The Effect of Mipomersen in the Management of Patients with Familial Hypercholesterolemia: A Systematic Review and Meta-Analysis of Clinical Trials

1.1. Description of the Condition

Familial hypercholesterolemia (FH) is an autosomal dominant disease that is caused by various mutations in a gene located on chromosome 19 whose function is to produce receptors located in sites that include liver. The function of the receptor is catching and removing low-density lipoproteins (LDL) from plasma. Different levels of mutations in the gene will lead to different severity of disturbances in LDL removal from the body [1]. Increased level of serum LDL (hypercholesterolemia) leads to major adverse effects in the cardiovascular system, including the coronary arteries as well as other organs [2]. The most severe form of the mutation leads to a complete lack of the receptor (receptor-negative mutation).

Mutations in two other genes can cause a similar phenotype; the first one is the gene related to apolipoprotein B-100 and the second one is related to the proprotein convertase subtilisin/kexin type 9 (PCSK9) [1].

In FH, malfunction of the receptor leads to the accumulation of LDL in various parts of the body, from the cornea to tendons, skin, and vessels [3,4]. The accumulation of cholesterol in the cornea causes corneal arcus, its accumulation in tendons causes xanthoma, and in the vessels, it can cause different degrees of atherosclerosis [1].

Familial hypercholesterolemia has two types; heterozygous type, which is autosomal dominant, exhibits milder forms of the disease, and its prevalence is about one in 500. The signs and symptoms of the disease, including xanthomas and coronary artery disease, are presented in later years of life (after the second decade) compared with the homozygous type [4]. The homozygous type is less prevalent (about one in a million), and the manifestations of the disease can be presented even in the first or second decade of life by very severe involvement of the vascular system, including the heart, which might need for coronary artery bypass graft surgery [2].

1.2. Description of the Intervention

Different therapeutic strategies have been used so far to treat FH; from dietary intervention with low effectiveness to pharmacological treatments and even invasive procedures such as plasma apheresis. In the late stage of the disease, the only effective treatment can be liver transplantation [5]. Pharmacological treatments include many drugs such as bile acid sequestrants and statins that have been used with different efficacy but mostly with lower than expected results, especially in patients with severe FH [6,7]. So the need for an add-on drug to enhance the effect of first-line treatments is suggested. New ideas include altering the absorption of cholesterol in the brush border of the intestine [7], affecting the related receptor on the liver [8], altering the pathway of producing cholesterol, or using inhibitors of proprotein convertase subtilisin/kexin type 9 (PCSK9) [9]. Apolipoprotein B-100 is an essential component of all atherogenic lipoproteins, including LDL. Altering the pathway of producing apolipoprotein B-100 can help in decreasing the production of LDL.

1.3. How the Intervention Might Work

Mipomersen is a new drug proposed for drug therapy of FH [10]. It is a second-generation antisense oligonucleotide that can inhibit the synthesis of apolipo-protein B-100 and thus decrease the production of LDL and lower LDL cholesterol [11,12]. The route of injection of mipomersen is subcutaneously in a formulation with 0.9% sodium chloride, which can lead to injection site reactions [13].

Because mipomersen targets apoB-100 mRNA in the liver and inhibits hepatic production of triglyceride-rich very low density lipoprotein (VLDL), the drug can result in the development of hepatic steatosis accompanied by elevated plasma liver transaminases [14]. The predominant distributes of mipomersen to the liver rather than the intestine results in a functional specificity for apo B100 synthesis in the liver compared with apo B48 synthesis in the intestine [15,16]. Nevertheless, investigators have reported the intestinal side effects of mipomersen because of inhibition of intestinal chylomicron production and subsequent fat malabsorption [17].” Mipomersen does not, however, have any effect on plasma PCSK9 levels [18].

1.4. Why It Is Important to Do This Review

Some studies have been done on the effect of mipomersen in patients with dyslipidemia, showing various effects as well as some side effects, including increased levels of liver enzymes [19,20]. Waldman et al. and Akdim et al. evaluated the effects of mipomersen in adult patients with heterozygous FH [21,22]. Raal and colleagues evaluated mipomersen as an add-on in patients with homozygous FH aged more than 12 years [15]. In a single-center placebo-controlled clinical trial, Visser and colleagues evaluated the effect of 200 mg mipomersen injection in patients with heterozygous FH [23]. However, there is no systematic review that combines the data from existing randomized clinical trials on patients with FH. Given the lack of synthesized evidence and given the seriousness of the effects of FH, it is imperative that this review be done to examine the effectiveness and safety of mipomersen. The results of this review can help clinicians, and future decision-makers be informed about the possible use of this drug. By considering the possible benefits as well as side effects, this review will assess the role of this drug as an add-on to the current therapeutic management of patients with FH. So, the objective of this review is to answer the question of what might be the effect of subcutaneous injection of mipomersen compared with placebo or no intervention in adults or pediatric patients with FH who are already on the treatment with first-line cholesterol-lowering drugs.

2.1. Criteria for Selecting Studies for This Review

2.2. Search Methods for Identification of Studies

2.3. Data Collection and Analysis

2.4. Data Synthesis

We pooled the results of trials using the generic inverse variance method or Mantel-Haenszel in Review Manager 5.3. We also used random-effects model. For safety analysis, studies with zero effects were excluded from the analysis.

2.5. Subgroup Analysis and Investigation of Heterogeneity

In case of detecting any heterogeneity, we planned to perform sensitivity analysis for studies with a high risk of bias. We planned to do subgroup analyses to determine if the magnitude of effect was related to studies at high versus the lower risk of bias.

2.6. Assessment of the Certainty of the Evidence

For all mentioned outcomes, we used the GRADE approach to evaluate the certainty of evidence. We used the following minimal important difference (MID) values when interpreting the importance of differences between the groups: for injection site reaction, we considered 25% increase in the number of patients with injection site reactions in the mipomersen group than the baseline risk, for increased in serum ALT level we considered even one more patient in 100 patients with increased at least 3 times more than ULN as MID, and for the decrease in serum LDL, we considered the MID as at least 25% decrease in LDL cholesterol after injection of mipomersen. For MACE, and liver toxicity, even one more event in 100 patients compared with the placebo group was considered important.

2.7. Presentation of the Results

For a better presentation of the main results, a table for “Summary of findings” was prepared using the GRADEPro software. Decisions for downgrading the certainty of evidence were mentioned in the footnotes.

3.1. Description of Studies and Results of the Search

The search was done in Jan 2020. Considering the search keywords, totally 219 studies were retrieved from different databases: Embase (161), Medline (24), and Web of Science (34). Searching other sources yielded 2 more studies. Excluding 40 duplicate studies led to 181 unique reports. The titles and abstracts were screened by two reviewers, which led to 34 full-text articles. The full texts were evaluated independently by two reviewers, and the disagreements were resolved by consensus. Finally, considering the inclusion and exclusion criteria and after in-depth evaluation, five articles were selected for inclusion in the analyses. Details of the process of screening and selection of articles are shown in Figure 1 (PRISMA diagram).

3.2. Included Studies

Table 1 shows the basic characteristics of the five included studies. In total 549 patients were included in all studies consisting of 369 patients in treating arms and 180 patients in the control arms. The studies were published between 2010 and 2019, and the follow-up period ranged from 6 weeks to 60 weeks. All but one study had done on adult patients with heterozygous FH [19,20,23,24]. The only one study which was done on patients with homozygous FH had mentioned in its inclusion criteria to include patients aged 12 years and more, but in practice, it could only include adults [11]. So the subgroup analysis between the adults and children that had been set up in priori could not be done. The total dose of mipomersen that had been used in almost all studies was about 200 mg per week. Akdim study had an escalating dose of 50 to 300 mg mipomersen per week in 4 arms, so the arm of 200 mg per week was selected to compare the effect on serum LDL; however, as the authors mentioned that “no evidence was found of a dose-dependent association in adverse events”, to compare injection site reactions and increased ALT, we combined all four intervention arms [19]. All studies but one were multicentre.

3.3. Excluded Studies

3.4. Risk of Bias in Included Studies

We assessed risk of bias in all included studies, which showed a low risk of bias in all domains. Table 3 shows the summarised risk of bias in various domains. Figure 2 shows the assessment of risk of bias in each individual study. Figure 3 shows a summary of the risk of bias by domain. Table S2 shows detailed judgments about the risk of bias in different studies.

3.5. Effects of Interventions and Assessment of the Evidence

3.6. Liver Toxicity

None of the included studies reported liver toxicity, so we could not evaluate this outcome.

3.7. Increased Serum ALT More Than 3 Times ULN

Four studies with 528 patients provided data for this outcome. The Visser’s study did not have any event so it was excluded from the analysis [23]. Figure 6 shows the related forest plot. RR of the pooled estimate was 5.19 and CI (1.01–26.69). This shows that risk of developing an increased level of ALT more than 3 ULN is about 5 times more in patients who received mipomersen compared with placebo. To assess the certainty of the evidence, we found that risk of bias across the studies was low. Statistical evaluation of heterogeneity shows that I² is 56% showing a moderate level of heterogeneity and P-value was 0.08, which is significant, but all CIs do overlap. Other than quantitatively and statistically thinking about the heterogeneity, we considered that the clinical question for this outcome is whether mipomersen can do harm by increasing ALT whose answer is yes because all the studies show degrees of harm by using mipomersen. So, clinically, the studies are not inconsistent, and we did not consider rating down the evidence for heterogeneity.

Considering the baseline control risk of 1.8 in 100 patients (3 divided by 169) in the placebo group to develop increased ALT more than 3 ULN and considering the RR and CI, the absolute risk in the mipomersen group would be about 7 more (0–45 more) patients to develop increased ALT compared with placebo. Because an increased level of ALT for at least 3 times more than ULN can be crucial for patients’ safety, we considered even one more patient having an increase in ALT as the MID, so the CI around the point estimate crosses the MID. This fact and the low event rate (77 in total) lead to degrees of imprecision for −2 levels. Comparing our PICO with the studies’ PICO and also considering that because of few number of studies we could not do a funnel plot, but our extensive search in various databases was comprehensive, so indirectness and publication bias were not any concern in the evaluation of certainty of evidence. So, in total, we considered a low level of evidence for this outcome which means that using mipomersen may result in a large increase in the ALT, more than 3 ULN.

3.8. Major Adverse Cardiac Events

Only one study evaluated MACE, so no meta-analysis was performed for this outcome [24]. In total 206 patients in the mipomersen group and 103 patients in the placebo group provided data for this outcome. Figure 7 shows the related plot. RR for this outcome was 0.67 with CI (0.24–1.87). The certainty of evidence was also assessed for MACE using the GRADE guideline. Risk of bias for all domains of this outcome was low, so the overall risk of bias was considered as low. Heterogeneity was not possible to be evaluated. Comparing the PICO of the study with our PICO did not show a significant difference. So, indirectness was not an issue for this outcome. Publication bias was not an issue because we searched all available databases and other resources to find relevant studies. However total event rate was very low. There were only 14 events in both groups, and risk ratio was 0.67 with a wide CI (0.24–1.87), which crosses the line of no difference considering the already set MID that even one more patient with MACE in the mipomersen group compared with placebo would be important. In absolute effect, as the baseline risk was developing MACE in 6 out of 100 patients in the placebo group, this wide CI means that mipomersen injection can lead to 2 fewer patients (5 fewer to 5 more patients) with MACE in the mipomersen group compared with placebo. So, this shows that the study is not precise for this outcome, and the level of evidence downgraded as −2 for this outcome and considered as low, which means that mipomersen may result in little or no difference in MACE.

3.9. Summary of Findings Table

Table 4 shows the summary of findings for different outcomes.

4.1. Summary of Main Results and Certainty of the Evidence

In patients with FH, subcutaneous injection of mipomersen probably reduces LDL compared with placebo. By pooling data yielded from 523 patients who were included in 6 RCTs, we found that the related evidence suffers from imprecision as the CI of pooled estimate crossed the MID. So, it had a moderate level of certainty. In contrast to the finding about serum LDL, we found that evidence for developing injection site reactions after mipomersen injection was high. Pooled data from five RCTs done on 549 patients showed a high level of evidence that the drug compared with placebo results in a large increase in injection site reaction. Regarding the outcome of increased in serum ALT at least more than 3 times ULN, we pooled the data of four RCTs that included 528 patients. The evidence related to this outcome suffered very severely from imprecision because of the low event rate and that CI crossed the MID. So, we had to downgrade the level of evidence to low, concluding that mipomersen may result in a large increase in the serum ALT.

We could not evaluate the effect of mipomersen on liver toxicity as no RCT had evaluated this outcome in patients with FH. There was only one RCT that reported MACE evaluated on a total of 309 patients. Very low event rate and the wide CI around the RR of comparing this outcome in both groups led us to downgrade the level of evidence because of imprecision to low and concluding that mipomersen may result in little to no difference in MACE.

4.2. Overall Completeness and Applicability of Evidence

Our review included RCTs that were done on adults. No RCT was found that compared mipomersen with placebo in children. Even the only study whose inclusion criteria was patients aged more than 12 years finally reported that the mean age of the included patients was 30 years [11]. So our reported evidence may not be applicable to children with FH. Secondly, all but one included RCTs in the meta-analysis had been done on patients with heterozygous FH. As this type of disease usually presents with milder signs and symptoms and lower serum levels of LDL, evidence reported here may not fully applicable to patients with homozygous FH.

4.3. Potential Biases in the Review Process

We tried to decrease the risk of bias in this review to the minimum by searching the related articles in all major databases. We did not limit our search to any specific language or region. All processes of article selection, screening, and assessment of the risk of bias were done by two reviewers to decrease the risk of bias. We found some abstracts presented in various congresses and tried to contact the authors for receiving the related data, but we received no response. This was not surprising because many of the research done in this field goes back to about 5–10 years ago, and the email addresses might change during this period. One of the limitations of this review might be the low number of studies done exclusively on patients with FH, which leads to a low event rate for some outcomes hence to imprecision while evaluating the level of evidence. Another limitation is that we could not evaluate the effect of mipomersen on the pediatric age group because we found no RCT done in this age group. However, the strength of our study is that this is a unique systematic review that pooled the data of RCTs exclusively done on patients with FH.

4.4. Agreements and Disagreements with Other Studies or Reviews

We could not find any systematic review on RCTs done on patients with FH to compare. Fogacci et al. [49] performed a systematic review on studies that were on the effects of mipomersen on lipoproteins; however, the studies they selected had various populations with various controlled designs. They concluded that there might be favorable effects for mipomersen, but in agreement with what we found, there were also huge concerns for safety. However, in reviewing studies done on volunteers and patients with hypercholesterolemia, including those with FH, Riccota et al. reported high effectivity and safety of mipomersen either as monotherapy or combined with statin therapy, which is in contrast to what we found [37]. In reviewing 4 phase III trials, Santos et al. reported 382 patients with different types of hypercholesterolemia, including two groups with FH who received mipomersen. They reported consistent effectiveness of the drug, but they did not approach the possible side effects [39].

5.1. Implications for Practice

Moderate certainty in the evidence of decreasing serum LDL by injection of mipomersen indicates that we are uncertain if this drug provides a clinically worthwhile effect in this regard.

Although the low level of evidence for increased serum level of ALT produces uncertainty about this side effect, the high level of evidence for injection site reactions as an important side effect of mipomersen that can affect the tolerability of patients receiving the drug can be an important concern. PCSK9 inhibitors, another add-on therapy for hypercholesterolemia, show similar injection site reactions as mipomersen [9]. Mipomersen, is not, however, associated with some adverse effects of PCSK9 inhibitors including hypersensitivity reaction and rare occurrence of leukocytoclastic vasculitis [50].

Lack of evidence for liver toxicity and low level of evidence for MACE may increase the ambiguities around the use of mipomersen.

5.2. Implications for Research

As there were no trials on children to be pooled, further research on the pediatric group with FH may affect the estimates of this review.

Future trials should preferably include more groups of patients with homozygous FH.