Hypertensive crisis

"Hypertensive Crisis" is an L1 condition from Section 2.1.8 in the second edition of the CICM Syllabus for the Second Part Examination. "Hypertensive Crisis" also appears in the renal section, where presumably it refers to specific renal-related causes of hypertensive emergencies (eg.  renal artery stenosis and scleroderma renal crisis) or to specific renal complications of hypertensive emergencies (i.e. renal failure due to the hypertension). And then there's phaeochromocytoma in the endocrine section, PRES in the neuro section, and hypertension in the post-op cardiac surgical patient, to say nothing of the extensive collection of antihypertensive pharmacology notes from the First Part exam. To tiptoe around each topic, trying to avoid duplication and overlap, is a road to madness. Instead, this chapter will intentionally celebrate and emphasise those areas of SEO-wrecking content duplication that could improve the reader's revision.

In summary:

No one paper could be singled out as the best source of truth on this subject, and the summary below had to be integrated from multiple sources. A list of recommended reading can be assembled from the syllabus keywords, knowing full well that a long list of recommended readings would be completely pointless for the time-poor candidate:

• Overview: Eliott (2024), or if you can't pirate the fulltext, Stewart (2023) or Rossi et al (2021)
• Epidemiology and aetiology: Paini et al (2018)
• Assessment and management: Jolly et al (2023) or Kulkarni et al (2023); van den Born et al (2019) for genuine medication suggestions
• Outcomes: Siddiqi et al (2023)

The need to finish this revision over the timescale of minutes or seconds should push the reader towards IBCC, where Josh Farkas does the same thing in readable point form. 

"You call that a crisis?"

Definitions of "hypertensive crisis" vary among the Royally Educated Societies. In general, each body of experts appears to have invented its own histrionic terminology for the varying scales of this problem, which include "urgency", "emergency", "crisis", and the confusing term "malignant hypertension" that seems to suggest that the other forms of hypertension are benign. Miller et al (2024), looking at the emergency medicine literature on the subject, produced this table of "inconsistent and sometimes misleading language" used to describe this phenomenon:

Looking at the definitions offered by the Various Learned Societies, one is somewhat reassured, as these are large groups of experts who all read the same evidence and data. The terminology is somewhat more consistent but the numbers, weirdly, are not:

Definition	ESC (2024)	AHA  (2018)	Australian guidelines (2016)
Hypertension	140 / 90	Stage 1: 130/80 Stage 2: 140/90	High-normal: 130/85 Grade 1(mild): 140/90 Grade 2 (moderate): 160/100 Grade 3 (severe): >180/110
Hypertensive urgency	"severe hypertension in patients without clinical evidence of acute organ damage."	"situations associated with severe BP elevation in otherwise stable patients without acute or impending change in target organ damage"	180 / 110 "severe blood pressure elevations ... that are not immediately life-threatening, but are associated with [symptoms] or moderate target organ damage"
Hypertensive emergency	180 / 120 "associated with evidence of new or worsening target organ damage"	180 / 110 "associated with acute HMOD"	220 / 140 "and acute target organ damage or dysfunction is present"
Malignant hypertension	"extreme BP elevation and acute microvascular damage (microangiopathy) affecting various organs"	- no mention	"severe hypertension with retinal haemorrhages and exudates plus papilloedema"
Hypertensive crisis	-no mention	- used colloquially	- used colloquially

As such, the CICM exam candidate, or whatever international reader happens to find this, can rest assured that whatever exam they plan to sit will not expect them to recall the terminology and definitions, because none are universally accepted. For instance, even within Europe, the British and Irish Hypertension Society deride the ESC nomenclature as unnecessarily emotional, noting that "the term ‘hypertensive urgency’ in patients who are asymptomatic is not necessarily useful as this can lead to undue anxiety". They prefer the calming power of the words "acute severe hypertension" instead. Overall, the origin of the term"crisis" is often attributed to Edward Freis (1969), but in fact it appears in the literature as far back as Holder (1929), which was only a few years after the establishment of hypertension as a concept by Theodore Janeway in 1913. Janeway himself was of course characteristically reserved in his language, as befit a gentleman scholar of his era, and would not have debased the dignity of his field with melodramatic terms like "crisis". His own observations of the phenomenon were splendidly understated:

"the medians of the living and the deceased groups do seem to show a certain unfavorable significance of blood-pressures above 200 mm... other things being equal, a systolic blood-pressure persistently above 200 mm. constitutes a certain presumption in favor of termination by uremia or apoplexy."

One will observe from this ancient record that, whatever the wording they use, most authors seem to report a consistent "presumption in favor of termination" associated with the same range of pressures. Somewhere around an SBP of 180-200, things begin to go wrong in a variety of ways. 

Epidemiology of hypertensive emergencies

What is so important about this specific 180 mmHg measurement? To take this a few steps back, in general the definitions of blood pressure normality are based entirely on population averages rather than any sort of physiological threshold, so we should not expect there to be any sort of hard physiological cutoff, below which nothing bad happens. Rather, the data is based on the risk of signiciant cardiovascular events, which are variably defined but which in general meet the pub test for significance (most people at the pup will agree that death, stroke, heart attack or AAA rupture are all meaningful undesirable outcomes). This risk appears to increase with both age and BP measurement. Wherever this is brought up in the papers, the reference is often Lewington (2002), a meta-analysis of about 1 million adults from 61 prospective studies, and the specific findings they refer to are shown below in blurry graphs stolen from the original pdf. As you can see, for the 40-49 year olds, the 10-year mortality rate from ischaemic heart disease is ~1% if their systolic is usually under 120, or about 10% if it is usually over 180; and the risk roughly doubles with each decade of life.
 

One will immediately note the depressing linearity of these plots. There does not actually seem to be a floor for this effect: Rapsomaniki et al (2014), looking at a cohort of about 1,200,000 patients, determined that the risk increase starts from a systolic blood pressure as low as 90 (though the Prospective Studies Collaboration of 2002 found that the risk in the sub-115 categories was extremely low). On reflection, one would have to agree that this is a completely normal pattern of behaviour for a normal distribution (as larger populations will inevitably spread to include weirder outliers), but it sure does interfere with creating meaningful cutoffs for definitions. Fortunately, each time one runs a trial of hypertension management, one typically uses a frequentist dichotomy for patient randomisation, and this calls for a threshold, which subsequent guideline-writers can borrow. On this basis we can say that patients with SBP over 140 seem to benefit from having their blood pressure reduced if they have other cardiovascular risk factors, whereas if the SBP is routinely in excess of 160, the benefit extend even to people without other comorbidities. The cutoffs we employ are therefore tied to outcomes data, which becomes relevant when you start using those cutoffs to decide who gets treatment. 

Which brings us back to the "hypertensive crisis" population. The 180mmHg threshold for systolic blood pressure seems to be the commonly agreed upper value for recommendations and classification systems, with any patients above this value captured into one amorphous category. This has the effect of amplifying the apparent risk, because you also end up analysing the data from patients who have an SBP of 250.  Either way, according to the distributions from Burt et al (1995), this is what about 3-4% of the middle-aged population look like when they walk in to their GP appointment:

But these people are not hurriedly transferred to the nearest ICU. Yes, something should be done about this blood pressure, but it can happen in the outpatient setting with tablets and lifestyle modification; they are not at any sort of immediate risk, their brain will not become oedematous tonight, nor will their kidneys crash, nor will their intima become necrotic. Their risk profile will evolve over decades. The CICM exam candidate would need to be able to make the distinction between these patients and those who require immediate blood pressure control with aggressive therapies. This group is defined by progressive blood-pressure-related organ damage which occurs over the timeframe of days or hours, for whom short term mortality may be extremely high. Katz et al (2009) reported a hospital mortality of 6.9% (for patients whose average length of hospital stay was only 5 days), with a subsequent 90-day mortality of ~4.6% for discharged survivors. This is what originally put "malignant" into hypertension: Keith, Wagener and Hernohan (1928) observed that in their case series of hypertensive patients, those with retinal involvement were all dead by the end of the second year. The term "malignant" was considered appropriate mostly because of the prognosis, but was actually translated  from an expression initially used by Volhard and Fahr in their 1914 treatise on hypertensive nephropathy; the original word was "bösartige",  a word which does not have a direct English equivalent, and which could have been equally well translated as "cruel" or "vicious" (it applies to tumours but is also used to describe malware, or the smile of a villain).

Aetiology of hypertensive emergencies

The attentive reader may at this stage object that mortality measurements for hypertension are meaningless unless they are qualified with a diagnosis. One will obviously have a better chance of a good outcome when one's blood pressure is elevated by some recreational cocaine, as compared to the hypertensive surge that follows brainstem herniation from a massive stroke. This naturally brings us to a conversation about the causes of hypertensive crises, which are a fair topic for some future SAQ. Fair, because hypertensive crises and malignant hypertension seem to have a different distribution of aetiologies as compared to the "usual" spectrum of hypertension, where up to 90% is described as "essential" (i.e. no specific treatable cause is found). The following list is borrowed from a mixed adult/paediatric review paper by Van Why & Pan (2022) and rearranged into a familiar framework:

Vascular causes: Aortic dissection Renal artery stenosis Subarachnoid haemorrhage Infectious causes: HIV Neoplastic causes: Tumours that compress the kidney or renal arteries: Wilms tumour (nephroblastoma), neuroblastoma, lymphoma. Tumours that compress the brainstem Drug-induced: Recreational, eg. cocaine and other sympathomimetics MAOI use Vitamin D intoxication (hypercalcemia) Calcineurin inhibitors (tacrolimus)  Non-adherence to regular meds Withdrawal from CNS depressants Neuroleptic malignant syndrome or serotonin syndrome Idiopathic or idiosyncratic Pre-eclampsia and eclampsia Fibromuscular dysplasia Seizure

Congenital: Polycystic kidney disease Medullary cystic kidney disease Nephronophthisis Von Hippel-Lindau syndrome  Congenital adrenal hyperplasia Gordon syndrome (pseudohypoaldosteronism) Williams syndrome, neurofibromatosis, Turner syndrome, middle aortic syndrome Autoimmune: Scleroderma renal crisis Systemic vasculitis with renal involvement (SLE, HSP, ANCA vasculitis) Haemolytic uremic syndrome Traumatic: TBI with raised ICP Spinal injury with dysautonomia Endocrine: Hyperaldosteronism Phaeochromocytoma or paraganglioma Thyrotoxicosis Hypercalcemia

It is, of course, not essential to make the diagnosis of any specific aetiology here: the need to manage the blood pressure is always a worthy objective in each of these conditions. The organ damage from sustained hypertension is the same, irrespective of whether it is caused by tacrolimus or SAH. 

Pathophysiology of hypertensive emergencies

Keith et al (1928), benefiting from the era when diagnosis was often determined by postmortem, were able to explore the bodies of their malignant hypertensive subjects to determine how this blood pressure elevation had contributed to the cause of their death. To their surprise, they found widespread damage:

"The most important and constant feature of these cases is the extreme diffuseness of the lesion...  Sometimes the kidney, the heart or the brain is more severely injured, but not any organ escapes."

The specific "lesion" they were referring to was a microvascular injury, involving the arterioles of all organ systems. There was marked hyperplasia of the intima and hypertrophy of the media and internal elastic lamina, with perivascular fibrosis and even intimal necrosis.

How did it get like this, one might sigh, looking at this microcirculation. Oh, the state of it.  Using the papers by van den Born et al (2011) and Varounis et al (2017), the following timeline can be constructed:

• Under normal circumstances, increased stretch and shear stress in the microcirculation lead to a compensatory vasoconstriction to regulate blood flow.
• For whatever reason, something interferes with autoregulation, and these small regulatory vessels fail to respond normally. 
• This failure of autoregulation leads to increased blood flow through these vessels, exposing smaller downstream arterioles to a much higher pressure and shear stress than what their usual design specifications would permit.
• These arterioles dilate abnormally and their endothelium is injured by the shear stress.
• This makes these smaller vessels more permeable, and plasma leaks into their intima, manifesting as fibrin deposits over time.
• A sufficiently high blood pressure also affects the perfusion of the vessel wall, which can lead to "necrotising arteriolitis", to borrow a phrase from Ault & Ellrodt (1985).
• The effect of this on the end organs is diminished perfusion. 
• Unfortunately one such organ is the juxtaglomerular apparatus, and poor perfusion here translates into an increased neurohormonal stimulus for the mechanisms that defend blood pressure, leading to a RAAS-driven spiralling increase of blood pressure.
• Apart from this, endothelial injury results in:
  • Oedema of the tissues, due to a purely hydrostatic Starling effect 
  • Platelet activation and microthrombosis ("thrombotic microangiopathy")
  • Microangiopathic haemolytic anaemia
  • As the result of these, also hyperfibrinolysis (leading to haemorrhage on various scales, micro and macro)

Clinical course of hypertensive emergencies

Considering that blood pressure is taken by many people at home, and certainly checked by the GP and triage nurse, one would not be expecting to see the natural course of a hypertensive crisis develop on their watch. Somebody would interfere with it at some stage, presumably by managing it with antihypertensives. Looking back on papers published before effective antihypertensives were developed, such as Koepsell et al (1950), the majority of patients succumbed to the disease within 1-2 years. Stroke, heart failure and renal failure were the main causes of death, but all the organ systems were affected in all of the patients. Which brings us to:

Complications and known sequelae of uncontrolled severe hypertension

To organise these into a familiar framework seems logical:

• A: Epistaxis; loss of airway reflexes due to CNS sequelae
• B: Pulmonary oedema
• C: Myocardial infarction,
       LV hypertrophy,
       Aortic dissection,
       acute mitral regurgitation
• D: Stroke, SAH, intracranial haemorrhage, seizures, PRES, retinal haemorrhage and blindness, carotid and vertebrobasilar dissection.
• E: Hypokalemia, hypernatremia, alkalosis (hyperaldosteronism)
• F: Renal failure
• G: Congestive hepatopathy 
• H: hypercoagulable state, but also hyperfibrinolysis, and MAHA
• I: endothelial activation, raised inflammatory markers

Looking for these end-organ effects is therefore the objective of the assessment:

Assessment of hypertensive emergencies

Excellent, and very similar, assessment sections are available from Benenson et al (2021),  Leiba et al (2016) and Kulkarni et al (2023).

History for the patient with a hypertensive crisis

• Risk factors:
  • Hypertension, compliance with the antihypertensives
  • Pregnancy
  • Recreational drug use
  • Chronic kidney disease
  • Solid organ transplant (tacrolimus, I am looking at you)
• Assessment of clinical symptoms:
  • Headaches, shortness of breath, palpitations and blurred vision, confusion
  • Palpitations, chest pain, abdominal/back pain of dissection
• Assessment of end organ damage:
  • Respiratory failure, haemoptysis, epistaxis
  • Confusion, seizures, visual disturbances
  • Oliguria, orthopnoea, unexplained weight gain
• History that suggests a specific cause:
  • Postprandial flushing or sweating (carcinoid syndrome)
  • Paroxysms of palpitations sweating and headache (phaeo)
  • Daytime somnolence, sleep apnoea
  • Weirdness like acromegaly or known congenital coarctation
  • Smoking, cholesterol, family history (atherosclerosis)
• Assessment of secondary contributors
  • Untreated pain, anxiety, agitation
  • Urinary retention or constipation

Examination findings in the patient with a hypertensive crisis

Though steeped in mostly paediatric lore, the paper by Van Why & Pan (2022) has by far the most comprehensive discussion of what to look for in a hypertensive crisis patient on examination. This was remixed with the classical Examination Medicine from Talley & O'Connor to create the strategy below:

• Airway:
  • Tonsillar enlargement
  • Features of crowding, presupposing a diagnosis of sleep apnoea
• Respiratory:
  • Pulmonary oedema
• Cardiovascular:
  • Radioradial and radiofemoral delay
  • Features of LVH (displaced, forceful apex beat)
  • Gallop rhythm (S3 or S4)
  • Renal and carotid bruits
• Neurological:
  • Fundoscopy *
  • An examination seeking to assess for focal neurology
• Endocrine/exposure:
  • Thyroid examination for toxic goitre
  • Cushingod features
• Renal:
  • Urinalysis (admittedly, a bedside test rather than an examination) for proteinuria or haematuria
• Abdominal/hepatic:
  • Ascites 
  • Pregancy 
  • Renal masses, adrenal masses
  • Abdominal aortic aneurysm or femoral artery aneurysm
• Haematological:
  • Petechial haemorrhages (DIC)
  • Hepatomegaly, pulsatile liver of RHF

*Fundoscopy occupies a weirdly elevated position in the assessment of hypertension mostly by virtue of being an examination of a conveniently exposed and highly functionally important vascular bed. The reason for this goes all the way back to Keith and Wagener, who depended on fundoscopy to discriminate malignant hypertension from the other types. Their definition called for Grade III or Grade IV retinopathy (flame-shaped haemorrhages and cotton-wool spots, or "AV nipping", "silver wiring" and papilloedema). The modern CICM exam candidate would of course revolt against this, as they would not have performed fundoscopy since medical school, and would have trouble locating the nearest working ophthalmoscope in their department. These should be reassured by being reminded that none of the Scientific Societies have included fundoscopy findings in their definitions of malignant hypertension, nor did they make any effort to define it quantitatively.

Investigations for the patient with a hypertensive crisis

• Bedside
  • ECG to look for features of LVH
• Biochemistry
  • Urinary protein
  • EUC/CMP for renal function and hypokalemia of Conn syndrome
  • TFTs for thyrotoxicosis, cortisol for Cushings
  • Troponin and BNP to assess cardiac involvement
  • βHCG to exclude pregnancy
  • 24-hr urinary catecholamines
• Imaging
  • Carotid doppler
  • Renal artery doppler
  • CT brain, to eliminate concerns about ICH
  • CT aortogram, to look for dissection (also incidentally images the adrenal glands and kidneys)
  • MRI brain, to confirm PRES

Though MRI is listed everywhere as the investigation of choice to identify PRES, one must question the wisdom of pushing the confused hypertensive patient into the MRI scanner, considering that the diagnosis of PRES will not make the control of blood pressure any more or less imperative, and that the inevitable necessary general anaesthetic will make the patient a lot more difficult to assess clinically.

Management of hypertensive crisis

Each of the Royally Scientific Societies has a slightly different approach to how they recommend the management for such patients, but in general they all agree that an abrupt drop in the blood pressure is not the goal. "An important principle is that a rapid lowering of BP can be more damaging to the patient than no change at all", quip the authors of the UK recommendations from 2023. The reason for this is the change in organ blood flow autoregulation that occurs with sustained hypertension; chronically tense vessels find it difficult to relax, which means they will remain constricted even as the pressure drops, resulting in hypoperfusion of the organs. Lower it slowly, the scholars caution.

How much should we adhere to this? The evidence was obviously enough for AHA and ESC to be convinced, but most of it is observational and retrospective. For example, one often-quoted supporting study is the aortic dissection case series by Al Adas et al (2018), who found a major increase in stroke risk (specifically, watershed infarcts) in patients whose blood pressure was precipitously dropped. There were nine such patients in the case series of 112; they had a MAP of around 140 in the ED which dropped to 80 in the ICU. Another paper used to support this assertion is O'Mailia et al (1987), which presented three cases of MI following the use of nifedipine, where  SBP was lowered from 180-230 mmHg to 80-85 mmHg. The most robust findings are probably from INTERACT2, where the group of intracerebral haemorrhage patients with aggressive BP control (110-140) had marginally higher rates of renal adverse events (9% vs 4%). 

There are exceptions to this rule, where you do want to drop the blood pressure fairly abruptly, but when one examines these exceptions closely, one begins to question whether they are really exceptional:

• Intracranial haemorrhage: from INTERACT2 (2015), it appears that "larger SBP reductions within the first hour after randomization were associated with lower risks of poor outcome"; but this was not a great heroic leap into normotension. The effect was greatest when an SBP reduction of ≥20 mmHg was achieved in the first hour and was then maintained for 7 days. For an ICH patient with an totally on-brand SBP of 220 mmHg, that would represent a 10% decrease, which is actually smaller than the usually recommended drop. 
• Pre-eclampsia and eclampsia: though there are concerns that the placenta will lose some blood flow from a too-quick drop in arterial pressure, there does not seem to be any real problem identified in the empirical literature. This makes logical sense, as the placental blood flow is one of those rare vascular beds in the body which has basically no intrinsic blood flow autoregulation, and so should be unaffected by the autoregulatory "drift" that makes other organs vulnerable to rapid decreases in pressure. The data in support of an abrupt drop is admittedly scant. All authors who mention this tend to reference Gupta et al (2018), wich was a retrospective observational study that noted a better outcome (1.9% vs 6.4% relative risk of composite endpoint unpleasantness) in patients who had something done for their blood pressure within 60 minutes. The authors did not report the actual drop in BP quantitatively, only that they viewed a reduction to below 160 mmHg as a satisfactory endpoint.
• Aortic dissection is listed alongside ICH and preeclampsia in the 2023 AHA recommendations, as one of the "compelling conditions, requiring rapid lowering of SBP, usually to <140 mm Hg, in the first hour of treatment". In fact their own Table 20 contradicts this with an even more aggressive target ("SBP ≤120 mm Hg should be achieved within 20 min") The attentive reader will have noted that this is the exact strategy that had led to watershed infarcts in the case series by Al Adas et al (2018). It is one thing to reconcile conflicting guidelines, but another entirely to have to handle the conflicting recommendations within the same guideline statement. 

Moreover, the reader must understand that RCTs conducted in hypertensive emergencies states have largely excluded patients with extremely high BP (typically > 220 mmHg SBP and/or >110 mmHg DBP).

Relevant guidelines and evidence in the management of hypertensive crisis

AHA (2017) have very clear positions:

• First hour:    Reduce SBP by no more than 25%.    
• Hours 2-6:    Reduce SBP to 160/100 mmHg.    
• Following 24-48 hrs:  Reduce SBP back to normal.

ESC (2024) refer to van den Born et al (2019) for this part of their guideline ("further comprehensive details on the clinical management of hypertensive emergencies are available elsewhere"). That paper has a very granular breakdown by condition, which is somewhat difficult to follow. In short:

• Malignant hypertension: "Several hours", MAP -20% to -25%
• Ischaemic stroke: 1 hr, MAP -15%
• Haemorrhagic stroke: immediately, SBP < 180
• Hypertensive encephalopathy: immediately, MAP -20% to -25%
• Acute MI, APO: immediately, SBP <140 mmHg
• Aortic dissection: immediately, SBP <120 mmHg
• Pre-eclampsia: immediately, SBP <160 mmHg

Beyond this, there is not a lot of guidance. Each Society offer some lists and tables of drugs considered useful, occasionally even with doses and common side effects, but these will not be useful to anybody who has passed the CICM First Part exam, as their pharmacology knowledge will already contain the necessary material. The position statement from the ESC (van den Born et al, 2019) goes somewhat further by tying the conditions to first and second line agents. These recommendations are:

• Labetalol and nicardipine (or nitroprusside and urapidil) for:
  • malignant hypertension,
  • hypertensive encephalopathy
  • stroke
  • ICH
  • Preeclampsia
• Labetalol and GTN (or urapidil alone) for acute MI
• Nitroprusside or nitroglycerine (with loop diuretic) or urapidil with loop diuretic for acute pulmonary oedema
• Esmolol and nitroprusside or nitroglycerine or nicardipine or labetalol or metoprolol (so, basically anything?) for aortic dissection

So, this urapidil-heavy guideline is great if you live in a country where urapidil is widely available, but that does not include Australia or the US. For a southern hemisphere perspective, one might expect Oh's Manual to come up with something careful deliberate and specific, but no - the 8th edition contains only a chapter on vasodilators and antihypertensives, saying only that "sustained hypertension may be treated acutely with incremental doses or infusions of short-acting drugs such as glyceryl trinitrate, sodium nitroprusside, labetalol, hydralazine, nicardipine or clonidine." The same drugs are then listed in different combinations for each of the conditions commonly associated with hypertension, in a manner that resembles van den Born et al.

This is obviously of little use to the CICM exam candidate.  What follows is an attempt to impose a rigid structure on the answer to the common question, "outline your management of this bösartige blood pressure":

Where a specific antihypertensive agent is preferred first-line, and why:

• Pre-eclampsia, magnesium, because it reduces the risk of seizures.
• Scleroderma renal crisis, captopril, because the mechanism is ACE-related.
• Aortic dissection, β-blockers, because anti-impulse.
• Phaeochromocytoma, not β-blockers, because unopposed α1 effects.
• Acute MI and APO, nitrates, because preload reduction.
• Stroke/ICH, not nitrates, because vasodilation and raised ICP
• Aortic dissection, not nitrates, because reflex tachycardia.
• Dysautonomia, clonidine, because centrally driven.
• Pain, clonidine, because co-analgesic effect with opioids

Where no specific antihypertensive agent is preferred:

• Labetalol can be first line, because:
  • Short acting, half-life 3-4 hrs
  • An IV formulation is available
  • Not renally cleared, so safe in AKI
  • No toxicity beyond the extension of the desired effect, i.e. hypotension and bradycardia
  • Not a cerebral vasodilator
  • Safe in pregnancy
  • Hits both α and β receptors, theoretically protecting any unrecognised phaeochromocytoma patients.
  • Order of magnitude cheaper than esmolol
• Clevidipine should be next in line, because
  • Ultra short acting (minutes)
  • Not renally cleared, so safe in AKI
  • Not a cerebral vasodilator as far as we know
  • Safe in pregnancy
  • Zero effect on cardiac contractility
  • No toxicity beyond the extension of the desired effect

There's nothing that can't be achieved with these two in combination, as they have a synergistic additive effect: clevidipine dilates the resistance vessels in the arterial circulation, and labetalol prevents any regulatory cardiovascular reflexes from antagonising this with increased heart rate and blood pressure.

When one runs out of room for labetalol because the heart rate drops too far, clevidipine will have a maximal effect. Nicardipine, another short-acting IV calcium channel blocker, is a well known alternative in parts of the world where clevidipine cannot reach for geo-economical reasons.

Escalation of antihypertensives

The cost of clevidipine being about $200 AUD per 200mg vial, and of IV labetalol being ~$30 for a 200mg vial, adds up to a potential cost of about a thousand dollars per day, purely for these two (considering the peak dose rates max out at around 800mg/day of clevidipine and about 3000mg/day of labetalol). Beyond that, one needs to consider that the patient whose blood pressure is reliably controlled on these agents is not going to be suitable for the ward. Clearly oral agents need to be incorporated into the program. Moreover, one needs to consider the possibility that the author's haughty exhortation about the undefeated efficacy of clevidipine and labetalol may not apply in 100% of cases, which means a third agent (or even a fourth, a fifth, etc) may need to be added. And because this is CICM and we like to make intelligent management choices, one should have a well-reasoned approach to this pathway, ideally one which does not consist of merely adding more agents of different classes in order of increasing toxicity.

• If the HR is over 60, there is still room for more β blocker.
  • This β blocker should be labetalol, ideally orally.
  • This can begin concurrently with the infusion, with the expectation that eventually the oral labetalol will take over from the IV.
  • The main reason to add another (non-labetalol) β blocker is if the α effects are becoming unpleasant, eg. decreased tone of intestinal genitourinary sphincters. This is not usually a problem.
  • Another potential reason is to make use of a non-β blocker effect, eg. where one may wish to manage the atrial arrhythmia with sotalol, or derive a long-term cardiac failure benefit from carvedilol or nebivolol
  • The main reason not to add more β-antagonist is undesirable additive effects on the AV node. Theoretically, complete heart block may develop.
  • Another reason not to add another different β blocker is the risk of actually increasing the peripheral resistance by β2 antagonist effects.  (43)
  • Mann (2017) describes some scientific basis for choosing among the multiple available β blockers. In short, if you are going to add another beta blocker, make it one which is either another mixed α+β agent like carvedilol, or a vasodilating NO-synthesis-enhancing agent like nebivolol. Steer away from classical agents like metoprolol or atenolol, which will enhance peripheral vasoconstriction.
• If the HR is under 60, there is still room for more vasodilator.
  • Which vasodilator you pick will be strongly dependent on which side effects are the most undesirable. In the modern era  this essentially excludes some agents. For example, there is virtually no situation in which sodium nitroprusside looks sufficiently attractive to be added to the labetalol/clevidipine combination therapy.
    The options include:
  • α1-antagonists like prazocin: there would, in most patients, be very few arguments against their use, making these a good low-toxicity second line option.  
  • ACE inhibitors, or ARBs, are used wherever the renin release is thought to be a major contributor (eg, malignant hypertension, scleroderma renal crisis, renal artery stenosis, and so on). In other circumstances they are probably third line, as AKI is common among hypertensive crisis patients, and most people (logically or not) would be reluctant to deploy them.
  • Nitrates, of the non-cyanide-releasing variety, may be a reasonable addition to the list of agents where there is good reason to venodilate the patient (for example, to reduce the preload on a failing ventricle). The tendency to tachyphylaxis and the various undesirable effects on oxygenation and ICP limit their use to acute hypertension scenarios where the patient is in perpetual APO.
  • Potassium channel activators like hydralazine have a role to play as fourth line agents, as well as short-acting rescue therapy to put out spot fires on the obs chart. These rank alongside the the nitrates as a relatively safe addition to a polypharmacy of antihypertensives. The reason for their placement so far behind sympatholytics ARBs and nitrates is nothing to do with their toxicity. Rather, this is perhaps the only way to use them, as without a host of other agents to block sympathetic and RAAS responses, hydralazine et al would generate a vigorous compensatory response from those systems, and soon become totally useless.
• But you might not want a vasodilator, if:
  • The patient has profound hypoxia (systemic agents like GTN and clevidipine enhance shunt by reversing hypoxic pulmonary vasoconstriction, making them poor choices for the severely hypoxic patient)
  • The patient has raised intracranial pressure (in which case you may wish to steer away from nitrates that are thought to raise it). Whether they actually raise it or not remains debated. The excellent article by Webb (2019) patiently unpacks the available data and concludes that, on balance of things, most vasodilator antihypertensives probably do not have much of an effect on cerebral blood flow or ICP. The impression that nitrates will destroy your brain with harmfully dysregulated blood flow comes from ancient manuscripts such as Cottrell et al (1978), who treated ten elective neurosurgical patients with sodium nitroprusside. Their CPP dropped into a range the moderns would consider unacceptable:
  • The patient has an ischaemic stroke: the same properties that make calcium channel blockers like nicardipine so favourable for raised ICP are also the properties that make them undesirable in the severely hypertensive stroke patient, or so is the theory has it. In that state, apparently, you do want to increase the flow to the flow-dysregulated ischaemic penumbra. Blumenfeld & Laragh (2001) review the fragile evidence base behind this statement and conclude that, on balance of things, centrally acting α2-agoinsts like clonidine seem to maintain cerebral blood flow better than calcium channel blockers.
• If you have maxed out the β blockers and vasodilators are unacceptable:
  • It's not hard to imagine a scenario where β blocker side effects  become embarrassing (the heart rate is too slow or the patient has occasional three-second sinus pauses) and vasodilators are  either forbidden by some aetiological factor or approaching their maximum safe doses.
  • Diuretics  may have a role to play in this group if fluid overload is concurrently a problem (or, conversely, if renal function is satisfactory and the patient could tolerate a lower volume). The specific scenarios that call for these are similar to the indications for nitrates, i.e. acute cardiac failure and poor preload tolerance by a recently ischaemic ventricle. The reason these are not resorted to earlier in the ladder of escalation is similar to hydralazine, i.e one does not wish for the RAAS and sympathetic nervous systems to become over-excited.
  • Centrally acting α2-agonists like clonidine could be introduced cautiously, minding that they will have sedating effects, rebound hypertension, undesirable gut-slowing side effects, and, in the case of methyldopa, the side effect of depression. Their main role is in the management of painful causes of hypertension (eg. aortic dissection), where they collaborate with opioids.