Efficacy of the Semont Maneuver in Benign Paroxysmal Positional Vertigo FREE

BENIGN PAROXYSMAL positional vertigo (BPPV) is a frequent cause of dizziness and, according to Nedzelski et al,¹ represents some 17% of vertigo cases investigated in an otoneurologic unit. The annual incidence of this disease is estimated at 64 cases per 100 000 people.² The most frequent cause is a mechanical disorder in the posterior semicircular canal; impairment stemming from the horizontal and anterior semicircular canal has been reported, but it is much rarer.³

Benign paroxysmal positional vertigo of the posterior semicircular canal (p-BPPV) occurs during rapid rotational motions of the head toward the affected ear when the neck is hyperextended. It can be clinically provoked using "triggering maneuvers," such as those described by Dix and Hallpike⁴ or Brandt and Daroff.⁵ It is characterized by a brief attack of rotational vertigo, neurovegetative disturbances of varying intensity, and a characteristic nystagmus. This nystagmus is linear or rotatory, geotropic, and lasts less than 1 minute; has a latency period between the maneuver that evokes it and its appearance; lessens when this maneuver is repeated (fatigability); and changes direction as soon as the patient returns to a sitting position at the end of the maneuver (reversibility).

Following the histopathological discovery in 1969 of basophilic deposits on the cupula of the posterior canal, Schuknecht⁶ proposed the theory of cupulolithiasis as the causal mechanism of BPPV. This theory is based on the hypothesis that detached otolithic debris from the utricular macula can deposit on the vestibular side of the posterior canal cupula, thus transforming it into a linear acceleration transducer, sensitive to the effects of gravity. In 1992, Parnes and McClure⁷ identified free-floating particles in the endolymph of the posterior canal, thus supporting a canalolithiasis mechanism as the cause of BPPV. Canalolithiasis is considered a more probable pathophysiological mechanism for BPPV than cupulolithiasis and provides a better explanation for the clinical characteristics of this disease.⁸ In most cases, the etiology of BPPV cannot be identified, but the symptoms are frequently preceded by head trauma or vestibular neuritis.^{9 - 10} In addition, numerous pathologic conditions associated with BPPV have been described in the literature, but a causal relationship has not been established.^{9 - 17} Benign paroxysmal positional vertigo has a tendency to spontaneously disappear, usually after a few weeks, but if no treatment is undertaken, it may persist several months or years. Relapses can likewise occur.

The current treatment of choice for p-BPPV is physiotherapy, and 2 types of maneuvers have been developed: (1) positioning maneuvers based on repeating the triggering maneuver, the aim of which is to disperse debris into the utricular cavity or to promote central habituation mechanisms^{5 ,18} and (2) liberatory maneuvers aimed at dislodging inorganic fragments located on the cupula of the posterior canal or floating in the canal, such as the maneuvers by Semont et al¹⁹ and Epley.²⁰

The aims of this retrospective study were to evaluate the efficacy of the Semont maneuver in treating p-BPPV and to assess the possible effect of various factors on the efficacy of this maneuver.

This retrospective study included 278 patients, examined between March 1, 1993, and March 1, 1999, who presented with a symptomatic, unilateral p-BPPV and who were exclusively treated with the Semont maneuver. There were 181 females and 97 males, with a median age of 55½ years (range, 13½ to 91½ years). The diagnosis of p-BPPV was made using the triggering maneuver of Brandt and Daroff⁵ as follows: the patient is seated on an examining table facing the examiner; then he or she is rapidly laid down on the affected side with the head rotated, nose facing up at 45° to the horizon, which provokes a vertigo with linear or rotatory (torsional), geotropic (ie, beating toward the undermost ear), or upbeating nystagmus, according to the position of the eyes in the orbit. Patients presenting with BPPV of the anterior and horizontal canals were excluded. Approximately 70% of patients sought medical attention within the first month following the onset of symptoms. Patients were classified according to the etiology of BPPV and subdivided into 4 categories similar to those of Baloh et al⁹ : idiopathic, traumatic, infectious, and miscellaneous.

The Semont maneuver was performed after the triggering maneuver: after maintaining the Brandt and Daroff position for approximately 2 minutes, the patient is rapidly returned to a contralateral decubitus position, nose down. If the maneuver is effective, a vertigo with nystagmus beating toward the uppermost ear is observed. After about 3 minutes, the patient is then gently raised to a seated position. Only a single maneuver was performed during any one session. The patient was then followed-up at intervals of approximately 1 week, and the maneuver was repeated if the vertigo persisted. The patient was considered cured when the symptoms had completely disappeared. Because p-BPPV can spontaneously resolve within a few weeks, a time limit of 30 days (allowing time to perform 4 Semont maneuvers) was fixed, after which recovery from p-BPPV was no longer attributed to the Semont maneuver.

The success of the Semont maneuver is described by a generalized linear model in which it is assumed that the probability of success at any given session is a fixed percentage of the probability that existed at the preceding maneuver. This type of model is flexible and requires only 2 factors, the initial success rate and the "deterioration" of this rate during repetitions that some patients require. If this coefficient of reproduction does not differ significantly from 100%, one can assume that the maneuver retains the same efficacy throughout the different repetitions.

To adapt the model to the data, the method of maximum likelihood was used, and the likelihood ratio test was used in a regression approach to determine the factors that played a significant role on the 2 factors of interest. The model was fitted at each level of the significant factors. To assess the quality of the fit, a classic goodness-of-fit χ² test was performed that compared the observed counts with those predicted by the model for each specific category of the significant factors. A multivariate analysis was also performed to determine which combinations of the significant factors were important. The level of significance was set at P<.05.

As an example, the likelihood of success after the third maneuver can be written as: (1 − P) × (1 − αP) × α²P, where P is the initial probability of success and α represents the effect of the maneuver repetition. The expression describes the failure of the first maneuver (1 − P), then the failure of the second maneuver (1 − αP), and the success of the third maneuver (α²P), where α² represents the double reduction of efficacy associated with the 2 previous attempts.

Table 1 gives the various etiologies and their distribution by sex. The female-male ratio was more than 2:1 for all etiologies except trauma, for which there were more males than females. Among the 278 patients studied, 251 (90.3%) were cured after a maximum of 4 Semont maneuvers. Figure 1 shows the distribution of patients healed at each Semont maneuver performed and the 9.7% (27 patients) who were not cured within 30 days. Table 2 presents the classification of patients as a function of the different factors studied, as well as the global efficacy of the Semont maneuver (cumulative rate of recovery after a maximum of 4 maneuvers). Patients were subdivided according to age into 3 categories, each containing approximately the same number of patients. Figure 2 shows that the efficacy recorded was 62.6% (174/278) for the first Semont maneuver, 55.8% (58/104) for the second, 28.3% (13/46) for the third, and 18.2% (6/33) for the fourth. The other curve in the figure is obtained from the linear model defined in the "Statistical Analysis" subsection of the "Methods" section. This model accurately describes the observations and can be considered as a smoothing of the data. The probability of recovery after the first maneuver is thus estimated at 64% and decreases by an α level of .71 for each repeated maneuver.

In the regression analysis, the duration of symptoms before consultation had a significant effect on the efficacy of the Semont maneuver (P<.001). Table 3 gives the raw data and the goodness of fit of the model for the different categories of duration. Figure 3 shows the level of recovery among the 4 duration categories as smoothed out by the model. Patients who waited longer than 6 months had the poorest probability of initial recovery (46.2%) and the worst deterioration rate (α = .42), while patients who consulted within 1 week after symptom onset had a nonsignificant decrease in efficacy (P = .18). The etiology of BPPV also had a significant effect on the efficacy of the Semont maneuver (P = .002). Table 4 gives the data and goodness of fit of the model for the different etiologies. The quality of the fit in Table 4 is not as good as that for Table 3. However, the low P = .03 for the idiopathic category is not to be considered as a rejection of the proposed model. Indeed, the goodness-of-fit test was performed 8 times (4 in Table 3 and 4 in Table 4), and a Bonferroni approach requires that P be compared with .05 divided by 8, which is .006; thus, P = .03 is far from significant. Figure 4 shows the recovery rate among the 4 etiological categories as smoothed out by the model. The traumatic category had the poorest probability of initial recovery (50.5%) and the worst deterioration rate (α = .62). For the infectious and miscellaneous categories, the deterioration was not significant (P = .20 and P = .70, respectively). Age (P = .12), sex (P = .06), and affected side (P = .20) had no significant effect on the efficacy of the Semont maneuver. A multivariate regression analysis showed that patients presenting with traumatic BPPV or BPPV lasting longer than 6 months had a recovery rate of 74.7% (59/79), while patients not having any of these factors had a recovery rate of 96.5% (192/199).

The Semont maneuver is effective, demonstrating a 90.3% recovery rate after 4 maneuvers. We assume that this is primarily a result of the treatment, although spontaneous resolution may occur in some patients during the evaluation period. Similar results have been reported by Serafini et al,¹⁶ who obtained a 94% success rate after 4 maneuvers. Other therapeutic maneuvers have proven effective in curing p-BPPV, such as the Epley²⁰ maneuver or its modifications as developed by various authors.^{21 - 27} However, studies that have compared the efficacy of some of these maneuvers with that of the Semont liberatory maneuver²³ or a slightly modified Semont maneuver ²¹ have not shown significant differences in the rates of patient recovery. Although the Semont liberatory maneuver was initially conceived based on the cupulolithiasis hypothesis, the way it is carried out can equally account for a repositioning of free-floating particles in the ampullofugal branch of the posterior canal (canalolithiasis) toward the utricule.²⁸

Our results show that most patients (83.5%) are cured after only 2 Semont maneuvers and that ensuing maneuvers only slightly improve (6.8%) the global recovery rate. Nevertheless, this supplementary recovery rate, apparently weak, represents 41.3% (19/46) of patients not cured after the first 2 maneuvers. It is, therefore, well worth the effort to repeat the maneuver several times. Conceptually, the probability of recovery could be the same for each Semont maneuver repeated on any given patient. However, it is noted that generally the more one repeats the maneuver the poorer the probability that the maneuver leads to recovery. Results obtained in this study support the hypothesis that, within any given category of patients, this decrease appears constant (represented by α in the model). However, some categories have an α that is not significantly different from 1, ie, no decrease in efficacy.

Among the various factors of interest investigated, the etiology of BPPV and the duration of symptoms before initiating treatment prove to have a significant effect on the efficacy of the Semont maneuver. It appears that traumatic BPPV and BPPV lasting more than 6 months are less responsive to the treatment. Indeed, when patients present with 1 of these factors, the level of success of the Semont maneuver is not more than 74.7%, while it increases to 96.5% if both factors are absent. In addition, the 8 patients who presented with traumatic BPPV that lasted longer than 6 months had a poorer prognosis for recovery. Indeed, only 3 of these were cured (results not shown), suggesting that the negative effects of these factors may be cumulative. It is possible that traumatic BPPV results from more severe otolithic or endolabyrinthine damage than that seen in nontraumatic BPPV. This would explain the stronger resistance to therapeutic maneuvers. The fact that patients seeking treatment later respond less favorably to the Semont maneuver has already been reported by others,¹⁶ but it remains unexplained. For these patients in whom the prognosis is not excellent with the Semont maneuver, it may be indicated to replace this liberatory maneuver with an alternate treatment, as suggested by some authors.²³ In light of our results, we suggest that the change should occur after the second Semont maneuver. On the other hand, the large group of patients who consulted within 1 week after the beginning of symptoms had a very good prognosis of healing (97.4%), because the probability of success remains the same at each maneuver. Therefore, for these patients, there is no indication to move to an alternate treatment. Two other categories of patient etiologies (infectious and miscellaneous) also showed a nondecreasing rate of success. However, their sizes were small (Table 2) and these results should be approached with caution.

The greater prevalence of women (2:1) among patients with BPPV, as well as the higher prevalence of women among patients with idiopathic BPPV (2.3:1), has been previously reported in the literature.^{9 - 10 ,14 ,16 ,29} The hypothesis of a hormonal effect on the formation of endolymphatic deposits, especially postmenopausal, has been proposed²⁹ but, again, has never been demonstrated. The predominance of traumatic BPPV among men, primarily noted in young individuals, is likely because of a risk behavior that is usually more frequently observed among these individuals and may explain the near significance (P = .06) of the effect of sex on the efficacy of the maneuver. Curiously, BPPV is preferentially localized to the right side (57.2%, P<.02). We cannot propose a rational explanation for this phenomenon, which has also been observed in other studies.^{10 ,16 ,29}

The Semont liberatory maneuver is an effective treatment for p-BPPV, curing 90.3% of patients in a maximum of 4 sessions and 83.5% after only 2 sessions. When the patient belatedly seeks medical attention (>6 months from symptom onset) or when BPPV has a traumatic origin, the success rate is less satisfactory (74.7%) than in patients without these conditions (96.5%). In the less satisfactory groups, pursuing the treatment beyond the second maneuver yields little benefit and may indicate the need to change the treatment. On the contrary, some groups have a good prognosis, such as patients consulting within 1 week after symptom onset, in whom the Semont maneuver is a treatment of choice.

Corresponding author and reprints: Raphaël Maire, MD, Unité d'Otoneurologie, Service d'Oto-rhino-laryngologie et chirurgie cervico-faciale, Centre Hospitalier Universitaire Vaudois (CHUV), 1011 Lausanne, Switzerland (e-mail: Raphael.Maire@chuv.hospvd.ch).

Accepted for publication October 24, 2002.