Evidence based treatment of “the dwindles” (persistent mild hypotension)
(This post was also published at CriticalCareNow on 11/1/2020)
In many ways de-escalating critical care is more challenging than initiating it. Approximately 27% of ICU patients require vasopressors during their stay, and for many, it can be a barrier to discharge. One term in the medical patois -'The Dwindles’ - aptly describes this clinical scenario: a patient who is ostensibly recovering from critical illness but whose blood pressure plummets when the low dose of vasopressor is weaned off. What can we do to treat 'The Dwindles’?
There are two common options: give a fluid bolus or continue vasopressors and wait. The first - more fluids - is usually not the best option. Few patients are hypovolemic after a few days in the ICU, and generally the effects of a fluid bolus are transient. Being patient is fine, but there are opportunity costs to waiting: risk of central line infections, more deconditioning while in the ICU, and greater cost of hospitalization. What else can we do? Here I discuss four strategies that can help liberate patients from vasopressors:
1.Choose a lower blood pressure goal
The Surviving Sepsis campaign recommends that patients with sepsis receive vasopressors to maintain a MAP goal of >65 mmHg. Three RCTs demonstrate that lower MAP goals may shorten the duration of vasopressor administration.
The SEPSISPAM Trial - a multi-site French study enrolled 776 patients with septic shock and randomized them to MAP goals of 65-75 or 80-85 mmHg. There was no difference in the primary endpoint of 90 day mortality, but patients randomized to the higher MAP goal were on vasopressors for 1 day longer (4.7 compared to 3.7 days) and had more than double the rate of atrial fibrillation.
The OVATION Trial - a multi-site Canadian study enrolled 118 patients with vasodilatory shock and randomized them to MAP 60-65 versus 75-80 mmHg. They found no difference in mortality between groups, though among elderly patients mortality was lower in the lower MAP group. Patients in the higher MAP group received vasopressors for a median of 2 days longer (5 vs 3 days).
The 65 Trial - a recent UK study enrolled 2600 elderly patients with vasodilatory hypotension receiving vasopressors and randomized them 1:1 to either a MAP goal of 60-65 or usual care. They found no difference in the primary endpoint of 90 day mortality but found that the median duration of vasopressors was shorter by a median of 5 hours (33 vs 38 hours). Importantly, there was no difference in the need for respiratory or renal support, suggesting that a lower blood pressure goal was not associated with greater risk of renal failure or respiratory failure.
Taken together, these studies suggest that we can safely use a target MAP of > 60 mmHg in most patients, and doing so may hasten liberation from vasopressors.
2. Wean vasopressors in the ‘correct’ order
For patients on multiple vasopressors, the order of tapering can be important. Studies have found a relative deficiency of endogenous vasopressin in patients with septic shock, and vasopressin is routinely used as a second line vasopressor.
Several studies have examined the rates of hypotension after vasopressor weaning. A recent meta-analysis published in 2020 identified 8 studies comparing weaning vasopressin (VP) versus norepinephrine (NE) first. 7/8 studies found a lower rate of rebound hypotension when NE was weaned first. Another meta-analysis identified 9 studies and also concluded that the rate of rebound hypotension was higher when VP was stopped first. Although not all studies reported the duration of vasopressors, 3/7 reported shorter duration of vasopressors when NE was weaned first, with a difference of between 5-33 hours in total vasopressor duration.
| Study | n | Design | Incidence of hypoTN after 1st pressor stopped | |
|---|---|---|---|---|
| VP First | NE First | |||
| Bauer | 50 | Retrospective | 56% | 16% |
| Curtis | 70 | Retrospective | 75% | 26.3% |
| Hammond | 154 | Retrospective | 67.8% | 10.9% |
| Bissell | 61 | Retrospective | 74% | 16.7% |
| Jeon | 78 | Prospective | 22.5% | 68.4% |
| Musallam | 80 | Retrospective | 62.2% | 28.6% |
| Sacha | 585 | Retrospective | 55% | 50% |
| Bredhold | 86 | Retrospective | 31% | 17% |
| Buckley | 69 | Retrospective | 82% | 48% |
| Payne | 81 | Retrospective | 37% | 13% |
Table 1 - Studies of vasopressor weaning sequence comparing norepinephrine (NE) and vasopressin (VP).
Therefore in patients on these two vasopressors, it is reasonable to wean norepinephrine first followed by stopping vasopressin. Although there are disadvantages of this approach (vasopressin is more expensive than norepinephrine), this may be an effective strategy to not only reduce the risk of rebound hypotension, but may also liberate dwindling patients from vasopressors sooner.
3.Consider corticosteroids in septic shock
Many ICU patients are believed to be relatively adrenally insufficient, which isboth a frequent cause of hypotension and associated with increased mortality. Numerous studies have demonstrated benefits to adding corticosteroids, including faster resolution of shock and shorter vasopressor duration. (IMO this is the reason vitamin C + thiamine + steroids has shown some benefits, but that is a rant for another time). If we look at the major RCTs of steroids in patients with sepsis we can see that the use of corticosteroids is consistently associated with a significantly shorter duration of vasopressor infusion.
| Study | n | Duration of | vasopressors (days) |
|---|---|---|---|
| Steroids | No steroids | ||
| Briegel 1999 | 40 | 2 | 7 |
| Annane 2002 | 299 | 7 | 9 |
| CORTICUS 2008 | 499 | 3.3 | 5.8 |
| ADRENAL 2018 | 3658 | 3 | 4 |
| APROCCHSS 2018 | 1241 | 11 | 13 |
Table 2 - Large RCTs of corticosteroids in ICU patients with sepsis.
Steroid use has become more frequent in ICUs in recent years, and a recent meta-analysis found a 7% relative reduction in 28-day mortality in sepsis patients who were treated with corticosteroids. However, there is significant heterogeneity in these studies, and whether or not to use steroids in all critically ill patients requiring vasopressors remains a matter for debate. PulmCrit has an excellent post on this topic. Whether you choose to use corticosteroids for all patients with vasodilatory shock requiring vasopressor or only for a subset (e.g. those requiring two or more vasopressors), there is compelling evidence that this practice shortens duration of vasopressor infusion, and is an effective dwindles treatment. This is not to suggest that all patients on vasopressors require corticosteroids; though there is clear evidence that if you choose to treat sepsis patients with corticosteroids, their need for vasopressors will resolve faster.
4.Add an oral vasopressor… or not
One strategy to wean from an intravenous vasopressor is to add an oral alternative. Several medications have been used for this purpose (midodrine, pseudoephedrine, and droxidopa), though all are off-label uses. Only midodrine is supported by prospective studies, and as we will see, those studies are hardly definitive.
| Oral vasopressor | Pharmacology | FDA status | Evidence | Cost |
|---|---|---|---|---|
| Midodrine | Selective alpha1 agonist | Off label | 5 studies 1 RCT complete 1 RCT in progress |
Medium $1-4 per pill |
| Phenylephrine | Selective alpha1 agonist | Off label | 1 study | Low $0.25 per pill |
| Pseudoephedrine | alpha and beta2 agonist | Off label | 2 studies | Low $0.03-0.05 per pill |
| Droxidopa | synthetic amino acid precursor | Off label | 0 studies | High $25-$75 per pill |
Table 3 - oral agents used for vasopressor effect; adapted from Poyant et al.
Of these options, midodrine was the most evidence-based and the most widely used option. Several single-site observational studies described the use of midodrine, though data on vasopressor duration was not reported in two of them and all found a significant incidence of bradycardia. Despite this meager data, the use of midodrine appeared to be increasing (confession: I used it all the time).
This all changed with the first, and so far only, prospective RCT: the MIDAS trial. The MIDAS trial was an international, double-blind, placebo-controlled trial of Midodrine 20 mg TID given to adults who were resuscitated and had a persistent low dose vasopressor requirement. It enrolled patients with many different etiologies of vasodilatory shock, so it may have been underpowered for specific etiologies like sepsis. Surprisingly, it found no benefit to adding midodrine in terms of duration of vasopressors or any secondary endpoints, though it did find that 7.6% of patients who received midodrine developed bradycardia.
| Study | n | Design | Location | Mean IV Duration | of Vasopressors (days) |
|---|---|---|---|---|---|
| IV vasopressor | IV vasopressor + midodrine | ||||
| Levine, J Crit Care 2013 | 20 | Prospective, observation | SICU | - | - |
| Whitson, Chest 2016 | 275 | Retrospective | 3.8 | 2.9 | |
| Rizvi, CCM 2018 | 663 | Retrospective | SICU, MICU | - | - |
| Balmes, unpublished | 49 | Retrospective | MICU | 1.6 | 2.3 |
| MIDAS, ICM 2020 | 132 | Prospective, placebo controlled RCT | SICU, MICU | 0.97 | 0.94 |
Table 4 - Duration of IV vasopressor infusions in studies comparing oral midodrine to standard of care
Does this mean that we should abandon altogether the use of midodrine for “the Dwindles”? The jury is still out and more RCTs are ongoing.
Conclusions
Persistent hypotension requiring vasopressors is a common situation in the ICU. Doing “nothing” is acceptable but entails exposing patients to increased risk of infections and delays their recovery. Four strategies are available, three of which are supported by evidence: (1) choose a MAP goal of >60, (2) wean NE before vasopressin, and (3) consider adding corticosteroids. Ongoing studies may support a role for oral vasopressors, but the current data is lacking.