Does an Increase in Supraspinatus Tendon Thickness After Swimming Increase the Likelihood of Future Shoulder Pain?

Abstract

Background

Increase in supraspinatus tendon thickness (STT) resulting from swimming practice has been observed in those with a history of shoulder pain. The magnitude of change in STT after a swimming session and its rate of recovery may be an indicator of future shoulder pain incidence.

Hypothesis

The supraspinatus tendons that demonstrate a greater increase in thickness as a result of swimming practice will have an increased likelihood of future shoulder pain in a cohort of competitive swimmers over a period of 6 months.

Design

Descriptive epidemiology study.

Level of Evidence:

Level 2b, individual cohort studies.

Methods:

A cohort of 50 nationally qualified swimmers aged between 14 and 22 years, from 3 open National Swimming Programs in Victoria, Australia, were recruited for this study. Ultrasonographic measurements of swimmers’ STT was obtained of both shoulders, before, immediately after, and 6 hours after a single swimming practice session. Data were recorded of any significant interfering shoulder pain at 3 and 6 months after the initial testing session.

Results

Stepwise logistic regression models indicated that significant predictors of the likelihood of experiencing significant interfering pain were sex [significant at 6 months; odds ratio (OR) 4.2] and the extent of change in STT immediately (OR 2.3 and 1.3 per mm at 3 and 6 months, respectively) and 6 hours postpractice (OR 1.9 and 1.5 per mm at 3 and 6 months, respectively).

Conclusion

The current data suggest larger increases in tendon thickness after training are associated with an increased likelihood of future shoulder pain.

Clinical Relevance

These data may be valuable for monitoring training load and identifying athletes who may have an increased likelihood of shoulder pain.

Shoulder pain is the most common musculoskeletal problem experienced by recreational and competitive swimmers.^10,18 Shoulder pain among swimmers results in significant morbidity, requires extensive rehabilitation, impedes competitive performance, and in some cases may lead to premature retirement from the sport.^2,10,18 A survey of 109 elite Australian swimmers found that 70% had experienced shoulder pain at some time during their career, ¹ with the incidence of swimming-related shoulder pain reported as high as 90%.^6,16,17 Numerous intrinsic risk factors for shoulder pain in swimmers have been identified, including joint laxity, internal/external rotation strength, previous history of injury, hours and kilometers swam per week, and competition level.^7,16

Supraspinatus tendinopathy is a common pathology observed among swimmers with studies reporting as many as 69% of all swimmers and 100% of international athletes tested have shown evidence of supraspinatus tendinopathy. ¹⁶ The incidence of tendinopathy among swimmers has been found to be strongly correlated with the hours and distance swum per week. ¹⁶ Although the incidence and severity of shoulder pain in swimmers is well established, the cause underpinning its occurrence is not fully understood. Therefore, appropriate tools to monitor and manage this condition have yet to be established.

Previous research has shown a greater degree of tendon thickening as a result of a single swimming practice in shoulders with a history of pain (HOP), compared with those without. ¹⁴ Further, it has been demonstrated the supraspinatus tendon responds differently to high intensity swimming practice compared to high volume practice, with a significantly greater increase in supraspinatus tendon thickness (STT) as a result of high-intensity practice requiring a longer period of recovery to regain prepractice thickness. ¹⁴ Given the significant morbidity associated with the development of symptomatic tendon pathology, the ability to predict and ultimately prevent the onset would be a useful tool for the sport medicine clinician. Asymptomatic rotator cuff changes are not uncommon among overhead and throwing athletes; however, from the studies available, there is limited evidence that tendon changes predict the development of future symptoms. In a cohort of Australian football athletes, thickening of the Achilles tendon has been associated with the development of tendon pain, ¹³ while an investigation which focused on the shoulders of cricketers found that increased dominant limb STT was a predictor of dominant shoulder injury. ⁴ Therefore, the change in tendon thickness may be a prelude to tendon degeneration, subsequent pain, and, finally, tendon injury or rupture.

Although these previous studies have demonstrated that tendon thickening is associated with the development of future pain, all previous work assessed rested tendon values.^4,5,8,13 To date, no studies have investigated the link between the change in tendon thickness in response to loading as a risk factor of future pain, injury, or dysfunction. Given the altered response to training observed in those with a HOP, ¹⁴ it is suggested that the magnitude of change in STT after a swimming session and its rate of recovery may also predict future incidence of shoulder pain among swimmers.

The specific aims of this descriptive injury study are to identify potential intrinsic and extrinsic tendon-related predictors for the development of shoulder pain in elite swimmers and to determine whether tendons which thicken to a greater degree as a result of swimming practice have an increased likelihood of future shoulder pain.

Methods

A descriptive epidemiology study was employed to investigate differences in shoulder pain incidence among elite Australian swimmers. A convenience sample of 50 nationally qualified (obtained Australian age or open qualifying times in the previous 12 months) swimmers aged between 14 and 22 years were recruited for this study from 3 of the top 5 ranked Australian swimming clubs. The study examined both shoulders (N = 100) of all swimmers included in this study. The study was conducted over a 6-month period. Demographic information presented in Table 1 was provided via a participant questionnaire before testing. Before the initial testing, all shoulders were scanned and stratified by pain history status into “HOP” and “pain free” groups as defined by the presence or absence of significant interfering pain (SIP) that caused the athlete to modify training or competition within the previous 6 months. No swimmers reported shoulder pain at the time of testing. Any swimmers with a history of shoulder surgery were excluded from the study. The study was approved by the University Human Research Ethics Committee of Queensland University of Technology and all participants read and signed the informed consent before participation.

Table 1.

Participant demographic information, main stroke, and training characteristics

Variables	HOP (N = 37)	PF (N = 63)
Age, y (mean ± SD)	17.0 ± 2.0	17.6 ± 2.1
Sex
Male	16 (43%)	34 (54%)
Female	21 (57%)	29 (46%)
Height, cm (mean ± SD)	174.1 ± 9.3	177.2 ± 9.6
Body mass, kg (mean ± SD)	65.2 ±8.5	69.6 ± 10.0
Main stroke
Freestyle	13 (35%)	23 (37%)
Backstroke	6 (16%)	14 (22%)
Breaststroke	8 (22%)	11 (17%)
Butterfly	4 (11%)	9 (14%)
IM	6 (16%)	6 (10%)
Training sessions per week (mean ± SD)	8.4 ± 0.7	8.3 ± 0.8

HOP, history of pain; IM, individual medley; PF, pain free.

Baseline Ultrasound Testing

Ultrasonographic measurements of swimmers’ acromiohumeral distance (AHD) and STT, in millimeters, were obtained using a Mindray DP-20 ultrasound machine in conjunction with a 38-mm 7.5-MHz Mindray 75L38EB linear transducer (Mindray). The ultrasound measurement protocol is outlined in previously published work. ¹⁴ All measurements were taken by the same researcher who has undergone advanced training in musculoskeletal ultrasound imaging. Previously published research from the preparatory phase of the study, using the same equipment in the same sample of swimmers, demonstrated exceptional test-retest reliability for the measurement of AHD [intraclass correlation (ICC) = 0.94; 95% confidence interval (CI) = 0.88-0.97; standard error of the mean (SEM) = 0.26 mm] and STT (ICC = 0.96; 95%CI = 0.98-1.00; SEM = 0.09 mm). ¹⁴

The initial testing was conducted across the 3 clubs within a 3-week period during January (early in long course season). All participants were tested on a Monday morning practice session to minimize variations in training load from previous practice sessions throughout the week. The effect of a single swimming practice was assessed across 3 testing sessions in which measures of STT and AHD were made. The first session (prepractice) was conducted after a 24-hour break from training before morning practice to standardize the effects of loading and establish rested values. The swimmers then conducted their prescribed (as determined by their coach) swimming practice [mean distance = 5467 m (min = 3800 m and max = 6750 m) and mean duration = 115 min]. The second testing session was conducted immediately after practice (postpractice). The third and final testing was conducted 6 hours after the completion of practice, immediately before their afternoon practice. Three images of AHD and STT were taken and measured at each session, with the mean of the 3 used for analysis. Coaches provided the details of the prescribed sessions, and the main investigator was present at all practice sessions.

Shoulder Pain Definition

Due to the pervasive and varied nature of the “swimmers’ shoulder” and the lack of agreement regarding its cause, the incidence of such conditions is commonly discussed in the literature as shoulder pain. The term SIP was used in this investigation because it describes shoulder pain that results in disruption to swimming training or competition. Specifically, SIP is defined as pain in the shoulder that interferes with effective quality of training or competition, or progress in training for more than 1 week.^11,17

Pain Incidence Follow-up

Head coaches were contacted via email at 3 and 6 months after the initial testing date. The time points of 3 and 6 months were chosen to align with typical phases of training within a periodized training program. Episodes of shoulder pain meeting the provided definition were reported by the coach. Incidence of SIP identified by coaches was then confirmed by the designated team clinician.

Statistical Analysis

Data were analyzed using SPSS 11.0 for Windows. Multiple stepwise logistic regression analysis was performed with incidence of SIP at 3 and 6 months as the dependent variable and the change (difference from prepractice) in STT immediately postpractice, change (difference from prepractice) in STT 6 hours postpractice, individual (height, weight, age, and sex) and training characteristics (main stroke, sessions per week, resting STT, AHD, and ratio), and previous HOP as the independent variables.

Results

The incidence of shoulder pain was 10% of all shoulders within the 3 months and 22% of all shoulders in 6 months after the initial ultrasound testing (data available in the Appendix, available in the online version of this article). Incidence of shoulder pain at 3 months posttesting was 7/50 (14%), which included 3 swimmers with bilateral shoulder pain and 4 swimmers with unilateral shoulder pain. At 6 months posttesting, the incidence was 17/50 (34%), which included 5 swimmers with bilateral shoulder pain and 12 swimmers with unilateral pain. At the 3-month follow-up 6/37 (16%), shoulders with a HOP had experienced another episode of SIP, and 13/37 (37%) had experienced another episode of SIP at the 6-month follow-up. Results from stepwise logistic regression model at 3 and 6 months are outlined in Tables 2 and 3, respectively. Significant predictors of the likelihood of experiencing SIP were sex (significant at 6 months; OR 4.2) and the extent of change in STT immediately (OR 2.3 and 1.3 per mm at 3 and 6 months, respectively) and 6 hours postpractice (OR 1.9 and 1.5 per mm at 3 and 6 months, respectively), after adjustment for all other factors. There were no significant associations between rested STT values or previous history of SIP and future shoulder pain.

Table 2.

Logistic regression, variables in the equation for SIP at 3 months

	B	SE	Sigma (P)	OR	95% CI		VIF
					Lower	Upper
Sex	19.573	4631.900	0.997	–	0	–	1.041
Change in STT (post-pre)	0.839	0.300	0.005*	2.3	1.2	4.1	1.083
Change in STT (6 hours, pre)	0.259	6.418	0.011*	1.9	1.1	3.2	1.042

B, beta coefficient; OR, odds ratio; SIP, significant interfering pain; STT, supraspinatus tendon thickness (mm); VIF, variance inflation factor.

^*Significant association.

Table 3.

Logistic regression, variables in the equation for SIP at 6 months

	B	SE	Sigma (P)	OR	95% CI		VIF
					Lower	Upper
Sex	1.436	0.630	0.023*	4.2	1.2	14.4	1.041
Change in STT (post-pre)	0.282	0.117	0.016*	1.3	1.0	1.6	1.083
Change in STT (6 hours, pre)	0.379	0.133	0.004*	1.5	1.1	1.8	1.042

B, beta coefficient; OR, odds ratio; SIP, significant interfering pain; STT, supraspinatus tendon thickness (mm); VIF, variance inflation factor.

^*Significant association.

Limitations

The current study utilized a retrospective design, which is weaker than that of a prospective study. While excellent test-retest reliability of the measurement of AHD and STT has been reported in a cited study utilizing the same population used in this study, no reliability testing was conducted in the present study.

The current investigation required coaches to report the incidence of pain and identify when training or competition had been modified. However, the rates reported by coaches are dependent on swimmers self-reporting pain; therefore, the incidence may still be influenced by the individual swimmers as well as the cultural standards at the clubs investigated.

Furthermore, the current study does not address time to occurrence of pain. The model indicates a likelihood of SIP within 3 or 6 months. SIP could have occurred anytime within these 3-month periods. Given the odds of experiencing SIP as a result of STT changes are greater at 3 months than 6 months, this measure may be more valuable for assessing short-term likelihood of pain.

Finally, no measures of shoulder internal and external rotation strength, training intensity, or duration data were taken during the follow-up period. The impact of varying training loads and intensities along with inadequate levels of strength may have contributed to the incidence of shoulder pain.

Discussion

Although previous studies have demonstrated that increases in anatomic tendon thickness are associated with the development of future pain, all previous work assessed rested tendon values.^4,5,8,13 To date, no studies have investigated the link between the change in tendon thickness in response to loading and future pain.

The primary aim of the current study was to identify whether changes in STT after swimming practice were associated with future incidence of shoulder pain in elite swimmers. The current study found the most significant predictors of experiencing SIP within 3 and 6 months of initial testing to be sex and change in STT immediately postpractice and 6 hours postpractice. Within 3 months of the initial testing, for every 1 mm increase in tendon thickness immediately posttraining, the odds of a swimmer experiencing SIP increased 2.3-fold (or 131.5%), whereas 1 mm of thickening at 6 hours posttraining increased the likelihood of experiencing significant shoulder pain increased 1.9-fold (or 92.7%). Within 6 months of the initial testing, the odds of a swimmer developing SIP were increased 1.3-fold (or 32.6%) for every 1 mm increase in STT immediately post swimming. At 6 hours posttesting, for every 1 mm of thickening the odds of a swimmer experiencing SIP increased by 1.5-fold (or 46.1%). With the previously reported SEM of STT being 0.09 mm a change of 1 mm can be considered a true change because it is outside the error of the test. Sex was a significant factor in the development of SIP in the current cohort, with no women reporting SIP within 3 months, and at 6 months posttesting the odds of a male swimmer being injured were 4.2 times higher than the odds of a female swimmer experiencing SIP. The significant association of sex may be a result of a number of factors, including swimmer’s rates of self-reporting ⁶ and potentially higher swimming speeds,^12,15 which may be an area for future investigation.

Some studies have suggested the significance of tendon thickening may represent a transient state or adaptive response by the tendon.^3,9 Tendon thickening may be seen as an indicator that a tendon is under stress from repetitive loading, but why some thickened tendons seem to compensate for or adapt to this stress and remain asymptomatic and others go on to develop pain remains unclear. The current study suggests the measure of change in STT in response to load is a potentially valuable tool in the monitoring and management of significant shoulder pain incidence in swimmers. These changes may also provide further information about the development of symptomatic tendon pathology. Furthermore, along with previous work, ¹⁴ the current study provides support for the use of ultrasound measures of STT in monitoring tendon responses to swimming training and thereby identifying swimmers at a higher risk of experiencing SIP.