Editor—The hereditary spastic paraplegias are characterised by progressive spasticity, predominantly affecting the lower limbs. They have been divided into “pure” or complicated forms.1 ,2 Four loci for autosomal dominant pure hereditary spastic paraplegia (ADPHSP) have been mapped, on chromosomes 2p (SPG4), 8q (SPG6), 14q (SPG3), and 15q (SPG6).1-3

Cognitive impairment may be present in apparently “pure” hereditary spastic paraplegia. Tedeschi et al 4 found subclinical cognitive deficits in seven patients with autosomal recessive or autosomal dominant pure hereditary spastic paraplegia, when compared to normal, unrelated controls. Webb and Hutchinson5 investigated 31 patients from 12 kindreds with ADPHSP (loci not described) and found subclinical cognitive impairment, apparently restricted to cases over 50 years old. In addition, the description of an SPG4 linked family with paraplegia complicated by subcortical dementia suggests that “pure” hereditary spastic paraplegia and hereditary spastic paraplegia complicated by frank dementia may be allelic disorders.6

As a pilot study, we investigated the presence of subclinical cognitive impairment in family members from four Welsh ADPHSP families (families 26, 27, 28, and 36) using the Mini-Mental State Examination. Diagnostic criteria and clinical details for these families have been described previously.7 None of the subjects had an existing diagnosis of overt dementia.

The three largest families, 26, 27, and 28, were genotyped by PCR amplification at microsatellite markers flanking and within the relevant ADPHSP candidate regions (table 1). Family 36 was not genotyped because of its small size. For families 26 and 28, two point lod scores do not support linkage to any of the four known loci, and multipoint linkage analysis generated lod scores of <−2 throughout all four candidate regions, formally excluding linkage. For family 27, two point and multipoint linkage analysis excluded linkage to the chromosome 15 locus. A positive lod score (1.54) was obtained with marker D14S288, although haplotype analysis (data not shown) made linkage at the chromosome 14 locus very unlikely, since it would require a double recombination in the 4 cM distance between D14S1013 and D14S269. Positive lod scores for all three chromosome 2 locus markers were obtained, peaking at 2.22 for D2S367. Dubéet al 8 have suggested that a lod score of 1.55 is sufficient to declare linkage at the chromosome 2 ADPHSP locus, based on the high prior probability of linkage at this locus, and so family 27 was classified as linked to the chromosome 2 ADPHSP locus.

The Mini-Mental State Examination (MMSE), a screening test for cognitive impairment that covers orientation, attention, memory, language, and praxic skills, was carried out in all consenting family members ⩾15 years old.9 Thirty five subjects from the four families were examined. Of these, 22 were affected by ADPHSP (five in family 26, eight in family 27, eight in family 28, and one in family 36) and 13 were unaffected but at risk, since they may later develop signs of ADPHSP. These “at risk” subjects represent a well matched and conservative control group. There was no significant difference in the ages of the affected (mean (1 SD) 37.5 (14.7) years) and at risk (mean (1 SD) 33.3 (12.3) years) groups. There was a significant difference in MMSE score between affected (mean 27.05) and at risk (mean 28.46) subjects (p=0.0086, Mann-Whitney U test). The difference in MMSE score between affected patients and controls was still significant (p=0.018) when two outliers with an MMSE score under 24 were excluded. To assess whether the lower MMSE scores were restricted to affected subjects >50 years, as previously reported, MMSE scores for affected subjects ⩽50 years old were compared to the controls. A significant difference in MMSE score remained (p=0.0045).

MMSE scores were compared between families 26, 27, and 28, and there was no significant difference (Kruskal-Wallis test), either when all subjects or only affected subjects were analysed. On the basis of the linkage results, the MMSE results for families 26 and 28 (linkage excluded at all four ADPHSP loci) were pooled and compared against family 27 (chromosome 2 linked). There was no significant difference between the results of these two groups, either when all subjects or only affected subjects were analysed (Mann-Whitney U test). Affected subjects in family 27 had lower MMSE scores compared to the pooled scores for at risk subjects from all families (p=0.0066) (Mann-Whitney U test). Similarly, affected subjects in families 26 and 28 had lower MMSE scores compared to the pooled scores for the at risk subjects from all families (p=0.043).

The MMSE is recognised to be a fairly crude test of cognitive function, but despite this we detected subclinical cognitive impairment in families with ADPHSP, confirming the results of two previous studies.4 ,5 As a new finding, we detected cognitive impairment in family members under the age of 50 years and our results also indicate that cognitive impairment in ADPHSP may not be confined to a single linkage group. These data suggest that further studies using more detailed neuropsychological tests, in ADPHSP patients from defined linkage groups, are required to define the precise nature of cognitive impairment in ADPHSP.