The two central themes of care for patients hospitalized for decompensated HF are decongestion and optimization of the therapies recommended for HF, but multiple other goals also need to be met. The coordinated care plan includes evaluation as necessary of the primary etiology of the heart disease and potential aggravating factors that would require specific intervention, both cardiac and noncardiac (34) (see Table 2).

All members of the team should contribute to an initial assessment of the likely outcome both in hospital and after discharge. A profile conferring high risk from factors that do not appear modifiable should trigger early discussion with the patient and family regarding anticipated outcomes and their priorities for remaining quality and quantity of life. Regardless of prognosis, all patients admitted to the hospital should have a designated surrogate decision maker, ideally identified in the outpatient setting and documented during admission. If not already done, however, this designation should be supervised by the inpatient care team.

Inpatient trajectories are primarily defined by the pace and extent of decongestion. Evaluation of the degree of clinical congestion is depicted in Figure 5. The usual goal is for complete decongestion, with absence of signs and clinical symptoms of elevated resting filling pressures (70,78,117,130).

Establishing an effective diuretic regimen is crucial for achieving decongestion. Usually patients require the first dose of intravenous (IV) diuretics at presentation or in the ED, and IV diuretics are continued throughout the hospitalization until effective decongestion warrants transition to oral diuretics before discharge.

Doses of various diuretics are shown in Table 7. If the patient is improving at the expected pace, IV diuretics should generally be continued until optimal decongestion is achieved, and then transitioned to the oral dose estimated for maintenance. If initial improvement is stalled, if there is inadequate improvement, or if the patient is worsening and the patient is still congested, IV diuretics should be increased. Usually, the IV loop diuretic dose can be increased by 50% to 100% until the total diuretic dose exceeds 400 to 500 mg of furosemide equivalent total daily dose. The doses should be increased until a response is apparent. When the response is brisk but transient, the frequency should be increased to 3 or 4 times daily. The DOSE (Diuretic Optimization Strategies Evaluation) trial did not demonstrate any improvement with continuous infusion of IV furosemide, but these patients were also less likely to require dose increases or the addition of a thiazide-type diuretic; additionally, those patients with chronic furosemide equivalents of over 240 mg daily were excluded (77). Consequently, some patients have been observed to respond better to continuous infusion. When high furosemide doses are not effective, metolazone can be added at 2.5 to 5 mg doses once or twice daily. Other thiazide diuretics can be added to loop diuretics and given intravenously if needed (Figure 6).

In this trajectory (Figure 7), the patient has stable vital signs and is making steady progress toward resolution of signs and symptoms of congestion (Table 3), without major complications, and with appropriate alignment of management strategies with goals of care. Some patients, especially those with a clear trigger that has been reversed, such as rapid atrial fibrillation, may have prompt improvement. Such patients may differ from those with chronic myocardial dysfunction.

Diuretic doses should be titrated as necessary (Figure 6, Table 7). When sufficient progress has occurred to render it likely that targets of decongestion will be reached, it is usually appropriate to initiate or up-titrate components of the GDMT regimen (see Sections 7.4 and 8.3). Throughout the hospitalization, it is important to tailor education to the patient’s needs and to continue to address and manage comorbidities.

Neurohormonal antagonists have dramatically improved outcomes for HFrEF. When possible, continuation of GDMT through hospitalization or initiation before discharge is associated with substantially better outcomes, both due to the benefit of the therapies and to the better prognostic profile of patients who can tolerate them (106). Expert advice concerning initiation of GDMT for chronic HF can be found in the 2017 ACC Expert Consensus Decision Pathway for Optimization of Heart Failure Treatment (13).

Most studies demonstrating safety and efficacy of GDMT, however, have enrolled stable patients, and specifically excluded those with a recent decompensation (with the exception of the early use of angiotensin-converting enzyme inhibitors [ACEIs] in CONSENSUS and beta blockers in COPERNICUS) (158,159). Hospitalization provides a pivotal opportunity to decrease risk and improve clinical trajectory in patients who respond well to diuresis and who have not previously received adequate trials of GDMT. This therapy modifies and frequently reverses disease progression (9). The introduction of GDMT during hospitalization for HF with reduced ejection fraction is thus a key target to reduce risk (9,160). This has been shown for ACEI, beta blockers, and most recently supported for angiotensin receptor–neprilysin inhibitor (ARNI). Patients with good early response to diuresis should be considered for addition of recommended therapies or up-titration toward trial targets for neurohormonal antagonist therapy as decongestion is approached, recognizing that the diuretic response may diminish acutely with increasing neurohormonal antagonism, particularly if blood pressure is lowered.

Most studies of GDMT have investigated the addition or titration of a single agent to stable background therapy. There are no bases of evidence for patients in whom beta blockers or ACEIs/angiotensin receptor blockers (ARBs) were decreased or discontinued during hospitalization, most commonly for hypotension, progressive kidney dysfunction, or use of intravenous inotropic therapy (161). The 2017 ACCF/AHA guidelines emphasize that “caution should be used when initiating beta blockers in patients who have required inotropes during their hospital course or when initiating ACEIs, ARBs or aldosterone antagonists in those patients who have experienced marked azotemia or are at risk for hyperkalemia” (26). For these reasons, expectations regarding the prescription and dosing of GDMT in patients with ADHF are more conservative and individualized than for stable patients in outpatient HF management. For patients with HFpEF, beyond diuretics, clinical trial evidence that medical therapy improves outcomes is limited, but it seems reasonable to titrate RAS inhibitors to desired blood pressures in hospital.

One important common principle is to start at a low dose and titrate slowly upward as tolerated (Table 1 in Yancy et al. [13]). High starting doses and/or overly aggressive titration can result in hypotension and worsening kidney function, setbacks that limit both decongestion and initiation of different components of GDMT.

This trajectory would represent a patient who has had some improvement in symptoms and signs of congestion but does not reach the targeted goals of decongestion (Figure 8). Such patients tend to have more advanced disease, a history of frequent hospitalizations, and worse baseline kidney function. They commonly have high outpatient diuretic doses, and kidney function may worsen progressively with diuresis, a pattern associated with residual congestion and worse outcomes (119). In some cases, a high calculated net fluid loss is not reflected in weight changes due to high unrecorded intake. Addition of loading doses of amiodarone for therapy of atrial or ventricular arrhythmias can stall the progress of diuresis. Approximately 30% to 40% of ADHF patients discharged from the hospital still have moderate to severe congestion at the time of release (77,135,137,138,173).

If the patient has improved but has continued symptoms and/or signs, it is important to ascertain whether the signs and symptoms are predominantly due to HF. Persistent symptoms may reflect comorbidities as detailed in Table 4, particularly chronic pulmonary, kidney, or liver disease.

This trajectory represents a patient who is not responding to therapy, who has either failed to improve at all or has worsened during hospitalization (Figure 9). Some patients who appear to have stalled, as described in the previous text, may progress into this category. Approximately 20% to 30% of patients have no improvement in their symptoms or signs during hospitalization (175,176), and similarly, 15% to 20% of clinical trial patients have worsening HF that needs rescue therapy with additional diuretics, IV vasoactive agents, or mechanical circulatory or respiratory support (77,81,89,135,137,138,177).

If a patient is not improving or worsening, additional diagnostic strategies or specialist consultation may be considered. Intensification of diuretic therapy is appropriate, even if kidney function has worsened, because congestion is usually the main problem (Figure 5). It is crucial to re-evaluate the level of care, which may warrant escalation, including admission to an intensive care unit. It is reasonable to consider invasive hemodynamic monitoring to clarify right and left heart filling pressures and vascular resistances.

Review of the long-term trajectory of the patient who continues to worsen may warrant accelerated discussion regarding prognosis and goals of care (see Section 12).

The transition point heralds a distinct phase of care that begins after the decompensation leading to admission has resolved or has been addressed within the limitations of the chronic clinical profile. The focus then shifts toward how best to maintain stability of compensation. This most commonly occurs when clinical assessment reveals complete resolution of congestion and diuretic therapy is switched from intravenous to oral dosing. Evidence suggests that many patients hospitalized with HF are discharged too early, before meeting criteria described in Section 7.1 and shown in Figure 5. The average length of stay in the U.S. has decreased to 4 days, compared with an average of at least 7 days in the rest of the world (179). The risk of readmission for HF has been linked to shorter lengths of stay, which may lead to incomplete decongestion, lack of appropriate titration of GDMT, and incomplete translation of plans to postdischarge care (21,104,180,181). As such, early discharge can lead to Excess Days In Acute Care (EDAC), a measure of acute care days within 30 days of discharge that is being incorporated into quality standards.

Verifying the effectiveness of oral diuretic therapy prior to discharge, as recommended in the guidelines (14,15,25), generally requires at least 24 hours of observation after discontinuation of intravenous diuretics. In a recent retrospective study, observing patients on their intended discharge diuretic regimen for ≥24 hours was associated with a significant reduction in 30- and 90-day HF readmissions (182). Discharge before 24 hours of stability on oral diuretics may occasionally be appropriate, particularly for well-known patients frequently hospitalized for whom the trigger for decompensation is obvious, net diuresis has been easily achieved, and early follow-up is available with a familiar clinician. Recent survey data from physicians at 1 center regarding attitudes toward discharge readiness questioned the utility of targeting complete decongestion and observing patients for a day on oral diuretics (183). However, their post-discharge practices were not surveyed and the readmission rates did not capture those admitted to other hospitals (87). Separation of the transition point from the discharge day acknowledges the sequential steps and personnel time usually required for the intricate steps of discharge coordination, which is particularly important in high-risk patients. While assessment of educational gaps and other challenges for discharge begins early after admission, a directed multidisciplinary alert triggered by identification of the transition point can focus and finalize plans to support stability and further optimization of recommended therapies in the outpatient setting (13).

The dominant role of recongestion in HF readmissions suggests that current strategies for implementing a discharge diuretic plan are not reliable. This likely relates both to inadequate dosing at the time of discharge and lack of an adequate response plan that includes both an increased diuretic dose and the right clinical trigger for its use. The regimen taken prior to admission and the intravenous IV dosing required to achieve negative balance in the hospital should both influence planning of a new discharge regimen.

Maintenance diuretic dosing should be planned recognizing that lower doses are required for fluid balance than for net diuresis, but also that fluid balance is usually harder to maintain at home than in the hospital, where patients have controlled intake and spend more time supine, a position that enhances renal blood flow. Torsemide and bumetanide are more reliably absorbed than furosemide and may be considered when daily furosemide doses are high. Kidney dysfunction may lead clinicians to underdose diuretics, despite evidence that transient worsening of creatinine during effective decongestion does not confer long-term decrements in kidney function (143,184) and that kidney dysfunction itself decreases diuretic responsiveness.

A rescue dosing plan should be included in the intended discharge regimen, to specify not only the increased diuretic therapy but also the personalized trigger that should prompt the rescue; patients should be encouraged to call their clinician if unsure, and to avoid delay in starting therapy. In the recent PIONEER-HF trial comparing initiation of sacubitril/valsartan to enalapril before hospital discharge, one-half of patients required an increase in diuretic dosing during the next 6 weeks (166). Reliance upon changes in weight or symptoms of congestion are most often used for this purpose, despite their low sensitivity and delayed kinetics for predicting decompensated HF (185). When possible, a patient’s sentinel symptom of congestion should be recalled and emphasized, with care taken not to instruct reliance upon the appearance of edema or orthopnea for patients who have never experienced them. Adjustments to therapy may include increases in the dose and/or frequency of loop diuretic therapy or one-time doses of thiazide-type diuretics for sequential nephron blockade. Increases in mineralocorticoid dosing are rarely effective for rescue in outpatients and should not be made when kidney function might be declining. Dosing recommendations are in Table 7. Patients should be encouraged to call their clinician for clarification if unsure of their rescue plan.

A decision should also be made regarding fluid restriction after discharge, which is frequently done in hospital to accelerate net fluid loss. Stringent fluid restriction may not be necessary in patients who respond to low diuretic doses and do not habitually have high fluid intake. Two liters (64 ounces) is the usual practical limit by consensus, particularly for patients taking many medications. Transition to oral diuretics should also trigger consideration of if and how potassium supplements should be prescribed for home. Need for supplementation is lower on oral diuretic dosing intended to maintain rather than decrease net fluid balance. Testing the oral potassium replacement schedule when switching to oral diuretics is preferable to relying upon potassium scale dosing until the time of discharge. Key considerations include changes in therapies that alter potassium elimination (ACEI, ARB, ARNI, and especially aldosterone antagonists) as well as conditions associated with increased risk of hyperkalemia (chronic kidney disease and diabetes mellitus). Patients who have undergone large volumes of diuresis without needing more potassium replacement in hospital require particular vigilance regarding risk of hyperkalemia after discharge.

Patients who have received all of the recommended therapies for their HFrEF have consistently better longterm outcome than patients who have not, for whom optimization of GDMT is a high priority during and after hospitalization (15). For those in whom prior GDMT therapy was held during hospitalization, reinitiation should be attempted in the absence of contraindications. Lower doses may be required when therapy is resumed. It is equally important to evaluate the degree to which the patient has progressed in achieving target doses (Table 1 in Yancy et al. [13]). When neurohormonal antagonist therapies have been reduced or interrupted during hospitalization, the up-titration of GDMT may need to continue through the outpatient follow-up visits. Extensive guidance for GDMT optimization after discharge is provided by the 2017 ACC Expert Consensus Decision Pathway for Optimization of Heart Failure Treatment (13).

The transition point provides clinicians with an additional opportunity to consider enhancement of GDMT for the outpatient setting (88,103,106,121,160). For patients lacking a recommended medication, initiation is indicated if careful history reveals no contraindications (89). Some patients who have a long or complex history may not recall the reason for previous discontinuation; of particular concern would be a history of angioedema. A step of up-titration toward target dose should again be considered if not already attempted, recognizing the need to limit the number of changes in the setting of recent decompensation.

The period between the transition point and discharge should include confirmation that the patient tolerates the planned GDMT regimen for discharge. Absorption and vasodilation often increase after decongestion and can necessitate a reduction in dosing. Confirmation of tolerability includes documentation of the absence of postural hypotension and the administration of all doses as scheduled, without any being held for hypotension or dizziness (179,182,186). Postural symptoms early after discharge can sometimes lead to indiscriminate reduction or discontinuation of recommended therapies. In addition, some medications may have been changed in the hospital due to formulary restrictions, so medicine reconciliation is crucial. Whether further titration as an outpatient is expected, and also what factors may limit such titration, are crucial pieces of information to disseminate to the clinicians assuming care after the hospitalization. There is a place for such information on the Focused Discharge Handoff in this document (Section 10), and clinicians are encouraged to start thinking about these issues and documenting plans early in the hospitalization, well before the discharge day itself. This is a critical part of the “to-do” list for early post-discharge follow-up.

Although most drug therapy decisions in ADHF will involve the optimization of core GDMT, some patients may be eligible for further optimization, such as the addition or titration of fixed dose long-acting nitrates and hydralazine in African Americans already receiving an ACEI, ARB, or ARNI and a beta blocker (13). One important caution regarding the optimization of GDMT is the premature addition of ivabradine in patients not receiving a maximally-tolerated dose of beta blocker therapy. Because beta blockers are sometimes decreased or stopped during an ADHF episode, an evaluation of tolerability may need to be deferred to long-term follow-up, as ivabradine was evaluated in outpatients and trial eligibility required persistently elevated heart rate on target or maximally tolerated doses of beta blockers (187). Digoxin may be considered in patients with advanced disease specifically for the purposes of symptomatic relief and reducing the risk of hospitalization as well as facilitation of resting rate control in atrial fibrillation, although caution should be exercised in those with renal impairment or other high-risk features for drug toxicity (e.g., advanced age, low body weight, female gender), as target serum concentrations are <1.0 ng/ml (188).

Medication regimens for comorbidities also merit consideration, particularly with respect to potential interactions with HF. The importance of considering the impact of medications for diabetes on cardiovascular risk is increasingly recognized (189). Although data for new therapies, including sodium-glucose cotransporter-2 inhibitors (-gliflozins) and glucagon-like peptide-1 receptor antagonists (-glutides) are confined to outpatients, inpatient initiation in some settings may permit an evaluation of tolerability and impact of concomitant HF therapy by a multidisciplinary team, and increase the potential for improved compliance (189). As diabetes therapy in the hospital is often limited to a sliding scale insulin regimen to avoid hypoglycemia, caution should be taken and information shared with the outpatient clinicians for patients discharged on different diabetic regimens. If sodium-glucose cotransporter-2 inhibitors are started in the hospital, careful consideration should be given to the dose of diuretic therapy, because these agents have potent osmotic diuretic effects. Regardless, close follow-up with a primary care provider, endocrinologist, or diabetes educator is recommended within 2 to 4 weeks for patients with hyperglycemia in the hospital, especially when medications have been changed (190).

The view of the trajectory and risk profile at admission may have included factors subsequently modified favorably during the HF hospitalization. Although the hospital trajectory has ideally been monitored throughout hospitalization, the transition node provides the last opportunity before discharge to re-evaluate the long-term prognosis. This review is important for the patient and family and for the clinicians who will provide care after discharge.

Favorable modification of risks from admission relates most often to the effectiveness of decongestion, to the enhancement of guideline-recommended therapies for patients with reduced LVEF ( Table 4), and to improvement in patient education for adherence. It is vital to recognize that the degree of clinical congestion at admission does not confer increased risk after discharge, as long as decongestion has been achieved (67,104). Regardless of how it is measured, multiple factors relating to the severity of congestion at discharge predict worse quality of life, rehospitalization, and mortality. These factors include not only the symptoms and signs individually or combined into congestion scores (67,70,104), but also the natriuretic peptide levels, with progressively higher risk conferred by high absolute levels or with failure to reduce levels by at least 30% (24,55,63,68,109–113).

Discharge with residual congestion may reflect different limitations. Regardless of ejection fraction, severe underlying renal disease can lead to diuretic refractoriness and persistent fluid retention. Repeated discharge with residual congestion may be unavoidable when education and follow-up support has not improved adherence to the outpatient regimen, which may need to include different motivational interventions (76). Prolonged hospitalization can be futile when brisk daily sodium and urine output are exceeded by even higher daily intake despite restrictions. In the contemporary era of increasing GDMT, patients hospitalized despite adherence may be in later stages of HF, particularly with right HF, cardiorenal limitations, or chronic hypoperfusion (161). Increasing information is urgently needed to recognize when decongestion goals should be modified and how care should be redesigned to decrease the risk of further decompensation after discharge with residual congestion (161).

For hospitalized patients in whom ACEI/ARBs were previously tolerated but then discontinued due to hypotension or kidney dysfunction (90), 1-year mortality may be as high as 50%, particularly if intravenous inotropic therapy is added. Discharge without beta blocker therapy is also associated with poor outcomes. If advanced hemodynamic instability precludes tolerability of neurohormonal antagonist therapies, the patient is on a downward trajectory and should be considered for advanced therapies or revision of goals of care. Another component of the discharge regimen that carries prognostic significance is the dose of loop diuretic (66). High doses required to maintain fluid balance indicate diuretic resistance, for which a major factor is chronic kidney disease (107,108).

Elements of risk that carry over from admission to discharge include advanced age, history of prior hospitalizations, and socioeconomic status (93–97). It remains unclear how often chronic patterns of nonadherence can be modified (108). Baseline kidney function remains a strong predictor of outcome, as transient changes are less relevant than the absolute levels of kidney function before and after discharge (120). New risk factors that can arise during the hospitalization include use of intravenous inotropic therapy, even if transient (20,125). Need for cardiopulmonary resuscitation or intubation are associated with a much higher risk of death within the next 6 months (66,81).

The risk assessment in the transition day should guide the priority for early follow-up. It is difficult to mandate the timing of follow-up as clinic personnel resources vary between institutions. However, residual congestion, discharge without ACEI/ARB/ARNI, discharge without beta blockers, or consideration for advanced therapies warrant the earliest clinic slots available for follow-up. Patients started on new medications in the hospital should be contacted every few days until their first follow-up visit, and should generally have electrolytes and renal function checked within a week. Instability of renal function and electrolytes prior to discharge also warrants repeat testing early after discharge, with the results sent to the receiving clinician identified in the hand-off form.

The follow-up phone call checklist (Figure 13) can be used on its own or integrated into the EHR. An important component of this assessment is to ensure that the patient (or caregiver) can verbalize understanding of the discussion (teach-reteach method) (196). A systematic approach with a checklist can help organize and streamline the phone call to ensure that it is comprehensive yet focused.

Whereas the structure of the post-discharge visit can vary between medical centers, depending on the resources available, we believe that the key components listed in Figure 14 should be considered, ideally linked to specific recommendations and potentially integrated into the EHR for ease of documentation. Of particular importance are evaluation as indicated by clinicians with expertise in HF, and other consultation such as from nutrition and social work, measurement of laboratories, management of comorbid conditions, and education. Outpatient health care services that have been assigned before discharge should be linked in, both to inform and be informed about the ongoing progress after discharge.

Management of Comorbidities.
HF patients are often readmitted for diagnoses other than HF, and so active comorbid conditions (Table 4) should be aggressively addressed at the time of the post-discharge visit.

Medication Reconciliation.
HF patients are often prescribed multiple chronic medications, and errors in prescription are particularly common during transitions of care. The early post-discharge period offers an opportunity for comprehensive patient-centered medication reconciliation and continued progress toward optimization of recommended medical and device therapies (https://www.cardiosmart.org/SDM/Decision-Aids/Find- Decision-Aids/Heart-Failure).

As HF medications frequently remain at suboptimal doses, perhaps due to “therapeutic inertia” (201,202), a standardized approach to medication titration may be ideal, as has been addressed in prior decision pathway documents (TreatHF). Optimization of dosages is addressed in the Trajectory and Transition nodes, and information about future plans and potential impediments to increased dosing forms part of the Focused Discharge Handoff (Figure 11).

Laboratory Testing.
Laboratory studies usually performed during the post-discharge visit include an assessment of kidney function when patients are in the transition period with medications.

Trajectory of Clinical Decline.
A subset of patients with HF will continue to have symptoms and rapid disease progression despite being on maximally tolerated GDMT and may even need down titration of neurohormonal blockade (196). When advanced treatment strategies such as transplantation or mechanical circulatory support may be options, referral to advanced HF specialists may be indicated or a shift in focus to palliative care may be appropriate for many stage D patients (203). The I-NEEDHELP algorithm can be useful to guide patient selection for referral follow-up to an advanced HF specialist (13).