- Email: [email protected]
Negative Fluid Balance as Predictor of Mortality
Second, the authors report a higher incidence of perioperative and postoperative complications in patients undergoing ST. As reported by the study of Gysin et al3 in 1999, the vast majority of these complications are clinically insignificant, consisting of nuisance bleedings and peristomal infections. On the other hand, life-threatening complications, although relatively infrequent, are more common in PDT. For example, the meta-analysis of Dulguerov et al4 in 1999 showed a higher incidence of death and major cardiorespiratory events in patients undergoing PDT. A higher incidence of death, loss of airway, and hypoxia in patients undergoing PDT was also reported by Porter and Ivatury.5 Third, there are no solid data to support the purported advantages of bedside PDT, which include reduced costs, decreased procedure time, and avoidance of the significant risk of transfer to the OR. In fact, a recent comparison of PDT to bedside ST failed to demonstrate any significant advantage of PDT regarding costs, complications, and procedure time. Taken together, these considerations suggest that PDT has a definite role in carefully selected patients. It should not be considered the procedure of choice for many critically ill patients for whom ST is the more appropriate procedure. Our current practice, therefore, is to perform PDT and ST at the bedside or in the OR, depending on the clinical circumstances and the physician’s preference and expertise. Finally, we agree that Dr. Freeman’s meta-analysis demonstrates that a well-designed, prospective, randomized study to compare complications, costs, and final outcome is required for a definitive answer to the question of whether to use ST vs PDT. Jeffrey D. Anderson, MD Reuven Rabinovici, MD Heidi L. Frankel, MD, FCCP Yale University School of Medicine New Haven, CT Correspondence to: Heidi L. Frankel, MD, FCCP, Yale University School of Medicine, 333 Cedar Street LH 118, Department of Surgery, New Haven, CT 06520-8062; e-mail: heidi.frankel@ yale.edu
References 1 Freeman BD, Isabella K, Lin N, et al. A meta-analysis of prospective trials comparing percutaneous and surgical tracheostomy in critically ill patients. Chest 2000; 108:1412– 1418 2 Brookes N, Howard D. Recent advances in intensive care: percutaneous tracheostomy may not be more effective than open technique. BMJ 2000; 320:1668 –1669 3 Gysin C, Dulguerov P, Guyot J, et al. Percutaneous versus surgical tracheostomy: a double-blind randomized trial. Annu Surg 1999; 230:208 –216 4 Dulguerov P, Gysin C, Perneger TV, et al. Percutaneous or surgical tracheostomy: a meta-analysis. Crit Care Med 1999; 27:1617–1625 5 Porter JM, Ivatury RR. Preferred route of tracheostomy— percutaneous versus open at bedside: a randomized prospective study in the SICU. Annu Surg 1999; 65:142–146 To the Editor: We wish to respond to the points raised by Anderson et al regarding potential limitations of percutaneous dilational tracheostomy (PDT) and our analysis.1 First, Anderson et al question the generalizability of our results. As with any clinical study, the conclusions of our metaanalysis are constrained by the selection criteria used in the individual trials on which our study was based. These trials excluded patients with contraindications for PDT, such as distorted neck anatomy, morbid obesity, and uncorrectable coagu1424
lopathy. Surgically created tracheostomy (SCT) remains the preferred method of providing an artificial airway in these settings, and the conclusions of our meta-analysis do not apply. In contrast, for patients meeting the inclusion criteria of these individual studies, our meta-analysis suggests that PDT has advantages relative to SCT.1 Second, Anderson et al question the significance and frequency of complications associated with PDT vs SCT. We found that PDT was associated with a lower rate of perioperative and postoperative bleeding and peristomal infection.1 We disagree that these complications, which potentially add to the expense and morbidity of patient care, are trivial. As Anderson et al point out, the findings of our meta-analysis conflict with those of Dulguerov et al,2 who reported a higher incidence of death and major cardiovascular complications associated with PDT. Differences in the nature of the studies included in these two analyses may explain their conflicting conclusions. Specifically, Dulguerov et al2 included both prospective and observational studies as well as those using a variety of techniques for performing PDT. In contrast, only prospective studies employing the method of Ciaglia et al3 were selected for our analysis. Third, a recently completed prospective study, which demonstrates that PDT may be performed more quickly and at lower cost than SCT,4 refutes the argument that there are no solid data to support the purported advantages of PDT. Finally, while meta-analysis remains a useful technique, we consider our findings preliminary. We agree with Anderson et al that additional, adequately powered studies comparing PDT and SCT are needed to completely answer the question as to the relative merits and limitations of these techniques. Bradley D. Freeman, MD Karen Isabella, RN Natatia Lin, MS Timothy G. Buchman, PhD, MD Washington University School of Medicine St. Louis, MO Correspondence to: Bradley D. Freeman, MD, Washington University School of Medicine, Department of Surgery, Suite 6104, Box 8109, St. Louis, MO 63110; e-mail: [email protected]
References 1 Freeman BD, Isabella K, Lin N, et al. A meta-analysis of prospective trials comparing percutaneous and surgical tracheostomy in critically ill patients. Chest 2000; 118:1412– 1418 2 Dulguerov P, Gysin C, Perneger TV, et al. Percutaneous or surgical tracheostomy: a meta-analysis. Crit Care Med 1999; 27:1617–1625 3 Ciaglia P, Firsching R, Syniec C. Elective percutaneous dilational tracheostomy. Chest 1985: 87:715–719 4 Freeman BD, Isabella K, Cobb JP, et al. A prospective randomized study comparing percutaneous and surgical tracheostomy in critically ill patients. Crit Care Med 2001; 29:926 –930
Negative Fluid Balance as Predictor of Mortality To the Editor: Alsous and coworkers (June 2000)1 reported that, in a study Communications to the Editor
of 36 patients in a medical ICU, a negative fluid balance of ⬎ 500 mL on any of the first 3 days of hospital admission was associated with better survival. In a retrospective analysis, we examined 80 critically ill surgical patients who were admitted to an intensive therapy unit (ITU) for mechanical ventilation and who had survived at least 3 days. There were 30 female and 50 male patients, and they had a median APACHE (acute physiology and chronic health evaluation) II score of 21 (range, 10 to 43) and a predicted percentage mortality of 40.8% (range, 10.1 to 90.5). Twenty-nine patients died within 30 days. There was no significant difference in APACHE II scores or predicted mortality rates between those patients who died compared with those who lived. In those patients who survived, there was a nonsignificant trend toward a higher proportion of patients who had a negative fluid balance ⬎ 500 mL on any of the first 3 days (12 of 51 patients), compared with those who died (2 of 29 patients; p ⫽ 0.072, two tailed). These results, in an unselected cohort of critically ill surgical patients, do not confirm the strong prognostic value of a negative fluid balance ⬎ 500 mL on any of the first 3 days of hospital admission. The reasons for the conflicting results are unclear, as the mean APACHE II scores in the previous and present study (25 vs 22, respectively) and the percentage of patients who died (56% vs 36%, respectively) were similar. In the present study, there were approximately twice the number of patients and twice the number of deaths. All the patients were either postoperative or had severe pancreatitis. This would have a major influence on the fluid requirements of the patients. Also, the normal stress response to surgery involves alterations in circulating hormone concentrations, with an antidiuretic response and water retention for the first 48 h after surgery. Therefore, it may be that, in the critically ill, postoperative surgical patient, the mobilization of excess fluid occurs at a later period of time than in medical patients. We conclude that a negative fluid balance, achieved by the third day after admission to the ITU, is not a significant predictor of death in the critically ill surgical patient, and therefore, it is not generally applicable to critically ill patients. Tara Quasim Donald C. McMillan, PhD John Kinsella Royal Infirmary Glasgow, United Kingdom Correspondence to: John Kinsella, FRCA, Department of Anesthesia, Royal Infirmary, 10 Alexandria Parade, Glasgow G-31 2 ER, United Kingdom; e-mail: [email protected]
Reference 1 Alsous F, Khamiees M, DeGirolamo A, et al. Negative fluid balance predicts survival in patients with septic shock: a retrospective pilot study. Chest 2000; 117:1749 –1754
demonstrates that there was a survival advantage for their postoperative patients who achieved a negative fluid balance in the first 3 days (relative risk ⫽ 1.45; 95% confidence interval ⫽ 1.08 to 1.94).2 The two studies differ merely in the magnitude of the observed survival advantage. This is to be expected for a number of reasons. The average amount of fluid administered to our patients with septic shock in the first day was 5.5 L. We have not been given the average amount of fluid administered to their patients on the first day but we doubt that the magnitude was similar. Another possibility is that different surgeries have variable rates of recovery based upon the gravity of the initial injury and factors pertaining to the procedure (eg, duration of anesthesia, size of the wound, site of the operation). Insofar as the dispersion of vascular recovery times is greater in a heterogenous group of postoperative patients, the dispersion of transition points from positive to negative balance would also be greater. The difference in observed death rate (56% in our study versus 21% in their cohort) may reflect differences in the severity of the systemic inflammatory response in the two patient groups—a factor that may impact the magnitude of vascular dilatation and/or leak and, therefore, outcomes. We are not provided with significant information about the characteristics of their patients. Differences in age, severity of illness, and baseline creatinine levels are bound to effect the magnitude of survival advantage, as demonstrated in Table 3 of our report.1 For instance, whereas patients with sepsis with negative fluid balance and APACHE II score ⬍ 20 were 1.3 times more likely to survive, those with negative balance and APACHE II score ⱖ 20 were 6 times more likely to survive. Thus, any meaningful differences in the proportional representation of the groups with APACHE II scores ⬍ 20 and ⱖ 20 would result in significant variations in the reported unadjusted risk ratios. Accordingly, we are delighted that our findings have been replicated in studies with postoperative patients. We hope that larger cohort studies will further validate these findings and will provide more precise estimates of the true magnitude of survival advantage in various populations of critically ill patients. Constantine A. Manthous, MD, FCCP Yaw Amoateng-Adjepong, MD, PhD Yale University School of Medicine New Haven, CT Correspondence to: Constantine A. Manthous, MD, FCCP, Bridgeport Hospital, West Tower 6, 267 Grant St, Bridgeport, CT 06610-2870; e-mail: [email protected]
References 1. Alsous F, Khamiees M, DeGirolamo A, et al. Negative fluid balance predicts survival in patients with septic shock: a retrospective pilot study. Chest 2000; 117:1749 –1754 2. Greenland S, Robbin JM. Estimation of common effect parameter from sparse follow-up data. Biometrics 1985; 41:55– 68
To the Editor: We appreciate the opportunity to respond to the interesting letter of Drs. Quasim, McMillan, and Kinsella. We are very pleased that these physicians took the time to interrogate our hypothesis.1 We published our observation because we are convinced that fluid fluxes are common in critically ill patients and that vascular leak and dilatation and the need for refilling are followed by a diuretic phase in those patients with sepsis who recover (it is inevitable). In our original report, we did not imply that these findings were conclusive, nor did we assert that they were applicable to any other than septic patients with sepsis. Nevertheless, a reanalysis of the data provided by Quasim et al
Does Omentoplasty Preclude Cardiac Retransplantation? To the Editor: Infectious complications are, with rejection, the main cause of morbidity and mortality in recipients of heart transplantation (HT). Between September 1984 and October 2000, we performed 514 HTs in 505 patients; of these, postoperative mediastinitis developed in 7 patients (1.4%). The mortality rate in this group of patients was 42%. CHEST / 120 / 4 / OCTOBER, 2001
1425
References 1 Freeman BD, Isabella K, Lin N, et al. A meta-analysis of prospective trials comparing percutaneous and surgical tracheostomy in critically ill patients. Chest 2000; 108:1412– 1418 2 Brookes N, Howard D. Recent advances in intensive care: percutaneous tracheostomy may not be more effective than open technique. BMJ 2000; 320:1668 –1669 3 Gysin C, Dulguerov P, Guyot J, et al. Percutaneous versus surgical tracheostomy: a double-blind randomized trial. Annu Surg 1999; 230:208 –216 4 Dulguerov P, Gysin C, Perneger TV, et al. Percutaneous or surgical tracheostomy: a meta-analysis. Crit Care Med 1999; 27:1617–1625 5 Porter JM, Ivatury RR. Preferred route of tracheostomy— percutaneous versus open at bedside: a randomized prospective study in the SICU. Annu Surg 1999; 65:142–146 To the Editor: We wish to respond to the points raised by Anderson et al regarding potential limitations of percutaneous dilational tracheostomy (PDT) and our analysis.1 First, Anderson et al question the generalizability of our results. As with any clinical study, the conclusions of our metaanalysis are constrained by the selection criteria used in the individual trials on which our study was based. These trials excluded patients with contraindications for PDT, such as distorted neck anatomy, morbid obesity, and uncorrectable coagu1424
lopathy. Surgically created tracheostomy (SCT) remains the preferred method of providing an artificial airway in these settings, and the conclusions of our meta-analysis do not apply. In contrast, for patients meeting the inclusion criteria of these individual studies, our meta-analysis suggests that PDT has advantages relative to SCT.1 Second, Anderson et al question the significance and frequency of complications associated with PDT vs SCT. We found that PDT was associated with a lower rate of perioperative and postoperative bleeding and peristomal infection.1 We disagree that these complications, which potentially add to the expense and morbidity of patient care, are trivial. As Anderson et al point out, the findings of our meta-analysis conflict with those of Dulguerov et al,2 who reported a higher incidence of death and major cardiovascular complications associated with PDT. Differences in the nature of the studies included in these two analyses may explain their conflicting conclusions. Specifically, Dulguerov et al2 included both prospective and observational studies as well as those using a variety of techniques for performing PDT. In contrast, only prospective studies employing the method of Ciaglia et al3 were selected for our analysis. Third, a recently completed prospective study, which demonstrates that PDT may be performed more quickly and at lower cost than SCT,4 refutes the argument that there are no solid data to support the purported advantages of PDT. Finally, while meta-analysis remains a useful technique, we consider our findings preliminary. We agree with Anderson et al that additional, adequately powered studies comparing PDT and SCT are needed to completely answer the question as to the relative merits and limitations of these techniques. Bradley D. Freeman, MD Karen Isabella, RN Natatia Lin, MS Timothy G. Buchman, PhD, MD Washington University School of Medicine St. Louis, MO Correspondence to: Bradley D. Freeman, MD, Washington University School of Medicine, Department of Surgery, Suite 6104, Box 8109, St. Louis, MO 63110; e-mail: [email protected]
References 1 Freeman BD, Isabella K, Lin N, et al. A meta-analysis of prospective trials comparing percutaneous and surgical tracheostomy in critically ill patients. Chest 2000; 118:1412– 1418 2 Dulguerov P, Gysin C, Perneger TV, et al. Percutaneous or surgical tracheostomy: a meta-analysis. Crit Care Med 1999; 27:1617–1625 3 Ciaglia P, Firsching R, Syniec C. Elective percutaneous dilational tracheostomy. Chest 1985: 87:715–719 4 Freeman BD, Isabella K, Cobb JP, et al. A prospective randomized study comparing percutaneous and surgical tracheostomy in critically ill patients. Crit Care Med 2001; 29:926 –930
Negative Fluid Balance as Predictor of Mortality To the Editor: Alsous and coworkers (June 2000)1 reported that, in a study Communications to the Editor
of 36 patients in a medical ICU, a negative fluid balance of ⬎ 500 mL on any of the first 3 days of hospital admission was associated with better survival. In a retrospective analysis, we examined 80 critically ill surgical patients who were admitted to an intensive therapy unit (ITU) for mechanical ventilation and who had survived at least 3 days. There were 30 female and 50 male patients, and they had a median APACHE (acute physiology and chronic health evaluation) II score of 21 (range, 10 to 43) and a predicted percentage mortality of 40.8% (range, 10.1 to 90.5). Twenty-nine patients died within 30 days. There was no significant difference in APACHE II scores or predicted mortality rates between those patients who died compared with those who lived. In those patients who survived, there was a nonsignificant trend toward a higher proportion of patients who had a negative fluid balance ⬎ 500 mL on any of the first 3 days (12 of 51 patients), compared with those who died (2 of 29 patients; p ⫽ 0.072, two tailed). These results, in an unselected cohort of critically ill surgical patients, do not confirm the strong prognostic value of a negative fluid balance ⬎ 500 mL on any of the first 3 days of hospital admission. The reasons for the conflicting results are unclear, as the mean APACHE II scores in the previous and present study (25 vs 22, respectively) and the percentage of patients who died (56% vs 36%, respectively) were similar. In the present study, there were approximately twice the number of patients and twice the number of deaths. All the patients were either postoperative or had severe pancreatitis. This would have a major influence on the fluid requirements of the patients. Also, the normal stress response to surgery involves alterations in circulating hormone concentrations, with an antidiuretic response and water retention for the first 48 h after surgery. Therefore, it may be that, in the critically ill, postoperative surgical patient, the mobilization of excess fluid occurs at a later period of time than in medical patients. We conclude that a negative fluid balance, achieved by the third day after admission to the ITU, is not a significant predictor of death in the critically ill surgical patient, and therefore, it is not generally applicable to critically ill patients. Tara Quasim Donald C. McMillan, PhD John Kinsella Royal Infirmary Glasgow, United Kingdom Correspondence to: John Kinsella, FRCA, Department of Anesthesia, Royal Infirmary, 10 Alexandria Parade, Glasgow G-31 2 ER, United Kingdom; e-mail: [email protected]
Reference 1 Alsous F, Khamiees M, DeGirolamo A, et al. Negative fluid balance predicts survival in patients with septic shock: a retrospective pilot study. Chest 2000; 117:1749 –1754
demonstrates that there was a survival advantage for their postoperative patients who achieved a negative fluid balance in the first 3 days (relative risk ⫽ 1.45; 95% confidence interval ⫽ 1.08 to 1.94).2 The two studies differ merely in the magnitude of the observed survival advantage. This is to be expected for a number of reasons. The average amount of fluid administered to our patients with septic shock in the first day was 5.5 L. We have not been given the average amount of fluid administered to their patients on the first day but we doubt that the magnitude was similar. Another possibility is that different surgeries have variable rates of recovery based upon the gravity of the initial injury and factors pertaining to the procedure (eg, duration of anesthesia, size of the wound, site of the operation). Insofar as the dispersion of vascular recovery times is greater in a heterogenous group of postoperative patients, the dispersion of transition points from positive to negative balance would also be greater. The difference in observed death rate (56% in our study versus 21% in their cohort) may reflect differences in the severity of the systemic inflammatory response in the two patient groups—a factor that may impact the magnitude of vascular dilatation and/or leak and, therefore, outcomes. We are not provided with significant information about the characteristics of their patients. Differences in age, severity of illness, and baseline creatinine levels are bound to effect the magnitude of survival advantage, as demonstrated in Table 3 of our report.1 For instance, whereas patients with sepsis with negative fluid balance and APACHE II score ⬍ 20 were 1.3 times more likely to survive, those with negative balance and APACHE II score ⱖ 20 were 6 times more likely to survive. Thus, any meaningful differences in the proportional representation of the groups with APACHE II scores ⬍ 20 and ⱖ 20 would result in significant variations in the reported unadjusted risk ratios. Accordingly, we are delighted that our findings have been replicated in studies with postoperative patients. We hope that larger cohort studies will further validate these findings and will provide more precise estimates of the true magnitude of survival advantage in various populations of critically ill patients. Constantine A. Manthous, MD, FCCP Yaw Amoateng-Adjepong, MD, PhD Yale University School of Medicine New Haven, CT Correspondence to: Constantine A. Manthous, MD, FCCP, Bridgeport Hospital, West Tower 6, 267 Grant St, Bridgeport, CT 06610-2870; e-mail: [email protected]
References 1. Alsous F, Khamiees M, DeGirolamo A, et al. Negative fluid balance predicts survival in patients with septic shock: a retrospective pilot study. Chest 2000; 117:1749 –1754 2. Greenland S, Robbin JM. Estimation of common effect parameter from sparse follow-up data. Biometrics 1985; 41:55– 68
To the Editor: We appreciate the opportunity to respond to the interesting letter of Drs. Quasim, McMillan, and Kinsella. We are very pleased that these physicians took the time to interrogate our hypothesis.1 We published our observation because we are convinced that fluid fluxes are common in critically ill patients and that vascular leak and dilatation and the need for refilling are followed by a diuretic phase in those patients with sepsis who recover (it is inevitable). In our original report, we did not imply that these findings were conclusive, nor did we assert that they were applicable to any other than septic patients with sepsis. Nevertheless, a reanalysis of the data provided by Quasim et al
Does Omentoplasty Preclude Cardiac Retransplantation? To the Editor: Infectious complications are, with rejection, the main cause of morbidity and mortality in recipients of heart transplantation (HT). Between September 1984 and October 2000, we performed 514 HTs in 505 patients; of these, postoperative mediastinitis developed in 7 patients (1.4%). The mortality rate in this group of patients was 42%. CHEST / 120 / 4 / OCTOBER, 2001
1425