Impact of Outpatient Administration of Cytarabine Consolidation Chemotherapy on Hospital Length of Stay in Adults with Acute Myeloid Leukemia

Background: Administration of 3-4 cycles of high/intermediate dose cytarabine (H/IDAC) reduces relapse of acute myeloid leukemia (AML) and bridges patients (pts) to allograft. There is no accepted standard model of safely delivering H/IDAC in these pts. To reduce reliance on inpatient (inpt) chemotherapy administration and safely avoid or forestall hospitalization during cycles, we developed an outpatient (outpt) protocol to deliver H/IDAC. Pts were assigned outpt or inpt treatment at their hematologist's discretion. We aimed to evaluate the benefits and safety of our model in this retrospective observational study.

Methods: Pts diagnosed with AML and treated with intensive induction chemotherapy between 2015 and 2020 were identified; included pts received ≥1 consolidation cycle of 1-3g/m² cytarabine over 3 hours BID, dose adjusted for age and renal function, on days 1, 3 & 5 of a planned 28-day cycle.

Outpts received H/IDAC via computerized ambulatory delivery device (CADD) pump and central venous access device (CVAD), attended days 1-6, then 3 times a week for blood product support until count recovery; elective readmission was not mandated. Pts were educated to present to hospital if unwell. Inpts were admitted for H/IDAC but could be discharged prior to count recovery. Antimicrobial prophylaxis consisted of antifungal agents and valaciclovir.

The primary endpoint was mean duration of hospital admission (LOS) per consolidation cycle. As pts may have been treated on an inpt and/or outpt basis, generalized estimating equations were used to compare LOS clustered by pt. Secondary outcomes included death during H/IDAC; intensive care unit (ICU) admission; time to admission; blood product use; cycle length; bacteremia; time to count recovery; relapse free survival (RFS) and overall survival (OS).

Results: Of 109 pts identified, 62 pts were included and 47 pts excluded (not AML=3; induction death=7; refractory leukemia=14; alternative=13; no consolidation=10).

Median age was 55.2 years (range 18.2-70.3); 31 (50%) males; ELN 2010 fav=23 (37.1%), Int-2/3 = 30 (48.4%), Adv=8 (12.9%); denovo AML in 60 (96.7%); and therapy related in 2 (3.2%). FLT3-ITD, NPM1 and bi-allelic CEBPA mutations detected in 11 (17.7%), 27 (43.5%) and 2 (3.2%) pts, respectively. Fifty-nine pts (95.1%) received 7+3 idarubicin induction and 7 (11.3%) received midostaurin. A total of 131 cycles of H/IDAC were administered; 81 (61.8%) were commenced as outpt and 50 (38.2%) as inpt. Twenty-seven (43.6%) pts commenced all cycles as an outpt.

Those commencing consolidation as an outpt spent, on average, 10.3 fewer days in hospital per cycle. The mean (SD) LOS for outpt and inpt was 11 (8.3) versus 21.3 (8) days (p<0.001), respectively. For each consolidation cycle, outpts spent less time in hospital (mean, [SD]): cycle 1, 12.6 (10.3) vs 20.7 (8.0) days (RR 1.6, p=0.007); cycle 2, 11.6 (7.2) vs 21.4 (8.3) days (RR 1.8, p=0.001); cycle 3, 8.4 (5.8) vs 24.6 (7.4) days (RR 2.9, p<0.001).

There were no deaths among inpts and one death in outpts (septic shock). ICU admission occurred in 8/81 (9.9%) cycles in outpts vs 1/50 cycles (2%) in inpts (p=0.12). Mean ICU LOS for outpts vs inpts in cycle 1 was 17.6 days vs 1.6; cycle 2, 5.1 vs 0 days; and cycle 3, 4.2 vs 0 days. Bacteremia was detected in 41 cycles (31.3%), with 38 isolates (gram pos=17, gram neg=21) in 31 outpt cycles (38.2%) and 15 isolates (gram pos=6, gram neg=9) in 10 inpt cycles (20%).

Median time to admission in outpts was 14 days (IQR, 3-21); 15 (18.5%) cycles did not result in pt admission. There were no differences in timing of admissions by cycle number nor time-point to prompt elective admission (Fig.1).

There was no difference in packed cell or platelet transfusion requirement between inpts and outpts. Time to blood count recovery was similar. No improvement was seen in cycle length. The proportion undergoing allogeneic transplant in first remission was 37.1%, and at any time, 59.7%. Over a median follow up of 3.3 years, median RFS was 2 years and OS, not reached.

Conclusion: In selected pts, outpt based AML H/IDAC consolidation therapy significantly reduced hospital length of stay compared with inpts. There was no impact on cycle length. Bacteremia and ICU admission rates were higher in outpts, and strategies such as remote monitoring or pre-emptive hospitalization could be considered. Our findings suggest outpt consolidation should be considered in a supported clinical setting.