Objective To investigate the effect of massive weight loss on (1) knee pain and disability, (2) low-grade inflammation and metabolic status and (3) joint biomarkers in obese patients with knee osteoarthritis (OA).
Methods 140 patients involved in a gastric surgery programme were screened for painful knee OA, and 44 were included (age 44 ± 10.3 years, body mass index (BMI) 50.7 ± 7.2 kg/m2). Clinical data and biological samples were collected before and 6 months after surgery.
Results Before surgery, interleukin 6 (IL-6) levels were correlated with levels of high-sensitivity C reactive protein (hsCRP) (p=0.006) and Helix-II (p=0.01), a biomarker of cartilage turnover, and the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) function score (p=0.03). Surgery resulted in substantial decrease in BMI (−20%). Levels of insulin and insulin resistance were decreased at 6 months. Knee pain decreased after surgery (24.5 ± 21 mm vs 50 ± 26.6 mm; p<0.001), and scores on all WOMAC subscales were improved. Levels of IL-6 (p<0.0001), hsCRP (p<0.0001), orosomucoid (p<0.0001) and fibrinogen (p=0.04) were decreased after surgery. Weight loss resulted in a significant increase in N-terminal propeptide of type IIA collagen levels (+32%; p=0.002), a biomarker of cartilage synthesis, and a significant decrease in cartilage oligomeric matrix protein (COMP) (−36%; p<0.001), a biomarker of cartilage degradation. Changes in COMP concentration were correlated with changes in insulin levels (p=0.02) and insulin resistance (p=0.05).
Conclusion Massive weight loss improves pain and function and decreases low-grade inflammation. Change in levels of joint biomarkers with weight loss suggests a structural effect on cartilage.
Statistics from Altmetric.com
Obesity is the main modifiable risk factor for the onset of knee osteoarthritis (OA).1,–,3 The strong association between body mass index (BMI) and OA of the knee is thought to be mainly due to an increase in mechanical loads to the tibiofemoral cartilage.4
The observation that obesity is also a risk factor for OA of non-weightbearing joints such as the hand5 has suggested that the link between overweight and OA might also occur through systemic inflammation. Adipose tissue may act as an endocrine organ, releasing several proinflammatory mediators and adipokines in blood that may participate in cartilage alteration in obese patients.6,–,9 With mass enlargement of fat, adipose tissue accumulates inflammatory cells, particularly macrophages, and secretes inflammatory cytokines such as interleukin 6 (IL-6), tumour necrosis factor α, serum amyloid A; high levels of leptin, resistin and visfatin; and low levels of adiponectin.10 11 In vitro and in vivo studies have shown that all these adipokines could affect cartilage homoeostasis.12,–,15 Additionally, some studies have suggested that metabolic risk factors such as diabetes mellitus, increased levels of triglycerides and/or cholesterol might also be associated with OA.16 17
Randomised controlled trials have evaluated the effects of diet weight loss on pain and function in overweight patients with knee OA.18,–,22 In those studies, the magnitude of the weight loss was mild to moderate, ranging from 5% to 11%. A recent meta-analysis of pooled data from four of these randomised controlled trials demonstrated that a moderate weight loss of about 5% in obese patients reduces functional disability and, to a lesser extent, pain.23 Few studies have explored the effects of moderate diet weight loss on systemic inflammation, knee OA structural progression and joint biomarkers.19 24 25 A 5% weight loss over 18 months, alone or in combination with exercise, reduced the serum concentrations of C reactive protein (CRP), IL-6 and tumour necrosis factor α soluble receptor 125 but did not attenuate OA progression, as assessed by sequential measurements of joint space width of the knee.19 The effect of exercise and a weight-loss intervention on the serum levels of joint biomarkers yielded inconclusive results.24 One explanation for these somewhat disappointing results might be that the magnitude of weight loss in these studies (5–10%) was not enough to sufficiently reduce systemic inflammation, knee joint loads and metabolic status impairment to demonstrate a benefit on cartilage turnover.
We aimed to investigate the effect of weight loss >10% in obese patients with knee OA on pain, disability, cartilage turnover and systemic inflammation. We chose the model of gastric surgery, which results in drastic weight loss, leading to significant metabolic changes with improvement in insulin sensitivity and lipid parameters and systemic inflammation.26 27
Patients and methods
The ethics committee of the Hôtel-Dieu Hospital approved the clinical investigations. All subjects gave their written informed consent before their inclusion in the study.
Subjects and study design
We screened 140 patients involved in a gastric surgery programme, who were prospectively recruited between 2006 and 2007 in the Department of Nutrition, Center of Reference for Medical and Surgical Care of Obesity, Pitié Salpêtrière Hospital (Paris, France). Patients met the criteria for obesity surgery: BMI ≥40 kg/m2, or ≥35 kg/m2 with at least one comorbidity (hypertension, diabetes mellitus, dyslipidaemia, obstructive sleep apnoea syndrome). Preoperative evaluation included medical history and physical, nutritional, metabolic, cardiopulmonary and psychological assessments. Subjects did not demonstrate acute or chronic inflammatory disease, infectious diseases, viral infection, cancer and/or known alcohol consumption (>20 g/day).
To be included, patients had to have radiographically confirmed knee OA (Kellgren/Lawrence (K/L) grade 2–4) with symptoms for at least 1 month, defined by a knee pain score of at least 30 mm on a 0–100 mm visual analogue scale (VAS). Exclusion criteria were a K/L grade of 1, inflammatory joint disease, chondrocalcinosis of the knee, current use of symptomatic slow-acting drugs, viscosupplementation within the past 6 months and oral corticosteroid treatment or intra-articular corticosteroid injection into any joint within the last month before the bariatric surgery. For all patients, the radiographic evidence of knee OA and eligibility criteria were verified by the same investigator (PR).
Of the 140 screened patients, 44 met the inclusion criteria and were enrolled. Patients' weight had to be stable (ie, variation within ±2 kg) for at least 3 months before surgery. The surgical procedure was laparoscopic Roux-en-Y gastric bypass (RYGB) (n=38) or laparoscopic adjustable gastric banding (n=6) and was performed in the department of surgery of Hôtel-Dieu Hospital.
Blood sampling and clinical evaluation
For all patients, clinical data and biological samples were collected just before surgery (ie, baseline) and at 6 months after surgery. Venous blood was collected in the morning (between 08:00 and 10:00) after a 12 h overnight fast. Serum samples were stored at −80°C for biological assays.
Outcomes for knee OA
The severity of the knee OA pain was evaluated using a continuous 100 mm VAS assessing the global level of pain in the target knee, regardless of the circumstances over the previous 48 h. Other outcomes included the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) subscores for pain, stiffness and disability, and the patient global assessment of the severity of knee OA measured on a 100 mm VAS. When both knees were painful, only the most painful knee was selected for the evaluation.
Lipid profile and insulin sensitivity calculation
Plasma glucose and insulin were measured on a Modular Hitachi system (Roche Diagnostics, Meylan, France) by the glucose oxidase method and an IMMULITE 2000 system (Siemens, La Garenne Colombe, France). Homoeostasis model assessment (HOMA) of insulin resistance (IR) was determined using the HOMA Calculator v2.2.2 (http://www.dtu.ox.ac.uk). This index has been validated in different populations in comparison with the euglycaemic-hyperinsulinaemic clamp values.28 Therefore, the HOMA represents a useful index for study of morbidly obese individuals in whom the evaluation of insulin sensitivity with the clamp technique has understandable technical limitations because of extreme BMI.
Measurements of body composition
Fat-free body mass and adiposity were determined by dual energy x-ray absorptiometry (GE Lunar Prodigy, Madison, Wisconsin, USA) before and 6 months after surgery.
Immunoassays for adipokines and inflammatory markers
Serum leptin and adiponectin were determined with a radioimmunoassay kit from Linco Research (St Louis, Missouri, USA), according to the manufacturer's recommendations. The sensitivity is 0.5 ng/ml and 0.8 µg/ml for leptin and adiponectin, respectively. Intra- and interassay coefficients of variation (CVs) are <4% and 9% for leptin and adiponectin, respectively. Serum levels of IL-6 were measured by a high-sensitivity ELISA system (Quantikine US; R&D Systems Europe, Lille, France). The sensitivity of this assay is <0.04 pg/ml, and intra- and interassay CVs are <8%. High-sensitivity CRP (hsCRP) and orosomucoid levels were measured with an IMMAGE automatic immunoassay system (Beckman–Coulter, Fullerton, California, USA). The sensitivity is 0.02 and 35 mg/dl, respectively; intra- and interassay CVs are <5% and 7.5%, respectively, for hsCRP and 4% and 6% for orosomucoid. Serum levels of fibrinogen were measured with use of a Star Diagnostica Stago system by Fibri-Prest (Parsippany, New Jersey, USA).
Immunoassays for joint biomarkers
Serum cartilage oligomeric matrix protein (COMP) is a biomarker of cartilage degradation and a potential prognostic indicator of joint OA damage.16 29 It was measured by ELISA (COMP ELISA Kit; AnaMar Medical, Lund, Sweden) with two monoclonal antibodies raised against different antigenic determinants of the COMP molecule. Intra- and interassay CVs are <10% and 12%, respectively. Serum hyaluronic acid (HA), a marker of synovial metabolism,30 was measured by ELISA (Corgenix, Broomfield, Colorado, USA) with a specific HA binding protein isolated from bovine cartilage. Intra- and interassay CVs are <7% and 9%, respectively.
Serum N-terminal propeptide of type IIA collagen (PIIANP), a marker of type II collagen synthesis,31 was measured by competitive ELISA (human PIIANP ELISA; Linco, El Paso, Texas, USA) with a polyclonal antibody raised against recombinant human type II procollagen exon 2 fusion protein. Intra- and interassay CVs are <8% and 14%, respectively. Serum collagen helical peptide (Helix-II), a marker of cartilage collagen turnover,32 was measured by competitive ELISA (SYNCART; Synarc, Lyon, France). Intra- and interassay CVs are both <15%. Assays for joint biomarkers were performed in a central specialised laboratory (Synarc) in batches, with the two samples for the same subject in the same run to reduce analytical variation.
Data are reported as mean (SD) or median (range) depending on distribution. The Wilcoxon test was used for comparisons between clinical outcomes and biological marker levels before and after weight loss. Correlations between relative changes in joint biomarker levels, clinical outcomes or changes in metabolic status and systemic inflammation were tested with the non-parametric Spearman correlation test. Multiple regression analysis was used to assess the independent association and contributions of changes in BMI, insulin and HOMA-IR with the dependant variable COMP. A two-sided significance level was fixed at 5%. All analyses involved use of SAS v9.2 (SAS Institute, Cary, North Carolina, USA). A two-tailed p value <0.05 was considered statistically significant.
Demographic characteristics of patients and effects of weight loss on metabolic status
Our study population consisted of 44 obese patients (mean age 44 ± 10.3 years, 36 women) with moderate to severe knee OA (K/L grade 2, 50%; grade 3, 35%; grade 4, 15%). Duration of knee OA symptoms was 5.2 ± 4.6 years. Mean BMI at age 20 was 31.1 ± 10.2 kg/m2 and was 50.7 ± 7.2 kg/m2 just before surgery (table 1). Bariatric surgery resulted in a substantial decrease in BMI (−20% of baseline values; p<0.0001), body weight (−20%; p<0.0001), fat mass (−21%; p<0.0001) and fat free mass (−9%; p<0.0001) at 6 months. Additionally, levels of circulating total cholesterol, triglycerides and insulin were significantly decreased after surgery (all p<0.01). Level of high-density lipoprotein cholesterol did not change (p=0.1). Finally, glycaemic status and insulin resistance, as evaluated by HOMA-IR, were significantly decreased 6 months after surgery (p<0.0001; table 1).
Effect of weight loss on knee OA symptoms
Massive weight loss greatly improved both pain and function in these obese patients with knee OA. After surgery, at 6 months, knee OA pain scores on the VAS decreased from 50 ± 26.6 to 24.5 ± 21 mm (−51%; p<0.001), and scores on all WOMAC subscales were improved: pain (−50%; p<0.001), stiffness (−47%; p<0.001) and function (−57%; p<0.001). Patient global assessment of the severity of the target knee OA was significantly decreased (−50%; p<0.001) (table 2).
Effect of weight loss on adipokine levels and systemic inflammation
The changes in levels of inflammation biomarkers after gastric surgery are shown in table 3. As expected, the serum levels of IL-6 (−26%; p<0.0001), hsCRP (−46%; p<0.0001), orosomucoid (−20%; p<0.0001) and fibrinogen (−5%; p=0.04) were all significantly decreased after surgery. Moreover, weight loss was associated with changes in adipokine levels: mean serum leptin concentration was decreased by 48% (63.2 ± 24.4 ng/ml vs 33±16.7 ng/ml; p<0.0001), and serum level of adiponectin was increased by 21% (7.9±4.6 µg/ml vs 9.9±7.7 µg/ml; p=0.03).
Changes in joint biomarkers with massive weight loss
Weight loss resulted in a significant increase (+32%) in serum level of PIIANP (443.6±257.5 ng/ml vs 586.4±239.4 ng/ml; p=0.002), whereas serum levels of Helix-II and HA were unchanged (p=0.98 and p=0.41, respectively). By contrast, the serum level of COMP was significantly decreased (−36%) after surgery: 10.5±3.5 UI/l vs 6.7 ± 2.2 UI/l (p<0.001) (table 4 and figure 1).
Correlation between biochemical markers and clinical outcomes at baseline
We searched for correlations between markers of systemic inflammation or joint biomarkers and symptoms (pain or disability) at baseline. We found a significant correlation between level of IL-6 and WOMAC function score (r=0.33; p=0.03). IL-6 level was also correlated with levels of hsCRP (r=0.42; p=0.006) and Helix-II (r=0.37; p=0.01). Levels of leptin or adiponectin were not correlated with clinical outcomes or joint biomarker levels. We also found a significant correlation between Helix II and hsCRP (r=0.63; p<0.0001) and orosomucoid (r=0.33; p=0.03). None of the other joint biomarkers correlated with inflammatory biomarkers or with outcomes for knee OA at baseline.
Correlation between changes in COMP or PIIANP levels and changes in clinical outcomes, systemic inflammation and metabolic status
Variation in COMP concentration was significantly correlated with changes in pain (VAS) (r=0.05; p=0.03) and WOMAC stiffness score (r=0.33; p=0.04). Change in COMP concentration was also correlated with change in BMI (r=0.48; p=0.001), insulin level (r=0.36; p=0.02) and HOMA-IR score (r=0.31; p=0.05) (table 5). However, the relationship between changes in COMP levels and changes in insulin levels or HOMA-IR score was not independent of changes in BMI because BMI was the only significant regressor found on multivariate regression analysis (β=0.42, p=0.04). Finally, change in the PIIANP level was not correlated with changes in clinical outcomes or biological marker levels.
Our study shows that a surgically induced mean weight loss of 20% in patients with severe obesity and knee OA can improve pain and function, decrease levels of metabolic parameters and low-grade inflammation, and result in a change in cartilage turnover as assessed by systemic biochemical markers.
Behavioural and pharmacological treatments of obesity usually result in short-term weight loss of approximately 5–10% body weight.33 According to a recent meta-analysis, the pooled effect sizes for improvement in pain and physical disability in patients with knee OA who lost an average of 6.1 kg were 0.2 (95% CI 0 to 0.39) and 0.23 (95% CI 0.04 to 0.42), respectively. Weight loss alone of <5% seems ineffective or poorly effective in alleviating OA knee pain in obese patients.34 Trials that have assessed the efficacy of surgically induced massive weight loss on knee OA symptoms are scarce and have not specifically included patients with well-defined radiographic evidence of knee OA, as in our study.35 36 Hooper et al, using a single-arm open study design, found that obese patients with knee OA who lost 29% of body weight between 6 and 12 months after bariatric surgery showed improved WOMAC pain, function and stiffness scores, by 51%, 74% and 64%, respectively.36 In our study, a 20% weight loss over 6 months decreased scores for pain by 50%, as assessed by a VAS or the WOMAC subscore and for function by 57%. The reductions in WOMAC pain, stiffness and function were significantly correlated among themselves and the WOMAC function correlated best with the patient global assessment of the severity of knee OA (data not shown). Although these data should be cautiously compared, they suggest that the greater the weight loss, the greater the benefit for pain and function.
Obesity is now recognised as a low-grade inflammatory disease. Elevated inflammatory protein levels in obese individuals suggest that inflammation may have a determinant role in connecting obesity to metabolic, hepatic and cardiovascular diseases.10 11 Whether this systemic inflammatory state has a role in the genesis of OA in obese patients is a subject of growing interest.37 Among a myriad of inflammatory mediators, IL-6 has been shown to be secreted substantially by adipose tissue and its level correlates with metabolic complications in some studies.10 Interestingly, here we found a significant correlation at baseline between circulating levels of IL-6 and WOMAC function score. We also observed that the level of IL-6, which mainly originates from adipose tissue in obese patients but which is also produced by infrapatellar fat pad within the joint,38 was correlated with Helix-II, a biomarker of cartilage turnover. By contrast, no other inflammatory protein or adipokine level or BMI were correlated with clinical outcomes before surgery. Our findings extend the results of recent work showing a significant association of IL-6 circulating levels and the prevalence and incidence of knee OA.39
As expected, bariatric surgery resulted in a significant increase in the serum level of adiponectin and a significant decrease in that of leptin, IL-6, hsCRP, orosomucoid and fibrinogen.27 40 However, none of the changes in these inflammatory biomarkers was correlated with changes in clinical outcomes, which suggests that the decrease in low-grade inflammation has no role or is of little importance in clinical improvement related to weight loss.
Because the morphology of our patients precluded carrying out repeated MRI or standard x-ray examinations for structural evaluation, we used biochemical markers of joints as surrogate markers to assess cartilage turnover.41 Massive weight loss resulted in a significant increase in the level of PIIANP (+32%), a marker of type II collagen synthesis, and a decrease in that of COMP (−36%), a marker of cartilage degradation, according to the BIPED classification.30 These results are the first to suggest a benefit of weight loss on both cartilage anabolism and catabolism. Of note, one recent study also found a positive correlation between moderate weight loss and changes in COMP levels.24 The mechanism by which weight loss decreases COMP levels is unknown. A decrease in knee joint load with weight loss42 may modulate the release of COMP, because this protein has been shown to be mechanosensitive.43 44
Changes in adipokine levels were not correlated with changes in joint biomarkers, suggesting that variation in leptin or adiponectin concentrations had little or no effect on cartilage homoeostasis in our patients. Bariatric surgery resulted in a sharp decrease in levels of cholesterol and triglycerides and insulin resistance. Interestingly, change in COMP level was correlated with changes in insulin resistance, but this correlation was not independent of change in BMI, probably because insulin resistance and BMI are closely related biologically. Although there is no formal clinical evidence for a link between diabetes mellitus and OA, several experimental data suggest a detrimental effect of insulin resistance on cartilage.45,–,48
Our study has some limitations that deserve attention. Our trial was conducted as an open exploratory study, and bias in evaluation of clinical outcomes may have occurred. Thus, our findings need to be replicated in confirmatory studies. The relatively small sample size may have missed some weak associations. Moreover, our findings on correlations do not imply causality between assessed variables, and thus should be carefully interpreted. The strengths of this study include the assessment of severely obese patients recruited from a centre of reference for medical and surgical care of obesity, which allows for a well-standardised biological and clinical evaluation of such patients before and after bariatric surgery.
In conclusion, our data show that massive weight loss (20%) in patients with symptomatic knee OA improves pain and function, decreases the low-grade inflammatory state and may modified cartilage turnover, which suggests a structural effect of substantial weight loss. The effect of changes in insulin resistance related to weight loss on cartilage homoeostasis needs further investigation.
The authors thank all the patients for their participation in this study. The authors also thank Mme Christine Baudouin, Dr Florence Marchelli and Mme Patricia Ancel involved in patient's recruitment, data collection and sampling at the Center of Research on Human Nutrition, Paris Pitié-Salpêtrière Hospital.
The first two authors contributed equally to this work.
Funding Assistance Publique-Hôpitaux de Paris (APHP) and Direction of Clinical Research (PHRC N°0702 and CRC N° P050318), which promoted and supported the clinical investigation, and a grant from the European community (Collaborative Project ADAPT, contract number HEALTH-F262008-201100). This work was also supported by the ‘Association Rhumatisme et Travail’ (Hôpital Lariboisière, Paris, France).
Competing interests None.
Patient consent Obtained.
Ethics approval The ethics committee of the Hôtel-Dieu Hospital approved the clinical investigations.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.