The 129 subjects included in the study were healthy volunteers of both genders with normal or slightly elevated total blood cholesterol (<7.5 mmol/L) and normal or slightly elevated blood triglyceride level (<4.0 mmol/L). Subjects with body mass index (BMI) >30 kg/m², hyperlipidemia, hypertension, coronary, peripheral or cerebral vascular disease were excluded from participating in the study. No concomitant medication intended to influence serum lipid level was permitted. All study subjects were informed verbally and in writing, and all subjects signed an informed consent form before entering the study. The study was approved by the Regional Ethics Committee.

The study was an open single-center, randomized, parallel group designed study. Screening of subjects (N = 149) was performed at the first visit to include subjects that satisfied the eligibility criteria (N = 129). These were randomized into three study groups. Seven participants were lost before the baseline visit. The remaining 122 participants were given either 3 g krill oil daily (N = 41), 1.8 g fish oil daily (N = 40) or no supplementation (N = 41) for a period of 7 weeks. A total of 115 participants finished the study. The disposition of the subjects is illustrated in Fig. 1. None of the subjects regularly ate fatty fish more than once a week prior to inclusion or during the 7-week intervention period. None were using cod liver oil or other marine n-3 supplements during the study or at least 2 months prior to inclusion. All the participants were instructed by a nutritionist to keep their regular food habits during the study.

The krill oil capsules contained processed krill oil extracted from Antarctic krill (Euphausia superba). The product was manufactured by Aker BioMarine ASA. Each capsule contained 500 mg oil that provided 90.5 mg EPA and DHA, and a total of 103.5 mg n-3 PUFAs. The capsules were made of gelatin softened with glycerol. The daily study dosage was six capsules (each of 500 mg oil). The comparator omega-3 fish oil product was manufactured by Peter Möller AS, Oslo, Norway. The daily study dosage was three capsules each containing 600 mg fish oil that provided 288 mg EPA and DHA, and a total of 330 mg n-3 PUFAs. The capsules were made of gelatin softened with glycerol. The fatty acid profile of the study products is presented in Table 1. The daily dose of EPA, DHA, and total n-3 PUFAs in the krill oil and fish oil groups is presented in Table 2. The daily EPA + DHA dose in the krill oil group was 62.8% of the dosage given in the fish oil group. DL-α-tocopheryl acetate (vitamin E), retinyl palmitate (vitamin A), and cholecalciferol (vitamin D) were added to the product.

Demographic characteristics (gender, age, height, and weight), concomitant medication, and medical history were recorded at the screening visit. In addition, all subjects went through a physical examination to confirm satisfaction of the eligibility criteria.

Changes in concomitant medication from screening, smoking and alcohol habits, and clinical symptoms before intake of the study products were also registered.

Blood from venipuncture was collected after an overnight fast (≥12 h) at baseline and at final visit. The subjects were instructed to refrain from alcohol consumption and from vigorous physical activity the day before the blood sampling. Serum was obtained from silica gel tubes (BD Vacutainer), kept at room temperature for at least 30 min until centrifugation at 1,300 × g for 12 min at room temperature. Serum analysis of total, low-density lipoprotein (LDL), and HDL-cholesterol, TG, apolipoprotein A1, and apolipoprotein B was performed at the routine laboratory at Department of Medical Biochemistry at the National Hospital, Norway using standard methods. Blood samples for analysis of hematology and serum biochemistry parameters including hemoglobin, leukocytes, erythrocytes, thrombocytes, hematocrit, glucose, calcium, sodium, potassium, urea, creatinine, alkaline phosphatase, alanine aminotransferase, bilirubin, albumin, and total protein were collected and analyzed at the routine laboratory at Department of Medical Biochemistry at the National Hospital, Norway using standard methods.

Plasma was obtained from ethylenediamine tetraacetic acid (EDTA) tubes (BD Vacutainer) kept on ice immediately and within 12 min centrifuged at 1,300 × g for 10 min at 10°C. Plasma samples were kept frozen at −80°C until analysis. Plasma fatty acid composition was analyzed by Jurilab Ltd., Finland using a slight modification of the method of Nyyssonen et al. [¹⁶]. Plasma (250 μL), fatty acids, and 25 μL internal standard (eicosane 1 mg/mL in isopropanol) were extracted with 6 mL methanol–chloroform (1:2), and 1.5 mL water was added. The two phases were separated by centrifugation, and the upper phase was discarded. To the chloroform phase, 1 mL methanol–water (1:1) was added, and this extraction was repeated twice. The chloroform phase was evaporated under nitrogen. For methylation, the remainder was treated with 1.5 mL sulfuric acid–methanol (1:50) at 85°C for 2 h. The mixture was diluted with 1.5 mL water and extracted with light petroleum ether. The fatty acids from the ether phase were determined using a gas chromatograph (Agilent Technologies 6890)/mass spectrometer (Agilent Technologies 5973) with electron impact ionization and a HP-5ms capillary column (Hewlett Packard). For retention time and quantitative standardization, fatty acids purchased from Nu-Chek-Prep (Elysian, MN, USA) were used. All work was carried out under a certified ISO 9001/2000 quality system.

Human plasma (100 μL) was diluted with 300 μL 2-propanol containing the internal standard tocol and butylated hydroxytoluene (BHT) as an antioxidant. After thorough mixing (15 min) and centrifugation (10 min, 4,000 × g at 10°C), an aliquot of 1 μL was injected from the supernatant into the high-performance liquid chromatography (HPLC) system. HPLC was performed with a HP 1100 liquid chromatograph (Agilent Technologies, Palo Alta, CA, USA) with a HP 1100 fluorescence detector (emission 295 nm, excitation 330 nm). Tocopherol isomers were separated on a 2.1 × 250 mm reversed-phase column. The column temperature was 40°C. A two-point calibration curve was made from analysis of a 3% albumin solution enriched with known concentration of tocopherols. Recovery is >95%, the method is linear from 1 to 200 μM at least, and the limit of detection is 0.01 μM. Relative standard deviation (RSD) is 2.8% (17.0 μM) and 4.6% (25.1 μM).

Fasting urine samples were analyzed for 8-iso-prostaglandin F_2α (8-iso-PGF_2α) by a highly specific and validated radioimmunoassay as described by Basu [¹⁷]. Urinary levels of 8-iso-PGF_2α were adjusted by dividing the 8-iso-PGF_2α concentration by that of creatinine.

Plasma interleukin-6 (IL-6), tumour necrosis factor-alpha (TNFα), monocyte chemotactic protein-1 (MCP-1), thromboxane B₂ (TxB₂), interferon-gamma (INFγ), soluble E-selectin and P-selectin, soluble intracellular adhesion molecule-1 (ICAM-1), and vascular cell adhesion molecule-1 (VCAM-1) were determined by Fluorokine^® MAP kits (R&D Systems, Inc., Minneapolis, MN, USA). Plasma leukotriene B₄ (LTB₄) and thromboxane B₂ (TxB₂) were assessed as described by Elvevoll et al. [¹⁸]. High-sensitivity C-reactive protein (hsCRP) evaluation was performed at the routine laboratory at Department of Medical Biochemistry at the National Hospital, Norway using standard method.

All continuous variables were summarized by product group and visit number and described using standard statistical measures, i.e., number of observations, mean, standard deviation (SD), median, minimum, and maximum. Absolute and percentage change from baseline to the week-7 visit are presented as summary statistics. All categorical (discrete, including ordinal) variables are presented in contingency tables showing counts and percentages for each treatment group at all time points. Continuously distributed efficacy laboratory parameters (lipids, EPA, DHA, and docosapentaenoic acid (DPA)) were analyzed by analysis of covariance (ANCOVA) using the following model: change in parameter value = baseline value + treatment + gender + age + error.

Change from baseline to week 7 was used as a dependent variable in the model. A linear model using SAS GLM with gender as fixed effect, subject as random effect, and baseline value and age as covariates was applied. A reduced ANCOVA model with baseline value and treatment was used for the secondary efficacy parameters. Significant treatment effects were analyzed by pairwise tests. Changes from baseline to end of intervention were tested by paired t-test.

Figure 1 shows the disposition of all subjects. One hundred fifteen of 129 randomized subjects completed the study. Withdrawal rates were similar in all three groups (three withdrawals in the fish oil group, six withdrawals in the krill oil group, and five withdrawals in the control group). Three subjects discontinued the study due to clinical symptoms (all in the krill group), three subjects violated the exclusion criteria (two in the fish oil group and one in the krill oil group), and one subject in each group was lost to follow-up. Two subjects in the control group were withdrawn due to concomitant treatments, and three subjects withdrew their consent (one subject in the krill oil group and two subjects in the control group). Clinical symptoms included symptoms of common cold or gastrointestinal symptoms. During database clean-up, it was detected that two subjects (in the krill oil group) had been allowed to enter the study although they violated the entry criteria. They where therefore excluded from the per protocol subjects. The statistical analyses of efficacy were performed on the data collected from 113 per protocol subjects (Fig. 1).

The study groups were comparable in terms of weight, height, BMI, gender, and age at baseline (Table 3). Vital signs including systolic and diastolic blood pressure and heart rate were within normal ranges. More females than males were included in all study groups.

Plasma levels of EPA, DHA, and DPA increased significantly from baseline to the end of the intervention phase in the groups receiving fish oil and krill oil, but not in the control group. The changes in EPA, DHA, and DPA differed significantly between the subjects supplemented with n-3 PUFAs and the subjects in the control group, but there was no significant difference in the change in any of the n-3 PUFAs between the fish oil and the krill oil groups (Table 4).

There were significant within-group changes in individual FAs from start to end of intervention, but no clear trends in changes in the plasma FA composition were apparent in any of the study groups (Table 4).

The level of arachidonic acid (C20:4n-6) increased from baseline in the krill group, whereas a decrease was observed in the fish oil group. The changes in arachidonic acid between the fish oil and the krill oil groups, and the control group differed significantly (p = 0.001). Pairwise comparisons showed that the mean increase in arachidonic acid in the krill oil group was significantly different from the mean decreases in the fish oil and control groups, but there was no significant difference between the mean changes in arachidonic acid level between the fish oil and control groups.

Small changes in the levels of HDL-cholesterol, LDL-cholesterol, and TG were observed in all study groups from start to end of the intervention phase, but only the within-group increase in LDL-cholesterol seen in the fish oil group (p = 0.039) was statistically significant. The tests comparing the differences between the study groups gave no statistically significant results (Table 5). The HDL-cholesterol/TG ratio and the change from start to end of the intervention were calculated for all study groups. No significant changes in the HDL-cholesterol/TG ratio from start to end of the interventions were detected in the fish oil or control groups. In the krill oil group, however, there was a significant increase in the HDL-cholesterol/TG ratio. The test for differences between the study groups gave no significant results (Table 5).

Although the interventions did not significantly change TG levels a reduction was seen in those subjects in the krill oil group having the highest baseline values (Fig. 2).

The changes in levels of Apo B-100 from baseline to end of study were minor in all study groups. Moreover, the test for differences between the study groups in changes in Apo A1 was not significant. However, the within-group changes of Apo A1 levels from start to end of the interventions were statistically significant in the krill oil group.

α-Tocopherol is considered an antioxidant, and an increase in PUFAs may lead to increased oxidative stress. Although α-tocopherol was added to both supplements, no significant change in levels of α-tocopherol was detected (Supplementary Table). A tendency towards a reduced level of α-tocopherol was observed in all study groups. F2-isoprostanes, formed from free-radical-induced peroxidation of membrane-bound arachidonic acid, are considered a reliable biomarker of oxidative stress. However no differences were observed in urine F2-isoprostane, suggesting that there was not an increase in oxidative stress. No significant changes were observed in levels of hsCRP, markers of inflammation or hemostasis (Supplementary Table).

Below is the link to the electronic supplementary material.

^aTest of within-group changes

^bTest comparing change from start to end of intervention between the fish oil, krill oil, and control groups

^aTest of within-group changes

^bTest comparing change from start to end of intervention between the fish oil, krill oil, and control groups