Using a rat model the characteristics of the sensory neurones of the dorsal-root ganglia (DRG) innervating the hip were investigated by retrograde neurotransport and immunohistochemistry.

Fluoro-Gold solution (FG) was injected into the left hip of ten rats. Seven days later the DRG from both sides between T12 and L6 were harvested. The number of FG-labelled calcitonin gene-related peptide-immunoreactive or isolectin B4-binding neurones were counted.

The FG-labelled neurones were distributed throughout the left DRGs between T13 and L5, primarily at L2, L3, and L4. Few FG-labelled isolectin B4-binding neurones were present in the DRGs of either side between T13 and L5, but calcitonin gene-related peptide-immunoreactive neurones made up 30% of all FG-labelled neurones.

Our findings may explain the referral of pain from the hip to the thigh or lower leg corresponding to the L2, L3 and L4 levels. Since most neurones are calcitonin gene-related peptide-immunoreactive peptide-containing neurones, they may have a more significant role in the perception of pain in the hip as peptidergic DRG neurones.

Synovitis in the subacromial bursa (SAB) and the gle-nohumeral joint (GHJ) is often seen in rotator cuff diseases. In order to clarify its significance, following studies were conducted. The mRNA expression levels of IL-1B, sIL-1ra and icIL-1ra and the amount of substance P in the SAB synovium were correlated with the degree of shoulder pain. The cytokine-mRNAs in the GHJ synovium expressed more significantly in full-thickness tears (perforating tears) than in non-perforating tears. Biochemical markers (MMP-1, MMP-3) in the GHJ fluid were significantly higher in massive cuff tears than in smaller tears. These findings suggest the possibility that SAB and GHJ synovitis in rotator cuff diseases are associated with shoulder pain and the development of glenohumeral arthropathy, respectively.

We reviewed 26 patients with 34 shoulders treated by the inferior capsular shift operation for inferior and multidirectional instability. The mean follow-up was 8.3 years. In total, 12 shoulders showed voluntary subluxation. Eight operations used an anterior and posterior approach, 11 were by the posterior route, and 15 shoulders had an anterior approach. In 30 shoulders (85%) the outcome was satisfactory and 20 (59%) scored good or excellent results on the Rowe system. Instability had recurred in nine shoulders (26%) from three months to three years after the operation. Six of the 12 shoulders with voluntary subluxation (50%) had recurrence, as against three of the other 22 (14%), a statistically significant difference. The operation is therefore not indicated for voluntary subluxation.

The 19 shoulders which had been assessed in 1987 at a mean of 3.5 years after surgery, were also reviewed in 1995 and found to have no significant changes in instability or Rowe score. This shows that the capsular shift appeared to have maintained its tension over an eight-year period. After the use of a posterior approach, 64% of the shoulders showed a posterolateral defect on radiographs of the humerus.