The association of physical activity, obesity and injury on the risk of knee osteoarthritis

Soutakbar, Hessam (2017) The association of physical activity, obesity and injury on the risk of knee osteoarthritis. PhD thesis, University of Nottingham.

PDF (Thesis - as examined) - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Download (10MB) | Preview



The association of physical activity, obesity and injury on the risk of knee osteoarthritis (OA)


1) To examine the effect of interactions between physical activity, obesity and injury on the incidence and progression of radiographic and symptomatic knee OA;

2) To establish age and gender specific normative data for knee pain, symptoms, function and knee related quality of life (QOL) as the clinical outcome measures in assessing people with knee OA and to examine their associations with OA risk factors including obesity, injury and physical activity.


1) Using existing cohort data from Osteoarthritis Initiative (OAI) and Multicenter Osteoarthritis Study (MOST) for interaction analyses

Participants without radiographic knee OA at baseline were followed for the incidence of radiographic and symptomatic knee OA. In OAI, the focus was on the tibiofemoral joints (TF) only, so TF-OA was defined as a knee with a Kellgren and

Lawrence (KL) grade 2 or greater. In MOST, knee OA was defined as a knee with TF-OA (KL ≥2) and/or patellofemoral- OA (osteophyte ≥2; or joint space narrowing ≥1 plus any cyst, osteophyte, or sclerosis using Osteoarthritis Research Society International atlas). The co-occurrence of radiographic knee OA and the frequent knee symptoms (pain, ache, or stiffness on most days of a month over the past 12 months) at the last follow-up was considered as the incidence of symptomatic knee OA.

Progression of radiographic knee OA was determined as either one grade increase in KL score or one grade worsening in joint space narrowing at the last follow-up, in participants with radiographic knee OA at baseline. For the progression of symptomatic knee OA, participants with frequent knee symptoms at baseline were included. An increase of greater than 9.29 points in the total Western Ontario and McMaster Universities Osteoarthritis Index score from baseline to last follow-up was considered as a cut-off point (minimal clinical important worsening) for considering a person with symptom progression. Body mass index (obese/non-obese), injury (yes/no), physical activity (active/inactive), age and gender data were also collected at baseline in both databases. The measures of interactions on both additive and multiplicative scales were computed using the generalized estimation equation.

2) Establishing age and gender specific reference values data for Knee Injury and Osteoarthritis Outcome Score (KOOS) and Oxford Knee Score (OKS)

Volunteer participants were recruited via a postal survey. From a list of 25,695 postcodes specified by Nottinghamshire local authorities and in the City of Nottingham, 2,500 postcodes were randomly selected. This was based on the proportion of the population in each local authority and in the City of Nottingham. 2,500 postcodes were then equally and randomly assigned into three age groups of 18-44, 45-69 and ≥70 years old. From each postcode assigned to the specific age group, one name and address was randomly selected. Participants were required to complete the questionnaire booklet once only. The questionnaire booklet consisted of the OKS and the KOOS questionnaires. It also collected information regarding participants’ age, gender, height, weight, history of injury and knee joint replacement and physical activity.


Interaction analysis

In both cohorts, active and inactive people had a similar risk of incident radiographic or symptomatic knee OA (p >0.05). This effect was not modified by obesity and/or injury in either cohort (p interactions >0.05). No significant interactions were also found between physical activity, obesity and injury on the risk of radiographic or symptomatic knee OA progression (p interaction >0.05). Obese people in both cohorts were significantly at a higher risk of incident radiographic and symptomatic knee OA when compared to non-obese people (p <0.01); injury also increased the incident risk of knee OA (p <0.01). There were some evidence of positive interactions between obesity and injury on the risk of incident knee OA. This reached statistical significance on additive and multiplicative scales in OAI (aOR-Symptomatic-multiplicative interaction: 2.83, 95%CI: 1.01 to 7.93; aOR-Symptomatic-additive interaction: 3.13, 95%CI: 0.05 to 6.21) and on additive scale in MOST (aOR- Radiological-additive interaction: 1.51, 95%CI: 0.10 to 2.93). There was no evidence of any statistically significant interaction between obesity and injury on the progressive risk of knee OA.

Reference values data

The overall response rate was 16.5% (n =414, 45% male, 55% female), with the highest in the middle age group with 24%, 18% in the old age and 8% in young age group. A significant dose response relationship was seen between increasing age and worsening scores of KOOS-Pain; KOOS- Activities of daily living (ADL); KOOS-QOL; and OKS (p <0.05). The median (M) and inter quartile range (IQ) in old, middle and young age groups were as follows: KOOS-Pain (M, IQ: 91.6, 58.3-100; 94.4, 77.7-100; 100, 80.5-100), KOOS- ADL (M, IQ: 91.1, 59.3-100; 98.5, 77.2-100; 100, 89.7-100), KOOS-QOL (M, IQ: 81.2, 43.7-100; 87.5, 62.5-100; 87.5, 68.7-100), and OKS (M,IQ: 42.3, 29-48; 46, 38-48; 47, 42- 48). The oldest age group had the worst scores in KOOS-Pain, KOOS-ADL; KOOS-QOL; and OKS compared to the young or middle age groups (p <0.05). However, the differences between young and middle age groups were not statistically significant in any KOOS or OKS scores (p >0.05). Data were also stratified by gender. There was no gender difference in any KOOS or OKS scores (p >0.05). Obesity and injury were also found as the strongest predictors for the worsening score in all KOOS and OKS subscale scores (p <0.05), whereas physical activity was significantly associated with a lower risk of knee related complaints (p <0.05).


Physical activity did not increase the risk of incident or progressive knee OA at any level of obesity and/or injury in middle aged and older people with or at high risk of knee OA. In addition, meeting the minimum physical activity guidelines was significantly associated with lower self-reported knee complaints evaluated by KOOS and OKS. Therefore, moderate levels of physical activity appears to be safe to recommend to the general population and people with or at high risk of knee OA regardless of obesity and injury status. There was also some modest evidence of positive interaction between obesity and injury on the risk of incident knee OA. Hence, weight gain prevention strategies may protect injured people against further increase in the risk of knee OA.

This study also provided normative data for KOOS and OKS. The self-reported knee complaints were found to vary with age (not gender) being highest in the oldest age group. This suggests that treatment outcomes in people with knee injury and knee OA should be compared against age-matched reference values from the general population.

Item Type: Thesis (University of Nottingham only) (PhD)
Supervisors: Edwards, Kimberley
Hodgson, Lisa
Scammell, Brigitte
Keywords: physical activity, obesity, exercise, injury, knee osteoarthritis, oa
Subjects: W Medicine and related subjects (NLM Classification) > WE Muscoskeletal system
Faculties/Schools: UK Campuses > Faculty of Medicine and Health Sciences > School of Clinical Sciences
UK Campuses > Faculty of Medicine and Health Sciences > School of Medicine
Item ID: 40281
Depositing User: Soutakbar, Hessam
Date Deposited: 29 Aug 2017 13:34
Last Modified: 13 Oct 2017 20:05

Actions (Archive Staff Only)

Edit View Edit View