Statin Effect on Sepsis Mortality

Can We Close the Book?
      FOR RELATED ARTICLE, SEE PAGE 805
      HMG-CoA (3-hydroxy-3-methylglutaryl-CoA) reductase inhibitors, commonly known as statins, have become some of the most widely used and best-selling pharmaceutical agents in history.
      • Jackevicius C.A.
      • Chou M.M.
      • Ross J.S.
      • Shah N.D.
      • Krumholz H.M.
      Generic atorvastatin and health care costs.
      One-half of American men and more than one-third of American women in the 65- to 74-year age group are prescribed statins.

      Centers for Disease Control and Prevention. Health, United States, 2010. http://www.cdc.gov/nchs/data/hus/hus10.pdf. Accessed October 15, 2017.

      Initially used for their lipid-lowering effects in coronary vascular disease, they are now known to have additional properties that could be of use in other disease processes. Such properties include effects on both inflammatory and anti-inflammatory cytokines, inducible nitric oxide synthase, and leukocyte adhesion and rolling.
      • Mermis J.D.
      • Simpson S.Q.
      HMG-CoA Reductase inhibitors for prevention and treatment of severe sepsis.
      These properties have attracted attention for the treatment of numerous diseases with inflammatory underpinnings, such as autoimmune disorders, multiple sclerosis, COPD, ARDS, and sepsis.
      • Smeeth L.
      • Douglas I.
      • Hall A.J.
      • Hubbard R.
      • Evans S.
      Effect of statins on a wide range of health outcomes: a cohort study validated by comparison with randomized trials.
      In addition, statins have been found to have antibacterial properties, which could be of added benefit to patients with sepsis.
      • Masadeh M.
      • Mhaidat N.
      • Alzoubi K.
      • Al-Azzam S.
      • Alnasser Z.
      Antibacterial activity of statins: a comparative study of atorvastatin, simvastatin, and rosuvastatin.
      Numerous studies have evaluated whether long-term statin use reduces mortality among patients in whom sepsis develops while they are taking statins. To date, the studies have been retrospective and observational, and they have largely been hampered by an inability to deal with certain important confounders, most important among them the healthy user effect.
      • Yende S.
      • Milbrandt E.B.
      • Kellum J.A.
      • et al.
      Understanding the potential role of statins in pneumonia and sepsis.
      Because statins are primarily used as preventive therapy for acute coronary artery events, they are taken by patients who have both the means and the will to engage in disease prevention and are not taken by patients who cannot afford them or who are less healthy or less concerned with disease prevention. It would, therefore, be expected that patients who are taking statins are on average healthier at the time that sepsis develops, with a lower predicted mortality—the healthy user effect.
      • Smeeth L.
      • Douglas I.
      • Hall A.J.
      • Hubbard R.
      • Evans S.
      Effect of statins on a wide range of health outcomes: a cohort study validated by comparison with randomized trials.
      An additional uncontrolled factor in these studies has been that it is difficult to tell whether patients who are prescribed statins are actually taking them regularly. Study results, including meta-analyses, have mostly suggested that statin use is associated with decreased sepsis mortality, but the two factors mentioned have left doubts about the reliability of these results.
      • Ouellette D.R.
      • Moscoso E.E.
      • Corrales J.P.
      • Peters M.
      Sepsis outcomes in patients receiving statins prior to hospitalization for sepsis: comparison of in-hospital mortality rates between patients who received atorvastatin and those who received simvastatin.
      • Janda S.
      • Young A.
      • Fitzgerald J.M.
      • Etminan M.
      • Swiston J.
      The effect of statins on mortality from severe infections and sepsis: a systematic review and meta-analysis.
      It would be an enormous undertaking to randomize patients to statins or no statins and track them until sepsis develops. The closest we have to that scenario is randomization of patients to receive statins or not on admission to the hospital with sepsis; studies of this nature have failed to reveal meaningful benefit.
      • Kruger P.S.
      • Harward M.L.
      • Jones M.A.
      • et al.
      Continuation of statin therapy in patients with presumed infection: a randomized controlled trial.
      • Truwit J.D.
      • Bernard G.R.
      • et al.
      National Heart, Lung, and Blood Institute ARDS Clinical Trials Network
      Rosuvastatin for sepsis-associated acute respiratory distress syndrome.
      In this issue of CHEST, Lee et al
      • Lee C.C.
      • Lee M.G.
      • Hsu T.C.
      • et al.
      A population-based cohort study on the drug-specific effect of statins on sepsis outcome.
      report a population-based study of statin use and sepsis outcomes from Taiwan. Using the National Health Insurance (NHI) database, they evaluated mortality among septic patients who were taking or not taking statins at the time they experienced sepsis. They undertook the study to determine whether the various classes of statins exhibited differential effects on sepsis outcome. The investigators used a randomly selected cohort of 1 million patients who were beneficiaries of the NHI in 2000 and tracked them through 2011. Sepsis episodes were identified by determining when hospital admissions included an acute infection associated with acute organ dysfunction according to International Classification of Diseases, ninth revision codes. Patients were followed for readmissions and until death or termination of the study in 2011. The primary outcome for the study was 30-day mortality, with secondary outcomes of 90-day mortality and development of acute respiratory failure. The authors matched statin users with non-statin users using propensity score matching consisting of 61 demographic and health-related variables of interest.
      Over the period of study, the authors identified 52,737 patients with sepsis, 3,598 of whom were taking a statin at the time that sepsis developed. Overall, 30-day sepsis mortality was 17.4% and 90-day mortality was 22.7%. The authors found that statin use in these circumstances was indeed associated with a mortality effect and that it was dependent on the class of statin taken. Using propensity score matching of statin takers and non-statin takers on 61 different variables, the odds ratio for mortality was 0.88 at 30 days and 0.93 at 90 days among statin users. In the propensity-matched analysis, atorvastatin and simvastatin were associated with reduced 30-day mortality, whereas rosuvastatin was not, although the differences among statins were less pronounced at 90 days. The observed associations of the statins with lower mortality were not aligned with their lipid-lowering capacity; atorvastatin and simvastatin were associated with reduced sepsis mortality, compared with rosuvastatin, which has the greatest lipid-lowering effect. However, this association with improved sepsis outcomes is aligned with the in vitro bactericidal capacity of the drugs.
      These results are of interest and represent an improvement over previous studies of the effect of statin use on sepsis mortality for a number of reasons. The ability to randomly sample a population-based cohort and to track patients longitudinally represents a substantial improvement over previous studies, which used what were essentially convenience samples. The healthy user effect was mitigated both by effective propensity-score matching of patients and by comparisons across statin groups. Because the NHI database contains prescription and refill information, the authors were able to determine how often prescriptions were refilled as a marker of whether statins were being taken as prescribed. Despite the retrospective and observational nature of the study, these factors make it the most reliable information to date, and the results likely represent the best sort of information that we will be able to achieve regarding the effect of long-term statin use on sepsis mortality.
      These findings may shed light on some heterogeneity of findings in previous similar studies, as statin type has not always been accounted for in the studies. It is unlikely that prospective randomized trials of statins for prevention of sepsis mortality will ever be undertaken owing to the sheer number of patients who would require randomization to have adequate numbers in whom sepsis actually develops. We believe that the next best thing to randomization and a prospective trial is exactly what the authors have done: identify a cohort, track them through time, even if not concurrently, and match cases to control subjects by propensity matching on important clinical characteristics. Nevertheless, this brings us to one aspect of the study that leaves the door open a crack for some doubt.
      Given that the statin effect is observed in patients receiving long-term therapy, that it has not been demonstrated when statins are initiated in the hospital, and that the effect appears to be associated with the in vitro bactericidal effect of the individual statin, one would expect that statins could also prevent the development of sepsis in individuals at risk. Unfortunately, Lee et al
      • Lee C.C.
      • Lee M.G.
      • Hsu T.C.
      • et al.
      A population-based cohort study on the drug-specific effect of statins on sepsis outcome.
      have not given us the data to determine whether that is the case. We know how many septic patients were taking statins but not how many total statin takers were in the cohort, so the incidence in statin users vs non-statin users cannot be calculated. Nevertheless, this may be an important hypothesis for future analysis using the current patient cohort. Although Lee et al
      • Lee C.C.
      • Lee M.G.
      • Hsu T.C.
      • et al.
      A population-based cohort study on the drug-specific effect of statins on sepsis outcome.
      have provided us with data of the highest quality that we can likely hope for, the book may not be quite closed—yet.

      References

        • Jackevicius C.A.
        • Chou M.M.
        • Ross J.S.
        • Shah N.D.
        • Krumholz H.M.
        Generic atorvastatin and health care costs.
        N Engl J Med. 2012; 366: 201-204
      1. Centers for Disease Control and Prevention. Health, United States, 2010. http://www.cdc.gov/nchs/data/hus/hus10.pdf. Accessed October 15, 2017.

        • Mermis J.D.
        • Simpson S.Q.
        HMG-CoA Reductase inhibitors for prevention and treatment of severe sepsis.
        Curr Infect Dis Rep. 2012; 14: 484-492
        • Smeeth L.
        • Douglas I.
        • Hall A.J.
        • Hubbard R.
        • Evans S.
        Effect of statins on a wide range of health outcomes: a cohort study validated by comparison with randomized trials.
        Br J Clin Pharmacol. 2009; 67: 99-109
        • Masadeh M.
        • Mhaidat N.
        • Alzoubi K.
        • Al-Azzam S.
        • Alnasser Z.
        Antibacterial activity of statins: a comparative study of atorvastatin, simvastatin, and rosuvastatin.
        Ann Clin Microbiol Antimicrob. 2012; 11: 13
        • Yende S.
        • Milbrandt E.B.
        • Kellum J.A.
        • et al.
        Understanding the potential role of statins in pneumonia and sepsis.
        Crit Care Med. 2011; 39: 1871-1878
        • Ouellette D.R.
        • Moscoso E.E.
        • Corrales J.P.
        • Peters M.
        Sepsis outcomes in patients receiving statins prior to hospitalization for sepsis: comparison of in-hospital mortality rates between patients who received atorvastatin and those who received simvastatin.
        Ann Intensive Care. 2015; 5: 9
        • Janda S.
        • Young A.
        • Fitzgerald J.M.
        • Etminan M.
        • Swiston J.
        The effect of statins on mortality from severe infections and sepsis: a systematic review and meta-analysis.
        J Crit Care. 2010; 25 (e7-22): 656
        • Kruger P.S.
        • Harward M.L.
        • Jones M.A.
        • et al.
        Continuation of statin therapy in patients with presumed infection: a randomized controlled trial.
        Am J Respir Crit Care Med. 2011; 183: 774-781
        • Truwit J.D.
        • Bernard G.R.
        • et al.
        • National Heart, Lung, and Blood Institute ARDS Clinical Trials Network
        Rosuvastatin for sepsis-associated acute respiratory distress syndrome.
        N Engl J Med. 2014; 370: 2191-2200
        • Lee C.C.
        • Lee M.G.
        • Hsu T.C.
        • et al.
        A population-based cohort study on the drug-specific effect of statins on sepsis outcome.
        Chest. 2018; 153: 805-815

      Linked Article