughtfully applied. It is the authors’ experience that dentists commonly call all abnormal sensations ‘paraesthesias’. This is incorrect and potentially misleading.

Table 2.   Nerve injury classificationsThumbnail image of
Table 3.   Definition of terms useful in describing altered neurological sensations
AnaesthesiaInsensitivity to all forms of stimulation
AnalgesiaAbsence of pain in response to stimulation that would normally be painful
HypoaesthesiaDiminished sensitivity to all forms of stimulation
HyperaesthesiaIncreased sensitivity to all forms of stimulation
ParaesthesiaUnusual abnormal but not painful spontaneous or evoked sensations (tingling or pricking sensations)
DysaesthesiaAny unpleasant abnormal sensation, either spontaneous or evoked, used to describe painful paraesthesia and burning neurogenic discomfort or pain
NeuropraxiaA condition in which a nerve remains in place after a severe injury although it no longer transmits impulses
AxonotmesisNerve injury characterized by disruption of the axon and myelin sheath but with preservation of the connective tissue fragments, resulting in degeneration of the axon distal to the injury site; regeneration of the axon is spontaneous and of good quality
NeurotmesisA peripheral nerve injury in which the nerve is completely disrupted by laceration or traction. It requires surgical approximation, with unpredictable recovery

The precise mechanisms of nerve injury from LA injection are controversial. The traditional concept is a direct injury by the needle traumatizing the nerve. This, however, is unlikely as the diameter of the mandibular nerve is up to 3 mm, and for lingual nerves up to 2 mm, but the diameter of the largest injection needle is less than 0.5 mm.9 In cadaver studies the injection needle easily transverses the nerve, usually between the fascicles. The potential danger of damage by the needle could be increased by the bevel being bent as it contacts the bone. This needle damage has been shown to occur in up to 80% of injections, with up to two-thirds of the bevels being hooked outwards.10 This could potentially transect multiple nerve fibres or even a fascicle when the needle is withdrawn. However, it would not feasibly transect the whole nerve. Usually, the whole dermatome is involved in prolonged anaesthesia following a LA block rather than a small segment which would occur from damage to individual axons or fascicles.

It has been suggested that an important clinical sign of nerve injury during a block is an electric-shock type sensation at the time of injection.9 It has also been suggested that injection injury is more likely if the area has already been anaesthetized. Thus, on a repeat injection, which directly hits the nerve, it would not result in an electric-shock sensation as the nerve is already partly anaesthetized.

An alternative hypothesis is that the needle damages the intraneural blood supply, resulting in a haematoma both within and without the nerve.11 Within the nerve, the resultant haematoma would compress nerve fibres within the epineurium. This occurs within 30 minutes and could cause compression of the nerve fibres. Dependent on the extent of breakdown products and scarring, this would determine the extent and length of sensory change. Haematoma formation outside the epineurium could compress the nerve and interfere with the segmental blood supply. If associated with direct needle damage to the medial pterygoid muscle, then this could also result in concurrent trismus.

The remaining concept is that the LA solution itself is neurotoxic to the nerve.12 This may be a direct chemical effect of the LA when deposited within the nerve fibres or indirectly by interfering with the blood supply to the nerve. Several studies have shown higher per use incidence of injury with prilocaine and articaine as compared to lidocaine.9,13,14 Both prilocaine (3%) and articaine (4%) are used in dentistry in a higher concentration than lidocaine (2%). Lidocaine, in a higher concentration than that usually used in dentistry, has also been shown to cause neurotoxic damage.15 The potential increased neurotoxicity of articaine was demonstrated by a sudden increase in nerve injuries after the introduction of articaine to Ontario, Canada and Denmark.13,14 The contention that articaine was neurotoxic was rejected, although there were no conflict of interest statements attached to that rebuttal.16

Despite the potential for nerve injury from LA injection, the overall incidence is low. The range of reported figures range from 1 in 26 762 to 1 in 160 57112 to 1 in 785 000.13 Approximately one-third of all LA-related adverse reactions result in dysaesthesia, which is higher than the reported range for direct surgical trauma to nerves.17 Thus, although rare, LA block damage does place a significant morbidity on the patient.

This paper reviews the literature regarding nerve injuries, presents recent data from South Australia and discusses the management of LA-related nerve injuries.


The total number of LA cartridges used in Australia was estimated by two methods. The first was to use published figures from the international literature and to correct it for the population and the year when the numbers were published. Secondly, all the pharmaceutical companies importing LA cartridges from their overseas parent companies and the dental supply companies who distribute LA cartridges to the profession in Australia were approached. Wide informal discussions were held with interested parties on current LA usage.

The number of prolonged anaesthesia cases from LA cases in 2009 was obtained by a questionnaire to all the practising oral and maxillofacial surgeons in South Australia.

The detailed records of one of the author’s (ANG) cases for 2009 were examined. This included the patient demographics, medical comorbidities, type of block, type and dose of LA agents, symptoms over time, results of objective testing, other investigations, management and outcome was recorded.


Over 200 million LA cartridges were reported to have been used in the United States in a 1990s publication.1 Over 70 million LA cartridges were reported to have been issued in the United Kingdom in a 1980s publication.18 Correction for population and time would indicate that approximately 11 million LA cartridges were issued in Australia in 2009. The estimated usage is presented in Table 4.

Table 4.   Calculated usage of local anaesthetic cartridges in Australia in 2009Thumbnail image of

None of the pharmaceutical companies provided any data. One replied indicating that they were not under statutory obligation to provide numbers but did advise that dental local anaesthetics were supplied to the profession by the dental supply companies. Extended email and telephone conversations were held with the dental supply houses who indicated they were under no statutory obligation to provide data and declined to do so on the grounds of ‘commercial confidentiality’.

All 13 registered specialist oral and maxillofacial surgeons (OMS) in South Australia responded, detailing a total of 15 cases of prolonged anaesthesia following LA injections. Cases of prolonged anaesthesia secondary to surgery were excluded. Of these, eight were seen by one OMS (ANG) with a known interest in pain and medico-legal issues. The results of these cases are reported in Table 5.

Table 5.   Peripheral abnormality following local anaesthetic. One author’s referral experience in 2009
CaseAgeGenderMedical stateInjection typeResponseActionOutcome
154MWell(L) Md blockPain intesiveCT6/52 numb dentition
2 Lignospan SpecialNumbness – lip, lip of face, trismus 30 mmVit B12less area
Spatula exercisesopen to 40mm resolving
240FWell(R) Mand block(R) lingual paraesthesiaVit B12at 5/52 resolved
4 carpulestaste normal
2 blocks lignocaine
370MHypertension(L) Md blocknumbness – (L) lipSpatulaResolved at 3/12
ControlledlignocaineTrismus 14 mmexercises
477FWell(R) Mand blocknumbness – (R) lipCTResolved at 3/12
lignocaineTrismus 28 mmSpatula
561MType I diabetes(L) Mand Blocknumb (L) lip.Observe lower left 6Resolved at 3/12
No neuropathyImmediate onset and burning painDermatonetooth extracted Infiltration
661FFit and well(L) mand block
Immediate numbness and pain

persistent numbness
Vit B12Return of taste and sensationResolved in 3/12
 loss of taste (L) lip and tongue 
739FThyroxine(R) Mand blockBurning pain at 2 years (R) tongueTongue looks normalNormal sensation area burns at 24/12
Bipolar lignocaine
864MFit and well(L) Mand block
Numb 24 hrslip and tongue loss of taste and burning some ulceratiionVit B1218/12 ongoing burning sensationno objective numbness
Aggressive letter from Insurers
Summary56.44M2 Thyroid5 (L)4 lip4 Vit B126 resolving
39–774F1 Diabetes3 (R)2 tongue3 CT2 ongoing
5 Fit and WellAll lignocaine2 lip and tongue
1 re-injection3 Trismus

The frequency of prolonged anaesthesia was established from 15 cases in one year in South Australia which is 8% of the Australian population. Assuming that half of the total cartridges are used for blocks, then this gives a frequency of 1 in 27 415.


This study shows that occasionally patients may suffer from prolonged effects of a LA injection for restorative treatment. This study shows a rate of approximately 1 in 27 415, consistent with the higher risk figures which have been previously published. A frequency of 1 in 750 000 was reported in 1995, although the authors felt this was an underestimate.5 Later, the same authors reported a much higher incidence of 1 in 26 762 to 1 in 160 571. In another study, involving students and junior staff, an incidence of prolonged lingual anaesthesia was found in 1 in 12 104. The same group had 126 cases in 12 104 of trismus, swelling and prolonged pain.19 A recent study based on voluntary reports in Ontario, Canada showed different incidence figures for different types of LA agents with 1 in 332 000 for prilocaine, 1 in 410 000 for articaine and 1 in 2 580 000 for lignocaine.20 The main deficiency of this study’s calculations of the incidence of prolonged anaesthetics was the refusal of the pharmaceutical companies and the supply houses to provide data on the quantity and type of local anaesthetics supplied to the profession. It must be noted that the quantity of all other S4 drugs in Australia, both medical and dental, is available.21 Thus, the figures presented in this paper are informed estimates only. They are also probably an underestimate as not all cases would be referred to an OMS.

When the adverse reaction of prolonged anaesthesia occurs, invariably the patient feels they have been wounded. This ranges from minor regret to considering medico-legal action. Usually, the patient feels that the cost of any referral or treatment should be borne by the practitioner who gave the block. It is also easy for the practitioner to further inflame the situation by unguarded comments or promises that it will get better soon.

Once a patient reports to the treating dentist that they are still suffering from anaesthesia following a LA injection, the patient must be examined for the distribution of the sensory loss and other signs. The patient should be regularly followed up until resolution. It is sound defensive practice to offer referral to a specialist with an interest and experience in nerve injuries.

In the author’s (ANG) group of cases, six were promptly referred and two after many months of complaint. All had been on antibiotics and analgesics without benefit. In this one-year period, all cases of prolonged anaesthesia related to lignocaine. In the previous and subsequent years, cases involving prilocaine and articaine were also seen. When first seen, a detailed history was recorded, both of the circumstances of the injection and of any medical comorbidities. Other neuralgic conditions, such as multiple sclerosis, were excluded. The area of sensory change was objectively mapped. Patients with other signs, particularly trismus, had a CT scan to exclude other pathology but none was found in this series. The nature of LA injury was outlined with the level of discussion varied dependent on the patient’s interest and comprehension. All were told that nerve injuries healed slowly, that there was up to two years over which recovery could occur and the sooner sensation returned, the better but complete recovery was unlikely after six months.

Patients with trismus were shown tongue spatula exercises to improve jaw opening. Multivitamin B, including B12, was recommended. This was on the basis that B vitamins are involved in nerve healing and function.22,23 They are also water soluble, so any excess was excreted. This made patients feel that they were having some active treatment. Patients were monitored at three-monthly intervals until resolution. Although all the patients had been given antibiotics, nerve injury is not an infection.

Six of the eight patients had resolution by three months. This is consistent with a neuropraxic-type injury. Two patients had persistent altered sensation. As time went by the numbness changed to a dysaesthetic burning-type sensation. Objectively, they had some return of sensation. These two patients were offered antineuralgic medication, carbamezepine or gabapentin, but declined on realizing that they would need to take the medication daily for an extended period and there were side effects. Analgesics are of minimal value for neuralgic pain. None of the nerve injuries had any indication for surgery.

One of the patients with prolonged burning sensation had his situation exacerbated. He was an intelligent retired man with a legal background. He had asked the dental practice if there was anything that could be done to help him. This request was forwarded to the practice insurer who responded with an aggressive letter denying all responsibility and that they would take him to court if he wrote again. At this stage the patient was seen by the author. The situation was mediated on the basis that undoubtedly the adverse reaction of prolonged anaesthesia followed the LA injection and also the patient was only seeking assistance with his further dental expenses, not compensation for injury. The matter was then simply settled, although he still has persistent symptoms. The matter could have been handled much better if in the first instance the nature of the request had been properly and sensitively explored.

All dentists need to be aware that prolonged anaesthesia may occur following LA blocks for restorative purposes. This risk of prolonged anaesthesia is not usually discussed during consent for treatment. However, if it occurs it should be documented early, factual information contained in this review and management advice given. Denial of responsibility or promises that it always quickly resolves should not be given. Early specialist referral is recommended.


The authors wish to thank the constructive suggestions of members of the Australian Dental Association’s Dental Therapeutics Committee.