Systematic review of statins in sepsis: There is no evidence of dose response.

OBJECTIVES: Sepsis is a common cause of morbidity and mortality and is associated with significant costs to the healthcare organizations. We performed a systematic review and meta-analysis to assess whether high or low-dose statin therapy improved mortality in patients with sepsis.
METHODS: The trials analyzed in this study were multicenter or single center randomized control studies using statins for sepsis in a hospital setting. The patients included were adults with suspected or confirmed infection.
INTERVENTIONS: This study found eight randomized controlled trials where participants were given either a statin or placebo daily for 14-28 days, the duration of their illness, or until their death or discharge, which ever occurred first. PRIMARY AND SECONDARY OUTCOMES MEASURED: This meta-analysis measured the effect of statin therapy on in hospital and 28 days mortality.
RESULTS: In unselected patients, there was no demonstrable difference in the 28 days mortality (relative risk [RR] 0.88 95% confidence interval [CI], 0.70-1.12 and P = 0.16). There was also no significant difference between statin versus placebo for in-hospital mortality (RR 0.98 95% CI, 0.85-1.14 P = 0.36). When the studies where divided into low-dose and high-dose groups, there were no statistically significant differences for in-hospital mortality between low-dose statin versus placebo for (RR 0.81 CI 0.44-1.49 P = 0.27) or high-dose statin versus placebo (RR 0.99 95% CI 0.85-1.16, P = 0.28). There was no significant difference in adverse effects between the high- and low-dose groups.
CONCLUSIONS: In this meta-analysis, we found that the use of statins did not significantly improve either in-hospital mortality or 28-day mortality in patients with sepsis. In the low-dose group, there were fewer quality multicenter studies; hence, conclusions based on the results of this subgroup are limited.

Introduction

Sepsis is characterized by an exaggerated inflammatory response to infection, resulting in a multisystem physiological, cellular, and organ dysfunction. In a European study, a typical episode of severe sepsis costs the health care organization approximately €25,000.[1] Assuming that we see 100,000 cases of severe sepsis per annum, this equates to a direct current cost to the national health services of England of over 2.5 billion every year. In the USA, the Centers for Disease Control and Prevention's National Center for Health Statistics estimates the number of hospital admissions attributed to sepsis increased from 621,000 in the year 2000 to 1,141,000 in 2008.[2] In 2011, sepsis resulted in an aggregate healthcare cost of $20.3 billion making it the most expensive condition treated in US hospitals.[3]

Mortality from sepsis has improved over the last decade but is still estimated to be 36% in Europe.[1] This warrants a search for novel therapeutic targets and preventative therapies in patient with sepsis. Statins are lipid-lowering drugs that inhibit 3-hydroxy-3 methyglutaryl coenzyme A reductase, with pleotropic mechanisms that may be beneficial in sepsis. They are hypothesized to possess a variety of benefits including anti-inflammatory, immunomodulatory, antioxidant, and antithrombotic effects.[45] In previous studies, statins have been shown to reduce the unregulated immune response, influence the gene transcription, and reduce the expression of mRNA to sepsis. The use of statins was associated with an improved mortality and morbidity following sepsis.[67] In animal studies, statins reduced the severity of sepsis.[89] It is also thought that statins exert a protective effect through down-regulation of toll-like receptor 4, inhibition of nuclear factor-kappa-B and protection against endothelial cell apoptosis.[1011] Despite the initial promise from animal data and early clinical trials, these benefits have not been reproduced in the large, well-designed, randomized controlled trials (RCTs) published recently.[1213]

This paper aims to assess the impact of both high and low dose statin therapy on in-hospital and 28-day mortality in patients with sepsis.

Methods

Search strategy for published studies with aggregate data

Electronic databases, including PubMed and EMBASE, were searched using a combination of keywords: “sepsis,” “intensive care” “statin” “simvastatin,” “atorvastatin,” and “rosuvastatin” to create a list of articles published before October 2015. The search was limited to articles published in the English language.

Study selection criteria

Articles were included if the study population was adult patients with a suspected or confirmed infection and reported 28 day or in-hospital mortality reported as a primary or secondary outcome. A total of 215 articles were identified. The following were excluded; 160 were not clinical trials, 2 post hoc analyses, 2 cell-based studies, 3 were not randomized control trials, and 31 were the wrong patient population. A total of 15 studies were fully assessed for eligibility. Of these, 3 were excluded as mortality was not reported as a primary or secondary outcome, and 4 were excluded as they were the wrong patient populations: 3 were postcardiac surgery patient populations and 1 was neurosurgical patients. Eight studies were included in the meta-analysis [Figure 1].

Figure 1

Studies included in this meta-analysis

Data extraction and statistical analysis

Data were extracted independently by two authors and analyzed using Review Manager (RevMan) [Computer program]. Version 5.3.[14] Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014. For each study, the characteristics of the study, number of participants, characteristics of included patients, selection criteria, drug and dose used, and outcomes observed in the study were extracted. Most studies reported in-hospital mortality and 28 days mortality outcomes. Some studies reported 28 days mortality only and others reporting in-hospital mortality only. To include all RCTs in this meta-analysis, the authors have performed a separate analyses on both 28 days mortality and in-hospital mortality for this reason. All studies had adequate randomization and blinding. The quality of the RCTs was evaluated using the method described in the Cochrane Handbook for Systematic Reviews of Interventions.[15] In each study, patients were given either a dose of statin (simvastatin 20/40/80 mg, rosuvastatin 40 mg or atorvastatin 20 mg) or a placebo. Low-dose statin was defined as simvastatin 20 mg or atorvastatin 20 mg. High dose was defined as simvastatin 40 mg or 80 mg or rosuvastatin 40 mg.[1617]

Results

We examined a total of 8 RCTs with a total of 2275 patients. Figures 2 and 3 show forest plots of 28-day mortality and in-hospital mortality, respectively. Table 1 shows the 8 RCTs and the characteristics of each study. Table 2 shows assessment of bias for each study. All studies used in the meta-analysis are listed in Appendix 1.

Figure 2

Forest plot of proportional effect on 28-day mortality. Forest plot of proportional effect of all studies (low and high doses) on 28-day mortality (Diamonds = totals and subtotals [95% confidence interval]. Squares = individual studies [horizontal lines are 95% confidence intervals]. Area of square proportional to amount of statistical information in that category. Risk ratios are weighted to represent heterogeneity)

Figure 3

Forest plot of proportional effect for in-hospital mortality of individual studies. Forest plot of proportional effect on in-hospital mortality (Diamonds = totals and subtotals [95% confidence interval]. Squares = individual studies [horizontal lines are 95% confidence intervals]. Area of square proportional to amount of statistical information in that category. Risk ratios are weighted to represent heterogeneity)

Table 1

Characteristics of included studies

Table 2

Assessment of risk of bias

All dose statin, 28-day mortality

Results from 1171 patients demonstrated that there was no significant difference between statin (102/576) and placebo (120/595) for 28 day mortality (relative risk [RR] 0.88 95% confidence interval [CI], 0.70–1.12 and P = 0.16), with a heterogeneity of the trials I2 = 43%, P = 0.16 [Figure 2].

All dose statin, in-hospital mortality

Results from 2175 patients demonstrated that there was no significant difference between statin (259/1086) versus placebo (265/1089) for in-hospital mortality (RR 0.98 95% CI, 0.85–1.14 P = 0.36) with low heterogeneity between the studies I2 = 9% P = 0.36. The authors excluded 83 patients from the Novack 2009 study from the analysis, as both arms of the trial had zero events [Figure 3].[18]

Low dose statin, in-hospital mortality

Results from 400 patients demonstrated no statistically significant difference between low-dosage statin use (22/198) versus placebo (27/202) for in-hospital mortality [RR 0.81 CI (0.44–1.49) P = 0.27, I2 = 16% Figure 4].

Figure 4

Forest plot of proportional effect for low dose statin, in-hospital mortality. Forest plot of proportional effect on in-hospital mortality (Diamonds = totals and subtotals [95% confidence interval]. Squares = individual studies [horizontal lines are 95% confidence intervals]. Area of square proportional to amount of statistical information in that category. Risk ratios are weighted to represent heterogeneity)

High dose statin, in-hospital mortality

Results from 1692 patients demonstrated that there was no significance between high-dose statin usage (237/846) versus placebo (239/846) for in-hospital mortality [RR 0.99 95% CI 0.85–1.16 P = 0.28, I2 = 21%, Figure 5].

Figure 5

Forest plot of proportional effect for high dose statin, in-hospital mortality. Forest plot of proportional effect on in-hospital mortality (Diamonds = totals and subtotals [95% confidence interval]. Squares = individual studies [horizontal lines are 95% confidence intervals]. Area of square proportional to amount of statistical information in that category. Risk ratios are weighted to represent heterogeneity)

Adverse effects in high dose and low dose groups

From all reported trials, patients with sepsis who received statins did not have a significantly higher incidence of adverse effects compared to placebo. In patients who were treated with high-dose statins, the incidence of adverse effects was higher (11.6% vs. 8.5%, P > 0.05), but this failed to reach statistical significance.

Discussion

This meta-analysis demonstrated that statin therapy does not reduce in-hospital or 28 days mortality in sepsis when compared to placebo. This meta-analysis also failed to show any significant difference in outcome of treating patients with low or high dose statins.

Statins were shown to be beneficial in animal model studies,[6] and early RCTs demonstrated benefits in critically ill patients with sepsis, however these studies were single center RCTs with a relatively small sample size.[19] In this study, we investigated whether the dose of statin was associated with adverse events. Statin therapy is associated with musculoskeletal side effects including myopathy, myositis, and rhabdomyolysis. Myalgia and arthralgia are reported in up to 5% of participants in clinical trials.[202122] Reported incidence of dose dependent adverse events such as deranged hepatic transaminases was 0.5–2%.[1620] In one study, there were a significant number of adverse events related to the intervention arm when using 80 mg of simvastatin.[23] However, across all studies in this analysis, the total incidence of adverse events in patients who were treated with high dose statins was not significantly different compared to placebo or to low dose statins.

In this up to date meta-analysis, we focused only on eligible RCTs to eliminate the risk of confounding variables often seen in observation studies. Compared to previous meta-analyses, we have not only included up to date RCTs but also tried to delineate the role of high and low-dose statins.[24] These findings are in contradiction to previous literature reviews and observational studies.[725] As our meta-analysis relies purely on RCTs, this limits the study to a relatively small sample size. This is particularly true for the low-dose statin subgroup of patients.

While preclinical and observational studies hypothesized the potential benefits of statin usage in sepsis, our meta-analysis contradicts this presumption. Although the beneficial effects in vitro and animal models have been well documented in literature,[611] these effects do not confer a detectable benefit in large human model RCTs. This could be due to the heterogeneity of the populations being studied; there is an inter subject variability in the host response due to age, existing comorbidities and genetic profile. Genes involved in sepsis have been analyzed for links between single-nucleoside polymorphisms and sepsis susceptibility, organ dysfunction and mortality.[26] Despite the heterogeneity of the populations being studied, there are no suggestions from this study that a specific population group would benefit from statin therapy.

Conclusions

This meta-analysis found no beneficial effect of statin therapy in the context of sepsis and the mortality of critically ill patients. Our findings are in contradiction to previous reviews of the literature and also to those of observational studies.[71925]

Financial support and sponsorship

Dr. Veenith is supported by NIAA clinical research training fellowship and Beverly Sackler Studentship.

Conflicts of interest

There are no conflicts of interest.