1. Lumbar spinal substance P and calcitonin-gene-related peptide derive from spillage out of the dorsal horns associated with activity of small primary sensory afferents (C- and Aδ-fibres). Cerebrospinal fluid and sural nerve levels of substance P and calcitonin-gene-related peptide have been measured in patients with diabetic polyneuropathy to determine whether differences in small primary sensory afferent activity are related to the presence or absence of painful symptoms.
2. Calcitonin-gene-related peptide was undetectable in the cerebrospinal fluid of the majority of diabetic patients (14 out of 22); it was lower overall in diabetic patients as compared with control subjects (P <0.01), it was lower in those diabetic patients with painless neuropathy (100% undetectable) as compared with those with painful neuropathy (50% undetectable; P <0.05) and it correlated conversely with warming threshold (r = 0.50; P <0.01).
3. Substance P showed no overall numerical inter-group differences or correlation with other measured variables, but six diabetic patients as compared with one control subject had undetectable cerebrospinal fluid levels and the proportion of patients with undetectable levels was higher in the group with painless neuropathy than in the group with painful neuropathy (P <0.05).
4. The levels of each peptide in cerebrospinal fluid correlated with its equivalent in sural nerve (P <0.01 for calcitonin-gene-related peptide and P <0.03 for substance P). Calcitonin-gene-related peptide correlated with substance P in the sural nerve (r = 0.84; P <0.002) and in the cerebrospinal fluid (r = 0.30; P <0.03).
5. We conclude that in sensory diabetic polyneuropathy the lumbar cerebrospinal fluid levels of substance P and calcitonin-gene-related peptide relate directly to their levels in sural nerve and appear to be reduced in the painless compared with the painful syndrome, suggesting a relatively spared and/or more active primary sensory afferent fibre population in the latter.