The epidemiology of the Lambert-Eaton myasthenic syndrome in the Netherlands

Patients and methods.

The study period ran from July 1, 1998, to July 1, 2003, and concerned the entire Netherlands, with 16,221,695 inhabitants by July 1, 2003. All eight university hospitals and eight large neurologic centers participated in this study. Patients were identified through standard hospital databases using codes 358.0 and 358.1 of the International Classification of Diseases, 9th revision, or using the Dutch neurologic coding system according to Kortbeek, or using the neuromuscular databases in the university hospitals. To overcome a possible referral bias we sent a letter to all neurologists in the Netherlands in a search for other LEMS patients, backed by additional attention to the aims of this study by publications in Dutch medical journals and presentations to neurologists and pulmonologists. Case finding was started on July 1, 1998.

After obtaining informed consent, one examiner (P.W.W.) interviewed and examined all patients. If this was not possible, clinical records were studied. We recorded demographic and clinical features, including date of onset of LEMS, date of first visit to a neurologist, signs and symptoms of LEMS, date of diagnosis of LEMS and of SCLC, results of diagnostic tests, any incorrect diagnoses, and date of death. EMG reports were evaluated, and an EMG with repetitive nerve stimulation was repeated whenever possible. In all but one patient, serum was available for testing of antibodies against P/Q-type voltage gated calcium channels. Patients with a diagnosis of LEMS were included if they had an abnormal EMG or P/Q-type voltage gated calcium channel antibodies, in addition to variable muscle weakness. Results of repetitive nerve stimulation were considered compatible with LEMS when an increment of more than 100% was obtained after either maximal voluntary muscle contraction or high frequency stimulation.

We calculated 95% CI assuming a Poisson distribution. Population figures for the Netherlands were based on data provided by the Statistics Netherlands, a department of the Ministry of Economic Affairs.5 We tested differences in clinical features between groups with and without SCLC with the Mann-Whitney test or Fisher’s exact test when appropriate, and considered a p value of < 0.05 significant.

Results.

Fifty-two patients were available for study (table), after excluding five cases in which diagnostic criteria were not met. We examined 36 patients at the Leiden University Medical Center and eight in their local hospital, leaving eight patients for whom only clinical records were reviewed. In patients with SCLC, the median interval between onset of LEMS-related symptoms and diagnosis of SCLC was 4.5 months (ranging from 41 months before to 4 months after SCLC diagnosis). Three patients without SCLC had a follow-up of less than 2 years after onset of LEMS. Eleven patients died before July 2003 (nine with SCLC). On July 1, 2003, 41 patients survived (eight with SCLC), resulting in a prevalence estimate of 2.5 per million inhabitants, which had almost doubled during our study period (1.3 × 10⁻⁶ by July 1, 1998). Dividing the Netherlands into four major regions as defined by the Statistics Netherlands, prevalence in the north was 3.0, the west 2.8, the east 1.8, and the south 2.8 per million inhabitants (differences not significant). Thirty-two new cases, including 15 with SCLC (47%), were ascertained during the 5-year study period (5 to 8 per year).

The median duration of disease until diagnosis of LEMS was higher in LEMS patients without SCLC than in those with SCLC (p < 0.001, see the table). Thirty patients (58%) had received an incorrect diagnosis before LEMS was established. This most often concerned myasthenia gravis (n = 17), polyneuropathy (n = 6), and myopathy (n = 3). In 9 patients (18%) a muscle biopsy had been performed.

Discussion.

The incidence of SCLC associated LEMS equaled that of LEMS without SCLC, whereas the prevalence was lower, as a result of poor survival from SCLC. The frequency of 47% of SCLC in newly diagnosed LEMS patients is in line with previous reports (42 to 61%).1–4⇓⇓⇓ SCLC is generally evident within the first 2 years after onset of LEMS.2,3⇓ As three of our patients without SCLC had a follow-up of less than 2 years, the frequency of underlying SCLC could be slightly higher. The duration of disease until LEMS diagnosis was significantly shorter in patients who had an underlying SCLC. The presence of SCLC could not explain this difference, because in all but one patient LEMS diagnosis preceded diagnosis of SCLC. Possibly, LEMS with underlying SCLC has a more progressive course, which could shorten both patient and doctor delay.

Epidemiologic data on LEMS are scarce. A Danish study on myasthenia gravis reported a low annual incidence of LEMS of 0.17 per million, but the study was not primarily aimed at LEMS and the prevalence was not calculated.6 The incidence and prevalence in the present study were similar to those in our regional study in the province of South Holland, comprising 1.7 million inhabitants,7 suggesting that the methods of case collection were equally appropriate nationwide, as is further supported by the comparable prevalences in four regions of the Netherlands. The participation of all university hospitals, together with the generally well-organized registration systems in Dutch hospitals, maximized case ascertainment.

However, the considerable doctor delay and incorrect diagnosis indicate that LEMS may be underdiagnosed. Furthermore, prevalence during our 5-year study period almost doubled, which suggests underdiagnosis at least before the start of our case collection. Better knowledge of the clinical features of LEMS and easier availability of the antibody assay in the Netherlands in the last years may have increased the recognition of LEMS.