Figure 26.1 Comparison of survival to hospital discharge in studies of patients with haematological malignancy (HM) admitted to intensive care unit (ICU), by the median year of the study period. HM, patients with haematological malignancy; HM (MV), mechanically ventilated patients with haematological malignancy.
Decision to admit
Early studies with high mortality led to interest in identifying factors that would help predict poor outcome and avoid inappropriate ICU admission. No variable in isolation is sufficiently reliable to solely exclude an individual patient from admission.
Indicators of a lower mortality risk:
Bacterial infection as the reason for admission Uninformative at admission:
Disease-related factors (e.g. type of HM, remission status, presence or length of neutropenia) do not predict short-term survival.
Severity-of-illness scores, e.g. Acute Physiology and Chronic Health Evaluation (APACHE) II score. Although the strongest predictors of outcome, they can only be applied after admission.
Autologous stem cell transplantation (Auto-SCT). In a systematic review, survival to hospital discharge in SCT patients and mechanically ventilated SCT patients improved from 12–23% and 4–6% pre-1998 to 30–61% and 18–26% post-1998 [4]. When analysed as a risk factor, Auto-SCT does not appear to provide further prognostic information.
Indicators of a higher mortality risk:
Increasing age (some but not all studies)
Poor performance status prior to hospital admission (from studies of unselected cancer patients)
Number and severity of organ failures at point of admission
Allogeneic SCT with respiratory failure and hepatic impairment (bilirubin ≥68 µmol/L) or uncontrolled graft-versus-host disease
A number of additional factors should be taken into consideration:
Understand local guidance on withdrawal or withholding of life-sustaining therapy. This includes the local institution’s policy (if available) and the professional and legal principles applicable to that state or country. The following discussion reflects recent guidance from the UK’s General Medical Council [5]. While differences exist to (and within) the USA, a recent review suggests that a shared decision-making model is increasingly being adopted [6].
Understand local, up-to-date audit data on outcomes of HM patients when deciding on the likely benefit of intervention. A limitation of published survival data is its applicability an individual unit, because of differences in patient selection criteria, type of patient and level of ICU support provided.
Is the patient sick enough?
Better survival has been noted in myeloma patients with ICU admission early in the course of a hospital admission. Worse survival is associated with late admission [7] and more advanced multi-organ failure, while in a prospective study of cancer patients referred for ICU admission, 21% of patients considered too well for ICU admission died before hospital discharge [8].
Patients with HM have higher severity-of-illness scores (indicating greater physiological derangement) at admission to ICU than general ICU patients, and this predicts higher mortality. This risk factor for poor outcome may potentially be modified by admitting HM patients to critical care units at an earlier stage. The use of physiological early warning scores and outreach services to trigger critical care assessment may reduce APACHE II scores and mortality, suggesting that this approach is beneficial [9].
Prompt admission allows non-invasive ventilation and intensive monitoring of physiological parameters before the development of irreversible organ failure.
Is the patient too sick?
In the same study of referred cancer patients, 30-day survival was 26% among the patients considered too sick for ICU admission [8] although similar survival has been reported in unselected patients refused ICU admission because too sick
Only gold members can continue reading. Log In or Register to continue
