- Email: [email protected]
The prognostic utility of dihomo-gamma-linolenic acid (DGLA) in patients with acute coronary heart disease
IJCA-25496; No of Pages 6 International Journal of Cardiology xxx (2017) xxx–xxx
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
The prognostic utility of dihomo-gamma-linolenic acid (DGLA) in patients with acute coronary heart disease☆ Dennis W.T. Nilsen a,c,⁎, Hildegunn Aarsetoey b, Volker Pönitz a, Trygve Brugger-Andersen a, Harry Staines d, William S. Harris e, Heidi Grundt b,c a
Division of Cardiology, Stavanger University Hospital, 4068 Stavanger, Norway Department of Medicine, Stavanger University Hospital, 4068 Stavanger, Norway Department of Clinical Science, University of Bergen, 5020 Bergen, Norway d Sigma Statistical Services, KY16 0BJ Balmullo, UK e Sanford School of Medicine, University of South Dakota, OmegaQuant Analytics, LLC, Sioux Falls, SD, United States b c
a r t i c l e
i n f o
Article history: Received 12 April 2017 Received in revised form 18 August 2017 Accepted 25 September 2017 Available online xxxx Keywords: Dihomo-gamma-linolenic acid (DGLA) Acute coronary syndrome (ACS) Prognosis Total mortality Myocardial infarction (MI) Stroke
a b s t r a c t Background: We previously investigated the prognostic utility of red blood cell (RBC) n-3 fatty acids (FAs) in survivors of an acute myocardial syndrome (ACS) but found no relationship with all-cause mortality and cardiac death or MI after two years. Here we extend our follow-up to 7 years, focusing on the potential predictive power of RBC n-6 FAs. Methods: We included 398 ACS patients presenting with increased troponin-T (TnT) levels for whom baseline RBC FA data were available. Cox regression analysis was used to relate the risk of future events to RBC n-6 FA levels, both continuously and by quartile. Results: At 7-year follow-up, 183 (46.0%) had died, 128 (32.2%) had experienced another MI and 24 (6.0%) had had a stroke. Death or MI occurred in 227 patients (57.0%); and death, MI or stroke in 235 patients (59.0%). In a multivariable Cox regression model for total death, the hazard ratio (HR) in the highest as compared to the lowest quartile of dihomo-γ-linolenic acid (DGLA) was 0.55 [95% confidence interval (CI), 0.35–0.88, p = 0.012, for death or MI [HR 0.62 (95% CI, 0.41–0.94), p = 0.025], and for the fully combined endpoint [HR 0.57 (95% CI, 0.38–0.86), p = 0.006]. Similar results were found in the per 1-SD analysis. No other RBC n-6 FAs significantly predicted these outcomes in multivariable models. Conclusion: RBC DGLA levels had significant independent prognostic value in post-ACS patients. These findings need confirmation, and the possible biochemical pathways by which higher DGLA membrane levels may be cardioprotective should be explored. © 2017 Published by Elsevier Ireland Ltd.
1. Introduction In a patient population from western Norway presenting with an acute coronary syndrome (ACS), red blood cell (RBC) n-3 fatty acid (FA) levels [i.e., the omega-3 index, which is the sum of eicosapentaenoic and docosahexaenoic acids], were unassociated with all-cause mortality and cardiac death or myocardial infarction (MI) after 2-years of follow-up [1]. We have now extended our analysis to 7-year follow-up and focused on the n-6 FAs in RBC membranes [linoleic acid (LA), dihomo-gamma linolenic acid (DGLA) and arachidonic acid (AA)]. Past studies have reported that higher intakes of LA [2], and higher blood levels of LA [3] and AA [4] are cardioprotective; ☆ The patient cohort is registered in ClinicalTrials.gov identifier: NCT00521976. ⁎ Corresponding author at: Department of Clinical Science, University of Bergen, 5020 Bergen, Norway. E-mail address: [email protected] (D.W.T. Nilsen).
and DGLA has been reported to have anti-inflammatory and antiproliferative properties, in addition to being a precursor of AA [5]. As for alpha linoleic acid (ALA) in the n-3 FA series, LA is considered to be an essential polyunsaturated FA (PUFA). LA (18:2n-6) can be desaturated to gamma-linolenic acid (GLA; 18:3n-6) which in turn is elongated into dihomo-gamma-linolenic acid (DGLA; 20:3n6), from which arachidonic acid (AA; 20:4n-6) may be synthesized upon further desaturation [6]. LA which is primarily derived from vegetable oils, such as corn, sunflower, safflower and soy, accounts for N85% to 90% of the dietary n-6 PUFAs, and the average daily intake in the adult population is estimated to approximately 15 g, representing between 6.5 and 7% of energy [7]. In tracer studies, the extent of conversion of LA to AA is very low, no more than about 0.2% [8]; approximately 0.15 g of preformed AA is ingested per day from meat, eggs and some fish [7]. GLA and DGLA are consumed in even smaller amounts than AA and their levels in RBC membranes are primarily determined by metabolism, not diet.
https://doi.org/10.1016/j.ijcard.2017.09.202 0167-5273/© 2017 Published by Elsevier Ireland Ltd.
Please cite this article as: D.W.T. Nilsen, et al., The prognostic utility of dihomo-gamma-linolenic acid (DGLA) in patients with acute coronary heart disease, Int J Cardiol (2017), https://doi.org/10.1016/j.ijcard.2017.09.202
2
D.W.T. Nilsen et al. / International Journal of Cardiology xxx (2017) xxx–xxx
Whereas the metabolic products of AA in the 2-series prostaglandins (PGE2) are mainly considered to promote inflammation and thrombosis, those of DGLA are generally considered to contain anti-inflammatory, antiproliferative and antithrombotic properties [6,9,10], and are formed by the action of cyclooxygenase (COX) 1 & 2 converting DGLA into prostaglandins of the 1-series (PGE-1) and into 15-(S)-hydroxy8,11,13-eicosatrenoic acid (15-HETrE) by the action of 15-lipoxygenase (LOX) [5]. There have been smaller trials investigating the biochemical effects of GLA supplementation which increases the levels of DGLA [11]. In supplementation studies conducted in humans [11,12], GLA is rapidly converted to DGLA. To what extent downstream metabolic products of DGLA are formed under normal conditions, is not known, and the relevance of this pathway to improved health has not been sufficiently investigated. The impact of the n-6 series on variables related to biochemistry and blood pressure in four primary prevention studies including 660 individuals was insufficient to claim an effect [13]. Randomized, controlled studies evaluating coronary heart disease (CHD) events following replacement of saturated fatty acids with PUFAs are burdened with design limitations, such as the lack of blinding and the ambiguity related to whether the effects are due to the removal of one FA class or the addition of another [7]. However, a meta-analysis [14] including 6 such trials indicated that PUFAs may lower the risk of CHD events by 24% in established CHD. We have evaluated the prognostic utility at 7-year follow-up of RBC levels of three individual FAs in the n-6 series: LA, DGLA and AA in patients admitted with chest-pain and a positive ACS biomarker. We hypothesize that one or more of these FAs may predict cardiovascular outcomes in patients hospitalized with an acute coronary syndrome (ACS). 2. Methods 2.1. Study population Our study population belonged to a prospective, observational study termed “Risk Markers in the Acute Coronary Syndrome” (Ref. ClinicalTrials.gov identifier: NCT00521976), designed to identify early risk markers of all-cause mortality and future coronary events following hospitalization with chest pain. A total of 871 patients were recruited at one hospital located in Stavanger, Norway, from November 2002 until October 2003 [15,16]. Of the 871 included patients, 471 had an ACS as defined by a troponin-T (TnT) value ≥0.02 μg/L, of which RBC samples were available in 398 patients. The prognostic utility of several RBC FAs, measured in per cent of total fatty acids, was evaluated in this population. The prognostic utility of brain natriuretic peptide (BNP) and high-sensitivity CRP (hsCRP), respectively, has previously been reported for the main population at 2-year [17] and at 7-year follow-up [18], and that of the omega-3 index has been investigated at 2-year follow-up [1]. The study was approved by the Regional Board of Research Ethics and the Norwegian Health Authorities and conducted in accordance with the Helsinki Declaration of 1975, as revised in 1983. 2.2. Study design Detailed clinical background information was collected in a case report form and reported at baseline (Table 1). During hospitalization 50% of the patients were referred to coronary angiography [1], of which 188 patients were revascularized. Percutaneous coronary intervention (PCI) was performed in 160 subjects and 28 subjects underwent coronary artery bypass grafting (CABG). At discharge all patients received optimal medical treatment according to current recommendations, aiming at N90% on both aspirin and a statin. Follow-up information was collected over the next 2 years by telephone interviews at 30 days and 6, 12, and 24 months. Extended follow-up was performed at 7-year followup. For patients not capable of responding or in case of out-of-hospital death, family members, general practitioners, and nursing staff were contacted for information. Hospital records were searched for detection and verification of reported events. No patients were lost to follow-up. The primary outcome measure was all-cause mortality. Secondary outcomes included recurrent myocardial infarction (MI) and ischemic stroke. Definition of a recurrent MI was based on symptoms of coronary ischemia associated with a typical rise and fall of TnT, exceeding 0.05 μg/L. Only the first recurrent MI or stroke was recorded during the observation period, and for non-survivors, death overruled either of these events.
2.3. Laboratory methods Venous blood samples were collected immediately following admission by direct venipuncture of an antecubital vein. Analyses of TnT, hematological variables, creatinine, glucose, and lipids were performed in serum as part of the hospital's routine. TnT was repeated on average 8 h following admission. Additional TnT measurements were performed with increasing values. RBC for analysis of total fatty acids were immediately prepared from citrated blood and stored at − 70 °C. At the end of the inclusion period samples were shipped dried on filter papers in sealed plastic bags on cold packs to Kansas City, where they again were kept for a short period at − 80 °C prior to measurements with flame ionization GC (GC9A, Shimadazu), as previously reported [19]. Fatty acids were reported as weight percent of total FA. The CV and stability of the FAs have previously been reported [20]. The FA composition in dried RBC spots reflects the FA composition of freshly thawed samples, and tests have been performed in relation to storage time, transport, and temperature [19]. Plasma brain natriuretic peptide (BNP) (Microparticle Enzyme Immunoassay, Abbott AxSYM) was measured in EDTA plasma, and serum was used for measurement of highsensitivity C-reactive protein (hsCRP) [Tina-quant C-reactive protein (latex) high sensitive assay, Roche Diagnostics], respectively [17]. Laboratory personnel had no access to clinical information.
2.4. Statistical analyses Cox regression analyses for the evaluation of follow-up events was applied to continuous values and quartiles of predictor variables at inclusion. A set of potentially confounding variables was included in the adjusted models in a forward stepwise procedure: age; sex; history of angina pectoris and cardiac failure; previous MI, percutaneous coronary intervention (PCI), or coronary bypass artery grafting (CABG); history of hypertension and diabetes mellitus (both insulin-dependent and noninsulin-dependent); current smoking; use of medication; baseline highsensitivity C-reactive protein (hsCRP), brain natriuretic peptide (BNP), creatinine, estimated glomerular filtration rate (eGFR), total cholesterol, HDL-cholesterol, triglycerides (TG), and hypercholesterolemia; peak TnT value; ST-elevation myocardial infarction (STEMI); non-ST-elevation myocardial infarction (NSTEMI). Missing values were excluded case wise. Body mass index (BMI) was also recorded, but not included in the model due to the number of missing values (n = 20). The hazard ratio (HR) for new cardiac events according to quartiles relative to the lowest quartile is presented with 95% CI. The chi-square test was applied when comparing proportions for qualitative variables among quartiles at baseline. Normally distributed quantitative variables are given as means ± SD and differences among quartiles at baseline evaluated through one way analysis of variance. In case of nonnormally distributed variables, the Kruskal-Wallis test was used. The results of nonparametric testing are given as median values with upper and lower quartiles. The area under the Receiver Operating Characteristic curve and asymptotic significance when tested for different to 0.5 were found. The statistical analyses were performed with the statistical package SPSS version 24.0. All tests were applied with a 2-sided significance level of 5%.
3. Results Of the 398 TnT-positive patients for whom RBC fatty acids were available, 65.1% were males, and the mean age (± SD) of the total population was 71.9 ± 13.0 years. At 2-year follow-up, 86 (21.6%) patients had died. Death or MI was recorded in 152 patients (38.2%). When including stroke in the combined endpoint, there were 162 (40.7%) events. DGLA was not found to be a significant predictor at 2 years, but a trend towards statistical significance was observed in relation to the combined endpoints consisting of total mortality or MI or stroke, when comparing Q4 to Q1 [HR 0.63 (95% CI, 0.39–1.02), p = 0.061]. As shown in Table 1, there was a linear decrease in age across the DGLA quartiles (p = 0.000) with patients in Q1, on average, 10 years older than those in Q4. Furthermore, there was a decrease in BNP, history of heart failure, NYHA class and use of diuretics from Q1 to Q4, whereas the opposite trend was demonstrated for peak TnT, STEMI and kidney function (with no patient on hemodialysis). By 7-year follow-up, 183 patients (46%) had died. Total mortality or MI was recorded in 227 (57%), and total mortality or MI or stroke in 235 patients (59%).
Please cite this article as: D.W.T. Nilsen, et al., The prognostic utility of dihomo-gamma-linolenic acid (DGLA) in patients with acute coronary heart disease, Int J Cardiol (2017), https://doi.org/10.1016/j.ijcard.2017.09.202
D.W.T. Nilsen et al. / International Journal of Cardiology xxx (2017) xxx–xxx
3
Table 1 Baseline characteristics according to quartiles of dihomo-gamma linolenic acid C20:3n-6 (DGLA), measured in per cent of total fatty acids in red blood cells. Variable
a
Median (25–75%)
Age, yr Male gender, % (n) Risk markers at baseline Tot. chol. (mmol/L)a HDL chol. (mmol/L)a Triglycerides (mmol/L)a Peak TNT (ng/mL)b hsCRP (mg/L)b BNP (pg/mL)b Risk factors Current smoking, % (n) Hypertension, % (n) IDDM, % (n) NIDDM, % (n) eGFR (ml/min/1.73 m2)a Hypercholesterolemia, % (n) History of coronary heart disease Angina pectoris, % (n) MI, % (n) CABG, % (n) PCI, % (n) History of heart failure CHF, % (n) NYHA 2–4, % (n) Admission data STEMI, % (n) NSTEMI, % (n) U-AP, % (n) Non-ACS, % (n) ASA, % (n) Clopidogrel, % (n) Statins % (n) Beta blockers, % (n) ACE-inhibitors, % (n) ARBs, % (n) Diuretics, % (n)
Total
Q1
Q2
Q3
Q4
1.04 (0.84–1.12)
1.30 (1.24–1.35)
1.54 (1.47–1.60)
1.89 (1.77–2.17)
p-Value
Cohorta
(n = 99)
(n = 100)
(n = 100)
(n = 99)
71.9 ± 13.0 65.1 (259)
76.5 ± 11.4 61.6 (61)
72.8 ± 11.8 64.0 (64)
71.5 ± 12.8 66.0 (66)
66.8 ± 14.2 68.7 (68)
0.000 0.758
5.21 ± 1.28 1.34 ± 0.44 1.66 ± 1.29 0.39 (0.09–1.98) 5.56 (2.24–16.78) 184 (53–460)
5.07 ± 1.47 1.41 ± 0.48 1.51 ± 1.24 0.22 (0.05–0.80) 6.7 (2.4–20.0) 352 (121–749)
5.14 ± 1.25 1.39 ± 0.48 1.59 ± 1.15 0.30 (0.09–1.16) 4.0 (1.7–14.4) 224 (54–558)
5.35 ± 1.07 1.36 ± 0.40 1.60 ± 0.88 0.69 (0.15–2.80) 6.1 (3.1–17.6) 137 (41–386)
5.26 ± 1.29 1.20 ± 0.35 1.93 ± 1.72 1.00 (0.16–3.27) 5.5 (2.0–13.5) 75 (35–224)
0.413 0.002 0.097 0.000 0.227 0.000
30.4 (121) 44.2 (176) 1.3 (5) 17.1 (68) 61.6 ± 23.5 44.5 (177)
24.2 (24) 41.4 (41) 2.0 (2) 16.2 (16) 53.8 ± 21.4 40.4 (40)
26.0 (26) 44.0 (44) 1.0 (1) 15.0 (15) 63.5 ± 27.2 50.0 (50)
37.0 (37) 52.0 (52) 1.0 (1) 20.0 (20) 62.0 ± 21.8 43.0 (43)
34.3 (34) 39.4 (39) 1.3 (1) 17.2 (17) 66.9 ± 21.6 44.4 (44)
0.140 0.295 0.892 0.809 0.001 0.575
42.0 (167) 32.7.(130) 8.8 (35) 7.8 (31)
47.5 (47) 37.4 (37) 10.1 (10) 7.1 (7)
45.0 (45) 40.0 (40) 13.0 (13) 8.0 (8)
41.0 (41) 30.0 (30) 6.0 (6) 11.0 (11)
34.3 (34) 23.2 (23) 6.1 (6) 5.1 (5)
0.0.260 0.051 0.230 0.467
30.9 (123) 27.6 (110)
40.4 (40) 38.4 (38)
36.0 (36) 27.0 (27)
23.0 (23) 22.0 (22)
24.2 (24) 23.2 (23)
0.016 0.040
27.4 (109) 54.8 (218) 16.8 (67) 1.0 (4) 33.4 (133) 1.8 (7) 30.7 (122) 32.7 (130) 19.1 (76) 17.8 (71) 34.2 (136)
12.1 (12) 61.6 (61) 25.3 (25) 1.0 (1) 38.4 (38) 1.0 (1) 30.3 (30) 22.2 (22) 19.2 (19) 20.2 (20) 46.5 (46)
25.0 (25) 58.0 (58) 14.0 (14) 3.0 (3) 33.0 (33) 4.0 (4) 33.0 (33) 44.0 (44) 26.0 (26) 15.0 (15) 40.0 (40)
32.0 (32) 51.0 (51) 17.0 (17) 0.0 (0) 35.0 (35) 2.0 (2) 29.0 (29) 35.0 (35) 15.0 (15) 22.0 (22) 24.0 (24)
40.4 (40) 48.5 (48) 11.1 (11) 0.0 (0) 27.3 (27) 0.0 (0) 30.3 (30) 29.3 (29) 16.2 (16) 14.1 (14) 26.3 (26)
0.000 0.217 0.048 0.111 0.408 0.169 0.940 0.009 0.193 0.387 0.001
HsCRP = high-sensitive C-reactive protein. BNP = B-type natriuretic peptide. IDDM = Insulin dependent diabetes mellitus. NIDDM = non-insulin dependent diabetes mellitus. eGFR = estimated glomerular filtration rate. CABG = coronary artery bypass graft. PCI = percutaneous coronary intervention. NYHA = New York Heart Association function classification of heart failure. ARBs = angiotensin II receptor betblockers. HDL = high density lipoprotein. TNT = troponin-T. MI = myocardial infarction. CHF = congestive heart failure. STEMI = ST-elevation myocardial infarction. NSTEMI = non-ST elevation myocardial infarction. U-AP = unstable angina pectoris. ACS = acute coronary syndrome. ASA = acetylsalicylic acid. ACE = angiotensinconverting-enzyme inhibitor. ARB = angiotensin receptor blocker. a Age, eGFR, total cholesterol, HDL cholesterol and triglycerides are given as mean (±Std. deviation). b Peak TnT, hsCRP and BNP are given as median (quartiles Q1–Q3).
3.1. Dihomo-gamma-linolenic acid (DGLA) and long-term (7-year) prognosis At 7-year follow-up the DGLA (%) levels were significantly lower among patients dying than in survivors; 1.33 ± 0.41 vs. 1.53 ± 0.44, p = 0.000. Kaplan–Meier survival curves for all-cause mortality according to quartiles of DGLA are presented in Fig. 1a. Receiver-operated characteristic (ROC) curve for all-cause mortality is shown in Fig. 1b. The area under the ROC for DGLA was 0.640 (p = 0.000). Results of the univariate analysis relating DGLA to prognosis are shown in Table 2. All p-values for the hazard ratios (HR) are statistically highly significant. DGLA quartiles were added to the model in a separate block after using a stepwise approach to identify significant predictor variables, as shown in Table 2. In the multivariable Cox regression model for total death within 7 years, the HR for DGLA in the highest quartile (Q4) as compared to the lowest (Q1) was 0.55 (95% CI, 0.35–0.88), p = 0.012, whereas applying continuous values of DGLA weakened the p-value, p = 0.055 (Table 2). The prognostic utility of DGLA for total mortality is mainly found in the extreme quartiles, when comparing the highest to the
other quartiles (p = 0.041), or the lowest against the other quartiles (p = 0.035). When applying continuous values of DGLA in the multivariable Cox regression model for total death or MI within 7-year follow-up, statistical significance was maintained [HR 0.69 (95% CI, 0.49–0.98), p = 0.037], which was also evident for total death or MI or stroke [HR 0.66 (95% CI, 0.47–0.93), p = 0.017], very similar to the results obtained in the quartile analysis (Table 2). 3.2. LA in relation to long-term prognosis In the univariate analysis, LA was found to be a highly significant predictor of all specified endpoints at 7-year follow-up, but lost its significance in the multivariable analysis. However, the p-value approached significance in the multivariable analysis for death, MI or stroke at 7 years; HR 0.70 95% CI (0.47–1.04), p = 0.075 (Table 3). 3.3. AA in relation to long-term prognosis The prognostic utility of AA was not significant for either the univariate or multivariable models.
Please cite this article as: D.W.T. Nilsen, et al., The prognostic utility of dihomo-gamma-linolenic acid (DGLA) in patients with acute coronary heart disease, Int J Cardiol (2017), https://doi.org/10.1016/j.ijcard.2017.09.202
4
D.W.T. Nilsen et al. / International Journal of Cardiology xxx (2017) xxx–xxx
Fig. 1. a. Kaplan–Meier plot of time to all-cause mortality during 7-year follow-up by DGLA quartiles. b. Receiver-operated characteristics curve for DGLA; area under the curve (AUC) 0.640 (p = 0.000).
3.4. Omega-3 index and long-term prognosis The omega-3 index was not identified as a predictor of risk for the study outcomes. In a multivariable Cox regression model for total death within 7 years, the HR for the omega-3 index in the highest quartile (Q4) as compared to the lowest (Q1) was 0.92 (95% CI, 0.59– 1.50), p = 0.74, for total death or MI the HR was 1.00 (95% CI, 0.67– 1.50), p = 1.00, and for total death or MI or stroke the HR was 1.10 (95% CI, 0.74–1.65), p = 0.63.
4. Discussion We have assessed the 7-year prognostic utility of several n-6 polyunsaturated FAs measured in RBCs from patients hospitalized with an ACS. At 2-years, there was a trend (p = 0.06, quartile 4 vs 1) for an inverse association between DGLA and the fully combined outcome metric.
After 7 years, DGLA was found to independently predict risk of total mortality, essentially of cardiac etiology. DGLA also predicted the combination of total mortality or MI, and that of the combination of total mortality or MI or stroke, respectively, after adjusting for potential confounding and/or mediating factors. These findings may reflect the trend seen for LA in the assessment of prognostic utility. Of interest, we noted a gradual increase in age and in BNP (heart failure), respectively, from upper through lower quartiles of DGLA. The opposite trend was observed for STEMI and peak TnT. Accordingly, DGLA levels seem to be clearly related to cardiac disease burden at baseline. The age difference from Q1 to Q4 of DGLA was 10 years, with the lowest levels of DGLA in Q1, which is in line with a previous report showing that LA, a precursor of DGLA, decreases with increasing age [21]. DGLA is only modestly converted into AA. It may in part be metabolized by COX1 to prostaglandin E1 (PGE1) and/or by 15-LOX into 15HETrE, both of which possess potent anti-inflammatory and antiproliferative properties in vitro in cells such as neutrophils, macrophages/
Table 2 Univariate and multivariable HRs (95% CI) for C20:3n-6 (DGLA) in relation to 1) total death, 2) total death or myocardial infarction (MI) and 3) total death or MI or stroke groups 7-year follow-up, comparing the highest to the lowest quartile of DGLA and continuous values of DGLA. All patients n = 398 C20:3n-6 (DGLA)
Total death Total death or MI Total death or MI or stroke
Quartiles univariatea
Quartiles multivariablea
Continuous values multivariablea
HR (95% CI) p-value
HR (95% CI) p-value
HR (95% CI) p-value
0.32 (0.20–0.50), 0.000 0.40 (0.27–0.59), 0.000 0.40 (0.27–0.58), 0.000
0.55 (0.35–0.88), 0.012 0.62 (0.41–0.94), 0.025 0.57 (0.38–0.85), 0.006
0.69 (0.47–1.01), 0.055 0.69 (0.49–0.98), 0.037 0.66 (0.47–0.93), 0.017
Age: 1.06 (1.04, 1.08), p = 0.000 CHF 2.08 (1.49, 2.91), p = 0.000 BNP (Q4 vs Q1) 2.24 (1.18, 4.23), p = 0.013 CHF: 2.07 (1.52, 2.82), p = 0.000 Age: 1.03 (1.01, 1.04), p = 0.000 BNP (Q4 vs Q1) 1.71 (1.03, 2.82), p = 0.037 CHF: 2.07 (1.53, 2.80), p = 0.000 Age: 1.03 (1.02, 1.05), p = 0.000 BNP (Q4 vs Q1) 1.71 (1.05, 2.81), p = 0.032
Age: 1.06 (1.04, 1.08), p = 0.000 CHF 1.99 (1.42, 2.77), p = 0.000 BNP (Q4 vs Q1) 2.56 (1.19, 4.26), p = 0.012 CHF: 1.99 (1.47, 2.71), p = 0.000 Age: 1.03 (1.02, 1.05), p = 0.000 BNP (Q4 vs Q1) 1.73 (1.05, 2.85), p = 0.032 CHF: 1.99 (1.47, 2.68), p = 0.000 Age: 1.03 (1.02, 1.05), p = 0.000 BNP (Q4 vs Q1) 1.75 (1.07, 2.85), p = 0.025
Predictor variables entered in the stepwise Cox model Total death Step 1 Step 2 Step 3 Total death or MI Step 1 Step 2 Step 3 Total death or MI or stroke Step 1 Step 2 Step 3
a HR (hazard ratio) [95% CI (Confidence Interval)]. Highest quartile (Q4) versus lowest quartile (Q1), and multivariable analysis based on continuous values. A set of potentially confounding variables was included in the adjusted models in a forward stepwise procedure: age; sex; history of angina pectoris and cardiac failure; previous MI, percutaneous coronary intervention (PCI), or coronary bypass artery grafting (CABG); history of hypertension and diabetes mellitus (both insulin-dependent and noninsulin-dependent); current smoking; use of medication; baseline high-sensitivity C-reactive protein (hsCRP), brain natriuretic peptide (BNP), creatinine, estimated glomerular filtration rate (eGFR), total cholesterol, HDLcholesterol, triglycerides (TG), and hypercholesterolemia; BMI; peak TnT value; ST-elevation myocardial infarction (STEMI); non-ST-elevation myocardial infarction (NSTEMI). Missing values were excluded case wise.
Please cite this article as: D.W.T. Nilsen, et al., The prognostic utility of dihomo-gamma-linolenic acid (DGLA) in patients with acute coronary heart disease, Int J Cardiol (2017), https://doi.org/10.1016/j.ijcard.2017.09.202
D.W.T. Nilsen et al. / International Journal of Cardiology xxx (2017) xxx–xxx Table 3 Univariate and multivariable HRs (95% CI) for LA (18:2n-6) in relation to 1) total death, 2) total death or myocardial infarction (MI) and 3) total death or MI or stroke groups 7year follow-up, comparing the highest to the lowest quartile. All patients n = 398 18:2n-6 (LA)
Total death Total death or MI Total death or MI or stroke
Quartiles univariatea
Quartiles multivariablea
HR (95% CI) p-value
HR (95% CI) p-value
0.59 (0.38–0.91), 0.018 0.55 (0.37–0.81), 0.002 0.55 (0.38–0.81), 0.002
0.79 (0.50–1.25), 0.313 0.72 (0.48–1.08), 0.113 0.70 (0.47–1.04), 0.075
a HR (hazard ratio) [95% CI (Confidence Interval)]. Highest quartile (Q4) versus lowest quartile (Q1). A set of potentially confounding variables was included in the adjusted models in a forward stepwise procedure, as accounted for in Table 2.
monocytes and epidermal cells [9]. PGE1 may retard platelet aggregation, theoretically reducing risk for thrombus formation [10], but its clinical relevance has not been documented, and this mechanism may not be active. Moreover, DGLA competes with AA for COX 2, so in theory, higher DGLA levels could slow the formation of PGE2 and leukotrienes from AA. These properties may retard cancer growth [22] and progression of atherothrombosis [10], possibly explaining the association between DGLA and total mortality and cardiovascular events we observed here. DGLA arises from elongation of GLA. Although DGLA was found here to predict future risk after an ACS, no evidence has been provided for cardiovascular benefits of either GLA supplementation (which increases the level of DGLA and its active eicosanoid metabolites) in primary prevention [13], nor does GLA have prognostic value itself. However, RCTs are limited and data on this question are scarce [7,13,14]. The effect of n-6 PUFA consumption on CHD events is somewhat mixed. The CHD benefits of n-6 PUFA (primarily LA) have been summarized in a 2009 American Heart Association Scientific Advisory [7]. The Advisory concluded that the consumption of at least 5% to 10% of energy from n-6 PUFAs reduces the risk of CHD relative to lower intakes. These conclusions were recently confirmed by a meta-analysis reporting protective effects of higher intakes of linoleic acid [2]. LA was found to be a significant predictor of cardiac mortality outcomes in the present study as well, but its predictive value was diminished after controlling for several other variables. Lending support to LA as a potential predictor, Shearer et al. [23] identified LA as one of 10 FAs in red blood cells (RBC) of ACS patients, matched according to age, gender and race, to be significantly related to ACS case status, whereas DGLA was not identified as a predictor in that study. Whether DGLA's prognostic utility reflects a causal relationship or not is unclear, and only RCTs with DGLA would be able to directly address the question. There are limitations that should be considered in our study. It is based on data from a single center where fish intake is atypically high (in the western coast of Norway), hence our results may not be representative for other populations. Our patients were relatively old (early 70s) at baseline, and with an additional 5 years of follow-up, we need to consider competing events in addition to cardiovascular events. At 2 years follow-up, malignant disease had been diagnosed in 9.4% of the total patient cohort [17]. However, as we are dealing with an ACS population at inclusion, our results should be mainly related to cardiac deaths, and our data do not allow us to assume that DGLA also may predict deaths from non-cardiovascular disease. There is a ten-year span across the quartiles of DGLA, but HRs from the linear analysis, where age was controlled for were not essentially different from those derived from the quartile analyses. Although the number of included patients was small, they experienced a relatively high rate of events due to the long-term follow-up, improving our ability to detect statistically significant relationships. Ideally, BMI,
5
previously accounted for [1], should also have been included in the present analysis, but this information lacked in twenty patients and affected the model. As the multivariable assessment of DGLA Q4 vs Q1 remained significant but less so (p = 0.031) after including BMI, we chose to keep the original model. With this trial design, we were unable to follow dietary/lifestyle/ medication changes over the entire follow-up period of 7 years and their potential effects on outcomes. Finally, unmeasured confounding factors may still be affecting the associations observed here. In conclusion, RBC DGLA levels were found to be a useful predictor of long-term outcomes in patients after hospitalization for an ACS, whereas the other major n-6 PUFAs did not render significant, independent prognostic information.
Acknowledgements Grants for this study were provided by the Western Norway Regional Health Authority, Norway. References [1] H. Aarsetøy, V. Pønitz, H. Grundt, H. Staines, W.S. Harris, D.W.T. Nilsen, (n-3) fatty acid content of red blood cells does not predict risk of future cardiovascular events following an acute coronary syndrome, J. Nutr. 139 (3) (2009) 507–513. [2] M.S. Farvid, M. Ding, A. Pan, Q. Sun, S.E. Chiuve, L.M. Steffen, W.C. Willett, F.B. Hu, Dietary linoleic acid and risk of coronary heart disease: a systematic review and meta-analysis of prospective cohort studies, Circulation 130 (2014) 1568–1578. [3] R.C. Block, W.S. Harris, K.J. Reid, J.A. Spertus, Omega-6 and trans fatty acids in blood cell membranes: a risk factor for acute coronary syndromes? Am. Heart J. 156 (6) (2008 Dec) 1117–1123, https://doi.org/10.1016/j.ahj.2008.07.014 (Epub 2008 Oct 14). [4] R. Showdhury, S. Warnakula, S. Kunutsor, F. Crowe, H.A. Ward, L. Johnson, O.H. Franco, A.S. Butterworth, N.G. Forouhi, S.G. Thompson, K.-T. Khaw, D. Mozaffarian, J. Danesh, E. Di Angelantonio, Association of dietary, circulating, and supplement fatty acids with coronary risk: a systematic review and meta-analysis, Ann. Intern. Med. 160 (2014) 398–406, https://doi.org/10.7326/M13-1788. [5] S. Sergeant, E. Rahbar, F.H. Chilton, Gamma-linolenic acid, dihomo-gamma linolenic, eicosanoids and inflammatory processes, Eur. J. Pharmacol. 785 (2016 Aug 15) 77–86, https://doi.org/10.1016/j.ejphar.2016.04.020. [6] Whitney Ellie, S.R. Rolfes, Understanding Nutrition, 11th ed. Thomson Wadsworth, California, 2008 154. [7] W.S. Harris, D. Mozaffarian, E. Rimm, P. Kris-Etherton, L.L. Rudel, L.J. Appel, M.M. Engler, M.B. Engler, F. Sacks, 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 119 (6) (2009 Feb 17) 902–907, https://doi.org/10.1161/circulationaha.108.191627 (Epub 2009 Jan 26). [8] N. Hussein, E. Ah-Sing, P. Wilkinson, C. Leach, B.A. Griffin, D.J. Millward, Long-chain conversion of [13C]linoleic acid and alpha-linolenic acid in response to marked changes in their dietary intake in men, J. Lipid Res. 46 (2) (2005 Feb) 269–280 (Epub 2004 Dec 1). [9] Y.-Y. Fa, R.S. Chapkin, Importance of dietary gamma-linolenic acid in human health and nutrition, J. Nutr. 128 (9) (1998 Sep) 1411–1414 (Review). [10] P.B.A. Kernoff, A.L. Willis, K.J. Stone, J.A. Davies, G.P. McNicol, Antithrombotic potential of dihomo-gamma-linolenic acid in man, Br. Med. J. 2 (6100) (1977 Dec 3) 1441–1444. [11] M.M. Johnson, D.D. Swan, M.E. Surette, J. Stegner, T. Chilton, A.N. Fonteh, F.H. Chilton, Dietary supplementation with gamma-linolenic acid alters fatty acid content and eicosanoid production in healthy humans, J. Nutr. 127 (8) (1997, Aug) 1435–1444. [12] R.B. Zurier, R.G. Rosetti, E.W. Jacobson, D.M. DeMarco, N.Y. Liu, J.E. Temming, B.M. White, M. Laposata, Gamma-linolenic acid treatment of rheumatoid arthritis. A randomized, placebo-controlled trial, Arthritis Rheum. 39 (11) (1996) 1808–1817. [13] L. Al-Khudairy, L. Hartley, C. Clar, N. flowers, L. Hooper, K. Rees, Omega 6 fatty acids for the primary prevention of cardiovascular disease, Cochrane Database Syst. Rev. 11 (2015 Nov. 16), CD011094. https://doi.org/10.1002/14651858.CD011094.pub2. [14] D.J. Gordon, W.B. Thomson, Lowering cholesterol and total mortality, in: B.M. Rifkin (Ed.), Lowering Cholesterol in High-risk Individuals and Populations, Marcel Dekker, Inc, New York, NY 1995, pp. 33–48. [15] V. Pönitz, T. Brügger-Andersen, D. Pritchard, H. Grundt, H. Staines, D.W. Nilsen, RACS Study Group, Activated factor XII type A predicts long-term mortality in patients admitted with chest pain, J. Thromb. Haemost. 7 (2) (2009 Feb) 277–287, https://doi.org/10.1111/j.1538-7836.2008.03248.x (Epub 2008 Dec 1). [16] P.A. Naesgaard, V. Pönitz, H. Aarsetoey, T. Brügger-Andersen, H. Grundt, W.S. Harris, H. Staines, D.W. Nilsen, Prognostic utility of vitamin D in acute coronary syndrome patients in coastal Norway, Dis. Markers 2015 (2015), 283178. https://doi.org/10. 1155/2015/283178. [17] T. Brügger-Andersen, V. Pönitz, F. Kontny, H. Staines, H. Grundt, M. Sagara, D.W. Nilsen, The long pentraxin 3 (PTX3): a novel prognostic inflammatory marker for
Please cite this article as: D.W.T. Nilsen, et al., The prognostic utility of dihomo-gamma-linolenic acid (DGLA) in patients with acute coronary heart disease, Int J Cardiol (2017), https://doi.org/10.1016/j.ijcard.2017.09.202
6
D.W.T. Nilsen et al. / International Journal of Cardiology xxx (2017) xxx–xxx
mortality in acute chest pain, Thromb. Haemost. 102 (3) (2009 Sep) 555–563, https://doi.org/10.1160/TH09-02-0137. [18] Ø.R. Mjelva, V. Pönitz, T. Brügger-Andersen, H. Grundt, H. Staines, D.W. Nilsen, Longterm prognostic utility of pentraxin 3 and D-dimer as compared to high-sensitivity C-reactive protein and B-type natriuretic peptide in suspected acute coronary syndrome, Eur. J. Prev. Cardiol. 23 (11) (2016 Jul) 1130–1140, https://doi.org/10. 1177/2047487315619733 (Epub 2015 Dec 3). [19] H. Aarsetøy, Pønitz, O.B. Nilsen, H. Grundt, Harris W.S. Nilsen DWT, Low levels of cellular omega-3 increases the risk of ventricular fibrillation during the acute ischemic phase of a myocardial infarction, Resuscitation 78 (2008) 258–264.
[20] W.S. Harris, J. Polreis, Measurement of the omega-3 index in dried blood spots, Ann. Clin. Lab. Res. 4 (2016) 4, https://doi.org/10.21767/2386-5180.1000137. [21] W.S. Harris, J.V. Pottala, S.A. Varvel, J.J. Borowski, J.N. Ward, J.P. McConnell, Erythrocyte omega-3 fatty acids increase and linoleic acid decreases with age: observations from 160,000 patients, Prostaglandins Leukot. Essent. Fat. Acids 88 (4) (2013 Apr) 257–263, https://doi.org/10.1016/j.plefa.2012.12.004 (Epub 2013 Jan 31). [22] X. Wang, H. Lin, Y. Gu, Multiple roles of dihomo-γ-linolenic acid against proliferation diseases, Lipids Health Dis. 11 (2012) 25. [23] G.C. Shearer, J.V. Pottala, J.A. Spertus, W.S. Harris, Red blood cell fatty acid patterns and acute coronary syndrome, PLoS One 4 (2009), e5444.
Please cite this article as: D.W.T. Nilsen, et al., The prognostic utility of dihomo-gamma-linolenic acid (DGLA) in patients with acute coronary heart disease, Int J Cardiol (2017), https://doi.org/10.1016/j.ijcard.2017.09.202
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
The prognostic utility of dihomo-gamma-linolenic acid (DGLA) in patients with acute coronary heart disease☆ Dennis W.T. Nilsen a,c,⁎, Hildegunn Aarsetoey b, Volker Pönitz a, Trygve Brugger-Andersen a, Harry Staines d, William S. Harris e, Heidi Grundt b,c a
Division of Cardiology, Stavanger University Hospital, 4068 Stavanger, Norway Department of Medicine, Stavanger University Hospital, 4068 Stavanger, Norway Department of Clinical Science, University of Bergen, 5020 Bergen, Norway d Sigma Statistical Services, KY16 0BJ Balmullo, UK e Sanford School of Medicine, University of South Dakota, OmegaQuant Analytics, LLC, Sioux Falls, SD, United States b c
a r t i c l e
i n f o
Article history: Received 12 April 2017 Received in revised form 18 August 2017 Accepted 25 September 2017 Available online xxxx Keywords: Dihomo-gamma-linolenic acid (DGLA) Acute coronary syndrome (ACS) Prognosis Total mortality Myocardial infarction (MI) Stroke
a b s t r a c t Background: We previously investigated the prognostic utility of red blood cell (RBC) n-3 fatty acids (FAs) in survivors of an acute myocardial syndrome (ACS) but found no relationship with all-cause mortality and cardiac death or MI after two years. Here we extend our follow-up to 7 years, focusing on the potential predictive power of RBC n-6 FAs. Methods: We included 398 ACS patients presenting with increased troponin-T (TnT) levels for whom baseline RBC FA data were available. Cox regression analysis was used to relate the risk of future events to RBC n-6 FA levels, both continuously and by quartile. Results: At 7-year follow-up, 183 (46.0%) had died, 128 (32.2%) had experienced another MI and 24 (6.0%) had had a stroke. Death or MI occurred in 227 patients (57.0%); and death, MI or stroke in 235 patients (59.0%). In a multivariable Cox regression model for total death, the hazard ratio (HR) in the highest as compared to the lowest quartile of dihomo-γ-linolenic acid (DGLA) was 0.55 [95% confidence interval (CI), 0.35–0.88, p = 0.012, for death or MI [HR 0.62 (95% CI, 0.41–0.94), p = 0.025], and for the fully combined endpoint [HR 0.57 (95% CI, 0.38–0.86), p = 0.006]. Similar results were found in the per 1-SD analysis. No other RBC n-6 FAs significantly predicted these outcomes in multivariable models. Conclusion: RBC DGLA levels had significant independent prognostic value in post-ACS patients. These findings need confirmation, and the possible biochemical pathways by which higher DGLA membrane levels may be cardioprotective should be explored. © 2017 Published by Elsevier Ireland Ltd.
1. Introduction In a patient population from western Norway presenting with an acute coronary syndrome (ACS), red blood cell (RBC) n-3 fatty acid (FA) levels [i.e., the omega-3 index, which is the sum of eicosapentaenoic and docosahexaenoic acids], were unassociated with all-cause mortality and cardiac death or myocardial infarction (MI) after 2-years of follow-up [1]. We have now extended our analysis to 7-year follow-up and focused on the n-6 FAs in RBC membranes [linoleic acid (LA), dihomo-gamma linolenic acid (DGLA) and arachidonic acid (AA)]. Past studies have reported that higher intakes of LA [2], and higher blood levels of LA [3] and AA [4] are cardioprotective; ☆ The patient cohort is registered in ClinicalTrials.gov identifier: NCT00521976. ⁎ Corresponding author at: Department of Clinical Science, University of Bergen, 5020 Bergen, Norway. E-mail address: [email protected] (D.W.T. Nilsen).
and DGLA has been reported to have anti-inflammatory and antiproliferative properties, in addition to being a precursor of AA [5]. As for alpha linoleic acid (ALA) in the n-3 FA series, LA is considered to be an essential polyunsaturated FA (PUFA). LA (18:2n-6) can be desaturated to gamma-linolenic acid (GLA; 18:3n-6) which in turn is elongated into dihomo-gamma-linolenic acid (DGLA; 20:3n6), from which arachidonic acid (AA; 20:4n-6) may be synthesized upon further desaturation [6]. LA which is primarily derived from vegetable oils, such as corn, sunflower, safflower and soy, accounts for N85% to 90% of the dietary n-6 PUFAs, and the average daily intake in the adult population is estimated to approximately 15 g, representing between 6.5 and 7% of energy [7]. In tracer studies, the extent of conversion of LA to AA is very low, no more than about 0.2% [8]; approximately 0.15 g of preformed AA is ingested per day from meat, eggs and some fish [7]. GLA and DGLA are consumed in even smaller amounts than AA and their levels in RBC membranes are primarily determined by metabolism, not diet.
https://doi.org/10.1016/j.ijcard.2017.09.202 0167-5273/© 2017 Published by Elsevier Ireland Ltd.
Please cite this article as: D.W.T. Nilsen, et al., The prognostic utility of dihomo-gamma-linolenic acid (DGLA) in patients with acute coronary heart disease, Int J Cardiol (2017), https://doi.org/10.1016/j.ijcard.2017.09.202
2
D.W.T. Nilsen et al. / International Journal of Cardiology xxx (2017) xxx–xxx
Whereas the metabolic products of AA in the 2-series prostaglandins (PGE2) are mainly considered to promote inflammation and thrombosis, those of DGLA are generally considered to contain anti-inflammatory, antiproliferative and antithrombotic properties [6,9,10], and are formed by the action of cyclooxygenase (COX) 1 & 2 converting DGLA into prostaglandins of the 1-series (PGE-1) and into 15-(S)-hydroxy8,11,13-eicosatrenoic acid (15-HETrE) by the action of 15-lipoxygenase (LOX) [5]. There have been smaller trials investigating the biochemical effects of GLA supplementation which increases the levels of DGLA [11]. In supplementation studies conducted in humans [11,12], GLA is rapidly converted to DGLA. To what extent downstream metabolic products of DGLA are formed under normal conditions, is not known, and the relevance of this pathway to improved health has not been sufficiently investigated. The impact of the n-6 series on variables related to biochemistry and blood pressure in four primary prevention studies including 660 individuals was insufficient to claim an effect [13]. Randomized, controlled studies evaluating coronary heart disease (CHD) events following replacement of saturated fatty acids with PUFAs are burdened with design limitations, such as the lack of blinding and the ambiguity related to whether the effects are due to the removal of one FA class or the addition of another [7]. However, a meta-analysis [14] including 6 such trials indicated that PUFAs may lower the risk of CHD events by 24% in established CHD. We have evaluated the prognostic utility at 7-year follow-up of RBC levels of three individual FAs in the n-6 series: LA, DGLA and AA in patients admitted with chest-pain and a positive ACS biomarker. We hypothesize that one or more of these FAs may predict cardiovascular outcomes in patients hospitalized with an acute coronary syndrome (ACS). 2. Methods 2.1. Study population Our study population belonged to a prospective, observational study termed “Risk Markers in the Acute Coronary Syndrome” (Ref. ClinicalTrials.gov identifier: NCT00521976), designed to identify early risk markers of all-cause mortality and future coronary events following hospitalization with chest pain. A total of 871 patients were recruited at one hospital located in Stavanger, Norway, from November 2002 until October 2003 [15,16]. Of the 871 included patients, 471 had an ACS as defined by a troponin-T (TnT) value ≥0.02 μg/L, of which RBC samples were available in 398 patients. The prognostic utility of several RBC FAs, measured in per cent of total fatty acids, was evaluated in this population. The prognostic utility of brain natriuretic peptide (BNP) and high-sensitivity CRP (hsCRP), respectively, has previously been reported for the main population at 2-year [17] and at 7-year follow-up [18], and that of the omega-3 index has been investigated at 2-year follow-up [1]. The study was approved by the Regional Board of Research Ethics and the Norwegian Health Authorities and conducted in accordance with the Helsinki Declaration of 1975, as revised in 1983. 2.2. Study design Detailed clinical background information was collected in a case report form and reported at baseline (Table 1). During hospitalization 50% of the patients were referred to coronary angiography [1], of which 188 patients were revascularized. Percutaneous coronary intervention (PCI) was performed in 160 subjects and 28 subjects underwent coronary artery bypass grafting (CABG). At discharge all patients received optimal medical treatment according to current recommendations, aiming at N90% on both aspirin and a statin. Follow-up information was collected over the next 2 years by telephone interviews at 30 days and 6, 12, and 24 months. Extended follow-up was performed at 7-year followup. For patients not capable of responding or in case of out-of-hospital death, family members, general practitioners, and nursing staff were contacted for information. Hospital records were searched for detection and verification of reported events. No patients were lost to follow-up. The primary outcome measure was all-cause mortality. Secondary outcomes included recurrent myocardial infarction (MI) and ischemic stroke. Definition of a recurrent MI was based on symptoms of coronary ischemia associated with a typical rise and fall of TnT, exceeding 0.05 μg/L. Only the first recurrent MI or stroke was recorded during the observation period, and for non-survivors, death overruled either of these events.
2.3. Laboratory methods Venous blood samples were collected immediately following admission by direct venipuncture of an antecubital vein. Analyses of TnT, hematological variables, creatinine, glucose, and lipids were performed in serum as part of the hospital's routine. TnT was repeated on average 8 h following admission. Additional TnT measurements were performed with increasing values. RBC for analysis of total fatty acids were immediately prepared from citrated blood and stored at − 70 °C. At the end of the inclusion period samples were shipped dried on filter papers in sealed plastic bags on cold packs to Kansas City, where they again were kept for a short period at − 80 °C prior to measurements with flame ionization GC (GC9A, Shimadazu), as previously reported [19]. Fatty acids were reported as weight percent of total FA. The CV and stability of the FAs have previously been reported [20]. The FA composition in dried RBC spots reflects the FA composition of freshly thawed samples, and tests have been performed in relation to storage time, transport, and temperature [19]. Plasma brain natriuretic peptide (BNP) (Microparticle Enzyme Immunoassay, Abbott AxSYM) was measured in EDTA plasma, and serum was used for measurement of highsensitivity C-reactive protein (hsCRP) [Tina-quant C-reactive protein (latex) high sensitive assay, Roche Diagnostics], respectively [17]. Laboratory personnel had no access to clinical information.
2.4. Statistical analyses Cox regression analyses for the evaluation of follow-up events was applied to continuous values and quartiles of predictor variables at inclusion. A set of potentially confounding variables was included in the adjusted models in a forward stepwise procedure: age; sex; history of angina pectoris and cardiac failure; previous MI, percutaneous coronary intervention (PCI), or coronary bypass artery grafting (CABG); history of hypertension and diabetes mellitus (both insulin-dependent and noninsulin-dependent); current smoking; use of medication; baseline highsensitivity C-reactive protein (hsCRP), brain natriuretic peptide (BNP), creatinine, estimated glomerular filtration rate (eGFR), total cholesterol, HDL-cholesterol, triglycerides (TG), and hypercholesterolemia; peak TnT value; ST-elevation myocardial infarction (STEMI); non-ST-elevation myocardial infarction (NSTEMI). Missing values were excluded case wise. Body mass index (BMI) was also recorded, but not included in the model due to the number of missing values (n = 20). The hazard ratio (HR) for new cardiac events according to quartiles relative to the lowest quartile is presented with 95% CI. The chi-square test was applied when comparing proportions for qualitative variables among quartiles at baseline. Normally distributed quantitative variables are given as means ± SD and differences among quartiles at baseline evaluated through one way analysis of variance. In case of nonnormally distributed variables, the Kruskal-Wallis test was used. The results of nonparametric testing are given as median values with upper and lower quartiles. The area under the Receiver Operating Characteristic curve and asymptotic significance when tested for different to 0.5 were found. The statistical analyses were performed with the statistical package SPSS version 24.0. All tests were applied with a 2-sided significance level of 5%.
3. Results Of the 398 TnT-positive patients for whom RBC fatty acids were available, 65.1% were males, and the mean age (± SD) of the total population was 71.9 ± 13.0 years. At 2-year follow-up, 86 (21.6%) patients had died. Death or MI was recorded in 152 patients (38.2%). When including stroke in the combined endpoint, there were 162 (40.7%) events. DGLA was not found to be a significant predictor at 2 years, but a trend towards statistical significance was observed in relation to the combined endpoints consisting of total mortality or MI or stroke, when comparing Q4 to Q1 [HR 0.63 (95% CI, 0.39–1.02), p = 0.061]. As shown in Table 1, there was a linear decrease in age across the DGLA quartiles (p = 0.000) with patients in Q1, on average, 10 years older than those in Q4. Furthermore, there was a decrease in BNP, history of heart failure, NYHA class and use of diuretics from Q1 to Q4, whereas the opposite trend was demonstrated for peak TnT, STEMI and kidney function (with no patient on hemodialysis). By 7-year follow-up, 183 patients (46%) had died. Total mortality or MI was recorded in 227 (57%), and total mortality or MI or stroke in 235 patients (59%).
Please cite this article as: D.W.T. Nilsen, et al., The prognostic utility of dihomo-gamma-linolenic acid (DGLA) in patients with acute coronary heart disease, Int J Cardiol (2017), https://doi.org/10.1016/j.ijcard.2017.09.202
D.W.T. Nilsen et al. / International Journal of Cardiology xxx (2017) xxx–xxx
3
Table 1 Baseline characteristics according to quartiles of dihomo-gamma linolenic acid C20:3n-6 (DGLA), measured in per cent of total fatty acids in red blood cells. Variable
a
Median (25–75%)
Age, yr Male gender, % (n) Risk markers at baseline Tot. chol. (mmol/L)a HDL chol. (mmol/L)a Triglycerides (mmol/L)a Peak TNT (ng/mL)b hsCRP (mg/L)b BNP (pg/mL)b Risk factors Current smoking, % (n) Hypertension, % (n) IDDM, % (n) NIDDM, % (n) eGFR (ml/min/1.73 m2)a Hypercholesterolemia, % (n) History of coronary heart disease Angina pectoris, % (n) MI, % (n) CABG, % (n) PCI, % (n) History of heart failure CHF, % (n) NYHA 2–4, % (n) Admission data STEMI, % (n) NSTEMI, % (n) U-AP, % (n) Non-ACS, % (n) ASA, % (n) Clopidogrel, % (n) Statins % (n) Beta blockers, % (n) ACE-inhibitors, % (n) ARBs, % (n) Diuretics, % (n)
Total
Q1
Q2
Q3
Q4
1.04 (0.84–1.12)
1.30 (1.24–1.35)
1.54 (1.47–1.60)
1.89 (1.77–2.17)
p-Value
Cohorta
(n = 99)
(n = 100)
(n = 100)
(n = 99)
71.9 ± 13.0 65.1 (259)
76.5 ± 11.4 61.6 (61)
72.8 ± 11.8 64.0 (64)
71.5 ± 12.8 66.0 (66)
66.8 ± 14.2 68.7 (68)
0.000 0.758
5.21 ± 1.28 1.34 ± 0.44 1.66 ± 1.29 0.39 (0.09–1.98) 5.56 (2.24–16.78) 184 (53–460)
5.07 ± 1.47 1.41 ± 0.48 1.51 ± 1.24 0.22 (0.05–0.80) 6.7 (2.4–20.0) 352 (121–749)
5.14 ± 1.25 1.39 ± 0.48 1.59 ± 1.15 0.30 (0.09–1.16) 4.0 (1.7–14.4) 224 (54–558)
5.35 ± 1.07 1.36 ± 0.40 1.60 ± 0.88 0.69 (0.15–2.80) 6.1 (3.1–17.6) 137 (41–386)
5.26 ± 1.29 1.20 ± 0.35 1.93 ± 1.72 1.00 (0.16–3.27) 5.5 (2.0–13.5) 75 (35–224)
0.413 0.002 0.097 0.000 0.227 0.000
30.4 (121) 44.2 (176) 1.3 (5) 17.1 (68) 61.6 ± 23.5 44.5 (177)
24.2 (24) 41.4 (41) 2.0 (2) 16.2 (16) 53.8 ± 21.4 40.4 (40)
26.0 (26) 44.0 (44) 1.0 (1) 15.0 (15) 63.5 ± 27.2 50.0 (50)
37.0 (37) 52.0 (52) 1.0 (1) 20.0 (20) 62.0 ± 21.8 43.0 (43)
34.3 (34) 39.4 (39) 1.3 (1) 17.2 (17) 66.9 ± 21.6 44.4 (44)
0.140 0.295 0.892 0.809 0.001 0.575
42.0 (167) 32.7.(130) 8.8 (35) 7.8 (31)
47.5 (47) 37.4 (37) 10.1 (10) 7.1 (7)
45.0 (45) 40.0 (40) 13.0 (13) 8.0 (8)
41.0 (41) 30.0 (30) 6.0 (6) 11.0 (11)
34.3 (34) 23.2 (23) 6.1 (6) 5.1 (5)
0.0.260 0.051 0.230 0.467
30.9 (123) 27.6 (110)
40.4 (40) 38.4 (38)
36.0 (36) 27.0 (27)
23.0 (23) 22.0 (22)
24.2 (24) 23.2 (23)
0.016 0.040
27.4 (109) 54.8 (218) 16.8 (67) 1.0 (4) 33.4 (133) 1.8 (7) 30.7 (122) 32.7 (130) 19.1 (76) 17.8 (71) 34.2 (136)
12.1 (12) 61.6 (61) 25.3 (25) 1.0 (1) 38.4 (38) 1.0 (1) 30.3 (30) 22.2 (22) 19.2 (19) 20.2 (20) 46.5 (46)
25.0 (25) 58.0 (58) 14.0 (14) 3.0 (3) 33.0 (33) 4.0 (4) 33.0 (33) 44.0 (44) 26.0 (26) 15.0 (15) 40.0 (40)
32.0 (32) 51.0 (51) 17.0 (17) 0.0 (0) 35.0 (35) 2.0 (2) 29.0 (29) 35.0 (35) 15.0 (15) 22.0 (22) 24.0 (24)
40.4 (40) 48.5 (48) 11.1 (11) 0.0 (0) 27.3 (27) 0.0 (0) 30.3 (30) 29.3 (29) 16.2 (16) 14.1 (14) 26.3 (26)
0.000 0.217 0.048 0.111 0.408 0.169 0.940 0.009 0.193 0.387 0.001
HsCRP = high-sensitive C-reactive protein. BNP = B-type natriuretic peptide. IDDM = Insulin dependent diabetes mellitus. NIDDM = non-insulin dependent diabetes mellitus. eGFR = estimated glomerular filtration rate. CABG = coronary artery bypass graft. PCI = percutaneous coronary intervention. NYHA = New York Heart Association function classification of heart failure. ARBs = angiotensin II receptor betblockers. HDL = high density lipoprotein. TNT = troponin-T. MI = myocardial infarction. CHF = congestive heart failure. STEMI = ST-elevation myocardial infarction. NSTEMI = non-ST elevation myocardial infarction. U-AP = unstable angina pectoris. ACS = acute coronary syndrome. ASA = acetylsalicylic acid. ACE = angiotensinconverting-enzyme inhibitor. ARB = angiotensin receptor blocker. a Age, eGFR, total cholesterol, HDL cholesterol and triglycerides are given as mean (±Std. deviation). b Peak TnT, hsCRP and BNP are given as median (quartiles Q1–Q3).
3.1. Dihomo-gamma-linolenic acid (DGLA) and long-term (7-year) prognosis At 7-year follow-up the DGLA (%) levels were significantly lower among patients dying than in survivors; 1.33 ± 0.41 vs. 1.53 ± 0.44, p = 0.000. Kaplan–Meier survival curves for all-cause mortality according to quartiles of DGLA are presented in Fig. 1a. Receiver-operated characteristic (ROC) curve for all-cause mortality is shown in Fig. 1b. The area under the ROC for DGLA was 0.640 (p = 0.000). Results of the univariate analysis relating DGLA to prognosis are shown in Table 2. All p-values for the hazard ratios (HR) are statistically highly significant. DGLA quartiles were added to the model in a separate block after using a stepwise approach to identify significant predictor variables, as shown in Table 2. In the multivariable Cox regression model for total death within 7 years, the HR for DGLA in the highest quartile (Q4) as compared to the lowest (Q1) was 0.55 (95% CI, 0.35–0.88), p = 0.012, whereas applying continuous values of DGLA weakened the p-value, p = 0.055 (Table 2). The prognostic utility of DGLA for total mortality is mainly found in the extreme quartiles, when comparing the highest to the
other quartiles (p = 0.041), or the lowest against the other quartiles (p = 0.035). When applying continuous values of DGLA in the multivariable Cox regression model for total death or MI within 7-year follow-up, statistical significance was maintained [HR 0.69 (95% CI, 0.49–0.98), p = 0.037], which was also evident for total death or MI or stroke [HR 0.66 (95% CI, 0.47–0.93), p = 0.017], very similar to the results obtained in the quartile analysis (Table 2). 3.2. LA in relation to long-term prognosis In the univariate analysis, LA was found to be a highly significant predictor of all specified endpoints at 7-year follow-up, but lost its significance in the multivariable analysis. However, the p-value approached significance in the multivariable analysis for death, MI or stroke at 7 years; HR 0.70 95% CI (0.47–1.04), p = 0.075 (Table 3). 3.3. AA in relation to long-term prognosis The prognostic utility of AA was not significant for either the univariate or multivariable models.
Please cite this article as: D.W.T. Nilsen, et al., The prognostic utility of dihomo-gamma-linolenic acid (DGLA) in patients with acute coronary heart disease, Int J Cardiol (2017), https://doi.org/10.1016/j.ijcard.2017.09.202
4
D.W.T. Nilsen et al. / International Journal of Cardiology xxx (2017) xxx–xxx
Fig. 1. a. Kaplan–Meier plot of time to all-cause mortality during 7-year follow-up by DGLA quartiles. b. Receiver-operated characteristics curve for DGLA; area under the curve (AUC) 0.640 (p = 0.000).
3.4. Omega-3 index and long-term prognosis The omega-3 index was not identified as a predictor of risk for the study outcomes. In a multivariable Cox regression model for total death within 7 years, the HR for the omega-3 index in the highest quartile (Q4) as compared to the lowest (Q1) was 0.92 (95% CI, 0.59– 1.50), p = 0.74, for total death or MI the HR was 1.00 (95% CI, 0.67– 1.50), p = 1.00, and for total death or MI or stroke the HR was 1.10 (95% CI, 0.74–1.65), p = 0.63.
4. Discussion We have assessed the 7-year prognostic utility of several n-6 polyunsaturated FAs measured in RBCs from patients hospitalized with an ACS. At 2-years, there was a trend (p = 0.06, quartile 4 vs 1) for an inverse association between DGLA and the fully combined outcome metric.
After 7 years, DGLA was found to independently predict risk of total mortality, essentially of cardiac etiology. DGLA also predicted the combination of total mortality or MI, and that of the combination of total mortality or MI or stroke, respectively, after adjusting for potential confounding and/or mediating factors. These findings may reflect the trend seen for LA in the assessment of prognostic utility. Of interest, we noted a gradual increase in age and in BNP (heart failure), respectively, from upper through lower quartiles of DGLA. The opposite trend was observed for STEMI and peak TnT. Accordingly, DGLA levels seem to be clearly related to cardiac disease burden at baseline. The age difference from Q1 to Q4 of DGLA was 10 years, with the lowest levels of DGLA in Q1, which is in line with a previous report showing that LA, a precursor of DGLA, decreases with increasing age [21]. DGLA is only modestly converted into AA. It may in part be metabolized by COX1 to prostaglandin E1 (PGE1) and/or by 15-LOX into 15HETrE, both of which possess potent anti-inflammatory and antiproliferative properties in vitro in cells such as neutrophils, macrophages/
Table 2 Univariate and multivariable HRs (95% CI) for C20:3n-6 (DGLA) in relation to 1) total death, 2) total death or myocardial infarction (MI) and 3) total death or MI or stroke groups 7-year follow-up, comparing the highest to the lowest quartile of DGLA and continuous values of DGLA. All patients n = 398 C20:3n-6 (DGLA)
Total death Total death or MI Total death or MI or stroke
Quartiles univariatea
Quartiles multivariablea
Continuous values multivariablea
HR (95% CI) p-value
HR (95% CI) p-value
HR (95% CI) p-value
0.32 (0.20–0.50), 0.000 0.40 (0.27–0.59), 0.000 0.40 (0.27–0.58), 0.000
0.55 (0.35–0.88), 0.012 0.62 (0.41–0.94), 0.025 0.57 (0.38–0.85), 0.006
0.69 (0.47–1.01), 0.055 0.69 (0.49–0.98), 0.037 0.66 (0.47–0.93), 0.017
Age: 1.06 (1.04, 1.08), p = 0.000 CHF 2.08 (1.49, 2.91), p = 0.000 BNP (Q4 vs Q1) 2.24 (1.18, 4.23), p = 0.013 CHF: 2.07 (1.52, 2.82), p = 0.000 Age: 1.03 (1.01, 1.04), p = 0.000 BNP (Q4 vs Q1) 1.71 (1.03, 2.82), p = 0.037 CHF: 2.07 (1.53, 2.80), p = 0.000 Age: 1.03 (1.02, 1.05), p = 0.000 BNP (Q4 vs Q1) 1.71 (1.05, 2.81), p = 0.032
Age: 1.06 (1.04, 1.08), p = 0.000 CHF 1.99 (1.42, 2.77), p = 0.000 BNP (Q4 vs Q1) 2.56 (1.19, 4.26), p = 0.012 CHF: 1.99 (1.47, 2.71), p = 0.000 Age: 1.03 (1.02, 1.05), p = 0.000 BNP (Q4 vs Q1) 1.73 (1.05, 2.85), p = 0.032 CHF: 1.99 (1.47, 2.68), p = 0.000 Age: 1.03 (1.02, 1.05), p = 0.000 BNP (Q4 vs Q1) 1.75 (1.07, 2.85), p = 0.025
Predictor variables entered in the stepwise Cox model Total death Step 1 Step 2 Step 3 Total death or MI Step 1 Step 2 Step 3 Total death or MI or stroke Step 1 Step 2 Step 3
a HR (hazard ratio) [95% CI (Confidence Interval)]. Highest quartile (Q4) versus lowest quartile (Q1), and multivariable analysis based on continuous values. A set of potentially confounding variables was included in the adjusted models in a forward stepwise procedure: age; sex; history of angina pectoris and cardiac failure; previous MI, percutaneous coronary intervention (PCI), or coronary bypass artery grafting (CABG); history of hypertension and diabetes mellitus (both insulin-dependent and noninsulin-dependent); current smoking; use of medication; baseline high-sensitivity C-reactive protein (hsCRP), brain natriuretic peptide (BNP), creatinine, estimated glomerular filtration rate (eGFR), total cholesterol, HDLcholesterol, triglycerides (TG), and hypercholesterolemia; BMI; peak TnT value; ST-elevation myocardial infarction (STEMI); non-ST-elevation myocardial infarction (NSTEMI). Missing values were excluded case wise.
Please cite this article as: D.W.T. Nilsen, et al., The prognostic utility of dihomo-gamma-linolenic acid (DGLA) in patients with acute coronary heart disease, Int J Cardiol (2017), https://doi.org/10.1016/j.ijcard.2017.09.202
D.W.T. Nilsen et al. / International Journal of Cardiology xxx (2017) xxx–xxx Table 3 Univariate and multivariable HRs (95% CI) for LA (18:2n-6) in relation to 1) total death, 2) total death or myocardial infarction (MI) and 3) total death or MI or stroke groups 7year follow-up, comparing the highest to the lowest quartile. All patients n = 398 18:2n-6 (LA)
Total death Total death or MI Total death or MI or stroke
Quartiles univariatea
Quartiles multivariablea
HR (95% CI) p-value
HR (95% CI) p-value
0.59 (0.38–0.91), 0.018 0.55 (0.37–0.81), 0.002 0.55 (0.38–0.81), 0.002
0.79 (0.50–1.25), 0.313 0.72 (0.48–1.08), 0.113 0.70 (0.47–1.04), 0.075
a HR (hazard ratio) [95% CI (Confidence Interval)]. Highest quartile (Q4) versus lowest quartile (Q1). A set of potentially confounding variables was included in the adjusted models in a forward stepwise procedure, as accounted for in Table 2.
monocytes and epidermal cells [9]. PGE1 may retard platelet aggregation, theoretically reducing risk for thrombus formation [10], but its clinical relevance has not been documented, and this mechanism may not be active. Moreover, DGLA competes with AA for COX 2, so in theory, higher DGLA levels could slow the formation of PGE2 and leukotrienes from AA. These properties may retard cancer growth [22] and progression of atherothrombosis [10], possibly explaining the association between DGLA and total mortality and cardiovascular events we observed here. DGLA arises from elongation of GLA. Although DGLA was found here to predict future risk after an ACS, no evidence has been provided for cardiovascular benefits of either GLA supplementation (which increases the level of DGLA and its active eicosanoid metabolites) in primary prevention [13], nor does GLA have prognostic value itself. However, RCTs are limited and data on this question are scarce [7,13,14]. The effect of n-6 PUFA consumption on CHD events is somewhat mixed. The CHD benefits of n-6 PUFA (primarily LA) have been summarized in a 2009 American Heart Association Scientific Advisory [7]. The Advisory concluded that the consumption of at least 5% to 10% of energy from n-6 PUFAs reduces the risk of CHD relative to lower intakes. These conclusions were recently confirmed by a meta-analysis reporting protective effects of higher intakes of linoleic acid [2]. LA was found to be a significant predictor of cardiac mortality outcomes in the present study as well, but its predictive value was diminished after controlling for several other variables. Lending support to LA as a potential predictor, Shearer et al. [23] identified LA as one of 10 FAs in red blood cells (RBC) of ACS patients, matched according to age, gender and race, to be significantly related to ACS case status, whereas DGLA was not identified as a predictor in that study. Whether DGLA's prognostic utility reflects a causal relationship or not is unclear, and only RCTs with DGLA would be able to directly address the question. There are limitations that should be considered in our study. It is based on data from a single center where fish intake is atypically high (in the western coast of Norway), hence our results may not be representative for other populations. Our patients were relatively old (early 70s) at baseline, and with an additional 5 years of follow-up, we need to consider competing events in addition to cardiovascular events. At 2 years follow-up, malignant disease had been diagnosed in 9.4% of the total patient cohort [17]. However, as we are dealing with an ACS population at inclusion, our results should be mainly related to cardiac deaths, and our data do not allow us to assume that DGLA also may predict deaths from non-cardiovascular disease. There is a ten-year span across the quartiles of DGLA, but HRs from the linear analysis, where age was controlled for were not essentially different from those derived from the quartile analyses. Although the number of included patients was small, they experienced a relatively high rate of events due to the long-term follow-up, improving our ability to detect statistically significant relationships. Ideally, BMI,
5
previously accounted for [1], should also have been included in the present analysis, but this information lacked in twenty patients and affected the model. As the multivariable assessment of DGLA Q4 vs Q1 remained significant but less so (p = 0.031) after including BMI, we chose to keep the original model. With this trial design, we were unable to follow dietary/lifestyle/ medication changes over the entire follow-up period of 7 years and their potential effects on outcomes. Finally, unmeasured confounding factors may still be affecting the associations observed here. In conclusion, RBC DGLA levels were found to be a useful predictor of long-term outcomes in patients after hospitalization for an ACS, whereas the other major n-6 PUFAs did not render significant, independent prognostic information.
Acknowledgements Grants for this study were provided by the Western Norway Regional Health Authority, Norway. References [1] H. Aarsetøy, V. Pønitz, H. Grundt, H. Staines, W.S. Harris, D.W.T. Nilsen, (n-3) fatty acid content of red blood cells does not predict risk of future cardiovascular events following an acute coronary syndrome, J. Nutr. 139 (3) (2009) 507–513. [2] M.S. Farvid, M. Ding, A. Pan, Q. Sun, S.E. Chiuve, L.M. Steffen, W.C. Willett, F.B. Hu, Dietary linoleic acid and risk of coronary heart disease: a systematic review and meta-analysis of prospective cohort studies, Circulation 130 (2014) 1568–1578. [3] R.C. Block, W.S. Harris, K.J. Reid, J.A. Spertus, Omega-6 and trans fatty acids in blood cell membranes: a risk factor for acute coronary syndromes? Am. Heart J. 156 (6) (2008 Dec) 1117–1123, https://doi.org/10.1016/j.ahj.2008.07.014 (Epub 2008 Oct 14). [4] R. Showdhury, S. Warnakula, S. Kunutsor, F. Crowe, H.A. Ward, L. Johnson, O.H. Franco, A.S. Butterworth, N.G. Forouhi, S.G. Thompson, K.-T. Khaw, D. Mozaffarian, J. Danesh, E. Di Angelantonio, Association of dietary, circulating, and supplement fatty acids with coronary risk: a systematic review and meta-analysis, Ann. Intern. Med. 160 (2014) 398–406, https://doi.org/10.7326/M13-1788. [5] S. Sergeant, E. Rahbar, F.H. Chilton, Gamma-linolenic acid, dihomo-gamma linolenic, eicosanoids and inflammatory processes, Eur. J. Pharmacol. 785 (2016 Aug 15) 77–86, https://doi.org/10.1016/j.ejphar.2016.04.020. [6] Whitney Ellie, S.R. Rolfes, Understanding Nutrition, 11th ed. Thomson Wadsworth, California, 2008 154. [7] W.S. Harris, D. Mozaffarian, E. Rimm, P. Kris-Etherton, L.L. Rudel, L.J. Appel, M.M. Engler, M.B. Engler, F. Sacks, 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 119 (6) (2009 Feb 17) 902–907, https://doi.org/10.1161/circulationaha.108.191627 (Epub 2009 Jan 26). [8] N. Hussein, E. Ah-Sing, P. Wilkinson, C. Leach, B.A. Griffin, D.J. Millward, Long-chain conversion of [13C]linoleic acid and alpha-linolenic acid in response to marked changes in their dietary intake in men, J. Lipid Res. 46 (2) (2005 Feb) 269–280 (Epub 2004 Dec 1). [9] Y.-Y. Fa, R.S. Chapkin, Importance of dietary gamma-linolenic acid in human health and nutrition, J. Nutr. 128 (9) (1998 Sep) 1411–1414 (Review). [10] P.B.A. Kernoff, A.L. Willis, K.J. Stone, J.A. Davies, G.P. McNicol, Antithrombotic potential of dihomo-gamma-linolenic acid in man, Br. Med. J. 2 (6100) (1977 Dec 3) 1441–1444. [11] M.M. Johnson, D.D. Swan, M.E. Surette, J. Stegner, T. Chilton, A.N. Fonteh, F.H. Chilton, Dietary supplementation with gamma-linolenic acid alters fatty acid content and eicosanoid production in healthy humans, J. Nutr. 127 (8) (1997, Aug) 1435–1444. [12] R.B. Zurier, R.G. Rosetti, E.W. Jacobson, D.M. DeMarco, N.Y. Liu, J.E. Temming, B.M. White, M. Laposata, Gamma-linolenic acid treatment of rheumatoid arthritis. A randomized, placebo-controlled trial, Arthritis Rheum. 39 (11) (1996) 1808–1817. [13] L. Al-Khudairy, L. Hartley, C. Clar, N. flowers, L. Hooper, K. Rees, Omega 6 fatty acids for the primary prevention of cardiovascular disease, Cochrane Database Syst. Rev. 11 (2015 Nov. 16), CD011094. https://doi.org/10.1002/14651858.CD011094.pub2. [14] D.J. Gordon, W.B. Thomson, Lowering cholesterol and total mortality, in: B.M. Rifkin (Ed.), Lowering Cholesterol in High-risk Individuals and Populations, Marcel Dekker, Inc, New York, NY 1995, pp. 33–48. [15] V. Pönitz, T. Brügger-Andersen, D. Pritchard, H. Grundt, H. Staines, D.W. Nilsen, RACS Study Group, Activated factor XII type A predicts long-term mortality in patients admitted with chest pain, J. Thromb. Haemost. 7 (2) (2009 Feb) 277–287, https://doi.org/10.1111/j.1538-7836.2008.03248.x (Epub 2008 Dec 1). [16] P.A. Naesgaard, V. Pönitz, H. Aarsetoey, T. Brügger-Andersen, H. Grundt, W.S. Harris, H. Staines, D.W. Nilsen, Prognostic utility of vitamin D in acute coronary syndrome patients in coastal Norway, Dis. Markers 2015 (2015), 283178. https://doi.org/10. 1155/2015/283178. [17] T. Brügger-Andersen, V. Pönitz, F. Kontny, H. Staines, H. Grundt, M. Sagara, D.W. Nilsen, The long pentraxin 3 (PTX3): a novel prognostic inflammatory marker for
Please cite this article as: D.W.T. Nilsen, et al., The prognostic utility of dihomo-gamma-linolenic acid (DGLA) in patients with acute coronary heart disease, Int J Cardiol (2017), https://doi.org/10.1016/j.ijcard.2017.09.202
6
D.W.T. Nilsen et al. / International Journal of Cardiology xxx (2017) xxx–xxx
mortality in acute chest pain, Thromb. Haemost. 102 (3) (2009 Sep) 555–563, https://doi.org/10.1160/TH09-02-0137. [18] Ø.R. Mjelva, V. Pönitz, T. Brügger-Andersen, H. Grundt, H. Staines, D.W. Nilsen, Longterm prognostic utility of pentraxin 3 and D-dimer as compared to high-sensitivity C-reactive protein and B-type natriuretic peptide in suspected acute coronary syndrome, Eur. J. Prev. Cardiol. 23 (11) (2016 Jul) 1130–1140, https://doi.org/10. 1177/2047487315619733 (Epub 2015 Dec 3). [19] H. Aarsetøy, Pønitz, O.B. Nilsen, H. Grundt, Harris W.S. Nilsen DWT, Low levels of cellular omega-3 increases the risk of ventricular fibrillation during the acute ischemic phase of a myocardial infarction, Resuscitation 78 (2008) 258–264.
[20] W.S. Harris, J. Polreis, Measurement of the omega-3 index in dried blood spots, Ann. Clin. Lab. Res. 4 (2016) 4, https://doi.org/10.21767/2386-5180.1000137. [21] W.S. Harris, J.V. Pottala, S.A. Varvel, J.J. Borowski, J.N. Ward, J.P. McConnell, Erythrocyte omega-3 fatty acids increase and linoleic acid decreases with age: observations from 160,000 patients, Prostaglandins Leukot. Essent. Fat. Acids 88 (4) (2013 Apr) 257–263, https://doi.org/10.1016/j.plefa.2012.12.004 (Epub 2013 Jan 31). [22] X. Wang, H. Lin, Y. Gu, Multiple roles of dihomo-γ-linolenic acid against proliferation diseases, Lipids Health Dis. 11 (2012) 25. [23] G.C. Shearer, J.V. Pottala, J.A. Spertus, W.S. Harris, Red blood cell fatty acid patterns and acute coronary syndrome, PLoS One 4 (2009), e5444.
Please cite this article as: D.W.T. Nilsen, et al., The prognostic utility of dihomo-gamma-linolenic acid (DGLA) in patients with acute coronary heart disease, Int J Cardiol (2017), https://doi.org/10.1016/j.ijcard.2017.09.202