Don’t put your Patient in a Box

The fourth rule of Expensive Scare Medicine is:

If you measure something and it is not normal, make it normal, if it is safe to do so.

Rule 4 is a tried-and-tested heuristic for looking after critically ill patients – it is useful so long as the last seven words are remembered with due emphasis.

‘Euboxia‘ is the pathophysiological state whereby ‘all boxes on a pathology print-out are in the normal range’ (Reade, 2009). Many people working in critical care settings experience an insatiable desire to achieve this state for their patients. Unfortunately finding a blood test abnormality does not mean that correcting it will make the patient better.

An example I came across recently concerns the ‘common knowledge’ (in other words, there is no convincing evidence that it is true) that having more than two relatives at the bedside of an ICU patient is a poor prognostic indicator (Reade, 2009). Unfortunately for euboxophiles, asking relatives to leave does not help the situation! The converse also appears to be true. A paucity of relatives relative to the severity of illness is suggestive of a bad outcome. Reade has termed this ‘relative relative insufficiency’, and correcting it by dragging relatives to the bedside doesn’t seem to be beneficial. Bearing this in mind, I’ll leave to you to decide if corticosteroid treatment is an efficacious therapy for ‘relative adrenal insufficiency’ in septic shock…

Indeed, euboxia may not just be unhelpful, sometimes it can put your patient in a box – a coffin-shaped box. Here are some examples of when the quest for euboxia can lead to nothing but trouble:

• lung protection strategies aimed at preventing ventilator-associated lung injury in acute respiratory distress syndrome (ARDS) don’t work if you try to correct the PaCO2 – permissive hypercapnia and moderate acidosis (e.g. pH >7.1) are acceptable and often necessary. [See Pulmonary Puzzle #006 Q4]
• Allowing PaCO2 to normalise after intubating a patient with severe metabolic acidosis can be fatal. Patients with severe illnesses like diabetic ketoacidosis or septic shock hyperventilate (resulting in hypocapnia) to compensate for their underlying metabolic acidosis. This has to be maintained when they are started on mechanical ventilation. [See EMCrit Podcast 3 – Intubating the patient with Severe Metabolic Acidosis]
• Toxicologists often find themselves embracing abnormality. Systemic alkalinsation with sodium bicarbonate or hyperventilation of the intubated patient (to pH 7.5-7.55) is a mainstay of the management of severe tricyclic antidepressant overdose. [See Toxicology Conundrum #022] ‘Dysboxia’ is also useful in the setting of salicylate overdose. Urinary alkalinsation with sodium bicarbonate increases the renal clearance of aspirin, and again, when intubated, hyperventilation should be maintained to prevent catastrophic effects from uncovering a decompensated metabolic acidosis.
• Over-enthusiastic correction of abnormalities that ‘should’ be ‘normal’ may cause harm. The patient with chronic hyponatremia runs the risk of central pontine myelinolysis if they are aggressively treated with hypertonic saline. [See NEJM Clinical Image – Central Pontine Myelinolysis] Similarly, excessive correction of hypertension, particularly in the patient who has just had a stroke, may result in cerebral hypoperfusion if the patient can no longer autoregulate blood flow at ‘normal’ blood pressures.
• In patients with penetrating trauma, excessive fluid resuscitation should generally be avoided. It may be better to aim for a ‘low-to-normal’ blood pressure (e.g. MAP 65 mmHg) that is adequate for organ perfusion and reduces the risk of further catastrophic hemorrhage prior to operative intervention. [See EMCrit Podcast 12 – Trauma Resus: Part I]
• In most critically ill patients with anaemia, transfusing significantly above a hemoglobin of 70 g/L does not seem to improve outcomes, and may lead to transfusion-related complications.
• Over-oxygenation of the patient with chronic obstructive pulmonary disease (COPD) reliant on hypoxemic drive is an oft-cited pitfall in critical care. Aiming for a ‘normal’ oxygen saturation can be harmful in other settings as well. Oxygen therapy for paraquat toxicity, or the oncology patient treated with bleomycin, may actually increase the free radical damage caused by reactive oxygen species (ROS) and contribute to severe pulmonary fibrosis. Aiming for oxygen delivery sufficient for, but not in excess of, organ requirements may be beneficial – even with sub-normal oxygen saturations (e.g. SaO2 in the high 80s). [see Oncology Quandary #003 Q4]

Indeed, ‘normality’ is an elusive concept. ‘Normal’ people can have values that lie outside of the ‘normal’ range (for instance, if the reference interval is +/- 2 standard deviations from the mean ‘only’ 95% of the ‘normal’ population will fall within this range). Being outside of the ‘normal’ range might simply mean that, yes, you are part of a bell curve and that, yes, we humans are a diverse lot – it doesn’t necessarily mean you are sick. Furthermore, ‘normal’ values can lead us astray. A patient with a ‘normal’ creatinine can still have acute kidney injury. A patient with a ‘normal’ white count and C-reactive protein (CRP) can still have septic shock or osteomyelitis. And finally, it is often difficult know exactly what ‘normal’ is – normality is sometimes simply a matter of convention and tradition.

After all, what exactly is health? What exactly is disease?

Conventions and traditions, I suppose, work blindly but surely for the preservation of the normal type…

Ford Maddox Ford, from The Good Soldier

Being guided by laboratory tests in preference to clinical judgment is a fool’s game. A game we can avoid playing by striving to truly understand the pathophysiology underlying our patients’ illnesses. Instead of looking at a patient and his or her illness as an impenetrable ‘black box’ we can think about what is happening inside the box and why. In doing so we may better help our patients, even when the evidence is contradictory or when the clinical picture is muddled.

Facts alone will not be of much service to you unless studied in connection with others and with the phenomena displayed during life

William Osler 1885

If you have a good example of when ‘euboxia goes bad’ leave a comment!

• Reade MC (2009). Should we question if something works just because we don’t know how it works? Critical Care and Resuscitation, 11 (4), 235-6 PMID: 20001869