The H-P VAS is a 170-mm visual analog scale containing the following verbal descriptors: none, faint, weak, mild, moderate, strong, intense and maximum possible.¹⁰ We asked subjects to place a mark on the line that corresponded with their assessment of pain at the injection site and the procedural site at various time points during the study.

After subjects completed training, they performed baseline assessments of their lip numbness under the guidance of a researcher. Subjects in the mandibular study also performed baseline assessments of their tongue numbness (numbness scores should have been rated as normal at this point) under the guidance of a researcher. Subjects completed the STAR questionnaire, and they and researchers evaluated smiling, speaking, drinking 3 ounces of water, and presence or absence of drooling. Each of these assessments should have been rated as normal at this point; if not, we excluded the subject from the study.

We then randomly assigned subjects to receive one of four FDA-approved dental local anesthetic solutions by dialing an interactive voice response system (IVRS) programmed to assign four vasoconstrictor-containing local anesthetics in a 6:1:1:1 ratio. The local anesthetics with vasoconstrictors were 2 percent lidocaine with 1:100,000 epinephrine, 4 percent articaine with 1:100,000 epinephrine, 4 percent prilocaine with 1:200,000 epinephrine and 2 percent mepivacaine with 1:20,000 levonordefrin. We based the randomization scheme and our choice of local anesthetics on usage patterns in the United States. It generally is recognized that lidocaine with 1:100,000 epinephrine is the most frequently used dental local anesthetic, and these four vasoconstrictor-containing local anesthetic solutions are the most commonly used for nonsurgical procedures in the United States.¹¹

Immediately before we administered the local anesthetic injection, we obtained vital signs, including sitting or supine and standing (for one minute) systolic and diastolic blood pressure levels and pulse counts. We also recorded subjects’ respiration rates and temperatures. We performed an examination of the oral cavity with an emphasis on the appearance of the planned injection and procedural sites. If needed, we applied topical anesthetic (20 percent benzocaine) to the injection site and administered the assigned local anesthetic injection in the amount of one cartridge by using standard techniques, including aspiration before injection. If pulpal and mucosal anesthesia was insufficient to perform the procedure, we administered a second injection of one cartridge of the same anesthetic at that same location. Supplemental buccal, lingual or palatal injections (up to one-half cartridge) could be administered during the procedure as long as the injection would not produce lip anesthesia, tongue anesthesia or both. Immediately after we administered the local anesthetic injections, subjects assessed the injection site pain by using the H-P VAS, and we performed an examination of the oral cavity with an emphasis on the appearance of the injection and procedural sites. Since this study required that there be residual anesthesia after the dental procedure was completed, we enforced a crucial caveat: the dental procedure had to be completed within 60 minutes of the local anesthetic injection, and the subjects had to rate the soft tissue of their lips as still being numb at the completion of the procedure. Otherwise, we excused subjects from the study.

Our randomization assignment of subjects to receive either PM or sham injections included stratification by study center, subject’s age (12–17 years, 18–64 years, 65 years and older), local anesthetic injections administered (lidocaine or other) and the number of full local anesthetic cartridges used (one or two). The instructions we received from IVRS were to retrieve a numbered study kit for the subject being treated.

The subject and all of the researchers recording efficacy and safety data after the administration of the PM and sham injections were blinded to the treatment assignment. However, with all other study personnel removed from the research suite, one researcher at each site blindfolded the subjects and administered PM through a standard dental syringe and needle or administered a sham injection in which the plastic needle cap attached to the syringe apparatus was simply pushed against the intraoral soft tissues. We administered the PM or sham injection in exactly the same sites and in the case of PM the same number of cartridges (one or two) as the previous local anesthetic injections. We recorded the following data immediately before we randomized subjects to receive either PM or sham injections and at the indicated postinjection times:

– blood pressure levels and pulse counts, including systolic and diastolic blood pressure and pulse recordings after the subject stood for one minute within five minutes and between 10 and 20 minutes;

– sitting or supine blood pressure and pulse recordings every 15 minutes for one hour and then every hour for the next four hours;

– lip and tongue palpation assessments beginning at 10 minutes, then every five minutes for five hours;

– the STAR questionnaire results every 30 minutes for five hours;

– smiling, speaking and drooling at 10 minutes and then every five minutes until all three assessments returned to normal according to the subject and the researcher; when these returned to normal, drinking 3 ounces of water and the other three functions every five minutes until all functional assessments returned to normal according to the subject and the researcher for two consecutive assessments; followed by all four functions every 30 minutes for the remainder of the five-hour observation period.

We assessed pain at the injection site and procedure site by having the subject use the H-P VAS immediately before the study drug (PM or sham) injection, immediately after the study drug injection, every 30 minutes for two hours and every 60 minutes for the remaining three hours and before any analgesic ingestion (ibuprofen or acetaminophen), if requested by the subject. We performed an oral cavity examination just before the injection of the study drug, immediately after the injection of the study drug, then every 15 minutes for one hour and then every hour for the remaining four hours of the observation period. We considered clinically significant oral cavity abnormalities and pain ratings greater than 54 mm on the H-P VAS (which corresponds to more than mild pain) to be adverse events (AEs). We recorded any AEs that the subject elicited spontaneously or the researcher observed when they occurred. We directly asked subjects once per hour if they experienced any AEs. We asked subjects (and parents or legal guardians in the case of minor subjects) 24 to 48 hours after the completion of the treatment visit if they were experiencing any AEs or had experienced any AEs since the completion of their study visit, and we asked them about any medications they may have taken.

Data analysis. The primary efficacy endpoint for the mandibular and maxillary studies was the time to recovery of normal lip sensation based on subjects’ reports of numbness every five minutes while they performed the standardized lip palpation procedures. We calculated the time to recovery of normal lip sensation as the number of minutes elapsed from the injection of study drug to the first of two consecutive times at which the patient reported normal sensation of the lip. We similarly calculated the secondary efficacy endpoints—including the time to return to a score of zero derived from the STAR questionnaire; time to return to normal on FAB of smiling, speaking, drooling and drinking; and, in the mandibular study, the time to return of normal tongue sensation using the standardized tongue palpation method.

We performed analyses of all efficacy data by using time-to-event analysis (Kaplan-Meier methodology) to compare the median times to recovery and their corresponding 95 percent confidence intervals (CIs) between treatment groups. All tests were two-sided with a significance level set at P = .05. We used a Weibull accelerated failure time (AFT) model to perform a secondary analysis of these efficacy endpoints. The AFT model "smoothes" time-to-event analysis curves, and it allowed us to determine the relative increase or decrease in time to normal sensation (the event-time ratio) for the two treatment groups. We summarized the safety data by tabulating AEs and any clinically significant changes in oral cavity assessments, as well as by graphically plotting changes in vital signs and intraoral or perioral pain at various times during the study procedures.

We determined the sample size (226 subjects who recovered normal sensation within five hours with a two-sided significance level of P = .05 and a statistical power of 90 percent) from the results of previous studies^5,12 (also S.E. Christensen, DDS, unpublished data, August 2004; G.F. Freer, DDS, and S. McGavin, unpublished data, May 2005). On the basis of the assumption that up to 6 percent of patients would not completely return to normal lip sensation by the end of the study (five hours after the PM or sham injection was administered), we increased the target enrollment for each study to 240 subjects.

The effect of phentolamine mesylate was an 85.0-minute reduction in median time to recovery of normal lower lip sensation.

	RESULTS

TOP ABSTRACT SUBJECTS, MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
We screened 247 patients for inclusion in the mandibular study. Of the 244 subjects who met the inclusion criteria and were randomized to receive a PM or sham injection, 163 were randomized to receive lidocaine and epinephrine and 81 were assigned to receive articaine and epinephrine, prilocaine and epinephrine or mepivacaine and levonordefrin (Table 1). All 244 randomized subjects reported that their lips still were numb immediately before the injection of the study drug, while only 196 subjects still reported numbness of their tongue at this same time point.

In the mandibular study, the median elapsed time between our administration of the local anesthetic and of the study drug was 44 minutes (range: 17–78 minutes) for subjects in the PM group and 47.5 minutes (range: 20–74 minutes) for subjects in the sham group. The median time to recovery of normal lower lip sensation as assessed by subjects and as analyzed by us via time-to-event analysis using Kaplan-Meier methodology is displayed in Figure 1 (page 1086). The median time to recovery of normal lower lip sensation was 70 minutes (95 percent CI: 65–80 minutes) for subjects in the PM group and 155 minutes (95 percent CI: 140–165 minutes) for subjects in the sham group. The difference between these times was significant (P < .0001). The effect of PM was an 85.0-minute reduction (54.8 percent) in median time to recovery of normal lower lip sensation. In addition, subjects in the PM group were more likely to recover normal sensation in the lower lip during the early periods after injection of the study drug than were subjects in the sham group. Within the first 30 minutes, 21 subjects in the PM group (17.2 percent) achieved normal sensation compared with only one subject in the sham group (0.8 percent). By 60 minutes, 50 subjects in the PM group (40.9 percent) had achieved normal sensation compared with nine subjects in the sham group (7.4 percent). By 90 minutes, 86 subjects in the PM group (70.5 percent) had achieved normal sensation compared with 16 subjects in the sham group (13.1 percent). At two hours, 99 subjects in the PM group (81.1 percent) had attained normal sensation compared with only 36 subjects in the sham group (29.5 percent). Only 23 subjects in the PM group (18.9 percent) required more than two hours to achieve normal sensation in the lower lip compared with 86 subjects in the sham group (70.5 percent), which included one subject who did not achieve normal lip sensation by the end of the five-hour observation period.

View larger version (60K): [in this window] [in a new window]   Figure 1. Median time to return of normal lower lip sensation as assessed via subjects’ using a standardized finger palpation procedure. Recovery curves for the phentolamine mesylate and sham groups are shown in Kaplan-Meier time-to-event analysis plots.

View larger version (55K): [in this window] [in a new window]   Figure 2. Median time to return of normal upper lip sensation as assessed via subjects’ using a standardized finger palpation procedure. Recovery curves for the phentolamine mesylate and sham groups are shown in Kaplan-Meier time-to-event analysis plots.

Figure 3 illustrates the return of normal tongue sensation in subjects in the mandibular study. The median times to recovery of normal sensation in the tongue were 60 minutes and 125 minutes for subjects in the PM and sham groups, respectively. The effect of PM was a 65-minute reduction (52.0 percent) in median time to recovery of normal sensation for subjects in the PM group compared with subjects in the sham group (P < .0001). Subjects in the PM group recovered at earlier times throughout the study (Figure 3 and Table 3). As shown in Table 3, time-to-event analysis revealed that recovery of the subjects’ perceptions of altered function, sensation and appearance as assessed via the STAR questionnaire and the recovery of function (smiling, speaking, drooling and drinking) as assessed by FAB was quicker in subjects in the PM group than in subjects in the sham group (all P < .0001). A strong correlation in the PM group was demonstrated among the efficacy endpoints of lip anesthesia, tongue anesthesia, FAB results, STAR questionnaire results, and lip and tongue anesthesia combined (Figure 4). As illustrated in Figure 4, the recovery of function as assessed by means of FAB occurred more rapidly than did subjective assessments on the STAR questionnaire and than the time when both lip and tongue anesthesia assessed together had returned to normal.

View larger version (61K): [in this window] [in a new window]   Figure 3. Median time to return of normal tongue sensation as assessed via subjects’ using a standardized finger palpation procedure. Recovery curves for the phentolamine mesylate and sham groups are shown in Kaplan-Meier time-to-event analysis plots. The sample sizes for this analysis differ from those for the lip, as not all patients had numb tongues just before the study drug injection was administered.

View larger version (59K): [in this window] [in a new window]   Figure 4. Time course of recovery in the mandible for patients treated with phentolamine meslyate for the various efficacy endpoints including lip anesthesia, the Soft Tissue Anesthesia Recovery (STAR) questionnaire, the Functional Assessment Battery (FAB), tongue anesthesia, and the combined recovery of lip and tongue anesthesia using Weibull accelerated failure time model plots.

Results from the H-P VAS analysis indicated that for the duration of the observation period, the majority of subjects in both treatment groups who had received a mandibular injection experienced either no oral pain (32 percent for subjects in the PM group; 38 percent for subjects in the sham group) or mild oral pain (58 percent for subjects in the PM group; 56 percent for subjects in the sham group). Less than 10 percent of subjects in each group reported having moderate oral pain (Figure 5). In the maxillary study, 40 percent of subjects in both treatment groups reported having no oral pain, while 51 percent of subjects in the PM group and 52 percent of subjects in the sham group experienced mild pain. Less than 10 percent of subjects in each group reported having moderate oral pain. No subjects in either study reported having severe pain. The average H-P VAS scores for the mandibular and maxillary studies on the 170-mm scale immediately after the study drug injections were administered were 11 and 7 mm, respectively, for the PM group and 5 mm for both studies in the sham group. These scores corresponded to a verbal description of faint pain for both treatment groups. Fourteen subjects in the mandibular study (eight in the PM group and six in the sham group) took analgesics, and five subjects in the maxillary study (two in the PM group and three in the sham group) took analgesics.

View larger version (63K): [in this window] [in a new window]   Figure 5. The mean values of the Heft-Parker visual analog scale (VAS) pain scores for subjects who underwent procedures in the mandibular arch at various times during the study. Recordings of greater than 54 millimeters were considered to be clinically significant and were adverse events (AEs). mm: Millimeters.

The mean values for systolic and diastolic blood pressure and heart rate in the mandibular and maxillary studies showed only small fluctuations from baseline over the entire treatment period, and these values were nearly identical for the two randomized treatment groups. Clinically significant orthostatic changes—defined as a decrease of greater than 30 mm in systolic or diastolic blood pressure or an increase in heart rate of greater than 30 beats per minute after standing for one minute—occurred in 4.1 percent and 3.3 percent of subjects in the PM group and 6.6 percent and 2.5 percent of subjects in the sham group in the mandibular and maxillary studies, respectively. Figure 6 shows nearly identical mean systolic blood pressure readings in the PM and sham groups at various times in various situations during the mandibular study. These times and situations were as follows: sitting or supine blood pressure after standing for one minute immediately before we administered the local anesthetic injection; sitting or supine blood pressure before we administered the study drug; blood pressure after standing for one minute within five minutes and within 10 to 20 minutes after we injected the study drug; sitting or supine blood pressure levels every 15 minutes for the first hour, then hourly through the remainder of the five-hour observation period; and immediately before discharge. We did not see any differences in AE profiles, pain assessments, oral cavity assessments or vital signs between subjects in the PM and sham groups.

View larger version (70K): [in this window] [in a new window]   Figure 6. Mean values of systolic blood pressure in subjects who underwent procedures in the mandibular arch at various times during the study. Except where indicated, all measurements were taken in either the supine or sitting position. Error bars indicate standard deviations. mm Hg: Millimeters of mercury.

	DISCUSSION

TOP ABSTRACT SUBJECTS, MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

PM appeared to work equally well in subjects in all age groups and with all of the local anesthetics tested except for two small (in terms of enrollment) subgroups after maxillary injections were administered. PM appeared to be less efficacious (a 23 percent reduction in the median duration of lip anesthesia) in subjects aged 12 to 17 years, and there was no observed effect after the administration of an injection of 4 percent prilocaine with 1:200,000 epinephrine. The number of subjects enrolled in both of these subgroups (24 and 27 subjects, respectively) was approximately one-seventh of that enrolled in the 18-to 64-year-old or the 2 percent lidocaine with 1:100,000 epinephrine maxillary subgroups (188 and 159 patients, respectively), in which we observed a 62 to 63 percent reduction in the median duration of lip anesthesia. The large variability and potential error inherent with analyzing small sample sizes may explain these apparent discrepancies.^13,14 The findings of a Phase II trial in adolescent and adult subjects comparing PM with placebo showed a 78-minute reduction (69 percent) in the median duration of lip anesthesia in subjects who had received maxillary injections of 4 percent prilocaine with 1:200,000 epinephrine; the sample size, however, was too small to make firm statistical conclusions.⁵ The effect of PM on solutions that do not contain vasoconstrictors, such as 3 percent mepivacaine plain and 4 percent prilocaine plain, needs to be explored because these local anesthetics are not profound vasodilators as lidocaine is and, thus, produce an extended duration of lip and tongue anesthesia in the absence of a vaso-constrictor.³

During the clinical development program of OraVerse and in discussion with the FDA, it was agreed that a sham injection would be used as the control instead of a placebo injection that penetrated the intraoral mucosa and injected fluid. This methodology gave us the opportunity to compare in a double-blinded fashion the pain induced from a PM injection that penetrated the oral mucosa and delivered the drug with that of what was essentially no injection at all. In addition, our mandibular and maxillary studies are different from postsurgical dental pain studies, which almost always incorporate look-alike placebo and active comparator drugs, partly because many approved active analgesic agents are available and the standard of care is to intervene pharmacologically to diminish the pain.^15–17

Subjects in the mandibular and maxillary studies could not discern a difference in the discomfort associated with the PM and sham injections; both groups of subjects had an average H-P VAS pain score corresponding to faint pain. This most likely was due to the fact that the areas of the PM and sham injections still were anesthetized from the previous local anesthetic injections.

When we compared the time of local anesthetic recovery—as determined by the various assessment tools we used in the studies—for subjects in the PM group, an interesting and clinically important trend emerged. As illustrated in Figure 4, a Weibull AFT model demonstrated that recovery of normal functions as assessed by FAB after the PM injection was administered occurs somewhat before the patients’ perceived total recovery on the STAR questionnaire and the recovery of normal sensation of the lip and tongue combined (that is, in Figure 4, curves to the left indicate more rapid recovery). We noticed a similar trend of function recovering before lip numbness or perceived recovery on the STAR questionnaire in the maxillary arch after the PM injection was administered (Table 3). The clinical significance of this ordering of recovery is that at the time subjects perceived that their soft-tissue numbness had dissipated, totally normal function had returned. Thus, if they smile, speak or drink at the time they subjectively feel their numbness is gone, their facial features and articulation should be normal and they should not experience an embarrassing situation such as having a beverage they are drinking dribble out of their mouths. More importantly, we are confident that the chances of patients’ inadvertently injuring their intraoral or perioral soft tissues during speaking, drinking or eating will have dissipated as well.

One group of AEs that we were interested in was those related to the cardiovascular system, specifically changes in blood pressure and pulse rate. Because of their vasodilating effects, PM and related -adrenergic-blocking agents administered at doses approved to treat pheochromocytoma and hypertension frequently produce orthostatic hypotension with concomitant reflex tachycardia, especially in people who are treatment-naïve (that is, those who are experiencing their first-ever exposure to a drug, in this case PM).^18,19 This is why we took standing blood pressure readings and pulse rates shortly after we administered the PM and sham injections. There were no differences in any of these measures between patients in the PM and sham groups, as demonstrated in Figure 6, and in our other blood pressure and heart rate evaluations. The fact that the dosages of PM (0.4–0.8 mg) we administered by submucosal intraoral injection in adults and adolescents were approximately six- to 12-fold less than those used intramuscularly or intravenously for other approved medical indications of the drug probably contributed to the lack of cardiovascular effects. In the dosages we used in our studies, we found that PM was well-tolerated overall by the subjects.

	CONCLUSIONS

TOP ABSTRACT SUBJECTS, MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
An injection of PM administered at the same volume and at the same site as a local anesthetic with vasoconstrictor significantly and safely reduced the duration of soft-tissue anesthesia and associated functional deficits in subjects who had undergone routine nonsurgical dental procedures. In subjects for whom a rapid return to normal oral function is desirable and significant postprocedural pain is not anticipated, PM is the first drug available to clinicians that has been shown to accelerate a safe return to normal oral soft-tissue function.