CORRESPONDENCE
end point of 20 percentage points (number need certainties about effectiveness of the procedures ed to treat, 5). No other treatment for angina regarding blinding.4 (including coronary-artery bypass grafting) has Johann Auer, M .D. achieved a benefit of such magnitude. Until a St. Josef Hospital Braunau larger trial is performed, I would not be so en Braunau, Austria thusiastic about the main findings of this study. [email protected] M arco A.S. Cordeiro, M .D., Ph.D.
Robert Berent, M.D.
Hospital Evangglico Goiano An^polis, Brazil
Herzreha Bad IscHI Bad IscHI, Austria
[email protected] No potential conflict o f interest relevant to this letter was re ported.
No potential conflict of interest relevant to this letter was re ported.
1. Poldermans D, Boersma E, Bax JJ, et al. The effect of bisoprolol on perioperative mortality and myocardial infarction in high-risk patients undergoing vascular surgery. N Engl J Med 1999;341:1789-94. 2. The Hypothermia after Cardiac Arrest Study Group. Mild therapeutic hypothermia to improve the neurologic outcome after cardiac arrest. N Engl J Med 2002;346:549-56. [Erratum, N Engl J Med 2002;346:1756J 3. The POISE Study Group. Effects of extended-release metoprolol succinate in patients undergoing non-cardiac surgery (POISE trial): a randomised controlled trial. Lancet 2008;371: 1839-47. 4. Nielsen N, Wetterslev J, Cronberg T, et al. Targeted tempera ture management at 33°C versus 36°C after cardiac arrest. N Engl J Med 2013;369:2197-206. DOI: 10.1056/NEJMcl503672
Verheye and colleagues investi gated the antianginal effects of a coronary-sinus reducing device. The authors conclude that nar rowing of the coronary sinus was associated with significant improvement in symptoms. However, limitations in the design and conduct of this trial can lead to inaccurate conclusions regarding ef ficacy. First, despite instructing the performing physicians to behave similarly during both the active procedure and the sham procedure, and despite offering either headsets playing music or sedation to mask the conversation, the effective ness of blinding might have been limited. In ad dition, intravenous heparin was used only in pa tients assigned to the treatment group; therefore, postprocedural laboratory testing might have un veiled the study assignment of the patients. Fur thermore, the viewpoints of staff members and investigators, who were not initially aware o f the study assignments, might have been compromised by the physician performing the procedure, who was aware o f the study assignment.1'3 Finally, the choice of improvement of angina class as the pri mary end point rather than more objective varia bles such as exercise time or wall motion during stress makes the results more vulnerable to un t o th e e d it o r :
N ENG LJ
M ED 372120
1. Sackett DL. Commentary — measuring the success of blind ing in RCTs: don’t, must, can’t or needn’t? Int J Epidemiol 2007; 36:664-5. 2. Schulz KF, Grimes DA. Blinding in randomised trials: hiding who got what. Lancet 2002;359:696-700. 3. Bhatt DL, Kandzari DE, O’Neill WW, et al. A controlled trial of renal denervation for resistant hypertension. N Engl J Med 2014;370:1393-401. 4. Kandzari DE, Lam LC, Eisenstein EL, et al. Advanced coro nary artery disease: appropriate end points for trials of novel therapies. Am Heart J 2001;142:843-51. DOI: 10.1056/NEJMC1503672
r e p ly : In the COSIRA trial, less than 3% of the patients were excluded because of unsuitable coronary-sinus anatomy for device implantation. The main anatomical reason for excluding patients was a large diameter of the sinus. The proximity o f the circumflex artery is not a problem since the device is implanted in a proximal segment of the coronary sinus. Diffi culties in advancing the catheter into the coro nary sinus owing to the presence of valves are usually overcome easily with the use of different types of guiding wires. If necessary, it is easy to dilate the narrowed central segment of the stainless steel reducing device with a simple balloon-dilatation technique. This approach may allow future passage of a coronary-sinus lead for CRT. Efficacy findings in phase 2 studies must be interpreted with caution because o f the small number o f patients. To obtain approval for therapies that improve symptoms (such as an gina), regulatory agencies such as the Food and Drug Administration typically require two posi tive trials with P values less than 0.05.1 We therefore agree with Cordeiro that the COSIRA trial may have failed to reject a true null hypoth esis and that a confirmatory trial is needed. The lack of approaches to assess efficacy in small trials performed in early phases o f development programs for devices is a considerable problem. the authors
N E J M .O R G
MAY 14 , 2015
1967
The N E W E N G L A N D J O U R N A L o / M E D I C I N E
Corroborative analyses are one possible solution.2 We conducted a post hoc corroborative efficacy analysis that combined the primary and second ary end points. We found that the device was consistently associated with improved symptoms, function, and quality of life, as compared with a sham intervention. Coronary-sinus reduction is a device-based therapy for improvement of symptoms in patients with severe angina who had exhausted conven tional therapies. Therefore, the primary end point of the COSIRA trial aimed to answer the question of whether the device is an effective treatment to improve symptoms of angina. Placebo treatment alone can result in sub stantial improvement in angina symptoms and in exercise duration. Our sham-controlled clini cal trial was designed to eliminate both patient bias and investigator bias in the interpretation of outcome end points. Extensive efforts were made so that the patients and the physician perform ing the follow-up were unaware of the treatment assignment. Implantation of the coronary-sinus reducing
device is a simple procedure, and there was not much difference in the length of the procedure between the sham group and the treatment group. Also, there was no difference in the post procedural care between the two groups. Shmuel Banai, M.D. Tel Aviv Medical Center Tel Aviv, Israel [email protected]
Stefan Verheye, M.D. Ph.D. Antwerp Cardiovascular Center Antwerp, Belgium
E. Marcjolicceur, M.D. Montreal Heart Institute Montreal, QC, Canada
Since publication of their article, the authors report no fur ther potential conflict o f interest. 1. Jolicoeur EM, Ohman EM, Temple R, et al. Clinical and re search issues regarding chronic advanced coronary artery dis ease part II: trial design, outcomes, and regulatory issues. Am Heart J 2008;155:435-44. 2. Hare JM, Bolli R, Cooke JP, et al. Phase II clinical research design in cardiology: learning the right lessons too well: observa tions and recommendations from the Cardiovascular Cell Ther apy Research Network (CCTRN). Circulation 2013;127:1630-5. DOI: 10.1056/NEJMcl503672
A c id - B a s e P r o b le m s in D ia b e tic K e to a c id o s is t h e e d i t o r : In their review article, Kamel and Halperin (Feb. 5 issue)1 discuss “delta-delta,” the increase in the plasma anion gap and the decrease in the plasma bicarbonate concentra tion, for detecting the presence o f coexisting acid-base disorders. The ratio is predicated on the assumption that in simple acidosis with an increased anion gap, the gap will increase by the same amount that the bicarbonate decreases. However, many laboratories currently measure electrolytes with the use of ion-selective elec trodes, and the range for the anion gap is usually given as approximately 3 to 11. For the bicarbon ate level, the laboratory in which I work provides a reference range of 21 to 31. So what does one use as the baseline value for each measurement? We typically do not have samples obtained from patients just before the onset of the acid-base disorder, so how does one know what baseline to use in calculating how much deviation has oc curred? For the anion gap, is it 3 or 11 or in be tween, and for the bicarbonate concentration, is it 21 or 31 or in between? This ambiguity leads
to
1968
N
EN G LJ
M ED 372120
me to question the value of “delta-delta” in assess ing acid-base problems. Markjoy, M.D.J.D. VA Pittsburgh Healthcare System Pittsburgh, PA [email protected]
No potential conflict of interest relevant to this letter was re ported. 1. Kamel KS, Halperin ML. Acid-base problems in diabetic keto acidosis. N Engl J Med 2015;372:546-54. DOI: 10.1056/NEJMcl502745 t h e e d i t o r : Kamel and Halperin state that sodium bicarbonate infusion may be considered in some patients with diabetic ketoacidosis to avoid possible deterioration in the hemodynamic status. This debated issue is supported by some older experimental studies, which suggest that severe acidosis might be associated with de creased cardiac contractility,1 but not by clinical studies. We assessed cardiac function in 10 pa tients with severe ketoacidosis on admission and during correction of acidosis and did not observe any alteration in the left ventricular shortening
to
N E J M .O R G
M A Y 1-4, 2 0 1 5
Copyright © Massachusetts Medical Society 2015.
end point of 20 percentage points (number need certainties about effectiveness of the procedures ed to treat, 5). No other treatment for angina regarding blinding.4 (including coronary-artery bypass grafting) has Johann Auer, M .D. achieved a benefit of such magnitude. Until a St. Josef Hospital Braunau larger trial is performed, I would not be so en Braunau, Austria thusiastic about the main findings of this study. [email protected] M arco A.S. Cordeiro, M .D., Ph.D.
Robert Berent, M.D.
Hospital Evangglico Goiano An^polis, Brazil
Herzreha Bad IscHI Bad IscHI, Austria
[email protected] No potential conflict o f interest relevant to this letter was re ported.
No potential conflict of interest relevant to this letter was re ported.
1. Poldermans D, Boersma E, Bax JJ, et al. The effect of bisoprolol on perioperative mortality and myocardial infarction in high-risk patients undergoing vascular surgery. N Engl J Med 1999;341:1789-94. 2. The Hypothermia after Cardiac Arrest Study Group. Mild therapeutic hypothermia to improve the neurologic outcome after cardiac arrest. N Engl J Med 2002;346:549-56. [Erratum, N Engl J Med 2002;346:1756J 3. The POISE Study Group. Effects of extended-release metoprolol succinate in patients undergoing non-cardiac surgery (POISE trial): a randomised controlled trial. Lancet 2008;371: 1839-47. 4. Nielsen N, Wetterslev J, Cronberg T, et al. Targeted tempera ture management at 33°C versus 36°C after cardiac arrest. N Engl J Med 2013;369:2197-206. DOI: 10.1056/NEJMcl503672
Verheye and colleagues investi gated the antianginal effects of a coronary-sinus reducing device. The authors conclude that nar rowing of the coronary sinus was associated with significant improvement in symptoms. However, limitations in the design and conduct of this trial can lead to inaccurate conclusions regarding ef ficacy. First, despite instructing the performing physicians to behave similarly during both the active procedure and the sham procedure, and despite offering either headsets playing music or sedation to mask the conversation, the effective ness of blinding might have been limited. In ad dition, intravenous heparin was used only in pa tients assigned to the treatment group; therefore, postprocedural laboratory testing might have un veiled the study assignment of the patients. Fur thermore, the viewpoints of staff members and investigators, who were not initially aware o f the study assignments, might have been compromised by the physician performing the procedure, who was aware o f the study assignment.1'3 Finally, the choice of improvement of angina class as the pri mary end point rather than more objective varia bles such as exercise time or wall motion during stress makes the results more vulnerable to un t o th e e d it o r :
N ENG LJ
M ED 372120
1. Sackett DL. Commentary — measuring the success of blind ing in RCTs: don’t, must, can’t or needn’t? Int J Epidemiol 2007; 36:664-5. 2. Schulz KF, Grimes DA. Blinding in randomised trials: hiding who got what. Lancet 2002;359:696-700. 3. Bhatt DL, Kandzari DE, O’Neill WW, et al. A controlled trial of renal denervation for resistant hypertension. N Engl J Med 2014;370:1393-401. 4. Kandzari DE, Lam LC, Eisenstein EL, et al. Advanced coro nary artery disease: appropriate end points for trials of novel therapies. Am Heart J 2001;142:843-51. DOI: 10.1056/NEJMC1503672
r e p ly : In the COSIRA trial, less than 3% of the patients were excluded because of unsuitable coronary-sinus anatomy for device implantation. The main anatomical reason for excluding patients was a large diameter of the sinus. The proximity o f the circumflex artery is not a problem since the device is implanted in a proximal segment of the coronary sinus. Diffi culties in advancing the catheter into the coro nary sinus owing to the presence of valves are usually overcome easily with the use of different types of guiding wires. If necessary, it is easy to dilate the narrowed central segment of the stainless steel reducing device with a simple balloon-dilatation technique. This approach may allow future passage of a coronary-sinus lead for CRT. Efficacy findings in phase 2 studies must be interpreted with caution because o f the small number o f patients. To obtain approval for therapies that improve symptoms (such as an gina), regulatory agencies such as the Food and Drug Administration typically require two posi tive trials with P values less than 0.05.1 We therefore agree with Cordeiro that the COSIRA trial may have failed to reject a true null hypoth esis and that a confirmatory trial is needed. The lack of approaches to assess efficacy in small trials performed in early phases o f development programs for devices is a considerable problem. the authors
N E J M .O R G
MAY 14 , 2015
1967
The N E W E N G L A N D J O U R N A L o / M E D I C I N E
Corroborative analyses are one possible solution.2 We conducted a post hoc corroborative efficacy analysis that combined the primary and second ary end points. We found that the device was consistently associated with improved symptoms, function, and quality of life, as compared with a sham intervention. Coronary-sinus reduction is a device-based therapy for improvement of symptoms in patients with severe angina who had exhausted conven tional therapies. Therefore, the primary end point of the COSIRA trial aimed to answer the question of whether the device is an effective treatment to improve symptoms of angina. Placebo treatment alone can result in sub stantial improvement in angina symptoms and in exercise duration. Our sham-controlled clini cal trial was designed to eliminate both patient bias and investigator bias in the interpretation of outcome end points. Extensive efforts were made so that the patients and the physician perform ing the follow-up were unaware of the treatment assignment. Implantation of the coronary-sinus reducing
device is a simple procedure, and there was not much difference in the length of the procedure between the sham group and the treatment group. Also, there was no difference in the post procedural care between the two groups. Shmuel Banai, M.D. Tel Aviv Medical Center Tel Aviv, Israel [email protected]
Stefan Verheye, M.D. Ph.D. Antwerp Cardiovascular Center Antwerp, Belgium
E. Marcjolicceur, M.D. Montreal Heart Institute Montreal, QC, Canada
Since publication of their article, the authors report no fur ther potential conflict o f interest. 1. Jolicoeur EM, Ohman EM, Temple R, et al. Clinical and re search issues regarding chronic advanced coronary artery dis ease part II: trial design, outcomes, and regulatory issues. Am Heart J 2008;155:435-44. 2. Hare JM, Bolli R, Cooke JP, et al. Phase II clinical research design in cardiology: learning the right lessons too well: observa tions and recommendations from the Cardiovascular Cell Ther apy Research Network (CCTRN). Circulation 2013;127:1630-5. DOI: 10.1056/NEJMcl503672
A c id - B a s e P r o b le m s in D ia b e tic K e to a c id o s is t h e e d i t o r : In their review article, Kamel and Halperin (Feb. 5 issue)1 discuss “delta-delta,” the increase in the plasma anion gap and the decrease in the plasma bicarbonate concentra tion, for detecting the presence o f coexisting acid-base disorders. The ratio is predicated on the assumption that in simple acidosis with an increased anion gap, the gap will increase by the same amount that the bicarbonate decreases. However, many laboratories currently measure electrolytes with the use of ion-selective elec trodes, and the range for the anion gap is usually given as approximately 3 to 11. For the bicarbon ate level, the laboratory in which I work provides a reference range of 21 to 31. So what does one use as the baseline value for each measurement? We typically do not have samples obtained from patients just before the onset of the acid-base disorder, so how does one know what baseline to use in calculating how much deviation has oc curred? For the anion gap, is it 3 or 11 or in be tween, and for the bicarbonate concentration, is it 21 or 31 or in between? This ambiguity leads
to
1968
N
EN G LJ
M ED 372120
me to question the value of “delta-delta” in assess ing acid-base problems. Markjoy, M.D.J.D. VA Pittsburgh Healthcare System Pittsburgh, PA [email protected]
No potential conflict of interest relevant to this letter was re ported. 1. Kamel KS, Halperin ML. Acid-base problems in diabetic keto acidosis. N Engl J Med 2015;372:546-54. DOI: 10.1056/NEJMcl502745 t h e e d i t o r : Kamel and Halperin state that sodium bicarbonate infusion may be considered in some patients with diabetic ketoacidosis to avoid possible deterioration in the hemodynamic status. This debated issue is supported by some older experimental studies, which suggest that severe acidosis might be associated with de creased cardiac contractility,1 but not by clinical studies. We assessed cardiac function in 10 pa tients with severe ketoacidosis on admission and during correction of acidosis and did not observe any alteration in the left ventricular shortening
to
N E J M .O R G
M A Y 1-4, 2 0 1 5
Copyright © Massachusetts Medical Society 2015.
