Systematic Review Efficacy of adjuvant ozone therapy in reducing postsurgical complications following impacted mandibular third-molar surgery A systematic review and meta-analysis

Abstract

Background
Results from several randomized controlled trials have shown a beneficial effect of ozone in reducing postsurgical complications after impacted mandibular third-molar surgery, but the literature is lacking a systematic review and meta-analysis.

Methods
The authors conducted this systematic review according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines after exclusion and inclusion criteria were applied and the following outcome parameters were evaluated: pain, swelling, trismus, quality of life, number of analgesics consumed, and adverse events. RevMan Cochrane Collaboration software, Version 5.3, was used to perform meta-analysis and the Grading of Recommendation Assessment, Development and Evaluation approach was used to rate the certainty of evidence.

Results
Patients who underwent adjuvant ozone application reported lower pain scores than patients in the control group at 24 hours after surgery (95% CI, –3.94 to –1.56) and at 7 days (95% CI, –1.67 to –0.78). Pooled analysis of all 4 included trials revealed a standardized mean difference (SMD) in swelling of –0.44 at 24 hours, 0.63 at 72 hours, and –0.87 at 7 days after surgery in the experimental group. Higher mean estimates in mouth opening were experienced by patients who received ozone at 24 hours (SMD, 2.74; 95% CI, –1.93 to 7.41; 4 studies, 133 patients), 72 hours (SMD, 2.77; 95% CI, –0.63 to 6.17; 4 studies, 133 patients), and 7 days after surgery (SMD, 1.42 SMD; 95% CI, –1.34 to 4.18; 4 studies, 133 patients).

Practical Implications
Evidence suggests that adjuvant ozone application can offer some benefit for reducing pain, improving quality of life, and decreasing mean intake of analgesics after impacted mandibular third-molar surgery, but it is not effective in reducing facial swelling and trismus, which paves the way for future research.

Full text

Systematic Review
Efficacy of adjuvant ozone therapy in
reducing postsurgical complications following
impacted mandibular third-molar surgery
A systematic review and meta-analysis
Kirti Chaudhry, MDS; Neeti Rustagi, MD; Rishi Bali, MDS; Shruti Khatana, MDS;
Shailendra Kumar, BDS; Amanjot Kaur, MDS; Pravin Kumar, MDS
ABSTRACT
Background. Results from several randomized controlled trials have shown a beneficial effect of
ozone in reducing postsurgical complications after impacted mandibular third-molar surgery, but the
literature is lacking a systematic review and meta-analysis.
Methods. The authors conducted this systematic review according to the Preferred Reporting
Items for Systematic Reviews and Meta-Analyses guidelines after exclusion and inclusion criteria
were applied and the following outcome parameters were evaluated: pain, swelling, trismus, quality
of life, number of analgesics consumed, and adverse events. RevMan Cochrane Collaboration
software, Version 5.3, was used to perform meta-analysis and the Grading of Recommendation
Assessment, Development and Evaluation approach was used to rate the certainty of evidence.
Results. Patients who underwent adjuvant ozone application reported lower pain scores than
patients in the control group at 24 hours after surgery (95% CI, e3.94 to e1.56) and at 7 days (95%
CI, e1.67 to e0.78). Pooled analysis of all 4 included trials revealed a standardized mean difference
(SMD) in swelling of e0.44 at 24 hours, 0.63 at 72 hours, and e0.87 at 7 days after surgery in the
experimental group. Higher mean estimates in mouth opening were experienced by patients who
received ozone at 24 hours (SMD, 2.74; 95% CI, e1.93 to 7.41; 4 studies, 133 patients), 72 hours
(SMD, 2.77; 95% CI, e0.63 to 6.17; 4 studies, 133 patients), and 7 days after surgery (SMD, 1.42
SMD; 95% CI, e1.34 to 4.18; 4 studies, 133 patients).
Practical Implications. Evidence suggests that adjuvant ozone application can offer some benefit
for reducing pain, improving quality of life, and decreasing mean intake of analgesics after impacted
mandibular third-molar surgery, but it is not effective in reducing facial swelling and trismus, which
paves the way for future research.
Key Words. Impacted mandibular third molar; systematic review; ozone; pain; swelling; trismus.
JADA 2021:152(10):842-854
https://doi.org/10.1016/j.adaj.2021.05.006
Removal of impacted mandibular third molars (IMTM) is one of the most common surgical
procedures performed in dental clinics and is associated with postsurgical complications
(PSCs), such as pain, swelling, and trismus.
1
PSCs can have a negative impact on patients’
quality of life due to increased workplace absence, follow-up appointments, expenses, and prolonged
medication use to address them.
2,3
Various pharmacologic and nonpharmacologic techniques have been developed to manage
PSCs.
4,5
The use of local or systemic corticosteroids and nonsteroidal anti-inflammatory drugs is
often recommended to reduce pain, swelling, and trismus.
6-8
However, their limited success and
associated adverse effects have led clinicians to search for nonpharmacologic adjunctive therapies,
for example, low-level laser therapy, ozone therapy, cryotherapy, pulsed electromagnetic energy
therapy, and acupuncture.
9,10
Ozone has a positive effect on the metabolism of various cellular components in tissue, promotes
hemostasis, inhibits bacterial growth, and increases supply of oxygen to local surgical sites.
11,12
It has
This article has an
accompanying online
continuing education
activity available at:
http://jada.ada.org/ce/home.
Copyright ª2021
American Dental
Association. All rights
reserved.
842 JADA 152(10) nhttp://jada.ada.org nOctober 2021

been shown to have beneficial effects on the healing and prevention of alveolar osteitis.
13
Results
from numerous randomized controlled trials (RCTs) have suggested a positive effect of ozone as an
adjunctive therapy in reducing PSCs, but consensus is still lacking.
14-17
The aim of our systematic review and meta-analysis was to compare the efficacy of ozone therapy
with a placebo or control in reducing PSCs associated with the surgical extraction of IMTMs.
METHODS
We performed this review in accordance to the latest Preferred Reporting Items for Systematic
Reviews and Meta-Analyses statement,
18
and the flow diagram is shown in Figure 1.
Literature search
Two authors independently searched MEDLINE, Cochrane, and Google Scholar using the key
words mandibular,third molar,extraction,impaction,removal,wisdom tooth,adjunctive therapy,and
ozone (Appendix; available online at the end of this article). We also searched for the registered
trials on trial registries and online databases of oral and maxillofacial surgery journals. The gray
literature was searched by means of reviewing the first 100 hits on Google Scholar. This literature
search was performed from February 1 through February 28, 2020, and was updated in May 2020.
The references of relevant articles were also examined to identify eligible studies.
Eligibility criteria
RCTs and quasi-RCTs comparing the efficacy of adjuvant ozone therapy with a placebo for reducing
the severity of PSCs after IMTM surgical extraction with a minimum follow-up of 72 hours with
standardized outcomes were included. The intervention of interest was application of any regimen
of ozone after IMTM surgical extraction because there is no standardized ozone therapy regimen.
Trials reporting the results of any of the outcomes (that is, pain, swelling, trismus, and adverse
effects) of interest were included. We placed no restrictions on patient characteristics, molar
impaction types, types of ozone regimen used, or the methods used to measure the outcomes of
interest.
Study selection and data extraction
Two reviewers (S. Khatana, S. Kumar) worked independently and screened the titles and abstracts
of identified studies for possible inclusion in the study and then screened full texts for potential
eligibility. Any disagreement was resolved by means of discussion with the third author (R.B.). The
other 2 reviewers (A.K, K.C.) extracted data from the included studies independently using the
standard data collection form 2013 08 12 of the Cochrane Collaboration.
19
If any study had more
than 1 intervention arm, then only the ozone intervention arm was included.
Reviewers abstracted the characteristics of included studies, depicting the study and population
characteristics, outcomes assessed (ozone application methods, pre- and postoperative
cointerventions, pain, swelling, trismus, quality of life, additional analgesic use, adverse effects), and
their timelines. Outcomes reported at 24 hours, 72 hours, and 7 days postoperatively were selected
for analysis.
Summary measures and data synthesis
For the outcome of pain, the mean of patient-reported postoperative pain scores in the intervention
and control group was used. Pain was measured in all 4 trials included
14-17
using a visual analog
scale. All of the authors used vernier calipers to measure interincisal mouth opening.
14-17
For
trismus outcomes, the mean change from baseline was assessed in the intervention and control
groups.
The investigators measured swelling from different landmarks on the face instead of a single
measurement and means and standard deviations were pooled to obtain a single swelling mea-
surement. Variability was observed across trials in the methods and devices used to measure the
swelling, and investigators used different statistics to report it. For assessing swelling, we used the
standardized mean difference (SMD) and its 95% CI to pool the results. For studies in which
the authors did not report data in a way that allowed pooling in meta-analysis, the results were
summarized narratively.
ABBREVIATION KEY
GRADE: Grading of
Recommendations
Assessment,
Development and
Evaluation.
IMTM: Impacted
mandibular third
molar.
NA: Not applicable.
PSC: Postsurgical
complication.
RCT: Randomized
controlled trial.
JADA 152(10) nhttp://jada.ada.org nOctober 2021 843

Data analysis
RevMan Cochrane Collaboration software (Version 5.3) was used to perform meta-analysis, and the
results were displayed via forest plots. Statistical significance was considered as P<.05. Statistical
heterogeneity was defined as P<.1 and I
2
>50%. The generic inverse variance method was chosen
a priori to pool the results using random-effects model because of clinical variability in the inter-
vention and trial designs to calculate the mean estimate of effects of ozone at 3 times for each
outcome of interest. Publication bias was not assessed quantitatively owing to the small number of
studies.
Sensitivity and subgroup analysis
Sensitivity analysis was done assuming no correlation between intervention and control arms in
split-mouth arms. This assumption that split-mouth trials and parallel trials are the samedwhich
provides the most conservative estimate of treatment effectdwas also adopted by Dawdy and
colleagues.
20
No subgroup analysis was done for the pooled outcomes.
Assessing risk of bias of included studies
Two reviewers (K.C., N.R.) assessed risk of bias of the included studies independently using the
Cochrane Handbook for Systematic Reviews of Interventions for guidance.
21
The risk of bias was
designated as high or low in the following areas: random sequence generation, allocation
concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete
outcome data, selective outcome reporting, and other bias. The third reviewer (R.B.) resolved any
disagreement.
Records excluded
(on the basis of title n = 144, on
the basis of abstract n = 5)
Full-text articles excluded,
with reasons
(n = 0)
Records after duplicates removed
(n = 153)
Records screened
(n = 153)
Full-text articles assessed
for eligibility
(n = 4)
Studies included in
qualitative synthesis
(n = 4)
Studies included in
quantitative synthesis
(meta-analysis)
(n = 4)
IdenticationScreeningEligibilityIncluded
No. of records identied through database
searching (MEDLINE = 143; Clinical Trials
Registry-India = 0; Google Scholar = 4;
Cochrane = 12; American Trial Register = 0;
World Health Organization International Clinical
Trials Registry = 0; Medical Research Council = 0;
National Health Service Trial Registers = 0)
Additional records identied through other
sources (online databases of oral and maxillofacial
surgery journals) (International Journal of Oral
and Maxillofacial Surgery = 2; Journal of Oral and
Maxillofacial Surgery = 7; Journal of
Cranio-Maxillofacial Surgery = 0; British Journal of
Oral and Maxillofacial Surgery = 0; Journal of
Maxillofacial Oral Surgery = 0)
Figure 1. Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow chart.
18
844 JADA 152(10) nhttp://jada.ada.org nOctober 2021

Dealing with missing data
All included trials were assessed for missing data, and the authors of relevant studies were contacted
to request access to the missing data. In instances for which there were randomly missing data, we
used the reported data.
Certainty of the evidence assessment
We used the Grading of Recommendations Assessment, Development and Evaluation (GRADE)
approach (eTable, available online at the end of this article) to assess the certainty of the estimates
of effect for each outcome.
22
Evidence Prime software (GRADEPro) was used to summarize key
information for all outcomes.
The potential limitations to ascertain the quality of evidence were considered in light of risk of
bias, inconsistency, imprecision, indirectness of evidence, and publication bias.
23-26
Imprecision was examined using the clinical thresholds that aimed to ascertain the magnitude of
effects considered as clinically relevant for each outcome. We determined similar thresholds for
outcomes assessed, as discussed by Dawdy and colleagues.
20
In the maxillofacial field, the minimal
clinically important difference
27
has not been determined for pain, swelling, or trismus.
RESULTS
Description of the included studies
The data extracted from the included studies
14-17
are summarized in Tables 1 through 3. A total of
133 records were screened and 4 RCTs published between 2013 and 2020 met the eligibility
criteria.
14-17
Three studies
15-17
had a split-mouth design (with crossover window being 2, 3, and 4
weeks, respectively), and 1 study had a parallel design.
14
Kazancioglu and colleagues
14,15
conducted
1 split-mouth design study and 1 parallel trial. Both studies followed similar protocol for experi-
mental and comparison groups and cointerventions administered. Variability was observed between
the other 2 included studies
16,17
for protocol followed for experimental group and the control group
and cointerventions administered by investigators to both group of participants (Table 2).
Variable modes of ozone application were used in the studies identified for detailed review. Two
studies used an extraoral ozone probe,
14,15
another used ozonated water irrigation,
17
and another
used topical ozone gel application.
16
Because no standard application for ozone is recommended in
the literature, all modes of ozone application were included in this review.
Characteristics of patients included
All 4 trials included patients with IMTMs with similar surgical difficulty index according to Pell and
Gregory classification.
14-17
The operative time was reportedly lower in Glória and colleagues
17
(15
minutes) for both experimental and control groups. Kazancioglu and colleagues
14,15
reported it was
approximately 22 minutes for the experimental group and 25 minutes for the control group in both
studies, and Sivalingam and colleagues
16
reported longer surgery time for the experimental group
(mean [standard deviation], 22 [14] minutes) than the control group (mean [standard deviation], 20
[12] minutes). In all of the studies, age and sex distribution, IMTM difficulty index, and mean
surgical time were comparable in both arms, indicating that patient factors had minimum con-
founding effect on the outcomes assessed.
Risk of bias
A tool from the Cochrane Collaboration
21
was used to assess the risk of bias of each study, as shown
in Figure 2. Low-quality reporting of methodological details influenced the risk-of-bias assessment.
The domains that lacked sufficient information to suggest that authors took precautionary measures
to reduce biases were judged as being high risk. High risk of selection bias was considered if unclear
or no descriptions of the methods were provided regarding random sequence generation and allo-
cation concealment of participants in the included trials. Lower risk of bias was observed for do-
mains regarding blinding of participants and personnel and blinding of outcomes assessment.
The studies in which participants were not blinded to the intervention through use of pseudo-
ozone generator
14
or topical application,
16
the subjective outcomes assessment (that is, pain and
quality of life) could have been affected because participants might have been aware of receiving the
ozone intervention. Low risk of attrition bias was observed among the trials due to the short-term
JADA 152(10) nhttp://jada.ada.org nOctober 2021 845

nature of the procedure and need for postoperative follow-up of included participants. The selective
outcome reporting bias was ascertained to be low as in all the included trials, investigators reported
the outcomes data with dispersion, which could be pooled for meta-analysis. Except for Glória and
colleagues,
17
who published their protocol in a US trial registry (ClinicalTrials.gov), none of the
study protocols were available in the public domain.
Effects of ozone
The meta-analysis was done for all outcomes at 24 hours, 72 hours, and 7 days after surgery. Tables 1
through 3 present summaries of the studies included, and final effect estimate and GRADE certainty
of evidence ratings for each outcome at each time are provided in the eTable (available online at
the end of this article).
Effects of Ozone on Pain
Investigators with all 4 trials
14-17
reported pain at 24 hours, 72 hours, and 7 days after surgery, with
the exception of Glória and colleagues,
17
who did not report finding at 7 days (Figure 3). Patients
who underwent adjuvant ozone application reported lower pain scores than patients in the control
group at 24 hours after surgery (e2.75 points on a 10-point scale in which 0 represents no pain and
10 represents unbearable pain), 95% CI, e3.94 to e1.56 points; 4 studies, 133 patients), at 72 hours
after surgery (e2.93 points; 95% CI, e3.77 to e2.08 points; 4 studies, 133 patients), and 7 days
after surgery (e1.22 points; 95% CI, e1.67 to e0.78 points; 3 studies, 113 patients) (Figure 3).
Sensitivity analysis revealed that the mean difference (MD) in pain scores between the ozone and
control arms was higher among patients recruited in the parallel trial than those recruited in the
split-mouth trial at 24 hours and 72 hours, but not after 7 days.
Effects of Ozone on Swelling
All 4 trials
14-17
were included to calculate mean estimate of swelling at all 3 times studied. Figure 4
shows that pooled analysis of all 4 trials revealed the SMD in swelling was e0.44 (95% CI, e2.61 to
Table 1. Characteristics of the included studies.
14-17
PARTCIPANTS, NO.
SURGERY TIME, MIN, MEAN
(STANDARD DEVIATION)
STUDY DESIGN
POPULATION
CHARACTERISTICS
AGE, Y, MEAN
(STANDARD
DEVIATION),
RANGE
MALE/
FEMALE,
NO. Intervention Comparison Intervention Comparison
Kazancioglu and
Colleagues,
14
2014
RCT*
Parallel, 3 arms
Group 1: 20
patients, low-level
laser therapy
Group 2: 20
patients, ozone
Group 3: 20
control patients
Impacted IIIb
surgical difficulty
grade
22.6 (2.3),
18-25
32/28 20 20 22 (9) for
ozonated
25 (11) for
control
23 (8) for low-
level laser
terapy
excluded
Kazancioglu and
Colleagues,
15
2014
RCT
Split-mouth,
crossover window
period of 2 wk
Impacted IIIb
surgical difficulty
grade
22.6 (2.3),
18-25
32/28 60 60 22 (9) for
ozonated
25 (11) for
control
Sivalingam
and
Colleagues,
16
2016
RCT
Split-mouth,
crossover
window period
of 3 wk
Similar difficulty
index as assessed by
grading system
25.6 (4.4) 17/16 33 33 22 (14) for
ozonated
20 (12) for
control
Glória and
Colleagues,
17
2020
RCT
Split-mouth,
crossover
window period
of 4 wk
Bilateral impacted
third molar with
class II position, type
B impaction
20.9 (Not
available)
8/12 20 20 15.65 (6.94) for
ozonated
15.90 (5.56)
for control
* RCT: Randomized controlled trial.
846 JADA 152(10) nhttp://jada.ada.org nOctober 2021

1.73; 4 studies, 133 patients) at 24 hours after surgery; 0.63 (95% CI, e1.42 to 2.69; 4 studies, 133
patients) at 72 hours after surgery; and e0.87 (95% CI, e2.31 to 0.56; 4 studies, 133 patients) at 7
days after surgery in the experimental group.
Sensitivity analysis revealed that the SMD in swelling between both arms was lower among
patients recruited in split-mouth trials than in those recruited in parallel trials at 24 hours, 72 hours,
and 7 days, but none of these were statistically significant (Figure 4).
Effects of Ozone on Mouth Opening or Trismus
All 4 trials
14-17
were included in the meta-analysis to study the effect of adjuvant ozone on trismus
compared with patients in the control arm for all 3 times studied (Figure 5). Higher mean estimates
were experienced by patients who received ozone at 24 hours after surgery (SMD, 2.74; 95% CI,
e1.93 to 7.41; 4 studies, 133 patients), 72 hours after surgery (SMD, 2.77; 95% CI, e0.63 to 6.17; 4
studies, 133 patients), and 7 days after surgery (SMD, 1.42; 95% CI, e1.34 to 4.18; 4 studies; 133
patients).
Sensitivity analysis revealed that MD in mouth opening between both arms was higher among
patients recruited in split-mouth trials than in those recruited in the parallel trials at 24 hours, 72
hours, and 7 days, but none of these were statistically significant (eTable, available online at the
end of this article).
Effects of Ozone on Quality of Life
Only 2 trials
14,15
used the 14-item Oral Health Impact Profile to measure the quality of life of
included participants, the score of which ranges from 0 through 56, with higher score indicating
poor quality of life. Pooled analysis revealed that patients undergoing ozone treatment experienced
Table 2. Summary of the studies included.
TREATMENT GIVEN
STUDY
RANDOMIZATION
METHOD Intervention Group Comparison Group
OZONE APPLICATION
PROTOCOL
COINTERVENTION
THERAPY COMMON
TO BOTH GROUPS
Kazancioglu and
Colleagues,
14
2014
Not reported Ozone generator with high-
frequency, 7.5-cm, deep-
tissue probe
Group 1: Gallium aluminium
arsenide diode laser, 808
nm applied using handpiece
at 100 mW for a total of
120 s. Received 12 J low-
level laser therapy at the
insertion point of masseter
immediately after surgery
and at postoperative 24
hours, 72 hours, and 7 days
Group 2: Control
Applied extraorally at the
insertion point of masseter
immediately after surgery
and at postoperative 24
hours, 72 hours, and 7 days
with an intensity of 80% for
10 s
1,000 mg of amoxicillin
and 550 mg of naproxen
sodium orally and 0.2%
chlorhexidine mouthrinse
(1 min, 3 times daily)
for 1 wk
An ice pack was applied to
the surgery area for at least
30 min
Kazancioglu and
Colleagues,
15
2014
Split-mouth design
Envelopes
Ozone generator with high-
frequency, 7.5-cm, deep-
tissue probe
Sham ozone therapy Applied extraorally at the
insertion point of masseter
immediately after surgery
and at postoperative 24
hours, 72 hours, and 7 days,
with an intensity of 80% for
10 s
1,000 mg of amoxicillin
and 550 mg of naproxen
sodium orally and 0.2%
chlorhexidine mouthrinses
(1 min, 3 times daily)
for 1 wk
An ice pack was applied to
the surgery area for at least
30 min
Sivalingam and
Colleagues,
16
2016
Split-mouth design
Lottery method
Topical ozone, no systemic
antibiotics
Gauze dressing, systemic
antibiotics 500-mg
amoxicillin capsule and 400
mg of metronidazole (Flagyl)
3 times/d for 5 d
Ozone gel filled the entire
socket and as a smear on
the vertical limb of the
incision after careful
isolation of the surgical site
using a scoop for 2 min, 2
times/d for 3 d
Postoperative analgesics
(400-mg ibuprofen tablet,
333-mg paracetamol
tablet) for 2 d, 3 times/d
Glória and
Colleagues,
17
2020
Split-mouth design
Opaque and sealed
envelopes
Ozonized, double-distilled
water as irrigating solution
Double-distilled water as
irrigating solution
Ozone double-distilled
water used as constant
irrigation using a 10-cc glass
hypodermic syringe coupled
to a 10-mm steel needle
twice per day and after
surgery to irrigate the socket
1 500 mg of sodium
dipyrone tablet every 6 h
and 1 100 mg-nimesulide
tablet every 12 h for 3 d
JADA 152(10) nhttp://jada.ada.org nOctober 2021 847

Table 3. Outline of outcomes assessment.
STUDY PAIN SWELLING TRISMUS QUALITY OF LIFE ANALGESICS USE
Kazancioglu and Colleagues,
14
2014
Method Yes*Yes
†‡
Yes (maximum interincisal distance
between 2 opposite right central
incisors via vernier caliper)
Yes (OHIP-14
§
) Yes (count)
Time, d 24 hours, 72 hours,
and 7 days
0,
{
24 hours, 72 hours,
and 7 days
0, 24 hours, 72 hours,
and 7 days
0, 24 hours, 72 hours,
and 7 days
NA
#
Kazancioglu and Colleagues,
15
2014
Method Yes*Yes
†‡
Yes (maximum interincisal distance
between 2 opposite right central
incisors via vernier calipers)
Yes (OHIP-14) Yes (count)
Time, d 1, 3, 5, 7 0, 1, 3, 5, 7 0, 1, 3, 5, 7 0, 1, 3, 5, 7 NA
Sivalingam and Colleagues,
16
2016
Method Yes*Yes
†‡
** Yes (interincisal mouth opening
via vernier caliper)
No Yes (count)
Time, d 24 hours, 72 hours,
and 7 days
24 hours, 72 hours,
and 7 days
0, 24 hours, 72 hours,
and 7 days
NA 3-5
Glória and Colleagues,
17
2020
Method Yes*Yes
†‡
** Yes (interincisal mouth opening
via digital caliper)
No No
Time, d 1, 2, 3 0, 1, 2, 3, 7 0, 1, 2, 3, 7 NA NA
* Visual analog scale (0-10). †Tragus to corner of mouth. ‡Tragus to pogonion. § OHIP-14: 14-Item Oral Health Impact Profile. {Day 0: Preoperative. # NA: Not
applicable. ** Lateral canthus to angle of mandible.
Random sequence generation (selection bias)
Allocation concealment (selection bias)
Blinding of participants and personnel (performance bias)
Blinding of outcome assessment (detection bias)
Incomplete outcome data (attrition bias)
Selective reporting (reporting bias)
Other bias
Glória and Colleagues,17 2020 ++++++ +
Kazancioglu and Colleagues,14 2014 +++ +– – –
Kazancioglu and Colleagues,15 2014 +++++– –
Sivalingam and Colleagues,16 2016 +++–+ +–
Study
Domain
Figure 2. Risk-of-bias summary based on review authors’judgments about each risk of bias for the included studies.
Green: Low risk of bias. Red: High risk of bias.
848 JADA 152(10) nhttp://jada.ada.org nOctober 2021

better quality of life than those in the control group, with MD ranging from 6 through 11 at the 3
times studied (Figure 6). Lower estimates of effect were observed at 24 hours (MD, e11.24; 95% CI,
e12.73 to e9.75; 2 studies, 80 patients), 72 hours (MD, e11.08; 95% CI, e12.08 to e10.08; 2
studies, 80 patients), and 7 days (MD, e6.85; 95% CI, e7.68 to e6.02; 2 studies, 80 patients) in
the intervention group. Results for sensitivity analysis were the same as the pooled estimates.
Heterogeneity: τ2 = 1.40; χ2
2 = 6.84, (P = .03); I2 = 71%
Test for overall effect: z = 2.80 (P = .005)
Heterogeneity: τ2 = 0.89; χ2
3 = 11.79, (P = .008); I2 = 75%
Test for overall effect: z = 4.53 (P < .00001)
Test for subgroup differences: χ2
1 = 0.64, (P = .42), I2 = 0%
Heterogeneity: Not applicable
Test for overall effect: z = 6.21 (P < .00001)
–3.3
–1.41
–3.37
0.5312
0.183
0.2563
73
60
60
133
53
60
60
113
73
60
60
133
–2.47 (–4.20 to –0.74)
–3.30 (–4.34 to –2.26)
–3.30 (–4.34 to –2.26)
–2.75 (–3.94 to –1.56)
–1.14 (–1.76 to –0.52)
–1.41 (–1.77 to –1.05)
–1.41 (–1.77 to –1.05)
–1.22 (–1.67 to –0.78)
–2.69 (–3.71 to –1.68)
–3.37 (–3.87 to –2.87)
–3.37 (–3.87 to –2.87)
–2.93 (–3.77 to –2.08)
73
133
60
60
53
113
60
60
73
133
60
60
Heterogeneity: τ2 = 0.45; χ2
2 = 6.78, (P = .03); I2 = 71%
Test for overall effect: z = 5.22 (P < .00001)
Heterogeneity: τ2 = 0.48; χ2
3 = 16.29, (P = .0010); I2 = 82%
Test for overall effect: z = 6.82 (P < .00001)
Test for subgroup differences: χ2
1 = 1.37, (P = .24), I2 = 27.0%
Heterogeneity: Not applicable
Test for overall effect: z = 13.15 (P < .00001)
Heterogeneity: τ2 = 0.16; χ2
1 = 4.37, (P = .04); I2 = 77%
Test for overall effect: z = 3.63 (P = .0003)
2.4.1 Split-mouth trials
Subtotal (95% CI)
2.4.2 Parallel trials
Kazancioglu and Colleagues,15 2014
Subtotal (95% CI)
Total (95% CI)
2.5.1 Split-mouth trials
Subtotal (95% CI)
2.5.2 Parallel trials
Kazancioglu and Colleagues,15 2014
Subtotal (95% CI)
Total (95% CI)
2.6.1 Split-mouth trials
00
0 Not estimable0Glória and Colleagues,17 2020
Subtotal (95% CI)
2.6.2 Parallel trials
Kazancioglu and Colleagues,15 2014
Subtotal (95% CI)
Total (95% CI)
Heterogeneity: τ2 = 0.12; χ2
2 = 9.56, (P = .008); I2 = 79%
Test for overall effect: z = 5.41 (P < .00001)
Test for subgroup differences: χ2
1 = 0.55, (P = .46), I2 = 0%
Heterogeneity: Not applicable
Test for overall effect: z = 7.70 (P < .00001)
–3.8
–2.03
0.7647
0.1391
20
33
–3.80 (–5.30 to –2.30)
–2.03 (–2.30 to –1.76)
20
33
Kazancioglu and Colleagues,14 2014
Sivalingam and Colleagues,16 2016
2.44 3.5162 20 2.44 (–4.45 to 9.33)20Glória and Colleagues,17 2020
–3.32
–2.18
0.3915
0.1971
20
33
–3.32 (–4.09 to –2.55)
–2.18 (–2.57 to –1.79)
20
33
Kazancioglu and Colleagues,14 2014
Sivalingam and Colleagues,16 2016
–2.91 3.7001 20 –2.91 (–10.16 to 4.34)20Glória and Colleagues,17 2020
–1.52
–0.88
0.2907
0.0965
20
33
100.0
31.5
31.5
68.5
100.0
34.0
34.0
66.0
100.0
33.9
33.9
66.1
40.7
25.0
2.8
35.6
29.2
1.3
40.6
25.4 –1.52 (–2.09 to –0.95)
–0.88 (–1.07 to –0.69)
20
33
Kazancioglu and Colleagues,14 2014
Sivalingam and Colleagues,16 2016
–100 –50
Favors (experimental) Favors (control)
500 100
–100 –50
Favors (experimental) Favors (control)
500 100
–100 –50
Favors (experimental) Favors (control)
500 100
Forest plot of comparison: Ozone versus placebo, outcome: Pain at 24 hours.
Experimental
Mean Difference SE Total Total Weight, %
Mean Difference
IV, Random, 95% CI
Mean Difference
IV, Random, 95% CIStudy or Subgroup
Control
Forest plot of comparison: Ozone versus placebo, outcome: Pain at 72 hours
Experimental
Mean Difference SE Total Total Weight, %
Mean Difference
IV, Random, 95% CI
Mean Difference
IV, Random, 95% CI
Study or Subgroup
Control
Forest plot of comparison: Ozone versus placebo, outcome: Pain at 7 days
Experimental
Mean Difference SE Total Total Weight, %
Mean Difference
IV, Random, 95% CI
Mean Difference
IV, Random, 95% CIStudy or Subgroup
Control
A
B
C
Figure 3. Forest plots for comparison of ozone versus control: pain at 24 hours (A), 72 hours (B), and 7 days (C). IV: Independent variable. SE: Standard
error.
JADA 152(10) nhttp://jada.ada.org nOctober 2021 849

Analgesics use
Pooled analysis revealed that analgesic use was significantly lower in the ozone group than in the
control group (weighted MD, e3.80; 95% CI, e4.38 to e3.22; I
2
¼0%). Results for sensitivity
analysis were the same as the pooled estimates as shown in Figure 7.
Safety of ozone treatment
No adverse effects were reported in any trials.
A
B
C
2.1.1 Split-mouth trials
Subtotal (95% CI)
Heterogeneity: τ2 = 5.93; χ2
2 = 90.60, (P < .00001); I2 = 98%
Test for overall effect: z = 0.82 (P = .41)
2.1.2 Parallel trials
Kazancioglu and Colleagues,15 2014
Subtotal (95% CI)
Total (95% CI)
Heterogeneity: τ2 = 4.78; χ2
3 = 148.15, (P < .00001); I2 = 98%
Test for overall effect: z = 0.40 (P = .69)
Test for subgroup differences: χ2
1 = 3.89, (P = .05), I2 = 74.3%
Heterogeneity: Not applicable
Test for overall effect: z = 7.87 (P < .00001)
1.678
0.4817
1.1428
0.2132
0.1853
0.1974
73
60
60
133
73
60
60
133
73
60
60
133
100.0
25.4
25.4
74.6
100.0
25.7
25.7
74.3
100.0
25.6
25.6
74.4
–1.16 (–3.95 to 1.63)
1.68(1.26 to 2.10)
1.68(1.26 to 2.10)
–0.44 (–2.61 to 1.73)
–1.34 (–2.96 to 0.28)
0.48 (0.12 to 0.84)
0.48 (0.12 to 0.84)
–0.87 (–2.31 to 0.56)
0.48 (–2.64 to 3.60)
1.14 (0.76 to 1.53)
1.14 (0.76 to 1.53)
0.63 (–1.42 to 2.69)
73
133
60
60
73
133
60
60
73
133
60
60
2.2.1 Split-mouth trials
Subtotal (95% CI)
Heterogeneity: τ2 = 7.44; χ2
2 = 102.65, (P < .00001); I2 = 98%
Test for overall effect: z = 0.30 (P = .76)
2.2.2 Parallel trials
Kazancioglu and Colleagues,15 2014
Subtotal (95% CI)
Total (95% CI)
Heterogeneity: τ2 = 4.26; χ2
3 = 130.85, (P < .00001); I2 = 98%
Test for overall effect: z = 0.60 (P = .55)
Test for subgroup differences: χ2
1 = 0.17, (P = .68), I2 = 0%
Heterogeneity: Not applicable
Test for overall effect: z = 5.79 (P < .00001)
2.3.1 Split-mouth trials
Subtotal (95% CI)
Heterogeneity: τ2 = 1.93; χ2
2 = 35.82, (P < .00001); I2 = 94%
Test for overall effect: z = 1.62 (P = .10)
2.3.2 Parallel trials
Kazancioglu and Colleagues,15 2014
Subtotal (95% CI)
Total (95% CI)
Heterogeneity: τ2 = 2.04; χ2
3 = 76.42, (P < .00001); I2 = 96%
Test for overall effect: z = 1.20 (P = .23)
Test for subgroup differences: χ2
1 = 4.64, (P = .03), I2 = 78.5%
Heterogeneity: Not applicable
Test for overall effect: z = 2.60 (P = .009)
Glória and Colleagues,17 2020 –0.195 0.3171 20 25.1 –0.20(–0.82 to 0.43)20
3.924
–2.5229
0.5599
0.3345
20
33 25.2
24.0 3.92 (2.83 to 5.02)
–2.52 (–3.18 to –1.87)
20
33
Kazancioglu and Colleagues,14 2014
Sivalingam and Colleagues,16 2016
0.1228 0.3166 20 25.2 0.12 (–0.50 to 0.74)20
Glória and Colleagues,17 2020
–1.0765
–2.9122
0.3407
0.3591
20
33 24.6
24.7 –1.08 (–1.74 to –0.41)
–2.91 (–3.62 to –2.21)
20
33
Kazancioglu and Colleagues,14 2014
Sivalingam and Colleagues,16 2016
–0.0606 0.3163 20 24.9 –0.06 (–0.68 to 0.56)20Glória and Colleagues,17 2020
–100 –50
Favors (experimental) Favors (control)
500 100
–100 –50
Favors (experimental) Favors (control)
500 100
–100 –50
Favors (experimental) Favors (control)
500 100
Forest plot of comparison: Ozone versus placebo, outcome: Swelling at 24 hours.
Experimental
Std. Mean Difference SE Total Total Weight, %
Std. Mean Difference
IV, Random, 95% CI
Std. Mean Difference
IV, Random, 95% CI
Study or Subgroup
Control
Forest plot of comparison: Ozone versus placebo, outcome: Swelling at 72 hours
Experimental
Std. Mean Difference SE Total Total Weight, %
Std. Mean Difference
IV, Random, 95% CI
Std. Mean Difference
IV, Random, 95% CI
Study or Subgroup
Control
Forest plot of comparison: Ozone versus placebo, outcome: Swelling at 7 days
Experimental
Std. Mean Difference SE Total Total Weight, %
Std. Mean Difference
IV, Random, 95% CI
Std. Mean Difference
IV, Random, 95% CIStudy or Subgroup
Control
Kazancioglu and Colleagues,14 2014
Sivalingam and Colleagues,16 2016
1.0169
–4.3574
0.3381
0.4621
20
33 24.5
25.0 1.02 (0.35 to 1.68)
–4.36 (–5.26 to –3.45)
20
33
Figure 4. Forest plots for comparison of ozone versus control: swelling at 24 hours (A), 72 hours (B), and 7 days (C). IV: Independent variable. SE:
Standard error. Std: Standard.
850 JADA 152(10) nhttp://jada.ada.org nOctober 2021

Quality of evidence
Only randomized, double-blind, placebo-controlled trials were included in our review. All of the
included trials reported the studied outcomes along with dispersion. Significant heterogeneity (I
2
>
75%) was identified in our analyses. This heterogeneity, the relatively small number of participants,
and the wide CIs of few outcomes resulted in a GRADE evidence certainty assessment of moderate
to low for most of our analyses. Only 1 split-mouth trial
15
and 1 parallel trial
14
assessed quality of
life, making it difficult to draw robust conclusions from this analysis. Owing to the limited number
of trials, publication bias could not be assessed using a funnel plot, and this issue was addressed by
means of citation searching.
Heterogeneity: τ2 = 10.88; χ2
2 = 11.27, (P = .004); I2 = 82%
Test for overall effect: z = 1.93 (P = .05)
Heterogeneity: τ2 = 19.67; χ2
3 = 41.83, (P < .00001); I2 = 93%
Test for overall effect: z = 1.15 (P = .25)
Test for subgroup differences: χ2
1 = 3.99, (P = .05), I2 = 74.9%
Heterogeneity: Not applicable
Test for overall effect: z = 0.56 (P = .58)
–0.5 0.8993
73
60
60
133
73
133
60
60
100.0
27.8
27.8
72.2 4.23 (–0.07 to 8.53)
–0.50 (–2.26 to 1.26)
–0.50 (–2.26 to 1.26)
2.74 (–1.93 to 7.41)
2.7.1 Split-mouth trials
Subtotal (95% CI)
2.7.2 Parallel trials
Kazancioglu and Colleagues,15 2014
Subtotal (95% CI)
Total (95% CI)
–100 –50
Favors (experimental) Favors (control)
500 100
3
7.66
1.3928
0.9116
20
33
20
33 27.7
26.3 3.00 (0.27 to 5.73)
7.66 (5.87 to 9.45)
Kazancioglu and Colleagues,14 2014
Sivalingam and Colleagues,16 2016
–0.18 3.3913 20 20 18.2 –0.18 (–6.83 to 6.47)Glória and Colleagues,17 2020
Forest plot of comparison: Ozone versus placebo, outcome: Mouth Opening at 24 hours
Experimental
Mean Difference SE Total Total Weight, %
Mean Difference
IV, Random, 95% CI
Mean Difference
IV, Random, 95% CIStudy or Subgroup
Control
0.9 1.1191
73
60
60
133
73
133
60
60
100.0
29.5
29.5
70.5 3.64 (–0.12 to 7.40)
0.90 (–1.29 to 3.09)
0.90 (–1.29 to 3.09)
2.77 (–0.63 to 6.17)
Heterogeneity: τ2 = 7.31; χ2
2 = 6.62, (P = .04); I2 = 70%
Test for overall effect: z = 1.90 (P = .06)
Heterogeneity: τ2 = 8.93; χ2
3 = 16.40, (P = .0009); I2 = 82%
Test for overall effect: z = 1.60 (P = .11)
Test for subgroup differences: χ2
1 = 1.52, (P = .22), I2 = 34.4%
Heterogeneity: Not applicable
Test for overall effect: z = 0.80 (P = .42)
2.8.1 Split-mouth trials
Subtotal (95% CI)
2.8.2 Parallel trials
Kazancioglu and Colleagues,15 2014
Subtotal (95% CI)
Total (95% CI)
–100 –50
Favors (experimental) Favors (control)
500 100
1.9
6.28
1.8102
0.8936
20
33
20
33 30.9
24.6 1.90 (–1.65 to 5.45)
6.28 (4.53 to 8.03)
Kazancioglu and Colleagues,14 2014
Sivalingam and Colleagues,16 2016
0.65 3.3577 20 20 14.9 0.65 (–5.93 to 7.23)Glória and Colleagues,17 2020
Forest plot of comparison: Ozone versus placebo, outcome: Mouth Opening at 72 hours
Experimental
Mean Difference SE Total Total Weight, %
Mean Difference
IV, Random, 95% CI
Mean Difference
IV, Random, 95% CIStudy or Subgroup
Control
0.2 0.664
73
60
60
133
73
133
60
60
100.0
33.4
33.4
66.6 2.07 (–1.21 to 5.36)
0.20 (–1.10 to 1.50)
0.20 (–1.10 to 1.50)
1.42 (–1.34 to 4.18)
Heterogeneity: τ2 = 5.03; χ2
2 = 5.08, (P = .08); I2 = 61%
Test for overall effect: z = 1.24 (P = .22)
2.9.1 Split-mouth trials
Subtotal (95% CI)
2.9.2 Parallel trials
Kazancioglu and Colleagues,15 2014
Subtotal (95% CI)
Total (95% CI)
Heterogeneity: τ2 = 5.49; χ2
3 = 15.85, (P = .001); I2 = 81%
Test for overall effect: z = 1.01 (P = .31)
Test for subgroup differences: χ2
1 = 1.08, (P = .30), I2 = 7.6%
Heterogeneity: Not applicable
Test for overall effect: z = 0.30 (P = .76)
–100 –50
Favors (experimental) Favors (control)
500 100
1.3
4.16
1.986
0.808
20
33
20
33 32.3
21.0 1.30 (–2.59 to 5.19)
4.16 (2.58 to 5.74)
Kazancioglu and Colleagues,14 2014
Sivalingam and Colleagues,16 2016
–1.97 3.0611 20 20 13.3 –1.97 (–7.97 to 4.03)
Glória and Colleagues,17 2020
Forest plot of comparison: Ozone versus placebo, outcome: Mouth Opening at 7 days
Experimental
Mean Difference SE Total Total Weight, %
Mean Difference
IV, Random, 95% CI
Mean Difference
IV, Random, 95% CIStudy or Subgroup
Control
A
B
C
Figure 5. Forest plots for comparison of ozone versus control: trismus or mouth opening at 24 hours (A), 72 hours (B), and 7 days (C). IV: Independent
variable. SE: Standard error.
JADA 152(10) nhttp://jada.ada.org nOctober 2021 851

DISCUSSION
Our meta-analysis, which included 4 RCTs
14-17
and a total sample of 133 patients, evaluated the
efficacy of ozone in reducing postoperative complications after IMTM extraction. Statistically
significant positive effects that were small showed improvement in pain and quality of life, and
reduction in number of analgesics consumed by the patients treated with ozone. However, the effect
of ozone on trismus and swelling after IMTM was statistically insignificant. Adverse events were not
found in any of the studies, which suggests that ozone appears to be a safe therapy. The quality of
evidence was moderate to low and should be interpreted with caution because the effects estimates
might have been influenced by several factors, including variable ozone regimens and patient
factors, especially pain relief, which is a subjective outcome and could be explained in part as
placebo effect. The difficulty of blinding the participants is associated with high risk of performance
bias.
28
These 4 trials
14-17
were conducted across varied geographical locations (that is, Brazil, India, and
Turkey), which might support the applicability of our findings across different practice settings and
patient populations. The populations studied in our review were comparable, with similar de-
mographics, mean surgical times, and difficulty indexes. Furthermore, the 3 split-mouth studies
15-17
had crossover trials with sufficient washout periods (2-4 weeks), reducing patient factors acting as
confounding variables. Nevertheless, there have been conflicting opinions on the split-mouth
design used in oral health care research. In a meta-epidemiologic study, Qiao and colleagues
29
concluded that split-mouth designs are similar to parallel designs if executed and analyzed prop-
erly. The split-mouth design used in our meta-analysis met the prerequisite that disease is stable with
Heterogeneity:τ2 = 0.00; χ2
1 = 0.68, (P = .41); I2 = 0%
Test for overall effect: z = 16.20 (P < .00001)
–100 –50
Favors (experimental) Favors (control)
500 100
Heterogeneity: τ2 = 0.00; χ2
1 = 0.95, (P = .33); I2 = 0%
Test for overall effect: z = 21.67 (P < .00001)
Mean Difference
IV, Random, 95% CI
Heterogeneity: τ2 = 0.00; χ2
1 = 0.41, (P = .52); I2 = 0%
Test for overall effect: z = 14.80 (P < .00001)
Test for subgroup differences: χ2
2 = 51.07, (P < .00001), I2 = 96.1%
1.4.1 At 7 days
Kazancioglu and Colleagues,14 2014
Subtotal (95% CI)
1.4.3 At 24 hours
Kazancioglu and Colleagues,15 2014
1.4.2 At 72 hours
Kazancioglu and Colleagues,14 2014
Subtotal (95% CI)
Kazancioglu and Colleagues,15 2014
Subtotal (95% CI)
Glória and Colleagues,17 2020
Kazancioglu and Colleagues,14 2014
Sivalingam and Colleagues,16 2016
Kazancioglu and Colleagues,15 2014
Study or Subgroup
100.0
72.1
27.9
100.0
73.4
26.6
Weight, %
Experimental
13.72
13.88
18.24
19.34
Mean
1.29
1.28
0.53
0.56
SD
20
60
80
20
60
80
Total
20.01
20.95
30.15
30.12
Mean
3.34
3.64
4.4
4.59
SD
20
60
80
20
60
80
Total
–6.29 (–7.86 to –4.72)
–7.07 (–8.05 to –6.09)
–6.85 (–7.68 to –6.02)
–11.91 (–13.85 to –9.97)
–10.78 (–11.95 to –9.61)
–11.08 (–12.08 to –10.08)
100.080 80 –11.24 (–12.73 to –9.75)
0 0 0 0 0 0 Not estimable
69.5
30.521.24
21.34
0
3.25
4.05
0
20
60
0
33.22
32.26
0
5.22
5.78
0
20
60
0
–11.98 (–14.67 to –9.29)
–10.92 (–12.71 to –9.13)
Not estimable
Mean Difference
IV, Random, 95% CI
Control
Figure 6. Forest plot for comparison of ozone versus control: quality of life at 24 hours, 72 hours, and 7 days. IV: Independent variable. SD: Standard
deviation.
Mean Difference
IV, Random, 95% CI
–100 –50
Favors (experimental) Favors (control)
500 100
100.0
Weight, %
Heterogeneity: τ2 = 0.00; χ2
1 = 0.00, (P = 1.00); I2 = 0%
Test for overall effect: z = 12.87 (P < .00001)
Experimental
Mean SD Total
113
Mean SD Total
113
1 0 33 1.32 0.48 33
25.0
4.6 2.1 20 8.4 1.6 20
75.0
Mean Difference
IV, Random, 95% CI
–3.80 (–4.38 to –3.22)
Not estimable
–3.80 (–4.96 to –2.64)
–3.80 (–4.47 to –3.13)
4.6 2.1 60 8.4 1.6 60
Study or Subgroup
Glória and Colleagues,17 2020
Kazancioglu and Colleagues,14 2014
Total (95% CI)
Kazancioglu and Colleagues,15 2014
Control
Figure 7. Forest plot for comparison of ozone versus control: total analgesic use. IV: Independent variable. SD: Standard deviation. SE: Standard error.
852 JADA 152(10) nhttp://jada.ada.org nOctober 2021

no period effect and no carryover effect, with a sufficient washout phase of 2 through 4 weeks. In
theory, any differences in outcomes between the 2 approaches are only due to their properties and
not participant-related factors.
One possible limitation could be probability of publication bias, which we tried to minimize by
means of searching all relevant online sources. It is still possible that some unpublished data or
articles in noneEnglish-language articles were missed, but this should not result in large bias, as
suggested by Moher and colleagues.
30
Although all 4 studies
14-17
had positive results as adjunctive therapies, this needs to be inter-
preted in the shadow of a high level of bias owing to small sample size. The 2 studies by Kazancioglu
and colleagues
14,15
have exactly the same population demographics, mean surgical time, and total
number of analgesics consumed, in totally different study designs, which is unexplainable and is an
independent finding that might question the validity of this meta-analysis. Furthermore, the
included studies used different ozone regimens in terms of states of matter and point and place of
application, which can potentially affect the outcomes studied and could be a cause of heterogeneity
of our meta-analysis. The small number of RCTs identified for inclusion in our review further
obviated the possibility of subgroup analysis.
There have been a few RCTs evaluating the efficacy of ozone therapy in some challenging
clinical scenarios, such as diabetic foot ulcers, second-degree burns, alveolar osteitis, and
bisphosphonate-related osteonecrosis of the jaw.
31-34
Ahmedi and colleagues
13
found an incidence
of 3.3% alveolar osteitis with use of prophylactic intra-alveolar ozone therapy compared with
16.67% with control participants in IMTM extractions. Dawdy and colleagues,
20
in their systematic
review of 21 RCTs, found that low-level laser therapy had negligible benefit and no adverse out-
comes. Our meta-analysis of 4 low-quality studies of ozone with no documented adverse effects
paves the way for future research to build robust evidence for or against ozone therapy as an
adjunctive in after IMTM extractions.
Our meta-analysis revealed several limitations that will guide future research in the form of high-
quality RCTs of ozone with placebos and possibly low-level laser therapy with larger sample size.
The choice of ozone regimen should be an important part of clinical decision making, but the
evidence from head-to-head trials comparing different ozone regimens for third-molar surgery is
scarce, and a pragmatic recommendation for a specific prescription requires additional robust
research to standardize the ozone therapy regimen to remove the heterogeneity in the intervention
delivered. Follow-up planning at specific times in a 1-week interval, along with standardization of
swelling evaluation, would help us analyze the outcomes in a reproducible manner. Split-mouth
designs have the advantage of optimizing resources if strict rules of implementation, statistical
analysis, and reporting are followed.
CONCLUSIONS
Evidence suggests that adjuvant ozone application can offer some benefit for reducing pain,
improving quality of life, and decreasing mean analgesic intake after IMTM extraction, but it is not
effective for reducing facial swelling and trismus. Considering the low quality of available evidence,
adequately powered, double-blind, placebo-controlled trials with appropriate randomization, allo-
cation concealment, and blinding of outcomes assessment, along with standardization of ozone
therapeutic regimen are needed in the future to substantiate efficacy of ozone after IMTM
extraction. n
SUPPLEMENTAL DATA
Supplemental data related to this article can be found at: https://doi.org/10.1016/j.adaj.2021.05.006.
Dr. Chaudhry is an associate professor, Oral and Maxillofacial Surgery,
Department of Dentistry, All India Institute of Medical Sciences, Jodhpur,
Rajasthan, India.
Dr. Rustagi is an associate professor, Department of Community Medi-
cine and Family Medicine, All India Institute of Medical Sciences, Jodhpur,
Rajasthan, India.
Dr. Bali is a professor, Department of Oral and Maxillofacial Surgery,
DAV Dental College, Yamunanagar, Haryana, India.
Dr. Khatana is a senior resident, Oral and Maxillofacial Surgery,
Department of Dentistry, All India Institute of Medical Sciences, Jodhpur,
Rajasthan, India.
JADA 152(10) nhttp://jada.ada.org nOctober 2021 853

Dr. Shailendra Kumar is a junior resident, Oral and Maxillofacial Surgery,
Department of Dentistry, All India Institute of Medical Sciences, Jodhpur,
Rajasthan, India.
Dr. Kaur is a senior resident, Oral and Maxillofacial Surgery, Department
of Dentistry, All India Institute of Medical Sciences, Basni, Jodhpur, Rajas-
than, 342005, India., email amanjotkaur1992@yahoo.com. Address cor-
respondence to Dr. Kaur.
Dr. Pravin Kumar is a professor and head, Department of Dentistry, All
India Institute of Medical Sciences, Jodhpur, Rajasthan.
Disclosures. None of the authors reported any disclosures.
The authors thank the trial authors who responded to us for the missing
data.
1. Bui CH, Seldin EB, Dodson TB. Types, frequencies,
and risk factors for complications after third molar
extraction. J Oral Maxillofac Surg. 2003;61(12):1379-1389.
2. Shugars DA, Gentile MA, Ahmad N, et al. Assess-
ment of oral healtherelated quality of life before and after
third molar surgery. J Oral Maxillofac Surg. 2006;64(12):
1721-1730.
3. Slade GD, Foy SP, Shugars DA, Phillips C,
White RP. The impact of third molar symptoms, pain, and
swelling on oral health-related quality of life. J Oral
Maxillofac Surg. 2004;62(9):1118-1124.
4. Markovic AB, Todorovic L. Postoperative analgesia
after lower third molar surgery: contribution of the use of
long-acting local anesthetics, low-power laser, and diclo-
fenac. Oral Surg Oral Med Oral Pathol Oral Radiol Endod.
2006;102(5):e4-e8.
5. Markovic A, Todorovic LJ. Effectiveness of dexa-
methasone and low-power laser in minimizing oedema
after third molar surgery: a clinical trial. Int J Oral Max-
illofac Surg. 2007;36(3):226-229.
6. Chugh A, Singh S, Mittal Y, Chugh V. Submucosal
injection of dexamethasone and methylprednisolone for
the control of postoperative sequelae after third molar
surgery: randomized controlled trial. Int J Oral Maxillofac
Surg. 2018;47(2):228-233.
7. Bamgbose BO, Akinwande JA, Adeyemo WL,
Ladeinde AL, Arotiba GT, Ogunlewe MO. Effects of co-
administered dexamethasone and diclofenac potassium
on pain, swelling and trismus following third molar sur-
gery. Head Face Med. 2005;1(1):11.
8. Almeida R de AC, Lemos CAA, de Moraes SLD,
Pellizzer EP, Vasconcelos BC. Efficacy of corticosteroids
versus placebo in impacted third molar surgery: systematic
review and meta-analysis of randomized controlled trials.
Int J Oral Maxillofac Surg. 2019;48(1):118-131.
9. Stocchero M, Gobbato L, De Biagi M, Bressan E,
Sivolella S. Pulsed electromagnetic fields for postoperative
pain: a randomized controlled clinical trial in patients
undergoing mandibular third molar extraction. Oral Surg
Oral Med Oral Pathol Oral Radiol. 2015;119(3):293-300.
10. Armond ACV, Glória JCR, dos Santos CRR,
Galo R, Falci SGM. Acupuncture on anxiety and in-
flammatory events following surgery of mandibular third
molars: a split-mouth, randomized, triple-blind clinical
trial. Int J Oral Maxillofac Surg. 2019;48(2):274-281.
11. Gelesko S, Long L, Faulk J, Phillips C, Dicus C,
White RP. Cryotherapy and topical minocycline as
adjunctive measures to control pain after third molar
surgery: an exploratory study. J Oral Maxillofac Surg. 2011;
69(11):e324-e332.
12. Brignardello-Petersen R, Carrasco-Labra A, Araya I,
Yanine N, Beyene J, Shah PS. Is adjuvant laser therapy
effective for preventing pain, swelling, and trismus after
surgical removal of impacted mandibular third molars? A
systematic review and meta-analysis. J Oral Maxillofac
Surg. 2012;70(8):1789-1801.
13. Ahmedi J, Ahmedi E, Sejfija O, Agani Z, Hamiti V.
Efficiency of gaseous ozone in reducing the development
of dry socket following surgical third molar extraction. Eur
J Dent. 2016;10(3):381-385.
14. Kazancioglu HO, Ezirganli S, Demirtas N. Compar-
ison of the influence of ozone and laser therapies on pain,
swelling, and trismus following impacted third-molar sur-
gery. Lasers Med Sci. 2014;29(4):1313-1319.
15. Kazancioglu HO, Kurklu E, Ezirganli S. Effects of
ozone therapy on pain, swelling, and trismus following
third molar surgery. Int J Oral Maxillofac Surg. 2014;43(5):
644-648.
16. Sivalingam VP, Panneerselvam E, Raja KVB,
Gopi G. Does topical ozone therapy improve patient
comfort after surgical removal of impacted mandibular
third molar? A randomized controlled trial. J Oral Max-
illofac Surg. 2017;75(1):51.e1-51.e9.
17. Glória JCR, Douglas-de-Oliveira DW, e Silva LDA,
Falci SGM, dos Santos CRR. Influence of ozonized water
on pain, oedema, and trismus during impacted third molar
surgery: a randomized, triple blind clinical trial. BMC Oral
Health. 2020;20(1):41.
18. Moher D, Liberati A, Tetzlaff J, Altman DG;
PRISMA Group. Preferred Reporting Items for Systematic
Reviews and Meta-Analyses: the PRISMA statement.
PLoS Med. 2009;6(7):e1000097.
19. Effective Practice and Organisation of Care (EPOC).
Data collection form: EPOC Resources for review authors.
Accessed February 15, 2020. http://epoc.cochrane.org/
epoc-specific-resources-review-authors
20. Dawdy J, Halladay J, Carrasco-Labra A, Araya I,
Yanine N, Brignardello-Petersen R. Efficacy of adjuvant
laser therapy in reducing postsurgical complications after
the removal of impacted mandibular third molars. JADA.
2017;148(12):887-902.e4.
21. Higgins JPT, Thomas J, Chandler J, et al, eds.
Cochrane Handbook for Systematic Reviews of In-
terventions, Version 6.0. Updated July 2019. Accessed
date is March 23, 2020. www.training.cochrane.org/
handbook
22. Guyatt GH, Oxman AD, Vist GE, et al; GRADE
Working Group. GRADE: an emerging consensus on
rating quality of evidence and strength of recommenda-
tions. BMJ. 2008;336(7650):924-926.
23. Guyatt GH, Oxman AD, Kunz R, et al; GRADE
Working Group. GRADE guidelines, 7: rating the quality
of evidencedinconsistency. J Clin Epidemiol. 2011;
64(12):1294-1302.
24. Guyatt GH, Oxman AD, Kunz R, et al. GRADE
guidelines, 6: rating the quality of evidencedimprecision.
J Clin Epidemiol. 2011;64(12):1283-1293.
25. Guyatt GH, Oxman AD, Kunz R, et al; GRADE
Working Group. GRADE guidelines, 8: rating the quality
of evidencedindirectness. J Clin Epidemiol. 2011;64(12):
1303-1310.
26. Guyatt GH, Oxman AD, Montori V, et al. GRADE
guidelines, 5: rating the quality of evidencedpublication
bias. J Clin Epidemiol. 2011;64(12):1277-1282.
27. Jaeschke R, Singer J, Guyatt GH. Measurement of
health status. Ascertaining the minimal clinically impor-
tant difference. Control Clin Trials. 1989;10(4):407-415.
28. do Nascimento-Júnior EM, dos Santos GMS,
Tavares Mendes ML, et al. Cryotherapy in reducing pain,
trismus, and facial swelling after third-molar surgery.
JADA. 2019;150(4):269-277.e1.
29. Qiao F, Fu Q-B, Guo Y-X, Zhang J. Risk factors for
the deterioration of oral health-related quality of life after
mandibular third molar removal. Int J Clin Exp Med. 2017;
10(2):3281-3288.
30. Moher D, Pham, Klassen TP, et al. What contri-
butions do languages other than English make on the
results of meta-analyses? J Clin Epidemiol. 2000;53(9):
964-972.
31. Khalifah MAA. A comparative study for the efficacy
of different forms of ozone as a treatment for alveolar
osteitis. Oral Surg. 2018;11(3):195-199.
32. Albatanony A, El-Khateep Y, Sadek SM,
Baghid AS. Evaluation of the effect of ozone therapy in
diabetic foot ulcers. Menoufia Med J. 2019;32(4):1272.
33. Agrillo A, Ungari C, Filiaci F, Priore P, Iannetti G.
Ozone therapy in the treatment of avascular
bisphosphonate-related jaw osteonecrosis. J Craniofac
Surg. 2007;18(5):1071-1075.
34. Campanati A, De Blasio S, Giuliano A, et al.
Topical ozonated oil versus hyaluronic gel for the treat-
ment of partial- to full-thickness second-degree burns: a
prospective, comparative, single-blind, non-randomised,
controlled clinical trial. Burns. 2013;39(6):1178-1183.
854 JADA 152(10) nhttp://jada.ada.org nOctober 2021

SEARCH STRATEGY
The search strategy used was the following: (Mandibular[MeSH Terms]) OR mandibular[Title/
Abstract]) OR third molar[MeSH Terms]) OR third molar[Title/Abstract]) OR impaction[MeSH
Terms]) OR impaction[Title/Abstract]) OR removal[Title/Abstract]) OR removal[MeSH Terms])
OR wisdom tooth[Title/Abstract]) OR wisdom tooth[MeSH Terms])) AND ((ozone*
[Title/Abstract]) OR ozone*[MeSH Terms]) AND (Humans[Mesh])
MeSH: Medical subject heading.
eTable. Ozone treatment compared with placebo for impacted molar.
CERTAINTY
ASSESSMENT NO. OF PATIENTS EFFECT
NO. OF
STUDIES
STUDY
DESIGN
RISK OF
BIAS Inconsistency Indirectness Imprecision
OTHER
CONSIDERATIONS
Ozone
Treatment Placebo
Relative
(95% CI)
Absolute
(95% CI) CERTAINTY
4Randomized
trials
Not
serious
Serious*Not serious Serious
†
Publication bias
strongly suspected
‡
133 133 NA
§
SMD
{
1.44
Lower (3.06 to
0.18)
Very low
4Randomized
trials
Not
serious
Serious*Not serious Serious
†
Publication bias
strongly suspected
‡
133 133 NA SMD 0.09
Lower (1.57 to 1.4)
Very low
4Randomized
trials
Not
serious
Serious*Not serious Serious
†
Publication bias
strongly suspected
‡
133 133 NA SMD 0.91
Lower (2.56 to
0.74)
Very low
4Randomized
trials
Not
serious
Serious*Not serious Serious
†
Publication bias
strongly suspected
‡
133 133 NA MD
#
1.42
Higher (1.34 to
4.18)
Very low
4Randomized
trials
Not
serious
Serious
†
Not serious Serious
†
Publication bias
strongly suspected
‡
133 133 NA MD 2.77
Higher (0.63 to
6.17)
Very low
4Randomized
trials
Not
serious
Serious*Not serious Serious
†
Publication bias
strongly suspected
‡
133 133 NA MD 2.74
Higher (1.93 to
7.41)
Very low
2Randomized
trials
Serious** Not serious Not serious Not
serious
††
Publication bias
strongly suspected
‡
80 80 NA MD 6.85 lower
(7.68 to 6.02)
Low
2Randomized
trials
Serious** Not serious Not serious Not
serious
††
Publication bias
strongly suspected
‡
80 80 NA MD 11.08
Lower (12.08 to
10.08)
Low
2Randomized
trials
Serious** Not serious Not serious Not serious Publication bias
strongly suspected
‡
80 80 NA MD 11.24
Lower (12.73 to
9.75)
Low
3Randomized
trials
Serious** Serious*Not serious Serious
†
Publication bias
strongly suspected
‡
113 113 NA MD 1.22 lower
(1.67 to 0.78)
Very low
4Randomized
trials
Serious** Serious*Not serious Not serious Publication bias
strongly suspected
‡
133 133 NA MD 2.93 lower
(3.77 to 2.08)
Very low
4Randomized
trials
Serious** Serious*Not serious Not serious Publication bias
strongly suspected
‡
133 133 NA MD 2.75 lower
(3.94 to 1.56)
Very low
2Randomized
trials
Serious** Not serious Not serious Not serious Publication bias
strongly suspected
‡
80 80 NA MD 0 e3.80 (4.38
to 3.22)
Low
* Unexplained statistical heterogeneity; however, the estimates of effect are not clinically different. †Sample size or number of events does not meet the optimal
information size as calculated, the CIs include the possibility of a small or no effect. ‡Small number of studies with small sample size with most positive results. § NA:
Not applicable. {SMD: Standardized mean difference. # MD: Mean difference. ** Investigators in most studies did not report methods for sequence generation and
allocation concealment or the reporting was unclear. Assessment of subjective outcome is likely to be influenced in trials even with blinding of participants and
outcomes, and had high risk of bias for selective outcomes reporting. †† Sample size or number of events does not meet the optimal information size as calculated.
JADA 152(10) nhttp://jada.ada.org nOctober 2021 854.e1