Key messages

• Weight loss exclusively through diet may improve functionality and quality of life (low certainty evidence).
• We are uncertain whether weight loss exclusively through diet reduces pain as the certainty of the evidence has been assessed as very low.
• Weight loss exclusively through diet probably has minimal or no adverse effects (moderate certainty evidence).

What is the evidence.
See evidence matrix  in Epistemonikos later

We found seven systematic reviews [1], [2], [3], [4], [5], [6], [7] including six primary studies reported in 14 references [8], [9] , [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], of which all corresponded to randomized trials.

What types of patients were included*

All trials included adults older than 18 years with a clinical and/or radiological diagnosis of symptomatic unilateral or bilateral knee osteoarthritis, mainly due to pain and functional limitation.

All trials included patients who presented some degree of overweight or obesity, with BMI values ​​greater than 28. Two of the trials [10], [21] recruited patients with a BMI greater than 27.

Three trials included patients with sedentary lifestyle [8], [9], [10], while the remaining trials did not provide this information [16], [20], [31].

What types of interventions were included*

All trials evaluated weight loss exclusively through diet, compared mainly against placebo, exercise, other less intensive diets and healthy lifestyle education sessions.

One trial [8] evaluated a diet divided into three phases (intensive phase, transition and maintenance), each with specific goals in addition to educational sessions on healthy lifestyles.

One trial [9] evaluated a diet that included the consumption of nutritional powder and weekly dietary educational sessions, compared to another diet with a higher energy content and education on a healthy diet.

One trial [20] evaluated a program that consisted of nutritional education and a very low energy diet (485 kcal/ day), compared to a low energy diet (810 kcal/day).

In one trial [10], the diet consisted of replacing regular meals of the day with meals adjusted in calories for the weight loss rate, with a goal of 10% reduction in the initial weight, compared against exercise.

One trial [16] evaluated a diet based on a reduction of 25% to 30% of the previous caloric level compared against placebo (they received neither diet nor exercise program).

In one trial [21] the diet consisted of a nutritional program for weight control, which included weekly diet diaries and information on fat intake and adequate proportions of vegetables, compared with standard care and non-specific health information.

What types of outcomes
were measured

The trials evaluated multiple outcomes, which were grouped by the systematic reviews as follows:

• Pain, measured through the pain subscale of the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), Visual Analogue Scale (VAS) and the pain subscale of the Knee Injury and Osteoarthritis Outcome Score (KOOS)
• Functionality measured with the WOMAC physical function or KOOS function in daily living
• Quality of life, through the Short Form Health Survey (SF-36) questionnaire and its physical subscale

The average follow-up of the trials was 60 weeks with a range between 8 weeks and 18 months.

To whom this evidence does and does not apply

• The results are applicable to adults (over 18 years) with overweight or obesity, without differences by sex, and with clinical or radiologic diagnosis of osteoarthritis in one or both knees with associated symptoms, mainly pain and loss of function. 
• This evidence is not applicable to patients who have received prior surgical treatment for knee osteoarthritis or who have achieved weight loss through surgery or drug use; these patients were not included in the evaluated trials.
• The available literature makes it difficult to determine the effects and to apply the information to decision-making, since there is wide variation in the diet in terms of duration and nutritional content, and there is not a common comparator among the different trials included in the reviews.

• We decided to include pain, functionality, mobility and quality of life as outcomes given its relevance for the decision-making process of knee osteoarthritis treatment. These outcomes are easy to measure clinically and they are part of the treatment expectations for both patients and clinicians. On the other hand, we made a search of relevant outcome sets for knee and hip osteoarthritis in the COMET Initiative (Core Outcome Measures in Effectiveness Trials), where one set [22] included the first three outcomes selected for this summary.
• Additionally, the adverse effects of the intervention were investigated, which were reported by a single trial [9]. According to this, the risks to be expected would be constipation, flatulence, dizziness and increased sensitivity to cold.

• The benefit of the diet itself in these patients translates into a possible improvement in functionality and quality of life. The effect on pain is uncertain due to the limitations of the existing evidence. 
• The risks, with a moderate certainty of the evidence, are probably of low frequency and severity, among them constipation, flatulence, dizziness and cold intolerance. Although there is a high degree of uncertainty, the balance between benefits and risks would be favorable for the diet in this context.

• Given the potential benefits, weighed against the slight risks described previously, it seems reasonable to consider weight loss through diet as a part of the treatment. This benefit could increase if it is associated with other measures that pursue the same objective: exercise, education, analgesia, physical therapy. 
• It is not possible to accurately determine the cost-effectiveness of the intervention, due to the existing uncertainty.

• Currently, most doctors recommend comprehensive weight management through programs that include both diet and exercise. Although the recommendation is well supported, problems arise when standardizing both the diet and the exercises to be implemented. Further trials are necessary to standardize the diets that achieve some benefit, which could be implemented without associated exercise in patients limited in their ability to perform physical activity, such as the elderly or disabled.
• On the other hand, the prescription of diet is not technically complex and is generally well received by overweight patients. However, the prolonged periods of time required to achieve some clinical benefit hinder the adherence.

Differences between this summary and other sources

• The reviews included in this summary, although differ in terms of the included trials, type of diet implemented and comparison studied, coincide in the beneficial effect of weight loss through diet for some outcomes such as functionality and quality of life. The conclusion of the reviews, both those that interpreted the differences as negative (not statistically significant) [3], [4], [5], and those that interpreted them as positive (statistically significant) [2], [6], [7] coincide with the conclusion of this summary: there is a possibility of benefit, but with a low certainty of evidence. 
• The international guidelines for knee osteoarthritis management of the American Academy of Orthopaedic Surgeons (AAOS) [23] recommend weight loss in symptomatic patients with a BMI greater than 25, mainly through diet associated with exercise. The recommendation of the Osteoarthritis Research Society International (OARSI) [24] is similar; suggesting programs for weight management in patients who have osteoarthritis of the knee and overweight, intending to achieve a reduction of 5% in 20 weeks, but in this recommendation diet alone is not mentioned as an option.

• Future evidence is likely to change the conclusions derived from this summary, given the uncertainty that currently exists. 
• There are at least two reviews on this topic that are in progress [25], [26], registered in the International Prospective Register of Systematic Reviews (PROSPERO).
• In addition, at least two ongoing trials [32], [33] evaluating the effect of diet alone were identified, according to the International Clinical Trials Registry Platform of the World Health Organization.