Aspiration Subsequent to a Pure Medullary Infarction: Lesion Sites, Clinical Variables, and Outcome

Background  Aspiration as a symptom of dysphagia and its apparent sequela, aspiration pneumonia, are common consequences of a stroke in the medulla. Previous reports that focused on dysphagia due to a medullary lesion were studies of single cases or a relatively small number of patients with multiple lesion loci. Moreover, the interval between the onset of stroke and the evaluation time of swallowing was not controlled and varied largely among patients. Thus, prediction of the swallowing status of patients with a medullary lesion has not been tenable.

Objectives  To investigate the relation between the loci of pure medullary lesions and aspiration, to examine swallowing function over time, and to explore clinical variables that can predict aspiration.

Methods  We investigated 23 patients with pure medullary infarctions using the videofluoroscopic swallowing study and compared the airway status findings with the lesion location as determined with magnetic resonance imaging. The patients were classified by 6 medullary lesion–level categories (lower, lower-middle, middle, lower-middle-upper, middle-upper, and upper) and by 5 intralevel lesion loci (inferior-dorsal, large inferior-dorsolateral, paramedian, midlateral, and dorsolateral). From the results of the videofluoroscopic swallowing studies, 2 patient groups were formed: one with aspiration and the other without aspiration. The clinical variables related to aspiration and outcome measures were also explored.

Results  Ten (44%) of the 23 patients manifested aspiration on swallowing: 9 (69%) of 13 with only middle-level lesions or lesions in multilevels, including the middle level; 1 (33%) of 3 with only upper-level lesions; and 0 (0%) of 7 with only lower-level medullary lesions. A lesion running the length of the middle and the lower medullary levels always resulted in aspiration. When an upper-level lesion was additionally involved, the incidence of aspiration depended on the horizontal extension of the lesion. We were able to discriminate the 2 patient groups with 95.7% accuracy using such variables as dysphonia, soft palate dysfunction, and facial hypesthesia. Most of the patients with aspiration symptoms due to a pure medullary infarction recovered rather quickly.

Conclusions  Medullary infarctions often cause aspiration, but the occurrence may depend on the levels along the neuraxis and intralevel lesion loci. When different lesion levels and loci and their related clinical findings are considered as possible variables, aspiration becomes predictable. The outcome data prove that systematic control of evaluation time of swallowing was critical as we engaged in this study, since many aspirators with pure medullary infarctions resolve their swallowing difficulties rather quickly.

THE ROLE of the medulla as a swallowing center has been well reported in numerous animal studies.^1-4 In fact, the studies have identified several internal anatomical regions within the medulla that are responsible for swallowing functions. They include a dorsal region consisting of the nucleus of the solitary tract and the middle to ventral regions corresponding to the reticular formation surrounding the nucleus ambiguus.

Although dysphagia is a common consequence of stroke in the medulla in general, the lesion location within the medulla seems to be critical as a determinant of occurrence of dysphagia. For example, lateral medullary syndrome is often documented as frequently related to dysphagia. From 51% to 100% of the patients with Wallenberg lateral medullary syndrome have been noted to experience some degree of swallowing difficulties.^5-7 On the contrary, a medial medullary lesion has been reported as being not common. To our knowledge, there is only one study⁸ that specifically commented on dysphagia related to that lesion. Kim et al⁸ noted that 2 (11%) of their 18 patients with a medial medullary lesion demonstrated dysphagia.

Dysphagia may be manifested as aspiration, defined as the tracheal invasion of foreign material or saliva below the true vocal folds.⁹ Aspiration itself can be seriously detrimental to patients who experience a stroke because of its apparent sequela, aspiration pneumonia. The underlying neurophysiological causes of aspiration are multifaceted, given that not only reduced functions in the oral mechanism but pharyngolaryngeal abnormalities also may result in tracheal invasion by foreign material under the vocal folds. Thus, it is of interest to investigate the neurologic lesion loci that relate to the dysfunction of the oral and pharyngolaryngeal swallowing mechanisms associated with aspiration.

Previous clinical studies that mainly focused on dysphagia secondary to medullary lesions were conducted with some methodological limitations. First, they were studies^10,11 of single cases or a relatively small number of patients with multiple lesion loci. For example, a study by Horner et al¹¹ included only 1 patient with a pure medullary lesion; the rest of the patients had multiple lesions, including lesions in the pons, mesencephalon, or thalamus. Studies of more focused topographical information may aid us in elucidating specific lesion sites associated with aspiration as a symptom of dysphagia. Second, the interval between the onset of stroke and the evaluation time of swallowing was not controlled and, as a consequence, it varied largely among patients. Since dysphagia often resolves as physiological changes occur after stroke, it is important that the time of evaluation within the poststroke interval be consistent among patients.

This study (1) investigates the relation between the loci of pure medullary lesions and dysphagia symptoms (ie, aspiration), (2) examines swallowing function over time, and (3) explores clinical variables that can predict aspiration in patients with a medullary lesion. Improved understanding of these relations will increase knowledge of neuropathophysiological features as they relate to poststroke outcomes, thereby facilitating more efficient use of health resources for this patient group.

We studied 23 consecutive patients with pure medullary infarctions. They were admitted to the Department of Neurology, Samsung Medical Center, Seoul, Korea, between January 3, 1995, and December 24, 1997. We excluded any patient with concomitant lacunae or infarctions in nonmedullary regions. One patient with an old lesion in the medulla was included. As shown in Table 1, the subjects consisted of 17 men and 6 women (age range, 27-89 years; mean, 62.6 years; SD, 14.9 years). They all were free of dementia as measured by the Korean Mini-Mental State Examination. Neurologic evaluations, bedside or clinical swallowing examinations, and videofluoroscopic swallowing (VFS) studies were conducted on them within 2 weeks after the onset of stroke.

The lesion location was determined by 2 neurologists (C.-S.C. and K.-H.L.) based on brain magnetic resonance imaging scan results (MR/GE–Signa, Milwaukee, Wis). The 2 neurologists were blinded to the clinical information and swallowing patterns of each patient. Three axial sections of the medulla were identified as the lower, the middle, and the upper levels.¹² The lower level with a closed fourth ventricle has a rather round shape. The middle level is the middle inferior olivary level. And, the upper level designates the pontobulbar junction. In Figure 1, "a" sections illustrate schematic drawings of the lower, middle, and upper levels of the medulla.¹³ In "b" sections of Figure 1, axial cuts were further classified into intralevel lesion loci, as modifications to the study by Sacco et al.⁶ At the lower level, inferior-dorsal and inferior-dorsolateral (IDL) lesions were identified. At the middle level, we classified paramedian (PM), midlateral (ML), dorsolateral (DL), and IDL lesions. At the upper level, PM and ML lesions were identified. A PM lesion can be triggered by the occlusion of penetrating arterioles arising from the anterior spinal artery, the upper vertebral artery, and/or the basilar junction.¹⁴ The PM area includes the medial lemniscus, and the nucleus of the hypoglossal nerve may be involved. An inferior-dorsal, ML, DL, or IDL lesion results from the occlusion of the vertebral artery or the posterior inferior cerebellar artery. The inferior-dorsal area is small and confined to the most dorsal area of the lower medulla. An ML lesion involves a notchlike diagonally oriented band in the retro-olivary medulla that may include the nucleus ambiguus. The DL area is dorsal to the ML area, and the 2 areas can be partially overlapped. The IDL area is an extensive region that may include numerous nuclei for swallowing functions, such as the nucleus ambiguus and the solitary nucleus.

Ten neurofunctional variables were assessed. These included facial palsy, facial hypesthesia, limb motor dysfunction, limb sensory dysfunction, soft palate dysfunction, absent gag reflex, gait ataxia, and cerebellar dysfunction. In addition, the presence of dysarthria as an articulatory impairment was rated from 0 to 4 (0 being normal; 4, severe).¹⁵ The presence of dysphonia characterized by breathy voice quality was also rated from 0 to 4 (0 being normal; 4, severe).

Each patient was situated in a commercially available chair (VESS Chair, Inc, West Allis, Wis) for the VFS studies. The protocol was to administer 3 mL and 9 mL of thin liquid barium to the patient. The instruction given to each patient was as follows, "Hold the barium until you are told to swallow." As the barium was fed, the lateral view from the anterior nasal spine to the cervical vertebrae was obtained. We also recorded the anteroposterior plane to observe asymmetrical functions of pharyngolaryngeal structures. Furthermore, we investigated the bolus-dependent outcome (aspiration) because it was thought to be clinically significant.

Among our 23 patients, there were 12 with lesions in the left medulla, 9 in the right, and 2 bilaterally. Approximately half of the 23 subjects (10 [44%]) exhibited aspiration. The mean postonset days at the time of the VFS studies for aspirators and nonaspirators were 4.9 and 7.7 days, respectively. The postonset days between the 2 groups showed no statistically significant difference.

As shown in Figure 2, seven (30%) of our 23 patients (patients 1-7) showed lesions only in the lower medullary level. In this subgroup, none showed aspiration. There were 6 patients (26%) (patients 8-13) with combined lower-middle lesions, and all of them were aspirators. One patient (4%) (patient 14) with a lesion confined to only the middle level showed aspiration. Four patients (17%) (patients 17-20) had lesions involving the middle and the upper medulla. Half of the 4 patients in this subgroup presented with aspiration. Among the 3 patients (13%) (patients 21-23) with only the upper lesions, 1 exhibited aspiration. Finally, 2 patients (9%) (patients 15 and 16) with lesions in all 3 (lower-middle-upper) levels had normal swallowing functions.

At the lower medullary level, none of the patients showed aspiration regardless of lesion loci.

All 6 patients with the lower-middle–level lesions were aspirators.

Patient 14 with an IDL lesion only at the middle level manifested aspiration. Two patients (patients 15 and 16) with lesions in the lower-middle-upper levels did not have swallowing dysfunction, and the lesion loci were PM. Among the 4 patients with the middle-upper–level lesions, one (patient 17) with a unilateral PM lesion and another (patient 18) with a bilateral PM lesion showed aspiration; 2 (patients 19 and 20) with a unilateral ML lesion showed no aspiration.

Two (patients 21 and 22) of 3 patients with only a unilateral upper ML lesion manifested no aspiration. The remaining one (patient 23) with a bilateral ML lesion was an aspirator.

Stepwise discriminative analysis revealed that among the 10 variables obtained from the neurologic and bedside swallowing examinations, there were 3 that tend to predict aspiration in patients with pure medullary infarcts. They included dysphonia (F=35.92, Wilks λ=0.37), facial hypesthesia (F=25.83, Wilks λ=0.28), and soft palate motor dysfunction (F=26.20, Wilks λ=0.20). As shown in Table 2, using the canonical discriminant functions, 95.7% of the original grouped subjects were correctly classified between the 2 groups (nonaspirators vs aspirators).

Table 3 shows the outcome information for each patient. Before the initial VFS study, 12 patients were oral eaters. The remaining 11 patients had nasogastric (NG) tubes inserted before the initial VFS study. Three (patients 7, 18, and 21) of the 11 patients with NG tubes resumed oral feeding immediately after the VFS study. Patients 7 and 21 were taught to implement such safety measures as chin tuck during swallowing, and were also placed on a diet modification program for a week as a precaution to the risk of aspiration.⁹ Although the 2 patients (patients 7 and 21) did not show aspiration, airway penetrations of barium were observed during the VFS study. Patient 18 with aspiration refused to continue the NG tube feeding mode and demanded oral intake. In our attempt to compromise, we placed her on postural and diet modification programs.

Among the remaining 8 patients with NG tubes inserted initially, patient 21 remained on NG tube feeding for the shortest duration (ie, 11 days). The longest duration of NG tube management was for patient 23. Follow-up was maintained for patient 23 for 110 days, during which he was continuously fed through an NG tube. However, the follow-up was terminated after he was transferred to another hospital. Patient 11 received NG tube feeding for 51 days until complete oral feeding became possible. All others were moved to safe oral feeding mode within 1 month without requiring further diet and/or postural modifications. One additional follow-up VFS study was done on every patient who manifested aspiration except for patient 8 (2 studies) before terminating the dysphagia treatment program.

Ten (44%) of our 23 patients with pure medullary lesions in this study manifested aspiration. The rate converges with those of 2 previous studies by Kim and coworkers^7,8; one study⁷ assessed 33 patients with lateral medullary syndrome, and the other⁸ assessed 18 patients with medial medullary syndrome. While 22 (43%) of their combined 51 patients manifested dysphagia, the definition of dysphagia was not clear. Dysphagia can be inclusive of aspiration and other dysphagic symptoms, such as residue in the pharynx or (airway) penetration, defined as a situation where the foreign materials enter the airway into the laryngeal vestibule, above or to the level of the vocal folds.⁹ When the 3 patients (patients 8, 20, and 21) with penetration are taken into account, the occurrence of "dysphagia" becomes as high as 56.5%.

From the results, it has been perceived that the incidence of aspiration may vary depending on medullary lesion levels and loci containing various cranial nuclei and tracts. When a patient has an inferior-dorsal lesion at the lower level of the medulla, the patient rarely manifests aspiration. This can be explained by the fact that the dorsal region of the lower medulla does not contain nuclei that are critical for swallowing function.

On the other hand, a PM lesion either in the lower or in the middle medulla damages the medially situated nucleus of the hypoglossal nerve, which is responsible for tongue functions. Tongue dysfunctions are primarily manifested as difficulty in manipulating or maintaining the bolus within the oral cavity and may result in premature spillage of boluses into the pharyngeal cavity and unprotected larynx, and in turn, aspiration. If a patient has either a unilaterally extensive or a bilateral PM lesion, the patient may become an aspirator. But, a patient with a narrow unilateral PM lesion can be aspiration free because the lesion does not invade the dorsal-medially located hypoglossal nucleus in the middle medulla. In a previous study⁸ of medial medullary infarction, dysphagia was observed in only 2 of 18 patients. One had a lower-middle-upper lesion bilaterally, and the other had an extensive unilateral middle-upper lesion, as in our patient 17.

Aspiration becomes rather apparent in vertically extensive IDL plus IDL or IDL plus DL lesion cases. Often, those lesions invade several structures that are significant for swallowing. For example, when the spinal trigeminal nucleus and the tract located DL in the lower and the middle medulla are damaged, a patient may become less sensitive to mucosal sensory information from various oral structures, such as the floor of the mouth, the gum, the tongue, and the palate. As a consequence, a bolus falls into the pharynx without triggering the normally automatic swallowing response, and the risk of aspiration increases. Reduced sensitivities in such pharyngolaryngeal areas as the base of the tongue and epiglottis due to a lesion at the nucleus of the solitary tract may also entail bolus overflow that leads to aspiration. The decreased sensory input of oral and pharyngolaryngeal areas has been observed in many patients who experienced a stroke.¹⁶

In addition to aspiration being generated as a result of sensitivity reduction, impaired motor function is also responsible for aspiration. The nervous system entities that are related to motor function are the nucleus ambiguus and the vagus nerve. They control the striated muscle movements of the soft palate, the base of the tongue, the pharynx, and the larynx. They can be invaded by the IDL plus IDL or the IDL plus DL lesion. The soft palate along with the base of the tongue serves as a piston to generate the intrabolus pressures that propel the bolus through the pharynx (not only over the vallecular spaces but also through the upper esophageal sphincter). Furthermore, pharyngeal constrictor paresis is often viewed in the context of disruption to the medullary swallowing center. When it induces pharyngeal motility disorders, it is a frequent cause of aspiration due to residue overflowing in the pharynx.^17,18 Likewise, the valving dysfunction of laryngeal structures, especially the vocal fold palsy, may cause aspiration.

An ML lesion in either the middle or the upper level was not detrimental to swallowing unless the lesion was bilateral, as in patient 23. The lesion extent is narrow and may reserve some areas of swallowing nuclei, such as the nucleus ambiguus and the solitary tract nucleus.

Three clinical variables were found in this study to be important in discriminating the patient groups, ie, aspirators vs nonaspirators. First, the presence of dysphonia was closely related to aspiration. When a patient manifests breathy quality of voice due to vocal fold palsy, we might be able to speculate a lesion to the nucleus ambiguus. In fact, all of our patients with aspiration had dysphonia. Second, soft palate dysfunction was also a predictor of aspiration. As previously mentioned, the motility of the levator veli palatini muscle for the soft palate is under the control of the vagal nerve and the nucleus ambiguus and is frequently compromised in the aspirators with a medullary lesion. Third, when a patient had a symptom of facial hypesthesia, the patient often demonstrated aspiration. As discussed earlier, reduced sensation of the mandibular-oral area is related to a lesion at the nucleus of the spinal trigeminal tract.

Overall, recovery from aspiration due to a pure medullary infarction occurs and does not seem to take an extended period. Most of the patients who initially started with NG tube feeding were able to convert to full oral feeding within a 2-month period after the onset. Only 1 patient, patient 23, needed a relatively extended period of recovery time. It might be ascribed to his experiencing a second stroke in the medulla at the time of his dysphagia evaluation, and this 2-stroke incident might have had a cumulative, detrimental effect on dysphagia. It was reported that the patient did not complain about dysphagia after his first stroke. Since the neural control of swallowing has bilateral representation, it is likely that more impaired swallowing patterns are observed secondary to bilateral lesions of the medulla.¹⁹

Contrary to the common perception, our findings reveal that a pure medullary lesion does not result in long-lasting dysphagia or aspiration. We speculate that patients described in the literature who showed limited recovery from dysphagia must have had cortical-subcortical lesions concomitant to brainstem or multiple brainstem lesions. In those with aspiration as manifested with the VFS studies, we had to implement diet and postural modifications to prevent them from developing aspiration pneumonia. In fact, none of our patients in the treatment program had pneumonia.

In summary, a medullary infarction often causes aspiration, but the occurrence may depend on the levels along the neuraxis and intralevel lesion loci. When different lesion levels and loci and their related clinical findings are considered as possible variables, aspiration becomes predictable. Such predictability may facilitate early intervention and result in improved patient outcomes, including health status or quality of life. Future studies may address these related issues. The outcome data from this investigation indicate that systematic control of evaluation time of swallowing was critical to demonstrate our finding that many aspirators with a pure medullary infarction resolve the swallowing difficulties rather quickly.

Accepted for publication November 9, 1999.

Reprints: Hyanghee Kim, PhD, Department of Neurology, Samsung Medical Center, 50 ILwon-dong, Kangnam-ku, Seoul 135-710, Korea (e-mail: hkim26@smc.samsung.co.kr).