An introduction to machine learning for classification and prediction.

Classification and prediction tasks are common in health research. With the increasing availability of vast health data repositories (e.g. electronic medical record databases) and advances in computing power, traditional statistical approaches are being augmented or replaced with machine learning (ML) approaches to classify and predict health outcomes. ML describes the automated process of identifying ("learning") patterns in data to perform tasks. Developing an ML model includes selecting between many ML models (e.g. decision trees, support vector machines, neural networks); model specifications such as hyperparameter tuning; and evaluation of model performance. This process is conducted repeatedly to find the model and corresponding specifications that optimize some measure of model performance. ML models can make more accurate classifications and predictions than their statistical counterparts and confer greater flexibility when modelling unstructured data or interactions between covariates; however, many ML models require larger sample sizes to achieve good classification or predictive performance and have been criticized as "black box" for their poor transparency and interpretability. ML holds potential in family medicine for risk profiling of patients' disease risk and clinical decision support to present additional information at times of uncertainty or high demand. In the future, ML approaches are positioned to become commonplace in family medicine. As such, it is important to understand the objectives that can be addressed using ML approaches and the associated techniques and limitations. This article provides a brief introduction into the use of ML approaches for classification and prediction tasks in family medicine.

10-Year Trends in Healthcare Spending among Patients with Schizophrenia in Alberta, Canada.

OBJECTIVES: Schizophrenia is characterized by high levels of disability often resulting in increased healthcare utilization and spending. With expanding healthcare costs across all healthcare sectors, there is a need to understand how healthcare spending has changed over time. We conducted a population-based study using administrative health data from Alberta, Canada, to describe changes in medical complexity and direct healthcare spending among patients with schizophrenia over a 10-year period. METHODS: A serial cross-sectional study from January 1, 2008, to December 31, 2017, was conducted to determine changes in demographic characteristics, medical complexity, and costs among all adults (18 years or older) with schizophrenia. Total healthcare spending and sector-specific costs attributable to hospitalizations, emergency department visits, practitioner billings, and prescriptions were calculated and compared over time. RESULTS: Over the 10-year period the contact prevalence of patients with schizophrenia increased from 0.6% (n = 16,183) to 1.0% (n = 33,176) within the province. There was a marked change in medical complexity with the number of patients living with 3 or more comorbidities increasing from 33.0% to 47.3%. Direct annual healthcare costs increased 2-fold from 321 to 639 million CAD (493 million USD) with a 7-fold increase in medication expenditures over the 10-year time frame. As of 2017, spending on pharmaceutical treatment surpassed hospitalizations as the leading spending category in this population. CONCLUSIONS: Healthcare spending among patients with schizophrenia continues to increase and may be partially attributable to growing rates of multimorbidity within this population. Although promising second-generation antipsychotic medications have entered the market, this has resulted in considerable changes in the distribution of healthcare spending over time. These findings will inform policy discussions around resource allocation and efforts to curb health spending while also improving care for patients with schizophrenia.

A biomarker-guided, prospective, phase 2 trial of pre-emptive graft-versus-host disease therapy using anti-thymocyte globulin.

BACKGROUND AIMS: Intensified immunosuppressive prophylaxis for graft-versus-host disease (GVHD) may be toxic and therefore warranted only in patients at high risk of developing GVHD. In patients who underwent allogeneic hematopoietic cell transplant at the authors' center, high serum soluble IL-2 receptor alpha (sIL-2Rα) and low IL-15 levels on day 7 post-transplant were found to predict a high risk of developing clinically significant GVHD (sGVHD), defined as grade 2-4 acute GVHD or moderate to severe chronic GVHD. METHODS: This was a prospective, phase 2 trial in which high-risk patients (serum sIL-2Rα >4500 ng/L or IL-15 <31 ng/L) received rabbit anti-thymocyte globulin (ATG) 3 mg/kg on day 8 post-transplant. Controls consisted of patients who had their sIL-2Rα/IL-15 levels measured but did not participate in the trial. A total of 68 trial patients and 143 controls were accrued to this study. The primary endpoint was incidence of sGVHD. RESULTS: There was a reduction in sGVHD in high-risk trial patients (received day 8 ATG) compared with high-risk controls (did not receive day 8 ATG) (sub-hazard ratio [SHR] = 0.48, P < 0.05). There was no significant difference between the groups in overall survival or relapse; however, there was a greater incidence of non-GVHD-associated non-relapse mortality in high-risk trial patients (SHR = 3.73, P < 0.05), mostly related to infections. This may be due in part to the biomarkers ineffectively stratifying GVHD risk. CONCLUSIONS: Pre-emptive ATG therapy is both feasible and effective at reducing sGVHD without increasing relapse. Further mitigation strategies are needed to reduce the risk of infection associated with intensified GVHD prophylaxis. This study was registered at ClinicalTrials.gov (NCT01994824).

Hematopoietic cell transplant outcomes after myeloablative conditioning with fludarabine, busulfan, low-dose total body irradiation, and rabbit antithymocyte globulin.

Optimal conditioning and graft-vs-host disease (GVHD) prophylaxis for hematopoietic cell transplantation (HCT) are unknown. Here, we report on outcomes after low toxicity, myeloablative conditioning consisting of fludarabine, busulfan, and 4 Gy total body irradiation, in combination with thymoglobulin and post-transplant methotrexate and cyclosporine. We retrospectively studied 700 patients with hematologic malignancies who received blood stem cells from 7 to 8/8 HLA-matched unrelated or related donors. Median follow-up of surviving patients was 5 years. At 5 years, overall survival (OS), relapse-free survival (RFS), and chronic GVHD/relapse-free survival (cGRFS) were 58%, 55%, and 40%. Risk factors for poor OS, RFS, and cGRFS were (1). high to very high disease risk index (DRI), (2). high recipient age, and (3). cytomegalovirus (CMV)-seropositive recipient with seronegative donor (D-R+). The latter risk factor applied particularly to patients with lymphoid malignancies. Neither donor other than HLA-matched sibling (7-8/8 unrelated) nor one HLA allele mismatch was risk factors for poor OS, RFS, or cGRFS. In conclusion, the above regimen results in excellent long-term outcomes. The outcomes are negatively impacted by older age, high or very high DRI, and CMV D-R+ serostatus, but not by donor unrelatedness or one HLA allele mismatch.

Mohs Micrographic Surgery Dermatopathology Concordance in Canada: A Single-Institution Experience.

BACKGROUND:: Mohs micrographic surgery (MMS) is a surgical modality that achieves high cure rates of nonmelanoma skin cancers but is dependent on accurate histologic examination of surgical margins. Therefore, quality assurance is essential to ongoing assessment of histological margins. OBJECTIVES:: To prospectively determine the concordance rate between a Mohs surgeon (MS) and dermatopathologist (DP) with respect to tumour status (ie, present or absent) and tumour type. Secondary end points were to determine the relationship between discordant interpretations and slide quality and to assess the feasibility of using an electronic webform for data collection. METHODS:: Ten percent (10%) of the planned MMS cases between January 2015 and March 2016 were randomly selected by a histotechnologist at the start of each month. The MS and DP were blinded to the chosen cases, and slides were reviewed independently at the beginning of the following month. Data were collected using an online webform. A blinded third party determined if there were discrepancies in interpretation, and any discordant slides were reviewed together and a consensus was reached. RESULTS:: A total of 270 slides from 54 total cases were reviewed. The overall tumour status concordance rate was 93.6%. Cohen's κ was 0.86. Tumour type concordance was 98.9%. No discrepancy required a change in patient care. All discrepant slides were from cases that required multiple stages. CONCLUSIONS:: This is the first study looking at MS-DP concordance in Canada, and our findings support the MS acting as his or her own pathologist.

Community paramedic point of care testing: validity and usability of two commercially available devices.

BACKGROUND: Community Paramedics (CPs) require access to timely blood analysis in the field to guide treatment and transport decisions. Point of care testing (POCT), as opposed to laboratory analysis, may offer a solution, but limited research exists on CP POCT. The purpose of this study was to compare the validity of two devices (Abbott i-STAT® and Alere epoc®) by CPs in the community. METHODS: In a CP programme responding to 6000 annual patient care events, a split sample validation of POCT against traditional laboratory analysis for seven analytes (sodium, potassium, chloride, creatinine, hemoglobin, hematocrit, and glucose) was conducted on a consecutive sample of patients. The difference of proportion of discrepant results between POCT and laboratory was compared using a two sample proportion test. Usability was analysed by survey of CP experience, a linear mixed effects model of Systems Usability Scale (SUS) adjusted for CP clinical and POCT experience, an expert heuristic evaluation of devices, a review of device-logged errors, and coded observations of POCT use during quality control testing. RESULTS: Of 1649 episodes of care screened for enrollment, 174 required a blood draw, with 108 episodes (62.1%) enrolled from 73 participants. Participants had a mean age of 58.7 years (SD16.3); 49% were female. In 4 of 646 (0.6%) comparisons, POCT reported a critical value but the laboratory did not; with no statistically significant (p = 0.323) difference between i-STAT® (0.9%;95%CI:0.0,1.9%) compared with epoc® (0.3%;95%CI:0.0,0.9%). There were no instances of the laboratory reporting a critical value when POCT did not. In 88 of 1046 (8.4%) comparisons the a priori defined acceptable difference between POCT and the laboratory was exceeded; occurring more often in epoc® (10.7%;95%CI:8.1,13.3%) compared with i-STAT® (6.1%;95%CI:4.1,8.2%)(p = 0.007). Eighteen of 19 CP surveys were returned, with 11/18 (61.1%) preferring i-STAT® over epoc®. The i-STAT® had a higher mean SUS score (higher usability) compared with epoc® (84.0/100 vs. 59.6/100; p = 0.011). There were no statistically significant differences in device logged errors between i-STAT® and epoc® (p = 0.063). CONCLUSIONS: CP programmes can expect clinically valid results from POCT. Device usability assessments should be considered with any local implementation as the two POCT systems have different strengths.

Low Counts of B Cells, Natural Killer Cells, Monocytes, Dendritic Cells, Basophils, and Eosinophils are Associated with Postengraftment Infections after Allogeneic Hematopoietic Cell Transplantation.

Hematopoietic cell transplant (HCT) recipients are immunocompromised and thus predisposed to infections. We set out to determine the deficiency of which immune cell subset(s) may predispose to postengraftment infections. We determined day 28, 56, 84, and 180 blood counts of multiple immune cell subsets in 219 allogeneic transplant recipients conditioned with busulfan, fludarabine, and Thymoglobulin. Deficiency of a subset was considered to be associated with infections if the low subset count was significantly associated with subsequent high infection rate per multivariate analysis in both discovery and validation cohorts. Low counts of monocytes (total and inflammatory) and basophils, and low IgA levels were associated with viral infections. Low plasmacytoid dendritic cell (PDC) counts were associated with bacterial infections. Low inflammatory monocyte counts were associated with fungal infections. Low counts of total and naive B cells, total and CD56(high) natural killer (NK) cells, total and inflammatory monocytes, myeloid dendritic cells (MDCs), PDCs, basophils and eosinophils, and low levels of IgA were associated with any infections (due to any pathogen or presumed). In conclusion, deficiencies of B cells, NK cells, monocytes, MDCs, PDCs, basophils, eosinophils, and/or IgA plasma cells appear to predispose to postengraftment infections.

Immune cell subset counts associated with graft-versus-host disease.

Graft-versus-host disease (GVHD) is a major transplantation complication. The purpose of this study was to measure immune cell subsets by flow cytometry early after transplantation (before median day of GVHD onset) to identify subsets that may play a role in GVHD pathogenesis. We also measured the subsets later after transplantation to determine which subsets may be influenced by GVHD or its treatment. We studied 219 patients. We found that acute GVHD (aGVHD) was preceded by high counts of CD4 T cells and CD8 T cells. It was followed by low counts of total and naive B cells, total and cytolytic NK cells, and myeloid and plasmacytoid dendritic cells. Chronic GVHD (cGVHD) was preceded by low counts of memory B cells. In conclusion, both CD4 and CD8 T cells appear to play a role in the pathogenesis of aGVHD. Generation of B cells, NK cells, and dendritic cells may be hampered by aGVHD and/or its treatment. Memory B cells may inhibit the development of cGVHD.

Identifying Unique Subgroups of High-Cost Patients With Schizophrenia: A Population-Based Study Using Latent Class Analysis.

Schizophrenia does not present uniformly among patients and as a result this patient population is characterized by a diversity in the type and amount of healthcare supports needed for daily functioning. Despite this, little work has been completed to understand the heterogeneity that exists among these patients. In this work we used a data-driven approach to identify subgroups of high-cost patients with schizophrenia to identify potentially actionable interventions for the improvement of outcomes and to inform conversations on how to most efficiently allocate resources in an already strained system. Administrative health data was used to conduct a retrospective analysis of "high-cost" adult patients with schizophrenia residing in Alberta, Canada in 2017. Costs were derived from inpatient encounters, outpatient primary care and specialist encounters, emergency department encounters, and drug costs. Latent class analysis was used to group patients based on their unique clinical profiles. Latent class analysis of 1659 patients revealed the following patient groups: (1) young, high-needs males early in their disease course; (2) actively managed middle-aged patients; (3) elderly patients with multiple chronic conditions and polypharmacy; (4) unstably housed males with low treatment rates; (5) unstably housed females with high acute care use and low treatment rates. This taxonomy may be used to inform policy, including the identification of interventions most likely to improve care and reduce health spending for each subgroup.

Rituximab Toxicity after Preemptive or Therapeutic Administration for Post-Transplant Lymphoproliferative Disorder.

Rituximab is commonly used as prevention, preemption, or therapeutically for post-transplant lymphoproliferative disorder (PTLD) after hematopoietic cell transplantation (HCT). Although it is generally assumed that rituximab toxicity (ie, infections resulting from hypogammaglobulinemia and neutropenia) is negligible in relation to mortality due to PTLD, limited evidence supports the validity of this assumption. We sought to determine the impact of rituximab on immunoglobulin levels, neutrophil count, infection density, and mortality outcomes. This study retrospectively analyzed 349 HCT recipients, 289 of whom did not receive rituximab and 60 of whom received rituximab preemptively or therapeutically at a median of 55 days post-transplantation. IgM, IgG, and IgA levels at 6 months and 12 months post-transplantation were lower in patients who received rituximab compared with those who did not (significant at P < .05 for IgM and IgA at 6 months and for IgM and IgG at 12 months). Rituximab recipients also had a higher incidence of severe neutropenia (<.5/nl) between 3 and 24 months (subhazard ratio [SHR], 2.3; P = .020). Regarding non-Epstein-Barr viral infections/PTLD, the rituximab group had a higher infection density between 3 and 24 months compared with the no-rituximab group (3.8 versus 1.6 infections per 365 days at risk; incidence rate ratio, 2.2; P < .001). The rituximab group also had a higher incidence of fatal infections (SHR, 3.1; P = .026), higher nonrelapse mortality (SHR, 2.4; P = .006), and higher overall mortality (hazard ratio, 1.7; P = .033). There were no significant between-group differences in the incidence of clinically significant graft-versus-host disease, graft failure, or relapse. Based on this study, rituximab given for PTLD is associated with substantial morbidity and mortality. Whether the benefit of preemptive rituximab outweighs the risk remains to be determined. © 2022 American Society for Transplantation and Cellular Therapy. Published by Elsevier Inc.

Health data governance for research use in Alberta.

Alberta has rich clinical and health services data held under the custodianship of Alberta Health and Alberta Health Services (AHS), which is not only used for clinical and administrative purposes but also disease surveillance and epidemiological research. Alberta is the largest province in Canada with a single payer centralised health system, AHS, and a consolidated data and analytics team supporting researchers across the province. This paper describes Alberta's data custodians, data governance mechanisms, and streamlined processes followed for research data access. AHS has created a centralised data repository from multiple sources, including practitioner claims data, hospital discharge data, and medications dispensed, available for research use through the provincial Data and Research Services (DRS) team. The DRS team is integrated within AHS to support researchers across the province with their data extraction and linkage requests. Furthermore, streamlined processes have been established, including: 1) ethics approval from a research ethics board, 2) any necessary operational approvals from AHS, and 3) a tripartite legal agreement dictating terms and conditions for data use, disclosure, and retention. This allows researchers to gain timely access to data. To meet the evolving and ever-expanding big-data needs, the University of Calgary, in partnership with AHS, has built high-performance computing (HPC) infrastructure to facilitate storage and processing of large datasets. When releasing data to researchers, the analytics team ensures that Alberta's Health Information Act's guiding principles are followed. The principal investigator also ensures data retention and disposition are according to the plan specified in ethics and per the terms set out by funding agencies. Even though there are disparities and variations in the data protection laws across the different provinces in Canada, the streamlined processes for research data access in Alberta are highly efficient.

Emergency medical services response time and mortality in an urban setting.

BACKGROUND: A common tenet in emergency medical services (EMS) is that faster response equates to better patient outcome, translated by some EMS operations into a goal of a response time of 8 minutes or less for advanced life support (ALS) units responding to life-threatening events. OBJECTIVE: To explore whether an 8-minute EMS response time was associated with mortality. METHODS: This was a one-year retrospective cohort study of adults with a life-threatening event as assessed at the time of the 9-1-1 call (Medical Priority Dispatch System Echo- or Delta-level event). The study setting was an urban all-ALS EMS system serving a population of approximately 1 million. Response time was defined as 9-1-1 call receipt to ALS unit arrival on scene, and outcome was defined as all-cause mortality at hospital discharge. Potential covariates included patient acuity, age, gender, and combined scene and transport interval time. Stratified analysis and logistic regression were used to assess the response time-mortality association. RESULTS: There were 7,760 unit responses that met the inclusion criteria; 1,865 (24%) were ≥8 minutes. The average patient age was 56.7 years (standard deviation = 21.5). For patients with a response time ≥8 minutes, 7.1% died, compared with 6.4% for patients with a response time ≤7 minutes 59 seconds (risk difference 0.7%; 95% confidence interval [CI]: -0.5%, 2.0%). The adjusted odds ratio of mortality for ≥8 minutes was 1.19 (95% CI: 0.97, 1.47). An exploratory analysis suggested there may be a small beneficial effect of response ≤7 minutes 59 seconds for those who survived to become an inpatient (adjusted odds ratio = 1.30; 95% CI: 1.00, 1.69). CONCLUSIONS: These results call into question the clinical effectiveness of a dichotomous 8-minute ALS response time on decreasing mortality for the majority of adult patients identified as having a life-threatening event at the time of the 9-1-1 call. However, this study does not suggest that rapid EMS response is undesirable or unimportant for certain patients. This analysis highlights the need for further research on who may benefit from rapid EMS response, whether these individuals can be identified at the time of the 9-1-1 call, and what the optimum response time is.

Primary Care: The Actual Intelligence Required for Artificial Intelligence to Advance Health Care and Improve Health.

With conversational agents triaging symptoms, cameras aiding diagnoses, and remote sensors monitoring vital signs, the use of artificial intelligence (AI) outside of hospitals has the potential to improve health, according to a recently released report from the National Academy of Medicine. Despite this promise, the success of AI is not guaranteed, and stakeholders need to be involved with its development to ensure that the resulting tools can be easily used by clinicians, protect patient privacy, and enhance the value of the care delivered. A crucial stakeholder group missing from the conversation is primary care. As the nation's largest delivery platform, primary care will have a powerful impact on whether AI is adopted and subsequently exacerbates health disparities. To leverage these benefits, primary care needs to serve as a medical home for AI, broaden its teams and training, and build on government initiatives and funding.

Factors associated with high health care spending among patients with schizophrenia. / Facteurs associés aux dépenses élevées en soins de santé chez les patients atteints de schizophrénie.

INTRODUCTION: Understanding the reasons for the wide variation in health care spending among patients with schizophrenia may benefit the development of interventions aimed at improving patient outcomes and health care spending efficiency. The aim of our study was to determine factors associated with high health care spending in the patient population. METHODS: A serial cross-sectional study used the administrative health records of residents of Alberta, Canada between 1 January 2008 and 31 December 2017 and provincial costing methodologies to calculate total health care spending and sector-specific costs. Factors that modified the odds of being a high cost (i.e. 95th percentile or higher) patient with schizophrenia were estimated using generalized estimating equations. RESULTS: This study captured 242 818 person-years of observations among 38 177 unique patients with schizophrenia. Increased odds of being a high-cost patient were associated with younger age (18-29 years), male sex, unstable housing status and requiring care from multiple medical specialties. The strongest estimated associations between high cost status and comorbidity were for metastatic cancer (OR = 2.26) and cirrhosis (OR = 2.07). In contrast, polypharmacy was associated with a decreased odds of being high cost compared with untreated patients. CONCLUSION: Factors associated with being a high-cost patient are the result of complex interactions between individual, structural and treatment-related factors. Efforts to improve patient outcomes and address rising health care costs must consider the value of allocating resources towards early detection and support of patients with schizophrenia along with the prevention/management of comorbidity.

Risk Factors for the Incidence of and the Mortality due to Post-Transplant Lymphoproliferative Disorder after Hematopoietic Cell Transplantation.

Post-transplant lymphoproliferative disorder (PTLD) is a potentially serious complication that occurs following hematopoietic cell transplantation (HCT), in which B cells transformed by Epstein-Barr virus (EBV) proliferate uncontrollably. It is unknown whether risk factors for the incidence of PTLD are identical to those for mortality due to PTLD, a clinically more important outcome. We sought to determine the risk factors influencing the incidence of PTLD and those influencing mortality due to PTLD in a cohort of 1184 allogenic HCT recipients. All patients were predisposed to PTLD, because their graft-versus-host disease (GVHD) prophylaxis included antithymocyte globulin. The overall PTLD incidence was 9.0%, and mortality due to PTLD was 1.1%. In multivariate analysis, risk factors for PTLD incidence include donor+/recipient- (D+/R-) EBV serostatus (subhazard ratio [SHR], 3.3; P = .002), use of a donor other than an HLA-matched sibling donor (non-MSD) (SHR, 1.7; P = .029), receipt of total body irradiation (TBI; SHR, 3.3; P = .008), and the absence of GVHD (SHR, 3.3; P < .001). The sole risk factor for mortality due to PTLD among all patients was D+/R- serostatus (SHR, 5.8; P = .022). Risk factors for mortality due to PTLD among patients who developed PTLD were use of a bone marrow (BM) graft (compared with peripheral blood stem cells [PBSCs]; SHR, 22.8; P < .001) and extralymphatic involvement (SHR, 14.6; P < .001). Interestingly, whereas the absence of GVHD was a risk factor for PTLD incidence, there was a trend toward the presence of GVHD as a risk factor for PTLD mortality (SHR, 4.2; P = .093). Likewise, whereas use of a BM graft was a risk factor for PTLD mortality, there was a trend toward use of a PBSC graft as a risk factor for PTLD incidence (SHR, 0.44; P = .179). Some risk factors for the incidence of PTLD are identical to the risk factors for mortality due to PTLD (ie, D+/R- serostatus), whereas other risk factors are disparate. Specifically, TBI was identified as a risk factor for PTLD incidence but not for PTLD mortality; the absence of GVHD was a risk factor for PTLD incidence, whereas the presence of GVHD was possibly a risk factor for PTLD mortality; and receipt of a PBSC graft was possibly a risk factor for PTLD incidence, whereas receipt of a BM graft was a risk factor for PTLD mortality.

Donor serostatus has an impact on cytomegalovirus-specific immunity, cytomegaloviral disease incidence, and survival in seropositive hematopoietic cell transplant recipients.

More cytomegalovirus (CMV)-specific T cells are transferred with grafts from CMV seropositive than seronegative donors. We hypothesized that seropositive recipients of grafts from seropositive donors (D+R+) have higher counts of CMV-specific T cells than seropositive recipients of grafts from seronegative donors (D-R+), and that this is clinically relevant in the setting of in vivo T cell depletion using rabbit-antihuman thymocyte globulin (ATG). We reviewed charts of 298 ATG-conditioned, seropositive recipients for CMV reactivation (pp65 antigenemia or CMV DNAemia above institutional threshold for preemptive therapy), recurrent CMV reactivation, CMV disease, and death. In 77 of these patients, we enumerated CMV-specific T cells. Median follow-up was 564 days. CMV-specific CD4+ and, to a lesser degree, CD8+ T cell counts were higher in D+R+ than D-R+ patients. D+R+ patients had lower cumulative incidence of CMV reactivation (21% versus 48%, P < .001), recurrent reactivation (4% versus 15%, P = .003), CMV disease (3% versus 13%, P = .005) and mortality (42% versus 56%, P = .006). We conclude that in the setting of in vivo T cell depletion using ATG, seropositive donors should be used for seropositive recipients. For scenarios where only seronegative donors are available, strategies to improve CMV-specific immunity (e.g., donor vaccination) should be explored.

High rabbit-antihuman thymocyte globulin levels are associated with low likelihood of graft-vs-host disease and high likelihood of posttransplant lymphoproliferative disorder.

Rabbit-antithymocyte globulin (ATG) given with conditioning has the potential to decrease the likelihood of graft-versus-host disease (GVHD) or graft failure and to increase the likelihood of relapse or infections. After a given ATG dose, serum ATG levels are variable. Here we determined ATG levels on days 7 and 28 in 153 patients whose conditioning included 4.5 mg/kg ATG (thymoglobulin). Median follow-up was 547 days (range: 14-1519, minimum for patients who have not died, relapsed, developed second malignancy, or had graft failure, 365). Both high day 7 levels and high day 28 levels were associated with low likelihoods of grade II-IV acute GVHD and chronic GVHD needing systemic immunosuppressive therapy, and a high likelihood of posttransplant lymphoproliferative disorder (PTLD). Patients with day 7 ATG levels above 0.803 mg/L had 0.52-fold risk of developing chronic GVHD needing systemic therapy (P = 0.012) and patients with day 7 ATG levels above 1.436 mg/L had 5.84-fold risk of developing PTLD (P = 0.001) compared to patients with lower ATG levels. There was no association of ATG levels with relapse, death, or non-PTLD infections. Association with graft failure could not be evaluated due to only 4 graft failures in the cohort. In conclusion, patients with slow clearance of ATG have a low risk of GVHD, but a high risk of PTLD. The clearance of this relatively low dose of ATG does not impact the likelihood of relapse, death, or non-PTLD infections.

Antirelapse effect of pretransplant exposure to rabbit antithymocyte globulin.

It remains unknown why rabbit antithymocyte globulin (ATG; Thymoglobulin) has not affected relapse after hematopoietic cell transplantation (HCT) in randomized studies. We hypothesized that high pre-HCT ATG area under the curve (AUC) would be associated with a low incidence of relapse, whereas high post-HCT AUC would be associated with a high incidence of relapse. We measured serum levels of ATG capable of binding to mononuclear cells (MNCs), lymphocytes, T cells, CD4 T cells, or CD33 cells. We estimated pre- and post-HCT AUCs in 152 adult recipients of myeloablative conditioning and blood stem cells. High pre-HCT AUCs of MNC- and CD33 cell-binding ATG were associated with a low incidence of relapse and high relapse-free survival (RFS). There was a trend toward an association of high post-HCT AUC of lymphocyte-binding ATG with a high incidence of relapse and low RFS. High pre-HCT AUCs were also associated with faster engraftment and had no impact on graft-versus-host disease (GVHD) or fatal infections. High post-HCT AUCs were associated with a low risk of GVHD, seemed associated with an increased risk of fatal infections, and had no impact on engraftment. In conclusion, pre-HCT AUC seems to have a positive, whereas post-HCT AUC seems to have a negative, impact on relapse.