The influence of high-efficiency particulate air filtration on mortality among multiple myeloma patients receiving autologous stem cell transplantation

Autologous stem cell transplantation (ASCT) continues to be the standard treatment for transplant-eligible multiple myeloma (MM) patients. A portion of MM patients received ASCT in an isolation room with high-efficiency particulate air (HEPA) filtration. The effectiveness of the HEPA filtration on reducing treatment-related mortality (TRM) is controversial. We enrolled patients with newly diagnosed MM in Taiwan between 2000 and 2017. The primary endpoint of the study was TRM, which was defined as death within 100 days after ASCT. A total of 961 MM patients received ASCT. Of them, 480 patients (49.9%) received ASCT in an isolation room with HEPA filtration (HEPA group). The median overall survival from ASCT was 7.52 years for the HEPA group and 5.88 years for the remaining patients (non-HEPA group) (p = 0.370). The 100-day mortality rate was 1.5% and 1.0% for the HEPA and non-HEPA groups, respectively. In the multivariate analysis, the 100-day mortality had no difference between the HEPA and non-HEPA groups (adjusted hazard ratio 1.65, 95% CI 0.52–5.23). The median cost for ASCT inpatient care was $13,777.6 and $6527.6 for the HEPA and non-HEPA groups, respectively (p < 0.001). Although half of MM patients in Taiwan received ASCT in HEPA room, it didn’t affect 100-day mortality.

www.nature.com/scientificreports/ The latest survey conducted between 2014 and 2015 showed that the use of HEPA-filtered rooms was 68% for ASCT recipients. The need for environmental HEPA filtration for patients receiving ASCT has not been established. IDSA guidelines suggest considering the use of HEPA-filtered rooms for ASCT recipients who develop prolonged neutropenia, which is the major risk factor of nosocomial aspergillosis 12 . Conversely, rapid engraftment with peripheral blood stem cells and the improvement of supportive care have made ASCT very safe, with a low treatment mortality rate 13,14 . Taiwan is located in southeastern Asia, with a warm humid climate, so ASCT patients in Taiwan might have a high rate of nosocomial infection 15 . Half of the MM patients in Taiwan received ASCT in an isolation room with HEPA filtration, although the efficacy of HEPA filtration for those patients has not been established. Therefore, we conducted a nationwide population-based study to evaluate the benefits and cost of the use of HEPA filtration for MM patients receiving ASCT.

Patients and methods
Study population. We used data from the Taiwan Cancer Registry, Cause of Death Data, and National Health Insurance Research Database (NHIRD). Data retrieval and analysis were carried out in the Health and Welfare Data Science Center (HWDC). Taiwan's NHIRD provides nationwide population-based data for health research. All patients with severe diseases, of which cancers are included, are enrolled in the Registry for Catastrophic Illness Patients (RCIP) and receive copayment exemption under the National Health Insurance (NHI) program. The integration of multiple NHI databases, including RCIP, NHI enrollment files, inpatient and outpatient databases, provides comprehensive information on NHI enrollment and utilization of healthcare resources, including examinations and treatment 16 . Cancer stages and treatment plans are available in the Taiwan Cancer Registry. All the patients' identification has been encrypted and can be analyzed only in the HWDC. This study has been approved by the Institutional Review Board of Taipei Veterans General Hospital (no. 2020-02-019AC). All methods for the study were performed in accordance with relevant guidelines and regulations of Taipei Veterans General Hospital in Taiwan. The institutional ethical committee waived the informed consent form.
Study cohort and study design. We enrolled patients with newly diagnosed MM in Taiwan between January 1, 2000 and December 31, 2017 from diagnosis codes according to the International Classification of Diseases, 9th revision, and Clinical Modification (ICD-9-CM) codes (203, 203.0X, and 203.1X) and 10th revision (ICD-10-CM) codes (C90.X). The diagnosis of MM had to be further verified by the RCIP, in which diagnosis was confirmed by pathologic reports. Patients under age 20, without ASCT, or receiving non-melphalan conditioning regimens were excluded. We identified MM patients who received ASCT in an isolation room with HEPA filtration as the HEPA group. Comparatively, those who received ASCT in a ward without HEPA filtration were identified as the non-HEPA group.
Endpoints. The primary endpoint of the study was treatment-related mortality, which was defined as death within 100 days following receiving ASCT, of which only the first ASCT was analyzed 17 . We used the National Cause of Death Data to identify the date and cause of death. The secondary endpoints included OS, medical expenditures within 100 days, length of stay in hospitals for the treatment course of ASCT, and emergency room visits and readmission within 14 days after discharge. Medical expenditures included all the expenditures of the treatment course of ASCT within 100 days.
Characteristics of the study population. The potential confounders considered in this study include age, sex, comorbidities, including hypertension, diabetes mellitus, chronic obstructive pulmonary disease, coronary artery disease (CAD), heart failure, end-stage renal disease (ESRD), cerebrovascular accidents, liver cirrhosis, and autoimmune diseases, disease stage, and socioeconomic status. Patients' socioeconomic status was categorized by degree of urbanization and level of monthly salary income stratified according to the previous work 18 .

Statistical analysis.
Patients' demographic and clinical characteristics were presented as the total number (n) and proportion (%) for categorical data, and medians and interquartile ranges (IQR) for continuous data. Patients' demographic data were compared by using the chi-square test for categorical variables, and the Mann-Whitney U test for continuous variables.
In the survival analysis, the Kaplan-Meier method was used for estimation of cumulative incidence of mortality, and differences between groups were tested using a log-rank test. Hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated using Cox proportional hazards models, controlling for potential confounding factors in the multivariate model. All factors with p < 0.1 in the univariate analysis were included in the multivariate analysis. Sensitivity analyses were performed using different cutoffs (60 days and one year after ASCT) to evaluate the mortality risk between the HEPA and non-HEPA groups.
Furthermore, to eliminate bias in selection, propensity score matching at a 1:1 ratio using greedy matching techniques was performed to match the HEPA and non-HEPA groups. Propensity scores were calculated using age, sex, comorbidities, stage, degree of urbanization, income level, and medications in a logistic regression model. Data management and all statistical analysis were performed using SAS 9.4 software (SAS Institute Inc., Cary, NC, USA) and STATA statistical software, version 15.1 (StataCorp, College Station, TX, USA). All statistically significant levels were set at p < 0.05. Treatment-related mortality rate. The 100-day mortality rate was 1.5% and 1.0% for the HEPA and non-HEPA groups, respectively. The crude hazard ratio (HR) for the 100-day mortality HEPA group was 1.39 (95% confidence interval [CI] 0.44-4.37, p = 0.576) compared to the non-HEPA group. In the univariate analysis, CAD (HR 5.80) and ESRD (HR 5.46) were associated with 100-day mortality. After adjusting for the variables found in univariate analysis, the 100-day mortality rate still had no difference between the HEPA and non-HEPA groups (adjusted HR 1.65, 95% CI 0.52-5.23, p = 0.399). Additionally, CAD and ESRD were significant risk factors of 100-day mortality in the multivariate analysis (  www.nature.com/scientificreports/ Propensity score-matched analysis. We conducted a propensity score-matched analysis to compare the HEPA and non-HEPA groups. We calculated the propensity scores for the likelihood of receiving ASCT in an isolation room with HEPA filtration using a multivariate logistic regression. We matched patients in the HEPA and non-HEPA groups with a 1:1 ratio (Supplemental Fig. S1). A total of 654 patients were matched. There was no statistically significant difference in the baseline characteristics between the two groups after matching (Supplemental Table S1). We further analyzed the primary endpoint and all secondary outcomes. There was also no statistically significant difference in the 100-day mortality rate (adjusted HR 1.83, 95% CI 0.54-6.27, p = 0.335) between the HEPA and non-HEPA groups. The 60-day, 1-year, and all-cause mortality rate, as well as emergency room visits and readmission within 14 days, were not statistically different (Supplemental Table S2). No overall survival difference between the HEPA and non-HEPA group were observed after matching (p = 0.089, Supplemental Fig. S2). However, the HEPA group had three more days of hospitalization and spent approximately $6500 more compared with the non-HEPA group (Supplemental Table S3).

Discussion
To the best of our knowledge, this is the first population-based study that compares early mortality rates of MM patients receiving ASCT in an isolation room with HEPA filtration (the HEPA group) to those in a standard ward (the non-HEPA group). Our study reveals no difference in 100-day mortality, OS, emergency room visits, or readmission rate between the HEPA and non-HEPA groups receiving ASCT. However, the healthcare cost was higher for the HEPA group, and the time from diagnosis to ASCT was marginally increased (p = 0.073).
Our results may help hematologists and healthcare administrators more appropriately allocate HSCT resources. We have systematically reviewed existing studies regarding ASCT in MM patients. We found many MM patients receiving ASCT in an isolation room with HEPA filtration 8,9,19,20 . Our study also reveals that half of the www.nature.com/scientificreports/ ASCT in MM patients was performed in an HSCT special ward with such facilities, which implies that many hematologists still believe that an isolation room with HEPA filtration can reduce infection and early-mortality rates in those patients. Our study discloses a very low mortality rate for MM patients receiving ASCT whether in an isolation room with HEPA filtration or not. The effect of HEPA filtration was to lower the nosocomial invasive fungal infection rate 12 . However, our previous study indicated low incidence of invasive fungal infection in MM patients 15 . Therefore, the insignificance of treatment-related mortality between the HEPA and non-HEPA group might be related to the low incidence of invasive fungal infection in MM patients.   22 . The 60-day, 100-day, and 1-year mortality rates of the ESRD patients in our cohort were 2.7%, 4.0%, and 8.7%, respectively.
ASCT should be performed with caution in MM patients with CAD. Stillwell et al. reported that TRM and one-year mortality were 5.6% and 15.3% in CAD patients receiving HSCT. Among them, 68.1% received ASCT 23 . Our study reveals that the 100-day mortality rate of CAD patients receiving ASCT was 3.2%, which was much higher than those without CAD. Saad et al. analyzed the records in the Center for International Blood and Marrow Transplant Research database and concluded that hematopoietic cell transplant comorbidity index (HCT-CI), including CAD and renal dysfunction, could predict survival in MM patients receiving ASCT 24 .
We found that the HEPA group had three more days of hospitalization and spent $7250 more for ASCT, in comparison with the non-HEPA group. However, all the short-term outcomes and long-term survival of the two groups had no difference. This population-based study shows that an isolation room with HEPA might not be necessary for MM patients receiving ASCT. Furthermore, ambulatory HSCT or even at-home HSCT, emphasizing on outpatient visits only or early discharge after stem cell transfusion, had been demonstrated to be feasible, safe, and cost-effective in several studies, especially for carefully selected MM patients undergoing ASCT [25][26][27] . A meta-analysis even showed a lower chance of febrile neutropenia and septicemia with outpatient ASCT 28 . The necessity of HEPA filtration for MM patients receiving ASCT should be re-examined.
Some studies have different definitions of early mortality and TRM in MM. Reece reported 100-day and one-year TRM mortality for patients receiving ASCT. Schmidt-Hieber, Kumar, Yin reported 100-day and oneyear TRM for allogeneic transplantation [32][33][34][35] . Schmidt-Hieber, Kröger, Kuruvilla used one year as the cutoff for TRM for patients receiving allogeneic transplantation [29][30][31] . Bringhen, Augustson, Larocca used 60 days as the cutoff for early mortality in MM studies [36][37][38] . We conducted sensitivity analysis showing that neither 60-day, 100-day, nor one-year mortality had any statistical difference between HEPA and non-HEPA groups. These Table 3. Comparison of outcomes between the HEPA and non-HEPA groups. ASCT autologous stem cell transplantation; IQR interquartile range. a The HEPA group received ASCT in an isolation room with highefficiency particulate air (HEPA) filtration.   www.nature.com/scientificreports/ results consistently demonstrate that ASCT in MM is fairly safe whether the patients are treated in an isolation room with HEPA or not. The non-HEPA group was older and had a higher proportion of ESRD. We conducted a propensity score-matched analysis. After matching, all the patient characteristics were similar. Both groups still had no difference in 60-day, 100-day, and one-year mortality, OS, emergency room visits, or readmission within 14 days. The HEPA group nevertheless had a higher medical expenditure and longer length of stay after matching. Saini et al. reported MM patients with t(11;14) had similar outcomes as those with normal cytogenetic and FISH studies in a propensity score-matched analysis 39 . Varma et al. reported that MM patients with 1q + /1p-were at significantly increased risk of progression or death compared to the propensity score-matched comparison group 40 . In the present study, the propensity score was defined as the conditional probability of receiving ASCT in an isolation room with HEPA. The calculated score was used to balance the covariates in the two groups and therefore reduced the bias 41 .
Our study has some limitations. Most of our patients didn't receive genetic studies by FISH, which are required according to the Revised International Staging System 42 . Second, this study has inherent limitations by using administrative data that did not provide smoking status, performance status, disease status prior to ASCT, comprehensive medications, and some essential laboratory data. Finally, the HEPA and non-HEPA groups were not randomly assigned, so confounding factors might exist. However, there was no survival difference between the two groups, although the non-HEPA group was slightly older and more of them had ESRD.
In conclusion, some hematologists believe an isolation room with HEPA filtration can reduce complications of hematologic patients receiving ASCT. We found that about 50% of the MM patients in Taiwan received ASCT in an isolation room with HEPA filtration, and it didn't affect 100-day mortality. This study may help clinicians and healthcare administrators utilize the limited resources of HSCT facilities. Further validation of our findings in other cohorts is warranted.