| 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 |
|||||||||
* From the Division of Pulmonary & Critical Care Medicine (Drs. Shorr, Moores, and Fitzpatrick) and the Division of Hematology & Oncology (Drs. Edenfield and Christie), Walter Reed Army Medical Center, Washington, DC.
Correspondence to: Andrew F. Shorr, MD, MPH, Pulmonary & Critical Care Medicine, Department of Medicine, Walter Reed Army Medical Center, Washington, DC 20307; e-mail: CPT_Andrew_Shorr{at}WRAMC1.AMEDD.ARMY.MIL
 |
Abstract |
|---|
 TOP Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Study objective: To determine the frequency of and risk factors for the use of MV in recipients of AHSCT and to identify predictors of survival in mechanically ventilated AHSCT patients.
Design: Retrospective, cohort analysis
Setting: Tertiary-care, university-affiliated medical center.
Patients: One hundred fifty-nine consecutive patients who underwent AHSCT.
Interventions: Patient surveillance and data collection.
Measurements and results: The primary outcome measure was the need for MV, and the secondary end point was survival after MV. Of 159 patients, 17 required MV (10.7%). Three variables were associated with the need for MV: increasing age, use of total body irradiation in the conditioning regimen, and treatment with amphotericin B. As a screening test to predict the need for MV, no risk factor had a sensitivity or specificity > 82%. Three of the 17 mechanically ventilated patients (17.6%) survived to discharge. Only the mean APACHE (acute physiology and chronic health evaluation) II score separated survivors from nonsurvivors (21.7 vs 31.4; p = 0.029). Both the duration of MV and the length of stay in the ICU were similar in survivors and nonsurvivors.
Conclusions: We conclude that MV is infrequently needed following AHSCT. Although survival after MV in these patients is limited, clinical variables do not reliably allow clinicians to prospectively identify patients destined to die.
Key Words: autologous bone marrow transplantation • hematopoietic stem cell transplantation • mechanical ventilation • survival • withdrawal of care
|
Introduction |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Several previous studies examining outcomes in patients undergoing BM transplantation observed that approximately 50% of patients develop some form of pulmonary complication and half of these individuals require care in an ICU.4 5 6 Crawford et al7 noted that 21% of all patients undergoing BM transplantation required support with mechanical ventilation (MV) and that certain pretransplant characteristics, such as patient age, hematologic malignancy in relapse, and human leukocyte antigen donor-recipient incompatibility, predicted the need for MV. In terms of outcomes from MV, reported short-term survival rates range from 0 to 11%.7 8 9 10 Long-term survival rates in BM transplant recipients after MV are also dismal, with approximately 5% of patients surviving 6 months.11 Efforts to identify factors that prospectively identify nonsurvivors, however, have met with limited success.
Previous studies exploring the need for and outcomes from MV in BM transplant recipients have focused on heterogeneous groups of patients and included individuals undergoing both AHSCT and allogeneic BM transplantation.4 5 6 7 8 Moreover, most of these studies come from major transplantation referral centers.7 11 Thus, given the severity of illness of these patients prior to transplantation and possible differences in institutional experience and care of the transplant recipient, it is unclear whether results from these centers are generalizable to other institutions. In short, because AHSCT use is expanding, confirmatory data from other institutions are needed to help guide the management of these patients. Therefore, we undertook a review of all patients undergoing AHSCT at our institution to identify factors that not only were predictive of the need for MV but were also associated with survival from MV.
|
Materials and Methods |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Pre-AHSCT chemotherapy conditioning regimens varied based on the patient's underlying diagnosis. The majority of patients with breast cancer and all those suffering from germ cell tumors received a regimen consisting of carboplatin (1,500 mg/m2), etoposide (1,200 mg/m2), and cytoxan (1,200 mg/kg). Approximately 11% of subjects with breast cancer (n = 11) were treated with a regimen of cytoxan (5,625 mg/m2), cisplatin (165 mg/m2), and 1,3-bis(2-chloroethyl)-1-nitrosourea (600 mg/m2); this regimen is similar to the STAMP I protocol. Patients suffering from lymphoma underwent conditioning with a combination of cytoxan (150 mg/kg), 1,3-bis(2-chloroethyl)-1-nitrosourea (60 mg/m2), and etoposide (1500 mg/m2), while patients with myeloma were treated with melphalan at a dose of either 140 mg/m2 or 200 mg/m2. Total body irradiation (TBI) was also employed for all patients receiving the lower melphalan dose and for some patients with lymphoma.
Patients were followed longitudinally to determine outcomes. The primary end point for this study was the need for MV, which was defined as use of machine-delivered tidal breaths through an endotracheal tube. No patient received noninvasive ventilation. Survival to hospital discharge after the initiation of MV was a secondary end point. No patient was admitted to the MICU immediately following an operative procedure, and no patient underwent MV for less than a 24-h period. If a patient was admitted to the MICU more than once, only data from the initial admission were used for analysis.
Clinical Variables
Data regarding patient characteristics, hospital course, and
transplant-specific variables were extracted from a computerized
hospital database and patient records. Patient characteristics analyzed
included age, sex, serum creatinine at time of admission for
transplantation, underlying diagnosis, cytomegalovirus (CMV)
serostatus, cardiac ejection fraction (EF), and spirometry. Age was
analyzed as both a continuous and a noncontinuous variable, with an age
of > 50 years arbitrarily chosen for comparisons. EF was measured by
either radionuclide scan or echocardiography. Spirometric testing was
done and interpreted in accordance with American Thoracic Society
guidelines. A pneumotach spirometer (Cybermedic, Inc., Ohio) was
used and the diffusing capacity for carbon monoxide was measured by the
single-breath method. Those factors related to the hospital course
which were examined included the time to neutrophil engraftment and the
use and duration of use of amphotericin B (ampho B). Time to
engraftment (days) was defined as the period between stem cell
reinfusion and the return of the absolute neutrophil count to > 500
cells/mL. Transplant-specific variables of interest were whether either
TBI along with alkylator therapy or alkylator therapy alone was
employed for the conditioning regimen. We also examined the
relationship between the source of the hematopoietic cells (BM and/or
PB) and the need for MV.
For patients receiving MV, we compared survivors and nonsurvivors with regard to each of the variables noted above. We further compared differences in severity of illness at admission in survivors and nonsurvivors based on the Acute Physiology and Chronic Health Evaluation (APACHE) II score and the Multiple Organ Dysfunction Score.12 13 Length of stay in the MICU and duration of MV were also measured. The reason for MV was also recorded.
Statistical Analysis
Associations between the study variables and either the need for
MV or survival from MV were analyzed by Student's t test
for continuous variables and by either the Fisher Exact Test or
2 test for noncontinuous variables. All tests
were two-tailed and a p value of less than 0.05 was assumed to
represent statistical significance. For variables that were
significantly associated with either study end point, standard risk
ratios were computed and then adjusted for possible interactions
between these variables. Stepwise logistic regression was used to
assess multivariate interactions between multiple variables and study
end points. Where appropriate, 95% confidence intervals (CIs) are
reported. All statistical analyses were performed using SPSS 7.0 (SPSS
Inc; Chicago, IL).
|
Results |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
As shown in Table 2 , three variables were significantly associated with the need for MV: age > 50 years, (adjusted p = 0.023), conditioning with TBI (adjusted p = 0.020), and ampho B use (adjusted p = 0.009). Age, when analyzed in a continuous fashion rather than as a noncontinuous variable, was also significantly associated with the use of MV (p = 0.013). CMV serostatus, alkylator use, and days to engraftment were not associated with the need for MV. The source of the hematopoietic cells (BM alone, PB alone, BM with PB) also did not influence the need for MV. Neither pretransplantation cardiac EF nor spirometric results differed between the two cohorts. For example, as shown in Table 3 the mean EF among patients who required MV was 62.9 ± 5.8%, compared with 64.9 ± 7.3% in patients who did not need ventilatory support. Additionally, the duration of neutropenia was similar in the two groups (10.5 days vs 12.3 days). There was a trend toward a longer duration of cytokine therapy (granulocyte colony-stimulating factor) predicting the need for MV, but this trend did not reach statistical significance. There did not appear to be a learning effect. In other words, year of transplant did not affect outcome, and the yearly proportion of patients requiring MV was similar throughout the study period. By stepwise logistic regression, adjusted risk ratios for the three variables associated with MV use were computed. These are shown in Table 4 . Ampho B use was associated with the greatest increased risk for MV (relative risk, 5.03; 95% CI, 1.49 to 17.02).
|
|
|
|
29 survived to discharge. Of the 14 patients who died, seven did
so after the withdrawal of supportive care. "Do not resuscitate"
orders were eventually written in nine cases, including the seven
patients in whom care was withdrawn. None of the survivors developed
multisystem organ failure, but all required prolonged (> 24 h)
treatment with vasopressors (norepinephrine and/or dopamine at > 5
µg/kg/min).
|
|
Discussion |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Crawford et al7 initially found that MV was needed in 21% of BM transplant patients, whereas in a later study, Rubenfeld and Crawford14 reported that 25% of transplant patients required MV.7 14 The use of MV in our population was lower (10.7%). This difference likely results from the distinct composition of the two study cohorts. Our patients received AHSCT, while the majority of the patients reported by Crawford et al underwent allogeneic transplant. Approximately 90% of the patients in their original study and 80% of the subjects in a subsequent report received allogeneic transplantation.7 11 This difference may reflect several factors. First, AHSCT generally results in fewer complications than allogeneic BM transplantation.15 Specifically, in AHSCT, graft-vs-host disease is not a clinical concern. However, Crawford et al7 did not find that the need for MV was significantly lower in AHSCT patients. Second, and more likely, the greater need for MV in prior studies likely represents referral bias. Specialized institutions such as the Fred Hutchinson Cancer Center may offer BM transplantation to patients whom otherwise would have been refused this option. The fact that some believe that increases in long-term survival following allogeneic BM transplantation and AHSCT reflect referral bias further highlights the significance of this issue.16 This potential for referral bias underscores the need for outcomes data from multiple centers. Unfortunately, other reports on MV in BM transplantation have focused solely on survival from MV and included no data regarding the overall need for MV.
In terms of risk factors for MV, our data confirm the findings of Crawford and colleagues7 that patient age is associated with the need for ventilatory support. In contrast to Crawford's data, we observed that TBI was a risk factor for MV. Although Crawford et al7 found that TBI use correlated with MV by univariate analysis, TBI was not a risk factor for MV in multivariate analysis. This discordance likely reflects the limited power of our study. In other words, our study cohort was smaller than the population reported by Crawford et al, and relatively few patients in our cohort (28.9 vs 89.4%) received TBI.
Ampho B therapy was also correlated with the use of MV. The significance of this relationship is unclear. Ampho B use could be a surrogate marker for severity of illness. Patients received ampho B because of either a documented fungal infection or because they had a prolonged neutropenic fever unresponsive to broad-spectrum antibiotics. That the duration of neutropenia as measured by time to engraftment was not correlated with the need for MV implies that ampho B therapy may not solely be a marker for severity of illness. Rather, ampho B itself may be toxic. For example, ampho B use, particularly when employed in addition to other nephrotoxic antibiotics such as aminoglycosides, can result in acute tubular necrosis. Ampho B therapy in the setting of granulocyte transfusion has also been reported to lead to acute lung injury.17 In rats, ampho B has been shown to lead to a neutrophil-independent form of acute lung injury,18 while in sheep, cyclooxygenase products have been found to play a significant role in ampho B pulmonary dysfunction.19 On a cellular level, Berliner et al20 demonstrated that ampho B enhanced pulmonary leukostasis. These data, coupled with clinical reports of acute lung injury in BM transplant patients who received ampho B, suggest that ampho B use is more than a surrogate for severity of illness.
Additionally, prior reports of interstitial pneumonitis early in the posttransplant period (ie, < 1 month after transplant) may reflect a form of lung injury similar to the one discussed above.21 It is difficult to compare these previously described cases of early posttransplant interstitial pneumonitis with the cases of acute lung injury we encountered. First, the previous literature on interstitial pneumonitis focuses predominantly on allogeneic BM transplant.21 Second, in most reports, the definition of interstitial pneumonitis is vague. Finally, recent research suggests that patients undergoing AHSCT may be at risk for a variant of acute lung injury related directly to the engraftment process.22
We observed that in-hospital mortality of patients undergoing MV after allogeneic BM transplantation or peripheral blood stem cell transplantation remained high. Limited survival from MV in this setting has been consistently reported. The initial reports noted that survival rates ranged from 0 to 8%, while more recent studies (published since 1992) concluded that 3 to 19% of transplant recipients survived MV.4 5 6 7 8 9 10 11 14 18 In 115 blood stem cell transplantation patients, Price and colleagues22 observed that approximately 19% of intubated patients lived to discharge. Rubenfeld and Crawford,14 in the largest published series (n = 3,635), noted an overall survival rate of 6.1%. However, Rubenfeld and Crawford14 also reported that survival rates increased over time, with 16% of MV patients living to discharge by 1992.14 The 17.6% survival rate in our cohort is consistent with this recent trend.
Some centers have recently reported lower mortality rates for AHSCT, particularly in breast cancer patients. Our immediate post-AHSCT mortality in breast cancer was 8.1% (95% CI, 3.8 to 15.7). Given that our study covers nearly a 7-year period and that our data are generally consistent with the trends noted above, we believe the observed rates of ventilatory failure and mortality are consistent with the results of other centers. Any observed variances in these parameters is unlikely to reflect differences in conditioning regimens.
Because MV in BM transplant recipients is fraught with complex emotional, economical, and ethical issues, efforts have been made to identify variables that will allow clinicians to prospectively determine which patients eventually die. For example, Faber-Langendoen et al9 ask rhetorically whether "ventilatory support should be initiated at all" in the BM transplant patient. They conclude that MV should not be considered the standard of care in certain BM transplant patients (eg, age > 40 years or MV within 90 days of transplant).9 Rubenfeld and Crawford14 noted that survival after MV was statistically associated with APACHE III score, patient age, and time between transplantation and intubation. They further observed that the combination of two or more clinical variables (lung injury, use of vasopressors, and hepatic and renal failure) led to dismal survival rates (< 2%).14 The value of severity-of-illness scores in the AHSCT population is unclear. Only two previous studies have specifically reported APACHE II scores. Afessa et al5 noted that APACHE II scores did not differentiate MICU survivors from nonsurvivors (25.0 ± 10.2 vs 27.3 ± 8.9), while Paz et al8 found that APACHE II scores were significantly lower in MICU survivors (15.8 ± 3.8 vs 21.2 ± 4.7). In our study, only the APACHE II score differentiated between survivors and nonsurvivors. The different findings may be related to differences in the composition of the various cohorts, as our study focused exclusively on patients receiving AHSCTs. These differences, however, suggest that severity-of-illness scores may not be helpful in AHSCT patients admitted to the MICU.
It is important to note that all of our survivors were > 40 years old and underwent MV < 90 days posttransplant. Additionally, each survivor had lung injury and at least one additional risk factor for mortality according to the definitions proposed by Rubenfeld and Crawford.14 Put simply, these two proposed guidelines for limiting MV failed to identify the AHSCT patients at our institution who would survive MV. Therefore, before adopting a plan to unilaterally limit a life-sustaining therapy, it is imperative that physicians systematically review the experience at their institutions. As Rubenfeld and Crawford14 commented, "Clinical prediction tools cannot easily be applied until they have been validated in other settings."
Our study has several limitations. First, its retrospective nature exposes our study to several forms of bias (eg, recall bias). However, since we reviewed prospectively collected data and focused on end points with clear definitions, the impact of potential bias should be small. Second, and more importantly, our study sample was small compared with those reported by Crawford et al,7 Rubenfeld and Crawford,14 and Price et al22 We attempted to compensate for this by relying on 95% CIs. For example, the 95% CI for the rate of MV use (6.5 to 16.8) did not overlap with the incidence of MV reported by Rubenfeld and Crawford14 (22.4 to 25.2).
In conclusion, we found that the need for MV in AHSCT is low and that certain risk factors (ie, age, TBI, and ampho B use) are associated with the use of MV. Although these variables have limited prognostic power, they may aid physicians in counseling patients prior to transplantation. For AHSCT patients, survival rates from MV remain poor. The recommendations of others regarding guidelines to be employed in limiting MV in BM transplant recipients are not necessarily applicable to other institutions.
|
Acknowledgements |
|---|
|
Footnotes |
|---|
The opinions expressed herein are not to be construed as official or as reflecting the policy of either the Department of the Army or the Department of Defense.
Received for publication November 13, 1998. Accepted for publication April 1, 1999.
|
References |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. O. Soubani, E. Kseibi, J. J. Bander, J. L. Klein, G. Khanchandani, H. P. Ahmed, and J. A. Guzman Outcome and Prognostic Factors of Hematopoietic Stem Cell Transplantation Recipients Admitted to a Medical ICU Chest, November 1, 2004; 126(5): 1604 - 1611. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. M. Kotloff, V. N. Ahya, and S. W. Crawford Pulmonary Complications of Solid Organ and Hematopoietic Stem Cell Transplantation Am. J. Respir. Crit. Care Med., July 1, 2004; 170(1): 22 - 48. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Rano, C. Agusti, N. Benito, M. Rovira, J. Angrill, T. Pumarola, and A. Torres Prognostic Factors of Non-HIV Immunocompromised Patients With Pulmonary Infiltrates* Chest, July 1, 2002; 122(1): 253 - 261. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. B. Bach, D. Schrag, D. M. Nierman, D. Horak, P. White Jr, J. W. Young, and J. S. Groeger Identification of poor prognostic features among patients requiring mechanical ventilation after hematopoietic stem cell transplantation Blood, December 1, 2001; 98(12): 3234 - 3240. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
