Fish Consumption and Risk of Rheumatoid Arthritis Among Women in Large Prospective Cohorts

Jeffrey A. Sparks¹, Shun-Chiao Chang², Bing Lu¹, Susan Malspeis¹, Karen H. Costenbader¹ and Elizabeth W. Karlson¹, ¹Division of Rheumatology, Immunology and Allergy, Brigham and Women's Hospital, Harvard Medical School, Boston, MA, ²Channing Division of Network Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, MA

Meeting: 2014 ACR/ARHP Annual Meeting

Keywords: Environmental factors, nutrition, pathogenesis and rheumatoid arthritis (RA)

Session Information

Title: Epidemiology and Public Health (ACR): Rheumatoid Arthritis Pathogenesis and Treatment

Session Type: Abstract Submissions (ACR)

Background/Purpose

Prior studies have suggested a protective effect of fish intake on RA. However, these studies were limited by potential recall bias, short follow-up, and small sample sizes. We aimed to evaluate the effect of fish consumption on RA development in two large cohorts, the Nurses’ Health Study (NHS) and NHSII.

Methods

We examined the effect of fish consumption on RA risk among women in two large, prospective cohorts. NHS is composed of 121,700 US nurses followed since 1976; NHSII is composed of 116,430 nurses followed since 1989. Lifestyle and environmental exposures were collected through biennial questionnaires. Diet and fish intake were assessed by a semi-quantitative food frequency questionnaire completed every 4 years. Participants who provided fish data at baseline in each cohort were analyzed. Incident RA cases were identified by screening questionnaire and validated by medical record review according to the 1987 ACR RA criteria. Cumulative average fish intake prior to RA development was used to represent long-term fish intake. Cox regression models estimated hazard ratios (HR) and 95% confidence intervals (CI) of fish intake and RA risk, adjusted for confounders, including age, total energy intake, smoking, body mass index, alcohol, and soda intake. HRs from the two cohorts were meta-analyzed using a random effects model.

Results

We validated 652 incident RA cases among 86,135 women in NHS and 322 incident RA cases among 92,984 women in NHSII with diet data. For women in NHSII, modest fish consumption (1-2 fish servings per week) was associated with a protective effect on RA (age-adjusted HR 0.63, 95% CI 0.42-0.96, Table) compared to never/rare fish consumption that was nearly significant in the multivariable model (HR 0.66, 95% CI 0.44-1.01). However, for women in NHS, modest fish consumption was not associated with RA (HR 0.86, 95% CI 0.57-1.32) compared to never/rare fish intake. In meta-analysis of the results, there was a suggestion of a protective effect for modest fish intake that was nearly statistically significant (HR 0.76, 95% CI 0.56-1.02). More frequent fish consumption was not statistically associated with RA (HR 0.94 for 2 to <3 servings/week; HR 0.88 for ≥3 servings per week compared to never/rare intake) and there was no statistically significant trend for fish consumption on RA risk.

Conclusion

In these large prospective studies of nearly 180,000 women, there was a suggestion of a protective effect of modest fish consumption (1-2 fish servings per week) on RA development compared to never/rare fish intake. This protective effect of fish intake on RA risk was more pronounced in NHSII than NHS, but was not statistically significant in meta-analysis of both cohorts. Cohort differences, such as age, secular diet trends, and age at RA onset, may explain these results or they may be due to chance. Further studies of nutritional components of fish, such as omega-3 fatty acids, may clarify the effect of fish intake on RA risk.

Table. Hazard ratios for RA development by categories of cumulative average updated fish intake serving frequency in NHS (n = 86,135), NHSII (n = 92,984), and meta-analysis of both cohorts.
	Never to <1 time/month	1 to 3 times/month	1 to <2 times/week	2 to <3 times/week	≥3 times/week	P for trend
NHS
No. of cases	25	57	210	180	180
Person-years	76,884	217,399	744,346	589,110	558,495
Age-adjusted HR (95% CI)*	1.0 (ref)	0.77 (0.48-1.25)	0.86 (0.57-1.31)	0.93 (0.61-1.43)	0.96 (0.63-1.47)	0.34
Multivariable HR (95% CI)**	1.0 (ref)	0.77 (0.47-1.24)	0.86 (0.57-1.32)	0.94 (0.62-1.44)	0.95 (0.62-1.47)	0.36
NHSII
No. of cases	29	88	107	56	42
Person-years	125,680	453,864	663,142	242,353	225,015
Age-adjusted HR (95% CI)*	1.0 (ref)	0.78 (0.51-1.18)	0.63 (0.42-0.96)	0.81 (0.51-1.28)	0.68 (0.42-1.11)	0.66
Multivariable HR (95% CI)**	1.0 (ref)	0.76 (0.50-1.15)	0.66 (0.44-1.01)	0.86 (0.55-1.37)	0.73 (0.44-1.19)	0.91
Meta-analysis***
Age-adjusted HR (95% CI)*	1.0 (ref)	0.78 (0.57-1.07)	0.74 (0.55-1.00)	0.87 (0.64-1.19)	0.83 (0.59-1.15)	0.60
Multivariable HR (95% CI)**	1.0 (ref)	0.78 (0.57-1.07)	0.76 (0.56-1.02)	0.91 (0.66-1.24)	0.85 (0.61-1.17)	0.49
Adjusted for age, questionnaire period, and total energy intake (continuous). Adjusted for age, questionnaire period, total energy intake (continuous), US region (New England, Mid-Atlantic, Southeast, Midwest, West), median household income (<$40k, $40k to <$50k, $50k to <$65k, $65k to <$80k, ≥$80k), cigarette smoking pack-years (never to 10, >10), body mass index (underweight/normal, overweight, obese), alcohol intake (never to <5, 5 to <10, ≥10 g/day), and sugar-sweetened soda intake (<1 serving per month, 1-4 per month, 2-6 per week, ≥1 per day). **Meta-analysis performed using DerSimonian and Laird random effects model.

Disclosure:

J. A. Sparks,
None;

S. C. Chang,
None;

B. Lu,
None;

S. Malspeis,
None;

K. H. Costenbader,
None;

E. W. Karlson,
None.

« Back to 2014 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/fish-consumption-and-risk-of-rheumatoid-arthritis-among-women-in-large-prospective-cohorts/