An 86 year-old man was referred to ICU because of oliguria, acidaemia and decreased conscious level. He had originally been referred by the general practitioner to the acute general medicine team with unexplained weight loss, malaise and reduced mobility, 19 days previously. He had a longstanding history of bronchiectasis and COPD. He had been able to mobilise independently around his house and garden until suffering a pneumonia several months before this admission, and since required a four-times-daily care package.

During the current admission the patient had been treated for a further pneumonia on the basis of new chest x-ray changes, breathlessness and raised inflammatory markers. He had also undergone a CT chest/abdomen/pelvis for the unexplained weight loss. This was consistent with chronic COPD and bronchiectasis but no other positive findings. A week prior to ICU referral he was found to have acute kidney injury (creatinine 280 µmol/mL, baseline 90 µmol/mL) which had failed to improve. In the 24 hours prior to referral had become progressively drowsy and oliguric.

The patient appeared frail, cachectic and oedematous. He groaned in response to voice and could not follow commands. He had Kussmaul breathing at a rate of about 18 breaths per minute with SaO₂ of 91% on 35% oxygen via facemask. Arterial blood gas showed pH 7.09, pCO₂ 7.1 kPa, pO₂ 9.1 kPa, base excess -9.3 mEq/L, lactate 1.3 mmol/L, glucose 8.7 mmol/L, creatinine 294 µmol/mL. His chest x-ray showed persistent bilateral patchy consolidation. He had a blood pressure of 98/55 mmHg with a pulse of 110 beats/min and cool peripheries. ECG showed sinus tachycardia. He was afebrile. Abdomen was soft and a urinary catheter had drained only 25 mL in the last 4 hours. Other than reduced responsiveness, neurological survey was non-diagnostic.

Evaluation of this patient revealed an elderly man who was severely unwell with acute kidney injury, probable sepsis, and a poor response to treatment to date. This was on the background of chronic suppurative lung disease, and diminished health for several weeks. No specific treatment limitations were in place. His next-of-kin was unaware of any prior expressed wishes and was under the impression that the patient would prefer active treatment. The referring team were of the opinion that intensive care should be considered.

Although no unifying diagnosis for this gentleman’s kidney injury had been identified, a single, rapidly-reversible condition was not apparent. The principal indication for intensive care was for renal replacement therapy for an unknown duration. In view of the status of his neurological, respiratory and cardiovascular systems, it was deemed that airway protection, invasive respiratory support and vasopressor treatment would almost certainly be required. His overall health status made the prospect of survival from a prolonged period of multi-organ support on intensive care highly unlikely. After discussion with the intensive care consultant and the referring consultant it was decided to withhold admission to the intensive care unit. Appropriate family discussions were held. The patient was actively managed on the ward for a further 12 hours, after which fluid management, antimicrobials and further investigation were ceased. He died the following day.

What uncertainties do we face when declining admission to intensive care?

Joel Meyer

Whether to admit or decline a patient for an intensive care bed is one of the most difficult decisions facing the intensivist. When a patient is denied intensive care, they may be being denied a whole range of specialty-specific benefits (such as high ratio nursing care, enhanced access to expertise, rapid diagnostics, harm minimisation strategies) over and above the specific organ therapies that they might require to survive. Most would agree that the decision process primarily centres around the individualised risk-benefit ratio for the patient in question. However, the scarcity of intensive care beds in many countries including the UK means that rationing has become an inescapable aspect of admission decision-making.

Refusal of patients from intensive care is common: in the UK, between 25% and 45% of referred patients are declined ¹. Observational cohort studies in the UK and Europe consistently report an excess mortality in patients refused intensive care compared to those admitted. Table 1 summarises the data from the seven largest published cohorts in which admission decisions have been studied ^2-8.

Study	Site	n	Admitted	Refused
Metcalfe 1997	6 ICUs UK	648	75% Mortality 37%	25% Mortality 46%
Sprung 1999	Single ICU Israel	382	76% Mortality 14%	24% Mortality 46%
Joynt 2001	Single ICU Hong Kong	624	62% Mortality 37%	38% Mortality 61%
Garrouste-Orgeas 2005	11 French ICUs	574	76% Mortality 28%	24% Mortality 39%
Iapichino 2010	11 European ICUs	7877	85% Mortality 28%	15% Mortality 39%
Sprung 2012	11 European ICUs	6796	82% Mortality 24%	18% Mortality 32%
Louriz 2012	Single ICU Morocco	398	64% Mortality 33%	36% Mortality 49%

Table 1. Summary of admission and refusal data from the seven largest published intensive care cohorts. The 28-day mortality rate of each admitted and refused group is shown in italics.

In each of these cohorts, factors including age, dependency, diagnosis, disease severity and other factors have been shown to associate independently with likelihood of admission or refusal to intensive care, as shown in table 2.

Study	Age	Dependency	Diagnosis	Illness severity	Bed state	Referral source
Metcalfe 1997	ü			ü		ü
Sprung 1999	ü		ü	ü	ü
Joynt 2001	ü		ü	ü
Garrouste-Orgeas 2005		ü	ü		ü	ü
Iapichino 2010	ü		ü	ü	ü
Sprung 2012	ü			ü
Louriz 2013	ü		ü	ü	ü

Table 2. Factors found to be independently associated with admission or refusal from intensive care in seven cohorts. 

These data are important because they reflect ‘real-world’ decision-making in parts of the world which broadly resemble UK practice and where intensive care beds are scarce (contrary to North America where the thresholds for admission to intensive care differ greatly). However, several aspects of the research methodology limit the applicability of these findings to individual clinical cases such as the one described above.

Firstly and foremost, the data are from observational studies. Conclusions about the overall effectiveness of intensive care admission cannot be reached because the characteristics of refused and admitted groups of patients are inherently different. Nevertheless, it is unlikely that all of the excess mortality, which is consistently observed across all the studies, can be explained by selection bias alone. It is important to continue to evaluate this kind of data, because, for ethical reasons, it is unlikely that randomised controlled studies of admission to intensive care will ever be performed.

Secondly, when regression analysis is used to identify independently associating factors of a clinical outcome, these factors must be validated in subsequent patient cohorts. Even then, such factors may allow outcome prediction for populations of patients, but caution must be used when prognosticating individual cases such as the patient described.

Thirdly, when interpreting the data, the reason for refusal from intensive care should not be assumed to be that the patient was too unwell to benefit. Most authors categorised reasons for refusal as either ‘too sick’, ‘too well’ or ‘other’. Amongst those five papers which reported separately the 28-day mortality of the ‘too well’ and the ‘too sick’ groups, these ranged from 8 to 14% and from 65 to 90% respectively ^3-7. This suggests that the decision processes being used by clinicians to refuse patients do have reasonable discriminatory power, but also indicates that even those patients from whom intensive care is withheld on the basis of being ‘too well’ have a significant mortality.

Three main models have been cited in the past as potential approaches to intensive care admission decisions, these are: the prioritisation/triage model (useful during increased demand e.g. pandemics), the diagnosis/disease model (admission is dependent on the pathological process), and the clinical parameters model (specified physiological, laboratory and radiological thresholds). These were alluded to in a document from the Society of Critical Care Medicine in 1999 ⁹. Since then, no institution in North America or otherwise has published any guidelines or protocols for navigating this complex area of practice. What does emerge from the literature is that caution must be used not to attach too much weight to individual factors such as age ⁷ or severity of illness ¹⁰ to the decision-making process. The tendency for doctors to be over-pessimistic with regard to admission decisions, particularly for patients with chronic lung disease, has been previously documented ¹¹. In summary, decisions ideally need to be objective and refined but with a degree of flexibility for each individual. Like in many other areas of medicine, the clinical acumen of the doctor is paramount and such acumen is developed over years of experience.

---

Conclusions

Clinicians’ ability to accurately identify those patients within the large and diverse pool of potential candidates who will benefit from admission to intensive care is hampered by a lack of high-quality evidence. The observational studies performed to date suggest that refusal from intensive care is common and associated with an increased mortality. It is impossible to determine whether the outcome of the patient described here might have been different if he had been admitted to intensive care, and whether the decision to refuse admission was correct. Intensive care doctors have to contend with the prospect that their admission decisions may not always be correct, and they should openly acknowledge the substantial uncertainty within this area of their practice.

---

References

1. Marik P. Should age limit admission to the intensive care unit? Am J Hosp Pall Care 2007;24(1):63.

2. Metcalfe M et al. Mortality among appropriately referred patients refused admission to intensive care units. Lancet 1997;350:7.

3. Sprung C et al. Evaluation of triage decisions for intensive care admission. Crit Care Med 1999;27(6):1073.

4. Joynt G et al. Prospective evaluation of patients refused admission to an intensive care unit: triage, futility and outcome. Int Care Med 2001;27:1459.

5. Garrouste-Orgeas M et al. Predictors of intensive care unit refusal in French intensive care units: A multiple-center study. Crit Care Med 2005;33(4):750.

6. Iapichino G et al. Reasons for refusal of admission to intensive care and impact on mortality. Int Care Med 2010;36:1772.

7. Sprung C et al. The Eldicus prospective, observational study of triage decision making in European intensive care units. Part II: Intensive care benefit for the elderly. Crit Care Med 2012;40(1):132.

8. Louriz M et al. Determinants and outcomes associated with decisions to deny or to delay intensive care unit admission in Morocco. Int Care Med 2012;38:830.

9. Guidelines for ICU admission, discharge and triage. Crit Care Med 1999;27(3):633.

10. Edbrooke D et al. Implications of ICU triage decisions on patient mortality: a cost-effectiveness analysis Crit Care 2011;15:R56.

11. Wildman M et al. Implications of prognostic pessimism in patients with chronic obstructive pulmonary disease or asthma admitted to intensive care in the UK within the COPD and asthma study (CAOS): multicentre observational cohort study. BMJ 2007;335:1132.