Etiology and Pathology
Trigeminal neuralgia may be produced in some instances by intracranial tumors, usually in the posterior fossa (Dandy, 1936; O'Connell, 1978; Destee et al., 1980). Revilla (1948) reviewed all the operated cases of tic douloureux in The Johns Hopkins Hospital and found that 5.1% were apparently produced by posterior fossa tumors (meningiomas, acoustic neurinomas, and epidermoids) that were not suspected before operation. Richards et al. (1983) found unsuspected tumors in three patients with trigeminal neuralgia among 52 patients who underwent posterior fossa exploration (5.8%). Olivecrona (1941) noted that it was the small tumors that caused tic and described them as being ``between the size of a pea and a hazelnut.'' The reports of Love and Woltman (1942) and Hamby (1943) also suggest that such lesions, when present, are small. Paradoxically, we have seen one patient with trigeminal neuralgia who had a meningioma in the contralateral middle fossa, and Mealey et al. (1965) reported a similar case.
About 2---4% of patients with trigeminal neuralgia suffer from multiple sclerosis (Harris, 1950; Stookey and Ransohoff, 1959), while 2---2.5% of patients with multiple sclerosis have a form of trigeminal neuralgia (Parker, 1928; Garcin et al., 1960). Rushton and Olafson (1965) found 35 cases of combined trigeminal neuralgia and multiple sclerosis in a review of 1735 cases of tic (2%) and 3880 cases of multiple sclerosis (0.9%) seen at the Mayo Clinic. The neuralgia tended to occur early in the course of the disease and was usually bilateral. According to Dalessio (1980), trigeminal neuralgia is almost never the first manifestation of the disease. In other respects, trigeminal neuralgia with multiple sclerosis is similar to the usual form of trigeminal neuralgia. Olafson et al. (1966) reported necropsy findings from one case and reviewed the histologic findings from other reported cases. In each, the plaque was located at the junction of the trigeminal nerve and the pons. Lazar and Kirkpatrick (1979) have used electron microscopy to identify a plaque of demyelination within the trigeminal nerve root just lateral to the pons in a patient with trigeminal neuralgia and other symptoms and signs of multiple sclerosis.
Other causes of trigeminal neuralgia include trauma, dental problems (Ratner et al., 1976; Shklar et al., 1976; Socransky et al., 1976), hydrocephalus (Maurice-Williams and Pilling, 1977; Tucker et al., 1978), and digitalis intoxication (Batterman and Gutner, 1948; Bernat and Sullivan, 1979). In 1963, Kerr reviewed the types of lesions that had been associated with trigeminal neuralgia. In addition to those mentioned above, he cited reports of aberrant arterioles, cirsoid aneurysms, a tortuous basilar artery, and, very rarely, an aneurysm of the petrous portion of the carotid artery. This report was generally ignored despite a previous paper by Sunderland (1948) demonstrating vascular compression of various cranial nerves in anatomic material. Although Dandy (1934) proposed that ``idiopathic'' trigeminal neuralgia was associated with compression of the trigeminal nerve by branches of the superior cerebellar artery at the root entry zone near the pons, it was not until Gardner (1962) and Jannetta and Rand (1966) reintroduced this concept, that its significance was appreciated (see also Jannetta, 1967, 1976). Since these initial reports, other investigators (Apfelbaum, 1977; Haines et al., 1979; Sato et al., 1979; Waga et al., 1979; Petty et al., 1980; Burchiel et al., 1981; Ferguson et al., 1981; Richards et al., 1983) have emphasized the findings of both arterial (normal and ectatic) and venous structures compressing the sensory root of the trigeminal nerve in patients with trigeminal neuralgia. Jannetta (1977) reported that of 88 patients with trigeminal neuralgia from vascular compression, compression or distortion of the root was caused by a superior cerebellar artery in 82, by an anterior inferior cerebellar artery in 4, and by a vein in 2, and Petty et al. (1980) reported vascular compression of the trigeminal root, primarily arterial, in 44 of 50 patients with trigeminal neuralgia. Although the vascular compression theory of trigeminal neuralgia remains controversial, the results of decompression surgery for this disorder (see below) support this concept.
Dandy (1936) believed that the sensory root is always the site of the lesion and showed that partial cutting of this root relieves the patient of pain and leaves other sensations intact or relatively so. Furthermore, it is only in true trigeminal neuralgia that trigger zones are present. Operations by the temporal route have seldom revealed pathology, since the pain of trigeminal neuralgia, which typically jumps from one branch of the nerve to another without affecting the entire sensory distribution of the original branch involved, is not seen in lesions of the Gasserian ganglion or of the peripheral branches of the nerve.
In 1957, List and Williams postulated that a central locus of pathology in the spinal nucleus of the trigeminal nerve was more probable than a peripheral one. Kerr (1963) presented cogent arguments against this theory and defended the peripheral etiology of the disease. He noted, among several points, that even the lesion in multiple sclerosis involves essentially peripheral fibers where the fifth nerve enters the pons. The studies by Gardner (1962), Jannetta (1977), and their colleagues regarding vascular compressive phenomena in such patients would appear to support the ``peripheral'' theory. In addition, Kerr and Miller (1966) examined specimens of the Gasserian ganglion removed at surgery in 12 cases of trigeminal neuralgia. Definite ultrastructural abnormalities were found in all but one specimen. Findings included pronounced proliferative and degenerative changes in myelin sheaths, swelling of axonal mitochondria, and minimal alteration of axons. Kruger et al. (1967) reviewed light and electron microscopic observations of Gasserian ganglia from patients with trigeminal neuralgia, and Kumagami (1974) also described electron microscopic changes in the trigeminal nerves of patients with trigeminal neuralgia. These changes included large, abnormal nerve fibers with hyperplastic myelin sheaths and disappearance of the myelin sheath with exposure of the axis cylinders to surrounding connective tissue. According to Loeser et al. (1977), however, no study has correlated changes within the trigeminal descending tract with those observed in the posterior root or Gasserian ganglion. Thus, it is impossible to state whether the changes observed in the root and ganglion are unrelated to, caused by, or cause the syndrome of trigeminal neuralgia.
Dott (1951) proposed that the mechanism by which touch stimuli trigger paroxysms is one of demyelination with short-circuiting into the pain pathway. Gardner (1962), in a discussion of the association between trigeminal neuralgia and hemifacial spasm, gave convincing arguments that the two conditions are of similar etiology. Trigeminal neuralgia, according to Gardner, exhibits the characteristics of a reflex that has become misrouted. It is a response to a stimulus that outlasts the stimulus by many seconds, is followed by a refractory period of seconds or minutes, and is stopped by a peripheral interruption of the afferent limb of the reflex arc. Since it is an afferent response to an afferent stimulus, it can be explained best by a ``short-circuit'' in which the action current of the evoked afferent barrage, transmitted by small, slowly conducting fibers in the sensory root, on reaching the point of compression, excites the pain fibers of similar size and rate of conduction.
According to Loeser et al. (1977), the clinical features of trigeminal neuralgia can be explained by a theory that utilizes two known physiologic mechanisms: the trigeminal dorsal root reflex (or ``presynaptic inhibition'') and impulse reflection or postpriming repetitive firing from a focal region of altered axonal size or myelination (Howe et al., 1976; Calvin et al., 1977).

Differential Diagnosis
It has already been stated that small tumors compressing the sensory root of the trigeminal nerve may be associated with trigeminal neuralgia and that multiple sclerosis may be responsible for facial pain in a small percentage of patients. In addition, the possibility of other rare neurologic abnormalities as well as drug intoxication (digitalis) must be considered. In such cases, the syndrome of facial pain is not that of ``classic'' trigeminal neuralgia. The pain may be constant, there may be changes in facial sensibility, and so-called trigger zones may be absent (Cusick, 1981). It should thus be apparent that all patients with ``trigeminal neuralgia'' should undergo complete neurologic and neuroradiologic evaluations. Only when these have been performed, can one consider appropriate therapy.