Lhermitte sign and urinary retention

Atypical presentation of oxaliplatin neurotoxicity in four patients




Regimens combining oxaliplatin with fluorouracil and folinic acid are standard therapeutic options for patients with metastatic colorectal carcinoma. Oxaliplatin has a good safety profile, although it is responsible for dose-limiting neurotoxicity typically consisting of two distinct clusters of symptoms. Cold-induced distal paresthesiae occur during or shortly after infusion in most patients and are usually transient and mild. A persistent sensory peripheral neuropathy may develop with prolonged treatment, eventually causing superficial and deep sensory loss, sensory ataxia and functional impairment.


The authors report four cases of atypical neurotoxicity induced by oxaliplatin in patients treated for metastatic colorectal carcinoma. Two patients were male and two were female, with an age range of 52–59 years.


Three patients experienced Lhermitte sign and two experienced urinary retention. In all cases, the cumulative dose of oxaliplatin was higher than 1000 mg (range, 1248–2040 mg). Brain and spinal magnetic resonance imaging was performed in two patients and was normal. Somatosensory evoked potentials were recorded in two patients and suggested cervical dorsal column dysfunction. Symptoms resolved a few weeks after discontinuation of oxaliplatin.


Lhermitte sign may be induced via a neurotoxic effect on the ascending axons of these T-shaped neurons. An atonic bladder may be the result of damage to the sensory portion of the sacral reflex arc, either in the dorsal roots, as for example in diabetic neuropathy, or in the posterior columns, as in tabes dorsalis. Alternatively, it may result from a paralysis of the parasympathetic fibers that control the bladder musculature. It is unclear at present whether the micturition difficulties observed in patients in the current series are due to sensory neuropathy or to autonomic neuropathy, event if the former hypothesis seems more likely, as autonomic neuropathy has not been previously observed with oxaliplatin, and its association with cisplatin is exceedingly rare and controversial. Cancer 2002;94:2434–40. © 2002 American Cancer Society.

DOI 10.1002/cncr.10500

In patients with metastatic colorectal carcinoma (CRC), chemotherapy has been shown to be effective for both symptom palliation and survival.1, 2 Combinations of high dose continuous 5 fluorouracil (5FU) and folinic acid are among the most active regimens. More recently, the third generation platinate oxaliplatin has shown significant activity, both as a single drug3–6 and in combination with fluorouracil.7, 8 Oxaliplatin containing regimens are now considered standard therapeutic options for metastatic CRC patients.9–11

Oxaliplatin has a very good safety profile, with low hematotoxicity, moderate, manageable gastrointestinal toxicity, and an absence of nephrotoxicity.12–15

The dose-limiting toxicity of oxaliplatin is neurosensory toxicity. In around 90% of patients oxaliplatin is associated with acute dysesthesiae and paresthesiae affecting the fingers, toes, perioral and oral region and pharyngolaryngeal tract, generally induced or aggravated by cold.12, 13, 15 These symptoms may appear at the first treatment cycles; they usually occur during or shortly after oxaliplatin infusion, are generally mild, and disappear within a few hours or days. In some patients, acute sensory symptoms can be accompanied by muscular contractions variously described as cramps or spasms. At high cumulative doses of oxaliplatin, a persistent sensory neuropathy can develop, with continuous, noncold related paresthesiae, superficial and deep sensory loss, and eventually sensory ataxia and functional impairment. A Grade 3 neuropathy12, 13, 15 has been reported in 15–18% of patients after cumulative doses of 780–850 mg/m2. Symptoms usually resolve a few weeks or months after oxaliplatin discontinuation.8, 16

Compared with the first generation platinate, cisplatin, the neurotoxicity profile of oxaliplatin shows some similarities and some differences.17– 19

Cisplatin neuropathy has been reported to affect around 50% of patients; it is typically cumulative and delayed, tends to worsen in the posttreatment period, and is usually irreversible.17–22 Severely affected patients develop disabling sensory ataxia, while motor neuropathy is absent or minimal. Audiometry shows cisplatin-induced ototoxicity in 75–100% of patients, which may be associated with tinnitus and hearing loss, most prominently in the high frequencies.17, 20, 23 More rarely, Lhermitte sign, an electric-like sensation induced by flexion of the neck, is present.24–27 Autonomic neuropathy has been occasionally reported but seems to be controversial.28 The central nervous system (CNS) appears to be spared, possibly because of the poor passage of platinates through the blood barrier;17 occasional reports of seizures in association with cisplatin29, 30 are difficult to interpret, as seizures may be induced indirectly, for example through electrolyte disturbances. Alternatively, it may be possible that cisplatin, a heavy metal compound, may occasionally achieve high CNS levels and cause neurotoxicity in a manner similar to that seen with other heavy metals.17, 29 Neurophysiologic investigations have shown reduced sensory nerve conduction velocity, prolongation or disappearance of sensory nerve latencies, and an absence or reduction in amplitude of sensory nerve potentials.17

As with cisplatin, the cumulative neurotoxicity observed with oxaliplatin presents as a sensory neuropathy sparing the motor neurons.17, 31 However, in contrast with cisplatin, oxaliplatin is not ototoxic. To our knowledge, Lhermitte sign has been reported only once in the literature,15 and there are no published reports of autonomic neuropathy or of involvement of the CNS in association with oxaliplatin. Further, and more impressively, the acute neurotoxicity typically observed with oxaliplatin has distinct characteristics and shows no resemblance to cisplatin. Indeed, cisplatin neurotoxicity has never been described as possibly induced by cold or as being accompanied by muscular contractions.

The pathophysiology of oxaliplatin neurotoxicity has been recently reviewed.31 There is some experimental evidence from various animal models that the acute neurosensory symptoms that occur during oxaliplatin administration may be linked to an increased neuronal excitability induced by a specific action of oxaliplatin on some isoforms of voltage-gated sodium channels located on the cellular membrane of sensory neurons.31, 32, 33 It is unclear at present whether the acute and chronic neurotoxicity observed with oxaliplatin is due to a common cellular mechanism. The cumulative sensory neuropathy that develops after chronic treatment has been well characterized by decreased nerve conduction in patients and in animals.31, 34 Experimental studies in rodents indicate that the dorsal root ganglia containing the cell bodies of sensory neurons are the most vulnerable structures, possibly because they are not protected by the blood-brain barrier.17, 31

We present four case reports of oxaliplatin-associated neurotoxicity presenting with atypical neurosensory symptoms (Lhermitte sign and micturition difficulties) and discuss their possible physiopathology in view of the current knowledge of the neurotoxicity of platinum salts.


The main clinical characteristics, symptoms, and treatments of the four patients are summarized in Table 1. The patients included 2 males and 2 females, aged 50–59 years, all with metastatic colorectal carcinoma. The primary tumor and the metastases were diagnosed synchronously in all cases.

Table 1. Characteristics of Patients, Symptoms, and Treatments
Patient no.1234
  1. F: female; M: male; FOLFOX: folonic acid and oxaliplatin; PR: partial response; MRI: magnetic resonance imaging; SEP: scmatosensory evoked potentials; Pt: patient.

Age at diagnosis (yrs)59505257
Primary tumor siteRight colonSigmoidSigmoidRectum
Metastatic sitesLiverLiverLiver and peritoneumLiver and lung
TNM stageT3N2M1T3N0M1T4N2M1TxNxM1
Response status
After 4 cycles of FOLFOXPRPRPRPR
Total number of FOLFOX cycles991212
Cumulative oxaliplatin dose (mg/sqm)1248146520401596
Reason for oxaliplatin withdrawalMicturition difficulties, perineal hypoesthesiaLhermitte sign, sensory ataxiaLhermitte sign, micturition difficulties, sensory ataxiaLhermitte sign, sensory ataxia
MRINormalNormalNot performedNot performed
SEPNot performedDorsal column dysfunctionNot performedDorsal column dysfunction
Followup after oxaliplatin withdrawalPt died after 3 monthsPt alive 3 years laterPt died after 6 monthsPt alive 1 year later

All patients received the same chemotherapy regimen, containing 5FU bolus 800 mg/ sqm and continuous infusion 1200 mg/sqm, L folinic acid 200 mg/sqm (LV5FU2), and oxaliplatin 85 mg/sqm (FOLFOX regimen) as a 48 hour bimonthly regimen, as previously described by de Gramont et al.8

Case Report 1

A 59 year old woman was treated for metastatic colon carcinoma. At the beginning of chemotherapy, she experienced dysesthesiae in the fingers and toes, induced by cold. The symptoms disappeared approximately six days after the end of each chemotherapy cycle. Moderate but continuous paresthesiae occurred after the seventh cycle. Despite this worsening of neurotoxicity, the same treatment was continued in consideration of the prolonged control of the metastatic disease. After the ninth cycle cancer progression was observed, with fever, weight loss, and a rise of tumor markers. At that time the patient reported a worsening of paresthesiae in the lower limbs, with S1 topography, bilateral hypoesthesiae, and a subjective cardboard feeling in the fingers with C6-C7 topography. Further, she complained of severe micturition difficulties. A neurologic examination revealed perineal hypoesthesia to touch and pinpricks. Patellar and ankle reflexes were hyperactive with a Babinski's sign on the right side. The other deep tendon reflexes and the left plantar stimulation were normal. There was no motor deficit. Cranial nerves, cerebellar tests, and cognition were normal.

To assess the possible presence of multifocal myeloradiculitis associated with carcinoma progression, magnetic resonance imaging of the spinal cord with and without gadolinium was performed; it was found to be normal. Lumbar puncture showed cerebrospinal fluid with normal cellular, protein, and glucose content. No evidence of central or peripheral nervous system carcinoma invasion was found.

Oxaliplatin was discontinued, and the patient received a further treatment cycle with a combination of high dose continuous 5 FU and folinic acid.

The patient was seen two months after oxaliplatin discontinuation. Neurologic signs and symptoms had consistently improved. There were no more pyramidal signs or micturition difficulties. Only a sensory deficit with S1 topography persisted. The patient died one month later. The neurologic findings did not appear to be tumor related. Other potential causes of neuropathy, such as alcohol abuse, diabetes, or metabolic dysfunction, were excluded. This unusual neurologic symptomatology was considered to be likely due to oxaliplatin neurotoxicity.

Case Report 2

During the first three cycles of FOLFOX, this 50 year old man experienced only transient, cold induced dysesthesiae. After the fourth cycle, moderate but continuous paresthesiae occurred. It was decided to continue the treatment, as the disease was well controlled. After nine cycles the patient described electric-like paresthesiae in both upper limbs provoked by cervical spine anteflexion (Lhermitte sign). He also complained of spontaneous distal paresthesiae and dysesthesiae in the fingers and the lower limbs. Deep tendon reflexes were absent at the knees and the ankles. Vibratory sensation was decreased in the feet, and touch sensation was slightly decreased in the fingers. Motor function, cranial nerves, cerebellar tests, and cognition were normal. Primary carcinoma as well as metastases remained well controlled.

Under these circumstances oxaliplatin was discontinued, and the patient received a 5FU and folinic acid-based chemotherapy regimen. During the following month, Lhermitte sign was still present, and the other neurologic symptoms worsened, with the occurrence of sensory ataxia. Brain and spinal cervical magnetic resonance imaging (MRI) was normal; the patient refused lumbar puncture. Median nerves somatosensory evoked potentials showed a decrease of the peripheral Erb's point (N9) responses amplitude with an absence of the P14 cervico-bulbar responses and delayed N20 cortical responses on each side. This pattern is consistent with peripheral and dorsal column dysfunction. Electromyography was not performed in this patient.

Five months later, Lhermitte sign disappeared and sensory ataxia improved, but paresthesiae and ankle deep tendon reflexes abolition persisted. The neurologic symptoms were regarded as due to oxaliplatin neurotoxicity.

A mediastinal irradiation was then performed, delivering a total dose of 50 Gy without major toxicity. Fifteen months after chemotherapy discontinuation, progression of liver involvement was shown by computed tomography (CT) scan, and the patient received irinotecan in combination with 5FU and folinic acid. A complete response was obtained. The patient is still alive one year later with no sign of evolutive disease.

Case Report 3

During the first cycles of chemotherapy, this 52 year old man experienced only distal and pharyngeal dysesthesiae induced by cold lasting about six days after infusion. After nine cycles, the severity and duration of cold related dysesthesiae increased, with the sensory testing remaining normal. Between the ninth and the tenth cycles, a treatment with oral carbamazepine (600 mg/day) was started; this allowed moderate improvement in the intensity and duration of dysesthesiae. From the eleventh cycle on, spontaneous distal paresthesiae and dysesthesiae appeared in association with decreased pinprick sensations in the fingers and toes. Moreover, the patient complained about severe micturition difficulties. Vibratory and touch sensation, motor testing, and deep tendon reflexes were normal, with the exception of absent ankle reflexes. Due to the evidence of tumor chemosensitivity, the patient received a twelfth chemotherapy cycle. After this last cycle, Lhermitte sign and sensory ataxia occurred, and all previously described neurologic abnormalities worsened as well.

Oxaliplatin was discontinued, and the patient received LV5FU2. Four months after oxaliplatin discontinuation, Lhermitte sign, urinary retention, and sensory ataxia had disappeared, although moderate distal sensory neuropathy was still present. The patient died six months from tumor progression after the administration of five cycles of LV5FU2 and a further third line chemotherapy with irinotecan.

Case Report 4

The patient, a 57 year old woman, experienced transient, cold induced distal and pharyngeal dysesthesiae from the third up to the eleventh FOLFOX cycle. During this period, a good quality tumor response was observed on liver and lung metastases (60% tumor reduction at CT scan). This patient participated in an open prospective study on the neurosensory toxicity of oxaliplatin. A complete clinical neurologic examination associated with median and tibial nerves sensory evoked potentials (SEPs) was performed before the first cycle and every two months from then on. Erb's point (N9), popliteal fossea, cervicobulbar (P14, P30), and cortical (N20, P39) responses were evaluated. These examinations remained normal up to the eleventh chemotherapy cycle. From the twelfth cycle on, the patient experienced moderate but continuous paresthesiae in the fingers, associated with numbness of the toes, sensory ataxia, and Lhermitte sign. Neurologic examination showed a slightly decreased vibratory sensation in the fingers and toes and a significant decrease of bilateral ankle deep tendon reflexes. Pinprick, cold. and touch sensations were normal. Cranial nerves, cerebellar tests. and cognitive functions were normal; SEP recordings showed reduced amplitude and delayed latency of peripheral responses at Erb's point (N9) and popliteal fossea with a delayed N9-P14 interval (used to assess dorsal column function). Ulnar, median, common peroneal, and tibial motor nerve conduction velocities were normal, as were the amplitudes of the evoked compound muscle action potentials and the sensory nerve conduction velocities of median and superficial peroneal nerves. These electrophysiologic findings are consistent with the cervical dorsal column dysfunction indicated by the presence of Lhermitte sign, which was attributed to oxaliplatin neurotoxicity. Oxaliplatin was discontinued and the patient was treated with a combination of continuous high dose 5FU and folinic acid. Three months later, Lhermitte sign disappeared, at the time of writing the patient was still responding to chemotherapy. Moderate distal sensory neuropathy was still present one year later.


The four cases reported in the current article presented with atypical neurologic signs and symptoms after oxaliplatin-containing chemotherapy for metastatic colorectal carcinoma. All patients received the same chemotherapy regimen containing 85 mg/sqm of oxaliplatin in a bimonthly schedule. None of the patients were known to have previous neurologic disease. All patients initially experienced transient, cold related dysesthesiae of the extremities, and after a few cycles of chemotherapy (7, 4, 11, and 12 cycles, respectively), permanent, non cold induced paresthesiae persisted between cycles.

In Patients 1, 2, and 3, oxaliplatin was not discontinued at that point, as normally recommended, in view of the prolonged control of the metastatic cancer. Severe neurotoxicity became apparent following additional chemotherapy cycles at cumulative doses of oxaliplatin of 1248 mg/sqm, 1465 mg/sqm, and 2040 mg/sqm, respectively. In addition to a worsening of the pre-existing peripheral neuropathy, manifested as perineal hypoesthesia (Patient 1) and sensory ataxia (Patients 2 and 3), patients developed unpredicted neurologic symptoms consisting of severe micturition problems (Patients 1 and 3) and Lhermitte sign (Patients 2 and 3).

In Patient 4, persistent paresthesias appeared following the 12th cycle of chemotherapy after a cumulative dose of oxaliplatin of 1596 mg/sqm, at the same time as Lhermitte sign and sensory ataxia.

In two cases (Patients 2 and 4), SEP recordings provided evidence for cervical dorsal column dysfunction.35–37

Following oxaliplatin withdrawal, the severe, atypical neurologic symptoms completely disappeared after 1, 5, 4, and 3 months, respectively. A moderate peripheral sensitive neuropathy with continuous paresthesiae persisted (Levi's scale Grade 2).

In all four patients, neurotoxicity was attributed to oxaliplatin. The presence of metastatic external pachymeningitis was excluded in two patients by normal MRI and cytology examination of the cerebrospinal fluid. Moreover, none of the patients had progressive disease at the time the symptoms occurred.

To the best of our knowledge, Lhermitte sign has been previously reported in association with oxaliplatin only by Brienza et al.15 in a pooled retrospective analysis of the peripheral neurotoxicity due to oxaliplatin from 9 studies on 682 metastatic CRC patients. In that study, the risk of developing severe neurotoxicity was 10% after 6 cycles (780 mg/sqm of oxaliplatin) and 50% after 9 cycles (1170 mg/sqm). A World Health Organization Grade > 2 peripheral neuropathy occurred in 152 patients. After discontinuation of treatment, symptomatic improvement was observed in 82% of patients (median followup, three to four months) and total disappearance of the symptoms in 46% (median followup, six to eight months). Somatosensory evoked potentials were recorded in 35 patients and showed evidence of sensory neuropathy. Two patients were reported as having Lhermitte sign. The physiopathology of this symptom was not discussed.

Lhermitte sign, also known as the barber's shop sign, should be more properly designated a symptom than a sign. This characteristic, electric-like sensation induced by forward flexion of the head was originally described by Marie and Chatelin in 1917 in a report on neurologic observations in soldiers wounded in World War 1 and was attributed to “minimal radicular lesions” following head or neck trauma.38 One year later, Babinski and Dubois described four additional cases in patients with neck trauma and discussed the sign's possible radicular or spinal origin.39 The importance of this phenomenon was brought to the attention of the neurologic world in a landmark paper by Lhermitte et al. in 1924 on a case of multiple sclerosis.40 According to that paper, the shock-like sensations are due to spinal lesions and are linked to demyelinating or traumatic lesions of the dorsal columns. Since that time, the presence of Lhermitte sign has been generally associated with the presence of spinal lesions and particularly with the demyelinating lesions of multiple sclerosis.

However, as was already stressed in 1949 by Alajouanine et al.,41 the presence of Lhermittte's sign does not necessary imply a lesion of the spinal cord. While the anatomic support of Lhermitte sign is clearly in the posterior columns, the sign can be due to hyperexcitability of the ascending sensory neurons due to mechanical or toxic causes in the absence of a spinal lesion. Alajouanine described a series of cases of compression of the spine by malformations of the atlas, vertebral malformations or fractures, disc herniation, posttraumatic or postinfectious meningeal adherences, syringomyelia, and tumors. In most of these cases, removal of the compression by surgery was followed by the disappearance of Lhermitte sign.

In the current case series, a likely hypothesis is that Lhermitte sign can be attributed to oxaliplatin-induced hyperexcitability of the large ascending sensory fibers, whose cell bodies are in the dorsal root ganglia.

This hypothesis is supported by the appearance of Lhermitte sign after high cumulative doses of oxaliplatin, in parallel with the development of a persistent sensory neuropathy with sensory ataxia, and by the reversibility of the symptom after oxaliplatin withdrawal; it is consistent with the observations obtained through sensory evoked potentials, pointing to involvement of the dorsal columns.

Two patients in the current case series experienced severe micturition difficulties, an adverse event that has not been previously reported in association with oxaliplatin.

An atonic bladder may be the result of damage to the sensory portion of the sacral reflex arc, either in the dorsal roots, as in diabetic neuropathy, or in the posterior colums, as in tabes dorsalis. Alternatively, it may result from a paralysis of the parasympathetic fibers that control the bladder musculature. The pelvic splanchnic nerves that originate in the sacral cord (S2–S4) terminate at ganglia in the wall of the bladder and in the internal sphincter muscle. Parasympathetic stimulation induces emptying of the bladder via the postganglionic parasympathetic neurons.

It is unclear at present whether the micturition difficulties observed in our patients are due to sensory neuropathy or to autonomic neuropathy, even if the former hypothesis seems more likely, as autonomic neuropathy has not been previously reported with oxaliplatin, and its association with cisplatin is rare and controversial.28

Both dorsal root ganglia and post-ganglionic autonomic neurons are not protected by the blood-brain barrier, which prevents entry of platinates into the central nervous system.

In Patient 1 of the current series, a unilateral Babinski's sign was recorded concurrently with the appearance of micturition difficulties and sensory deficit at the extremities and in the perineal area. While the exact neural mechanism of Babinski's sign is poorly understood, this sign is always associated with pyramidal (cortico-spinal) tract injuries, and its presence in our patient remains unexplained.


At cumulative doses higher than 1000 mg/sqm, oxaliplatin may occasionally induce severe, atypical neurotoxic symptoms, such as micturition disturbances and Lhermitte sign, that may mimic spinal cord disease.

In three of our four patients, severe neurotoxicity was preceded by paresthesias persisting between cycles. In these cases, a reduction in the dose of oxaliplatin may have prevented the occurrence of severe neurotoxicity; however, in the last case, the worsening of neuropathy occurred without previous aggravation and would have been impossible to prevent.

In the current cases, Lhermitte sign and urinary retention were reversible after oxaliplatin discontinuation.

Since SEP recording may allow detection of infraclinical symptoms of posterior columns toxicity, we have recently initiated a prospective evaluation of the potential interest of this technique to improve the clinical followup of patients receiving oxaliplatin chemotherapy for CRC.