Table 1 Characteristics and key results from reviewed studies

Prospective studies

 2022; [7]

Prospective

87,951 ICU patients, 15% died or received therapy limitation and were further analyzed

36 countries of South Africa

9% (1199 of the 15%)

N/A

N/A

N/A

 1995; [19]

Prospective

26 ICU patients with ADs

USA

16%

N/A

- In 34% AD was discovered 1–12 days after ICU transfer

- Controversies among care providers about meeting ADs in 34%

In 4% [1 patient] physician adhered to AD for most treatments but struggled with withdrawing MV

Cross-sectional studies

 2017; [17]

Cross-sectional study

998 ICU patients

Germany

51% (39% powers of attorney; 29% with ADs)

N/A

40% stated that they had given the relevant document to the hospital, yet such documents were present in the patient’s hospital record for only 23% (88/385 powers of attorney (23%) and 93/293 ADs (32%)

40% of ADs and 44% of powers of attorney that were present in the hospital records were poorly interpretable due to incomplete ADs

 2016; [20]

Cross-sectional study

331 ICU physicians and nurses

Spain

N/A

N/A

- 91% didn't check for ADs and 90% were unaware of ADs

N/A

 1995; [22]

Cross-sectional survey

879 ICU physicians

USA

N/A

N/A

N/A

- 34% of physicians would continue MV against wishes (belief in patient's chance of recovery, patient's best interest, fear of litigation or legality)

- 11% of physicians refused to discontinue MV despite patient request and capability to decide

Systematic review

 2020; [15]

Systematic review

35,717 neurocritically ill  ICU patients

Switzerland

39% ADs and/or healthcare agents

In reference to ADs, care was adapted in 71%, withheld or withdrawn in 58%, and resuscitation was withheld in every fourth patient

N/A

N/A

Retrospective studies

 2021; [11]

Retrospective

400 deceased neuro-ICU patients

Germany

22% (68/310) of patients who died after withdrawal or withholding of life-sustaining therapy

No difference in timing of withdrawal or withholding of life-sustaining therapy, nor in treatment intensity

N/A

N/A

 2021; [16]

Retrospective

229 deceased ICU patients (123 in 2009; 106 in 2019)

Germany

9% in 2009; 26% in 2019

N/A

N/A

N/A

 2020; [14]

Retrospective

489 deceased ICU or IMC patients

Germany

12%

Less likely to receive CPR and MV (by trend, therapy was more often limited and withdrawn)

N/A

N/A

 2019; [13]

Retrospective

16`945 ICU patients (1536 died)

USA

40% of deceased patients

Lower odds of CPR in the last hour of life

N/A

N/A

 2017; [18]

Retrospective

143 diseased stroke patients (focus on life-sustaining treatments)

Germany

29%

- DNR order implemented for one patient referring to AD

- In 21/35 patients additional waiver of ICU measures was implemented, with 11 out of 21 referencing ADs

- Comfort care initiated in 12/35, with 9 out of 12 referencing AD

29% had ADs, but only 25% were available

- Refusal of treatment was respected, when AD was found to be applicable, with exceptions in continuation of nutrition (1 patient) and hydration (3 patients)

- 22/35 objected cardiopulmonary resuscitation, 19/35 mechanical ventilation, and 26/35 nutrition

- 33/35 wanted treatment for pain or discomfort even if it hastens death

- 34/35 patients had therapy limited

- ADs lacked specificity, (16/35 being applicable)

 2014; [12]

Retrospective

477 deceased ICU patients

Germany

13%

- No difference regarding withholding or withdrawing life-sustaining therapies, lengths of ICU stay

- More DNR orders

- CPR received less often

N/A

- 50% (32/64) of ADs were considered valid and factored in treatment decisions

- Patients received MV, nutrition and circulatory support despite refusing it

 2011; [8]

Retrospective

1121 ICU patients

USA

16%

No differences in care, end-of-life management, and outcomes

N/A

N/A

 2010; [3]

Retrospective (on data of a survey)

3746 deceased patients via proxy responses (proportion treated on ICUs not reported)

USA

68% requiring decision-making and lacking the capacity (incl. power of attorney)

- Less likely to receive all care possible and more likely to receive limited treatment

- Patients with a power of attorney were less likely to die in hospital

N/A

- 43% required decision-making in their final days and 70% of them lacked decision-making capacity, especially neurocritically ill

- Subjects with living wills preferred limited (93%) or comfort care (96%)

- The majority desiring limited (83%) or comfort care (97%) had care in line with preferences

 2007; [21]

Retrospective

500 random patients (focus on life-sustaining treatments)

USA

95%

N/A

5% were mislabeled as having an AD

78% preferred to avoid general life support

 2001; [10]

Retrospective

270 ICU patients

USA

27%

No differences in care, except for more DNR orders in first 72 h, shorter ICU stay/costs

N/A

N/A

 1998; [9]

Retrospective

401 ICU patients

USA

5%

No differences with or without ADs

Medical personnel are unaware of AD

11% with ADs rejecting CPR received CPR