Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Jul 1.
Published in final edited form as: Arthritis Care Res (Hoboken). 2024 Apr 1;76(7):993–1005. doi: 10.1002/acr.25319

Fatty acid intake and polyunsaturated fatty acid biomarkers and risk of total knee or hip arthroplasty among older women in the Women’s Health Initiative

Tonya Orchard 1,*, Eric McLaughlin 1, Timothy Winschel 1, Aladdin Shadyab 2, Deepika Laddu 3, Mara Vitolins 4, Florina Constantinescu 5, Rebecca Jackson 1,#
PMCID: PMC11209811  NIHMSID: NIHMS1971615  PMID: 38412867

Abstract

Objective:

The objective was to determine whether baseline fatty acid intake and erythrocyte omega-3 and omega-6 polyunsaturated fatty acids (PUFAs) can predict risk for total hip arthroplasty (THA) and total knee arthroplasty (TKA) in older women.

Methods:

This was a prospective analysis of 34,990 women in the Women’s Health Initiative (WHI). Dietary fatty acids were estimated from food frequency questionnaires. Imputed erythrocyte PUFAs were available in a sub-cohort of 3428 women. Arthroplasty (THA and TKA), used as a surrogate of severe osteoarthritis, was identified via linked Medicare data. Cox proportional hazards models were constructed to estimate risk of arthroplasty.

Results:

Risk of THA was associated with higher intake of arachidonic acid, [multivariable hazard ratio (HR) quartile (Q)4 vs Q1: 1.16; 95% CI: 1.01, 1.34; p for linear trend = 0.03] and higher intake of eicosapentaenoic acid + docosahexaenoic acid [(EPA+DHA), HR Q4 vs Q1: 1.20; 95% CI:1.05, 1.39; p for linear trend=0.003]. There was a linear trend (p=0.04) for higher risk of THA with higher erythrocyte EPA+DHA in BMI-adjusted models; however, there was no significant difference in THA by quartiles of erythrocyte EPA+DHA (p=0.10). Dietary fatty acids and erythrocyte PUFAs were not significantly associated with risk of TKA.

Conclusions:

Higher baseline intakes of arachidonic acid and EPA+DHA were associated with a modestly higher risk of THA. No association was found between fatty acids and TKA. Further research in populations with direct measures of osteoarthritis severity is needed to better understand the importance of PUFAs in modulating osteoarthritis and arthroplasty risk.

INTRODUCTION

Osteoarthritis (OA) is the most common form of joint disease and a leading cause of disability in older adults in the United States.1 The hallmark of OA is articular cartilage breakdown and subchondral bone changes often leading to joint pain, stiffness, and loss of mobility. There are currently no FDA-approved treatments that delay the onset or progression of OA. Arthroplasty (i.e. joint replacement) represents the final treatment option for many patients, and has been used as a proxy outcome for severe OA.2,3 Most arthroplasty procedures (93%) involve replacement of the knee or hips. In 2013, there were approximately 723,000 knee replacements and 493,700 partial or total hip replacements in the United States. In 98% of knee replacements, OA was the principal or first diagnosis. Nearly two thirds of knee replacements and over half of hip replacements were performed on women. Approximately 80% of knee and hip replacements were in non-Hispanic White individuals, while approximately 7% were in non-Hispanic Black individuals.4 Rates of both total hip arthroplasty (THA) and total knee arthroplasty (TKA) are projected to more than double by the year 2030 compared to rates in the National Inpatient Sample 2000–2014, underscoring the growing impact on public health and health care systems.5

Some studies have suggested that chronic low-grade inflammation may be associated with risk of OA incidence and progression. Inflammatory cytokines are elevated in serum and synovial fluids of OA patients 6, and inflammation of synovial tissues is associated with cartilage destruction 7,8, as well as progression and severity of OA.9,10 There is considerable evidence that dietary fatty acid intake and composition of polyunsaturated fatty acids (omega-3 (n-3) PUFAs and omega-6 (n-6) PUFAs) distinctly influence the maintenance of bone and joint health and are implicated in the development and progression of OA. Elevated metabolites of the n-6 PUFA, arachidonic acid (AA), have been shown to promote inflammation within the arthritic joint.11 AA metabolites are elevated in patients with symptomatic knee OA and have been shown to predict disease severity in patients with symptomatic knee OA 12 or radiographic OA.13 In contrast, n-3 PUFAs produce metabolites that are typically more anti-inflammatory14,15, and reduce cartilage-degrading enzymes and inflammatory cytokines in in vitro models.16,17 Recently, low endogenous and dietary n-3 PUFAs in mouse models were linked to more rapid OA onset 18, and severity, as well as more synovitis.19 Emerging data also suggest that endogenous n-3 PUFAs may decrease the loss of key structural components of joint cartilage.13,20

However, there is limited prospective data from human cohorts addressing the association of fatty acid intake or biomarkers with OA progression or severity. Analysis of self-reported dietary data from 2092 adults enrolled in the Osteoarthritis Initiative (OAI) revealed that higher intake of total PUFAs and monounsaturated fatty acids were associated with less radiographic progression of knee OA, while higher intakes of total fat or saturated fatty acids were associated with greater progression of knee OA.21 Although these results suggest that dietary fatty acids might modulate OA symptoms and progression, to our knowledge, the association of fatty acid intake or PUFA biomarkers with arthroplasty, an indicator of the most severe cases of OA, has not been prospectively examined.

Therefore, the aims of this study were to: 1) examine the association of dietary intake of fatty acids with risk of TKA and THA due to OA in older women in the Women’s Health Initiative (WHI), and 2) examine the association of red blood cell polyunsaturated fatty acids (RBC PUFAs), objective biomarkers of marine sourced n-3 PUFA intake (EPA and DHA) and AA intake, with risk for TKA or THA in a sub-sample of the WHI cohort.

PATIENTS AND METHODS

Study population

The WHI cohort was comprised of 161,808 racially diverse postmenopausal women, recruited between 1993 and 1998. The WHI consisted of an observational study (OS) and three randomized clinical trials (CT); a hormone therapy trial (HT) with estrogen or estrogen alone versus placebo, a low-fat dietary modification trial (DM) and a calcium plus vitamin D supplement versus placebo trial (CAD). A detailed description of the study design and methods have been previously published.22

The present study sample for the main analysis included CT and OS participants with valid baseline food frequency questionnaire (FFQ) data who were 65 years or older at the time of their FFQ assessment and were enrolled in Medicare Fee-For-Service (FFS: Medicare A or A+B) at the time the FFQ was completed. Women were excluded if they reported a history of hip or other arthroplasty or rheumatoid arthritis at baseline or had TKA or THA but no diagnosis of OA in CMS data, were in active treatment arm of the DM Trial, had missing Medicare enrollment information, no Medicare FFS coverage after baseline FFQ, or greater than one year gap in Medicare FFS coverage from time of baseline FFQ until December 31, 2016. Our final analytic cohort consisted of 34,990 women (Figure 1).

FIGURE 1:

FIGURE 1:

Open in a new tab

Flow diagram for selection of study population. 1, original enrollment file contains Medicare information for 145,753/161,808 participants. 2, using primary, secondary, or tertiary OA diagnosis. 3, coverage includes intervals of HMO, part B only, and/or not enrolled. 4, no women in this cohort had a self-reported history of oral daily glucocorticoid use. Legend: WHI, women’s health initiative; CT, clinical trial; OS, observational; DM, dietary modification; RA, rheumatoid arthritis; FFQ, food frequency questionnaire; FFS, fee-for-service

A sub analysis was conducted among participants of the WHI Memory Study (WHIMS)23, with complete RBC PUFA data at baseline (n= 6912). 24,25 The FA compositions of the RBC samples were multiple imputed using regression calibration equations as previously described.24 Using similar exclusion criteria as the main analysis, 3428 women, each with 10 imputed RBC PUFA values, were included in the sub-analysis.

Assessment of dietary fatty acids and PUFA biomarkers

Dietary intake was assessed using semi-quantitative FFQ administered at baseline visit. FFQ included 126 questions of which 19 questions were designed to improve the precision of fat intake estimates, and two summary questions addressing usual intake of added fats.26 Nutrient intake was estimated from the FFQ using Nutrition Data System for Research, version 30, Minneapolis, MN.27 Data on fish oil or PUFA supplements were not collected as part of the WHI protocol; thus, only fatty acid intake data from the FFQ are included in dietary analysis.

RBC PUFAs, analyzed for the WHIMS cohort, were used as an objective biomarker of EPA+DHA intake and a measure of n-3 PUFA status.24 Measurement of RBC PUFAs were determined via gas chromatography.25 The protocol for blood collection in WHI has been previously described.24 Briefly, fasting blood samples were collected from consenting WHI participants at baseline visit, processed for plasma, buffy coat and RBC fractions and stored at −70 degrees C within 2 hours of collection. Samples were shipped on dry ice, to the central WHI biorepository in Rockville, MD, where they were stored at −80 degrees C. RBC samples were then shipped on dry ice to the sample processing lab in Seattle, WA, where they were aliquoted for use by investigators. During the aliquoting process, RBC samples were inadvertently stored at −20 degrees C for approximately 10–30 days. To correct for real or potential degradation of highly unsaturated PUFAs in the RBC samples used for this analysis, Pottala et al. 24, used multiple imputation to replace each fatty acid value in the dataset with ten instances from a distribution of possible values. These corrected values were used in all statistical analyses in this study.

Ascertainment of Arthroplasty

Arthroplasty outcomes were obtained using the linked Medicare FFS data (specifically, the MedPAR dataset through December 31, 2016). Arthroplasty was defined as total knee (TKA) and/or total hip replacement (THA). Arthroplasty for OA was identified using the following diagnostic codes: TKA based on ICD-9 surgical procedure code 81.54 and/or THA based on ICD-9 surgical procedure code 81.51 in combination with a primary, secondary, or tertiary diagnosis of OA (ICD-9-CM 715.xx). For participants with multiple events, we considered time to the first event in the analyses.

Assessment of Covariates

Data on age, race/ethnicity, education, physical activity, smoking status, history of arthritis and joint pain were obtained from self-reported questionnaires administered at the baseline WHI visit.22 History of chronic diseases including cancer, diabetes mellitus, emphysema, stroke, myocardial infarction, congestive heart failure, and peripheral arterial disease, was self-reported at baseline; for this analysis, we calculated a comorbidity score (0,1,2,3+) based on the number of chronic disease conditions reported. Information on use of medications and supplements of interest [non-steroidal anti-inflammatory medications (NSAID), aspirin, and calcium and vitamin D)] was extracted from data collected by clinic interviewers at baseline. Calcium and vitamin D intake from the diet, as well as alcohol consumption, were estimated from the FFQ. Body mass index (BMI) was calculated from anthropometric data measured at baseline by trained clinic staff; weight was measured to the nearest 0.1 kg and height to the nearest 0.5 cm using research quality scales and stadiometers.

Statistical Analysis:

Multivariable Cox proportional hazards regression models were used to assess associations between dietary FAs and arthroplasty, first as a composite outcome of TKA and THA and then TKA and THA separately in the whole cohort (n=34,990). In these models, participants were followed from the date of the baseline assessment to the date of arthroplasty or end of enrollment in Medicare FFS, last date of available Medicare data (December 31, 2016), or death, whichever came first. For women with short (≤ 1 year) gaps in Medicare FFS coverage, follow-up time was broken up into multiple intervals to represent periods of FFS coverage.

For all models, participant age was used as the time scale, and the baseline hazard was stratified by WHI trial membership (OS, Estrogen-alone Trial, Estrogen + Progestin Trial, DM, Estrogen-alone/Estrogen + Progestin Trial + DM) to account for demographic and diet differences between groups (e.g. women in the OS tended to have more comorbidities; women in the DM trial consumed greater than 30% of kcal from fat at baseline). The dietary fatty acids of primary interest in this analysis were eicosapentaenoic acid (EPA) + docosahexaenoic acid (DHA), arachidonic acid (AA), linoleic acid (LA), and alpha linolenic acid (ALA) measured at baseline; secondary fatty acids included total n-3 PUFA, total n-6 PUFA, saturated fatty acids (SFA), monounsaturated fatty acids (MUFA), and PUFA. Dietary fatty acids were adjusted for energy density prior to statistical analysis. All dietary fatty acids are presented as % of total energy intake. For all the models described below, fatty acids were first considered as continuous variables and also categorized into quartiles. We first considered univariable (unadjusted) relationships between each fatty acid and outcomes. Next, we fit a BMI-adjusted model; finally, we fit a multivariable model using the covariates previously described and identified a priori for adjustment based on clinical relevance. In the BMI and multivariable-adjusted models, we first assessed the presence of two-way interactions with the fatty acid of interest. For models with FA as quartiles, we tested both for overall differences between quartiles as well as for the presence of a linear trend among increasing quartiles of FA. Estimated hazard ratios (HR) and 95% confidence intervals (CI) are presented for each FA quartile (Q2, Q3, Q4) vs. the lowest quartile (Q1).

In the sub-analysis of RBC PUFAs (n=3428), Cox proportional hazards regression models were used to assess the relationships between blood PUFAs and arthroplasty. The confidence interval (CI) for the BMI-adjusted hip and knee models alone was wide, reflecting more sparse data. Therefore, we did not perform multivariable adjustment when considering total hip and knee replacements separately. PUFAs of primary interest were RBC EPA+DHA, AA, LA, and ALA; other model statistical considerations are like those described above. Using imputed RBC PUFA data from Pottala et al.24, final combined HR and CI from all models run separately for each of the ten imputations were calculated using SAS’s MIANALYZE procedure.

All analyses were performed using SAS/STAT software version 14.3 (SAS v9.4 for Windows, SAS Institute, Inc., Cary, NC).

RESULTS

Characteristics of WHI women with or without arthroplasty

Demographic characteristics of the sample population are reported in Table 1. Of the 34,990 women in the sample, 1851 (5.3%) had a THA and 3176 (9.0%) had TKA during the follow-up period. The average length of follow-up time for women from baseline was 12.3 years (SD=6.9 years; range: 0.003 to 23.3 years). Mean (SD) age at baseline was 70.1 (3.7) years and mean age at time of arthroplasty was 78.3 (5.1) years for THA and 77.8 (4.9) years for TKA. Seventy percent of women who had arthroplasty self-reported a diagnosis of arthritis (non-rheumatoid), 30% reported moderate joint pain and 9% reported severe joint pain at baseline. By comparison, 49% of women without arthroplasty reported a diagnosis of arthritis, 17% reported moderate joint pain and 4% reported severe joint pain at baseline. Women with arthroplasty were more likely to have graduated from college, to drink alcohol, consume greater amounts of calcium and vitamin D, and to have obesity (p <0.001 for all), but were similar to women without arthroplasty in self-reported physical activity (p=0.644). Additionally, 71% of women with arthroplasty reported no co-morbid chronic conditions compared to 68% of women without arthroplasty.

Table 1:

Demographic characteristics of WHI women with or without arthroplasty

Variable1 Level No Arthroplasty (n =29,963) Arthroplasty Total (n=34,990) p-value2
Hip (n=1,851) Knee (n=3,176) Total (n=5,027)
Age Baseline Mean (SD) 70.2 (3.7) 69.7 (3.4) 69.4 (3.4) 69.5 (3.4) 70.1 (3.7) <0.001
Race Ethnicity Missing <1% <1% <1% <1% <1% <0.001
White 88% 95% 93% 93% 89%
Black 6% 3% 4% 4% 6%
Other 5% 2% 3% 3% 5%
HT Trial E+P control 5% 6% 5% 6% 5% 0.097
E+P intervention 6% 5% 5% 5% 5%
E control 4% 4% 4% 4% 4%
E intervention 3% 3% 4% 4% 3%
Not in HRT 82% 82% 82% 82% 82%
DM Trial Not in DM 18% 18% 21% 20% 18% <0.001
Control 82% 82% 79% 80% 82%
CaD Trial Not in CaD 8% 9% 10% 9% 9% 0.002
Intervention 9% 9% 10% 10% 9%
Control 83% 82% 80% 81% 83%
Arthritis Missing 1% <1% 1% 1% 1% <0.001
No 50% 34% 26% 29% 47%
Yes 49% 65% 73% 70% 52%
Joint Pain Missing 1% 1% 1% 1% 1% <0.001
Did Not Occur 29% 17% 12% 14% 27%
Mild 48% 49% 44% 46% 48%
Moderate 17% 26% 33% 30% 19%
Severe 4% 8% 10% 9% 5%
Education 0 <1% <1% <1% <1% <1% <0.001
1 5% 3% 4% 4% 5%
2 19% 16% 19% 18% 19%
3 37% 35% 38% 37% 37%
4 37% 47% 38% 41% 38%
BMI Missing (n) 291 13 31 44 335 <0.001
Median 26.1 26.9 28.6 28 26.3
BMI Category Missing 1% 1% 1% 1% 1% <0.001
<25 40% 34% 22% 26% 38%
25–29 36% 37% 37% 37% 36%
30+ 23% 29% 41% 36% 25%
Physical activity (MET-hours per week) Missing 4% 4% 4% 4% 4% 0.644
None 13% 11% 14% 13% 13%
>0 – 3.75 13% 11% 14% 13% 13%
3.75 – 8.75 20% 21% 20% 20% 20%
9.75 – 17.5 23% 24% 23% 23% 23%
≥ 17.5 27% 29% 27% 28% 27%
NSAID No 65% 60% 59% 59% 64% <0.001
Yes 35% 40% 41% 41% 36%
Aspirin No 74% 74% 74% 74% 74% 0.367
Yes 26% 26% 26% 26% 26%
Dietary + Supplement Vitamin D (mcg) Missing 0 0 0 0 0 <0.001
Median 8.7 10.8 10 10.3 8.9
Min 0 0.2 0.1 0.1 0
Max 117.3 54.3 56.1 56.1 117.3
Dietary + Supplement Calcium (mcg) Missing 3 0 0 0 3 <0.001
Median 1052 1160.9 1153.9 1155.3 1066.8
Min 83.6 154.6 132.2 132.3 83.6
Max 8526.1 5713.5 9490.1 9490.1 9490.1
Alcohol Missing 1% <1% 1% 1% 1% <0.001
No or past 31% 22% 28% 26% 30%
<1/month – <7/week 55% 59% 57% 58% 56%
7+/week 13% 18% 14% 16% 13%
Smoking Missing 2% 2% 1% 1% 1% <0.01
None 52% 50% 54% 53% 52%
Former 41% 45% 42% 43% 41%
Current 5% 4% 2% 3% 5%
Sum of Chronic Conditions Missing 7% 7% 7% 7% 7% <0.001
None 68% 73% 70% 71% 69%
1 20% 18% 19% 19% 20%
2 4% 2% 2% 2% 3%
3 – 7 1% <1% <1% <1% 1%
Open in a new tab
1

Note that all frequencies smaller than 11 have been masked per CMS data usage requirements.

2

p-value comparing No arthroplasty versus arthroplasty.

3

Education variable levels: 0 = none-some high school (HS); 1 = HS diploma – GED; 2 = Vocational, training school, some college, or associate degree; 3 = college graduate or more.

Legend: HT, hormone therapy; DM, dietary modification; CaD: calcium and vitamin D; BMI, body mass index; NSAID, nonsteroidal anti-inflammatory agents.

Self-reported intake of dietary fatty acids as a percentage of total energy at baseline in women with and without arthroplasty is presented in Supplementary Table 1. Similar to national norms 28, in this sample of older women, MUFAs were the primary component of dietary fat intake (median 12% kcal; range 1.03–34.22%), followed by SFAs (median 10.35% kcals; range 1.25–34.62%), and PUFAs (median 6.39% kcal; range 0.71–22.19%). Consumption of major fatty acid classes (i.e., SFA, MUFA and PUFA) and individual PUFAs (i.e., n-3 and n-6 PUFAs), did not differ appreciably between arthroplasty groups. Supplementary Table 2 details fatty acid intake in quartiles for the entire sample at baseline. Median n-6 PUFA intake varied from 3.74% kcal (range 0.61%–4.46%) in the lowest quartile (Q1) to 7.97% kcal (range 6.92%–19.7%) in the highest quartile (Q4). Majority of women in Q2 and Q4, and all the women in Q3, consumed n-6 PUFAs within the Acceptable Macronutrient Distribution Range (AMDR), which is 5%–10% kcal.29 Median n-3 PUFA intake varied from 0.49% kcal in Q1 to 1.10% kcal in Q4, with median EPA+DHA intake of 0.2% kcal in Q1 to 0.13% kcal in Q4. The AMDR for n-3 PUFAs is 0.6% – 1.2% kcal, with approximately 0.06%–0.12% kcal from EPA and/or DHA.29 Similar to n-6 PUFA intake, the majority of women in Q2 and Q4, and all of the women in Q3 fell within the AMDR for total n-3 PUFA consumption. However, consumption of EPA+DHA was below the AMDR for over half of the women in our sample.29

Dietary fatty acids and risk of arthroplasty

In unadjusted and multivariable models for risk of THA by dietary fatty acids (Table 2), higher intake of AA was associated with a modestly higher risk of THA (multivariable HR Q4 vs Q1: 1.16; 95% CI: 1.01, 1.34; p for linear trend =0.03). In unadjusted, BMI-adjusted, and multivariable models, EPA+DHA intake was associated with modestly higher risk of THA (multivariable HR for Q4 vs Q1: 1.20; 95% CI:1.05, 1.39; p for linear trend=0.003).

Table 2.

Cox proportional hazard ratios for total hip arthroplasty based on dietary fatty acid intake quartiles1

Dietary Fatty Acid3 Quartile Univariable BMI-Adjusted Multivariable2
HR (95% CI) HR (95% CI) HR (95% CI) p-value4
SFA
pU= 0.861
pB= 0.923
pM= 0.688		 Q2 1.04 (0.92, 1.18) 1 (0.88, 1.13) 1 (0.87, 1.15) 0.535
Q3 1.05 (0.92, 1.2) 1 (0.87, 1.14) 1.08 (0.94, 1.24)
Q4 1.05 (0.92, 1.2) 0.96 (0.84, 1.1) 1.03 (0.88, 1.19)
MUFA
pU= 0.883
pB= 0.514
pM= 0.936		 Q2 0.99 (0.87, 1.13) 0.97 (0.86, 1.11) 0.99 (0.87, 1.13) 0.577
Q3 1.02 (0.89, 1.17) 0.98 (0.85, 1.12) 1.02 (0.89, 1.18)
Q4 0.97 (0.84, 1.11) 0.9 (0.79, 1.04) 1.04 (0.89, 1.2)
PUFA
pU= 0.996
pB= 0.808
pM= 0.899		 Q2 0.99 (0.87, 1.13) 0.97 (0.86, 1.11) 1.02 (0.89, 1.17) 0.492
Q3 0.99 (0.87, 1.13) 0.96 (0.84, 1.09) 1.02 (0.88, 1.17)
Q4 0.98 (0.86, 1.12) 0.94 (0.82, 1.07) 1.06 (0.91, 1.22)
Total n-6
pU= 0.853
pB= 0.478
pM= 0.998		 Q2 0.99 (0.87, 1.12) 0.97 (0.85, 1.1) 1.01 (0.88, 1.16) 0.916
Q3 0.98 (0.86, 1.11) 0.95 (0.83, 1.08) 1.01 (0.88, 1.17)
Q4 0.94 (0.83, 1.08) 0.9 (0.79, 1.03) 1.01 (0.87, 1.17)
LA
pU= 0.873
pB= 0.508
pM= 0.996		 Q2 0.99 (0.87, 1.13) 0.97 (0.86, 1.11) 1.01 (0.89, 1.16) 0.872
Q3 0.98 (0.86, 1.11) 0.95 (0.83, 1.08) 1.01 (0.88, 1.17)
Q4 0.95 (0.83, 1.08) 0.9 (0.79, 1.03) 1.01 (0.88, 1.17)
AA
pU= 0.174
pB= 0.666
pM= 0.189		 Q2 1.08 (0.95, 1.23) 1.06 (0.93, 1.21) 1.08 (0.94, 1.24) 0.030
Q3 1.13 (0.99, 1.29) 1.08 (0.95, 1.23) 1.12 (0.97, 1.29)
Q4 1.15 (1.01, 1.30) 1.07 (0.94, 1.23) 1.16 (1.01, 1.34)
Total n-3
pU= 0.411
pB= 0.369
pM= 0.286		 Q2 0.95 (0.83, 1.08) 0.93 (0.82, 1.06) 0.95 (0.83, 1.1) 0.211
Q3 1.06 (0.93, 1.2) 1.03 (0.91, 1.18) 1.08 (0.94, 1.24)
Q4 0.98 (0.86, 1.11) 0.95 (0.83, 1.08) 1.05 (0.92, 1.21)
ALA
pU= 0.209
pB= 0.165
pM= 0.294		 Q2 0.96 (0.84, 1.09) 0.94 (0.82, 1.07) 0.98 (0.85, 1.12) 0.6
Q3 1.06 (0.93, 1.2) 1.03 (0.91, 1.17) 1.1 (0.96, 1.26)
Q4 0.93 (0.81, 1.06) 0.9 (0.79, 1.03) 1 (0.87, 1.15)
EPA+DHA
pU <0.001
pB <0.001
pM= 0.023		 Q2 1.06 (0.92, 1.21) 1.06 (0.92, 1.22) 1.01 (0.87, 1.17) 0.003
Q3 1.22 (1.07, 1.39) 1.23 (1.08, 1.41) 1.13 (0.98, 1.3)
Q4 1.28 (1.13, 1.46) 1.29 (1.14, 1.48) 1.2 (1.05, 1.39)
Open in a new tab
1

Baseline hazard stratified by trial membership, and age was the time scale

2

Multivariable model contains race, BMI, physical activity, education, history of arthritis(non-rheumatoid), history of joint pain, NSAID/Aspirin use, Calcium and Vitamin D (diet + supplement), alcohol, smoking, and number of chronic conditions.

3

The p-values represent differences between quartiles based on model pU(univariable), pB(bmi-adjusted), pM(multivariable). Pairwise comparisons performed with Q1 as the reference quartile.

4

p-value for linear trend of the multivariable analysis.

Legend: HR, hazard ratio; 95% CI, 95% confidence interval; SFA, saturated fatty acids; MUFA, monounsaturated fatty acids; PUFA, polyunsaturated fatty acids; n-6, omega-6 fatty acids; LA, linoleic acid; AA, arachidonic acid; n-3, omega-3 fatty acids; ALA, alpha linolenic acid; EPA+DHA, eicosatetraenoic acid + docosahexaenoic acid.

Hazard ratios for TKA are presented in Table 3. Risk of TKA increased with higher SFA and MUFA intake in unadjusted models (HR Q4 vs Q1: 1.19; 95% CI: 1.07,1.32 for both SFA and MUFA). However, BMI and multivariable models for risk of TKA revealed no significant associations based on SFA or MUFA intake (p for linear trend 0.752 and 0.454, respectively).

Table 3.

Cox proportional hazard ratios for total knee arthroplasty based on dietary fatty acid intake in quartiles1

Dietary Fatty Acid3 Quartile Univariable BMI-Adjusted Multivariable2
HR (95% CI) HR (95% CI) HR (95% CI) p-value4
SFA
pU <0.001
pB= 0.122
pM= 0.044		 Q2 1.16 (1.05, 1.28) 1.04 (0.94, 1.15) 1.03 (0.93, 1.15) 0.752
Q3 1.24 (1.12, 1.38) 1.08 (0.97, 1.2) 1.11 (0.99, 1.24)
Q4 1.19 (1.07, 1.32) 0.96 (0.87, 1.07) 0.96 (0.85, 1.08)
MUFA
pU= 0.014
pB= 0.832
pM= 0.827		 Q2 1.11 (1, 1.23) 1.04 (0.94, 1.15) 1.04 (0.94, 1.16) 0.454
Q3 1.13 (1.01, 1.25) 1 (0.9, 1.11) 1.03 (0.92, 1.16)
Q4 1.19 (1.07, 1.32) 0.99 (0.89, 1.1) 1.05 (0.94, 1.18)
PUFA
pU= 0.369
pB= 0.805
pM= 0.652		 Q2 1.08 (0.97, 1.19) 1.02 (0.92, 1.13) 1.03 (0.92, 1.14) 0.207
Q3 1.06 (0.96, 1.17) 0.98 (0.89, 1.09) 1.04 (0.93, 1.16)
Q4 1.09 (0.98, 1.21) 0.97 (0.88, 1.08) 1.07 (0.96, 1.2)
Total n-6
pU= 0.099
pB= 0.703
pM= 0.397		 Q2 1.11 (1.01, 1.23) 1.05 (0.95, 1.16) 1.06 (0.95, 1.18) 0.094
Q3 1.08 (0.98, 1.2) 1 (0.91, 1.11) 1.07 (0.96, 1.2)
Q4 1.13 (1.02, 1.25) 1 (0.91, 1.11) 1.1 (0.98, 1.23)
LA
pU= 0.106
pB= 0.738
pM= 0.393		 Q2 1.11 (1, 1.23) 1.05 (0.95, 1.16) 1.05 (0.95, 1.17) 0.09
Q3 1.08 (0.98, 1.2) 1 (0.9, 1.11) 1.07 (0.96, 1.19)
Q4 1.13 (1.02, 1.25) 1 (0.91, 1.11) 1.1 (0.99, 1.23)
AA
pU= 0.011
pB= 0.266
pM= 0.603		 Q2 1.12 (1.01, 1.24) 1.04 (0.94, 1.15) 1.05 (0.94, 1.17) 0.929
Q3 1.18 (1.07, 1.3) 1.04 (0.94, 1.15) 1.07 (0.95, 1.18)
Q4 1.13 (1.03, 1.25) 0.95 (0.86, 1.06) 1.1 (0.9, 1.12)
Total n-3
pU= 0.441
pB= 0.171
pM= 0.881		 Q2 1 (0.91, 1.1) 0.95 (0.86, 1.05) 0.96 (0.87, 1.07) 0.952
Q3 0.99 (0.9, 1.1) 0.95 (0.86, 1.05) 1 (0.9, 1.11)
Q4 0.93 (0.84, 1.03) 0.89 (0.81, 0.99) 0.98 (0.88, 1.1)
ALA
pU= 0.361
pB= 0.224
pM= 0.432		 Q2 0.98 (0.89, 1.08) 0.94 (0.85, 1.04) 0.98 (0.88, 1.09) 0.562
Q3 1.04 (0.95, 1.15) 0.99 (0.9, 1.1) 1.07 (0.96, 1.18)
Q4 0.96 (0.87, 1.06) 0.91 (0.83, 1.01) 1.01 (0.9, 1.12)
EPA+DHA
pU =0.056
pB =0.214
pM= 0.762		 Q2 1.07 (0.97, 1.18) 1.05 (0.95, 1.16) 1.02 (0.92, 1.13) 0.781
Q3 1.11 (1.01, 1.22) 1.1 (0.99, 1.21) 1.03 (0.93, 1.15)
Q4 0.99 (0.89, 1.09) 1 (0.91, 1.11) 0.98 (0.88, 1.09)
Open in a new tab
1

Baseline hazard stratified by trial membership, and age was the time scale

2

Multivariable model contains race, BMI, physical activity, education, history of arthritis(non-rheumatoid), history of joint pain, NSAID/Aspirin use, Calcium and Vitamin D (diet + supplement), alcohol, smoking, and number of chronic conditions.

3

p-values represent differences between quartiles based on model pU(univariable), pB(bmi-adjusted), pM(multivariable). Pairwise comparisons performed with Q1 as the reference quartile.

4

p-value for linear trend

Legend: HR, hazard ratio; 95% CI, 95% confidence interval; SFA, saturated fatty acids; MUFA, monounsaturated fatty acids; PUFA, polyunsaturated fatty acids; n-6, omega-6 fatty acids; LA, linoleic acid; AA, arachidonic acid; n-3, omega-3 fatty acids; ALA, alpha linolenic acid; EPA+DHA, eicosatetraenoic acid + docosahexaenoic acid.

Cox proportional hazards models for risk of arthroplasty overall (knee + hip) by dietary fatty acid intake are presented in Supplementary Table 3. In unadjusted models, saturated fatty acid (SFA) and arachidonic acid (AA) intake were both associated with increased risk for arthroplasty. However, after adjusting for BMI or multivariable models, significance of HRs was attenuated. EPA+DHA intake was associated with higher risk of arthroplasty in unadjusted and BMI-adjusted models. However, after multivariable adjustment the associations were attenuated. There were no significant linear associations of fatty acid intake with arthroplasty in all multivariable models.

Characteristics of women in the WHIMS sub-cohort

Demographic characteristics of the sub-cohort of WHIMS participants are reported in Supplementary Table 4. There were 2917 women with no arthroplasty and 511 women with arthroplasty in this sub-cohort. Like the total sample cohort, approximately 5.3% (n=183) of WHIMS participants reported having THA and 9.6% (n=511) reported TKA. Mean age of participants at baseline was 70.3 (3.7) years. Sixty-five percent of women with arthroplasty reported a diagnosis of non-rheumatoid arthritis at baseline, compared to 47% of women without arthroplasty. Moderate joint pain was reported by 31% and severe joint pain was reported by 6% of women who later had arthroplasty versus 17% reporting moderate joint pain and 3% reporting severe joint pain in the group who did not have arthroplasty. Like the total sample, women with arthroplasty were more likely to have graduated from college, drink alcohol, consume greater amounts of calcium and vitamin D, have obesity, and report no other chronic disease conditions compared to women who did not have arthroplasty.

Baseline concentrations of RBC fatty acids and quartile ranges from 10 imputations per participant are reported in Supplementary Tables 5 and 6. Although there are no established “normal” ranges for RBC fatty acids, sample means for SFA (37.9%), MUFA (13.3%), PUFA (43.6%), total n-6 (35.7%) and total n-3 (7.9%) PUFAs were comparable to data from other large cohorts of adults in the United States 30 31, and 95% CIs were small. Mean RBC PUFA values between women with arthroplasty versus no arthroplasty did not vary appreciably (43.7% versus 43.6%, respectively). There was no difference in mean RBC AA content between women with and without arthroplasty (6.9% in both groups). RBC EPA+DHA content was 5.4% in women with arthroplasty vs. 5.2% in women without arthroplasty.

RBC fatty acids and risk of arthroplasty

Cox proportional hazards models for RBC fatty acids and total arthroplasty (knee + hip) are shown in Supplementary Table 7. There were no significant associations or linear trends between RBC fatty acids and risk of arthroplasty overall. Unadjusted and BMI-adjusted hazards models for RBC fatty acids and THA are presented in Table 4. Consistent with the dietary FA analysis, there was a significant linear trend for higher risk of THA in women with higher RBC EPA+DHA (p for linear trend = 0.04). In unadjusted and BMI-adjusted models, women in the highest quartile (Q4) of RBC EPA+DHA had increased risk of THA when compared to women in Q1 (BMI-adjusted HR: 1.71; 95% CI: 1.07, 2.75). However, the CIs for each individual quartile vs. Q1 after BMI adjustment were large and with considerable overlap between quartiles, indicating no significant difference between quartiles of RBC EPA+DHA (p=0.102) (Supplementary Table 8). There were no significant associations found between RBC fatty acids and TKA (Table 5).

Table 4.

Cox proportional hazard ratios for total hip arthroplasty based on quartiles of red blood cell fatty acids in WHIMS participants1

RBC Fatty Acid Quartile2 Univariable BMI-Adjusted
HR (95% CI) HR (95% CI) p-value3
SFA Q2 0.82 (0.47, 1.43) 0.81 (0.47, 1.42) 0.703
Q3 0.88 (0.52, 1.48) 0.86 (0.51, 1.43)
Q4 0.93 (0.59, 1.47) 0.9 (0.57, 1.42)
MUFA Q2 1.24 (0.6, 2.58) 1.23 (0.6, 2.53) 0.586
Q3 1.13 (0.58, 2.19) 1.13 (0.59, 2.18)
Q4 1.08 (0.64, 1.83) 1.05 (0.62, 1.78)
PUFA Q2 1.1 (0.65, 1.85) 1.11 (0.66, 1.88) 0.666
Q3 1.17 (0.72, 1.92) 1.2 (0.73, 1.98)
Q4 1.2 (0.78, 1.85) 1.23 (0.8, 1.9)
Total n-6 Q2 0.81 (0.48, 1.39) 0.8 (0.47, 1.36) 0.976
Q3 0.86 (0.52, 1.4) 0.84 (0.51, 1.37)
Q4 0.79 (0.5, 1.25) 0.8 (0.5, 1.26)
LA Q2 0.92 (0.5, 1.67) 0.9 (0.5, 1.62) 0.425
Q3 0.89 (0.57, 1.4) 0.87 (0.56, 1.36)
Q4 0.73 (0.44, 1.2) 0.74 (0.44, 1.22)
AA Q2 0.88 (0.49, 1.61) 0.87 (0.48, 1.6) 0.506
Q3 0.87 (0.55, 1.37) 0.86 (0.54, 1.37)
Q4 1.03 (0.66, 1.62) 1.05 (0.67, 1.64)
Total n-3 Q2 0.98 (0.53, 1.83) 1.01 (0.53, 1.92) 0.064
Q3 1.26 (0.74, 2.14) 1.31 (0.76, 2.26)
Q4 1.58 (1, 2.5) 1.65 (1.03, 2.63)
ALA Q2 1.11 (0.67, 1.82) 1.1 (0.67, 1.84) 0.399
Q3 0.98 (0.52, 1.84) 0.99 (0.54, 1.84)
Q4 0.98 (0.52, 1.47) 0.89 (0.52, 1.52)
EPA+DHA Q2 1.05 (0.63, 1.75) 1.07 (0.64, 1.78) 0.040
Q3 1.31 (0.78, 2.2) 1.34 (0.79, 2.28)
Q4 1.67 (1.04, 2.69) 1.71 (1.07, 2.75)
Open in a new tab
1

Baseline hazard was stratified by trial membership, and age was the time scale.

2

Pairwise comparisons performed with Q1 as the reference quartile.

3

p-value for linear trend

Note: due to the small number of events, multivariable models are not included.

Legend: HR, hazard ratio; 95% CI, 95% confidence interval; SFA, saturated fatty acids; MUFA, monounsaturated fatty acids; PUFA, polyunsaturated fatty acids; n-6, omega-6 fatty acids; LA, linoleic acid; AA, arachidonic acid; n-3, omega-3 fatty acids; ALA, alpha linolenic acid; EPA+DHA, eicosatetraenoic acid + docosahexaenoic acid.

Table 5.

Cox proportional hazard ratios for total knee arthroplasty based on quartiles of red blood cell fatty acids in WHIMS participants1

RBC Fatty Acid Quartile2 Univariable BMI-Adjusted
HR (95% CI) HR (95% CI) p-value3
SFA Q2 1 (0.62, 1.62) 0.95 (0.59, 1.54) 0.948
Q3 1.15 (0.77, 1.72) 1.07 (0.71, 1.61)
Q4 1.07 (0.72, 1.59) 0.96 (0.65, 1.43)
MUFA Q2 1.19 (0.74, 1.91) 1.17 (0.74, 1.86) 0.471
Q3 1.22 (0.81, 1.82) 1.18 (0.79, 1.75)
Q4 1.4 (0.99, 2) 1.35 (0.94, 1.93)
PUFA Q2 1.03 (0.69, 1.54) 1.08 (0.72, 1.61) 0.488
Q3 0.93 (0.61, 1.43) 0.97 (0.64, 1.47)
Q4 0.87 (0.6, 1.28) 0.94 (0.64, 1.37)
Total n-6 Q2 1.04 (0.69, 1.56) 1.02 (0.67, 1.54) 0.786
Q3 1.04 (0.69, 1.58) 1.01 (0.65, 1.55)
Q4 0.96 (0.68, 1.34) 0.96 (0.69, 1.35)
LA Q2 0.96 (0.64, 1.44) 0.96 (0.64, 1.44) 0.785
Q3 1.06 (0.72, 1.55) 1.06 (0.72, 1.56)
Q4 0.95 (0.69, 1.33) 1.01 (0.72, 1.42)
AA Q2 1.06 (0.74, 1.51) 1.07 (0.75, 1.53) 0.578
Q3 0.97 (0.67, 1.4) 0.98 (0.68, 1.4)
Q4 0.96 (0.68, 1.34) 0.97 (0.69, 1.37)
Total n-3 Q2 1.18 (0.78, 1.77) 1.21 (0.79, 1.86) 0.47
Q3 0.98 (0.64, 1.5) 1.02 (0.67, 1.57)
Q4 0.99 (0.68, 1.43) 1.08 (0.73, 1.58)
ALA Q2 0.99 (0.67, 1.45) 1.04 (0.71, 1.53) 0.678
Q3 1.01 (0.72, 1.41) 1.08 (0.77, 1.51)
Q4 1.03 (0.64, 1.67) 1.14 (0.71, 1.82)
EPA+DHA Q2 1.1 (0.7, 1.72) 1.12 (0.71, 1.77) 0.529
Q3 0.93 (0.64, 1.35) 0.95 (0.65, 1.39)
Q4 0.94 (0.65, 1.35) 1 (0.69, 1.45)
Open in a new tab
1

Baseline hazard was stratified by trial membership, and age was the time scale.

2

Pairwise comparisons performed with Q1 as the reference quartile.

3

p-value for linear trend

Note: due to the small number of events, multivariable models are not included.

Legend: HR, hazard ratio; 95% CI, 95% confidence interval; SFA, saturated fatty acids; MUFA, monounsaturated fatty acids; PUFA, polyunsaturated fatty acids; n-6, omega-6 fatty acids; LA, linoleic acid; AA, arachidonic acid; n-3, omega-3 fatty acids; ALA, alpha linolenic acid; EPA+DHA, eicosatetraenoic acid + docosahexaenoic acid.

Sensitivity Analyses

Since women who had severe OA but chose not to have arthroplasty were potentially different than those who chose arthroplasty, we performed sensitivity analyses excluding 1040 women with self-reported non-rheumatoid arthritis + severe joint pain + no arthroplasty. In hazards models evaluating dietary intake of fatty acids and risk of arthroplasty, THA and TKA, results were nearly identical to the original analyses. These analyses were repeated in the WHIMS sub-cohort excluding 83 women with self-reported non-rheumatoid arthritis + severe joint pain + no arthroplasty, with no appreciable differences in results from the main study findings (data not shown).

DISCUSSION

Chronic low-grade inflammation may contribute to osteoarthritis incidence and progression to severe disease.9,10 Dietary components that alter inflammation, such as fatty acids, may modulate risk of osteoarthritis.32 Results of this study indicated that higher baseline AA intake and higher baseline EPA+DHA intake were associated with greater risk of THA among older women. There were no associations with fatty acid intake or biomarkers and TKA.

Our finding that higher self-reported intake of AA was associated with higher risk for THA is not unexpected. Dietary intake of AA, as well as other fatty acids, can impact the composition of phospholipid membranes in tissue and blood.11,33 AA metabolites, such as prostaglandin E2 (PGE2), predict disease severity in patients with symptomatic knee OA 12 and have been shown to be involved in OA joint inflammation and cartilage degradation.11,34 Additionally, higher plasma AA is associated with synovitis in patients with radiographic OA.13 However, we did not find evidence that RBC AA was associated with risk for THA or TKA. This may be because RBC AA is influenced by metabolism of linoleic acid as well as long-term dietary intake of AA.35 The role of linoleic acid in arthritic incidence and severity is complex and may differ depending upon the food sources, and therefore the isomers of linoleic acid consumed.36 RBC linoleic acid has been associated with lower inflammation and lower risk for development of inflammatory conditions such as rheumatoid arthritis.37 Further investigation is warranted into metabolites of AA such as PGE2 as well as interactions between a variety of dietary fatty acids in relation to OA.

Contrary to our hypothesis, higher intake of the n-3 PUFAs, EPA+DHA, was associated with modestly higher risk of THA. Based on prior research38, WHI participants had very low intake of marine n-3 PUFAs overall (median of 20 mg/day in lowest quartile and 130 mg/day in highest quartile of intake) and low concentrations of EPA+DHA in the blood.38,39 Because of the lack of variation in the range of EPA+DHA consumed by this cohort of older women, we had limited ability to evaluate benefit, if any, of meeting recommendations for EPA+DHA intake (i.e. ~500 mg/day) promoted by multiple health organizations.39 It is possible that any benefit of n-3 PUFAs in relation to OA may require a threshold of consumption not obtained by the women in this sample. Indeed, several studies using EPA+DHA supplementation suggest that participants supplemented with EPA+DHA have fewer THA or TKA compared to those not supplemented.40 Additionally, use of oils from marine sources have shown potential in improving pain of people with knee OA.41,42 It is possible that women with more symptomatic OA (i.e. more pain) at baseline may have increased consumption of n-3 PUFA rich foods in response to symptoms. Diets high in EPA and DHA, such as the Mediterranean diet, also have been associated with improvements of pain in OA patients.43 However, other observational studies found no significant associations of fatty acids with incidence of OA.44,45 The difference in findings between intervention and observational studies may suggest that the dose of n-3 PUFAs needed to benefit individuals with OA exceeds usual intake of n-3 PUFAs consumed without supplementation, especially in participants living in areas that have low intakes of marine foods. In a prior case-control study of hip fracture within WHI, mean EPA+DHA consumption was estimated to be 130 mg/d, similar to the highest quartile of intake in this sample, but still well below health recommendations.38,39

SFA intake was associated with higher risk of TKA in unadjusted models, but after adjusting for BMI and other covariates, this relationship was attenuated. This may be because SFAs, specifically longer chain SFAs like palmitic and stearic acid, contribute to both obesity and osteoarthritic changes in bone and cartilage.46 Although associations of SFA with risk of TKA were attenuated with multivariable analysis, unadjusted associations between SFA and risk of TKA agree with reported findings from the Osteoarthritis Initiative.21

We found relationships with THA in this study however, not with either TKA or with total arthroplasty. We found that EPA+DHA and AA were associated with THA while in unadjusted models SFA were associated with TKA. We would expect there to be differences in which fatty acids impact risk of THA or TKA as the etiology of OA of the hip and knee differ. For example, obesity is a primary risk factor for knee OA, but the relationship between obesity and hip OA is less clear.47 SFA intake has been shown to be associated with obesity which increases weight-bearing on the knee joint, potentially impacting the need for TKA 46 Heightened local inflammation of the periarticular muscle surrounding the hip joint has been found in a significant sub-set of patients undergoing surgery for THA.48 It is possible that dietary intake of PUFAs may play a role in regulating this local inflammatory milieu.

It is not surprising that we found no association of fatty acids with total arthroplasty (i.e. the calculated variable based on THA+TKA). As is seen in the general population, there were almost twice as many TKAs compared to THAs in our sample. Thus, results of the TKA analyses drive the results for total arthroplasty, especially because significant relationships between PUFAs and THA were modest in our analysis. Ultimately, fatty acids may affect knee and hip OA in unique ways based on differences in risk factors, the inflammatory environment and pathophysiology of the arthritic joints.

This study has several limitations to be discussed. Due to the observational nature of the study, we were unable to establish causality. Additionally, arthroplasty was used as an objective surrogate for severe OA because >95% of THA or TKA procedures are performed for this diagnosis, but we acknowledge that this is not a perfect proxy. We attempted to address this by including appropriate covariates in statistical models and completed sensitivity analyses showing that results did not significantly differ. However, patients with severe OA could choose not to undergo arthroplasty for a variety of reasons (e.g. family support, fear of surgery, etc) and some of these reasons could not be captured with statistical adjustment. Another limitation is the low consumption of n-3 PUFAs by women in the WHI at baseline and no data on fatty acid supplement use. N-3 supplement use was low in the population during the initiation of WHI 38,49. However, over the course of the study follow-up period, dietary supplement use increased in older adults in the United States, and by 2017–2018 nearly 22% of adults over 60 years reported taking n-3 supplements. The impact of this change in n-3 supplement use was not able to be measured in the WHI participants. Finally, dietary data was collected through a self-reported questionnaire which is vulnerable to misclassification or reporting bias.50 In conclusion, associations between dietary FAs and RBC FAs with risk of THA and TKA, surrogate measures of severe OA, were variable. Higher baseline dietary intakes of the n-6 PUFA, AA, and the n-3 PUFAs, EPA+DHA, were moderately associated with risk for THA. Although baseline dietary intake of EPA+DHA was low in this sample and supplemental n-3 PUFA data were not available, the association of EPA+DHA and THA were supported by RBC EPA+DHA data. No associations were observed between baseline dietary FA intake and risk of TKA or between baseline RBC FAs with overall risk of arthroplasty or risk of TKA. These results suggest a potential role of fatty acids in risk of THA and TKA. Further investigation using randomized control trials and cohorts focused specifically on OA, diet, and dietary supplement intake, is needed to better elucidate the relationship between fatty acids and severity of OA.

Supplementary Material

Supinfo

Significance & Innovations.

  • Incidence of severe osteoarthritis and subsequent arthroplasty is predicted to increase over the next decades which will have great impacts on public health and health care systems; thus, there is a need to understand preventative measures.

  • The Women’s Health Initiative combined with Medicare data, provided a large diverse cohort of postmenopausal women to investigate the association of dietary fatty acids and red blood cell biomarkers with incidence of total hip and total knee arthroplasty.

  • We identified that dietary intakes of arachidonic acid and eicosapentaenoic acid + docosahexaenoic acid showed modest association with risk of total hip arthroplasty after adjustment for demographic and clinical covariates.

  • Our results demonstrated that there is a need for further investigation into the relationship between fatty acids and risk of arthroplasty and osteoarthritis. Further studies should include controlled conditions for dietary intake of fatty acids.

ACKNOWLEDGEMENTS

The WHI program is funded by the National Heart, Lung, and Blood Institute, National Institutes of Health, U.S. Department of Health and Human Services through contracts HHSN268201600018C, HHSN268201600001C, HHSN268201600002C, HHSN268201600003C, and HHSN268201600004C.

Program Office: (National Heart, Lung, and Blood Institute, Bethesda, Maryland) Jacques Rossouw, Shari Ludlam, Joan McGowan, Leslie Ford, and Nancy Geller Clinical Coordinating Center: (Fred Hutchinson Cancer Research Center, Seattle, WA) Garnet Anderson, Ross Prentice, Andrea LaCroix, and Charles Kooperberg Investigators and Academic Centers: (Brigham and Women’s Hospital, Harvard Medical School, Boston, MA) JoAnn E. Manson; (MedStar Health Research Institute/Howard University, Washington, DC) Barbara V. Howard; (Stanford Prevention Research Center, Stanford, CA) Marcia L. Stefanick; (The Ohio State University, Columbus, OH) Rebecca Jackson; (University of Arizona, Tucson/Phoenix, AZ) Cynthia A. Thomson; (University at Buffalo, Buffalo, NY) Jean Wactawski-Wende; (University of Florida, Gainesville/Jacksonville, FL) Marian Limacher; (University of Iowa, Iowa City/Davenport, IA) Jennifer Robinson; (University of Pittsburgh, Pittsburgh, PA) Lewis Kuller; (Wake Forest University School of Medicine, Winston-Salem, NC) Sally Shumaker; (University of Nevada, Reno, NV) Robert Brunner Women’s Health Initiative Memory Study: (Wake Forest University School of Medicine, Winston-Salem, NC) Mark Espeland

We would like to thank the efforts of Amy Lehman, MAS, and Dr. Michael Pennell, PhD for their statistical support.

REFERENCES

  • 1.Centers for Disease Control and Prevention. Updated September 1, 2011. Accessed July 25, 2013, 2013. http://www.cdc.gov/arthritis/basics/osteoarthritis.htm [Google Scholar]
  • 2.Cirillo DJ, Wallace RB, Wu L, Yood RA. Effect of hormone therapy on risk of hip and knee joint replacement in the Women’s Health Initiative. Multicenter Study Randomized Controlled Trial Research Support, N.I.H., Extramural. Arthritis and rheumatism. Oct 2006;54(10):3194–204. doi: 10.1002/art.22138 [DOI] [PubMed] [Google Scholar]
  • 3.Dougados M, Gueguen A, Nguyen M, et al. Requirement for total hip arthroplasty: an outcome measure of hip osteoarthritis? Clinical Trial Multicenter Study. J Rheumatol. Apr 1999;26(4):855–61. [PubMed] [Google Scholar]
  • 4.United States Bone and Joint Initiative: The Burden of Musculoskeletal Diseases in the United States (BMUS). United States Bone and Joint Initiative. Updated 2022. Accessed April 7, 2022. https://www.boneandjointburden.org/fourth-edition/iiib70/joint-pain-and-joint-replacement [Google Scholar]
  • 5.Singh JA, Yu S, Chen L, Cleveland JD. Rates of Total Joint Replacement in the United States: Future Projections to 2020–2040 Using the National Inpatient Sample. J Rheumatol. Sep 2019;46(9):1134–1140. doi: 10.3899/jrheum.170990 [DOI] [PubMed] [Google Scholar]
  • 6.Farah H, Wijesinghe SN, Nicholson T, et al. Differential Metabotypes in Synovial Fibroblasts and Synovial Fluid in Hip Osteoarthritis Patients Support Inflammatory Responses. Int J Mol Sci. Mar 17 2022;23(6)doi: 10.3390/ijms23063266 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Kapoor M, Martel-Pelletier J, Lajeunesse D, Pelletier JP, Fahmi H. Role of proinflammatory cytokines in the pathophysiology of osteoarthritis. Review. Nature reviews Rheumatology. Jan 2011;7(1):33–42. doi: 10.1038/nrrheum.2010.196 [DOI] [PubMed] [Google Scholar]
  • 8.Hill CL, Hunter DJ, Niu J, et al. Synovitis detected on magnetic resonance imaging and its relation to pain and cartilage loss in knee osteoarthritis. Research Support, N.I.H., Extramural. Annals of the rheumatic diseases. Dec 2007;66(12):1599–603. doi: 10.1136/ard.2006.067470 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Guermazi A, Hayashi D, Roemer FW, et al. Synovitis in Knee Osteoarthritis Assessed by Contrast-enhanced Magnetic Resonance Imaging (MRI) is Associated with Radiographic Tibiofemoral Osteoarthritis and MRI-detected Widespread Cartilage Damage: The MOST Study. The Journal of Rheumatology. March 1, 2014. 2014;41(3):501–508. doi: 10.3899/jrheum.130541 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Berenbaum F Osteoarthritis as an inflammatory disease (osteoarthritis is not osteoarthrosis!). Review. Osteoarthritis Cartilage. Jan 2013;21(1):16–21. doi: 10.1016/j.joca.2012.11.012 [DOI] [PubMed] [Google Scholar]
  • 11.Sheibanie AF, Khayrullina T, Safadi FF, Ganea D. Prostaglandin E2 exacerbates collagen-induced arthritis in mice through the inflammatory interleukin-23/interleukin-17 axis. Arthritis and rheumatism. Aug 2007;56(8):2608–19. [DOI] [PubMed] [Google Scholar]
  • 12.Attur M SA, Aliferis CF, Li Z, Krasnokutsky S, Samuels J, Greenberg JD, Patel J, Oh C, Lu QA, Ramirez R, Todd J, Abramson SB. Inflammatory genomic and plasma biomarkers predict progression of symptomatic knee OA (SKOA). Abstract. Osteoarthritis and Cartilage. 2013 2012;20:S34–S35. [Google Scholar]
  • 13.Baker KR, Matthan NR, Lichtenstein AH, et al. Association of plasma n-6 and n-3 polyunsaturated fatty acids with synovitis in the knee: the MOST study. Multicenter Study Research Support, N.I.H., Extramural. Osteoarthritis Cartilage. May 2012;20(5):382–7. doi: 10.1016/j.joca.2012.01.021 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Poulsen RC, Gotlinger KH, Serhan CN, Kruger MC. Identification of inflammatory and proresolving lipid mediators in bone marrow and their lipidomic profiles with ovariectomy and omega-3 intake. American journal of hematology. Jun 2008;83(6):437–45. [DOI] [PubMed] [Google Scholar]
  • 15.Harris WS, Mozaffarian D, Rimm E, et al. Omega-6 Fatty Acids and Risk for Cardiovascular Disease. A Science Advisory From the American Heart Association Nutrition Subcommittee of the Council on Nutrition, Physical Activity, and Metabolism; Council on Cardiovascular Nursing; and Council on Epidemiology and Prevention. Circulation. January 26, 2009. 2009:CIRCULATIONAHA.108.191627. doi: 10.1161/circulationaha.108.191627 [DOI] [PubMed] [Google Scholar]
  • 16.Curtis CL, Hughes CE, Flannery CR, Little CB, Harwood JL, Caterson B. n-3 Fatty Acids Specifically Modulate Catabolic Factors Involved in Articular Cartilage Degradation. Journal of Biological Chemistry. January 14, 2000. 2000;275(2):721–724. [DOI] [PubMed] [Google Scholar]
  • 17.Zainal Z, Longman AJ, Hurst S, et al. Relative efficacies of omega-3 polyunsaturated fatty acids in reducing expression of key proteins in a model system for studying osteoarthritis. Osteoarthritis Cartilage. Jul 2009;17(7):896–905. [DOI] [PubMed] [Google Scholar]
  • 18.Huang M-j, Wang L, Jin D-d, et al. Enhancement of the synthesis of n-3 PUFAs in fat-1 transgenic mice inhibits mTORC1 signalling and delays surgically induced osteoarthritis in comparison with wild-type mice. Annals of the rheumatic diseases. September 1, 2014 2014;73(9):1719–1727. doi: 10.1136/annrheumdis-2013-203231 [DOI] [PubMed] [Google Scholar]
  • 19.Wu C-L, Jain D, McNeill JN, et al. Dietary fatty acid content regulates wound repair and the pathogenesis of osteoarthritis following joint injury. Ann Rheum Dis. July 10, 2014 2014;doi: 10.1136/annrheumdis-2014-205601 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Wann AK, Mistry J, Blain EJ, Michael-Titus AT, Knight MM. Eicosapentaenoic acid and docosahexaenoic acid reduce interleukin-1beta-mediated cartilage degradation. Research Support, Non-U.S. Gov’t. Arthritis research & therapy. 2010;12(6):R207. doi: 10.1186/ar3183 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Lu B, Driban JB, Xu C, Lapane KL, McAlindon TE, Eaton CB. Dietary Fat and Progression of Knee Osteoarthritis Dietary Fat Intake and Radiographic Progression of Knee Osteoarthritis: Data from the Osteoarthritis Initiative. Arthritis care & research. Jun 6 2016;doi: 10.1002/acr.22952 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Design of the Women’s Health Initiative clinical trial and observational study. The Women’s Health Initiative Study Group. Controlled clinical trials. Feb 1998;19(1):61–109. [DOI] [PubMed] [Google Scholar]
  • 23.Shumaker SA, Reboussin BA, Espeland MA, et al. The Women’s Health Initiative Memory Study (WHIMS): a trial of the effect of estrogen therapy in preventing and slowing the progression of dementia. Controlled clinical trials. Dec 1998;19(6):604–21. doi: 10.1016/s0197-2456(98)00038-5 [DOI] [PubMed] [Google Scholar]
  • 24.Pottala JV, Espeland MA, Polreis J, Robinson J, Harris WS. Correcting the effects of −20 degrees C storage and aliquot size on erythrocyte fatty acid content in the Women’s Health Initiative. Research Support, N.I.H., Extramural. Lipids. Sep 2012;47(9):835–46. doi: 10.1007/s11745-012-3693-y [DOI] [PubMed] [Google Scholar]
  • 25.Harris WS, Luo J, Pottala JV, et al. Red blood cell polyunsaturated fatty acids and mortality in the Women’s Health Initiative Memory Study. J Clin Lipidol. Jan-Feb 2017;11(1):250–259 e5. doi: 10.1016/j.jacl.2016.12.013 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Patterson RE, Kristal AR, Tinker LF, Carter RA, Bolton MP, Agurs-Collins T. Measurement characteristics of the Women’s Health Initiative food frequency questionnaire. Annals of epidemiology. Apr 1999;9(3):178–87. [DOI] [PubMed] [Google Scholar]
  • 27.Schakel SF, Sievert YA, Buzzard IM. Sources of data for developing and maintaining a nutrient database. Journal of the American Dietetic Association. Oct 1988;88(10):1268–71. [PubMed] [Google Scholar]
  • 28.Agriculture USDo. What we eat in America, NHANES 2017–2018. Accessed January 30, 2023, 2023. https://www.ars.usda.gov/northeast-area/beltsville-md-bhnrc/beltsville-human-nutrition-research-center/food-surveys-research-group/docs/wweia-data-tables/ [Google Scholar]
  • 29.The National Academies Press. [Google Scholar]
  • 30.Harris WS, Pottala JV, Varvel SA, Borowski JJ, Ward JN, McConnell JP. Erythrocyte omega-3 fatty acids increase and linoleic acid decreases with age: observations from 160,000 patients. Prostaglandins Leukot Essent Fatty Acids. Apr 2013;88(4):257–63. doi: 10.1016/j.plefa.2012.12.004 [DOI] [PubMed] [Google Scholar]
  • 31.Harris WS, Pottala JV, Lacey SM, Vasan RS, Larson MG, Robins SJ. Clinical correlates and heritability of erythrocyte eicosapentaenoic and docosahexaenoic acid content in the Framingham Heart Study. Atherosclerosis. Dec 2012;225(2):425–31. doi: 10.1016/j.atherosclerosis.2012.05.030 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Messier SP, Mihalko SL, Legault C, et al. Effects of intensive diet and exercise on knee joint loads, inflammation, and clinical outcomes among overweight and obese adults with knee osteoarthritis: the IDEA randomized clinical trial. Jama. Sep 25 2013;310(12):1263–73. doi: 10.1001/jama.2013.277669 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Seyberth HW, Oelz O, Kennedy T, et al. Increased arachidonate in lipids after administration to man: effects on prostaglandin biosynthesis. Clin Pharmacol Ther. Nov 1975;18(5 Pt 1):521–9. doi: 10.1002/cpt1975185part1521 [DOI] [PubMed] [Google Scholar]
  • 34.Abshirini M, Ilesanmi-Oyelere BL, Kruger MC. Potential modulatory mechanisms of action by long-chain polyunsaturated fatty acids on bone cell and chondrocyte metabolism. Progress in lipid research. Jul 2021;83:101113. doi: 10.1016/j.plipres.2021.101113 [DOI] [PubMed] [Google Scholar]
  • 35.Lasserre M, Mendy F, Spielmann D, Jacotot B. Effects of different dietary intake of essential fatty acids on C20:3 omega 6 and C20:4 omega 6 serum levels in human adults. Lipids. Apr 1985;20(4):227–33. doi: 10.1007/bf02534193 [DOI] [PubMed] [Google Scholar]
  • 36.Muhlenbeck JA, Butz DE, Olson JM, Uribe-Cano D, Cook ME. Dietary Conjugated Linoleic Acid-c9t11 Prevents Collagen-Induced Arthritis, Whereas Conjugated Linoleic Acid-t10c12 Increases Arthritic Severity. Lipids. Apr 2017;52(4):303–314. doi: 10.1007/s11745-017-4241-6 [DOI] [PubMed] [Google Scholar]
  • 37.de Pablo P, Romaguera D, Fisk HL, et al. High erythrocyte levels of the n-6 polyunsaturated fatty acid linoleic acid are associated with lower risk of subsequent rheumatoid arthritis in a southern European nested case-control study. Ann Rheum Dis. Jul 2018;77(7):981–987. doi: 10.1136/annrheumdis-2017-212274 [DOI] [PubMed] [Google Scholar]
  • 38.Orchard TS, Cauley JA, Frank GC, et al. Fatty acid consumption and risk of fracture in the Women’s Health Initiative. Am J Clin Nutr. Dec 2010;92(6):1452–60. doi: 10.3945/ajcn.2010.29955 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Gebauer SK, Psota TL, Harris WS, Kris-Etherton PM. n-3 fatty acid dietary recommendations and food sources to achieve essentiality and cardiovascular benefits. Am J Clin Nutr. Jun 2006;83(6 Suppl):1526S–1535S. [DOI] [PubMed] [Google Scholar]
  • 40.Alfaddagh A, Elajami TK, Saleh M, Elajami M, Bistrian BR, Welty FK. The effect of eicosapentaenoic and docosahexaenoic acids on physical function, exercise, and joint replacement in patients with coronary artery disease: A secondary analysis of a randomized clinical trial. J Clin Lipidol. Jul-Aug 2018;12(4):937–947.e2. doi: 10.1016/j.jacl.2018.03.080 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Hill CL, March LM, Aitken D, et al. Fish oil in knee osteoarthritis: a randomised clinical trial of low dose versus high dose. Multicenter Study Randomized Controlled Trial Research Support, Non-U.S. Gov’t. Annals of the rheumatic diseases. Jan 2016;75(1):23–9. doi: 10.1136/annrheumdis-2014-207169 [DOI] [PubMed] [Google Scholar]
  • 42.Stonehouse W, Benassi-Evans B, Bednarz J, Vincent AD, Hall S, Hill CL. Krill oil improved osteoarthritic knee pain in adults with mild to moderate knee osteoarthritis: a 6-month multicenter, randomized, double-blind, placebo-controlled trial. Am J Clin Nutr. Sep 2 2022;116(3):672–685. doi: 10.1093/ajcn/nqac125 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Sadeghi A, Zarrinjooiee G, Mousavi SN, Abdollahi Sabet S, Jalili N. Effects of a Mediterranean Diet Compared with the Low-Fat Diet on Patients with Knee Osteoarthritis: A Randomized Feeding Trial. Int J Clin Pract. 2022;2022:7275192. doi: 10.1155/2022/7275192 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Felson DT, Misra D, LaValley M, et al. Fatty acids and osteoarthritis: the MOST study. Osteoarthritis Cartilage. Jul 2021;29(7):973–978. doi: 10.1016/j.joca.2021.03.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Loef M, Ioan-Facsinay A, Mook-Kanamori DO, et al. The association of plasma fatty acids with hand and knee osteoarthritis: the NEO study. Osteoarthritis Cartilage. Feb 2020;28(2):223–230. doi: 10.1016/j.joca.2019.10.002 [DOI] [PubMed] [Google Scholar]
  • 46.Sekar S, Shafie SR, Prasadam I, et al. Saturated fatty acids induce development of both metabolic syndrome and osteoarthritis in rats. Sci Rep. Apr 18 2017;7:46457. doi: 10.1038/srep46457 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Garstang SV, Stitik TP. Osteoarthritis: epidemiology, risk factors, and pathophysiology. Am J Phys Med Rehabil. Nov 2006;85(11 Suppl):S2–11; quiz S12–4. doi: 10.1097/01.phm.0000245568.69434.1a [DOI] [PubMed] [Google Scholar]
  • 48.Bridges SL Jr., Sun D, Graham ZA, McAdam JS, Mayo ED, Bamman MM. Muscle Inflammation Susceptibility: A Potential Phenotype for Guiding Precision Rehabilitation After Total Hip Arthroplasty in End-Stage Osteoarthritis. Hss j. Nov 2023;19(4):453–458. doi: 10.1177/15563316231190402 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Timbo BB, Ross MP, McCarthy PV, Lin CT. Dietary supplements in a national survey: Prevalence of use and reports of adverse events. Journal of the American Dietetic Association. Dec 2006;106(12):1966–74. [DOI] [PubMed] [Google Scholar]
  • 50.Ravelli MN, Schoeller DA. Traditional Self-Reported Dietary Instruments Are Prone to Inaccuracies and New Approaches Are Needed. Front Nutr. 2020;7:90. doi: 10.3389/fnut.2020.00090 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supinfo
NIHMS1971615-supplement-Supinfo.docx (61.2KB, docx)

RESOURCES