Investigating Racial Differences in Diet Benefits for Knee Osteoarthritis

Knee osteoarthritis (OA) is the most prevalent form of arthritis and race is a risk factor for poor outcomes. African-Americans (AAs) report greater OA-related disability and pain severity compared to their Non-Hispanic White (NHW) counterparts. These disparities are reinforced through social and biological mechanisms, ultimately resulting in dramatic racial disparities in pain experience and associated quality of life. Low-carbohydrate diets (LCDs) reduce inflammation and pain independent of weight loss, but significant racial differences exist in metabolism that are rarely addressed in diet interventions. The overall objective of the proposed study is to determine whether the beneficial effects of an LCD for knee OA pain are related to race. The investigators will recruit 20 adult women (65-75) with knee OA with equal representation across racial groups (10 AA, 10 NHW). Following one week of diet and pain self-report, the investigators will assess quality of life, depression, experienced pain and evoked pain. Participants will be placed on a LCD wherein all meals and snacks will be delivered weekly after consult with study personnel. Participants will return every 3 weeks for testing during the 12-week intervention with blood drawn at baseline and at the conclusion of the 12-week diet. Blood will be assayed for oxidative stress markers. This will be the first assessment of racial differences in the efficacy of a LCD to reduce knee OA pain. Objective 1: To determine whether the LCD reduces pain after 12 weeks. Hypothesis: The LCD will significantly reduce evoked and self-reported pain. Objective 2: To determine whether the benefits of the LCD differ based on race. Hypothesis 1: The LCD will reduce evoked and self-reported pain more in AA than in NHW. Hypothesis 2: AAs will experience greater improvements in depression, quality of life, pain interference and show more weight loss than NHWs. Objective 3: To determine whether the LCD has a differential impact on oxidative stress by race. Hypothesis 1: The LCD will significantly reduce oxidative stress over 12 weeks. Hypothesis 2: AAs will show greater reductions in oxidative stress than NHWs. The reduction in oxidative stress will be correlated with reduction in evoked pain.

A telephone screening interview will be used to assess each participant's OA status and assessment of the reported duration of knee OA, current and past treatments, comorbid conditions, current medication use, dietary conditions and other exclusion criteria. Eligible, participants will be invited to the testing facilities in Campbell Hall for informed consent and baseline measures. Following this visit, participants will record daily food consumption for one week as well as daily pain ratings. After one week, participants will visit the Clinical Research Unit for fasted blood draw and additional testing. Immediately following this session, participants will be asked to select one week of meals and snacks from a menu of commercially-available meals. Study personnel will record the choices and order the meals to be delivered to the participants to initiate the intervention. During the 12-week intervention, participants will be contacted by study personnel weekly to place food orders and be instructed to record any beverages consumed during the week. Every three weeks, participants will return for testing. At the end of 12 weeks, blood will be taken prior to testing/debriefing. Participants. The investigators will recruit 20 adult women (65-75) with knee OA with equal representation across racial groups (10 AA, 10 NHW). Peak prevalence rates for OA are at the 65-75 years of age, so the investigators feel confident that there will be no trouble recruiting this population. Participants will be recruited using existing databases, community flyers and community outreach. The feasibility trial used prescribed diets and had an attrition rate of 0% in the LCD group. Here, provision of all of the food for the intervention is expected to have lower rates of attrition. Inclusion criteria will include: diagnosis of knee OA; pain in at least 4/7 days/week for the past 3 months; age between 65-75; average daily consumption of >100 g carbohydrates; understanding of verbal and written English; self-identification as either AA or NHW; BMI between 25 and 40 kg/m2. Exclusion criteria will be the following: diabetes; unwillingness to follow prescribed diets; recent weight change (>4 kg in past month); currently on a diet; history of eating disorders or other psychiatric disorders; digestive diseases; difficulty chewing or swallowing; reliance on others for meal preparation; cardiovascular or pulmonary disease; daily opioid pain medications; use of medications known to alter metabolism or digestion (e.g., proton-pump inhibitors); use of anti-hypertensive medications that affect glucose tolerance; use of tobacco; participation in extreme exercise, knee replacement. Diet Intervention. All foods will be provided under the direction of study personnel and will be delivered weekly to participants' home address. Weekly contact will maintain retention in the intervention and improve adherence. The Dietary Guidelines for Americans suggests 225-325 g of carbohydrates/day. Therefore, those participants consuming less than 100 g/day would be considered as consuming a reduced-carbohydrate diet and will be excluded. Participants are directed to reduce their total (not net) carbohydrate intake to ≤ 40 g/day. Meals will be offered such that no combination of chosen meals will exceed our limit. Fats will not be restricted, nor will protein (meats, eggs). Fruits will be restricted and vegetables permitted in limited quantities (2 cups/day of leafy greens, 1 cup/day non-starchy vegetables, etc.). Participants will be instructed as to the types and quantities of beverages that are permitted to accompany the LCD. Daily or almost-daily consumption of sugar-sweetened beverages was associated with lowered optimism in chronic pain sufferers and greater risk for depression in healthy women. Artificial carbohydrate-free sweeteners (stevia or sucralose) will be permitted, but powdered sweeteners (aspartame, saccharin, stevia, sucralose) can only be used in limited quantities as they contain maltodextrin (1 g of rapidly digesting carbohydrate). LCDs are known to be safe and first-line treatments for diabetes. LCDs are also known to reduce inflammatory biomarkers to a greater extent than low-fat diets. In fact, a LCD resulted in improved insulin sensitivity as well as reduced triglycerides even when weight loss was accounted for. Anthropocentric Measures. Body weight, height, waist circumference, blood pressure, and heart rate will be assessed. Pain-Specific Questionnaires. Pain and disability will be measured using the Brief pain Inventory (BPI) short form and the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC). The BPI is used to assess the severity of pain and the degree to which that pain interferes with daily activities. This inventory also allows for reports on medications used to treat pain. The WOMAC is commonly used in studies of arthritis and has been validated in numerous clinical trials. Specifically, the WOMAC allows for assessment of pain and interference, but also is specific to aspects of stiffness in the joints not accessed by other surveys. The investigators have previously utilized these measures for chronic pain sufferers. These assessments will be given at baseline and every 3 weeks. Evoked Pain Testing. Functional tests will be carried out by study personnel. Participants will rate the intensity and unpleasantness of the pain in their affected knee (0-100 scale) prior to and immediately following the tests. The numerical rating scale will be anchored at 0 (no pain) and 100 (worst pain imaginable). Questionnaires will be given between each task to allow for a rest period. Task order will be randomized. Temporal summation will be assessed on the patellar of the affected knee using a nylon monofilament (Touch test Sensory Evaluator 6.65) calibrated to bend at 300 g of pressure. Participants will provide a pain rating following a single contact of the monofilament, after which they will provide another pain rating following a series of 10 contacts (at a rate of one contact per second). The change in pain ratings for single versus multiple contacts reflects temporal summation. Repeated chair stands is a portion of the Short Performance Physical Battery. Participants are asked to stand from a sitting position five times in a row as fast as possible with arms crossed. The time to completion and number of successful stands is scored as the degree of ability. Timed Up-and-Go is a common task for evaluating pain interference of everyday activities. From a seated position, participants will be asked to get up, walk 10 feet, turn and return to the chair. Time to complete the task will be recorded. Timed walk will be the final task included. As in the Short Performance Physical Battery, a distance of 15 feet will be marked out on the floor. Participants will complete the distance walking at normal gait speed twice and the time to complete each walk will be recorded. Quality of Life and Emotional State. Quality of Life will be measured using the short form 36 (SF-36) every 3 weeks. The SF- 36 measures general health status and quality of life across eight domains that are relevant for assessment of a diet intervention. The domains include physical functioning, role limitations due to physical health, bodily pain, general health perceptions, vitality, social functioning, role limitations due to emotional problems, and mental health. The SF-36 has demonstrated reliability and validity in older adult populations and in diet intervention studies. Depression and mood will be assessed using the Patient Health Questionnaire (PHQ-9), which assesses the DSM-IV criteria for depression and can measure depression severity and response to intervention. Participants with PHQ-9 scores >15 will be referred to University of Alabama at Birmingham (UAB) Psychiatry Services for evaluation, as will participants identifying suicidal ideation of any duration on question 9. Oxidative Stress. At baseline and at the end of the intervention, 10 ml of blood will be taken by a trained research nurse. Sera will be isolated, aliquoted, and frozen at -80°C. Serum samples will be analyzed for TBARS (oxidative stress) by ELISA using commercially-available kits. Statistical Analysis. Data analysis will begin with an Intention-To-Treat (ITT) analysis to account for participant attrition through inclusion of all randomized participants, as recommended by the Food and Drug Administration. This analysis will be followed by an analysis of per-protocol completers (adherence) and will begin with calculation of measures of central tendency (mean, median) and dispersion (variance, interquartile range) for all study variables. Repeated measures analysis of variance (ANOVA) will be used to determine the efficacy of the diet across time for evoked and self-reported pain measures and oxidative stress.To assess efficacy of the intervention across races, this will be a cross-sectional analysis using change from baseline to 12-weeks. Generalized linear models will be used to assess the main effects of race on changes in daily pain and evoked pain, as well as depression and quality of life and oxidative stress. Pearson correlations will be calculated to assess the relationship between daily pain and oxidative stress within each group at baseline and after 12 weeks. For each outcome measure of interest, normality assumptions will be assessed, distributions (e.g. normal, binary, Poisson) explored, and transformations employed when appropriate. Based on our preliminary results showing significant changes in the LCD group with 8 participants, the investigators feel that n=10/group is appropriate. Additionally, with 90% power and alpha set to 0.05, a total sample of 20 will allow detection of an effect size as low as 0.4 for our primary comparisons. With 80% power, the investigators will be able to detect effect sizes as low as 0.34. Clinically-Meaningful Differences. Group mean differences are not always reflective of clinically-meaningful differences at the individual level. Therefore a post-hoc analysis will be carried out using published clinically- meaningful differences in: (1) WOMAC pain, (2) WOMAC disability, and baseline pain intensity score. Briefly, a reduction of ≥1.5 (pain) or ≤6.0 (disability) is considered clinically-meaningful, as is a reduction of ≥1.7 on an 11-point rating scale. Conversely, an increase of ≥2.2 (pain) or ≥6.0 (disability) is considered worsening, as is an increase of ≥2.2 on an 11-point scale. Generalized linear mixed models will be used to estimate the odds of reporting improvement or worsening between AA and NHW participants.

All participants assigned to a single diet intervention. Between-groups differences and within-group changes will be assessed.

The Western Ontario and McMaster Osteoarthritis Index (WOMAC) pain score is a 0-20 score with higher scores reflecting more severe pain. Questions 1-5 are summed to provide a WOMAC pain score. Change scores will be calculated.

The Brief Pain Inventory (BPI) pain score is a 0-40 score with higher scores reflecting more pain. Questions 3-6 are scored on 0-10 and are summed to provide an overall pain score. Change scores will be calculated.

Before and after the Timed-Up-And-Go (TUG) task, participants will be asked to rate the intensity of pain in their knee on a 0-100 scale. The difference in the ratings will be considered the evoked pain score. Change scores will be calculated.

The Western Ontario and McMaster Osteoarthritis Index (WOMAC) function score is a 0-68 score with higher scores reflecting greater impairment in function. The 17 items assess difficulty performing specific tasks and are scored 0-4. These scores are summed to provide a WOMAC function score. Change scores will be calculated.

The Brief Pain Inventory (BPI) pain interference score is a 0-90 score with higher scores reflecting more pain. Question 9A-I are scored on 0-10 and are summed to provide an overall pain interference score. Change scores will be calculated.

Before and after the Temporal Summation (TS) task, participants will be asked to rate the intensity of pain in their knee on a 0-100 scale. The difference in the ratings will be considered the evoked pain score. Change scores will be calculated.

Before and after the Repeated Chair Stand (RCS) task, participants will be asked to rate the intensity of pain in their knee on a 0-100 scale. The difference in the ratings will be considered the evoked pain score. Change scores will be calculated.

The Short Form 36 (SF-36) quality of life score is a 0-150 score with higher scores reflecting poorer quality of life. Scores in each of the 7 sections (general health, limitations, physical health, emotional health, social activities, pain and energy) are summed to provide an overall quality of life score. Change scores will be calculated.

The Patient Health Questionnaire 9 (PHQ-9) depression score is a 0-27 score with higher scores reflecting more severe depression. The 9 items are scored on a 0-3 scale and are summed to provide an overall depression score. Change scores will be calculated.

The time to complete the Timed-Up-And-Go (TUG) task will be measured with a stopwatch. Change scores will be calculated.

The time to complete the Repeated Chair Stand (RCS) task will be measured with a stopwatch. Change scores will be calculated.

Before and after the Timed Walk (TW) task, participants will be asked to rate the intensity of pain in their knee on a 0-100 scale. The difference in the ratings will be considered the evoked pain score. Change scores will be calculated.

The time to complete the Timed Walk (TW) task will be measured with a stopwatch. Change scores will be calculated.

Thiobarbituric acid-reactive substances (TBARS) will be assessed via assay purchased from R&D Systems. The sensitivity of the assay is reported to be 0.0-17.0 uM. In our hands, the minimum sensitivity has been found to be 1.1 uM. Higher levels are thought to correspond to greater oxidative stress. Change scores will be calculated.

Inclusion Criteria: diagnosis of knee OA pain in at least 4/7 days/week for the past 3 months age between 65-75 average daily consumption of >100 g carbohydrates understanding of verbal and written English self-identification as either AA or NHW BMI between 25 and 40 kg/m2 Exclusion Criteria: diabetes unwillingness to follow prescribed diets recent weight change (>4 kg in past month) currently on a diet history of eating disorders or other psychiatric disorders digestive diseases difficulty chewing or swallowing reliance on others for meal preparation cardiovascular or pulmonary disease daily opioid pain medications use of medications known to alter metabolism or digestion (e.g., proton-pump inhibitors) use of anti-hypertensive medications that affect glucose tolerance use of tobacco participation in extreme exercise knee replacement

Following our extended analysis period, de-identified data will be made available, as well as the variable key, to the scientific community. Before this time, a request for data sharing will be assessed by study personnel (PI and Co-Is) as needed. These requests will be evaluated based on scientific merit and overlap with the aims of the current study.

De-identified data can be made available on encrypted servers for researchers whose research plans do not overlap with the aims of the current project. Requests will be assessed by the PI and Co-Is as needed.

Strath LJ, Jones CD, Philip George A, Lukens SL, Morrison SA, Soleymani T, Locher JL, Gower BA, Sorge RE. The Effect of Low-Carbohydrate and Low-Fat Diets on Pain in Individuals with Knee Osteoarthritis. Pain Med. 2020 Jan 1;21(1):150-160. doi: 10.1093/pm/pnz022.