Global management of patients with knee osteoarthritis begins with quality of life assessment: a systematic review

We identified original articles using electronic searches of MEDLINE, EMBASE, Cochrane and PsycINFO databases. Literature review start date was unrestricted and end date was January 23, 2017. Searches were not limited by language. However, no eligible study was found in non-English languages. The keywords used were “knee osteoarthritis” AND “quality of life” OR “life quality” OR “wellbeing” OR “well-being” OR “short form 36” OR “knee injury and osteoarthritis outcome score” OR “koos” OR “koos-qol” OR “euroqol” OR “assessment of quality of life” OR “qualitymetric” OR “whoqol-100” OR “quality of wellbeing” OR “rosser” OR “osteoarthritis quality of life scale” OR “osteoarthritis knee and hip quality of life” OR “arthritis impact measurement scale” and all shorter forms and variations. The systematic review was conducted following Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines [9].

An example of the tag flowchart for inclusion and exclusion can be found in online Additional file 1: Figure S1. Only abstracts or articles reporting original data on QoL of KOA patients were included. Inclusion criteria were QoL associated to one or more demographic factors (e.g., age, gender), lifestyle factors (e.g., functional independence), or comorbidity factors (e.g., diabetes, obesity) and compared with a reference population, or control group. The control group was composed of individuals without KOA. There were no other restrictions on the comparison control group. There was no restriction on age, gender, language, or year of publication. Review articles, protocols for clinical trials, commentaries, editorials, proceedings summaries, or instrument development summaries were excluded from this review. Articles that described unspecified knee pain, pre-intervention anterior cruciate ligament repair, hip osteoarthritis (OA), spine OA or any study that combined KOA patients with other cohorts of patients and did not collect, analyse, or report KOA-specific data separately (e.g., a population defined as “hip or KOA” or “hip and/or KOA”) were also excluded. Three reviewers independently assessed each reference against pre-specified inclusion and exclusion criteria using a two-stage process: first, titles and abstracts, and, second, full-text articles. Any queries were resolved during a consensus meeting.

A single reviewer, using a pre-piloted extraction form, obtained the data for each eligible article. Study characteristics including publication details (author and year), participant characteristics (age, sex, body mass index [BMI] and number of participants in each group), instruments, treatments applied in the intervention and control groups, and a summary of main findings were extracted from each included study for subsequent review amongst all three reviewers.

The quality assessment of each article was based on a modified version of the Cochrane quality appraisal tool [10]. The individual assessment criteria are presented in Fig. 1. Strength of consistency between raters was not scored yet individual and average quality assessment results are included in online Additional file 2. One point was allocated for each of the 13 quality appraisal criteria. The maximum score was 13 (indicating high quality), with the lowest possible score being zero. The methodological quality of each study was rated as low (0–4 points), moderate (5–9 points), or high (10–13 points).

A total of 9143 articles were initially identified (Fig. 2); 4863 articles from EMBASE, 2792 from PubMed, 1279 from Cochrane and 209 from PsycINFO. A total of 610 articles were selected after initial title and abstract screening. After manual searches, full text review and removal of duplicates, 62 articles were included for final data extraction. Studies were labelled by first author and year of publication (Table 1). Year of publication ranged from 1995 to 2017.

Most of the studies were conducted in Europe (n = 20), followed by the North, Central and South America (n = 16), Asia (n = 12) and other (n = 14). The 74% (n = 46) of studies were cross-sectional in design, followed by other designs comprising 6% (n = 4) prospective cohorts, 5% (n = 3) prospective, 3% (n = 2) randomized controls trials, (n = 1) population based cohorts, (n = 1) retrospective, etc. The main instruments used to assess QoL were the SF-36 (n = 31) followed by EQ-5D (n = 6) (Table 1). The results from the quality appraisal tool had all articles scoring moderate to high quality scores; no study scored lower than 8 points. Four (6%) of the articles [25, 28, 54, 60] were classified as moderate quality, with the remaining 58 (94%) articles classified as having high quality [11‐24, 26, 27, 29‐53, 55‐59, 61‐72].

The total KOA population was 24,706 of which 93.4% (n = 23,079) were female [11‐72]. 11 articles included only females [19, 26, 28, 37, 41, 43, 53, 55, 56, 60, 63] and one article included only males [64]. When the male and female only articles were removed, the total KOA population still comprised 93.8% female. The mean number of KOA patients per study was 560, with sample size varying between 12 [43] and 7977 [31] and the median number was 101. The mean age across all studies was 65 years (range 47 to 85) and 68.2 (range 56.9 to 71.1) years for those studies including only females [19, 26, 28, 37, 41, 43, 53, 55, 56, 60, 63] and 50 (range 50 to 69) for the one study enrolling only males (Table 1) [64].

Studies reported worse QoL for the KOA group, regardless of measurement used to assess QoL [12, 14, 22, 28, 31, 33, 38, 46, 53, 60, 61, 63, 65‐67, 69]. Lower QoL scores were mostly reported with increasing age [25, 26, 32, 35, 44, 53] yet Jenkins reported higher QoL in older patients [57]. When QoL scores were compared based on gender, females with KOA reported worse QoL scores and psychosocial variables [49, 58, 66]. In online Additional file 3: Table S1 additional results are presented for studies comparing QoL of KOA patients with a reference population by instrument.

Weight was reported as effecting QoL in KOA patients [11, 20, 46, 67, 70]. Vulcano, Elbaz and Visser reported a higher BMI aggravated KOA patient symptoms [11, 46, 51]. Gomez-Neto found a negative impact on functional capacity in obese KOA patients; however, found no difference reported in QoL [20]. de Leeuw, reported that the median pre-operative QoL scores for obese patients were significantly lower than for non-obese [70].

Calvacante et al. found that KOA in older women can promote a decline in time spent performing physical activity and functional fitness with decline in QoL and an increase in sitting time [26]. Wantanabe, reported reduced physical activity resulted in worse QoL and also reported too much physical activity may exacerbate the development of KOA [60]. Strength training (ST) had a positive effect in KOA patients [17]. Reid found that muscle power is an independent determinant of QoL in KOA [34].

Coleman evaluated a self-management educational program delivered by health care professionals, reporting an improved QoL in KOA patients assessed at 8 weeks and 6 months based on WOMAC and SF-36 measures [48].

Higher educational level and higher total mindfulness scores were reported to improve patients’ perception of QoL [13, 23, 28]. Alkan et al. reported that approximately 70% of the study participants had low-middle education, resulting in poor QoL in this group [38]. Another study reported lower Health-Related Quality of Life (HRQL) scores for KOA females with chronic physical or mental health conditions [62]. Poverty also worsened QoL in KOA patients [45]. Psychological distress and depression worsened QoL [23, 28, 42, 54]. Patients lacking familial relationships reported worse QoL [50].

The majority of QoL findings with surgical intervention were reported as dependents of demographic variables, which showed an effect on QoL after surgical outcomes. Williams et al. [47], for example, reported that patient satisfaction was lower among patients younger than 55 years of age compared to older patients. Vulcano, reported a higher BMI was associated with worse outcomes [46]. Perez-Prieto, reported patients with depression had less improvement that non-depressed patients after surgical intervention, reporting lower QoL scores [42]. Lower socioeconomic groups undergoing surgical intervention also reported worse QoL [45]. Patients participating in sports pre- and post- surgical intervention reported higher QoL scores [40].

There are limitations to acknowledge and use as opportunities to improve quality of future OA research and reporting to have a more meaningful impact for OA patients. The database search did not include SPORTDiscus or Cumulative Index to Nursing and Allied Health Literature (CINAHL) which may have provided additional QoL study data, particularly assessing QoL related to healthy weight, physical activity, educational and psychosocial factors. The search was limited to 2017 and additional studies may have been published thereafter reporting on QoL in KOA patients. Caution should be used when interpreting these findings, as studies were included which may not have been scientifically rigorous upon review, yet methodological reporting sufficed to meet the data extraction criteria. For example, a moderate to high score could be obtained for providing an adequate description of most criteria even if the sample size and statistical methods weren’t methodologically and analytically rigorous. Articles were included whether power analysis was performed on QoL as the primary outcome or not; thus, sample size may not have been sufficient to find differences when not powered for QoL as the primary outcome. Risk of bias was not calculated. The methodological heterogeneity of the studies limited unbiased comparisons and quantitative syntheses was not permissible.