| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Guest Access | Sign In via User Name/Password
|
|
|
|||||||
Guest Access | Sign In via User Name/Password |
|||||||||
Division of Pulmonary, Critical Care, and Sleep Medicine, Health Science Center, Memphis, TN
Correspondence to: Gianfranco Umberto Meduri, MD, FCCP, Division of Pulmonary, Critical Care, and Sleep Medicine, Health Science Center 956 Court Ave, Room H316, Memphis, TN 38163; e-mail: umeduri{at}utmem.edu
To the Editor:
The perceptive observations outlined in the five letters provide the opportunity to do the following: (1) introduce additional data showing how methylprednisolone treatment improved organ dysfunction irrespective of underlying shock; (2) clarify the impact of open-label methylprednisolone treatment on data interpretation; (3) expand on the relationship between biological response and reduction in lung injury score (LIS); (4) explain the use of a 2:1 randomization; (5) review the consistent positive findings of five randomized trials in acute lung injury (ALI); and (6) correct the mistaken reporting of mortality in the text.
Shock at Baseline and Primary Outcome
The Web repository1 of the article included the results of a series of stepwise logistic regression analyses adjusting for the effect of confounders and baseline differences (including pressors) on the primary variable of the study. When controlling for confounders and baseline differences, treatment effect remained significant (p = 0.002). Among patients with and without shock, improvement in the primary variable was observed: methylprednisolone vs placebo, 67% vs 23% (p = 0.03) and 71% vs 47% (p = 0.09), respectively.
Although intention-to-treat analyses are standard for large randomized trials, smaller phase 2 trials may be biased by protocol violations or withdrawals, and a "per-protocol" analysis is recommended to reflect scientific methods of the protocol.2 In the per-protocol analysis of this study (n = 79),1 the blinded data safety monitoring board removed an equal proportion of patients in each group, including eight patients with shock (reported in the article). The findings of the per-protocol analysis are shown in Tables 1, 2 . At study entry, the two groups had similar proportions of patients with shock, and by day 7 the treated group had twice the proportion of patients alive and improved (87% vs 42%; relative risk [RR], 2.1; 95% confidence interval [CI], 1.3 to 3.4; p < 0.001).1 The significant findings observed by day 7 in the intention-to-treat and per-protocol analysis were similar, increasing confidence in the trial results.2 In conclusion, the positive effects of treatment were not significantly affected by the presence of shock at study entry or undertreatment of this subgroup.
|
|
The statement that open-label treatment obscured detection of harm caused by steroid (Fremont and Rice3) is not supported by the findings of our study or the literature.145 In our study, open-label treatment was not associated with deterioration in lung function, while a 1-point reduction in LIS was observed in 60% of patients. One patient had neuromuscular weakness, improved LIS, was extubated on ARDS day 27, and survived hospital admission. Infections developing before, during, and after open-label methylprednisolone were amply reported in the article.2
Similar to our original study,4 the ARDS Network Trial5 found that methylprednisolone treatment of unresolving ARDS was associated with a significant improvement in lung mechanics and PaO2/fraction of inspired oxygen (FIO2) and a significant reduction in plasma interleukin-6, BAL neutrophilia, duration of mechanical ventilation, and ICU stay. Combining the survival data from the three trials (Fig 1 )145 for patients randomized before day 14 (n = 245), methylprednisolone treatment significantly reduced mortality (35 of 144 patients [24%] vs 40 of 101 patients [39%]; RR, 0.62; 95% CI, 0.43 to 0.90; p = 0.01). In these three trials,145 methylprednisolone treatment increased ventilation-free days at day 28 by 5.6 days (95% CI, 3.5 to 7.7; p < 0.0001). These data suggest that the late introduction (day 7 to day 9) of open-label methylprednisolone treatment in patients with unresolving ARDS likely favored the control group and decreased the effect size (in comparison to control subjects) observed after day 7 in those randomized to methylprednisolone.
|
Multiple studies (n = 15, including the original trial by Bernard and collaborators6; reviewed by Meduri7) have shown that improvement in LIS or its components by day 7 of ARDS correlates with improved survival. Reversal of hypoxemia (PaO2/FIO2 ratio > 300) by day 7 was associated with significantly lower mortality rates (43% vs 97%, p < 0.001).5 Similarly, we reported that improvement in LIS by day 7 correlated with improved hospital survival (R = 0.59; p < 0.001).1 Even the cited ARDS Network Trial8 reported that low tidal volume ventilation led to a reduction in systemic inflammation and improvement in PaO2/FIO2 ratio by day 7 despite a lower PaO2/FIO2 value at baseline.
Activation of the glucocorticoid receptor by endogenous or exogenous glucocorticoids is the most important down-regulator of inflammation. The significant physiologic and laboratory improvement in pulmonary and extrapulmonary organ function observed by study day 7 reflects the positive modulatory effects of prolonged methylprednisolone administration on nuclear factor-
B–mediated pathways (inflammation, coagulation, and tissue repair) that affect histologic and physiologic changes at the tissue level.9 Translational research has shown that prolonged methylprednisolone in ARDS is associated with a significant reduction in pulmonary and circulating levels of markers of inflammation and fibrogenesis not shown by any other intervention.10 Consequently, the accelerated resolution of respiratory failure observed during prolonged methylprednisolone treatment of early ARDS led to a significant increase in mechanical ventilation-free days by day 28 (16.5 ± 10.1 days vs 8.7 ± 10.2 days, p = 0.001),1 which is similar to those previously reported in unresolving ARDS by our group (16 ± 2 days vs 6 ± 2 days, p = 0.005)4 and the ARDS Network (11.2 ± 9.4 days vs 6.8 ± 8.5 days, p < 0.001).5 This effect is far greater than the one observed with the recommended low tidal volume ventilation8 or conservative strategy of fluid management.11
Rationale for Using a 2:1 Randomization
Unequal randomization is useful when trying to gain knowledge about the response to a new drug or, in this case, a new application for an old drug. In an efficacy study with a 2:1 randomization scheme, the higher number of subjects exposed to treatment may result in a narrower CI and improve the precision of the treatment effect. This design requires more patients to achieve the same power and effect as would be necessary with a 1:1 randomization design, and therefore does not bias the study toward methylprednisolone.
Risks/Benefits of Prolonged Glucocorticoid Treatment: Results of Five Randomized Trials
Five randomized trials (n = 518) have investigated prolonged glucocorticoid treatment (hydrocortisone, 200 to 400 mg/d; methylprednisolone, 1 mg/kg/d) in early ALI (PaO2/FIO2 < 300),12 early ARDS (PaO2/FIO2 < 200),113 and unresoloving ARDS (methylprednisolone, 2 mg/kg/d).45 These trials consistently reported that treatment was associated with significant improvement in PaO2/FIO2,1451213 and a significant reduction in markers of systemic inflammation,1451213 BAL neutrophilia,514 duration of mechanical ventilation,1451213 and ICU stay.14512
For decades, intensivists have routinely administered higher daily doses of methylprednisolone for acute exacerbation of asthma or COPD and carefully balanced the important antiinflammatory effect of treatment with potential complications. The most reliable data on actual complication rates associated with prolonged glucocorticoid treatment (in low-to-moderate doses) is from randomized trials in sepsis1516 and ARDS.1451213 Similar to findings in septic shock,1516 glucocorticoid treatment in ARDS was not associated with increased rates of GI bleeding or nosocomial infections.1451213 Glycemic spikes observed when glucocorticoids are administered as a bolus,5 are mitigated when they are administered as an infusion.117 Finally, systemic inflammation-associated neuromuscular weakness is uncommon with low-dose steroids if concomitant paralysis is avoided.112 Neuromuscular weakness is a known independent predictor of prolonged weaning,18 yet the ARDS Network trial reported that among the 43 patients with weakness, those randomized to methylprednisolone (n = 25) had a significant (p = 0.003) and sizeable (11 days) reduction in duration of mechanical ventilation (Table 7, supplemental appendix).5 Finally, when glucocorticoid treatment is initiated before day 14 of ARDS (n = 468),1451213 mortality is significantly decreased (84 of 252 patients [33%] vs 108 of 216 patients [50%]; RR, 0.71; 95% CI, 0.50 to 0.87; p = 0.001), with 6 patients needed to treat to save one life.
Incorrect Reporting of Mortality
We express regret for an error in reporting "mortality" that has unfortunately generated misunderstanding (Segel19). The results for survival reported in the Tables and Figures are correct. In the text, however, the word mortality was used incorrectly in place of survival four times.
The text on page 956 should read, "Survival by day 7 for patients with catecholamine-dependent shock was similar (80% vs 76.9%)." Twelve of 15 patients (80%) and 10 of 13 patients (76.9%) with shock survived to day 7 in the treated and control groups, respectively. The text on page 957 should read, "ICU survival for patients with catecholamine-dependent shock was 73% vs 46% (p = 0.24), and for patients without shock was 81% vs 67% (p = 0.29). In per-protocol analysis (Web repository), ICU survival for patients with catecholamine-dependent shock was 90% vs 71% (p = 0.07)." The number of patients with shock who survived ICU admission was 11 of 15 patients (73%) and 6 of 13 patients (46%) in the treated and control groups, respectively. The text on page 958 should read, "Survival rates at 2,6, and 12 months were 76% vs 61% (p = 0.13), 67% vs 46% (p = 0.07), and 63.5% vs 46%, respectively."
In conclusion, methylprednisolone treatment has a strong benefit/risk ratio when it is applied in conjunction with measures demonstrated to reduce morbidity associated with glucocorticoids.14 These measures include the following: (1) intensive infections surveillance, (2) avoidance of paralytic agents, and (3) avoidance of rebound inflammation with premature discontinuation of treatment that may lead to physiologic deterioration and reintubation. Correct use of this inexpensive and highly effective antiinflammatory therapy is associated with a significant improvement in patient-centered outcome variables, irrespective of shock, and decreases health-care cost associated with ARDS.20 The relevance of these findings to public health and health-care economics urges investment in clinical investigation of this inexpensive and highly effective antiinflammatory therapy.
Acknowledgements
"There is no illumination in speculation" as used in the title of this letter is quoted with permission from Rev. Kenneth V. Leeper Sr.
Footnotes
The author has no conflict of interest to disclose.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
