Effect of vitamin D and/or omega-3 fatty acid supplementation on stroke outcomes: a randomized trial

Abstract

Background:

Among stroke patients, low serum 25-hydroxyvitamin D (25[OH]D) predicts poor outcomes. In mice, higher omega-3 (n-3) fatty acid intake diminishes brain damage after stroke. However, few studies have examined effects of supplementation with vitamin D or n-3 prior to stroke on post-stroke outcomes.

Methods:

We used data from the VITamin D and OmegA-3 TriaL which randomized middle-aged and older men and women without cardiovascular disease to vitamin D₃ (2000 IU/day) and/or marine n-3 fatty acids (1 g/day) and followed them for incident stroke events. Individuals experiencing a non-fatal stroke were mailed questionnaires assessing functional limitations (the physical performance scale adapted from Nagi) and physical disability (the modified Katz Activities of Daily Living and Rosow-Breslau Functional Health scales). We used logistic regression to analyze associations between randomized treatment and limitations on each scale.

Results:

290 individuals experienced their first stroke during the trial. 197 stroke survivors completed the stroke outcomes questionnaire a median of 1.4 years after diagnosis. We observed no associations between randomized treatment to vitamin D and functional limitations (odds ratio [OR]=1.01, 95% confidence interval [CI]: 0.52, 1.97) or physical disability (Rosow-Breslau - OR=0.92, 95% CI: 0.50, 1.67; Katz - OR=1.03, 95% CI: 0.31, 3.42). Those randomized to n-3 had a non-significantly lower risk of functional limitations (OR=0.55, 95% CI: 0.28, 1.09) and physical disability (Rosow-Breslau - OR=0.56, 95% CI: 0.31, 1.02; Katz - OR=0.32, 95% CI: 0.50, 1.67).

Conclusion:

Vitamin D or omega-3 fatty acids supplementation prior to stroke did not result in significantly improved post-stroke outcomes.

Stroke is a leading cause of morbidity and mortality worldwide. To reduce the morbidity burden of stroke, research has focused on finding factors that may reduce the incidence of stroke or improve post-stroke outcomes. Recently, there has been growing interest in the roles that both vitamin D and omega-3 (n-3) fatty acids may play in the prevention of cardiovascular disease events.

The recently completed VITamin D and OmegA-3 TriaL (VITAL)(1) observed no effect of vitamin D or omega-3 fatty acid supplementation on the primary prevention of total stroke. (2, 3) Although supplementation with vitamin D or n-3 fatty acids did not reduce stroke incidence, these agents may affect stroke outcomes. Previous studies among stroke patients have shown that lower levels of serum 25-hydroxyvitamin D (25[OH]D) at hospital admission are associated with greater stroke severity, larger infarct size, poorer functional outcomes, and higher risk of death.(4–8) However, these studies have not tested whether vitamin D supplementation in initially healthy populations affects outcomes post-stroke. Vitamin D may affect stroke outcomes through its neuroprotective properties(9) and by promoting neuroplastic changes.(4)

Data also suggest a role of n-3 fatty acids in the primary prevention of stroke. A meta-analysis of observational studies found that eating fish was associated with lower risk of stroke among healthy individuals(10), but less is known about the effect of n-3 fatty acids on stroke outcomes. Animal studies suggest that dietary supplementation with n-3 fatty acids, particularly docosahexaenoic acid (DHA), may offer protection against the immune response or diminish brain damage in ischemic stroke and could potentially result in improved stroke outcomes.(11–13) However, there are limited data examining the effects of n-3 fatty acid supplementation on stroke outcomes among humans.

Using data from the recently completed VITamin D and OmegA-3 TriaL (VITAL)(1), we tested whether vitamin D or marine n-3 fatty acid supplementation reduces the risk of functional limitations and physical disability after stroke.

Materials and Methods

The data that support the findings of this study are available from the corresponding author upon reasonable request. VITAL is a randomized, double-blind, placebo-controlled 2×2 factorial trial designed to test the effects of vitamin D₃ (cholecalciferol, 2,000 IU/day or 50 mcg) and marine n-3 fatty acid supplements (1 g/day fish-oil capsule containing 840 mg of n-3 fatty acids including eicosapentaenoic acid [EPA, 460 mg] + docosahexaenoic acid [DHA, 380 mg]) in the primary prevention of cardiovascular disease and cancer in a multiethnic population of 25,871 individuals.(1) The study enrolled men aged ≥50 years and women aged ≥55 years living in the United States without a history of cancer (except non-melanoma skin cancer), myocardial infarction, stroke, transient ischemic attack, or coronary revascularization. Participants were sent study pills in monthly calendar packs containing either active and/or placebo pills via mail and received the study questionnaires described below via mail or email. Prior to randomization, approximately 65.5% of participants provided a blood sample. Randomized treatment ended on December 31, 2017 with a median of 5.3 years of follow-up. The main results of the trial are reported elsewhere.(2, 3) VITAL was approved by the Institutional Review Board of Partners Healthcare/Brigham and Women’s Hospital, and the study agents have received Investigational New Drug Approval from the U.S. Food and Drug Administration. This study was registered at clinicaltrials.gov (NCT04070833).

At baseline of VITAL, all participants completed a comprehensive questionnaire which collected information on demographic, medical history, lifestyle, dietary habits such as fish consumption, and clinical risk factors. The questionnaire assessed baseline physical function and limitations in: vigorous or moderate activities; lifting or carrying groceries; climbing stairs; bending, kneeling, or stopping; walking >1 mile; walking several blocks; or walking one block. Respondents were asked if they received help with feeding themselves, dressing, getting in and out of bed, and taking a bath or shower. If the participant indicated experiencing a limitation in these items or needing help with these activities, they were categorized as having a functional limitation at baseline.

At 6 months and yearly thereafter, participants received follow-up questionnaires that asked about new medical diagnoses, including incident stroke events. If a participant reported a stroke event, we asked for permission to obtain copies of their medical records from this event. An Endpoints Committee of physicians, including a vascular neurologist, blinded to treatment assignment reviewed the medical records from self-reported stroke events to confirm or disconfirm the case according to the Trial of Org 10172 in Acute Stroke Treatment (TOAST) criteria and categorized the stroke events as ischemic or hemorrhagic subtypes.(14) Only individuals whose stroke was later confirmed by the Endpoints Committee are included in our analyses. The annual questionnaires also asked participants about adherence to the randomized trial intervention. Participants who reported taking at least two-thirds of the trial capsules on the annual questionnaire prior to completing the stroke outcomes questionnaire were considered to have adhered to the trial regimen.

Prior to the end of the randomized treatment phase of the trial in December 2017, we received additional funding in August 2016 to assess among stroke survivors a broad assessment of functional limitations and physical disability. For stroke events which were self-reported and confirmed before August 2016, we mailed a questionnaire (referred to as the “stroke outcomes questionnaire”) to assess current functional limitations and physical disability. For stroke events which were self-reported after August 2016, we mailed the same questionnaire at the time of the self-report.

Functional limitations were assessed through the physical performance scale adapted from Nagi(15), and physical disability with the modified Katz Activities of Daily Living (ADL) Scale(16) and the Rosow-Breslau Functional Health Scale.(17) These measures have high test-retest reliability.(18–23) The adapted Nagi scale asks respondents if they experience limitations in: pulling or pushing large objects; stooping, crouching, or kneeling; reaching or extending arms above shoulder level; reaching or extending arms below shoulder level; writing or handling or fingering small objects; standing in one place for long periods; and sitting for an hour. Possible response options were: no, not limited; yes, limited a little; yes, limited a lot; and don’t do under doctor’s orders or unable to do. The Rosow-Breslau Functional Health Scale asks if the respondent needs help with the following gross mobility tasks: walking 0.5 mile; walking up and down stairs to the second floor; and doing heavy work around the house. Possible response options were: by myself without help; with some help; and completely unable to do this. The modified Katz ADL scale asks about needing help with: bathing; dressing; eating; getting in and out of a chair; and walking 50 yards. Possible response options were: by myself without help; with some help; and completely unable to do this. For each scale, first, we dichotomized responses to each item in a particular scale as unlimited (no limitation or no help needed) versus limited (any amount of limitation or help needed). Then, each scale was dichotomized into unlimited (no limitation in any of the items) versus limited (limitation in at least one item).(24)

Statistical Analysis

We used logistic regression to examine the effects of randomized treatment assignments in VITAL on functional limitations and physical disability using separate models for each outcome as defined above. Only individuals with a completed stroke outcomes questionnaire contributed to these analyses. We performed analyses for total stroke as well as for the 172 ischemic strokes only. We performed a sensitivity analysis in which we restricted to those individuals who did not report functional limitations at baseline in VITAL. We also performed a sensitivity analysis restricting to those who were adherent to their randomized treatment assignment.

We performed a sensitivity analysis in which we used inverse probability of treatment weights to adjust for potential differences between treatment groups since our analysis conditions on experiencing a stroke event. We first developed a propensity score for randomized treatment assignment at the time of stroke including information on age at randomization, sex, body mass index, physical activity, statin use, alcohol consumption, history of hypertension, history of high cholesterol, race, and smoking status to predict treatment assignment. We then used weighted logistic regression to estimate the association between randomized treatment and our outcomes. We performed an additional sensitivity analysis in which fatal stroke events (N=42) were included.

We explored whether age (<66.7 years versus ≥66.7 years, the median age of the full VITAL cohort at baseline), sex (men versus women), or race (white versus non-white) modified the effect of vitamin D or n-3 fatty acids by running models stratified by each factor. For vitamin D supplementation, we ran models stratified by baseline 25-hydroxy-D levels (<31 ng/ml versus ≥31 ng/ml) and body mass index (<25, 25 to <30, and ≥30 kg/m²). For n-3 fatty acid supplementation, we ran models stratified by baseline fish consumption (<1.5 servings/week versus ≥1.5 servings/week). We formally tested for interaction by including an interaction term between randomized treatment assignment and each factor in separate models.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Results

During a median of 5.3 years of follow-up, 290 first stroke events occurred in VITAL (227 ischemic strokes, 44 hemorrhagic strokes, and 19 unknown subtype). Forty-two strokes were fatal events and an additional 11 deaths occurred before we began mailing the stroke outcomes questionnaires and therefore these participants could not complete the stroke outcomes questionnaire. Of the 237 individuals who were eligible to receive the stroke outcomes questionnaire, 199 individuals returned the questionnaire and 197 individuals completed the questions assessing functional limitations and physical disability (79.4% of non-fatal stroke events). The median time from stroke diagnosis to questionnaire return was 1.6 years (interquartile range: 0.6 to 2.5 years). Of the 197 individuals who completed the questionnaire, 123 individuals had a stroke event which was self-reported and confirmed prior to August 2016. The median time from stroke diagnosis to questionnaire return was 2.3 years for these individuals. Among the 74 individuals whose stroke events were self-reported and confirmed after August 2016, the median time from stroke diagnosis to questionnaire return was 0.6 years.

Table 1 presents the baseline characteristics of the individuals who returned the stroke outcomes questionnaire by randomized treatment assignment. The groups were similar at baseline, although those on active vitamin D supplements were slightly more likely to be smokers at baseline compared with those on placebo vitamin D supplements.

Table 1.

Baseline characteristics of individuals who experienced a stroke and completed the questionnaire assessing functional limitations and physical disability.

	Placebo Vitamin D (N=104)	Active Vitamin D (N=93)	p-value	Placebo n-3 (N=98)	Active n-3 (N=99)	p-value
Age at randomization, mean (SD)	71.4 (6.4)	70.7 (7.5)	0.45	71.5 (6.9)	70.6 (7.0)	0.35
Age at stroke, mean (SD)	74.5 (6.7)	73.8 (7.7)	0.47	74.7 (7.2)	73.6 (7.1)	0.30
Female sex, N (%)	53 (51.0)	46 (49.5)	0.83	47 (48.0)	52 (52.5)	0.52
Race, N (%)			0.41			0.30
White	85 (81.7)	81 (87.1)		86 (87.8)	80 (80.8)
Black	15 (14.4)	11 (11.8)		11 (11.2)	15 (15.2)
Other	4 (3.8)	0		1 (1.0)	3 (3.0)
BMI (mean, SD)	27.8 (5.9)	28.1 (5.6)	0.71	28.1 (5.5)	27.8 (6.0)	0.75
Smoking status, N (%)			0.03			0.18
Never	47 (45.2)	44 (47.3)		39 (39.8)	52 (52.5)
Past	52 (50.0)	38 (40.9)		51 (52.0)	39 (39.4)
Current	2 (1.9)	10 (10.8)		6 (6.1)	6 (6.1)
Alcohol consumption, N (%)			0.59			0.88
Never	35 (33.7)	28 (30.1)		33 (33.7)	30 (30.3)
<1/week	12 (11.5)	7 (7.5)		8 (8.2)	11 (11.1)
1–6/week	34 (32.7)	38 (40.9)		35 (35.7)	37 (37.4)
Daily	22 (21.2)	19 (20.4)		21 (21.4)	20 (20.2)
Physical activity, N(%)			0.75			0.95
<200 kcal/week	13 (12.5)	11 (11.8)		13 (13.3)	11 (11.1)
200- <600 kcal/week	15 (14.4)	19 (20.4)		17 (17.4)	17 (17.2)
600- <1500 kcal/week	28 (26.9)	24 (25.8)		26 (26.5)	26 (26.3)
≥1500 kcal/week	46 (44.2)	38 (40.9)		40 (40.8)	44 (44.4)
Total fish and seafood <1.5 servings/week, N (%)	53 (51.0)	51 (54.8)	0.58	52 (53.1)	52 (52.5)	0.94
Baseline 25(OH)D level <31 ng/ml, N(%)	32 (30.8)	31 (33.3)	0.98	27 (27.6)	36 (36.4)	0.13
History of hypertension, N (%)	74 (71.2)	56 (60.2)	0.11	66 (67.4)	64 (64.7)	0.69
Statin use, N (%)	40 (38.5)	35 (37.6)	0.78	40 (40.8)	35 (35.4)	0.39
History of high cholesterol, N (%)	43 (41.4)	42 (45.2)	0.63	43 (43.9)	42 (42.2)	0.89
History of diabetes, N(%)	23 (22.1)	15 (16.1)	0.30	20 (20.4)	18 (18.2)	0.72
Time in years from stroke to receipt of questionnaire, median (IQR)	1.4 (0.6, 2.4)	1.5 (0.5, 2.7)	0.78	1.4 (0.6, 2.5)	1.4 (0.4, 2.7)	0.88

Numbers may not add to 100% due to missing data

We observed no associations between randomized treatment to vitamin D supplementation and either functional limitations or physical disability among the stroke survivors (Table 2). Those randomized to n-3 had lower risk of functional limitations (OR=0.55, 95% CI: 0.28, 1.09) and physical disability (OR=0.56, 0.31, 1.02 for the Rosow-Breslau scale and OR=0.32, 95% CI: 0.08, 1.24 for the Katz ADL scale) although these effects were not statistically significant (Table 2). Results from analyses restricting to the 189 individuals who did not report functional limitations at baseline were similar. Due to the low number of hemorrhagic strokes (N=24), we were unable to perform analyses among individuals with hemorrhagic stroke. Among individuals with ischemic stroke (N=172), the effects of vitamin D and n-3 fatty acid supplementation on stroke outcomes were similar to the total stroke results (supplemental table 1). A sensitivity analysis restricting to the 174 individuals who adhered to both randomized treatment assignments showed similar findings to the main analysis (results not shown). Results from our sensitivity analysis using inverse probability of treatment weights to adjust for potential differences between treatment groups were similar to the primary analyses (supplemental table 2). This was expected given that the parent VITAL study observed no effect of vitamin D (hazard ratio=0.95, 95% CI: 0.76, 1.20) or n-3 fatty acids (hazard ratio=1.04, 95%CI: 0.83, 1.31) supplementation on the primary prevention of total stroke. When we included fatal stroke events in the same category as those reporting limitations in each scale, we observed somewhat attenuated effect estimates for n-3 (Supplemental Table 3).

Table 2.

Associations between randomized treatment assignments and functional limitations or physical disability (N=197)

	Nagi Scale		Rosow-Breslau Scale		Katz ADL scale
	n*	OR (95% CI)	n*	OR (95% CI)	n*	OR (95% CI)
Placebo Vitamin D	27	Ref	53	Ref	6	Ref
Active Vitamin D	24	1.01 (0.52, 1.97)	45	0.92 (0.50, 1.67)	6	1.03 (0.31, 3.42)
Placebo Omega-3	31	Ref	56	Ref	9	Ref
Active Omega-3	20	0.55 (0.28, 1.09)	42	0.56 (0.31, 1.02)	3	0.32 (0.08, 1.24)

^*n=number of individuals reporting limitation or disability in that scale

Our stratified analyses are shown in Supplemental Tables 4 and 5. Vitamin D was associated with non-significant decreases in the risk of functional limitations and physical activity among individuals aged 66.7 years or older, among white individuals, and among those with BMI <25 kg/m². No consistent effects were seen for sex or baseline 25(OH)D levels. For n-3 fatty acids, we observed a non-significant decrease in risk of functional limitations and physical disability across most strata.

Discussion

Using data from the VITAL study, we observed that randomized supplementation with vitamin D or n-3 fatty acids prior to stroke did not result in significant improvements in post-stroke outcomes among middle-age and older individuals. However, a trend towards improvement in outcomes was suggested for n-3 fatty acids.

No prior studies have examined the association of vitamin D supplementation before stroke and effect on stroke outcomes. Observational studies among individuals with stroke have observed that lower levels of serum 25[OH]D at the time of stroke are associated with greater stroke severity, larger infarct size, and poor post-stroke outcomes.(4–8) However, there are several important limitations to these studies. First, due to their observational nature, they may suffer from uncontrolled confounding, particularly by factors like body mass index or baseline health status. Second, they do not address the question of whether supplementation with vitamin D may improve stroke outcomes. Some studies have suggested that vitamin D supplementation after stroke may improve post-stroke outcomes, particularly among those with low vitamin D levels. For example, two small unblinded studies among individuals with vitamin D insufficiency or deficiency who experienced an ischemic stroke suggested that vitamin D supplementation after stroke may result in larger improvements in the Scandinavian Stroke Scale(25) or lower modified Rankin Scale scores(26) than individuals assigned to no supplementation. A blinded trial among 132 individuals with hemiplegia due to ischemic stroke observed that an intramuscular injection of 300,000 IU vitamin D resulted in increased activity levels and accelerated balance recovery but did not affect ambulation or motor recovery compared with a saline injection.(27) In contrast, a blinded trial among 100 hospitalized patients undergoing rehabilitation after acute stroke observed no significance differences in the gain in Barthel Index score between those randomized to receive 2000 IU/day of vitamin D3 and those randomized to placebo.(28)

The current study addressed whether pre-stroke supplementation with vitamin D may improve stroke outcomes. Most individuals in this cohort were not vitamin D deficient at baseline (mean baseline 25[OH]D was 30.8 ng/mL),(1) but results did not differ between those above and below the median. We administered vitamin D orally and did not inject mega-doses intramuscularly as was done in some of the prior studies which observed significant effects. Unfortunately, we were unable to measure changes in serum vitamin D levels over time among the stroke survivors. However, prior analysis among a subset of VITAL participants demonstrated that compared with baseline, mean 1-year 25(OH)D increased by 40% (in absolute terms, by 11.9 ng/mL) in the active group and decreased by <2% (0.7 ng/mL) in the placebo group.(29)

N-3 fatty acids have been hypothesized to have protective effects on stroke. Several animal studies have suggested that treatment with n-3 fatty acids before and after induced cerebral ischemia may improve outcomes and result in improved neurogenesis, angiogenesis, and oligodendrogenesis.(11–13) Although our results did not achieve statistical significance, we observed some evidence that pre-stroke supplementation with n-3 fatty acids was associated with reductions in the risk of functional limitations and physical disability a median of 1.4 years after stroke. In contrast, most prior studies among humans have examined only whether supplementation with n-3 fatty acids after stroke influence post-stroke outcomes. The Supplementation en Folates et Omega-3 trial found no effect of post-stroke supplementation with 600 mg eicosapentaenoic and docosahexaenoic acids (2:1 ratio) on health-related quality of life measured by the Medical Outcomes Study 36-Item Short Form Health Survey (SF-36).(30) Two small studies observed non-significant protective effects of n-3 fatty acid supplementation on functional outcomes as measured by the GOSE scale, the Barthel Index, and the Rivermead Mobility Index,(31, 32) although the quality of the evidence from these studies was judged to be of very low quality in a recent Cochrane review.(33) In addition, these studies only assessed post-stroke supplementation with n-3 fatty acids and could not assess how supplementation prior to the event may influence outcomes.

Prior randomized studies of n-3 fatty acids on stroke incidence have not reported on stroke outcomes. For example, the recently completed REDUCE-IT trial did observe a significant effect of 2 g of icosapent ethyl, a highly purified and stable EPA ethyl ester, twice daily on the risk of ischemic stroke (HR=0.64; 95% CI: 0.49, 0.85) among individuals with a history of cardiovascular disease or with diabetes and other vascular risk factors.(34) However, this study did not provide additional information on short or long-term stroke outcomes. The differences in study population eligibility as well as the formulation and dose of the interventions tested makes it difficult to determine if the results seen in the present study would have been observed in REDUCE-IT.

In conclusion, although we did not find significant effects of pre-stroke supplementation with vitamin D or n-3 fatty acids on post-stroke outcomes, further exploration of the effect of higher doses of n-3 fatty acids on stroke outcomes may be warranted given the recent evidence on the effects of high dose icosapent ethyl on stroke incidence and the trends towards improved post-stroke outcomes observed in the current study. Additionally, it is unknown if beginning supplementation with high dose n-3 fatty acids after the stroke event has already occurred could improve outcomes.