Quantity of Research Available
The literature search yielded 482 citations. Upon screening titles and abstracts, 406 citations were excluded and 22 potentially relevant articles were retrieved for full-text review. No additional potentially relevant reports were retrieved from the grey literature and by hand search. Of the 22 potentially relevant reports 18 were excluded. Four RCTs6–9 met the inclusion criteria. The process of study selection is outlined in the PRISMA flowchart (Appendix 1).
Summary of Study Characteristics
A detailed summary of the included study and guidelines is provided in Appendix 2.
Dental Pain
Jena et al., 20136
A randomized, double-blinded study design was conducted. A total of 100 patients (63 males) between the ages of 18 and 65 years in acute pain with mandibular molar teeth diagnosed as acute irreversible pulpitis were included. Patients must have been in good health with vital mandibular molar teeth actively experiencing moderate-to-severe pain (≥85 mm) as determined by a Heft-Parker Visual Analogue Scale (VAS). Patients also must have had a prolonged response to cold testing. Patients were excluded if they had an allergy to ibuprofen, ketorolac, etodolac, aceclofenac, or paracetamol, if they had no response to cold testing or periradicular pathosis (other than a widened periodontal ligament), if they had a history of significant medical problems, gastrointestinal problems, syndrome of nasal polyps, angioedema or bronchospastic reactivity to aspirin or other NSAIDs, taken central nervous system (CNS) depressants (including alcohol or any analgesic medications) within the last 48 hours or if they were pregnant. Patients were randomized to one of five groups: 1) placebo with sugar coated pills, 2) Ibuprofen (600 mg) (n=20), 3) Ketorolac (10 mg) (n=20), 4) combination of etodolac with paracetamol (400 mg + 500mg) (n=20), and 5) combination of aceclofenac with paracetamol (100 mg + 500mg) (n=20). Before receiving administration of anesthesia, patients underwent cold testing using Green Endo ice spray to determine level of pain using the 170mm Heft-Parker VAS scale with 0 mm representing no pain and 170mm the worst possible pain. Treatment was provided 30 minutes prior to the administration of anesthesia. Inferior alveolar nerve block (IANB) was administered using 2% lidocaine with 1:100,000 adrenaline. The teeth were isolated with a rubber dam, and a conventional access opening was initiated. The treatment consisted of three phases: access into dentin, access into the pulp chamber, and instrumentation of the canals. Patients were instructed to rate pain during endodontic treatment with the Heft-Parker VAS. Statistical testing using analysis of variance (ANOVA) was used to compare between group differences.
Aggarwal et al., 20107
A randomized, double-blinded study design was conducted. A total of 69 patients (14 males) between the ages of 21 and 35 years who reported to the dental emergency department with acute pain with mandibular molar teeth were included. Patients must have had active pain in a mandibular molar, prolonged response to cold testing with an ice stick and an electric pulp tester, and absence of any periapical radiolucency on radiographs, with the exception for a widened periodontal ligament and a vital coronal pulp on access opening. Patients were excluded if they had an allergy, sensitivity, or contraindications to any opioid or non-opioid analgesic, history of active peptic ulcer within the preceding 12 months, history of bleeding problems or anticoagulant use within the last month, pregnant or breast-feeding, a history of known or suspected drug abuse, and if NSAIDs were taken within 12 hours prior to the administration of the study medication. Patients were also excluded if they had active pain in more than 1 mandibular molar. Patients were randomized to one of three groups: 1) placebo capsules (n=24), 2) Ibuprofen (300 mg) (n=22), and 3) Ketorolac (10 mg) (n=23). All patients received standard IANB injection using 1.8mL of 2% lidocaine with 1:200,000 epinephrine one hour after oral administration of study treatment. Patients were asked to rate their pain on the 170mm Heft-Parker VAS 15 minutes after initial IANB. Patients were excluded from the study when the block was considered unsuccessful (if lip numbness was not recorded within 15 minutes). Treated teeth were isolated with a rubber dam. During the procedure, patients were instructed to raise their hand when pain was experienced, and treatment was stopped. Patients were again asked to rate the pain on Heft-Parker VAS. The extent of access preparation and/or instrumentation was recorded as “within dentin”, “within pulpal space”, and “instrumentation of canals”. Statistical testing for anesthetic success (defined as Heft-Parker VAS score > 54 mm) was conducted using chi-square tests.
Non-dental/Non-cancer Pain
Ortiz et al., 20108
A randomized, double-blinded study design was conducted. A total of 49 patients (proportion of males/females not provided) between the ages of 18 and 55 years with closed ankle fractures hospitalized in the trauma service were included. Patients must had acute pain ≥ 5 cm according to a 10-cm VAS ranging from 0 (no pain) to 10 (worst pain), good health determined by clinical history, without sanguineous dyscrasias or hypersensitivity to study treatment. No specific exclusion criteria were provided. At 24 hours post-treatment, patients rated their pain with a Likert scale (0 representing complete relief; no pain during treatment; 1 representing slight relief, pain intermittently throughout the study, which is very tolerable; 2 representing moderate relief, pain intermittently throughout the study, which causes inconvenience and discomfort to the patient, but not leaving the study; and 3 representing no pain subsided with treatment) at baseline and were randomized to one of three groups: 1) ketorolac (10mg) (n=15), 2) etoricoxib (60 mg) (n=17), and 3) diclofenac (70 mg) (n=17) all taken orally twice daily. Patients’ measurements for pain were recorded at 0, 2, 4, 8, 12 and 24 hours using the VAS.
Kaeding et al., 20049
A randomized study design was conducted. A total of 50 patients (34 males) between the ages of 14 and 46 years who underwent acute or chronic anterior cruciate ligament (ACL) reconstruction surgery were included. No specific inclusion criteria were provided. Patients were excluded if they were pregnant women, had a history of gastrointestinal bleeding, ulceration, were allergic to NSAIDs, had renal disease or coagulation disorders, or if they were undergoing multi-ligament reconstruction or meniscal repair. Patients were randomized to one of two groups: 1) rofecoxib (50 mg orally while in the preoperative holding area and 50 mg orally every morning thereafter for 5 days) (n=25), or 2) ketorolac (30 mg intravenously intraoperatively and 10 mg orally four times daily for first 5 postoperative days) (n=25). At baseline, patients rated their pain using a VAS ranging from 1 (extremely poor) to 5 (extremely good, no pain) in the preoperative holding area. All patients received standardized anesthesia both intraoperatively and postoperatively and were prescribed oxycodone (1 or 2 tablets orally every 3 hours as needed for breakthrough pain). Patients’ pain scores (VAS) and severity of side effects were measured at 5 days post-surgery. Statistical analyses were conducted using Student t tests for continuous data, and Fisher exact test for discrete data.
Summary of Critical Appraisal
The strengths and limitations of included studies are summarized in Appendix 3.
In regards to the RCTs pertaining to dental pain, the methodological quality of Jenna et al.6 was poor as the randomization and concealment methods were not adequately described, blinding procedures were not adequately described, and baseline patients characteristics were not well reported. No sample size calculation was provided, thus it remains unclear whether the study was adequately powered to detect meaningful differences. The proportion of patients actually completing the study was unclear. The RCT by Aggarwal et al.7 was of high methodology quality as an appropriate and clearly focused research question was posed, the only difference between treatment groups was the treatment under investigation, and randomization, concealment methods, and blinding procedures were well described. A sample size calculation was provided and indicated that the study was adequately powered. A list of both inclusion and exclusion criteria was provided. However, results did not demonstrate statistical significance.
Both non-dental/non-cancer RCTs were of poor methodological quality. The RCT by Ortiz et al.8 did not adequately describe randomization, concealment methods and blinding procedures. Baseline characteristics were not well reported, and it was unclear how many patients were originally randomized in the study. Statistical testing was performed, though the tests used were not specified and patient exclusion criteria were not provided. The RCT by Kaeding et al.9 did not blind participants which may have impacted the subjective pain outcome measures. Though randomization methods were well described, and baseline patient characteristics were reported, statistical measures of dispersion and confidence intervals were not provided. Specific inclusion criteria were not provided and none of the results demonstrated statistical significance. Oxycodone use among the ketorolac group was greater compared with the rofecoxib group, thus it is unclear whether this affected pain score results.
Summary of Findings
A summary of the main study findings can be found in Appendix 4.
As seen in Appendix 4 Table 1, The RCT by Jena et al.6 (n=100) revealed, despite having the highest mean (SD) Heft-Parker VAS score before local anesthesia, the ketorolac group had the lowest mean (SD) Heft-Parker VAS score during endodontic treatment (27.80 [47.02]) compared with the ibuprofen (46.30 [54.23]), etodolac plus paracetamol (46.50 [51.10]), aceclofencac (39.85 [52.02]) and placebo (62.35 [55.37]) groups. Differences between groups were not statistically significant. Although no statistical analyses were performed for pain type, the proportion of patients who experienced “no pain” was greatest in the ketorolac group (70%) compared to the other treatment groups, while the lowest proportion of patients experiencing “severe pain” (Hefter-Parker VAS score 86mm to 170mm) was also in the ketorolac group (20%).
As seen in Appendix 4 Table 2, The RCT by Aggarwal et al., 20107 (n=69) revealed that the proportion of patients experiencing “mild pain” (Heft-Parker VAS score < 54 mm) was lower in the ibuprofen group (27%) during endodontic treatment. The proportion of patients experiencing “moderate pain” (Heft-Parker VAS score 54mm to 114mm) during endodontic treatment was lowest among the ibuprofen group for within dentin (25%), and placebo group for within pulpal space (12%) and for instrumentation of canals (12%) . The proportion of patients experiencing “severe pain” (Heft-Parker VAS score > 114 mm) during endodontic treatment was lower among the placebo group for within dentin (18%), and the ketorolac group for within pulpal space (7%) and for instrumentation of canals (7%). Between-group differences were not statistically significant.
As seen in Appendix 4 Table 3, The RCT by Ortiz et al. 20108 (n=49) revealed that etoricoxib group had the lowest (less pain) mean (SE) VAS pain score (20.1 [4.3]) compared with the other treatment groups at 24 hours post-treatment, while the mean (SE) Likert scale scores were both similarly greater (less pain) among the ketorolac (1.13 [0.8]) and etoricoxib (1.13 [0.9]) compared with placebo (1.07 [0.7]) at 24 hours post-treatment. Though it was unclear what statistical methods were used, the investigators stated that between-group differences were not statistically significant.
As seen in Appendix 4 Table 4, The RCT by Kaeding et al. 20049 (n=50) revealed that rofecoxib group had greater (less pain) mean VAS scores (3.57) compared with the ketorolac group (3.49) at 5 days post-operation. The mean change from baseline (0.13) and overall daily pain mean VAS score (4.09) was also greater in the rofecoxib group. The rofecoxib group used less oxycodone (mean change of −2.52 pills per day) compared with the ketorolac group (mean change of −1.86 pills per day). The proportion of patients experiencing incision-site bleeding was greater among the ketorolac group (28%) compared with the rofecoxib group (8%), while a greater proportion of patients experienced nausea in the rofecoxib group (28%) compared with the ketorolac group (16%). Between-group differences were not statistically significant.
Limitations
As the majority of included were of poor methodological quality, results should be interpreted with caution. Given the lack of detail regarding the randomization, concealment methods and blinding processes in several of the included studies, the validity of the subjective pain outcome results is uncertain. With four studies being retrieved, the literature pertaining to effectiveness and safety of oral ketorolac compared with other NSAIDS is limited. No studies measuring long-term effects such as health-related quality of life were retrieved. Furthermore, three of the four included studies were conducted in either India or Mexico, thus generalizability of the findings to the Canadian population is uncertain.
