Chronic Cough as the Presenting Symptom of Hydrocephalus

Selmin Karatayli-Ozgursoy, MD; Jacob Dominik, MD; Benjamin Eidelman, MD; Juan C. Guarderas, MD


South Med J. 2010;103(6):574-577. 

In This Article


Excluding common causes of chronic cough should always be the first objective, although it may require the collaboration of experts in many disciplines. In this case, clinical evaluation uncovered signs indicating a neurological disorder, but the cough was determined to be of neurogenic origin. This assessment was made only after exhaustive investigations failed to reveal a systemic cause and, in particular, after a positive response to treatment of the hydrocephalus was observed. Thus, it is important to review the central mechanisms of cough genesis with particular reference to hydrocephalus as an etiological condition in neurogenic cough.

Coughing is typically thought of as a protective reflex triggered by several inflammatory or mechanical changes, and/or inhalation of chemical and mechanical irritants to the airways. Sensory nerve receptors responding to these stimuli are defined regarding their conductive properties as rapidly adapting receptors, slowly adapting receptors, or C-fibre receptors. Afferent nerve activation in the external auditory meatus, pharynx and esophagus can also initiate cough.[2] Afferent nerve fibers from cough receptors converge via the vagus nerve in the nucleus tractus solitarius (NTS), and the NTS is connected to the neurons in the central respiratory generator which coordinates the efferent cough response.[3] In this model, the cough reflex is produced by a convergent, but plastic, brainstem neural network.

Circumstantial evidence suggests that coughing in humans involves higher brain regulation over the basic reflex network at the brainstem level.[4] Coughing can be initiated voluntarily.[5] Placebo-induced suppression of cough,[6] or cough suppression during sleep or under general anesthesia[7–9] might be related to the modulation of cortical control. An enhanced cough reflex has been reported in central nervous system disorders, including cerebellar ataxia and motor neuron disease.[1,10] Chronic cough can also be a presenting symptom of a type-I Arnold-Chiari malformation. Kramer et al[11,13] described two cases of cough in Arnold-Chiari I; however, in these cases, the cough precipitated other neurologic symptoms such as numbness, weakness, or cervical pain, presumably via transiently-increased intracranial pressure. In these cases, it is not clear whether the cough was a consequence of the Chiari or whether it resolved after surgical decompression. In this patient, the MRI demonstrated mildly low-lying cerebellar tonsils and a minimally-crowded posterior fossa (Fig. 1). It is conceivable that this may have accounted for the patient's chronic cough. The relief of hydrocephalus by shunting could have reduced downward pressure on the posterior fossa, diminishing medullary compression and improving the cough. However, our patient denied any neck pain or headaches, alone or associated with the cough, which one might expect if this was indeed the mechanism.[11] Also, lumbar puncture in this instance might increase tonsillar displacement which should make the cough worse, but it did not. The patient did have lumbar punctures performed prior to presentation at this institution. The combination of obstructive hydrocephalus and a low-pressure state due to lumbar puncture could have accounted for the mildly low-lying cerebellar tonsils. The cough, however, had been present for almost one year before the lumbar punctures. That may suggest that the lumbar punctures (LP) were the cause of the mild Chiari-like appearance, but this was not the mechanism responsible for the chronic cough.

Neurologic disorders have also rarely been noted to present with chronic cough as the initial symptom. Diffusion-weighted imaging (DWI) changes in the putamen, caudate, and surrounding cortical areas led them to speculate whether these areas could have some role in the generation of the chronic cough.[12] In this patient, the hydrocephalus with enlargement of the lateral and third ventricles affected periventricular subcortical motor fibers, causing a mild upper-motor-neuron pattern, bilateral lower-extremity weakness, hyperreflexia, and gait apraxia. Local compression of subcortical fibers and interruption of cortical-basal-ganglia connections by ventricular expansion may be involved in the generation of our patient's chronic cough. Although speculative, it is remarkable that previous patients with chronic cough as the presenting symptom of neurologic disease demonstrate known pathology in the same subcortical regions.[12,13]Additional evidence suggesting a role for a cortically-based cough center arises out of reports of impairment of voluntary cough responses in patients with strokes involving the frontoparietal region[15,16] in Parkinson disease,[17] as well as multiple system atrophy.[21]

On initial presentation, it can be difficult to differentiate primary aqueductal stenosis (AS) from normal pressure hydrocephalus (NPH).[19] Tisell compared the clinical presentations of AS and NPH and found no difference in the presence of gait disturbances, cognitive function, or urinary incontinence after adjusting for age. The only significant difference was the presence of headache in the AS patients preoperatively, and greater improvement with surgery.[21] On our initial evaluation, it was considered that the patient's presentation was more consistent with NPH. However, MRI studies revealed narrowing of the cerebral aqueduct and sparing of the fourth ventricle, which is consistent with primary aqueductal stenosis.[20]

Although LP is frequently used as a diagnostic aid in NPH to help determine a patient's likely response to shunting, it is usually contraindicated in cases of obstructive hydrocephalus, in which the trapped ventricular CSF can act like a space-occupying lesion.[22] This patient had already undergone a diagnostic LP at an outside institution and tolerated it well with some transient questionable improvement. Given this fact, as well as the similarity of the clinical presentation to that of NPH, we performed a high-volume LP with video. The patient showed significant improvement in gait after the high-volume LP.

Traditionally, treatment options for AS provide a route for outflow of entrapped CSF performed via a VP shunt. More recently, endoscopic third ventriculostomy (ETV) has become a preferred option in many centers.[19] After the neurosurgery consultation, our patient had a VP shunt. Some anesthetic drugs are known to decrease cough reflex in patients undergoing surgery. Therefore, one might consider that the anesthetic drugs administered at the time of shunt placement may have had some impact on the cough. If so, this effect could have continued only until the end of the half-life of these drugs. However, elimination, or at least significant improvement of the cough, was demonstrated for almost one full year after shunt placement.

Ventricular dilatation in our patient might have caused loss or impairment of higher inhibitory mechanisms of the cough reflex. It is of relevance to examine the possible mechanisms whereby hydrocephalus may have induced chronic cough. In this patient, both the neurological symptoms and cough significantly improved after the operation, indicating that compression rather than irreversible pathological changes were operative in the development of the clinical condition.