Prevalence of hip and knee osteoarthritis in Europe: a systematic review and meta-analysis

Abstract

Introduction

Osteoarthritis (OA) is a leading cause of pain and disability worldwide, with the hip and knee being the most commonly affected major joints. These forms of OA contribute substantially to individual suffering, reduced quality of life, and increased healthcare costs. While global estimates of OA prevalence are available, there is a paucity of reliable, region-specific data for Europe. This systematic review aims to address the following questions: (1) What is the prevalence of hip and knee OA in Europe? (2) How does this prevalence vary across different regions? (3) How does the reported prevalence differ depending on the diagnostic methods used?

Methods

A systematic literature search was conducted in PubMed and Embase for studies published between January 1, 2000, and November 14, 2024. Study quality and risk of bias were assessed independently using the Joanna Briggs Institute (JBI) checklist. Pooled prevalence estimates were calculated using random-effects meta-analysis, and subgroup analyses were performed by diagnostic method and geographic region. The review was prospectively registered with PROSPERO (CRD42022302976) and conducted in accordance with PRISMA guidelines.

Results

Twenty nine studies were included, comprising data from 6,767,340 individuals for hip OA and 6,805,777 for knee OA. The pooled prevalence of hip OA in Europe was estimated at 6% (95% CI: 3–9%), and for knee OA, 10% (95% CI: 7–14%). Subgroup analysis by diagnostic method revealed that prevalence was highest for hip OA when using Kellgren–Lawrence grade ≥ 2 (11.5%, 95% CI: 5.0–24.5%) and for knee OA when using American College of Rheumatology (ACR) criteria (14.5%, 95% CI: 7.6–25.9%). Subgroup analysis by European region showed that hip OA prevalence ranged from 2% in Eastern Europe to 7% in Southern Europe, while knee OA ranged from 7% in Northern Europe to 19% in Eastern Europe.

Conclusion

Hip and knee OA are common in Europe, though prevalence varies widely across studies and diagnostic approaches.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12891-026-09493-7.

Introduction

By 2030, one in six people globally will be over the age of 60, according to the WHO [1]. With this demographic shift comes a rise in chronic musculoskeletal disorders, particularly osteoarthritis (OA), which is a leading cause of pain, disability, and reduced quality of life worldwide [2]. Among the major joints affected, the hip and knee are most commonly involved, placing a growing burden on individuals and healthcare systems through increased demand for orthopedic services and long-term management strategies [2, 3].

OA is multifactorial, with age as a primary risk factor, but also influenced by obesity, genetic predisposition, smoking, and physical inactivity [2, 4]. Importantly, symptom severity is not determined solely by joint pathology. Psychological, cultural, and social factors also affect how patients experience and report OA-related pain and disability [5].

Although previous studies have reported global estimates of OA prevalence, they often aggregate data across continents, overlooking important regional differences [6, 7]. While some reviews include European data, they frequently lack detailed stratification by joint site, diagnostic method, and geographic subregions within Europe. Because OA can be defined using radiographic, clinical, or self-reported criteria, each reflecting structural, symptomatic, and perceived aspects of the disease, differences in diagnostic approach may influence reported prevalence. As such, current evidence does not provide a clear, comprehensive picture of how OA prevalence differs across European populations, despite well-known differences in healthcare access, lifestyle, diet, and aging demographics across countries.

A clearer understanding of these regional and methodological variations is essential for targeted healthcare planning, as countries or regions with higher prevalence may require proportionally greater investment in preventive, rehabilitative, and surgical services.

This systematic review aims to address the following questions: (1) What is the prevalence of hip and knee OA in Europe? (2) How does this prevalence vary across different regions? (3) How does the reported prevalence differ depending on the diagnostic methods used.

Methods

This systematic review, examining the prevalence of hip and knee OA in Europe, was prospectively registered with PROSPERO (CRD42024610014) and conducted in accordance with the PRISMA guidelines [8].

Eligibility criteria

Studies were eligible if the study population consisted of adults aged 18 and older residing in Europe. Studies had to follow a cross-sectional, cohort, observational, register, or case–control design and report prevalence data. Additionally, only studies published between January 1, 2000, and November 14, 2024 were included, to ensure that prevalence estimates reflected contemporary European populations and healthcare contexts.

Studies were excluded if they were hospital-based or not available in English. Studies that did not specify the method of diagnosis or failed to report prevalence directly, such as those presenting only adjusted percentages without absolute numbers, were also excluded. Furthermore, studies focusing exclusively on high-risk occupational or activity groups (e.g., athletes, soldiers, farmers) were excluded to avoid overestimation of prevalence in the general adult population.

Search strategy and screening

A comprehensive search strategy was conducted across two databases, Embase and PubMed (Fig. 1). The detailed search strategy is available in Appendix A.

Fig. 1.

Prisma flow chart [8]

A two-phase screening process took place (Fig. 1) using the Rayyan software [9]. In the first phase, studies were screened based on their title and abstract to determine if they met the inclusion criteria. Studies that passed the initial screening underwent a full-text review in the second phase. Screening was performed independently by the first two authors, and any disagreements were resolved through discussion or, when necessary, by consulting the senior author. The bibliography of selected articles was also screened by the first two authors for relevant articles that met the inclusion criteria.

Data extraction

Data were collected on various study characteristics, including author, publication year, country, study design, and sample size. Additionally, population demographics such as age, sex, and subgroup characteristics were recorded, along with details on OA definition, such as the criteria used for diagnosis, and prevalence data, including age- and sex-specific rates where available. Study quality indicators were also assessed. To ensure accuracy and consistency, data were extracted independently by two reviewers using a standardized form and entered into a digital database, with discrepancies resolved by consensus and consultation with the senior author.

Quality assessment

The quality of the included studies was assessed using the Joanna Briggs Institute (JBI) critical appraisal checklist for prevalence studies [10]. This checklist evaluates methodological rigor and the potential for bias in study design, conduct, and analysis. Key criteria included the sampling frame, participant recruitment, sample size adequacy, study setting, data analysis, measurement bias, outcome assessment, and response rate.

The first two authors independently conducted the quality appraisal, and disagreements were resolved through discussion with the senior author. A threshold of five or more affirmative responses out of nine on the JBI checklist was prespecified to ensure inclusion of studies that met a minimum methodological standard while maintaining sufficient comparability and breadth of evidence across data sources (Table 1).

Table 1.

Quality assessment using the Joanna Briggs Institute (2014) critical appraisal checklist for prevalence studies [10]

Authors	1. Was the sample frame appropriate to address the target population?	2. Were study participants sampled in an appropriate way?	3. Was the sample size adequate?	4. Were the study subjects and the setting described in detail?	5. Was the data analysis conducted with sufficient coverage of the identified sample?	6. Were valid methods used for the identification of the condition?	7. Was the condition measured in a standard, reliable way for all participants?	8. Was there appropriate statistical analysis?	9. Was the response rate adequate, and if not, was the low response rate managed appropriately?	Overall score
Andrianakos et al. (2006) [11]	YES	YES	YES	NO	UN	YES	YES	YES	YES	7/9
Arslan et al.—hip (2022) [12]	YES	UN	YES	YES	YES	YES	YES	YES	YES	8/9
Arslan et al.—knee (2022) [13]	YES	UN	YES	YES	YES	YES	YES	YES	YES	8/9
Bedson et al. [14] [14]	YES	YES	YES	NO	NO	YES	UN	UN	YES	5/9
Bikbov et al. [15] [15]	YES	UN	YES	YES	YES	UN	UN	YES	YES	6/9
Blanco et al. (2020) [16]	YES	YES	YES	NO	YES	YES	UN	YES	YES	7/9
Carmona et al. [17] [17]	YES	UN	YES	NO	YES	YES	UN	YES	YES	6/9
Cvijetiae et al. (2000)	YES	UN	YES	YES	YES	YES	YES	YES	YES	8/9
Cunha-Miranda et al. [18] [18]	YES	UN	YES	YES	YES	UN	UN	YES	YES	6/9
Duncan et al. [19] [19]	YES	YES	YES	YES	UN	UN	UN	YES	YES	6/9
Grotle et al. (2008) [20]	YES	YES	YES	NO	YES	NO	NO	YES	YES	6/9
Guillemin et al. [21] [21]	YES	YES	YES	NO	NO	UN	YES	YES	YES	6/9
Hashmi et al. (2024) [22]	YES	YES	YES	YES	YES	YES	YES	YES	YES	9/9
Ingvarsson et al. (2001) [23]	YES	YES	YES	NO	YES	YES	YES	UN	YES	7/9
Inoue et al. (2000) [24]	YES	YES	YES	NO	NO	YES	YES	UN	YES	6/9
Jacobsen et al. (2004) [25]	YES	YES	YES	UN	YES	YES	YES	UN	YES	7/9
Jordan et al. (2003) [26]	YES	YES	YES	NO	UN	YES	UN	UN	YES	5/9
Kajos et al. [27] [27]	UN	YES	YES	NO	UN	YES	YES	UN	YES	5/9
Mannoni et al. [28] [28]	YES	YES	YES	NO	UN	YES	YES	YES	YES	7/9
Pereira et al. (2015) [29]	YES	YES	YES	YES	NO	YES	NO	NO	YES	6/9
Postler et al. (2018) [30]	YES	YES	YES	YES	YES	YES	YES	YES	YES	9/9
Poulsen et al. (2011) [31]	YES	UN	YES	NO	NO	YES	YES	NO	YES	5/9
Quintana et al. (2008) [32]	YES	YES	YES	YES	UN	UN	YES	YES	YES	7/9
Rodriguez-Veiga et al. (2019) [33]	YES	YES	YES	YES	YES	YES	UN	YES	YES	8/9
Spitaels et al. (2019) [34]	YES	UN	YES	YES	YES	YES	YES	YES	YES	8/9
Summanen et al. [35] [35]	YES	UN	YES	YES	YES	YES	YES	YES	YES	8/9
Swain et al. (2020) [36]	YES	UN	YES	NO	UN	YES	YES	UN	YES	5/9
Visser et al. [37] [37]	YES	UN	YES	YES	YES	YES	YES	YES	YES	8/9
Wills et al. (2011) [38]	YES	UN	YES	YES	YES	YES	YES	UN	YES	7/9

Statistical analysis

Statistical analyses were conducted using R (version 4.4.1) [39]. Meta-analyses of prevalence data were performed using the meta and metafor packages. Proportions and 95% confidence intervals (CIs) were calculated on a logit-transformed scale (PLOGIT) using inverse variance weighting. A random-effects model with Hartung-Knapp adjustment was applied. Subgroup analyses were conducted based on European regions (Western, Eastern, Northern, Southern Europe) as defined in the UN geoscheme, and by method of diagnosis (Kellgren–Lawrence (KL) grade ≥ 2, American College of Rheumatology (ACR) criteria, self-reported, doctor-diagnosed (including ICD codes)) [40]. Separate meta-analyses were conducted within each diagnostic category. For all analyses, heterogeneity was assessed using the I² statistic and interpreted as low (0–25%), moderate (26–50%), substantial (51–75%), or considerable (> 75%). Forest plots were generated to visualize prevalence estimates, subgroup summaries, and heterogeneity.

Results

A total of 29 studies met the inclusion criteria and were included in the analysis. The pooled prevalence of OA in Europe was estimated at 6% for hip OA (95% CI: 3–9%) and 10% for knee OA (95% CI: 7–14%). Data were heterogeneous for both hip- and knee OA (I² = 100% and I² = 100%, respectively).

Characteristics of included studies

The methodological quality of included studies was generally moderate to high. All studies met the predefined inclusion threshold of five or more affirmative responses on the JBI checklist. Of the 29 included studies, 14 studies reported data on the prevalence of both hip and knee OA. In total, 21 studies reported data on hip OA, comprising a combined sample of 6,767,340 individuals [11, 12, 15, 16, 18–25, 27–32, 35, 36, 41], while 23 studies reported on knee OA, with a total of 6,805,777 individuals [11, 13–22, 26, 28–30, 32–38, 41]. Tables 2 and 3 provide an overview of the selected studies and the data collected for hip and knee OA, respectively.

Table 2.

Hip OA—Study characteristics and participants demographics [11, 12, 15, 16, 18–25, 27–32, 35, 36, 41]

Study	Country	Sample (n)	Females (n)	Hip OA (n)	Females with hip OA (n)	BMI (kg/m2)	Age	Method of diagnosis	Time of data collection	Type of study	Sampling method
Andrianakos et al. (2006) [11]	Greece	8,740	N/A	80	N/A	N/A	≥ 19 Mean age 46.9, SD = 17.7	ACR	1996–1999	Cross-sectional	Stratified and random sampling
Arslan et al. (2022) [12, 13]	Netherlands	767,159	398,827	31,444	20,308	N/A	≥ 30 Mean age OA 68.2, SD = 11.7	ICPC code L89	2019	Registry-based study	N/A
Bikbov et al. [15] [15]	Russia	5,899	3,319	157	N/A	N/A	Mean age 59.0 ± 10.7	Self-reported (interview)	2015–2017	Cross-sectional	N/A
Blanco et al. (2020) [16]	Spain	3,336	N/A	396	N/A	N/A	≥ 40	Self-reported (telephonic questionnaire)	2000	Cross-sectional	Multistage and stratified random cluster sampling
Cvijetiae et al. (2000)	Croatia	610	306	140	57	Females: 28.6 ± 4.3 Males: 28.1 ± 3.4	> 45 Females: Mean age 63.6, SD = 10.9Males: Mean age 62.6, SD = 10.4	KL grade ≥ 2	1981–1983	Cross-sectional	Stratified sampling
Cunha-Miranda et al. [18]	Portugal	1,039	563	23	N/A	N/A	> 45	Self-reported (face to face interviews)	2011	Cross-sectional	Stratified sampling
Duncan et al. [19]	UK	1,029	660	180	131	N/A	Mean age 85	Doctor-diagnosed	2006	Cohort-study	Stratified sampling
Grotle et al. (2008) [20]	Norway	3,266	1,786	179	111	N/A	24–76	Self-reported- doctor diagnosed	2004	Cross-sectional	Stratified sampling
Guillemin et al. [21]	France	21,195	13,789	728	N/A	N/A	Participants: 40–49: 6856 50–59: 6600 60–69: 5144 70–75: 2595	Self-reported (validated questionnaire)	2007–2009	Population-based study	Random sampling
Hashmi et al. (2024) [22]	UK	285,947	148,793	10,698	5,895	27.3 ± 4.7	Mean age 52.7, SD = 7.1	ICD-9/10 codes	2022	Cross-sectional	N/A
Ingvarsson et al. (2001) [23]	Iceland	1,517	N/A	165	88	N/A	Mean age 68	KL grade ≥ 2	1990–1996	Population-based cohort study	N/A
Inoue et al. (2000) [24]	France	401	118	19	3	N/A	20–39: 42 males, 44 females 40–59: 102 males, 41 females 60–79: 139 males, 33 females	KL grade ≥ 3	1994	Cross-sectional	N/A
Jacobsen et al. (2004) [25]	Denmark	3,792	2,293	275	118	N/A	22–93 < 60: 1448 ≥ 60: 2344	KL grade ≥ 2	1994	Cross-sectional	Stratified random sampling
Kajos et al. [27]	Hungary	100,000	N/A	2,238	N/A	N/A	≥ 30	ICD-10 codes	2018	Database-study	N/A
Mannoni et al. [28]	Italy	697	406	81	N/A	N/A	≥ 65 Mean age 74.1, SD = 6.8	ACR	N/A	Cross-sectional	Stratified sampling
Pereira et al. (2015) [29]	Portugal	775	510	53	N/A	N/A	Mean age 56.6, SD = 15.4	Self-reported (structured questionnaire)	2005–2008	Cohort-study	Stratified random sampling
Postler et al. (2018) [30]	Germany	2,728,100	1,744,067	266,538	182,336	N/A	With OA: (mean age 74.9) 60–69: 68,784/1090251 70–79: 121,442/1097789 80–89: 63,609/446057 90–99: 12,447/91770  ≥ 100: 256/2233	ICD-10 codes	2014	Cross-sectional	De-identified claims data
Poulsen et al. (2011) [31]	Denmark	1000	N/A	192	N/A	N/A	≥ 40	Radiographic signs	2007	Cross-sectional	N/A
Quintana et al. (2008) [32]	Spain	7,577	4,264	1,398	N/A	N/A	60–69: 3730 70–79: 2886  ≥ 80: 961	Self-reported (KHOA-SQ questionnaire)	2002–2003	Cross-sectional	Stratified random sampling
Summanen et al. [35]	Finland	1,134,643	N/A	9,040	4,902	28.7	Mean age OA 56.8	ICD-10 codes	2020	Registry-based study	N/A
Swain et al. (2020) [36]	UK	1,690,618	N/A	25,359	N/A	N/A	≥ 20	ICD-10 codes	2017	Longitudinal	N/A

Table 3.

Knee OA—Study characteristics and participants demographics [11, 13–22, 26, 28–30, 32–38, 41]

Study	Country	Sample (n)	Females (n)	Knee OA (n)	Females with knee OA (n)	BMI (kg/m2)	Age	Method of diagnosis	Time of data collection	Type of study	Sampling method
Andrianakos et al. (2006) [11]	Greece	8,740	N/A	547	N/A	N/A	≥ 19 Mean age 46.9, SD = 17.7	ACR	1996–1999	Cross-sectional	Stratified and random sampling
Arslan et al. (2022) [12, 13]	Netherlands	767,159	398,827	48,221	30,929	N/A	≥ 30 Mean age OA 66.8, SD = 11.9	ICPC code L90	2019	Registry-based study	N/A
Bedson et al. [14]	England	6,102	N/A	146	97	N/A	≥ 45 Mean age OA 66.8, SD = 8.6	Doctor-diagnosed	2000	Case–control	Stratified sampling
Bikbov et al. [15]	Russia	5,899	3,319	1,147	N/A	N/A	Mean age 59.0 ± 10.7	Self-reported (interview)	2015–2017	Cross-sectional	N/A
Blanco et al. (2020) [16]	Spain	3,336	N/A	649	N/A	N/A	≥ 40	Self-reported (telephonic questionnaire)	2000	Cross-sectional	Multistage and stratified random cluster sampling
Carmona et al. [17]	Spain	2,192	1,178	223	165	N/A	≥ 20 With OA: 20–29: 2/463 30–39: 3/439 40–49: 13/371 50–59: 32/326 60–69: 88/313 70–79: 69/205  ≥ 80: 16/75	ACR	2000	Cross-sectional	Cluster sampling
Cvijetiae et al. (2000)	Croatia	610	306	43	30	Females: 28.6 ± 4.3 Males: 28.1 ± 3.4	> 45 Females: Mean age 63.6, SD = 10.9 Males: Mean age 62.6, SD = 10.4	KL grade ≥ 2	1981–1983	Cross-sectional	Stratified sampling
Cunha-Miranda et al. [18]	Portugal	1,039	563	65	N/A	N/A	> 45	Self-reported (face to face interviews)	2011	Cross-sectional	Stratified sampling
Duncan et al. [19]	UK	1,029	660	315	217	N/A	85	Doctor-diagnosed	2006	Cohort-study	Stratified sampling
Grotle et al. (2008) [20]	Norway	3,266	1,786	233	141	N/A	24–76	Self-reported- doctor diagnosed	2004	Cross-sectional	Stratified sampling
Guillemin et al. [21]	France	21,195	13,789	1,093	N/A	N/A	40–49: 6856 50–59: 6600 60–69: 5144 70–75: 2595	Self-reported (validated questionnaire)	2007–2009	Population-based study	Random sampling
Hashmi et al. (2024) [22]	UK	285,947	148,793	18,578	8,955	27.3 ± 4.7	Mean age 52.7, SD = 7.1	ICD-9/10 codes	2022	Cross-sectional	N/A
Jordan et al. (2003) [26]	England	4,566	N/A	828	N/A	N/A	≥ 55	Doctor-diagnosed	2003	Cross-sectional	Stratified sampling
Mannoni et al. [28]	Italy	697	406	159	N/A	N/A	≥ 65 Mean age 74.1, SD = 6.8v	ACR	N/A	Cross-sectional	Stratified sampling
Pereira et al. (2015) [29]	Portugal	775	510	106	N/A	N/A	Mean age 56.6, SD = 15.4	Self-reported (structured questionnaire)	2005–2008	Cohort-study	Stratified random sampling
Postler et al. (2018) [30]	Germany	2,728,100	1,744,067	425,760	302,862	N/A	With OA: (mean age 74.9) 60–69: 118,748/1090251 70–79: 186,035/1097789 80–89: 101,535/446057 90–99: 19,047/91770  ≥ 100: 395/2233	ICD-10 codes	2014	Cross-sectional	De-identified claims data
Quintana et al. (2008) [32]	Spain	7,577	4,264	2,275	N/A	N/A	60–69: 3730 70–79: 2886  ≥ 80: 961	Self-reported (KHOA-SQ questionnaire)	2002–2003	Cross-sectional	Stratified random sampling
Rodriguez-Veiga et al. (2019) [33]	Spain	707	398	206	N/A	29.1 ± 9.3	> 40 Mean age 61.8, SD = 23.3	ACR	N/A	Cross-sectional	Random sampling
Spitaels et al. (2019) [34]	Belgium	123,261	64,420	5,049	3,232		Mean age OA 56.9	ICPC-2 code	2015	Registry-based study	N/A
Summanen et al. [35]	Finland	1,134,643	N/A	42,028	23,849	30.4	Mean age OA 56.6	ICD-10 codes	2020	Registry-based study	N/A
Swain et al. (2020) [36]	UK	1,690,618	N/A	49,028	N/A	N/A	≥ 20	ICD-10 codes	2017	Longitudinal	N/A
Visser et al. [37]	Netherlands	5,284	2,794	991	685	29.9 (27.8–32.8)	56	ACR	2008–2012	Cross-sectional	Stratified sampling
Wills et al. (2011) [38]	UK	3,035	1,563	302	194	N/A	53	ACR	1999	Cohort-study	N/A

Regional prevalence of OA in Europe

Hip OA

The included studies were distributed across various European regions: two from Eastern Europe (Hungary and Russia) [15, 27], eight from Northern Europe (one from Finland, one from Norway, three from the UK, one from Iceland, and two from Denmark) [19, 20, 22, 23, 25, 31, 35, 36], seven from Southern Europe (two from Spain, one from Greece, one from Italy, one from Croatia, and two from Portugal) [11, 16, 18, 28, 29, 32, 41], and four from Western Europe (one from the Netherlands, one from Germany, and two from France) [12, 21, 24, 30]. A detailed breakdown is presented in the forest plot shown in Fig. 2.

Fig. 2.

The forest plot of the prevalence of hip OA in Europe [11, 12, 15, 16, 18–25, 27–32, 35, 36, 41]

Southern Europe reported the highest prevalence of hip OA at 7% (95% CI: 2–20%), while Eastern Europe showed the lowest prevalence at 2% (95% CI: 1–7%).

Knee OA

One study was included from Eastern Europe (Russia) [15], eight from Northern Europe (one from Finland, one from Norway, and six from the UK) [14, 19, 20, 22, 26, 35, 36, 38], nine from Southern Europe (four from Spain, one from Greece, one from Italy, one from Croatia, and two from Portugal) [11, 16–18, 28, 29, 32, 33, 41], and five from Western Europe (one from Belgium, two from the Netherlands, one from Germany, and one from France) [13, 21, 30, 34, 37].

The highest prevalence of knee OA was observed in Eastern Europe at 19% (95% CI: 18–20%), Northern Europe had the lowest reported prevalence at 7% (95% CI: 3–16%). A detailed analysis is presented in the forest plot shown in Fig. 3.

Fig. 3.

The forest plot of the prevalence of knee OA in Europe [11, 13–22, 26, 28–30, 32–38, 41]

Prevalence of OA in Europe per diagnostic method

Hip OA

OA was diagnosed using various methods across the included studies (Table 4): two studies employed the ACR (American College of Rheumatology) criteria, seven relied on doctor-diagnosed OA, five used radiographic assessment with Kellgren-Lawrence (KL) grade greater than two, and seven studies were based on self-reported OA. The highest prevalence was observed in studies using KL-grade criteria, at 11.5%, while the lowest prevalence was reported in studies using the ACR criteria, at 3.4%.

Knee OA

Across the included studies, OA was diagnosed using a range of methods (Table 5): six studies applied the ACR criteria, ten studies relied on doctor-diagnosed OA, one study used radiographic assessment with KL grade greater than two, and six studies were based on self-reported OA. The highest prevalence was observed in the ACR group at 14.5, while the lowest was reported in the KL grade study at 7.1%.

Discussion

The primary aim of this systematic review was to provide an overview of the prevalence of hip and knee OA across Europe. A total of 29 studies were included, encompassing data from over 6.7 million individuals for hip OA and 6.8 million individuals for knee OA.

Prevalence of hip and knee OA in Europe

The pooled prevalence of hip OA in Europe was estimated at 6% (95% CI: 3–9%), while knee OA prevalence was higher at 10% (95% CI: 7–14%). However, prevalence estimates varied substantially depending on the diagnostic criteria employed. The use of the KL grading system with a threshold of ≥ 2 yielded the highest prevalence for hip OA (11.5%, 95% CI: 5.0–24.5%), while the lowest prevalence for knee OA was found with the same criterion (7.1%, 95% CI: 5.3–9.5%). These findings highlight the influence of diagnostic criteria on reported prevalence [42].

​The KL classification is the most widely used radiographic tool for grading OA [43]. However, studies have consistently shown that KL grades do not always align with symptom severity. De Polo et al. found that pain intensity scores reported by patients with knee OA are largely unrelated to radiographic severity grades and that patients with higher OA grades use similar types of pain medication than patients with lower OA grades [44].

Differences in diagnostic definitions have major implications for epidemiological interpretation. Radiographic OA captures structural changes, often preceding symptoms, while self-reported or doctor-diagnosed OA may better reflect symptomatic disease burden relevant to healthcare use. Consequently, comparisons across studies using divergent definitions should be made cautiously.

Geographic variation in hip and knee OA prevalence among European regions

Marked regional variation in OA prevalence was observed across Europe. Hip OA prevalence ranged from 2% in Eastern Europe to 7% in Southern Europe, while knee OA ranged from 7% in Northern Europe to 19% in Eastern Europe.

Multiple contextual factors are likely to contribute to these disparities. Differences in demographic profiles, such as age distribution, alongside the varying prevalence of risk factors such as obesity, physical inactivity, and occupational exposures, are important considerations [2].

Healthcare system characteristics may also influence observed prevalence. A study focused on Nordic European countries reported substantial variation in the number of orthopedic surgeons per capita and annual hip arthroplasty volumes, despite otherwise comparable socioeconomic contexts. Such differences suggest that diagnostic thresholds, healthcare access, and system-level priorities may shape who receives a diagnosis and when [45].

Behavioral and psychosocial determinants further modulate OA detection and symptom severity. Cultural attitudes toward aging and joint pain, fear of medical diagnosis or surgery, and adaptive behaviors such as self-medication or avoidance of care can delay diagnosis. These behaviors are often influenced by socioeconomic status, healthcare accessibility, and previous interactions with medical systems. As a result, certain populations may be systematically underdiagnosed, leading to underrepresentation in prevalence data [46–48].

Implications

Given the projected rise in the prevalence of OA in the coming years, it is imperative that current epidemiological findings inform and shape public health strategies [2, 49]. Targeted efforts to mitigate established risk factors are essential to reducing disease incidence. Moreover, early identification and intervention could play a crucial role in slowing progression [50].

When translating prevalence findings into healthcare planning, the choice of diagnostic definition becomes highly relevant. Symptom-based or doctor-diagnosed OA provides estimates that most closely reflect clinical demand and resource utilization, while radiographic definitions capture the broader structural burden of disease, including individuals who may not yet seek medical care. Both perspectives are valuable: the former for immediate service planning, and the latter for anticipating future needs and guiding preventive strategies. Ensuring alignment between epidemiological definitions and healthcare objectives will enhance the utility of prevalence research for policy development and long-term planning.

Implementing effective conservative management plans at the primary care level may alleviate some of the anticipated burden on secondary and tertiary healthcare services [51]. Ultimately, such measures can contribute to improved patient outcomes and enhanced quality of life for individuals affected by OA.

Limitations

This review has several limitations. First, substantial heterogeneity across studies, with I² values exceeding 99% in some subgroups, limits the precision of the pooled estimates. This heterogeneity likely stems from differences in study design, diagnostic criteria, and population characteristics. The quality of studies included in the analysis varied, and inconsistent reporting of recruitment procedures may have introduced selection bias, affecting the accuracy of prevalence estimates. As such, reported means should be interpreted as indicative summaries rather than exact population values.

Second, the potential confounding between diagnostic methods and geographic regions should be considered. Certain regions may preferentially use radiographic, clinical, or self-reported criteria, which may explain part of the observed regional variation in prevalence. Furthermore, differences in age distributions across study populations could have influenced regional prevalence estimates, but we were unable to adjust for demographic differences due to limited reporting of age-stratified data.

Third, publication bias could not be assessed, as the number of studies per subgroup was small, and heterogeneity was high. Additionally, the exclusion of hospital-based cohorts and high-risk groups (e.g., athletes, soldiers) means the estimates are not directly generalizable to these subgroups.

Finally, the inclusion period (2000–2024) spans more than two decades, during which demographic, diagnostic, and risk-factor trends may have shifted. Although only a small number of early studies were included, temporal changes should be considered when interpreting pooled estimates. The regional classification used follows the United Nations geoscheme, but we acknowledge that this broad grouping may not fully reflect the geographic, cultural, or genetic diversity within Europe. Furthermore, the lack of more studies from Eastern Europe reduces the precision of regional estimates.

Despite these limitations, this review fills a critical gap in the literature by providing region-specific prevalence estimates for hip and knee OA in Europe. Previous reviews have typically pooled data across continents or large geographic regions, obscuring intra-continental differences. Our study offers more granular insights into the burden of OA by region and diagnostic method, which are directly applicable to healthcare planning. As Europe faces significant demographic shifts, with an aging population and rising rates of modifiable risk factors like obesity and inactivity, accurate and up-to-date prevalence data is essential for forecasting demand for orthopedic services, planning surgical capacity, and implementing preventive strategies.

Conclusion

This review found that knee and hip OA are common in Europe. However, substantial variation exists across studies, influenced by differences in diagnostic methods, study design, and regional healthcare contexts. Although some population-based data are available, they remain limited, and individuals who do not seek healthcare are often underrepresented in the current literature. To produce more accurate and inclusive estimates, future studies should explicitly aim to capture this overlooked segment of the population.

Abbreviations

OA

Osteoarthritis

ACR

American College of Rheumatology

KL grade

Kellgren–Lawrence grade

WHO

World Health Organization

JBI

Joanna Briggs Institute

CIs

Confidence intervals

Authors’ contributions

I.C. and I.V. performed the screening and data collection and drafted the manuscript. H.K. contributed to the study structure and preparation of figures and tables. M.B. contributed to methodological input and manuscript revision. C.V. and J.D. supervised the study. All authors reviewed and approved the final manuscript.

Funding

The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

Data availability

All data generated or analyzed during this study are included in this published article and its supplementary information files.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors have no relevant financial or non-financial interests to disclose.