Pain as a First Manifestation of Paraneoplastic Neuropathies: A Systematic Review and Meta-Analysis

A systematic computer-based literature search was conducted on April 6th, 2017 on PubMed database. For the search, we used three Medical Subject Headings (MeSH) terms. Term A was “neuronopathy” or “ganglionopathy” or “neuropathy”, term B was “paraneoplastic” and term C was “first manifestation” or “first symptom” or “initial symptom” or “first manifestations” or “first symptoms” or “initial symptoms” or “presenting symptom” or “presenting symptoms” or “pain” or “painful”. Limitations included language to be English and full text to be available. We also perused the reference lists of the papers in order to find papers not found through the search strategy.

To be included in the review, the articles had to meet the following criteria:

The following exclusion criteria were applied:

Data were extracted from each study in a structured coding scene using Microsoft Excel and included information on the article identification, year of publication, evaluation period, total number of subjects, gender, age, first manifestation of the PPN, presence of pain secondary to the neuropathy, neurophysiological type of neuropathy, course of symptoms, type of cancer, and presence of anti-neuronal antibodies.

We used the International Association for the Study of Pain (IASP) definitions to classify pain as neuropathic or not.

This article is based on previously conducted studies and does not involve any new studies of human or animal subjects performed by any of the authors.

A database was developed using the IBM SPSS Statistics (version 23.0 for Mac). Frequencies and descriptive statistics were examined for each variable. The primary outcomes of interest were the proportion of patients who experienced pain as a first manifestation of the PPN and the proportion of patients who experienced pain during their PNS. Further comparisons between groups were made using Student’s t test for continuous data and Chi-square test for categorical data. A value of P < 0.05 was considered to be statistically significant.

This search strategy resulted in the identification of 126 articles. After the eligibility assessment, 81 articles were excluded. In total, 45 papers met the inclusion criteria and were used for this review [3‐47]. These studies were published between 1984 and 2017. Figure 1 illustrates the study selection process.

Full clinical and neurophysiological data were further extracted from 42 papers and involved 92 patients with PPN (54.5% males). The age of the patients ranged from 18 to 81 years (mean 60.0 ± 12.2 years). The demographic and clinical characteristics of these patients are summarized in Table 1.

Although PPN is said to be the commonest PNS (34.9%), followed by encephalomyelitis (23.8%) and cerebellar degeneration (20.6%) [45], there is sparse epidemiological data on the prevalence of neurophysiologically confirmed PPN. The reasons for this are two-fold. The available studies to date are either in cohorts of patients with known malignancy and neuropathic symptoms or in patients who are seropositive for well-characterized paraneoplastic antibodies. A correct prevalence study of PPN should include a large cohort of patients with newly diagnosed malignancy all of whom should be evaluated for the presence of PN with nerve conduction studies, even in the absence of neuropathic symptoms and paraneoplastic antibodies.

Lucchinetti et al. [47] estimated that out of 162 patients seropositive for anti-Hu antibodies (anti-neuronal antibodies usually associated to small-cell lung carcinoma), 72.8% presented with neuropathic symptoms. Most commonly, the patients presented with pure sensory (40.1%), followed by mixed sensory and motor (30.2%) and pure motor symptoms (2.4%) [47]. However, no neurophysiological data were provided in this study.

In another series of 150 patients with a biopsy confirmed SCLC, Elrington et al. [6] estimated that the prevalence of sensory symptoms was 16% and the prevalence of motor symptoms 44%. However, a definite neuropathy was confirmed with nerve conduction studies only in one patient (0.7% of the total study population).

Graus et al. [46] presented a series of 200 patients, seropositive for anti-Hu antibodies, and estimated that 54% were present with predominantly sensory neuropathy and 4.5% with sensorimotor neuropathy. However, no information was given about whether nerve conduction studies were done in all patients, and what type of neuropathy these patients presented with.

The commonest early manifestation of PPN is sensory loss (67.4%), followed by pain (41.3%), weakness (22.8%), and sensory ataxia (20.7%). In 39.1% of the cases, the presenting complaints included more than one symptom. However, in 13.0% of the cases, pain was the sole first manifestation of the PPN. During the course of the PPN, 57.6% of the patients reported pain secondary to the neuropathy. In the majority of these patients (96.2%), the pain was directly related to the neuropathy (considered as neuropathic) as described in the papers included in this review; however no validated tools were used to systematically assess and record other neuropathic symptoms accompanying pain. A minority of patients reported pain indirectly related to the neuropathy (not neuropathic), secondary to painful muscle cramps.

Most reports classified PPN as acute, sub-acute or chronic/progressive. Acute were those neuropathies that evolved within 1 month, sub-acute were those that evolved between 1 and 6 months, and chronic the neuropathies that slowly kept progressing beyond 6 months. Based on these reports, we estimated that PPN are sub-acute in the majority of cases (64.1%), followed by chronic (22.8%) and acute (13.0%).

Large fiber neuropathy can be axonal or demyelinating. In axonal neuropathy, axons are affected usually in proportion to their length (length dependent). In demyelinating neuropathy, the myelin sheath around axons is affected, and as a result, the ability of the axons to speedily conduct electrical impulses is impaired. The latter leads to slow or no conduction (conduction block). Sensory neuronopathy, or ganglionopathy, is another large fiber peripheral neuropathy, in which the cell bodies of the sensory neurons located in the dorsal root ganglia are affected. Finally, in pure motor neuropathy, there is involvement of only the lower motor neurons [48].

Demyelinating neuropathies were reported in 13% of PPN cases. Among the axonal PPN (87%), the most common neurophysiological type was asymmetrical sensory ganglionopathy (27.2%), followed by symmetrical sensorimotor neuropathy (23.9%), symmetrical sensory neuropathy (21.7%), pure motor neuropathy (7.6%), and mononeuritis multiplex (6.5%). Figure 2 illustrates how common is pain based on the neurophysiological type of the PPN.

Though rare, the most painful type of PPN is mononeuritis multiplex. Mononeuritis multiplex is known to occur in many illnesses including certain types of systemic vasculitis [49]. Indeed, in the majority of patients with paraneoplastic mononeuritis multiplex a nerve biopsy was performed, showing evidence for vasculitis [36, 39].

Although pure motor neuropathy is the least common, it can still be painful. The major difference in the pain reported in such cases is that it is muscular in nature, most commonly because of cramps [40].

Out of 76 patients with painful PPN tested for anti-Hu antibodies, 47 were positive (61.8%). Other antibodies that have been associated with PPN are anti-amphiphysin and anti-CV2 antibodies. Interestingly enough, patients with anti-Hu antibodies were more likely to present with pain as a first manifestation compared to patients without anti-Hu antibodies (53.2 vs. 24.1%, p = 0.013). Also, patients with anti-Hu antibodies were more likely to present with pain secondary to the PPN at any point compared to patients without anti-Hu antibodies (74.5 vs. 41.4%, p = 0.004).

Lung cancer was the commonest cause of PPN (45.7%), followed by hematological malignancies (16.3%) and gastro-intestinal tract malignancies (12.0%). The majority of patients seropositive for anti-Hu had lung cancer (66.0%). Detailed data regarding type of malignancy are shown in Table 1. Cancer type was not related significantly to pain either as a first manifestation or during the course of the PPN (p > 0.05).

PNS can be the first manifestation of cancer relapse [8, 17, 20] and they often precede the clinical or radiological diagnosis of a local relapse or distant metastases. Patients with a history of cancer presenting with peripheral neuropathic pain, not otherwise explained, can be diagnosed with PNS when circulating paraneoplastic antibodies are detected or a PET scan is positive.

Management of pain as a result of paraneoplastic neuropathy does not differ from the published guidelines on the management of neuropathic pain [50]. However, it is commonly reported that treatment for the underlying malignancy (including tumor resection and/or chemotherapy) can improve the symptoms [3, 11, 28, 30, 37].