How common it is, and who gets itStrong evidence
Most tennis injuries are in the legs, then the arm and the trunk, and most are overuse problems: 80% of 4,047 injuries in Dutch club players (average age 49). In 528 American recreational league players (average age 47), 41% of injuries were in the arm, mostly overuse, with no difference by skill level. How often that is the shoulder rather than the elbow or wrist hasn't been counted in recreational players.
The shoulder rises up the list with level and age. At 12 French Opens it was 11.9% of 750 injuries, mostly tendon problems; at 6 Australian Opens it was the commonest region in women; it was the second commonest specific site in a season of professional women, and the second commonest region in 101 Australian elite juniors (3.1 injuries per 1,000 hours). In 301 Swedish competitive players aged 13 to 19, shoulder injuries came at 1.12 per 1,000 hours of tennis. A review sums it up: shoulder pain is more common in high-level players, elbow problems in recreational ones. Older club players often have rotator cuff tears that don't trouble them.
Sources 1, 2, 3, 4, 5, 6, 7, 8, 9, 10
What the serve asks of the shoulderEmerging evidence
The serve is the most strenuous stroke in tennis, with the highest muscle activity in the shoulder and forearm, a review says. In 20 players at the 2000 Olympics, the upper arm turned back to 172 degrees before whipping forwards, at about 2,420 degrees a second in men, and the men's shoulders took about 608 newtons of compression. The highest forces come as the arm cocks back and starts forwards; after the ball is struck, the muscles at the back of the shoulder brake the arm, an enormous demand on their lowering strength. Elite players serve about 40 times a set.
The legs and trunk drive the serve. Servers with good leg drive have lower loads at the front of the shoulder, and players who bent the front knee less served as fast but with higher twisting loads at the shoulder. In expert players followed for two seasons, those who later had arm overuse injuries had poorer energy flow from the trunk to the arm. Kick serves loaded the back of the shoulder more than flat or slice serves in 7 college players, while in 12 high-performance men flat and kick serves loaded the shoulder similarly, and the authors pointed to the repetition of either serve over the years.
Sources 11, 12, 13, 14, 15, 16, 17, 18, 19, 20
Telling the causes apartEmerging evidence
The NHS lists pain that is worse when using the arm under tendonitis, bursitis and impingement; a tingling, weak arm with a shoulder that clicks or locks under instability; and sudden very bad pain after an injury under a dislocation, a broken bone or a torn tendon. The International Tennis Federation (ITF) says impingement in players is most often caused by strain from many serves and high forehands, with muscle imbalance and changes in how the shoulder blade moves. Reviews name rotator cuff problems, internal impingement, SLAP tears and the biceps tendon as the usual shoulder diagnoses in tennis, and biceps tendon changes were common in professional players' shoulders during the season.
A dead arm, catching or lost serve speed point deeper, to the labrum or a loose shoulder. The American Academy of Orthopaedic Surgeons (AAOS) says that from about 30 to 40 years of age, fraying at the top of the labrum can be a normal part of ageing. A physiotherapist sorts the causes by examination; clinical tests for SLAP tears are poor (our SLAP tear guide).
Sources 21, 22, 11, 23, 24, 25
Range, strength and the shoulder blade: the evidence is mixedMixed evidence
The racket shoulder turns inwards less than the other one. Across 25 studies of 18,534 competitive players, inward turning averaged 53 degrees on the racket side against about 63 on the other; the loss grew with years of tournament play in 39 elite players, and elite juniors had less total rotation on the racket side. One match cut inward turning by 1.3% and outward turning strength by 4.8% in 26 professionals.
Whether this matters is unclear, and the evidence is mixed and mostly from baseball, so this guide doesn't use it as a rule. Professionals with past shoulder pain had less inward turning in both shoulders, not a bigger difference between sides. In 301 adolescent players followed weekly for a year, shoulder range didn't predict shoulder complaints or injuries, and stronger rotators went with more problems, not fewer. Across overhead athletes, mostly baseball players, one meta-analysis found no range measure significantly linked with injury, while another, 75% baseball players and only 47 tennis players, found more loss in injured athletes.
An abnormal-looking shoulder blade raised the chance of later shoulder pain from 25% to 35% in 419 athletes, but a consensus says its exact role isn't clearly defined, and in handball, risk factors found in 206 players weren't confirmed in 329 others. In our reading, these are adaptations a physiotherapist checks and trains, not proven causes.
Sources 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37
Rackets and stringsMixed evidence
No study has linked racket weight, string tension, string type or grip size to shoulder pain or injury. In 5 players, a light racket needed more shoulder power and higher turning loads than a heavy one for a similar ball speed, and in 8 players a racket that resisted twisting more raised some shoulder and elbow loads. Both are small lab studies with no injury outcome, and one author of the second worked for a racket maker. Advice on grip size and strings comes from tennis elbow, as our tennis elbow guide explains, so this guide doesn't advise a racket or string change for the shoulder.
Sources 38, 39
When a scan or test mattersStrong evidence
Scans find a lot in shoulders that don't hurt: rotator cuff tears in 34% of 96 people without symptoms, and 54% of those over 60; in 40% of 20 elite overhead athletes' pain-free racket or throwing shoulders and none of the other side, with no problems over 5 years, the authors advising that a magnetic resonance imaging (MRI) scan alone shouldn't be the basis for an operation; in 36% of the racket-side shoulders of 53 club players aged 60 to 83, without worse pain or function; and in 46% of 113 high-level players aged 35 and over, most without pain.
A scan helps after a fall with sudden weakness, when a tear may need repair, or when surgery is being considered (our rotator cuff tear guide); an X-ray checks a suspected break or dislocation.
Sources 40, 41, 9, 10
How it usually goesEmerging evidence
The NHS says shoulder pain can take 6 months or longer to recover, and that self-care usually takes 2 weeks to start helping; shoulder impingement usually gets better in a few weeks or months, and NHS inform asks you to talk to a healthcare professional if it hasn't improved within 6 weeks. In the Holmgren trial, 12 weeks of specific exercise gave 69% a good outcome at 3 months, against 24% with general exercise. A dislocated shoulder usually takes 6 to 12 weeks to recover, and up to 16 weeks for some sports. No one can promise a date.
Sources 21, 42, 43, 44, 45
