Pages Navigation Menu

Meet 'em, greet 'em, treat 'em and street 'em

SGEM#81: Sore Mouth (Lidocaine for Oral Ulcers)

SGEM#81: Sore Mouth (Lidocaine for Oral Ulcers)

Podcast Link: SGEM81
Date:  June 24th, 2014 

Guest Skeptics: Meghan Groth (@EMPharmGirl). Meghan is the emergency medicine pharmacy specialist at Fletcher Allen Health Care in Burlington, Vermont and an adjunct professor of pharmacy practice at the Albany College of Pharmacy and Health Sciences. Her professional interests include resuscitation and acute neurologic emergencies. In her free time, you can find her teaching Les Mills BodyPump classes at the local gym.

Case Scenario: Two year old comes to the ER with a rash on his hands and feet. It is associated with painful mouth ulcers. You diagnose the child with hand, foot and mouth disease. The parents are concerned about dehydration. You estimate the child to be mildly dehydrated. You know from SGEM#12 that ondansetron helps hydrate the child with vomiting but this is different.

Question: Does fluid intake improve when 2% viscous lidocaine is applied to oral ulcers?

Background: Children with infected mouth ulcers is a common emergency department presentation. Most of these can be easily diagnosed clinically. The cause is often viral infections like gingivostomatitis, ulcerative pharyngitis, herpangina and hand, foot and mouth disease.

Children usually present because of pain and decreased oral intact. Oligoanalgesia (poor pain management) is a big problem in emergency departments. Children represent a group that is less likely to receive adequate analgesia. (Brown et al, Selbst and Clark).

Goldman et al Pediatrics 2008 published a helpful article describing the degree of dehydration in children ranging from mild, moderate to severe.

Screen Shot 2012-11-15 at 4.45.16 PM

Dehydration can usually be treated effectively with oral rehydration. For more information on visit this site on Oral Rehydration Therapy. The Canadian Pediatric Society also has a algorithm for oral rehydration.

Article:Hooper et al. Topical lidocaine to improve oral intake in children with painful infectious mouth ulcers: a blinded, randomized, placebo-controlled trial. Ann Emerg Med 2014.

  • Population: Pediatric patients aged 6 months to 8 years (n=100) with the mouth ulcers (gingivostomatitis, ulcerative pharyngitis, herpangina or hand, foot and mouth disease) and decreased PO intake (parent/guardian “not drinking well” and <10ml/kg of body fluid in preceding 2 hours).
  • Intervention: A single oral dose of 2% lidocaine (weight based at 0.15 mL/kg), patients instructed to gargle and spit (if able) or swallow if not able to follow instructions
  • Comparison: Placebo arm administered a topical methylcellulose/cherry flavored solution in an identical fashion
  • Outcome: Amount of oral fluid ingested within 60 minutes after study drug administration (mL/kg)
    • Excluded: Hypersensitivity to lidocaine or other amide local anesthetics. Diseases in which elevated levels of lidocaine may be dangerous (epilepsy, impaired cardiac conduction, bradycardia, or impaired hepatic or renal function). Severe dehydration or toxic needing immediate resuscitation. Patients with >2 episodes of vomiting before ED arrival, dental disease, mouth trauma, or malignancy. Patients on cardiac or other drugs with possible interactions with lidocaine. Patients who already had >1 dose of topical anesthetic for the same illness. Pre-existing upper airway obstruction or swallowing difficulties, If they had received analgesia <1hr before enrollment. Non-English speaking.

Authors Conclusions: Viscous lidocaine is not superior to a flavored gel placebo in improving oral intake in otherwise healthy children with painful infectious mouth ulcers. It appears that staff coaching and possibly the coating effect of oral topical agents alone can increase oral intake.”

checklist-cartoonQuality Check List for Systematic Reviews:

  1. The study population included or focused on those in the ED. Yes
    • Comment: These were all pediatric ED patients presenting to the Royal Children;s Hospital in Melbourne, Australia
  2. The patients were adequately randomized. Yes 
    • Block randomization with block sizes of 2 and 4
  3. The randomization process was concealed. Unsure?
    • Numbered bottles by pharmacist independent of the study.
    • Patient given two bottles (A and B)
      • A- contained study drug or placebo
      • B- contained the alternative mixture (lidocaine or placebo)
    • After 60min the treating clinician could give the second bottle if clinically indicated. So all participants int he study had access to lidocaine.
  4. The patients were analyzed in the groups to which they were randomized. Yes 
  5. The study patients were recruited consecutively (i.e. no selection bias). No 
    • Convenience sample when research assistants or investigators were in the ED
  6. The patients in both groups were similar with respect to prognostic factors. Unsure 
  7. All participants (patients, clinicians, outcome assessors) were unaware of group allocation. Unsure? 
  8. All groups were treated equally except for the intervention. Yes 
  9. Follow-up was complete (i.e. at least 80% for both groups). Yes 
    • One patient left the ED before completion of the study
  10. All patient-important outcomes were considered. Yes
  11. The treatment effect was large enough and precise enough to be clinically significant. No

Key Results:  No difference in the amount of fluid intake at 60 minutes between groups. Lidocaine group 8.5 mL/kg (IQR 4.1-13.8 mL/kg) vs. placebo group 9.3 mL/kg (IQR 3.1-15.2 mL/kg). Difference in medians 0.8 mL/kg (95% CI -2.52 to 3.26 mL/kg)

  • Secondary outcomes: Similar oral intake at multiple time points measured, similar utility of adjunct analgesics. 40% of patients (n=20) in each group required “rescue” treatment. Bottle B 14% (n=7) in lidocaine group, 6% (n=3) in placebo group required admission for fluid administration via either NG or IV route. Longer term outcomes detailed in methods (adverse events, ED LOS) not reported in results section.
Meghan Groth

Meghan Groth

Comments: This is the first study to evaluate the efficacy of topical lidocaine for painful mouth ulcers in children. The randomization was adequate, but patients were enrolled as a convenience sample as research investigators were only present in the ED about 50 hours per week. Additionally, 40% of patients in each arm received the other “treatment” at 60 minutes.

There is concern about the randomization being concealed. The placebo may have looked, tasted and smelled like the lidocaine viscus but it would not have caused oral numbness. If the child starts talking funny and drooling parents/guardians may have figured it out which group their child was in. This information could have been passed along to the research staff even unintentionally.

If there was unblinding the bias would have probably been towards lidocaine treatment. Such that there was no difference between groups this potential unblinding actually strengthens the conclusion to accept the null hypothesis.

One way researchers could have tested for unblinding would have been to ask the parents/guardians and the researchers which group they felt the child was assigned. If they were able to tell more than random chance than randomization was not concealed.

Another issue with this paper was the original intent to report the primary outcome in terms of mean fluid intake at 60 minutes, with the comparison between groups as mean difference. An interim analysis revealed a skewed data set, thus the primary outcome was ultimately reported as a median with an interquartile range instead.

nerd glassesTalk Nerdy to Me? Mean and Median can both be used to describe the “centre” of a dataset. Which one to use depends on the symmetry of the dataset being measured.

A mean is simply the average. You add all the data points up and divide by the number of data points. However, if your data is skewed this can represent a problem. The mean can be highly influences by one or two outliers. It is only representative of the centre if the distribution of the data is symmetric. The mean can be affected by any single change to the data

As an example take the following dataset: 1, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 1000

Mean is 1046/11 = 95.1

The median denotes the value lying at the midpoint of the dataset. It is is not influenced by outlying measurements. So for asymmetrical data or skewed data the median might better represent the centre of the data. The median does not change with any single data point change. In the example given the median would be 5. Which do you think better represents the centre of the data?


Both groups may not have been equal with prognostic factors. More patients in the placebo group received topical analgesics in the 24 hours prior to being enrolled in the study. The authors did not report a p value, and did not comment on whether or not this was a significant difference. However, it has the potential to introduce bias as patients in the placebo group may have been more “comfortable” at baseline and thus more likely to increase their oral intake during the study period, a difference which would have been attributed to the intervention.

The authors comment that it may have been coaching from the medical staff and/or simply the coating of the ulcers with a liquid preparation that encouraged an increase in oral intake in this study. I think this is a reasonable conclusion. They did however, only study topical lidocaine by itself.

In US EDs, a “magic mouthwash” preparation is commonly prescribed for painful mouth ulcers, with includes diphenhydramine and an antacid along with lidocaine. Many patients in this cohort (50% in the lidocaine group, 44% in the placebo group) required adjunct analgesia; this appears to be a helpful treatment in this condition.

Although adverse events were not reported in this study, there have been serious adverse events associated with oral lidocaine use (see poison center data from Curtis LA, Dolan TS, Seibert HE. Are one or two dangerous? Lidocaine and topical anesthetic exposures in children. J Emerg Med 2009;37:32-39), especially in young children that may not be able to follow instructions to “spit” after gargling.

fda-warningBlack Box Warning FDA June 2014: “oral viscous lidocaine 2 percent solution should not be used to treat infants and children with teething pain.”

Author’s conclusion compared to our conclusion: This study calls into question the routine utility of topical lidocaine for painful mouth ulcers. Even though there are a few methodological flaws to the study, it seems as though the conclusion is reasonable.

The Bottom Line: Viscous lidocaine is not superior to a placebo gel in improving oral intake in children with painful infectious mouth ulcers.

Case Resolution: Educate the parents about the diagnosis of hand, foot and mouth disease. Reassure them it is a common viral illness in children < 5 years of age. It usually is mild in nature with a low grade fever and rash lasting about 1 week. There is no specific treatment but taking over the counter medication for pain and fever may help. Encourage your son to drink liquids is important to prevent dehydrated.

Clinical Application: Given the questionable efficacy over placebo demonstrated in this study, along with the potential to cause harm, it may be time to rethink the use of topical lidocaine for painful mouth ulcers in this population.

What do I tell my patients? There appears to be evidence that topical lidocaine is not better than a placebo for increasing oral intake (and preventing dehydration) in children with painful mouth ulcers. Supportive care with analgesic therapy, along with encouragement by parents and/or ED staff are useful adjuncts.

Keener Kontest: Last weeks winner was Femke Friesema from the Netherlands. Fremke knew the couple who discovered tranexamic acid was Shosuke and Utako Okamoto (husband and wife).

If you want to play the Keener Kontest this week then listen to the podcast for the question. Email me your answer at with “keener” in the subject line. The first person to correctly answer the question will receive a cool skeptical prize.

BEEM LogoUpcoming conferences:

  • QueBEEM Quebec City September 29th and 30th
  • EuroBEEM Prague  December 5th and 6th

Remember to be skeptical of anything you learn,

even if you heard it on the Skeptics’ Guide to Emergency Medicine.