Alla Y. Grigorian, MD, PhD

Reversible impairment of brain function in the setting of cirrhosis defines hepatic encephalopathy (HE). HE is associated with significantly decreased survival,[1] and patients with HE have poor outcomes whether HE occurs in isolation or in conjunction with acute‐on‐chronic liver failure.[2] A large multicenter study comparing cirrhotics with and without HE also found that those with a history of HE were hospitalized more frequently.[2]

The presentation of HE is variable, and diagnosis remains clinical. Subtle manifestations of HE persist between episodes, even if gross cognitive function normalizes.[3] Retrospective data suggest the effects of serial bouts of HE may be cumulative, because even with appropriate treatment, the severity of impairment correlates with the number of prior episodes.[3] Even minimal manifestations of hepatic encephalopathy correlate with reduced quality of life.[4]

The West Haven score is the most validated scoring system.[5] Higher grades of HE correlate with significantly increased mortality,[2] but due to difficulties differentiating stages 0 and 1, these criteria remain somewhat controversial. The Spectrum of Neurocognitive Impairment in Cirrhosis (SONIC) has been proposed as an alternate conceptualization of HE as a continuous spectrum rather than discrete stages.[6] Table 1 shows findings associated with various West Haven and SONIC stages. Both systems include covert and overt encephalopathy. Covert correlates with West Haven grades 0 to 1, and consists mainly of subtle findings that require specialized psychometric testing to detect. The SONIC system terms demonstrable but subclinical manifestations minimal HE.[6] Overt HE includes West Haven grades 2 through 4, and refers to objective findings that can be reliably detected on clinical evaluation.[7] Whereas specific numeric scores are used largely for research purposes, classifying HE as covert or overt is clinically useful.

Although blood ammonia levels correlate well across populations, they are not diagnostically useful for individuals, because considerable overlap exists between patients with no HE and those with severe encephalopathy.[8] Ammonia levels also do not predict HE development.[9] Brain imaging is of limited utility, but may be prudent with abrupt decompensation, focal neurologic findings, or poor response to therapy.[10] A recent single‐center review of head computed tomography in cirrhotic patients presenting with altered level of consciousness found a low incidence of intracranial hemorrhage (ICH).[11] The number needed to scan was 293 patients to detect a single ICH. Only 1 patient out of 316 had ICH when fever, trauma, and focal neurological findings were excluded. The presence of acute ICH was not associated with platelet count, coagulopathy, creatinine, or Model for End‐Stage Liver Disease score.

PRECIPITANTS

Initial evaluation of patients with suspected HE must confirm the presence of HE and identify potentially reversible precipitants. Infection, bleeding, and metabolic derangements (including renal injury, hypovolemia, and hyponatremia) are common precipitants.[12] Searching for precipitants is heavily stressed in the 4‐pronged approach recommended by the American Association for the Study of Liver Disease,[7] as summarized in Table 2. Common precipitants are grouped into episodic and recurrent causes. Episodic causes are those that represent discrete insults with specific, short‐term treatments. Recurrent causes are those that are likely to require active management over time. These distinctions may help inform different approaches for initial or recurrent episodes of HE; in practice, much overlap exists.

Diuretic use has been clearly correlated with incidence of HE.[2] Although diuretic usage may be an indicator of more advanced liver disease, their use can also contribute to HE via increased risk of hypovolemia and dysnatremia.[2] Accordingly, caution is necessary when using diuretics to manage patients with HE and refractory ascites. These findings have led some to suggest serial paracentesis may be preferable to diuretics in this population.[2]

MANAGEMENT

The mainstay of HE treatment is administration of the nonabsorbable disaccharide lactulose. Lactulose is part of nearly all regimens because it is effective, easily titrated, and inexpensive.[13] It is efficacious orally or as an enema.[14] Lactulose increases both cognitive function and quality of life,[15] and is effective for prophylaxis and treatment of all stages of HE.[16, 17]

Rifaximin is often used as an adjunct to lactulose, particularly in cases of recurrent HE. Small trials have associated rifaximin with increased quality of life[18] and cognitive function.[19] The largest randomized trial of rifaximin was a double‐blind, placebo‐controlled trial in patients with multiple episodes of overt HE during the prior 6 months.[20] Lactulose was used concomitantly in approximately 91% of patients. At the end of the 6‐month study, rifaximin was associated with a 58% relative risk reduction in overt HE recurrence and roughly 50% reduction in HE‐related hospitalization. The numbers needed to treat were 4 patients to prevent 1 overt HE episode and 9 to prevent 1 HE‐related hospitalization.[20]

A meta‐analysis of 264 patients included in published, high‐quality trials found rifaximin monotherapy to be similar to nonabsorbable disaccharides in both efficacy and incidence of diarrhea, but with significantly less abdominal pain.[21] This analysis was limited by significant heterogeneity among trials. A larger, more recent systematic review and meta‐analysis of 19 studies (both published and unpublished) found rifaximin to be effective for treatment, secondary prophylaxis, and possibly decreased mortality.[22] Of note, this meta‐analysis included placebo studies as well as studies using varying doses of lactulose or other antibiotics as controls. Despite this variability, the authors concluded that the control used in the individual trials did not significantly affect the aggregate results.[22] In the largest individual study to show a mortality benefit, improvement seemed to be driven by decreased rates of sepsis when rifaximin was used as an adjunct to lactulose.[23] Cost is a barrier to use, as rifaximin has not proven to be cost‐effective as monotherapy instead of lactulose.[24] Many insurers will facilitate adjunctive rifaximin with prior authorization, and the manufacturer offers assistance programs.[25]

Other adjuncts, including laxatives,[26] antibiotics,[12] branched‐chain aminoacids,[27] and acarbose[28] have far less evidentiary support and require further study prior to incorporation into clinical practice.[26] A recent study showed polyethylene glycol to perform similar to lactulose, but the studied volume of 4 L daily may make routine use impractical.[29] Dietary protein restriction has been shown in a prospective randomized controlled trial to accelerate body muscle breakdown without affecting HE,[30] so is best avoided.

ISSUES PERTINENT TO HOSPITAL MANAGEMENT

Concurrent HE frequently complicates inpatient management of acute pain. Acetaminophen below 3 g daily for short‐term use is safe,[31] but may be insufficient. Non‐steroidal anti‐inflammatory agents are best avoided given risks for renal dysfunction and bleeding.[32] Although a direct connection between opiate use and HE remains unproven, these agents are problematic because they can cause both sedation and constipation. Nonetheless, they are often needed for pain control. Oxycodone has a more desirable side effect profile than other narcotics. We often prescribe doses every 6 hours initially to account for decreased hepatic metabolism. Morphine has active metabolites that can accumulate in cirrhotics, so morphine use is best avoided.[32] Fluctuations in cognition may help distinguish narcosis from HE; specifically, narcosis causes chronic somnolence worst shortly after an opiate dose, whereas HE causes alterations in sleep‐wake cycles including insomnia.[32] Frequent adjustment of opiate dose and frequency may be required to balance analgesia with unwanted sedation and constipation.

Decisional capacity frequently complicates care of patients with cirrhosis. Patients may decline therapy because of dissatisfaction with bowel frequency, but such lapses in adherence likely contribute to HE recurrence. Patients with overt HE are often incapable of making decisions based on informed consent. If such patients have inadequate social support to ensure medical attention if symptoms progress, then mandatory treatment is reasonable. This may include involuntary administration of medications via rectal or nasogastric tube. Once cognition improves enough that he or she can reliably articulate risks, benefits, and alternatives of declining therapy, then it is reasonable to allow them to do. Subspecialty consultation with psychiatry or ethics may be useful in such situations.

For cirrhotics admitted for management of nonhepatic issues (particularly operations or invasive procedures), vigilance is needed to monitor for HE during hospitalization. Patients with HE have increased risk of falls and impaired driving, which may lead to admission onto surgical services.[4] Changes in diet, medications, bowel function, and environment may all contribute to encephalopathy. HE occurring during admission for other diagnoses still requires prompt titration of lactulose. Routine inquiry about bowel function and sleep quality are likely to help identify trouble early.

Placement of transvenous intrahepatic portosystemic shunt (TIPS) increases the risk for HE via introduction of neurotoxins directly into the systemic circulation. These patients can typically be treated medically,[33] but are likely to require increased lactulose dosage. TIPS revision may be necessary for patients with treatment‐refractory HE, but retrospective evidence suggests this is rarely necessary.[33] In that study, only a single patient out of 81 with post‐TIPS HE required TIPS closure.

Under the International Classification of Disease, 10th Revision, a diagnosis of HE is often most consistent with metabolic encephalopathy (G93.41).[34] It may also be coded as chronic hepatic failure without coma (K7210) or chronic hepatic failure with coma (K7211).[35] Whenever possible, specifying the underlying liver disease (eg, hepatitis C virus, alcohol) is preferable.

TRANSITIONING TO OUTPATIENT CARE

HE patients are usually ready for community living once their cognition has improved enough to reliably take medications. Key aspects of HE management need to be communicated clearly to patients and caregivers. Barriers to optimal outpatient care mostly relate to lactulose adherence. Stressing the direct correlation between insufficient bowel movements and HE progression may enhance adherence. All patients need a lactulose titration plan including when doses can be skipped and when additional doses are needed. Even minimal symptoms of HE need to be addressed,[36] and specific vigilance for alterations in sleep‐wake cycles needs to be adopted. Table 3 is an example of a lactulose titration plan that can be used at discharge. These plans should be included in discharge documents and within communication to outpatient healthcare providers. Close follow‐up with a hepatology specialist is ideal to ensure appropriate lactulose use, answer questions that arise upon return home, and address other concerns related to cirrhosis.

Although specific interventions to decrease readmission have not been studied in this population, best practices from other populations (such as medication self‐management, follow‐up plans, and red flags to be on watch for[37]) likely apply. Defining optimal strategies to decrease readmission is an opportunity for hospitalists to contribute to standardization of care for these patients.

CONCLUSIONS

HE is a common but very treatable complication of cirrhosis. Various metabolic insults may precipitate HE, and hospitalists should seek to reverse contributing factors whenever possible. Lactulose titrated to ensure adequate bowel output is the cornerstone of both therapy and prevention for HE. Adjunctive use of rifaximin improves many outcomes. Patient education about manifestations of HE and medication titration is crucial to achieving smooth transition to the outpatient setting.

Disclosure

Nothing to report.

Reversible impairment of brain function in the setting of cirrhosis defines hepatic encephalopathy (HE). HE is associated with significantly decreased survival,[1] and patients with HE have poor outcomes whether HE occurs in isolation or in conjunction with acute‐on‐chronic liver failure.[2] A large multicenter study comparing cirrhotics with and without HE also found that those with a history of HE were hospitalized more frequently.[2]

The presentation of HE is variable, and diagnosis remains clinical. Subtle manifestations of HE persist between episodes, even if gross cognitive function normalizes.[3] Retrospective data suggest the effects of serial bouts of HE may be cumulative, because even with appropriate treatment, the severity of impairment correlates with the number of prior episodes.[3] Even minimal manifestations of hepatic encephalopathy correlate with reduced quality of life.[4]

The West Haven score is the most validated scoring system.[5] Higher grades of HE correlate with significantly increased mortality,[2] but due to difficulties differentiating stages 0 and 1, these criteria remain somewhat controversial. The Spectrum of Neurocognitive Impairment in Cirrhosis (SONIC) has been proposed as an alternate conceptualization of HE as a continuous spectrum rather than discrete stages.[6] Table 1 shows findings associated with various West Haven and SONIC stages. Both systems include covert and overt encephalopathy. Covert correlates with West Haven grades 0 to 1, and consists mainly of subtle findings that require specialized psychometric testing to detect. The SONIC system terms demonstrable but subclinical manifestations minimal HE.[6] Overt HE includes West Haven grades 2 through 4, and refers to objective findings that can be reliably detected on clinical evaluation.[7] Whereas specific numeric scores are used largely for research purposes, classifying HE as covert or overt is clinically useful.

Although blood ammonia levels correlate well across populations, they are not diagnostically useful for individuals, because considerable overlap exists between patients with no HE and those with severe encephalopathy.[8] Ammonia levels also do not predict HE development.[9] Brain imaging is of limited utility, but may be prudent with abrupt decompensation, focal neurologic findings, or poor response to therapy.[10] A recent single‐center review of head computed tomography in cirrhotic patients presenting with altered level of consciousness found a low incidence of intracranial hemorrhage (ICH).[11] The number needed to scan was 293 patients to detect a single ICH. Only 1 patient out of 316 had ICH when fever, trauma, and focal neurological findings were excluded. The presence of acute ICH was not associated with platelet count, coagulopathy, creatinine, or Model for End‐Stage Liver Disease score.

PRECIPITANTS

Initial evaluation of patients with suspected HE must confirm the presence of HE and identify potentially reversible precipitants. Infection, bleeding, and metabolic derangements (including renal injury, hypovolemia, and hyponatremia) are common precipitants.[12] Searching for precipitants is heavily stressed in the 4‐pronged approach recommended by the American Association for the Study of Liver Disease,[7] as summarized in Table 2. Common precipitants are grouped into episodic and recurrent causes. Episodic causes are those that represent discrete insults with specific, short‐term treatments. Recurrent causes are those that are likely to require active management over time. These distinctions may help inform different approaches for initial or recurrent episodes of HE; in practice, much overlap exists.

Diuretic use has been clearly correlated with incidence of HE.[2] Although diuretic usage may be an indicator of more advanced liver disease, their use can also contribute to HE via increased risk of hypovolemia and dysnatremia.[2] Accordingly, caution is necessary when using diuretics to manage patients with HE and refractory ascites. These findings have led some to suggest serial paracentesis may be preferable to diuretics in this population.[2]

MANAGEMENT

The mainstay of HE treatment is administration of the nonabsorbable disaccharide lactulose. Lactulose is part of nearly all regimens because it is effective, easily titrated, and inexpensive.[13] It is efficacious orally or as an enema.[14] Lactulose increases both cognitive function and quality of life,[15] and is effective for prophylaxis and treatment of all stages of HE.[16, 17]

Rifaximin is often used as an adjunct to lactulose, particularly in cases of recurrent HE. Small trials have associated rifaximin with increased quality of life[18] and cognitive function.[19] The largest randomized trial of rifaximin was a double‐blind, placebo‐controlled trial in patients with multiple episodes of overt HE during the prior 6 months.[20] Lactulose was used concomitantly in approximately 91% of patients. At the end of the 6‐month study, rifaximin was associated with a 58% relative risk reduction in overt HE recurrence and roughly 50% reduction in HE‐related hospitalization. The numbers needed to treat were 4 patients to prevent 1 overt HE episode and 9 to prevent 1 HE‐related hospitalization.[20]

A meta‐analysis of 264 patients included in published, high‐quality trials found rifaximin monotherapy to be similar to nonabsorbable disaccharides in both efficacy and incidence of diarrhea, but with significantly less abdominal pain.[21] This analysis was limited by significant heterogeneity among trials. A larger, more recent systematic review and meta‐analysis of 19 studies (both published and unpublished) found rifaximin to be effective for treatment, secondary prophylaxis, and possibly decreased mortality.[22] Of note, this meta‐analysis included placebo studies as well as studies using varying doses of lactulose or other antibiotics as controls. Despite this variability, the authors concluded that the control used in the individual trials did not significantly affect the aggregate results.[22] In the largest individual study to show a mortality benefit, improvement seemed to be driven by decreased rates of sepsis when rifaximin was used as an adjunct to lactulose.[23] Cost is a barrier to use, as rifaximin has not proven to be cost‐effective as monotherapy instead of lactulose.[24] Many insurers will facilitate adjunctive rifaximin with prior authorization, and the manufacturer offers assistance programs.[25]

Other adjuncts, including laxatives,[26] antibiotics,[12] branched‐chain aminoacids,[27] and acarbose[28] have far less evidentiary support and require further study prior to incorporation into clinical practice.[26] A recent study showed polyethylene glycol to perform similar to lactulose, but the studied volume of 4 L daily may make routine use impractical.[29] Dietary protein restriction has been shown in a prospective randomized controlled trial to accelerate body muscle breakdown without affecting HE,[30] so is best avoided.

ISSUES PERTINENT TO HOSPITAL MANAGEMENT

Concurrent HE frequently complicates inpatient management of acute pain. Acetaminophen below 3 g daily for short‐term use is safe,[31] but may be insufficient. Non‐steroidal anti‐inflammatory agents are best avoided given risks for renal dysfunction and bleeding.[32] Although a direct connection between opiate use and HE remains unproven, these agents are problematic because they can cause both sedation and constipation. Nonetheless, they are often needed for pain control. Oxycodone has a more desirable side effect profile than other narcotics. We often prescribe doses every 6 hours initially to account for decreased hepatic metabolism. Morphine has active metabolites that can accumulate in cirrhotics, so morphine use is best avoided.[32] Fluctuations in cognition may help distinguish narcosis from HE; specifically, narcosis causes chronic somnolence worst shortly after an opiate dose, whereas HE causes alterations in sleep‐wake cycles including insomnia.[32] Frequent adjustment of opiate dose and frequency may be required to balance analgesia with unwanted sedation and constipation.

Decisional capacity frequently complicates care of patients with cirrhosis. Patients may decline therapy because of dissatisfaction with bowel frequency, but such lapses in adherence likely contribute to HE recurrence. Patients with overt HE are often incapable of making decisions based on informed consent. If such patients have inadequate social support to ensure medical attention if symptoms progress, then mandatory treatment is reasonable. This may include involuntary administration of medications via rectal or nasogastric tube. Once cognition improves enough that he or she can reliably articulate risks, benefits, and alternatives of declining therapy, then it is reasonable to allow them to do. Subspecialty consultation with psychiatry or ethics may be useful in such situations.

For cirrhotics admitted for management of nonhepatic issues (particularly operations or invasive procedures), vigilance is needed to monitor for HE during hospitalization. Patients with HE have increased risk of falls and impaired driving, which may lead to admission onto surgical services.[4] Changes in diet, medications, bowel function, and environment may all contribute to encephalopathy. HE occurring during admission for other diagnoses still requires prompt titration of lactulose. Routine inquiry about bowel function and sleep quality are likely to help identify trouble early.

Placement of transvenous intrahepatic portosystemic shunt (TIPS) increases the risk for HE via introduction of neurotoxins directly into the systemic circulation. These patients can typically be treated medically,[33] but are likely to require increased lactulose dosage. TIPS revision may be necessary for patients with treatment‐refractory HE, but retrospective evidence suggests this is rarely necessary.[33] In that study, only a single patient out of 81 with post‐TIPS HE required TIPS closure.

Under the International Classification of Disease, 10th Revision, a diagnosis of HE is often most consistent with metabolic encephalopathy (G93.41).[34] It may also be coded as chronic hepatic failure without coma (K7210) or chronic hepatic failure with coma (K7211).[35] Whenever possible, specifying the underlying liver disease (eg, hepatitis C virus, alcohol) is preferable.

TRANSITIONING TO OUTPATIENT CARE

HE patients are usually ready for community living once their cognition has improved enough to reliably take medications. Key aspects of HE management need to be communicated clearly to patients and caregivers. Barriers to optimal outpatient care mostly relate to lactulose adherence. Stressing the direct correlation between insufficient bowel movements and HE progression may enhance adherence. All patients need a lactulose titration plan including when doses can be skipped and when additional doses are needed. Even minimal symptoms of HE need to be addressed,[36] and specific vigilance for alterations in sleep‐wake cycles needs to be adopted. Table 3 is an example of a lactulose titration plan that can be used at discharge. These plans should be included in discharge documents and within communication to outpatient healthcare providers. Close follow‐up with a hepatology specialist is ideal to ensure appropriate lactulose use, answer questions that arise upon return home, and address other concerns related to cirrhosis.

Although specific interventions to decrease readmission have not been studied in this population, best practices from other populations (such as medication self‐management, follow‐up plans, and red flags to be on watch for[37]) likely apply. Defining optimal strategies to decrease readmission is an opportunity for hospitalists to contribute to standardization of care for these patients.

CONCLUSIONS

HE is a common but very treatable complication of cirrhosis. Various metabolic insults may precipitate HE, and hospitalists should seek to reverse contributing factors whenever possible. Lactulose titrated to ensure adequate bowel output is the cornerstone of both therapy and prevention for HE. Adjunctive use of rifaximin improves many outcomes. Patient education about manifestations of HE and medication titration is crucial to achieving smooth transition to the outpatient setting.

Disclosure

Nothing to report.

Reversible impairment of brain function in the setting of cirrhosis defines hepatic encephalopathy (HE). HE is associated with significantly decreased survival,[1] and patients with HE have poor outcomes whether HE occurs in isolation or in conjunction with acute‐on‐chronic liver failure.[2] A large multicenter study comparing cirrhotics with and without HE also found that those with a history of HE were hospitalized more frequently.[2]

The presentation of HE is variable, and diagnosis remains clinical. Subtle manifestations of HE persist between episodes, even if gross cognitive function normalizes.[3] Retrospective data suggest the effects of serial bouts of HE may be cumulative, because even with appropriate treatment, the severity of impairment correlates with the number of prior episodes.[3] Even minimal manifestations of hepatic encephalopathy correlate with reduced quality of life.[4]

The West Haven score is the most validated scoring system.[5] Higher grades of HE correlate with significantly increased mortality,[2] but due to difficulties differentiating stages 0 and 1, these criteria remain somewhat controversial. The Spectrum of Neurocognitive Impairment in Cirrhosis (SONIC) has been proposed as an alternate conceptualization of HE as a continuous spectrum rather than discrete stages.[6] Table 1 shows findings associated with various West Haven and SONIC stages. Both systems include covert and overt encephalopathy. Covert correlates with West Haven grades 0 to 1, and consists mainly of subtle findings that require specialized psychometric testing to detect. The SONIC system terms demonstrable but subclinical manifestations minimal HE.[6] Overt HE includes West Haven grades 2 through 4, and refers to objective findings that can be reliably detected on clinical evaluation.[7] Whereas specific numeric scores are used largely for research purposes, classifying HE as covert or overt is clinically useful.

Although blood ammonia levels correlate well across populations, they are not diagnostically useful for individuals, because considerable overlap exists between patients with no HE and those with severe encephalopathy.[8] Ammonia levels also do not predict HE development.[9] Brain imaging is of limited utility, but may be prudent with abrupt decompensation, focal neurologic findings, or poor response to therapy.[10] A recent single‐center review of head computed tomography in cirrhotic patients presenting with altered level of consciousness found a low incidence of intracranial hemorrhage (ICH).[11] The number needed to scan was 293 patients to detect a single ICH. Only 1 patient out of 316 had ICH when fever, trauma, and focal neurological findings were excluded. The presence of acute ICH was not associated with platelet count, coagulopathy, creatinine, or Model for End‐Stage Liver Disease score.

PRECIPITANTS

Initial evaluation of patients with suspected HE must confirm the presence of HE and identify potentially reversible precipitants. Infection, bleeding, and metabolic derangements (including renal injury, hypovolemia, and hyponatremia) are common precipitants.[12] Searching for precipitants is heavily stressed in the 4‐pronged approach recommended by the American Association for the Study of Liver Disease,[7] as summarized in Table 2. Common precipitants are grouped into episodic and recurrent causes. Episodic causes are those that represent discrete insults with specific, short‐term treatments. Recurrent causes are those that are likely to require active management over time. These distinctions may help inform different approaches for initial or recurrent episodes of HE; in practice, much overlap exists.

Diuretic use has been clearly correlated with incidence of HE.[2] Although diuretic usage may be an indicator of more advanced liver disease, their use can also contribute to HE via increased risk of hypovolemia and dysnatremia.[2] Accordingly, caution is necessary when using diuretics to manage patients with HE and refractory ascites. These findings have led some to suggest serial paracentesis may be preferable to diuretics in this population.[2]

MANAGEMENT

The mainstay of HE treatment is administration of the nonabsorbable disaccharide lactulose. Lactulose is part of nearly all regimens because it is effective, easily titrated, and inexpensive.[13] It is efficacious orally or as an enema.[14] Lactulose increases both cognitive function and quality of life,[15] and is effective for prophylaxis and treatment of all stages of HE.[16, 17]

Rifaximin is often used as an adjunct to lactulose, particularly in cases of recurrent HE. Small trials have associated rifaximin with increased quality of life[18] and cognitive function.[19] The largest randomized trial of rifaximin was a double‐blind, placebo‐controlled trial in patients with multiple episodes of overt HE during the prior 6 months.[20] Lactulose was used concomitantly in approximately 91% of patients. At the end of the 6‐month study, rifaximin was associated with a 58% relative risk reduction in overt HE recurrence and roughly 50% reduction in HE‐related hospitalization. The numbers needed to treat were 4 patients to prevent 1 overt HE episode and 9 to prevent 1 HE‐related hospitalization.[20]

A meta‐analysis of 264 patients included in published, high‐quality trials found rifaximin monotherapy to be similar to nonabsorbable disaccharides in both efficacy and incidence of diarrhea, but with significantly less abdominal pain.[21] This analysis was limited by significant heterogeneity among trials. A larger, more recent systematic review and meta‐analysis of 19 studies (both published and unpublished) found rifaximin to be effective for treatment, secondary prophylaxis, and possibly decreased mortality.[22] Of note, this meta‐analysis included placebo studies as well as studies using varying doses of lactulose or other antibiotics as controls. Despite this variability, the authors concluded that the control used in the individual trials did not significantly affect the aggregate results.[22] In the largest individual study to show a mortality benefit, improvement seemed to be driven by decreased rates of sepsis when rifaximin was used as an adjunct to lactulose.[23] Cost is a barrier to use, as rifaximin has not proven to be cost‐effective as monotherapy instead of lactulose.[24] Many insurers will facilitate adjunctive rifaximin with prior authorization, and the manufacturer offers assistance programs.[25]

Other adjuncts, including laxatives,[26] antibiotics,[12] branched‐chain aminoacids,[27] and acarbose[28] have far less evidentiary support and require further study prior to incorporation into clinical practice.[26] A recent study showed polyethylene glycol to perform similar to lactulose, but the studied volume of 4 L daily may make routine use impractical.[29] Dietary protein restriction has been shown in a prospective randomized controlled trial to accelerate body muscle breakdown without affecting HE,[30] so is best avoided.

ISSUES PERTINENT TO HOSPITAL MANAGEMENT

Concurrent HE frequently complicates inpatient management of acute pain. Acetaminophen below 3 g daily for short‐term use is safe,[31] but may be insufficient. Non‐steroidal anti‐inflammatory agents are best avoided given risks for renal dysfunction and bleeding.[32] Although a direct connection between opiate use and HE remains unproven, these agents are problematic because they can cause both sedation and constipation. Nonetheless, they are often needed for pain control. Oxycodone has a more desirable side effect profile than other narcotics. We often prescribe doses every 6 hours initially to account for decreased hepatic metabolism. Morphine has active metabolites that can accumulate in cirrhotics, so morphine use is best avoided.[32] Fluctuations in cognition may help distinguish narcosis from HE; specifically, narcosis causes chronic somnolence worst shortly after an opiate dose, whereas HE causes alterations in sleep‐wake cycles including insomnia.[32] Frequent adjustment of opiate dose and frequency may be required to balance analgesia with unwanted sedation and constipation.

Decisional capacity frequently complicates care of patients with cirrhosis. Patients may decline therapy because of dissatisfaction with bowel frequency, but such lapses in adherence likely contribute to HE recurrence. Patients with overt HE are often incapable of making decisions based on informed consent. If such patients have inadequate social support to ensure medical attention if symptoms progress, then mandatory treatment is reasonable. This may include involuntary administration of medications via rectal or nasogastric tube. Once cognition improves enough that he or she can reliably articulate risks, benefits, and alternatives of declining therapy, then it is reasonable to allow them to do. Subspecialty consultation with psychiatry or ethics may be useful in such situations.

For cirrhotics admitted for management of nonhepatic issues (particularly operations or invasive procedures), vigilance is needed to monitor for HE during hospitalization. Patients with HE have increased risk of falls and impaired driving, which may lead to admission onto surgical services.[4] Changes in diet, medications, bowel function, and environment may all contribute to encephalopathy. HE occurring during admission for other diagnoses still requires prompt titration of lactulose. Routine inquiry about bowel function and sleep quality are likely to help identify trouble early.

Placement of transvenous intrahepatic portosystemic shunt (TIPS) increases the risk for HE via introduction of neurotoxins directly into the systemic circulation. These patients can typically be treated medically,[33] but are likely to require increased lactulose dosage. TIPS revision may be necessary for patients with treatment‐refractory HE, but retrospective evidence suggests this is rarely necessary.[33] In that study, only a single patient out of 81 with post‐TIPS HE required TIPS closure.

Under the International Classification of Disease, 10th Revision, a diagnosis of HE is often most consistent with metabolic encephalopathy (G93.41).[34] It may also be coded as chronic hepatic failure without coma (K7210) or chronic hepatic failure with coma (K7211).[35] Whenever possible, specifying the underlying liver disease (eg, hepatitis C virus, alcohol) is preferable.

TRANSITIONING TO OUTPATIENT CARE

HE patients are usually ready for community living once their cognition has improved enough to reliably take medications. Key aspects of HE management need to be communicated clearly to patients and caregivers. Barriers to optimal outpatient care mostly relate to lactulose adherence. Stressing the direct correlation between insufficient bowel movements and HE progression may enhance adherence. All patients need a lactulose titration plan including when doses can be skipped and when additional doses are needed. Even minimal symptoms of HE need to be addressed,[36] and specific vigilance for alterations in sleep‐wake cycles needs to be adopted. Table 3 is an example of a lactulose titration plan that can be used at discharge. These plans should be included in discharge documents and within communication to outpatient healthcare providers. Close follow‐up with a hepatology specialist is ideal to ensure appropriate lactulose use, answer questions that arise upon return home, and address other concerns related to cirrhosis.

Although specific interventions to decrease readmission have not been studied in this population, best practices from other populations (such as medication self‐management, follow‐up plans, and red flags to be on watch for[37]) likely apply. Defining optimal strategies to decrease readmission is an opportunity for hospitalists to contribute to standardization of care for these patients.

CONCLUSIONS

HE is a common but very treatable complication of cirrhosis. Various metabolic insults may precipitate HE, and hospitalists should seek to reverse contributing factors whenever possible. Lactulose titrated to ensure adequate bowel output is the cornerstone of both therapy and prevention for HE. Adjunctive use of rifaximin improves many outcomes. Patient education about manifestations of HE and medication titration is crucial to achieving smooth transition to the outpatient setting.

Disclosure

Nothing to report.