Effect of Over-the-Counter Fish-Oil Administration on Plasma Lp(a) Levels in an End-Stage Renal Disease Population

Study Population 

The ESRD patients associated with the Central Texas Nephrology Associates who were undergoing chronic hemodialysis participated in this study. Exclusion criteria for this study included a life expectancy of less than 6 months (based on physician prognosis), pregnancy, a history of hemodialysis noncompliance, previous medication noncompliance, no desire to participate in the study, or age of less than 18 years. One-hundred patients 18 years of age and older with a history of dialysis and medication compliance were identified, and 40 patients agreed to participate. Five participants from the control group (omega-6 fatty acids, or n-6) and two participants from the n-3 group did not comply with the supplement protocol and were excluded from the final analysis, for a total sample size of 33 at study completion. However, all noncompliant participants were encouraged to continue supplementation for the duration of the study, in compliance with the Declaration of Helsinki and “intent to treat” guidelines. There were no participant reports of CVD-related symptoms, and specifically, no findings of palpitations, syncope, clinically significant episodes of arrhythmia, or other cardiac manifestations during the study. Patients meeting eligibility criteria were informed of the requirements of the study and signed informed-consent statements, in compliance with the human-subjects guidelines of the university and clinics for the protection of human subjects in research before the intervention.

Experimental Design and Procedures 

The study was conducted using a double-blind, permuted-block, randomized, and controlled experimental design, with a four-block permuted randomization procedure. Each “block” had an equal number of experimental and control group selections (two each), with the order of blocks permuted. Microsoft Excel was used to generate the random number used to assign blocks and randomize group selection. Participants were randomly placed into the n-3 fatty acid (experimental, n=18) or control (n=15) group. Participants in the experimental group consumed two 1-g soft-gel capsules of over-the-counter (OTC) fish-oil concentrate with each meal, or 6g (six capsules) per 24hours containing 160mg of eicosapentaenoic acid (EPA; 0.96g/day) and 100mg of docosahexaenoic acid (DHA; 0.6g/day). The control group consumed two 1-g soft-gel capsules of corn oil with each meal, to accumulate 6g (six capsules) per 24hours. Both groups were instructed to take the supplement in the same way. Outcome variables were measured at baseline and at 6 months. Furthermore, all patients consumed vitamin supplements that contained 15mg of B6, 12mg of B12, and 2.5mg of folic acid.

Supplement Distribution and Compliance 

Participants' regular renal dietitians distributed study materials and monitored supplement and dietary compliance. Dietitians issued a 1-month supply six times during the study period for each patient. Moreover, the dietitian and clinic physician consulted with the patients three times per week during dialysis sessions throughout the study period. Participants were considered compliant if they consumed 90% of the n-3 fatty-acid soft-gel supplements provided, based on the common practice of “pill-counting.”22, 23 The renal dietician was responsible for monitoring compliance with all patients.

Dialysis Protocol 

The dialysis clinics used polysulfone membranes, with 95% using a Fresenius-160 dialyzer, and 5% using a Fresenius-180 dialyzer (Fresenius Medical Care, Ogden, Utah), at a dialysate flow rate of 800cm3/minute and a mean blood flow rate of 376.49cm3/minute. The dialysis dose for both n-3 fatty-acid and control-oil groups entailed a Kt/V range of 1.1 to 2.0, with means of 1.35 and 1.38, respectively. All patients ran 4-hour dialysis duration per unit protocol.

Lipid Analysis 

Venous blood samples were collected at baseline and after completion of the supplement intervention 6 months later. A phlebotomist collected approximately 20mL of blood from each participant (after fasting for 12hours), using standardized venipuncture techniques in the antecubital vein. Venous samples, intended for serum analysis, were centrifuged and immediately placed in a cold-storage unit and sent for assay. Lipid profiles were assessed using gel electrophoresis with whole-blood samples sent to Quest Diagnostics (Dallas, TX). The Lp(a) was obtained and is reported in micromoles per liter (μmol/L), as well as milligrams per deciliter (mg/dL).

Fish-Oil and Control Composition 

Fish oil and corn oil (control) were quality-assured and quality-controlled by Royal Numico Research B.V. (Greenville, SC). Fish-oil soft gels were packaged in a 1-g capsule that contained 160mg of EPA and 100mg of DHA, and 0.9IU of d-α tocopherol as an antioxidant. The control soft gels were packaged in a 1-g capsule of corn oil (94% unsaturated fat, 6% saturated fat). The fish-oil and corn-oil supplements were indistinguishable from each other, and were “minted” to have the same flavor. Patients, physicians, and researchers were blinded to which supplement patients were consuming. Corn oil was used as a control in a number of studies associated with n-3 supplementation and lipid changes.24, 25, 26, 27

Statistical Analysis 

We summarized all baseline demographic information using frequencies and percentages for categorical variables, and means and standard deviations for continuous variables. We assessed the primary research question of whether the fish-oil intervention had a different effect on Lp(a) concentrations compared with the control, by using a repeated-measures analysis of variance (ANOVA) and testing for significance the multivariate Wilks' λ statistic for the group-by-time interaction. This allows for appropriate control of within-individual variability of the data, while ascertaining whether there is a difference in Lp(a) response over time as a result of the intervention. Of secondary interest, if the first test did not produce significant results, was the issue of whether a time effect or group effect existed within the data.

A secondary aim of the study was to explore the qualities that affect Lp(a) levels, given the findings of the first portion of the study. We used three repeated-measures ANOVA models with age, race, and gender as predictors. This necessitated removing the three Hispanic individuals from the data, because the available sample size was inadequate for a subgroup analysis. The use of three separate models was necessary, given that the sample size of the study could not sustain a model with all covariates and interactions. All data summaries and statistical results were produced using SAS, version 9.2 (SAS Institute, Cary, NC). We performed all comparisons using a type I error rate of 0.05.

Limitations 

A few limitations of the study design are worth noting. First, compliance to the dietary protocol in our study was assessed using the standard method of pill-counting. Although good compliance was reported in all participants retained for analysis (>90%), future studies should use more objective measures, such as in vivo testing to confirm patient compliance. Second, although supplements were quality-controlled and quality-assured by the manufacturer before shipping, OTC supplements can have significant variation in content. Therefore, the amounts of DHA and EPA per capsule may have varied, thereby decreasing the likelihood of detecting a significant difference in Lp(a) concentration. Third, in terms of assessing cardiovascular risk, the way in which Lp(a) was measured in our study, although widely practiced, ignores the variation in apo(a) gene size when quantifying Lp(a) levels. Several studies showed that Lp(a) particles that carry small apo(a) molecules are associated with increased cardiovascular risk,34, 35 especially in ESRD patients,36 and therefore may provide limited information on true cardiovascular risk. Lastly, the large intra-individual variability of our dataset may have limited our ability to observe a significant effect of fish oil on Lp(a). Values of Lp(a) in our dataset ranged from 0.02 to 3.26 μmol/L (0.6 to 91.3mg/dL) at baseline, and 0.02 to 3.81 μmol/L (0.5 to 106.9mg/dL) after supplementation, with intra-individual before-and-after differences from −2.40 to 0.92 μmol/L (−67.1 to 25.8mg/dL). However, given the current assessment technology, such variability may be common. Marcovina et al. showed that the biological variability for Lp(a) is high, ranging anywhere from 1% to 51%, depending on Lp(a) concentration.37 Future research should focus on the nature of the variability of Lp(a), i.e., whether such variability is largely biological in nature, or reflects a flaw in the measurement process. In the present study, a larger sample size could have overcome the loss in power due to the sizable variability of data.