Surgeons who are interested in the treatment of shoulder disorders are well aware that many patients with symptomatic rotator cuff tear report, over time, that shoulder pain has lessened or even disappeared.

Goldberg et al. [10] evaluated, with the simple shoulder test (SST), subjective symptoms of 46 patients with symptomatic lesion conservatively treated. The assessments were made every 6 months and for a period of 2.5 years. Fifty-nine percent of patients in this series had an improvement in symptoms that was maintained throughout the follow-up period. Bokor et al. [11] have re-evaluated 53 patients with symptomatic rupture after a mean period of 7.6 years; among these, 74 % reported having no pain or have only mild discomfort; but only 56 % had a satisfactory “UCLA score.” A similar result was observed by Hawkins and Dunlop [12]. In their series, consisting of 33 patients, 58 % had achieved an improvement in terms of pain and range of motion after an average period of 3.8 years.

Maman et al. [13] evaluated 33 patients with symptomatic tear and observed that, at a median follow-up of 2 years, the lesion had increased in size in 52 % of cases and that the increase occurred predominantly after 18 months (50 %) and rarely before this time (19 %). Obviously, tear progression was also correlated with patient age and state of muscles fatty degeneration of the respective broken tendons. Safran et al. [14], followed over time (mean follow-up 29 months), 51 symptomatic patients. Also in this series, 49 % of cuff tears became larger. Forty-three percent of the lesions had no significant changes, while in 8 % of tears a reduction of its initial size was present. On the other side, in 24 % of initially healthy shoulders, a cuff injury was diagnosed. Mall et al. [5] report that patients with symptomatic rupture have a higher probability that their lesion size may increase (18 %) than the asymptomatic patients (5 %). The analysis of these data leads to the assumption that: (a) approximately half of the patients with symptomatic rotator cuff tear undergo a progressive increase in the size of the lesion and that (b) this increase occurs after a mean period of 2 years.

In contrast with from Maman et al. [13] and Safran et al. [14], Fucentese et al. [15] observed that, at a mean follow-up of 3.5 years, in only 25 % of 24 symptomatic patients (mean age 54 years) with a tear less than 1.6 cm, conservatively treated, a progression of the lesion occurred. In addition, in the cases of a progression, it did not compromise the tear repairability.

Therefore, these data suggest that young, symptomatic, and with a small tear patients can initially be treated conservatively as the risk that a progression of the lesion occurs is still low.

Zingg et al. [16] examined clinically, with RX and RM, 19 patients with a massive tear, mean age of 64 years, with mild symptoms and poor functional claims. Four years after the first observation, patients maintained a satisfactory functional status although they presented greater degenerative joint changes.

Natural history of symptomatic and asymptomatic partial tears has been studied by Mall et al. [5]. The authors followed over time (mean follow-up: 2 years) 30 patients with asymptomatic partial lesion; 20 of them were asymptomatic, the others started to have pain. During ultrasound examination, none of partial tears had become full-thickness ones, while the worsening had occurred in 40 % of the lesions which became symptomatic. Pain, as for full-thickness lesions, appeared closely related to the increase of tear size.

Maman et al. [13] re-evaluated, after 2 years, 30 patients with symptomatic partial-thickness lesion diagnosed by MRI. In this series, only 10 % of the lesions increased in size (>5 mm); this data is significantly lower than the one reported for full-thickness lesions (50 %). No difference was found between progression of the partial tear on bursal side compared to that on the articular side.

These data suggest that partial-thickness tears have low tendency to progress in the first 2 years; therefore, especially at the beginning, the best treatment of patients with this type of lesion should be conservative. However, as it is possible that these lesions remain asymptomatic for a long time, and therefore undetected, in front of an occasional diagnosis it is incorrect to state that they cannot become full-thickness tears in the short period.

Some authors [17–19] studied the possibility of a spontaneous healing of rotator cuff tendons in animal models. These studies provide histological information on the reparative tissue, but may be altered by the fact that the shoulder of laboratory animals is subject to load and that the tendon tissue may, as in humans, have different degrees of individual degeneration that largely depend on age, but also on heredity, metabolic disorders, peripheral microcirculation disorders, and anatomical variations. Therefore, to reduce the possibility that results may be influenced by individual factors, the animals should be taken from the same farrow.

In the study by Hirose et al. [17] carried out on 4 rabbits, no spontaneous healing of supraspinatus tear, of about 12 mm, was found after 3 weeks. On other 16 rabbits, a lesion of 5 mm was then practiced. The macroscopic and histological examination at 1, 2, 3, and 4 weeks (performed for each interval of tie on 4 rabbits) showed the presence of reparative tissue in increasing proportion.

Carpenter et al. [18], in a study conducted on mice, produced a 2 mm² tear of the supraspinatus tendon on both shoulders. Furthermore, tendon tissue adjacent to the lesion on the left shoulder was subjected to refrigeration so that it had a reduced capacity to repair. Twelve mice were sacrificed at 3 weeks and an equal number, at 6 and 12 weeks. At the last follow-up, 78 % of the animals still had the tendon defect; furthermore, the mechanical strength properties of the repaired tissue were lower than those of the healthy one.

In another study on mice, Gimbel et al. [19] studied the repair tissue present after 12 weeks from the detachment of the supraspinatus tendon. In all the samples, only scar tissue was observed.

Clinical studies on humans provide conflicting results. Yamaguchi et al. [4] observed that none of the patients with symptomatic or asymptomatic lesion had a reduction in the size of the lesion after an average period of 5 years.

Weber [20] performed an arthroscopic or open debridement, associated with an acromioplasty, on 55 patients with a partial-thickness rotator cuff tear. Tendon healing has never been observed at a second arthroscopic evaluation carried out after 2–7 years. In a study conducted by Kartus et al. [21], there was no healing at a mean follow-up of 101 months in patients treated with arthroscopic acromioplasty for a partial-thickness rotator cuff tear.

Similarly, Massound et al. [22] performed an acromioplasty and an arthroscopic debridement in 114 patients with a small full-thickness lesion. At the follow-up (2–5 years), none of the 25 patients who underwent reoperation because of the beginning of pain had a healing or a reduction in tear size.

Among 40 patients re-evaluated after 14 months with arthrography by Yamanaka and Matsumoto [23] suffering for a symptomatic partial tear of the articular side, 4 (10 %) had experienced a healing and other 4 (10 %) a reduction in tear size.