The prevalence of neuropsychiatric syndromes in systemic lupus erythematosus

Patients and methods.

Results.

We studied 39 women and seven men, with a mean age of 45 ± 13 years (range 20 to 64). The mean disease duration since diagnosis of SLE was 14 ± 8 years (range 2 to 37). The mean number of ACR SLE criteria fulfilled was 5.5 ± 1.2 (range 4 to 9). The mean level of education was 13 ± 3 years (range 8 to 18). The mean ECLAM score was 1.6 (range 0 to 5), indicating a very mild disease activity. At the time of the study, 25 patients (54%) were using glucocorticoids, seven (15%) were on antimalarials, five (11%) on azathioprine, two (4%) on methotrexate, and one (2%) on cyclosporin. Nineteen patients (41%) were without specific medication. For those who had taken steroids, the mean cumulative lifetime dose of prednisolone was 26 g (range 1 to 89). A total of four patients (9%) had never used corticosteroids.

At least one NP syndrome was identified in 42 patients (91%). Five patients (11%) had one, 12 patients (26%) had two, 15 patients (33%) had three, four patients (9%) had four, four patients (9%) had five, and two patients (4%) had six NP syndromes, whereas only in four patients was no evidence of NP syndromes detected (table 2).

The most frequent single manifestation was cognitive dysfunction, detected in 37 patients (81%). Memory dysfunction was the most affected domain registered in 20 patients (43%), followed by deficits in simple attention, visuospatial processing, and psychomotor speed, each in 12 patients (26%). Only 14 of 37 patients (38%) with cognitive dysfunction by testing had cognitive problems. Headache was diagnosed in 25 patients (54%). Depression, consisting of a major depressive-like episode of mostly mild severity in 18 and a mood disorder with mixed features in two patients, was detected in 43% of patients. Clinical findings possibly related to polyneuropathy were detected in 13 patients, but only four patients had subjective symptoms. ENMG findings consistent with polyneuropathy were present in only three of them. Seven patients (15%) had a history of cerebrovascular disease. No intracerebral hemorrhages were recorded. Mean time since the diagnosis of SLE to the occurrence of the first stroke or transient ischemic attack was 11 years (range 1 to 18). In four patients cerebrovascular disease was concomitant with systemic disease activity.

One patient (2%) had a demyelinating syndrome. The neurologic symptoms appeared with subacute onset 18 years after the diagnosis of SLE. In CSF examination, IgG index was elevated and eight oligoclonal bands were present, indicating intrathecal synthesis. Increased signal in periventricular white matter and at the level of corona radiata on T2-weighted MRI scans was observed. Both aCL (151 IgG phospholipid unit) and B2GPAbG (333 standard IgG anti-beta 2 GPI unit) levels were in the high positive range.

There was an association between history of headache and current steroid use (χ² = 4.114; p = 0.043). Compared with other patients with SLE, those with a history of epileptic seizures (mean 65.0g vs 23.0g; t = 3.319, p = 0.002) or cerebrovascular disease (mean 45.4g vs 22.7g; t = 2.288, p = 0.028) had a higher cumulative lifetime dose of steroids. No other associations between occurrence of NP manifestation and steroid use were noticed.

Discussion.

In our study of 46 patients with SLE, 91% were identified to have at least one NP syndrome, whereas in previous studies, the prevalence of NPSLE was significantly lower, ranging from 14 to 75%.15-20⇓⇓⇓⇓⇓

The main explanation for the exceptionally high prevalence of NP involvement in our study is most likely that the ACR nomenclature system includes a wide array of different neurologic and psychiatric features as diagnostic elements. In most former studies the diagnosis of NPSLE was entirely based on major neurologic and psychiatric involvement, such as cerebrovascular event, neuropathy, movement disorder, transverse myelitis, seizure, meningitis, organic brain syndrome, and psychosis.15-16⇓ Patients with SLE also present with a number of minor neurologic and psychiatric problems such as headache, mild depression, anxiety, and cognitive complaints that have now been included in the new nomenclature system. If these minor manifestations and also ENMG-negative polyneuropathy were excluded, the prevalence of NPSLE in our series would be 46% instead of 91%.

Another important difference compared with former studies is in patient selection. To our knowledge, no population-based study on the prevalence of NPSLE has been published so far. In the previous studies patients have been gathered from academic referral centers likely to be biased in terms of disease severity in comparison with the general population. Furthermore, much milder forms of the disease are now being recognized than a few decades ago.20

Cognitive impairment was the most frequent finding, occurring in 37 patients (80%). Previously reported prevalence rates have been lower, ranging from 21 to 66%.2,21-26⇓⇓⇓⇓⇓⇓ According to the ACR nomenclature, individual patients were defined as cognitively impaired if they demonstrated an impairment in one or more of eight areas of cognitive ability. In 70% of our patients with cognitive impairment, the dysfunction was classified as mild, implying that the impairment was established in only one or two cognitive domains. Eleven patients had three or more impaired areas, fulfilling the criteria for cognitive impairment used in other studies.21 Based on this definition, only 24% of our study population would have been classified as having a cognitive impairment. This figure is in accordance with prevalence rates of 21% in a study of 70 unselected patients with SLE21 and 26% in a study of 73 patients.22

In our study population, 25 patients (54%) had had headache, with lifetime prevalence of migraine and tension headache being 39 and 15%. The prevalence of tension headache in our study was actually slightly lower than in the healthy population.27-29⇓⇓

The reported prevalence rates of psychiatric manifestations in SLE have varied from 20 to 70%.22,30⇓ Recent reports have focused on the occurrence of milder psychiatric features such as depression and anxiety, whereas, in contrast to older studies, low prevalence of psychosis has been observed.20,22⇓ Our results corroborate this finding: 80% of patients with mood disorder had mild depression requiring no medication, whereas none had experienced any episode of psychosis.

Cerebrovascular disease was diagnosed in seven (15%) of our patients with SLE. In earlier studies, the frequency of stroke varied from 3 to 19%.15-19,31⇓⇓⇓⇓⇓ Patients with SLE who have had a stroke or transient ischemic attack are at high risk for recurrent cerebral ischemia.32 Two of our seven patients with cerebrovascular disease had a recurrence of stroke or transient ischemic attack, and one had a chronic multifocal disease. A significantly higher cumulative lifetime dose of steroids in patients with SLE with a history of cerebrovascular disease obviously reflects the severity of this subgroup.

A mild, symmetric distal sensory or sensorimotor neuropathy is the most common form of peripheral neuropathies in SLE.16 A 28% frequency of polyneuropathy in the current study is certainly overestimated, because the ACR nomenclature system does not demand an ENMG confirmation of polyneuropathy. The corresponding prevalence of ENMG-confirmed polyneuropathy in our series was 7%, in accordance with the previous studies.15-18,33-34⇓⇓⇓⇓⇓

One of the patients fulfilled the ACR nomenclature criteria for a demyelinating syndrome which may be reminiscent of MS.35-36⇓ Our patient had neurologic and MRI findings similar to MS, but a high positive titer of aCL together with the history of thrombocytopenia and deep venous thrombosis strongly favors the diagnosis of antiphospholipid syndrome. Oligoclonal banding has been described to occur in patients with secondary antiphospholipid syndrome caused by SLE.36 We suggest that future revisions of the ACR nomenclature system specifically exclude antiphospholipid syndrome in the diagnosis of a demyelinating syndrome.

The main weakness of the study is the small sample size, and therefore the results should to be interpreted with caution. As a cross-sectional study, the validity of disease history depends on the documentation of past disease episodes. We think that this is unlikely to have a substantial effect on the prevalence rates of the NP syndromes, but reliable discrimination between primary NPSLE, secondary NPSLE, and concurrent disease process becomes problematic. It has been estimated that two-thirds of NP manifestations in SLE are not directly related to NPSLE but are due to secondary causes such as drugs, infections, and metabolic complications of the disease.37,38⇓ Prospective analyses would help in discriminating between primary and secondary NPSLE.