Abstract
Background:
To date, there have not been any epidemiologic studies that have evaluated the association between swimming over a lifetime and knee health.
Objective:
We aimed to evaluate the relationship of a history of swimming with knee pain, radiographic knee OA (ROA), and symptomatic knee OA (SOA).
Design:
Cross-sectional retrospective study
Setting:
Multi-center observational cohort
Participants:
Respondents to the historical physical activity survey within the Osteoarthritis Initiative with knee radiographs and symptom assessments
Methods:
In this retrospective study nested within the Osteoarthritis Initiative (OAI), we performed logistic regression with the predictor being swimming over a lifetime and over particular age ranges.
Main Outcome Measurements:
Person-based definitions of frequent knee pain, ROA and SOA.
Results:
2637 participants were included, with a mean age of 64.3 years (SD 8.9), BMI of 28.4 kg/m2 (SD 4.9), and 44.2% male.
Over a lifetime, the adjusted prevalence measures for frequent knee pain, ROA, and SOA for any v. no history of swimming were 36.4% (33.4 – 39.5%) v. 39.9% (37.4 – 42.5%), 54.3% (51.0 – 57.6%) v. 61.1% (58.4 – 63.7%), and 21.9% (19.4 – 24.7%) v. 27.0% (24.7 – 29.4%) respectively.
Conclusions:
This is the first epidemiologic study to indicate that swimming is potentially beneficial towards knee health, particularly when performed earlier in life (before age 35). Future prospective studies are needed to confirm these findings and to better scrutinize the associations in older age groups.
Introduction
Osteoarthritis (OA) is the most common arthritis worldwide and a major cause of disability and pain. It ranks high among the causes of disability, medical visits, and both prescription and nonprescription pharmaceutical use among individuals suffering from musculoskeletal conditions1–3. The knee is a commonly affected joint4,5. Known risk factors for knee OA include female gender6–8, increased age and BMI6,8,9, prior history of trauma8,10, prior ACL tears and meniscal surgery/damage11,12. Chronic mechanical overloading has the potential to physically damage structures within the knee13–16, and can lead to arthritic changes within the joint.
Swimming as a sport has not been studied in any randomized controlled clinical trials in knee OA, but aquatic exercises which have similar attributes to swimming have been studied in arthritis17,18. The potential benefits of aquatic exercises include heightened sensory input from water turbulence, pressure and temperature19. Additionally, aquatic exercises are felt to provide muscle relaxation and decreased joint compression that may be related to the temperature and buoyancy of being in the water19. Due to the non-weight bearing nature of swimming, it is often considered to be a preferred exercise in people who have knee OA or other sources of knee pain19,20. Despite this current clinical practice, to date there are no epidemiologic studies that have evaluated the relationship between a lifetime exposure of swimming and knee health. Therefore, systematic evaluation of an association between swimming and knee OA is needed.
The Osteoarthritis Initiative (OAI) is a cohort recruited from the community irrespective of their swimming status. In this cohort, more than 2,000 participants completed a historical physical activity survey and had standardized assessments of knee health related outcomes. We have previously used this dataset to evaluate the relationship of running21 and American football22 with knee OA. In this study, we were interested in evaluating the relationship between swimming and knee pain, radiographic knee OA (ROA), and symptomatic knee OA (SOA) in a rigorous manner. We hypothesized that leisure swimming would be associated with less frequent knee pain, ROA, and SOA.
Materials and Methods
Study Design
This is a cross-sectional study nested within the OAI, a multi-center observational cohort of knee OA. An ancillary study to the OAI funded the administration of a modified version of the Historical Physical Activity Survey23 at the OAI 96-month visit. The OAI 48-month visit had the largest number of standardized knee radiographs read for OA severity that was closest temporally to the 96-month visit, when the historical physical activity survey was administered and therefore, we evaluated outcomes primarily at the OAI 48-month visit.
The Osteoarthritis Initiative (Parent Study)
The OAI is a prospective study of men and women aged 45 to 79 who were recruited between 2004 and 2006. The four clinical centers were Memorial Hospital of Rhode Island / Brown University, Ohio State University, University of Pittsburgh, and the University of Maryland (subcontract site: Johns Hopkins University). There were 3 groups of participants targeted for this cohort, (1) those who had no evidence of knee OA or symptoms, (2) those who had risk factors for SOA but did not meet the criteria for this designation in either knee24, (3) and those who had SOA in at least one knee. Inclusion into the middle group was based on a complicated age-specific eligibility criteria where with greater age, fewer putative risk factors were required with a goal in mind to balance effectiveness in enriching each stratum with incident events of OA and facilitating recruitment24. All participants had standardized knee radiographs obtained and knee symptom assessments.
Approval was obtained from the institutional review board at each participating OAI site and at Baylor College of Medicine. Each participant provided written informed consent.
Historical Physical Activity Survey Instrument
Participants were asked to complete a self-administered modified version of the Historical Physical Activity Survey instrument23 prior to the OAI 96-month visit. In this instrument, participants reviewed 37 leisure time physical activities, including “swimming”, and were asked to identify which activities they performed at least 10 times in their lives for at least 20 minutes within a given day. The questionnaire asked about four distinct age ranges: 12–18, 19–34, 35–49, and ≥50 years old, specifically to identify up to three of the most frequently performed activities during each of those age ranges. Modifications from the original Historical Physical Activity Survey involved the questionnaire being self-administered, similar to what was previously done by Chasan-Taber25. Also, in order to limit response burden, ordinal categories were used for each of the frequency and duration selections and comments were limited to the three activities endorsed as most frequent in each age range. Individuals who indicated swimming as a top 3 activity were categorized as swimmers during that particular time period. The “Any history of swimming” category included participants who were swimmers in at least one age range. Data on the number of swimming sessions per age period was established through additional questions ascertaining the number of years, months per year, and sessions per month the participants engaged in the activity.
Knee Radiographs
At the OAI 48-month visit standardized bilateral, weight-bearing, posteroanterior semi-flexed radiographs of knees were obtained and read. If readings were not available at this time then readings from the most proximate prior visit were used instead (36-, 24-, or 12-month, or baseline). Radiographic OA severity was scored utilizing central readers26, with the Kellgren/Lawrence (K/L) grades (0–4) based on the Osteoarthritis Research Society International Atlas27. ROA was defined as KL ≥ 2. The reliability for these readings (read-reread) was substantial28: weighted kappa for intrarater agreement ranged from 0.70 – 0.80.
Pain Assessment
At the parent study 48-month visit, participants were asked to self-report knee-specific pain using the following question: “During the last 12 months, have you had pain, aching, or stiffness in or around your right/left knee on most days for at least one month? By most days, we mean more than half the days of a month.” If the response was not available then the response from the most proximate prior visit were used instead (36-, 24-, or 12-month, or baseline).
Covariates
Participant date of birth and date of 48-month visit were used to calculate age. Body mass index was measured at the 36-month OAI visit (closest visit to the 48-month visit where both height and weight were measured) and was calculated using the standard formula: weight divided by height squared (kg/m2). History of knee injuries and total knee replacements (TKRs) were self-reported at baseline and at all annual visits up to the 48-month visit. We also calculated participants’ body mass index at age 25 using reported height and weight at age 25.
Statistical Analysis
Logistic regression was performed with predictors being swimming status defined as any history of swimming, and history of swimming during particular age periods (ages 12–18, 19–34, 35–49, and ≥50 years).
Outcome definitions.
We used person-based outcomes of frequent knee pain, ROA and SOA. Frequent knee pain was defined by endorsing knee pain in at least one knee. ROA was defined as a K/L grade 2 or higher in at least one knee. SOA was defined as having at least one knee with both frequent knee pain and ROA. If participants had one knee with the outcome of interest, they were classified as having the outcome, even if data from the contralateral knee was missing. If participants had one knee without the outcome of interest and the data for the contralateral knee was missing, those participants were excluded from the analyses. Participants with total knee replacements (TKRs) were classified as having all 3 outcomes of interest. If outcomes from the 48 month visit were not available, then we used data from the most proximate visit that occurred before the 48 month visit and carried those values forward.
Exposure Definitions:
For any history of swimming, and each of the 4 age periods, exposure to swimming was explored in 2 ways. First, swimming status was dichotomized into swimmers v. non-swimmers. Second, participants were separated into four groups: (1) non-swimmers and (2) low, (3) medium, and (4) high tertiles of swimming based on an estimate of the number of sessions people participated in. We performed analyses both unadjusted and adjusted for covariates of age, sex, BMI, reported BMI at age 25, all leisure activities significantly associated with swimming during the relevant age range, and prior history of knee injury.
We also evaluated swimming status based on the number of age periods people swam during the earlier two age ranges as an additional method of measuring dose response. We excluded the latter to age ranges to minimize the likelihood of confounding by indication as there are already current recommendations to advise swimming as a physical activity for those who have arthritis19,20.
Participants without data from the modified Historical Physical Activity Survey were not included in the analyses. Trends were tested using the Cochrane-Armitage test29,30. Analyses were performed using SAS, version 9.4. P-values less than 0.05 were considered statistically significant.
Results
4,796 participants were enrolled in the OAI database. 842 did not return to the 96 month visit (17.6%). Out of the 3,954 participants who returned for the 96 month visit, 699 participants had this visit before we began administering the modified Historic Physical Activity Survey (17.7%) and therefore did not complete a survey. Of the remaining 3,255 participants, 618 did not complete the modified Historic Physical Activity Survey (19.0%). This left 2,637 of the original 4,796 who completed the modified Historic Physical Activity Survey (55.0%)21. Those who did not complete the lifetime historical physical activity question tended to be older, female, more likely to have radiographic OA and knee pain, and more likely to have a history of TKR and injury.
55.8% of the 2,637 participants were female. The mean age was 64.3 (SD 8.9) years, and the mean BMI was 28.4 (SD 4.9) kg/m2 (see Table 1). The demographic characteristics were similar between included males and females with the exception that the history of knee injury was more common in males (53.5%) than females (43.5%). 1,062 out of 2,637 participants (40.3%) participated in swimming at some point in their lives. Of those swimmers, 738 (28.0%) identified swimming in the 12–18 age range, 519 (19.7%) in the 19–34 age range, 409 (15.5%) in the 35–49 age range, and 363 (13.8%) in the ≥ 50 age range. Activities that were positively correlated with any history of swimming included bicycling/spinning, canoeing/rowing/kayaking, dancing, gardening/yardwork, ice skating, sailing, snow skiing, table tennis/ping pong, singles tennis, water aerobics, yoga, and walking. Activities that were negatively correlated included baseball/softball, basketball, bowling, fishing, and football. Of the 2,637 participants, 10% required imputation of data. For radiographic OA status, 193 participants had data imputed from the 36 month visit, 55 from the 24 month visit, and 44 from the 12 month visit. For frequent knee pain, 17 participants required imputation, 9 from the 36 month visit, 5 from the 24 month visit, and 1 from the 12 month visit.
Table 1.
Characteristics of those with no history of swimming, any history of swimming, all participants, and those excluded from these analyses.
| Non-Swimmers (n = 1575) | Swimmers (n =1062) | All Participants (n = 2637) | OAI Participants seen at 96-month visit before 9/12/12 who did not complete the historic physical activity survey (n = 699) | OAI Participants eligible for historic physical activity survey, but did not complete questionnaire (n = 618) |
|---|---|---|---|---|
| 64.1 (9.0) | 64.7 (8.9) | 64.3 (8.9) | 65.4 (8.5) | 67.0 (9.4) |
| 48.5% | 37.9% | 44.2% | 32.5% | 38.8% |
| 28.7 (4.8) | 28.1 (4.9) | 28.4 (4.9) | 28.7 (5.1) | 29.1 (5.2) |
| 23.2 (3.2) | 22.8 (3.3) | 23.0 (3.2) | 22.7 (3.5) | 23.1 (3.4) |
| 40.1% | 36.9% | 39.3% | 50.4% | 48.5% |
| 59.5% | 53.9% | 57.3% | 65.7% | 62.9% |
| 29.4% | 24.4% | 27.4% | 37.0% | 38.7% |
| 4.6% | 3.2% | 4.0% | 7.0% | 6.2% |
| 48.0% | 47.0% | 47.6% | 55.5% | 47.6% |
Knee-based characteristic (within person measure)
Exposure to any history of swimming was associated with lower prevalence of having frequent knee pain, 37.2% (34.3 – 40.1%) v. 41.2% (38.8 – 43.6%). However after adjusting for age, sex, BMI, BMI at age 25, leisure physical activities that significantly correlated with swimming during the relevant time frame, and prior knee injury this difference in prevalence was no longer statistically significant 36.4% (33.4 – 39.5%) v. 39.9% (37.4 – 42.5%). The results for all other age ranges were not statistically significant except for the age range of 19–34 where the p-value for trend was statistically significant at p=0.02 but this was no longer statistically significant once we the model was adjusted (see Table 2).
Table 2.
Prevalence of Frequent Knee Pain in Non-Swimmers compared to Swimmers (dichotomous). Prevalence of frequent knee pain among those participating in 3 levels of activity: low, middle, and high. Results are presented unadjusted and adjusted with 95%CIs.
| Swimming Time Period | Unadjusted Prevalence of Frequent Knee Pain (95% CI) | Adjusted* Prevalence of Frequent Knee Pain (95% CI) |
|---|---|---|
| Any History of Swimming | ||
| Non-Swimmers (n = 1567) | 41.2% (38.8 – 43.6%) | 39.9% (37.4 – 42.5%) |
| Swimmers (n = 1055) | 37.2% (34.3 – 40.1%) | 36.4% (33.4 – 39.5%) |
| Low (n = 398) | 39.2% (37.3 – 41.1%) | 37.2% (35.1 – 39.3%) |
| Middle (n = 304) | 37.5% (34.8 – 40.3%) | 36.0% (33.1 – 39.0%) |
| High (n = 353) | 35.8% (31.2 – 40.0%) | 35.9% (31.7 – 40.4%) |
| p for trend = 0.05 | p for trend=0.8 | |
| Ages 12 – 18 years old | ||
| Non-Swimmers (n = 1889) | 40.5% (38.3 – 42.7%) | 39.4% (37.1 – 41.8%) |
| Swimmers (n = 733) | 37.2% (33.8 – 40.8%) | 36.5% (32.9 – 40.2%) |
| Low (n = 295) | 39.1% (37.0 – 41.2%) | 36.0% (33.8 – 38.4%) |
| Middle (n = 223) | 38.1% (34.7 – 41.5%) | 38.3% (34.7 – 42.0%) |
| High (n = 215) | 37.1% (32.1 – 42.3%) | 37.2% (32.0 – 42.7%) |
| p for trend = 0.3 | p for trend=0.8 | |
| Ages 19 – 34 years old | ||
| Non-Swimmers (n = 2107) | 40.7% (38.6 – 42.8%) | 39.6% (37.4 – 41.8%) |
| Swimmers (n = 515) | 35.0% (31.0 – 39.2%) | 34.8% (30.6 – 39.2%) |
| Low (n = 181) | 37.9% (35.7 – 40.2%) | 37.8% (35.3 – 40.3%) |
| Middle (n = 142) | 35.3% (31.6 – 39.3%) | 32.9% (29.1 – 37.0%) |
| High (n = 192) | 32.8% (27.5 – 38.6%) | 33.7% (28.1 – 39.8%) |
| p for trend = 0.02 | p for trend=0.2 | |
| Ages 35 – 49 years old | ||
| Non-Swimmers (n = 2215) | 39.8% (37.8 – 41.9%) | 39.0% (36.8 – 41.1%) |
| Swimmers (n = 407) | 38.0% (33.5 – 42.9%) | 37.2% (32.4 – 42.3%) |
| Low (n = 143) | 38.7% (36.3 – 41.3%) | 37.4% (34.7 – 40.2%) |
| Middle (n = 142) | 37.6% (33.3 – 42.2%) | 35.9% (31.3 – 40.7%) |
| High (n = 122) | 36.5% (30.2 – 43.5%) | 36.3% (29.6 – 43.5%) |
| p for trend = 0.4 | p for trend=0.5 | |
| Ages > 50 years old | ||
| Non-Swimmers (n = 2260) | 39.7% (37.7 – 41.7%) | 38.7% (36.6 – 40.8%) |
| Swimmers (n = 362) | 39.0% (34.1 – 44.1%) | 38.4% (33.3 – 43.8%) |
| Low (n = 132) | 39.2% (36.6 – 41.8%) | 39.5% (36.6 – 42.5%) |
| Middle (n = 115) | 38.6% (34.0 – 43.4%) | 37.7% (32.8 – 42.8%) |
| High (n = 115) | 38.1% (31.3 – 45.4%) | 37.8% (30.6 – 45.5%) |
| p for trend = 0.7 | p for trend=0.8 |
Adjusted for age, sex, BMI, BMI at age 25, history of knee injury, and all other leisure activities correlated with swimming during the respective age ranges.
For the outcome of ROA, any history of swimming was associated with a decreased prevalence. Those with a history of swimming had an unadjusted ROA prevalence of 54.4% (51.4 – 57.4%) compared to those without having a prevalence of 60.3% (57.8 – 62.7%). The p-for trend across the groups of swimmers was also statistically significant with a p = 0.007. The adjusted results were similar except that the p for trend was higher at 0.02. For the age range-specific analyses, the 12–18 and the 19–34 age range unadjusted results were statistically significant for a lower prevalence of ROA among those who had a history of swimming. However, for the adjusted results, only the 19–34 age range results remained statistically significant (see Table 3).
Table 3.
Prevalence of Radiographic Osteoarthritis (ROA) in Non-Swimmers compared to Swimmers (dichotomous). Prevalence of ROA among those participating in 3 levels of activity: low, middle, and high. Results are presented unadjusted and adjusted with 95%CIs.
| Swimming Time Period | Unadjusted Prevalence of ROA (95% CI) | Adjusted* Prevalence of ROA (95% CI) |
|---|---|---|
| Any History of Swimming | ||
| Non-Swimmers (n = 1554) | 60.3% (57.8 – 62.7%) | 61.1% (58.4 – 63.7%) |
| Swimmers (n = 1052) | 54.4% (51.4 – 57.4%) | 54.3% (51.0 – 57.6%) |
| Low (n = 397) | 57.4% (55.5 – 59.5%) | 55.7% (53.5 – 57.9%) |
| Middle (n = 305) | 55.1% (52.2 – 57.9%) | 54.0% (50.8 – 57.2%) |
| High (n = 350) | 52.7% (48.4 – 57.0%) | 54.1% (49.4 – 58.8%) |
| p for trend=0.007 | p for trend=0.02 | |
| Ages 12 – 18 years old | ||
| Non-Swimmers (n = 1873) | 59.7% (57.5 – 61.9%) | 60.3% (57.8 – 62.7%) |
| Swimmers (n = 733) | 53.3% (49.7 – 56.9%) | 53.4% (49.5 – 57.3%) |
| Low (n = 296) | 56.7% (54.6 – 58.9%) | 52.7% (50.3 – 55.3%) |
| Middle (n = 222) | 54.1% (50.6 – 57.6%) | 55.0% (51.2 – 58.8%) |
| High (n = 215) | 51.5% (46.1 – 56.8%) | 52.8% (47.1 – 58.6%) |
| p for trend=0.01 | p for trend=0.2 | |
| Ages 19 – 34 years old | ||
| Non-Swimmers (n = 2093) | 59.2% (57.1 – 61.3%) | 59.9% (57.6 – 62.2%) |
| Swimmers (n = 513) | 52.4% (48.1 – 56.7%) | 52.3% (47.6 – 57.1%) |
| Low (n = 179) | 56.1% (53.8 – 58.4%) | 57.1% (54.4 – 59.7%) |
| Middle (n = 144) | 53.1% (49.2 – 57.1%) | 51.0% (46.6 – 55.4%) |
| High (n = 190) | 50.1% (44.2 – 56.0%) | 50.7% (44.1 – 57.2%) |
| p for trend=0.007 | p for trend=0.01 | |
| Ages 35 – 49 years old | ||
| Non-Swimmers (n = 2201) | 57.9% (55.8 – 59.9%) | 58.3% (56.0 – 60.5%) |
| Swimmers (n = 405) | 58.0% (53.2 – 62.7%) | 59.2% (53.9 – 64.3%) |
| Low (n = 143) | 57.9% (55.3 – 60.4%) | 57.6% (54.6 – 60.5%) |
| Middle (n = 140) | 57.8% (53.2 – 62.3%) | 57.5% (52.3 – 62.4%) |
| High (n = 122) | 57.8% (50.8 – 64.4%) | 60.5% (52.9 – 67.6%) |
| p for trend=1.0 | p for trend=0.9 | |
| Ages > 50 years old | ||
| Non-Swimmers (n = 2248) | 58.0% (55.9 – 60.0%) | 58.4% (56.1 – 60.6%) |
| Swimmers (n = 358) | 57.5% (52.4 – 62.6%) | 58.0% (52.3 – 63.4%) |
| Low (n = 129) | 58.0% (55.3 – 60.7%) | 56.9% (53.7 – 60.0%) |
| Middle (n = 114) | 58.1% (53.2 – 62.8%) | 56.1% (50.7 – 61.4%) |
| High (n = 115) | 58.2% (50.8 – 65.2%) | 61.0% (53.0 – 68.4%) |
| p for trend=0.9 | p for trend=0.7 |
Adjusted for age, sex, BMI, BMI at age 25, history of knee injury, and all other leisure activities correlated with swimming during the respective age ranges.
Any history of swimming was associated with decreased prevalence of having SOA, with a prevalence of 24.4% (21.9 – 27.1%) v. 29.5% (27.3 – 31.8%) with a significant p for trend across the levels of swimming. The fully adjusted model showed a similar direction, just meeting statistical significance, with a prevalence of 21.9% (19.4 – 24.7%) v. 27.0% (24.7 – 29.4%) but the p for trend across the levels of swimming was 0.08, which did not meet the level of statistical significance. The p for trend of the 19–34 age range across the levels of swimming for both the crude and adjusted results were significant (see Table 4).
Table 4.
Prevalence of Symptomatic Osteoarthritis (SOA) in Non-Swimmers compared to Swimmers (dichotomous). Prevalence of SOA among those participating in 3 levels of activity: low, middle, and high. Results are presented unadjusted and adjusted with 95%CIs.
| Swimming Time Period | Unadjusted Prevalence of SOA (95% CI) | Adjusted* Prevalence of SOA (95% CI) |
|---|---|---|
| Any History of Swimming | ||
| Non-Swimmers (n = 1554) | 29.5% (27.3 – 31.8%) | 27.0% (24.7 – 29.4%) |
| Swimmers (n = 1052) | 24.4% (21.9 – 27.1%) | 21.9% (19.4 – 24.7%) |
| Low (n =397) | 26.9% (25.2 – 28.7%) | 23.3% (21.4 – 25.2%) |
| Middle (n = 305) | 24.7% (22.3 – 27.3%) | 21.5% (19.0 – 24.1%) |
| High (n = 350) | 22.7% (19.3 – 26.5%) | 21.2% (17.7 – 25.1%) |
| p for trend=0.007 | p for trend=0.08 | |
| Ages 12 – 18 years old | ||
| Non-Swimmers (n = 1873) | 28.7% (26.7 – 30.8%) | 26.4% (24.3 – 28.6%) |
| Swimmers (n = 733) | 24.3% (21.3 – 27.5%) | 22.2% (19.2 – 25.5%) |
| Low (n = 296) | 26.7% (24.8 – 28.7%) | 22.0% (20.0 – 24.1%) |
| Middle (n = 222) | 25.2% (22.3 – 28.5%) | 24.3% (21.2 – 27.6%) |
| High (n = 215) | 23.8% (19.7 – 28.7%) | 22.7% (18.5 – 27.7%) |
| p for trend=0.1 | p for trend=0.9 | |
| Ages 19 – 34 years old | ||
| Non-Swimmers (n = 2093) | 28.6% (26.7 – 30.5%) | 26.2% (24.2 – 28.2%) |
| Swimmers (n = 513) | 22.8% (19.4 – 26.7%) | 21.1% (17.7 – 25.0%) |
| Low (n = 179) | 25.7% (23.7 – 27.8%) | 24.3% (22.1 – 26.7%) |
| Middle (n = 144) | 23.1% (19.8 – 26.7%) | 19.8% (16.7 – 23.3%) |
| High (n = 190) | 20.6% (16.3 – 25.8%) | 19.7% (15.3 – 25.0%) |
| p for trend=0.007 | p for trend=0.05 | |
| Ages 35 – 49 years old | ||
| Non-Swimmers (n = 2201) | 27.7% (25.8 – 29.6%) | 25.4% (23.5 – 27.4%) |
| Swimmers (n = 405) | 26.2% (22.1 – 30.7%) | 23.8% (19.8 – 28.5%) |
| Low (n = 143) | 26.8% (24.5 – 29.1%) | 24.0% (21.7 – 26.6%) |
| Middle (n = 140) | 25.9% (22.0 – 30.1%) | 22.7% (18.9 – 27.0%) |
| High (n = 122) | 25.0% (19.5 – 31.4%) | 23.4% (17.8 – 30.0%) |
| p for trend=0.4 | p for trend=0.5 | |
| Ages > 50 years old | ||
| Non-Swimmers (n = 2248) | 27.7% (25.9 – 29.6%) | 25.3% (23.4 – 27.3%) |
| Swimmers (n = 358) | 26.0% (21.7 – 30.8%) | 23.6% (19.3 – 28.5%) |
| Low (n = 129) | 26.7% (24.4 – 29.2%) | 24.3% (21.8 – 27.0%) |
| Middle (n = 114) | 25.8% (21.7 – 30.3%) | 22.2% (18.2 – 26.7%) |
| High (n = 115) | 24.9% (19.1 – 31.7%) | 23.4% (17.5 – 30.5%) |
| p for trend=0.4 | p for trend=0.2 |
Adjusted for age, sex, BMI, BMI at age 25, history of knee injury, and all other leisure activities correlated with swimming during the respective age ranges.
We additionally evaluated swimming status based on the number of age periods people swam including only the earlier two age ranges as an additional method of measuring dose response. The p-for trend for unadjusted analyses were statistically significant for all three outcomes, frequent knee pain, ROA, and SOA though only the ROA results remained statistically significant in the fully adjusted model (see Table 5).
Table 5.
Prevalence of Frequent Knee Pain, Radiographic OA, and Symptomatic Knee OA based on the counts of swimming participated in within the earlier age periods (12 – 18 and 19 – 34). Results are presented unadjusted and adjusted with 95%CIs.
| Frequent Knee Pain | Unadjusted Prevalence (95% CI) | Adjusted* Prevalence (95% CI) |
|---|---|---|
| Number of time periods of swimming | ||
| 0 (n = 1730) | 40.1% (38.7 – 43.2%) | 40.0% (37.6 – 42.4%) |
| 1 (n = 536) | 38.0% (35.7 – 40.3%) | 35.7% (33.3 – 38.2%) |
| 2 (n = 356) | 35.0% (30.9 – 39.4%) | 35.7% (31.3 – 40.3%) |
| p for trend=0.03 | p for trend=0.6 | |
| Radiographic OA | ||
| Number of time periods of swimming | ||
| 0 (n = 1730) | 60.0% (57.8 – 62.3%) | 60.0% (58.5 – 63.4%) |
| 1 (n = 536) | 55.6% (53.2 – 57.9%) | 53.9% (51.3 – 56.5%) |
| 2 (n = 356) | 51.0% (46.5 – 55.5%) | 52.2% (47.3 – 57.1%) |
| p for trend=0.0008 | p for trend=0.03 | |
| SOA | ||
| Number of time periods of swimming | ||
| 0 (n = 1730) | 29.1% (27.1 – 31.2%) | 26.9% (24.7 – 29.1%) |
| 1 (n = 536) | 25.5% (23.5 – 27.8%) | 22.0% (19.9 – 24.2%) |
| 2 (n = 356) | 22.2% (18.8 – 26.1%) | 21.3% (17.8 – 25.3%) |
| p for trend=0.005 | p for trend=0.2 |
Adjusted for age, sex, BMI, BMI at age 25, history of knee injury, and all other leisure activities correlated with swimming during the respective age ranges.
Discussion
In this first epidemiologic study of swimming as it relates to knee OA, people with any history of swimming had a statistically significant lower adjusted prevalence measure of ROA though admittedly, the absolute difference was modest with a difference of 5.7%. The result for SOA was just missed the level of statistical significance though the dose-response analyses were significant. There was no clear difference by swimming status in regards to difference in frequent knee pain. We observed similar findings with a modest dose-responses in the younger age ranges but not the older age ranges. These findings support the notion that swimming may be beneficial towards knee health, particularly in younger age ranges. In the older age ranges, swimmers and non-swimmers did not have significant differences in the outcomes of interest. Further prospective studies are needed to address whether there are associations between swimming and frequent knee pain, ROA, and SOA, particularly in these age groups.
A limitation to our study is the retrospective nature of the modified Historical Physical Activity Survey and the inclusion of multiple activities, limiting acquisition of more detailed information about swimming and there is a possibility that we may be observing a “healthy, active” effect rather than an effect that is specific to swimming. However, were this the case, and we were observing a “healthy, active” effect, then we would find the same pattern of association with all physical activities that were evaluated using this instrument. Instead, we have published a prior study evaluating running where there was a decreased risk of frequent knee pain and SOA but not for ROA21. Additionally, we have also reported that American football was associated with a greater prevalence of frequent knee pain, ROA, and SOA and most of the participation of that sport was during the youngest age range22. Thus, although the ascertainment of data on swimming is imperfect, it is likely that we are evaluating a particular effect of swimming. To date, this is the best data that currently exists evaluating swimming over a life time as it relates to knee OA. Our study provides initial data evaluating the association between swimming and knee OA that supports the need for future studies evaluating the relationship in a prospective longitudinal manner.
Swimming might be associated with lower BMIs which would serve as potential mechanism towards reducing chronic loading through weight-bearing joints on a regular basis. We acknowledge that within the OAI, we did not see a difference in BMIs between those who did and did not swim, but this is potentially related to the fact that this is an observational study where the exposure of swimming was not ascertained prospectively.
Some theoretical benefits of swimming for knee health include that it facilitates muscle strengthening and endurance training without bearing weight on joints that normally experience load-bearing, such as the hip, knee, ankle and joints of the feet. It is also possible that it provides similar benefits to what has been seen in aquatic therapy including improved sensory input from water turbulence, pressure and temperature, muscle relaxation and decreased joint compression19. Swimming may have other benefits including improved proprioception and maintenance of range of motion of joints.
It is also known that swimming is an aerobic activity that reduces all-cause mortality by improving both endurance and cardiovascular health31 and there is a growing idea that there might be a link between cardiovascular health and OA32–36. There are also other purported benefits to swimming including improved mental health37. It has been established that depression is associated with symptoms in patients with knee OA38, and that treatment of the depression can improve symptoms from OA39.
Although future longitudinal studies are needed to firmly establish the relationship between swimming and knee health, findings from this study do indicate a decreased association with ROA and perhaps SOA, particularly when people participated in the activity at younger ages. This is pertinent as swimming is the fourth most common sports activity in the United States40.
In summary, this is the first epidemiologic study to indicate that swimming is potentially beneficial towards knee health, particularly when performed earlier in life (before age 35). Future prospective studies are needed to evaluate populations at risk for developing knee OA and confirm that swimming is beneficial from a knee health perspective. Nevertheless, it is reassuring that the findings from our study support the likelihood that swimming may, in fact, be protective of knee OA.
Acknowledgments
Dr. Lo is supported by K23 AR062127, an NIH/NIAMS funded mentored award, providing support for design and conduct of the study, analysis, and interpretation of the data; and preparation and review of this work. Dr. Suarez-Almazor is supported by K24 AR053593, funded by NIH/NIAMS. This work is supported in part with resources at the VA HSR&D Center for Innovations in Quality, Effectiveness and Safety (#CIN 13-413), at the Michael E. DeBakey VA Medical Center, Houston, TX. The Osteoarthritis Initiative is a public-private partnership comprised of five contracts (N01-AR-2-2258; N01-AR-2-2259; N01-AR-2-2260; N01-AR-2-2261; N01-AR-2-2262) funded by the National Institutes of Health, a branch of the Department of Health and Human Services, and conducted by the Osteoarthritis Initiative Study Investigators. Private funding partners include Merck Research Laboratories; Novartis Pharmaceuticals Corporation, GlaxoSmithKline; and Pfizer, Inc. Private sector funding for the Osteoarthritis Initiative is managed by the Foundation for the National Institutes of Health.
The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Arthritis and Musculoskeletal and Skin Diseases, the National Institutes of Health, or the Department of Veterans Affairs.
References
- 1.Rothfuss J, Mau W, Zeidler H, Brenner MH. Socioeconomic evaluation of rheumatoid arthritis and osteoarthritis: a literature review. Semin Arthritis Rheum. 1997;26(5):771–779. [DOI] [PubMed] [Google Scholar]
- 2.Guccione AA, Felson DT, Anderson JJ. Defining arthritis and measuring functional status in elders: methodological issues in the study of disease and physical disability. Am J Public Health. 1990;80(8):945–949. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Ling SM, Fried LP, Garrett ES, Fan MY, Rantanen T, Bathon JM. Knee osteoarthritis compromises early mobility function: The Women’s Health and Aging Study II. J Rheumatol. 2003;30(1):114–120. [PubMed] [Google Scholar]
- 4.Kellgren JH, Lawrence JS. Radiological assessment of osteo-arthrosis. Ann Rheum Dis. 1957;16(4):494–502. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Murphy L, Schwartz TA, Helmick CG, et al. Lifetime risk of symptomatic knee osteoarthritis. Arthritis Rheum. 2008;59(9):1207–1213. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Kellgren JH, Lawrence JS. Osteo-arthrosis and disk degeneration in an urban population. Ann Rheum Dis. 1958;17(4):388–397. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Losina E, Weinstein AM, Reichmann WM, et al. Lifetime Risk and Age at Diagnosis of Symptomatic Knee Osteoarthritis in the US. Arthritis Care and Research (Hoboken). 2013;65(5):703–711. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Cooper C, Snow S, McAlindon TE, et al. Risk factors for the incidence and progression of radiographic knee osteoarthritis. Arthritis Rheum. 2000;43(5):995–1000. [DOI] [PubMed] [Google Scholar]
- 9.Felson DT, Naimark A, Anderson J, Kazis L, Castelli W, Meenan RF. The prevalence of knee osteoarthritis in the elderly. The Framingham Osteoarthritis Study. Arthritis Rheum. 1987;30(8):914–918. [DOI] [PubMed] [Google Scholar]
- 10.Driban JB, Eaton CB, Lo GH, Ward RJ, Lu B, McAlindon TE. Knee injuries are associated with accelerated knee osteoarthritis progression: Data from the Osteoarthritis Initiative. Arthritis Care Res (Hoboken). 2014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Roos H, Adalberth T, Dahlberg L, Lohmander LS. Osteoarthritis of the knee after injury to the anterior cruciate ligament or meniscus: the influence of time and age. Osteoarthritis Cartilage. 1995;3(4):261–267. [DOI] [PubMed] [Google Scholar]
- 12.Englund M, Roos EM, Lohmander LS. Impact of type of meniscal tear on radiographic and symptomatic knee osteoarthritis: a sixteen-year followup of meniscectomy with matched controls. Arthritis Rheum. 2003;48(8):2178–2187. [DOI] [PubMed] [Google Scholar]
- 13.Sharma L, Song J, Felson DT, Cahue S, Shamiyeh E, Dunlop DD. The role of knee alignment in disease progression and functional decline in knee osteoarthritis. Jama. 2001;286(2):188–195. [DOI] [PubMed] [Google Scholar]
- 14.Chehab EF, Favre J, Erhart-Hledik JC, Andriacchi TP. Baseline knee adduction and flexion moments during walking are both associated with 5 year cartilage changes in patients with medial knee osteoarthritis. Osteoarthritis Cartilage. 2014;22(11):1833–1839. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Chang A, Hayes K, Dunlop D, et al. Thrust during ambulation and the progression of knee osteoarthritis. Arthritis Rheum. 2004;50(12):3897–3903. [DOI] [PubMed] [Google Scholar]
- 16.Sharma L, Chang AH, Jackson RD, et al. Varus Thrust and Incident and Progressive Knee Osteoarthritis. Arthritis Rheumatol. 2017;69(11):2136–2143. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.McNeal RL. Aquatic therapy for patients with rheumatic disease. Rheum Dis Clin North Am. 1990;16(4):915–929. [PubMed] [Google Scholar]
- 18.Hall J, Skevington SM, Maddison PJ, Chapman K. A randomized and controlled trial of hydrotherapy in rheumatoid arthritis. Arthritis Care Res. 1996;9(3):206–215. [DOI] [PubMed] [Google Scholar]
- 19.Westby MD. A health professional’s guide to exercise prescription for people with arthritis: a review of aerobic fitness activities. Arthritis Rheum. 2001;45(6):501–511. [DOI] [PubMed] [Google Scholar]
- 20.Alkatan M, Baker JR, Machin DR, et al. Improved Function and Reduced Pain after Swimming and Cycling Training in Patients with Osteoarthritis. J Rheumatol. 2016;43(3):666–672. [DOI] [PubMed] [Google Scholar]
- 21.Lo GH, Driban JB, Kriska AM, et al. Is There an Association Between a History of Running and Symptomatic Knee Osteoarthritis? A Cross-Sectional Study From the Osteoarthritis Initiative. Arthritis Care Res (Hoboken). 2017;69(2):183–191. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Lo GH, McAlindon TE, Kriska AM, et al. Participation in American Football is Associated with Increased Risk for Knee Pain and Osteoarthritis: Data from the Osteoarthritis Initiative. Paper presented at: American College of Rheumatology Annual Scientific Meeting; 21 October 2018, 2018; Chicago, IL. [Google Scholar]
- 23.Kriska AM, Sandler RB, Cauley JA, LaPorte RE, Hom DL, Pambianco G. The assessment of historical physical activity and its relation to adult bone parameters. Am J Epidemiol. 1988;127(5):1053–1063. [DOI] [PubMed] [Google Scholar]
- 24.Nevitt MC, Felson DT, Lester G. The Osteoarthritis Initiative: Protocol for the Cohort Study. 2006; https://oai.epi-ucsf.org/datarelease/docs/StudyDesignProtocol.pdf. Accessed 11/10/2017, 2017. [Google Scholar]
- 25.Chasan-Taber L, Erickson JB, McBride JW, Nasca PC, Chasan-Taber S, Freedson PS. Reproducibility of a self-administered lifetime physical activity questionnaire among female college alumnae. Am J Epidemiol. 2002;155(3):282–289. [DOI] [PubMed] [Google Scholar]
- 26.Initiative O. Central Reading of Knee X-rays for Kellgren & LawrenceGrade and Individual Radiographic Features of Tibiofemoral Knee OA. 2016; https://oai.epi-ucsf.org/datarelease/SASDocs/kXR_SQ_BU_descrip.pdf.
- 27.Altman RD, Gold GE. Atlas of individual radiographic features in osteoarthritis, revised. Osteoarthritis Cartilage. 2007;15 Suppl A:A1–56. [DOI] [PubMed] [Google Scholar]
- 28.Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics. 1977;33(1):159–174. [PubMed] [Google Scholar]
- 29.Cochran WG. Some Methods for Strengthening the Common X2 Tests. Biometrics. 1954;10(4):417–451. [Google Scholar]
- 30.Armitage P. Tests for Linear Trends in Proportions and Frequencies. Biometrics. 1955;11(3):375–386. [Google Scholar]
- 31.Chase NC, Sui X, Blair SN. Comparison of the Health Aspects of Swimming With Other Types of Physical Activity and Sedentary Lifestyle Habits. International Journal of Aquatic Research and Education. 2008;2(2):151–161. [Google Scholar]
- 32.Lo GH, McAlindon TE, Katz JN, et al. Systolic and pulse pressure associate with incident knee osteoarthritis: data from the Osteoarthritis Initiative. Clin Rheumatol. 2017;36(9):2121–2128. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Yoshimura N, Muraki S, Oka H, et al. Accumulation of metabolic risk factors such as overweight, hypertension, dyslipidaemia, and impaired glucose tolerance raises the risk of occurrence and progression of knee osteoarthritis: a 3-year follow-up of the ROAD study. Osteoarthritis Cartilage. 2012;20(11):1217–1226. [DOI] [PubMed] [Google Scholar]
- 34.Clockaerts S, Van Osch GJ, Bastiaansen-Jenniskens YM, et al. Statin use is associated with reduced incidence and progression of knee osteoarthritis in the Rotterdam study. Ann Rheum Dis. 2012;71(5):642–647. [DOI] [PubMed] [Google Scholar]
- 35.Kadam UT, Blagojevic M, Belcher J. Statin use and clinical osteoarthritis in the general population: a longitudinal study. J Gen Intern Med. 2013;28(7):943–949. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Gierman LM, Kuhnast S, Koudijs A, et al. Osteoarthritis development is induced by increased dietary cholesterol and can be inhibited by atorvastatin in APOE*3Leiden.CETP mice--a translational model for atherosclerosis. Ann Rheum Dis. 2014;73(5):921–927. [DOI] [PubMed] [Google Scholar]
- 37.Berger BG, Owen DR. Mood alteration with yoga and swimming: aerobic exercise may not be necessary. Percept Mot Skills. 1992;75(3 Pt 2):1331–1343. [DOI] [PubMed] [Google Scholar]
- 38.Kroenke K, Bair MJ, Damush TM, et al. Optimized antidepressant therapy and pain self-management in primary care patients with depression and musculoskeletal pain: a randomized controlled trial. JAMA. 2009;301(20):2099–2110. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Chappell AS, Desaiah D, Liu-Seifert H, et al. A double-blind, randomized, placebo-controlled study of the efficacy and safety of duloxetine for the treatment of chronic pain due to osteoarthritis of the knee. Pain Pract. 2011;11(1):33–41. [DOI] [PubMed] [Google Scholar]
- 40.Prevention CfDCa. Health Benefits of Water-based Exercise. 2016; https://www.cdc.gov/healthywater/swimming/swimmers/health_benefits_water_exercise.html.