The association between shoulder stiffness and glycemic control is controversial. Laslett et al. [8] have observed an increased incidence of shoulder pain in diabetic patients with higher hemoglobin A_1c levels and previous eye surgery. This result was not confirmed by others [5, 7, 9].

Causes of this predisposition to shoulder disorders remain unclear. The possible pathomechanisms include the alteration of collagen synthesis and overaccumulation of advanced glycation end products [7]. It was hypothesized that the link between shoulder pain or disability and diabetes is to be excessive glycosylation of connective tissue, particularly collagen [10]. In 1999, Monnier et al. [11] observed that the patients who had had long-term intensive treatment of their diabetes showed to have lower levels of skin collagen glycosylation, glycoxidation, and cross-linking. Salmela et al. [12] revealed in that group of patients also slower rates of accumulation of advanced glycosylation end-points (AGEs) than the patients treated under conventional regimes. The formation of AGEs is associated with microvascular complication in diabetes. They have been observed to decrease vascular elasticity in experimental studies [2]. Siu et al. [13] noted that diabetic patients have higher proinflammatory IL-1β levels in the subacromial fluid than nondiabetic patients. In 2002, Mentink et al. [14] performed a murine study and observed that tendons may be directly affected by nonenzymatic glycosylation processes which change collagen cross-link. The biosynthesis of collagen is characterized by modifications such as glycosylation of the polypeptide chains which are unique to collagen and some other proteins [15]. Collagen modification due to glucose fixation on free amino groups is characterized by an altered solubility, increased resistance to enzymatic action and variations in cross-linking [15].

In 2008, Longo et al. [16], in a case-control study performed on 194 patients, observed that normal, but in the higher part of the normal range, increasing plasma glucose levels may be a risk factor for rotator cuff tear. In fact, according to the statistics, patients with a cuff tear had significantly higher fasting plasma glucose levels within the normoglycemic range than a controlled group represented by patients who had undergone arthroscopic meniscectomy for a meniscal tear and with no evidence of shoulder pathology.

Miranda et al. [17] observed that patients with only insulin-treated diabetes had an increased risk of chronic rotator cuff tendinitis. These data suggest type I diabetes as the underlying condition. However, since type II diabetes is more common than type I, it is likely that patients with type II diabetes dominate in the studies over those who do not have specified diabetes type.

In 1999, Duh and Aiello [18] proposed that vascular endothelial growth factor (VEGF) – a glycoprotein that plays an important role in neovascularization and increases vascular permeability – can cause endocrine changes responsible for systemic neovascularization in diabetes mellitus, thereby inducing microangiopathy. Yanagisawa et al. [19] have found that expression of this growth factor in subacromial bursa induces neovascularization and synovial proliferation in patients with rotator cuff disease, causing shoulder pain. In 2003, Handa et al. [20] examined 67 patients (14 type II diabetics and 53 nondiabetics) with rotator cuff diseases (36 with complete rotator cuff tears; 20 with partial tears and 11 with subacromial bursitis) and investigated whether VEGF had been also involved in the development of shoulder contracture in diabetics with rotator cuff disease. The mean preoperative active forward flexion significantly differed between two groups (104° in diabetics and 125 in no diabetics). Specimens of subacromial bursa were obtained from all patients during surgery. Expression and localization of VEGF were evaluated by using the reverse transcriptase chain reaction and immunohistochemistry, respectively. Number of vessels was evaluated based on CD34 immunoreactivity. Results showed that VEGF was expressed in significantly more diabetics (14/14 = 100 %) than in no diabetics (37/53 = 70 %) and that it was localized in both vascular endothelial cells and synovial lining cells. The mean number of VEGF-positive vessels and the vessels area were also significantly greater in diabetics. Using shoulder joint contracture Keating’s method [21], authors observed that shoulder contracture was significantly greater in diabetics (8/14 = 57 %) than in no diabetics (9/53 = 14 %). Furthermore, they observed that contracture was also significantly correlated with VEGF mRNA expression. These results indicate that shoulder contracture is associated with diabetes and with the soluble isoforms of VEGF.

Correlation between diabetes and rotator cuff tear seems to be confirmed by our data (unpublished data). In fact, out of a series consisting of 381 consecutive patients who underwent arthroscopic repair for a postero-superior rotator cuff tear, 49 (12.8 %) were suffering from diabetes mellitus. Out of these patients, 10 (2.6 %) were insulin-dependent and 39 (10.2 %) non-insulin-dependent diabetics. In the control group, consisting of 220 subjects without shoulder disorders, the percentage of patients with diabetes was 8.1 % (insulin dependent = 1.8 %; non-insulin dependent = 6.3 %).

Figs. 1 and 2 show the percentage of diabetics among the patients with rotator cuff tear of different size. Data indicate that the percentage of diabetic patients increases along with increasing of tear size, and moreover, the increase is particularly evident among insulin-dependent diabetics. Usually, the latter patients have a long history of hyperglycemia, with a scarcely controlled blood sugar level, only responsive to hexogen insulin.