Intensive care medicine has dramatically evolved over the last 15 years, after reports from many landmark trials.¹ Updated strategies for mechanical ventilation² and “bundles” of strategies to optimize hemodynamic therapy³ have reduced the rates of morbidity and death from deadly critical conditions such as the adult respiratory distress syndrome (ARDS) and sepsis.

Despite these important improvements in short-term outcomes, it is increasingly recognized that intensive care unit (ICU) survivors suffer considerable long-term complications that affect their usual functioning.⁴ Recently, the Society of Critical Care Medicine convened a conference in which these long-term complications were named the “post-intensive care syndrome.”⁵

Quality of life, particularly its physical component, is considerably lower after a stay in the medical or surgical ICU.^6–8 Posttraumatic stress disorder, depression, and sexual dysfunction are consistently reported years after ICU discharge.^9–13

Perhaps the most frequently unrecognized complication in ICU survivors is cognitive impairment. Current data suggest that neurocognitive impairment after an ICU stay is common and that it persists 6 years or more after hospital discharge.

Hopkins et al^14,15 analyzed 10 cohort studies of long-term cognitive impairment after an ICU stay; 5 of them focused on patients with ARDS. The prevalence of cognitive impairment was as high as 78% at hospital discharge, 46% at 1 year, and 25% 6 years after discharge.^15,16 Of the cognitive domains compromised, memory was the most often affected, followed by executive function and attention.^14,17

Interestingly, data suggest that cognition may improve somewhat in the first 6 to 12 months after ICU discharge.¹⁵ Therefore, if we can detect it early on and promptly refer patients for cognitive therapy, we may eventually improve the prognosis of this disabling complication.

This review will focus on how to evaluate, prevent, and treat cognitive impairment in patients who survive an ICU stay.

COGNITIVE IMPAIRMENT AFTER A STAY IN THE ICU

The association between ICU stay and neurocognitive dysfunction is poorly understood. Potential causes include hypoxemia,¹⁸ hypotension, ¹⁹ hyperglycemia,¹⁴ and—an area of growing interest and evolving research—sedation and delirium.²⁰

Patients on mechanical ventilation are commonly given sedatives and analgesics to prevent anxiety and pain.²¹ However, these medications are strongly associated with delirium.²² In fact, recent studies found that benzodiazepines have an independent, dose-related, temporal association with delirium, with some reports describing a 20% increase in delirium per milligram of benzodiazepine.²³ In another study, which included medical and surgical ICU patients, use of morphine was the strongest predictor of delirium, with a sixfold increase in odds over a period of 5 months.²⁴

Delirium is important to prevent, diagnose, and treat, since it has a direct association with the development of long-term cognitive impairment.^22,25 A review of studies that included 1,885 medical and surgical patients found that those who developed delirium during an ICU stay were three times more likely to have cognitive dysfunction when assessed 3 years later.²⁰

Whether delirium is a primary disorder associated with cognitive impairment or if it only represents an underlying process leading to poor cognitive outcomes is unknown. As delirious patients are more likely to be older, to be mechanically ventilated, to require more sedation, and, in particular, to be sicker, the association between delirium and cognitive impairment may reflect the relationship between these risk factors and poor cognitive outcomes.²⁶

Glucose and its relationship with cognitive function is another topic of investigation. A secondary analysis of a study that included ARDS survivors revealed that blood glucose values higher than 153 mg/dL, higher glucose variability, and duration of mechanical ventilation were associated with cognitive sequelae.^27,28

Other studies focused on mechanical ventilation. In one study,²⁹ one-third of patients who had been mechanically ventilated showed signs of neurocognitive impairment when they were evaluated 6 months after hospital discharge.

Mild cognitive impairment differs from cognitive impairment after an ICU stay

Cognitive impairment after ICU discharge does not follow the same pattern as mild cognitive impairment, and some authors consider these two types of cognitive impairment to be unrelated.

While mild cognitive impairment is progressive and associated with aging, cognitive impairment in ICU survivors develops rapidly after acute illness and is usually related to numerous pathologic and neurochemical pathways.

For example, the neurotransmitter acetylcholine is thought to be involved in cognitive function as well as neuroplasticity of the motor cortex. In a model of cognitive impairment after stroke, activity of the cholinergic system was reduced.^30,31 Further, in a study in rats, Baskerville et al³² showed that experience-dependent plasticity could be completely blocked by damaging the cholinergic neurons in the nucleus basalis of Meynert, thereby affecting memory and other functions supported by this pathway.

Another implicated pathway involves dopamine. Of interest, dopamine augmentation has been shown to enhance simple motor memories and to improve procedural learning. Understanding of these neurochemical alterations opens opportunities for investigation of drug therapies.

ASSESSMENT TOOLS

Cognitive impairment is important to detect in ICU survivors because it predicts poor outcomes from rehabilitation. A study of stroke patients found that those with cognitive alterations immediately after the stroke were less likely to be discharged home or to be living at home 6 months after discharge.³³

A possible explanation may be that affected patients cannot fully participate in rehabilitation activities, owing to impairment in executive function, inability to remember therapy instructions, or disruption of implicit and explicit learning. Indeed, some authors consider cognitive impairment after acquired brain injury to be the most relevant surrogate marker of rehabilitation potential. Consequently, manipulation or enhancement of cognition may directly affect rehabilitation outcomes.³⁴

Disagreement about terminology and diagnostic criteria creates a problem for health care providers working with patients with potential cognitive impairment. Numerous systems have been proposed to define this condition; in fact, Stephan et al³⁵ reviewed the literature and found no fewer than 17. None of them is specific for cognitive impairment after an ICU stay.

Petersen et al³⁶ in 1999 proposed initial criteria for mild cognitive impairment that included the following:

• A memory complaint
• Normal general cognitive functioning
• Normal activities of daily living
• Memory impairment in relation to age and education
• No dementia.

Later, other areas of impairment besides memory were recognized, such as language, attention, perception, reasoning, and motor planning.³⁷ Therefore, mild cognitive impairment is currently classified into subtypes, which include amnestic (affecting single or multiple domains) and nonamnestic (also affecting single or multiple domains).³⁸

In clinical practice, impairment of specific cognitive domains may be challenging to detect, and neuropsychological testing is often needed. Cognitive screening tests can detect impairment across a restricted range of cognitive abilities, while more comprehensive assessments address each of the primary domains of cognition.³⁹ Formal testing provides normative and validated data on cognition performance and severity.

The Montreal Cognitive Assessment⁴⁰ is popular, comprehensive, used in a variety of professions in diverse types of facilities (acute care, rehabilitation, and skilled care facilities), and brief (taking 11 minutes to administer). It evaluates orientation, memory, language, attention, reasoning, and visual-constructional abilities. The maximum score is 30; cognitive impairment is defined as a score of less than 26. It has a sensitivity of 90% and a specificity of 87%.

The Folstein Mini-Mental State Examination (MMSE) is the most commonly used of the noncomprehensive tests in clinical practice.⁴¹ It assesses orientation, memory, language, attention, and praxis. It has a maximum score of 30 points; the cutoff score for cognitive impairment is 24 points or less.

A limitation of the MMSE is that its sensitivity is very low, ranging from 1% to 49%.^42,43 The MMSE scores of patients with cognitive impairment overlap considerably with those of age-matched healthy controls.³⁹ Conversely, the MMSE’s specificity is usually high, ranging from 85% to 100%.⁴²

Moreover, the MMSE poses copyright issues, an important consideration when selecting a test. In 2001, the authors of the MMSE transferred all intellectual property rights to Psychological Assessment Resources, which has exclusive rights to publish, license, and manage all intellectual property rights in all media and languages. Photocopying and using the MMSE without applying for permission from and paying this company ($1.23 per use) constitutes copyright infringement. Therefore, health care providers and researchers have been using other tests to evaluate cognition.

Other tests of cognition assess individual domains. Interestingly, studies of long-term cognitive impairment after ICU admission used these tests to define outcomes.²⁵ Specific tests include:

• The Digit Span and the Trailmaking Test A (used to assess attention and orientation)²⁵
• The Rey Auditory Verbal Learning Test (used to evaluate verbal memory)
• The Complex Figure Test (helpful in defining visual-spatial construction and delayed visual memory)
• The Trailmaking Test B (also included in the Montreal Cognitive Assessment; assesses executive functioning).

Besides formal testing, an informal battery is often recommended to provide additional information. An informal evaluation includes word definition, reading and verbal fluency, reading comprehension, and performance of instrumental activities of daily living. Observing as patients perform tasks of daily living provides therapists with a vast amount of information, as these tasks require using multiple cognitive processes. Therefore, if a functional breakdown occurs during this assessment, the clinician needs to identify the domain or specific level of cognitive dysfunction involved in that deficit.⁴⁴

PREVENTIVE STRATEGIES

Strategies for minimizing the long-term effects of cognitive impairment have mostly focused on preventing it.

During the ICU stay, optimizing hemodynamic, glucose, and oxygenation levels may prevent future long-term complications.¹⁸

Also, the association between sedation, delirium, and consequent cognitive impairment (see above) has led many investigators to apply the “ABCDE” bundle of strategies.^25,45,46 Specifically, ABCDE stands for awakening and breathing, choice of sedatives with fewer adverse effects, daily delirium monitoring, and early mobility exercise. These strategies have been shown in randomized controlled trials to prevent delirium; however, they have not been proved to prevent cognitive impairment.

Awakening and breathing

In the Awakening and Breathing Controlled Trial,⁴⁷ patients in the intervention group (ie, those who had their sedatives interrupted every morning to see if they would awaken, and if so, if they could breathe on their own) were extubated 3 days sooner than those in the control group (who underwent daily trials of spontaneous breathing, if deemed safe). Also, ICU and hospital length of stay were shorter by 4 days. Best of all, over 1 year, the mortality rate was lower by 14 absolute percentage points.

Choice of sedatives

Often, mechanically ventilated patients are given benzodiazepines, opiates, and propofol (Diprivan).²¹ Dexmedetomidine (Precedex), a newer agent, is an alpha-2 agonist and may offer advantages over the others.

To date, three randomized controlled trials have assessed the effect of dexmedetomidine in terms of outcomes associated with delirium, and one trial evaluated its association with intellectual capacity in ICU patients.

The Maximizing Efficacy of Targeted Sedation and Reducing Neurological Dysfunction (MENDS) trial randomized patients on mechanical ventilation to receive either dexmedetomidine or lorazepam (Ativan).⁴⁸ Dexmedetomidine-treated patients had 4 more days alive without delirium or coma (7 vs 3 days, P = .01).

Subsequently, the Safety and Efficacy of Dexmedetomidine Compared With Midazolam (SEDCOM) trial compared dexmedetomidine and midazolam (Versed) in mechanically ventilated patients. Those who received dexmedetomidine had a lower incidence of delirium (54% vs 76%, P < .001), and 2 fewer days on mechanical ventilation.⁴⁹

Reade et al⁵⁰ evaluated time to extubation in already delirious patients randomized to receive either dexmedetomidine or haloperidol (Haldol). Those receiving dexmedetomidine had a shorter time to extubation as well as a shorter ICU length of stay.

The Acute Neuroscience Intensive Care Sedation Trial⁵¹ evaluated intellectual capacity in neurological ICU patients sedated with either dexmedetomidine or propofol. This randomized, double-blind trial included 18 brain-injured and 12 non-brain-injured intubated patients. In a crossover protocol, each received the combination of fentanyl (Sublimaze) and propofol and the combination of fentanyl and dexmedetomidine.

Cognition was evaluated using the Adapted Cognitive Exam (ACE), which assesses intellectual capacity through orientation, language, registration, attention, calculation, and recall. This 10-minute examination does not require verbal communication, as it relies on the ability to respond to yes-or-no questions and perform simple motor tasks. The maximum possible score is 100 points.

Interestingly, while on propofol, the patients’ adjusted ACE scores went down by a mean of 12.4 points, whereas they went up by 6.8 points while on dexmedetomidine. Even though brain-injured patients required less sedation than non-brain-injured patients, the effect of dexmedetomidine and propofol did not change.⁵¹

In summary, these studies suggest that all sedatives are not the same in their short-term and intermediate-term outcomes.

In our practice, we use dexmedetomidine as our first-line sedation therapy. In patients with hemodynamic instability, we use benzodiazepines. We reserve propofol for very short periods of intubation or for hemodynamically stable patients who cannot be sedated with dexmedetomidine.

Daily delirium monitoring

As mentioned above, delirium affects many patients on mechanical ventilation, and it is highly underrecognized if valid tests are not used.⁵² Therefore, it is critically important to be familiar with the tests for assessing delirium. Of these, the Confusion Assessment Method for the ICU is probably the one with the best performance, with a sensitivity of 93% to 100% and a specificity of 98% to 100%.^53,54

Early mobilization

A landmark study paired the awakening and breathing strategy with early mobilization through physical and occupational therapy in the ICU.⁵⁵ Patients in the intervention group had a higher rate of return to independent functional status upon hospital discharge and a shorter duration of mechanical ventilation and delirium.

In conclusion, even though direct prevention of cognitive dysfunction is a challenging task, the ABCDE approach targets individual risk factors for delirium, which is an important contributor to cognitive impairment. Whether the ABCDE bundle directly affects the development of cognitive impairment requires further investigation.

COGNITIVE THERAPIES

The cognition-focused intervention most often described is cognitive training. Cognitive training is delivered in individual or group sessions in which the patient practices tasks targeting different domains, such as memory, language, and attention. Outcomes are often assessed in terms of improvement in test scores or effects on everyday functioning. Unfortunately, because of heterogeneity among cognitive training interventions and studied populations, we cannot yet make strong evidence-based recommendations for clinical practice.

Martin et al⁵⁶ in 2011 reviewed cognition-based interventions for healthy older people and people with mild cognitive impairment and found 36 relevant studies. Of these, only 3 were in patients with mild cognitive impairment, while the rest were in healthy older people.^56–58 Overall, the only available data were related to the memory domain, and outcomes were mostly associated with immediate recall of words, paragraphs, and stories. Based on this, cognitive therapy is currently considered justified, as most patients with cognitive impairment after an ICU stay have memory problems.

Zelinski et al⁵⁹ conducted a randomized, controlled, double-blind study comparing outcomes in an intervention group that underwent a computerized cognitive training program with those in a control group that viewed videos on a variety of topics such as literature, art, and history. The intervention, based on brain plasticity, aimed to improve the speed and accuracy of auditory information processing and to engage neuromodulatory systems. Some of the secondary outcomes favored the intervention group. These outcomes were related mostly to measures of overall memory, such as immediate and delayed recall, but also to a composite outcome that included letter-number sequencing and the digit span backwards test.

Despite these encouraging results, it is worth mentioning that these studies were not performed in patients with cognitive impairment associated with ICU admission. Therefore, the applicability and effectiveness of such therapies in post-ICU patients remains unknown.

Patients with posttraumatic brain injury and stroke have also been extensively studied in regard to the development of cognitive impairment.³⁴ These patients probably represent a better standard for comparison, as their cognitive impairment does not necessarily progress.

The effect of cognitive rehabilitation on the recovery in these patients depends on adaptation and remediation. Adaptation describes a patient’s ability to compensate for functional impairment.³⁴ This can be divided into internal and external adaptation. Internal adaptation requires the patient to recognize his or her cognitive limitation in order to adapt the to the environment accordingly. External adaptation entails getting help from devices or relatives (eg, phone calls) to achieve desired goals (eg, taking medication at scheduled times). Again, to adapt, the patient needs to be able to recognize his or her affected cognitive domain. Unfortunately, this is not always the case.

Remediation refers to the actual regaining of a lost ability. To stimulate neural plasticity, the patient is required to experience and repeat targeted skill-building activities.³⁸ There is evidence that patients are more likely to regain lost ability by repeating the practice frequently during a short period of time.⁶⁰

From the physician’s perspective, evaluating and identifying deficits in particular cognitive domains may help in designing a remediation plan in partnership with a cognitive therapist.

Cognitive rehabilitation in ICU survivors

The Returning to Everyday Tasks Utilizing Rehabilitation Networks (RETURN) study focused on cognitive and physical rehabilitation in post-ICU patients.⁶¹ This pilot study included 21 ICU survivors with cognitive or functional impairment at hospital discharge. Eight patients received usual care and 13 received a combination of in-home cognitive, physical, and functional rehabilitation over a 3-month period with a social worker or a master’s-level psychology technician.

Interventions included six in-person visits for cognitive rehabilitation and six televisits for physical and functional rehabilitation. Cognitive training was based on the goal-management training (GMT) protocol.⁶² This strategy attempts to improve executive function by increasing goal-directed behavior and by helping patients learn to be reflective before making decisions and executing tasks. The GMT model consists of sessions that build on one another to increase the rehabilitation intensity. During each session, goals are explained and participants perform increasingly challenging cognitive tasks.

Cognitive outcomes were evaluated using the Delis-Kaplan Tower Test to evaluate executive function by assessing the ability to plan and strategize efficiently. The patient is required to move disks across three pegs until a tower is built. The object is to use the fewest moves possible while adhering to two rules: larger disks cannot be placed on top of smaller ones, and disks must be moved one at a time, using only one hand.

At 3 months there was a significant difference between groups, with the intervention group earning higher tower test scores than controls did (median of 13 vs 7.5).

The Activity and Cognitive Therapy in the Intensive Care Unit (ACT-ICU) trial is another pilot study that will attempt to assess the feasibility of early cognitive rehabilitation in ICU survivors. This study will combine early mobilization with a cognitive intervention, and its primary outcome is executive function (with the tower test) at 3 months after discharge.⁶³

DRUG THERAPY

Some medications have been tested to assess whether they reduce the risk of progression from adult traumatic brain injury to cognitive impairment. These drugs augment dopamine and acetylcholine activity.

Methylphenidate (Ritalin), a dopaminergic drug, was studied in two trials. The first was a double-blind trial in 18 patients with posttraumatic brain injury. Memory was found to improve, based on the Working Memory Task Test. However, due to the small number of participants, no further conclusions were obtained.⁶⁴

The second trial, in 19 patients with posttraumatic brain injury, had a double-blind crossover design. Attention, evaluated by the Distraction Task Test, improved with the use of methylphenidate.⁶⁵ Again, the small number of patients precludes generalization of these results.

Donepezil (Aricept), a cholinergic drug, was evaluated in four clinical trials in posttraumatic brain injury patients^66–69; each trial included 21 to 180 patients. The trials evaluated the drug’s effect on memory and attention through a variety of tools (Paced Auditory Serial Addition Test; Wechsler Memory Scale; Boston Naming Test; Rey Auditory Verbal Learning Test; Complex Figure Test; and Reaction Time–Dual Task). Interestingly, donepezil was associated with large improvements in objective assessments of attention and memory. Despite methodologic flaws, such as a lack of blinding in one of these studies⁶⁹ and an open-label design in two of them,^66,68 of the drugs available, donepezil presents the strongest evidence for use in cognitive impairment after traumatic brain injury.⁷⁰

Intensive care medicine has dramatically evolved over the last 15 years, after reports from many landmark trials.¹ Updated strategies for mechanical ventilation² and “bundles” of strategies to optimize hemodynamic therapy³ have reduced the rates of morbidity and death from deadly critical conditions such as the adult respiratory distress syndrome (ARDS) and sepsis.

Despite these important improvements in short-term outcomes, it is increasingly recognized that intensive care unit (ICU) survivors suffer considerable long-term complications that affect their usual functioning.⁴ Recently, the Society of Critical Care Medicine convened a conference in which these long-term complications were named the “post-intensive care syndrome.”⁵

Quality of life, particularly its physical component, is considerably lower after a stay in the medical or surgical ICU.^6–8 Posttraumatic stress disorder, depression, and sexual dysfunction are consistently reported years after ICU discharge.^9–13

Perhaps the most frequently unrecognized complication in ICU survivors is cognitive impairment. Current data suggest that neurocognitive impairment after an ICU stay is common and that it persists 6 years or more after hospital discharge.

Hopkins et al^14,15 analyzed 10 cohort studies of long-term cognitive impairment after an ICU stay; 5 of them focused on patients with ARDS. The prevalence of cognitive impairment was as high as 78% at hospital discharge, 46% at 1 year, and 25% 6 years after discharge.^15,16 Of the cognitive domains compromised, memory was the most often affected, followed by executive function and attention.^14,17

Interestingly, data suggest that cognition may improve somewhat in the first 6 to 12 months after ICU discharge.¹⁵ Therefore, if we can detect it early on and promptly refer patients for cognitive therapy, we may eventually improve the prognosis of this disabling complication.

This review will focus on how to evaluate, prevent, and treat cognitive impairment in patients who survive an ICU stay.

COGNITIVE IMPAIRMENT AFTER A STAY IN THE ICU

The association between ICU stay and neurocognitive dysfunction is poorly understood. Potential causes include hypoxemia,¹⁸ hypotension, ¹⁹ hyperglycemia,¹⁴ and—an area of growing interest and evolving research—sedation and delirium.²⁰

Patients on mechanical ventilation are commonly given sedatives and analgesics to prevent anxiety and pain.²¹ However, these medications are strongly associated with delirium.²² In fact, recent studies found that benzodiazepines have an independent, dose-related, temporal association with delirium, with some reports describing a 20% increase in delirium per milligram of benzodiazepine.²³ In another study, which included medical and surgical ICU patients, use of morphine was the strongest predictor of delirium, with a sixfold increase in odds over a period of 5 months.²⁴

Delirium is important to prevent, diagnose, and treat, since it has a direct association with the development of long-term cognitive impairment.^22,25 A review of studies that included 1,885 medical and surgical patients found that those who developed delirium during an ICU stay were three times more likely to have cognitive dysfunction when assessed 3 years later.²⁰

Whether delirium is a primary disorder associated with cognitive impairment or if it only represents an underlying process leading to poor cognitive outcomes is unknown. As delirious patients are more likely to be older, to be mechanically ventilated, to require more sedation, and, in particular, to be sicker, the association between delirium and cognitive impairment may reflect the relationship between these risk factors and poor cognitive outcomes.²⁶

Glucose and its relationship with cognitive function is another topic of investigation. A secondary analysis of a study that included ARDS survivors revealed that blood glucose values higher than 153 mg/dL, higher glucose variability, and duration of mechanical ventilation were associated with cognitive sequelae.^27,28

Other studies focused on mechanical ventilation. In one study,²⁹ one-third of patients who had been mechanically ventilated showed signs of neurocognitive impairment when they were evaluated 6 months after hospital discharge.

Mild cognitive impairment differs from cognitive impairment after an ICU stay

Cognitive impairment after ICU discharge does not follow the same pattern as mild cognitive impairment, and some authors consider these two types of cognitive impairment to be unrelated.

While mild cognitive impairment is progressive and associated with aging, cognitive impairment in ICU survivors develops rapidly after acute illness and is usually related to numerous pathologic and neurochemical pathways.

For example, the neurotransmitter acetylcholine is thought to be involved in cognitive function as well as neuroplasticity of the motor cortex. In a model of cognitive impairment after stroke, activity of the cholinergic system was reduced.^30,31 Further, in a study in rats, Baskerville et al³² showed that experience-dependent plasticity could be completely blocked by damaging the cholinergic neurons in the nucleus basalis of Meynert, thereby affecting memory and other functions supported by this pathway.

Another implicated pathway involves dopamine. Of interest, dopamine augmentation has been shown to enhance simple motor memories and to improve procedural learning. Understanding of these neurochemical alterations opens opportunities for investigation of drug therapies.

ASSESSMENT TOOLS

Cognitive impairment is important to detect in ICU survivors because it predicts poor outcomes from rehabilitation. A study of stroke patients found that those with cognitive alterations immediately after the stroke were less likely to be discharged home or to be living at home 6 months after discharge.³³

A possible explanation may be that affected patients cannot fully participate in rehabilitation activities, owing to impairment in executive function, inability to remember therapy instructions, or disruption of implicit and explicit learning. Indeed, some authors consider cognitive impairment after acquired brain injury to be the most relevant surrogate marker of rehabilitation potential. Consequently, manipulation or enhancement of cognition may directly affect rehabilitation outcomes.³⁴

Disagreement about terminology and diagnostic criteria creates a problem for health care providers working with patients with potential cognitive impairment. Numerous systems have been proposed to define this condition; in fact, Stephan et al³⁵ reviewed the literature and found no fewer than 17. None of them is specific for cognitive impairment after an ICU stay.

Petersen et al³⁶ in 1999 proposed initial criteria for mild cognitive impairment that included the following:

• A memory complaint
• Normal general cognitive functioning
• Normal activities of daily living
• Memory impairment in relation to age and education
• No dementia.

Later, other areas of impairment besides memory were recognized, such as language, attention, perception, reasoning, and motor planning.³⁷ Therefore, mild cognitive impairment is currently classified into subtypes, which include amnestic (affecting single or multiple domains) and nonamnestic (also affecting single or multiple domains).³⁸

In clinical practice, impairment of specific cognitive domains may be challenging to detect, and neuropsychological testing is often needed. Cognitive screening tests can detect impairment across a restricted range of cognitive abilities, while more comprehensive assessments address each of the primary domains of cognition.³⁹ Formal testing provides normative and validated data on cognition performance and severity.

The Montreal Cognitive Assessment⁴⁰ is popular, comprehensive, used in a variety of professions in diverse types of facilities (acute care, rehabilitation, and skilled care facilities), and brief (taking 11 minutes to administer). It evaluates orientation, memory, language, attention, reasoning, and visual-constructional abilities. The maximum score is 30; cognitive impairment is defined as a score of less than 26. It has a sensitivity of 90% and a specificity of 87%.

The Folstein Mini-Mental State Examination (MMSE) is the most commonly used of the noncomprehensive tests in clinical practice.⁴¹ It assesses orientation, memory, language, attention, and praxis. It has a maximum score of 30 points; the cutoff score for cognitive impairment is 24 points or less.

A limitation of the MMSE is that its sensitivity is very low, ranging from 1% to 49%.^42,43 The MMSE scores of patients with cognitive impairment overlap considerably with those of age-matched healthy controls.³⁹ Conversely, the MMSE’s specificity is usually high, ranging from 85% to 100%.⁴²

Moreover, the MMSE poses copyright issues, an important consideration when selecting a test. In 2001, the authors of the MMSE transferred all intellectual property rights to Psychological Assessment Resources, which has exclusive rights to publish, license, and manage all intellectual property rights in all media and languages. Photocopying and using the MMSE without applying for permission from and paying this company ($1.23 per use) constitutes copyright infringement. Therefore, health care providers and researchers have been using other tests to evaluate cognition.

Other tests of cognition assess individual domains. Interestingly, studies of long-term cognitive impairment after ICU admission used these tests to define outcomes.²⁵ Specific tests include:

• The Digit Span and the Trailmaking Test A (used to assess attention and orientation)²⁵
• The Rey Auditory Verbal Learning Test (used to evaluate verbal memory)
• The Complex Figure Test (helpful in defining visual-spatial construction and delayed visual memory)
• The Trailmaking Test B (also included in the Montreal Cognitive Assessment; assesses executive functioning).

Besides formal testing, an informal battery is often recommended to provide additional information. An informal evaluation includes word definition, reading and verbal fluency, reading comprehension, and performance of instrumental activities of daily living. Observing as patients perform tasks of daily living provides therapists with a vast amount of information, as these tasks require using multiple cognitive processes. Therefore, if a functional breakdown occurs during this assessment, the clinician needs to identify the domain or specific level of cognitive dysfunction involved in that deficit.⁴⁴

PREVENTIVE STRATEGIES

Strategies for minimizing the long-term effects of cognitive impairment have mostly focused on preventing it.

During the ICU stay, optimizing hemodynamic, glucose, and oxygenation levels may prevent future long-term complications.¹⁸

Also, the association between sedation, delirium, and consequent cognitive impairment (see above) has led many investigators to apply the “ABCDE” bundle of strategies.^25,45,46 Specifically, ABCDE stands for awakening and breathing, choice of sedatives with fewer adverse effects, daily delirium monitoring, and early mobility exercise. These strategies have been shown in randomized controlled trials to prevent delirium; however, they have not been proved to prevent cognitive impairment.

Awakening and breathing

In the Awakening and Breathing Controlled Trial,⁴⁷ patients in the intervention group (ie, those who had their sedatives interrupted every morning to see if they would awaken, and if so, if they could breathe on their own) were extubated 3 days sooner than those in the control group (who underwent daily trials of spontaneous breathing, if deemed safe). Also, ICU and hospital length of stay were shorter by 4 days. Best of all, over 1 year, the mortality rate was lower by 14 absolute percentage points.

Choice of sedatives

Often, mechanically ventilated patients are given benzodiazepines, opiates, and propofol (Diprivan).²¹ Dexmedetomidine (Precedex), a newer agent, is an alpha-2 agonist and may offer advantages over the others.

To date, three randomized controlled trials have assessed the effect of dexmedetomidine in terms of outcomes associated with delirium, and one trial evaluated its association with intellectual capacity in ICU patients.

The Maximizing Efficacy of Targeted Sedation and Reducing Neurological Dysfunction (MENDS) trial randomized patients on mechanical ventilation to receive either dexmedetomidine or lorazepam (Ativan).⁴⁸ Dexmedetomidine-treated patients had 4 more days alive without delirium or coma (7 vs 3 days, P = .01).

Subsequently, the Safety and Efficacy of Dexmedetomidine Compared With Midazolam (SEDCOM) trial compared dexmedetomidine and midazolam (Versed) in mechanically ventilated patients. Those who received dexmedetomidine had a lower incidence of delirium (54% vs 76%, P < .001), and 2 fewer days on mechanical ventilation.⁴⁹

Reade et al⁵⁰ evaluated time to extubation in already delirious patients randomized to receive either dexmedetomidine or haloperidol (Haldol). Those receiving dexmedetomidine had a shorter time to extubation as well as a shorter ICU length of stay.

The Acute Neuroscience Intensive Care Sedation Trial⁵¹ evaluated intellectual capacity in neurological ICU patients sedated with either dexmedetomidine or propofol. This randomized, double-blind trial included 18 brain-injured and 12 non-brain-injured intubated patients. In a crossover protocol, each received the combination of fentanyl (Sublimaze) and propofol and the combination of fentanyl and dexmedetomidine.

Cognition was evaluated using the Adapted Cognitive Exam (ACE), which assesses intellectual capacity through orientation, language, registration, attention, calculation, and recall. This 10-minute examination does not require verbal communication, as it relies on the ability to respond to yes-or-no questions and perform simple motor tasks. The maximum possible score is 100 points.

Interestingly, while on propofol, the patients’ adjusted ACE scores went down by a mean of 12.4 points, whereas they went up by 6.8 points while on dexmedetomidine. Even though brain-injured patients required less sedation than non-brain-injured patients, the effect of dexmedetomidine and propofol did not change.⁵¹

In summary, these studies suggest that all sedatives are not the same in their short-term and intermediate-term outcomes.

In our practice, we use dexmedetomidine as our first-line sedation therapy. In patients with hemodynamic instability, we use benzodiazepines. We reserve propofol for very short periods of intubation or for hemodynamically stable patients who cannot be sedated with dexmedetomidine.

Daily delirium monitoring

As mentioned above, delirium affects many patients on mechanical ventilation, and it is highly underrecognized if valid tests are not used.⁵² Therefore, it is critically important to be familiar with the tests for assessing delirium. Of these, the Confusion Assessment Method for the ICU is probably the one with the best performance, with a sensitivity of 93% to 100% and a specificity of 98% to 100%.^53,54

Early mobilization

A landmark study paired the awakening and breathing strategy with early mobilization through physical and occupational therapy in the ICU.⁵⁵ Patients in the intervention group had a higher rate of return to independent functional status upon hospital discharge and a shorter duration of mechanical ventilation and delirium.

In conclusion, even though direct prevention of cognitive dysfunction is a challenging task, the ABCDE approach targets individual risk factors for delirium, which is an important contributor to cognitive impairment. Whether the ABCDE bundle directly affects the development of cognitive impairment requires further investigation.

COGNITIVE THERAPIES

The cognition-focused intervention most often described is cognitive training. Cognitive training is delivered in individual or group sessions in which the patient practices tasks targeting different domains, such as memory, language, and attention. Outcomes are often assessed in terms of improvement in test scores or effects on everyday functioning. Unfortunately, because of heterogeneity among cognitive training interventions and studied populations, we cannot yet make strong evidence-based recommendations for clinical practice.

Martin et al⁵⁶ in 2011 reviewed cognition-based interventions for healthy older people and people with mild cognitive impairment and found 36 relevant studies. Of these, only 3 were in patients with mild cognitive impairment, while the rest were in healthy older people.^56–58 Overall, the only available data were related to the memory domain, and outcomes were mostly associated with immediate recall of words, paragraphs, and stories. Based on this, cognitive therapy is currently considered justified, as most patients with cognitive impairment after an ICU stay have memory problems.

Zelinski et al⁵⁹ conducted a randomized, controlled, double-blind study comparing outcomes in an intervention group that underwent a computerized cognitive training program with those in a control group that viewed videos on a variety of topics such as literature, art, and history. The intervention, based on brain plasticity, aimed to improve the speed and accuracy of auditory information processing and to engage neuromodulatory systems. Some of the secondary outcomes favored the intervention group. These outcomes were related mostly to measures of overall memory, such as immediate and delayed recall, but also to a composite outcome that included letter-number sequencing and the digit span backwards test.

Despite these encouraging results, it is worth mentioning that these studies were not performed in patients with cognitive impairment associated with ICU admission. Therefore, the applicability and effectiveness of such therapies in post-ICU patients remains unknown.

Patients with posttraumatic brain injury and stroke have also been extensively studied in regard to the development of cognitive impairment.³⁴ These patients probably represent a better standard for comparison, as their cognitive impairment does not necessarily progress.

The effect of cognitive rehabilitation on the recovery in these patients depends on adaptation and remediation. Adaptation describes a patient’s ability to compensate for functional impairment.³⁴ This can be divided into internal and external adaptation. Internal adaptation requires the patient to recognize his or her cognitive limitation in order to adapt the to the environment accordingly. External adaptation entails getting help from devices or relatives (eg, phone calls) to achieve desired goals (eg, taking medication at scheduled times). Again, to adapt, the patient needs to be able to recognize his or her affected cognitive domain. Unfortunately, this is not always the case.

Remediation refers to the actual regaining of a lost ability. To stimulate neural plasticity, the patient is required to experience and repeat targeted skill-building activities.³⁸ There is evidence that patients are more likely to regain lost ability by repeating the practice frequently during a short period of time.⁶⁰

From the physician’s perspective, evaluating and identifying deficits in particular cognitive domains may help in designing a remediation plan in partnership with a cognitive therapist.

Cognitive rehabilitation in ICU survivors

The Returning to Everyday Tasks Utilizing Rehabilitation Networks (RETURN) study focused on cognitive and physical rehabilitation in post-ICU patients.⁶¹ This pilot study included 21 ICU survivors with cognitive or functional impairment at hospital discharge. Eight patients received usual care and 13 received a combination of in-home cognitive, physical, and functional rehabilitation over a 3-month period with a social worker or a master’s-level psychology technician.

Interventions included six in-person visits for cognitive rehabilitation and six televisits for physical and functional rehabilitation. Cognitive training was based on the goal-management training (GMT) protocol.⁶² This strategy attempts to improve executive function by increasing goal-directed behavior and by helping patients learn to be reflective before making decisions and executing tasks. The GMT model consists of sessions that build on one another to increase the rehabilitation intensity. During each session, goals are explained and participants perform increasingly challenging cognitive tasks.

Cognitive outcomes were evaluated using the Delis-Kaplan Tower Test to evaluate executive function by assessing the ability to plan and strategize efficiently. The patient is required to move disks across three pegs until a tower is built. The object is to use the fewest moves possible while adhering to two rules: larger disks cannot be placed on top of smaller ones, and disks must be moved one at a time, using only one hand.

At 3 months there was a significant difference between groups, with the intervention group earning higher tower test scores than controls did (median of 13 vs 7.5).

The Activity and Cognitive Therapy in the Intensive Care Unit (ACT-ICU) trial is another pilot study that will attempt to assess the feasibility of early cognitive rehabilitation in ICU survivors. This study will combine early mobilization with a cognitive intervention, and its primary outcome is executive function (with the tower test) at 3 months after discharge.⁶³

DRUG THERAPY

Some medications have been tested to assess whether they reduce the risk of progression from adult traumatic brain injury to cognitive impairment. These drugs augment dopamine and acetylcholine activity.

Methylphenidate (Ritalin), a dopaminergic drug, was studied in two trials. The first was a double-blind trial in 18 patients with posttraumatic brain injury. Memory was found to improve, based on the Working Memory Task Test. However, due to the small number of participants, no further conclusions were obtained.⁶⁴

The second trial, in 19 patients with posttraumatic brain injury, had a double-blind crossover design. Attention, evaluated by the Distraction Task Test, improved with the use of methylphenidate.⁶⁵ Again, the small number of patients precludes generalization of these results.

Donepezil (Aricept), a cholinergic drug, was evaluated in four clinical trials in posttraumatic brain injury patients^66–69; each trial included 21 to 180 patients. The trials evaluated the drug’s effect on memory and attention through a variety of tools (Paced Auditory Serial Addition Test; Wechsler Memory Scale; Boston Naming Test; Rey Auditory Verbal Learning Test; Complex Figure Test; and Reaction Time–Dual Task). Interestingly, donepezil was associated with large improvements in objective assessments of attention and memory. Despite methodologic flaws, such as a lack of blinding in one of these studies⁶⁹ and an open-label design in two of them,^66,68 of the drugs available, donepezil presents the strongest evidence for use in cognitive impairment after traumatic brain injury.⁷⁰

Intensive care medicine has dramatically evolved over the last 15 years, after reports from many landmark trials.¹ Updated strategies for mechanical ventilation² and “bundles” of strategies to optimize hemodynamic therapy³ have reduced the rates of morbidity and death from deadly critical conditions such as the adult respiratory distress syndrome (ARDS) and sepsis.

Despite these important improvements in short-term outcomes, it is increasingly recognized that intensive care unit (ICU) survivors suffer considerable long-term complications that affect their usual functioning.⁴ Recently, the Society of Critical Care Medicine convened a conference in which these long-term complications were named the “post-intensive care syndrome.”⁵

Quality of life, particularly its physical component, is considerably lower after a stay in the medical or surgical ICU.^6–8 Posttraumatic stress disorder, depression, and sexual dysfunction are consistently reported years after ICU discharge.^9–13

Perhaps the most frequently unrecognized complication in ICU survivors is cognitive impairment. Current data suggest that neurocognitive impairment after an ICU stay is common and that it persists 6 years or more after hospital discharge.

Hopkins et al^14,15 analyzed 10 cohort studies of long-term cognitive impairment after an ICU stay; 5 of them focused on patients with ARDS. The prevalence of cognitive impairment was as high as 78% at hospital discharge, 46% at 1 year, and 25% 6 years after discharge.^15,16 Of the cognitive domains compromised, memory was the most often affected, followed by executive function and attention.^14,17

Interestingly, data suggest that cognition may improve somewhat in the first 6 to 12 months after ICU discharge.¹⁵ Therefore, if we can detect it early on and promptly refer patients for cognitive therapy, we may eventually improve the prognosis of this disabling complication.

This review will focus on how to evaluate, prevent, and treat cognitive impairment in patients who survive an ICU stay.

COGNITIVE IMPAIRMENT AFTER A STAY IN THE ICU

The association between ICU stay and neurocognitive dysfunction is poorly understood. Potential causes include hypoxemia,¹⁸ hypotension, ¹⁹ hyperglycemia,¹⁴ and—an area of growing interest and evolving research—sedation and delirium.²⁰

Patients on mechanical ventilation are commonly given sedatives and analgesics to prevent anxiety and pain.²¹ However, these medications are strongly associated with delirium.²² In fact, recent studies found that benzodiazepines have an independent, dose-related, temporal association with delirium, with some reports describing a 20% increase in delirium per milligram of benzodiazepine.²³ In another study, which included medical and surgical ICU patients, use of morphine was the strongest predictor of delirium, with a sixfold increase in odds over a period of 5 months.²⁴

Delirium is important to prevent, diagnose, and treat, since it has a direct association with the development of long-term cognitive impairment.^22,25 A review of studies that included 1,885 medical and surgical patients found that those who developed delirium during an ICU stay were three times more likely to have cognitive dysfunction when assessed 3 years later.²⁰

Whether delirium is a primary disorder associated with cognitive impairment or if it only represents an underlying process leading to poor cognitive outcomes is unknown. As delirious patients are more likely to be older, to be mechanically ventilated, to require more sedation, and, in particular, to be sicker, the association between delirium and cognitive impairment may reflect the relationship between these risk factors and poor cognitive outcomes.²⁶

Glucose and its relationship with cognitive function is another topic of investigation. A secondary analysis of a study that included ARDS survivors revealed that blood glucose values higher than 153 mg/dL, higher glucose variability, and duration of mechanical ventilation were associated with cognitive sequelae.^27,28

Other studies focused on mechanical ventilation. In one study,²⁹ one-third of patients who had been mechanically ventilated showed signs of neurocognitive impairment when they were evaluated 6 months after hospital discharge.

Mild cognitive impairment differs from cognitive impairment after an ICU stay

Cognitive impairment after ICU discharge does not follow the same pattern as mild cognitive impairment, and some authors consider these two types of cognitive impairment to be unrelated.

While mild cognitive impairment is progressive and associated with aging, cognitive impairment in ICU survivors develops rapidly after acute illness and is usually related to numerous pathologic and neurochemical pathways.

For example, the neurotransmitter acetylcholine is thought to be involved in cognitive function as well as neuroplasticity of the motor cortex. In a model of cognitive impairment after stroke, activity of the cholinergic system was reduced.^30,31 Further, in a study in rats, Baskerville et al³² showed that experience-dependent plasticity could be completely blocked by damaging the cholinergic neurons in the nucleus basalis of Meynert, thereby affecting memory and other functions supported by this pathway.

Another implicated pathway involves dopamine. Of interest, dopamine augmentation has been shown to enhance simple motor memories and to improve procedural learning. Understanding of these neurochemical alterations opens opportunities for investigation of drug therapies.

ASSESSMENT TOOLS

Cognitive impairment is important to detect in ICU survivors because it predicts poor outcomes from rehabilitation. A study of stroke patients found that those with cognitive alterations immediately after the stroke were less likely to be discharged home or to be living at home 6 months after discharge.³³

A possible explanation may be that affected patients cannot fully participate in rehabilitation activities, owing to impairment in executive function, inability to remember therapy instructions, or disruption of implicit and explicit learning. Indeed, some authors consider cognitive impairment after acquired brain injury to be the most relevant surrogate marker of rehabilitation potential. Consequently, manipulation or enhancement of cognition may directly affect rehabilitation outcomes.³⁴

Disagreement about terminology and diagnostic criteria creates a problem for health care providers working with patients with potential cognitive impairment. Numerous systems have been proposed to define this condition; in fact, Stephan et al³⁵ reviewed the literature and found no fewer than 17. None of them is specific for cognitive impairment after an ICU stay.

Petersen et al³⁶ in 1999 proposed initial criteria for mild cognitive impairment that included the following:

• A memory complaint
• Normal general cognitive functioning
• Normal activities of daily living
• Memory impairment in relation to age and education
• No dementia.

Later, other areas of impairment besides memory were recognized, such as language, attention, perception, reasoning, and motor planning.³⁷ Therefore, mild cognitive impairment is currently classified into subtypes, which include amnestic (affecting single or multiple domains) and nonamnestic (also affecting single or multiple domains).³⁸

In clinical practice, impairment of specific cognitive domains may be challenging to detect, and neuropsychological testing is often needed. Cognitive screening tests can detect impairment across a restricted range of cognitive abilities, while more comprehensive assessments address each of the primary domains of cognition.³⁹ Formal testing provides normative and validated data on cognition performance and severity.

The Montreal Cognitive Assessment⁴⁰ is popular, comprehensive, used in a variety of professions in diverse types of facilities (acute care, rehabilitation, and skilled care facilities), and brief (taking 11 minutes to administer). It evaluates orientation, memory, language, attention, reasoning, and visual-constructional abilities. The maximum score is 30; cognitive impairment is defined as a score of less than 26. It has a sensitivity of 90% and a specificity of 87%.

The Folstein Mini-Mental State Examination (MMSE) is the most commonly used of the noncomprehensive tests in clinical practice.⁴¹ It assesses orientation, memory, language, attention, and praxis. It has a maximum score of 30 points; the cutoff score for cognitive impairment is 24 points or less.

A limitation of the MMSE is that its sensitivity is very low, ranging from 1% to 49%.^42,43 The MMSE scores of patients with cognitive impairment overlap considerably with those of age-matched healthy controls.³⁹ Conversely, the MMSE’s specificity is usually high, ranging from 85% to 100%.⁴²

Moreover, the MMSE poses copyright issues, an important consideration when selecting a test. In 2001, the authors of the MMSE transferred all intellectual property rights to Psychological Assessment Resources, which has exclusive rights to publish, license, and manage all intellectual property rights in all media and languages. Photocopying and using the MMSE without applying for permission from and paying this company ($1.23 per use) constitutes copyright infringement. Therefore, health care providers and researchers have been using other tests to evaluate cognition.

Other tests of cognition assess individual domains. Interestingly, studies of long-term cognitive impairment after ICU admission used these tests to define outcomes.²⁵ Specific tests include:

• The Digit Span and the Trailmaking Test A (used to assess attention and orientation)²⁵
• The Rey Auditory Verbal Learning Test (used to evaluate verbal memory)
• The Complex Figure Test (helpful in defining visual-spatial construction and delayed visual memory)
• The Trailmaking Test B (also included in the Montreal Cognitive Assessment; assesses executive functioning).

Besides formal testing, an informal battery is often recommended to provide additional information. An informal evaluation includes word definition, reading and verbal fluency, reading comprehension, and performance of instrumental activities of daily living. Observing as patients perform tasks of daily living provides therapists with a vast amount of information, as these tasks require using multiple cognitive processes. Therefore, if a functional breakdown occurs during this assessment, the clinician needs to identify the domain or specific level of cognitive dysfunction involved in that deficit.⁴⁴

PREVENTIVE STRATEGIES

Strategies for minimizing the long-term effects of cognitive impairment have mostly focused on preventing it.

During the ICU stay, optimizing hemodynamic, glucose, and oxygenation levels may prevent future long-term complications.¹⁸

Also, the association between sedation, delirium, and consequent cognitive impairment (see above) has led many investigators to apply the “ABCDE” bundle of strategies.^25,45,46 Specifically, ABCDE stands for awakening and breathing, choice of sedatives with fewer adverse effects, daily delirium monitoring, and early mobility exercise. These strategies have been shown in randomized controlled trials to prevent delirium; however, they have not been proved to prevent cognitive impairment.

Awakening and breathing

In the Awakening and Breathing Controlled Trial,⁴⁷ patients in the intervention group (ie, those who had their sedatives interrupted every morning to see if they would awaken, and if so, if they could breathe on their own) were extubated 3 days sooner than those in the control group (who underwent daily trials of spontaneous breathing, if deemed safe). Also, ICU and hospital length of stay were shorter by 4 days. Best of all, over 1 year, the mortality rate was lower by 14 absolute percentage points.

Choice of sedatives

Often, mechanically ventilated patients are given benzodiazepines, opiates, and propofol (Diprivan).²¹ Dexmedetomidine (Precedex), a newer agent, is an alpha-2 agonist and may offer advantages over the others.

To date, three randomized controlled trials have assessed the effect of dexmedetomidine in terms of outcomes associated with delirium, and one trial evaluated its association with intellectual capacity in ICU patients.

The Maximizing Efficacy of Targeted Sedation and Reducing Neurological Dysfunction (MENDS) trial randomized patients on mechanical ventilation to receive either dexmedetomidine or lorazepam (Ativan).⁴⁸ Dexmedetomidine-treated patients had 4 more days alive without delirium or coma (7 vs 3 days, P = .01).

Subsequently, the Safety and Efficacy of Dexmedetomidine Compared With Midazolam (SEDCOM) trial compared dexmedetomidine and midazolam (Versed) in mechanically ventilated patients. Those who received dexmedetomidine had a lower incidence of delirium (54% vs 76%, P < .001), and 2 fewer days on mechanical ventilation.⁴⁹

Reade et al⁵⁰ evaluated time to extubation in already delirious patients randomized to receive either dexmedetomidine or haloperidol (Haldol). Those receiving dexmedetomidine had a shorter time to extubation as well as a shorter ICU length of stay.

The Acute Neuroscience Intensive Care Sedation Trial⁵¹ evaluated intellectual capacity in neurological ICU patients sedated with either dexmedetomidine or propofol. This randomized, double-blind trial included 18 brain-injured and 12 non-brain-injured intubated patients. In a crossover protocol, each received the combination of fentanyl (Sublimaze) and propofol and the combination of fentanyl and dexmedetomidine.

Cognition was evaluated using the Adapted Cognitive Exam (ACE), which assesses intellectual capacity through orientation, language, registration, attention, calculation, and recall. This 10-minute examination does not require verbal communication, as it relies on the ability to respond to yes-or-no questions and perform simple motor tasks. The maximum possible score is 100 points.

Interestingly, while on propofol, the patients’ adjusted ACE scores went down by a mean of 12.4 points, whereas they went up by 6.8 points while on dexmedetomidine. Even though brain-injured patients required less sedation than non-brain-injured patients, the effect of dexmedetomidine and propofol did not change.⁵¹

In summary, these studies suggest that all sedatives are not the same in their short-term and intermediate-term outcomes.

In our practice, we use dexmedetomidine as our first-line sedation therapy. In patients with hemodynamic instability, we use benzodiazepines. We reserve propofol for very short periods of intubation or for hemodynamically stable patients who cannot be sedated with dexmedetomidine.

Daily delirium monitoring

As mentioned above, delirium affects many patients on mechanical ventilation, and it is highly underrecognized if valid tests are not used.⁵² Therefore, it is critically important to be familiar with the tests for assessing delirium. Of these, the Confusion Assessment Method for the ICU is probably the one with the best performance, with a sensitivity of 93% to 100% and a specificity of 98% to 100%.^53,54

Early mobilization

A landmark study paired the awakening and breathing strategy with early mobilization through physical and occupational therapy in the ICU.⁵⁵ Patients in the intervention group had a higher rate of return to independent functional status upon hospital discharge and a shorter duration of mechanical ventilation and delirium.

In conclusion, even though direct prevention of cognitive dysfunction is a challenging task, the ABCDE approach targets individual risk factors for delirium, which is an important contributor to cognitive impairment. Whether the ABCDE bundle directly affects the development of cognitive impairment requires further investigation.

COGNITIVE THERAPIES

The cognition-focused intervention most often described is cognitive training. Cognitive training is delivered in individual or group sessions in which the patient practices tasks targeting different domains, such as memory, language, and attention. Outcomes are often assessed in terms of improvement in test scores or effects on everyday functioning. Unfortunately, because of heterogeneity among cognitive training interventions and studied populations, we cannot yet make strong evidence-based recommendations for clinical practice.

Martin et al⁵⁶ in 2011 reviewed cognition-based interventions for healthy older people and people with mild cognitive impairment and found 36 relevant studies. Of these, only 3 were in patients with mild cognitive impairment, while the rest were in healthy older people.^56–58 Overall, the only available data were related to the memory domain, and outcomes were mostly associated with immediate recall of words, paragraphs, and stories. Based on this, cognitive therapy is currently considered justified, as most patients with cognitive impairment after an ICU stay have memory problems.

Zelinski et al⁵⁹ conducted a randomized, controlled, double-blind study comparing outcomes in an intervention group that underwent a computerized cognitive training program with those in a control group that viewed videos on a variety of topics such as literature, art, and history. The intervention, based on brain plasticity, aimed to improve the speed and accuracy of auditory information processing and to engage neuromodulatory systems. Some of the secondary outcomes favored the intervention group. These outcomes were related mostly to measures of overall memory, such as immediate and delayed recall, but also to a composite outcome that included letter-number sequencing and the digit span backwards test.

Despite these encouraging results, it is worth mentioning that these studies were not performed in patients with cognitive impairment associated with ICU admission. Therefore, the applicability and effectiveness of such therapies in post-ICU patients remains unknown.

Patients with posttraumatic brain injury and stroke have also been extensively studied in regard to the development of cognitive impairment.³⁴ These patients probably represent a better standard for comparison, as their cognitive impairment does not necessarily progress.

The effect of cognitive rehabilitation on the recovery in these patients depends on adaptation and remediation. Adaptation describes a patient’s ability to compensate for functional impairment.³⁴ This can be divided into internal and external adaptation. Internal adaptation requires the patient to recognize his or her cognitive limitation in order to adapt the to the environment accordingly. External adaptation entails getting help from devices or relatives (eg, phone calls) to achieve desired goals (eg, taking medication at scheduled times). Again, to adapt, the patient needs to be able to recognize his or her affected cognitive domain. Unfortunately, this is not always the case.

Remediation refers to the actual regaining of a lost ability. To stimulate neural plasticity, the patient is required to experience and repeat targeted skill-building activities.³⁸ There is evidence that patients are more likely to regain lost ability by repeating the practice frequently during a short period of time.⁶⁰

From the physician’s perspective, evaluating and identifying deficits in particular cognitive domains may help in designing a remediation plan in partnership with a cognitive therapist.

Cognitive rehabilitation in ICU survivors

The Returning to Everyday Tasks Utilizing Rehabilitation Networks (RETURN) study focused on cognitive and physical rehabilitation in post-ICU patients.⁶¹ This pilot study included 21 ICU survivors with cognitive or functional impairment at hospital discharge. Eight patients received usual care and 13 received a combination of in-home cognitive, physical, and functional rehabilitation over a 3-month period with a social worker or a master’s-level psychology technician.

Interventions included six in-person visits for cognitive rehabilitation and six televisits for physical and functional rehabilitation. Cognitive training was based on the goal-management training (GMT) protocol.⁶² This strategy attempts to improve executive function by increasing goal-directed behavior and by helping patients learn to be reflective before making decisions and executing tasks. The GMT model consists of sessions that build on one another to increase the rehabilitation intensity. During each session, goals are explained and participants perform increasingly challenging cognitive tasks.

Cognitive outcomes were evaluated using the Delis-Kaplan Tower Test to evaluate executive function by assessing the ability to plan and strategize efficiently. The patient is required to move disks across three pegs until a tower is built. The object is to use the fewest moves possible while adhering to two rules: larger disks cannot be placed on top of smaller ones, and disks must be moved one at a time, using only one hand.

At 3 months there was a significant difference between groups, with the intervention group earning higher tower test scores than controls did (median of 13 vs 7.5).

The Activity and Cognitive Therapy in the Intensive Care Unit (ACT-ICU) trial is another pilot study that will attempt to assess the feasibility of early cognitive rehabilitation in ICU survivors. This study will combine early mobilization with a cognitive intervention, and its primary outcome is executive function (with the tower test) at 3 months after discharge.⁶³

DRUG THERAPY

Some medications have been tested to assess whether they reduce the risk of progression from adult traumatic brain injury to cognitive impairment. These drugs augment dopamine and acetylcholine activity.

Methylphenidate (Ritalin), a dopaminergic drug, was studied in two trials. The first was a double-blind trial in 18 patients with posttraumatic brain injury. Memory was found to improve, based on the Working Memory Task Test. However, due to the small number of participants, no further conclusions were obtained.⁶⁴

The second trial, in 19 patients with posttraumatic brain injury, had a double-blind crossover design. Attention, evaluated by the Distraction Task Test, improved with the use of methylphenidate.⁶⁵ Again, the small number of patients precludes generalization of these results.

Donepezil (Aricept), a cholinergic drug, was evaluated in four clinical trials in posttraumatic brain injury patients^66–69; each trial included 21 to 180 patients. The trials evaluated the drug’s effect on memory and attention through a variety of tools (Paced Auditory Serial Addition Test; Wechsler Memory Scale; Boston Naming Test; Rey Auditory Verbal Learning Test; Complex Figure Test; and Reaction Time–Dual Task). Interestingly, donepezil was associated with large improvements in objective assessments of attention and memory. Despite methodologic flaws, such as a lack of blinding in one of these studies⁶⁹ and an open-label design in two of them,^66,68 of the drugs available, donepezil presents the strongest evidence for use in cognitive impairment after traumatic brain injury.⁷⁰

The management of hypertension has advanced significantly in the last few decades. But the race for more effective means to control this epidemic and its associated complications is far from won. A high percentage of patients in the United States have hypertension that is uncontrolled. Most of these belong to the most rapidly growing demographic group in the United States, ie, the elderly.

It is estimated that more than 70% of medical practice will be directed to geriatric needs in the coming years. It is therefore very important for clinicians to be comfortable with managing hypertension in the elderly.

A GROWING PROBLEM IN AN AGING POPULATION

Between 1980 and 2009, the US population age 65 and older increased from 25.6 million to 39.6 million, of which 42% are men and 58% women.¹ This number is expected to reach 75 million by the year 2040. People over 85 years of age are the fastest growing subset of the US population.² As many as 50% of people who were born recently in countries such as the United States, the United Kingdom, France, Denmark, and Japan will live to celebrate their 100th birthday.³

According to the Framingham Heart Study, by age 60 approximately 60% of the population develops hypertension, and by 70 years about 65% of men and about 75% of women have the disease. In the same study, 90% of those who were normotensive at age 55 went on to develop hypertension. The elderly also are more likely to suffer from the complications of hypertension and are more likely to have uncontrolled disease.

Compared with younger patients with similar blood pressure, elderly hypertensive patients have lower cardiac output, higher peripheral resistance, wider pulse pressure, lower intravascular volume, and lower renal blood flow.⁴ These age-related pathophysiologic differences must be considered when treating antihypertension in the elderly.

IS TREATING THE ELDERLY BENEFICIAL?

Most elderly hypertensive patients have multiple comorbidities, which tremendously affect the management of their hypertension. They are also more likely than younger patients to have resistant hypertension and to need multiple drugs to control their blood pressure. In the process, these frail patients are exposed to a host of drug-related adverse effects. Thus, it is relevant to question the net benefit of treatment in this age group.

Many studies have indeed shown that treating hypertension reduces the risk of stroke and other adverse cardiovascular events. A decade ago, Staessen et al,⁵ in a meta-analysis of more than 15,000 patients between ages 62 and 76, showed that treating isolated systolic hypertension substantially reduced morbidity and mortality rates. Moreover, a 2011 meta-analysis of randomized controlled trials in hypertensive patients age 75 and over also concluded that treatment reduced cardiovascular morbidity and mortality rates and the incidence of heart failure, even though the total mortality rate was not affected.⁶

Opinion on treating the very elderly (≥ 80 years of age) was divided until the results of the Hypertension in the Very Elderly trial (HYVET)⁷ came out in 2008. This study documented major benefits of treatment in the very elderly age group as well.

The consensus, therefore, is that it is appropriate, even imperative, to treat elderly hypertensive patients (with some cautions—see the sections that follow).

GOAL OF TREATMENT IN THE ELDERLY

Targets for blood pressure management have been based primarily on observational data in middle-aged patients. There is no such thing as an ideal blood pressure that has been derived from randomized controlled trials for any population, let alone the elderly. The generally recommended blood pressure goal of 140/90 mm Hg for elderly hypertensive patients is based on expert opinion.

Moreover, it is unclear if the same target should apply to octogenarians. According to a 2011 American College of Cardiology/American Heart Association (ACC/AHA) expert consensus report,⁸ an achieved systolic blood pressure of 140 to 145 mm Hg, if tolerated, can be acceptable in this age group.

An orthostatic decline in blood pressure accompanies advanced age and is an inevitable adverse effect of some antihypertensive drugs. Accordingly, systolic blood pressure lower than 130 and diastolic blood pressure lower than 70 mm Hg are best avoided in octogenarians.⁸ Therefore, when hypertension is complicated by coexisting conditions that require a specific blood pressure goal, it would seem reasonable to not pursue the lower target as aggressively in octogenarians as in elderly patients under age 80.

Having stated the limitations in the quality of data at hand—largely observational—it is relevant to mention the Systolic Blood Pressure Intervention trial (SPRINT).⁹ This ongoing randomized, multicenter trial, launched by the National Institutes of Health, is assessing whether maintaining blood pressure levels lower than current recommendations further reduces the risk of cardiovascular and kidney diseases or, in the SPRINT-MIND substudy, of age-related cognitive decline, regardless of the type of antihypertensive drug taken. Initially planning to enroll close to 10,000 participants over the age of 55 without specifying any agegroup ranges, the investigators later decided to conduct a substudy called SPRINT Senior that will enroll about 1,750 participants over the age of 75 to determine whether a lower blood pressure range will have the same beneficial effects in older adults.

Given the limitations in the quality and applicability of published data (coming from small, nonrandomized studies with no long-term follow-up), SPRINT is expected to provide the evidence needed to support standard vs aggressive hypertension control among the elderly. The trial is projected to run until late 2018.

MANAGEMENT APPROACH IN THE ELDERLY

Blood pressure should be recorded in both arms before a diagnosis is made. In a number of patients, particularly the elderly, there are significant differences in blood pressure readings between the two arms. The higher reading should be relied on and the corresponding arm used for subsequent measurements.

Lifestyle interventions

Similar to the approach in younger patients with hypertension, lifestyle interventions are the first step to managing high blood pressure in the elderly. The diet and exercise interventions in the Dietary Approaches to Stop Hypertension (DASH) trial have both been shown to lower blood pressure.^10,11

Restricting sodium intake has been shown to lower blood pressure more in older adults than in younger adults. In the DASH trial,¹² systolic blood pressure decreased by 8.1 mm Hg with sodium restriction in hypertensive patients age 55 to 76 years, compared with 4.8 mm Hg for adults aged 23 to 41 years. In the Trial of Nonpharmacologic Interventions in the Elderly (TONE),¹³ in people ages 60 to 80 who were randomized to reduce their salt intake, urinary sodium excretion was 40 mmol/day lower and blood pressure was 4.3/2.0 mm Hg lower than in a group that received usual care. Accordingly, reducing salt intake is particularly valuable for blood pressure management in the salt-sensitive elderly.¹⁴

Drug therapy

The hypertension pandemic has driven extensive pharmaceutical research, and new drugs continue to be introduced. The major classes of drugs commonly used for treating hypertension are diuretics, calcium channel blockers, and renin-angiotensin system blockers. Each class has specific benefits and adverse-effect profiles.

It is appropriate to start antihypertensive drug therapy with the lowest dose and to monitor for adverse effects, including orthostatic hypotension. The choice of drug should be guided by the patient’s comorbid conditions (Table 1) and the other drugs the patient is taking.¹⁵ If the blood pressure response is inadequate, a second drug from a different class should be added. In the same manner, a third drug from a different class should be added if the blood pressure remains outside the optimal range on two drugs.

The average elderly American is on more than six medications.¹⁶ Some of these are for high blood pressure, but others interact with antihypertensive drugs (Table 2), and some, including nonsteroidal anti-inflammatory drugs (NSAIDs) and steroids, directly affect blood pressure. Therefore, the drug regimen of an elderly hypertensive patient should be reviewed carefully at every visit. The Screening Tool of Older Person’s Prescriptions (STOPP), a list of 65 rules relating to the most common and most potentially dangerous instances of inappropriate prescribing and overprescribing in the elderly,¹⁷ has been found to be a reliable tool in this regard, with a kappa-coefficient of 0.75. Together with the Screening Tool to Alert Doctors to Right [ie, Appropriate, Indicated] Treatment (START),¹⁷ which lists 22 evidence-based prescribing indicators for common conditions in the elderly, these criteria provide clinicians with an easy screening tool to combat polypharmacy.

Given the multitude of factors that go into deciding on a specific management strategy in the elderly, it is not possible to discuss individualized care in all patients in the scope of one paper. Below, we present several case scenarios that internists commonly encounter, and suggest ways to approach each.

CASE 1: SECONDARY HYPERTENSION

A 69-year-old obese man who has hypertension of recent onset, long-standing gastroesophageal reflux disease, and benign prostatic hypertrophy comes to your office, accompanied by his wife. He has never had hypertension before. His body mass index is 34 kg/m². On physical examination, his blood pressure is 180/112 mm Hg.

We start with this case to emphasize the importance of considering causes of secondary hypertension in all patients with the disease (Table 3).¹⁸ Further workup should be pursued in those who appear to have “inappropriate” hypertension. This could present as refractory hypertension, abrupt-onset hypertension, hypertension that is first diagnosed before age 20 or after age 60, or loss of control over previously well-controlled blood pressure. Secondary hypertension must always be considered when the history or physical examination suggests a possible cause.

Renal artery stenosis increases in incidence with age. Its prevalence is reported to be as high as 50% in elderly patients with other signs of atherosclerosis such as widespread peripheral artery disease.¹⁹

Obstructive sleep apnea also commonly coexists with hypertension and its prevalence also increases with age. In addition, elderly patients with obstructive sleep apnea have a higher incidence of cardiovascular complications, including hypertension, than middle-aged people.²⁰ Numerous studies have found that the severity of obstructive sleep apnea corresponds with the likelihood of systemic hypertension.^21–23 Randomized trials and meta-analyses have also concluded that effective treatment with continuous positive airway pressure reduces systemic blood pressure,^24–27 although by less than antihypertensive medications do.

A causal relationship between obstructive sleep apnea and hypertension has not been established with certainty. It is recommended, however, that patients with resistant hypertension be screened for obstructive sleep apnea as a possible cause of their disease.

Other causes of secondary hypertension to keep in mind when evaluating patients who have inappropriate hypertension include thyroid disorders, alcohol and tobacco use, and chronic steroid or NSAID use. Pheochromocytoma and adrenal adenoma, though possible, are less prevalent in the elderly.

Case continued

Physical examination in the above patient revealed an epigastric systolic-diastolic bruit, a sign that, although not sensitive, is highly specific for renal artery stenosis, raising the suspicion of this condition. Duplex ultrasonography of the renal arteries confirmed this suspicion. The patient underwent angiography and revascularization, resulting in a distinct fall in, but not normalization of, his blood pressure.

Though we do not intend to detail the diagnostic approaches and treatments for the various causes of secondary hypertension, we need to briefly mention those for renal artery stenosis.

According to the 2006 ACC/AHA guidelines on peripheral artery disease,²⁸ testing for renal artery stenosis is indicated only if a subsequent corrective procedure is a viable option.

Renal arteriography remains the gold standard for diagnosing renal artery stenosis. However, noninvasive imaging has largely replaced it.

Duplex Doppler ultrasonography, compared with angiography, has a sensitivity of 84% to 98% depending on operator experience, and a specificity of 62% to 99% for detecting renal artery stenosis.²⁹ Some of its limiting factors are the time needed to do the study, its steep learning curve and operator-dependence, and interference with the results by body fat and intestinal gas.

Computed tomographic angiography has a sensitivity of over 90% for detecting renal artery stenosis, and its specificity has been shown to be as high as 99% in certain studies.²⁹ Use of contrast can be a limiting factor in some clinical settings.

Magnetic resonance angiography also offers a sensitivity of 90% to 100% and specificities of 76% to 94% for detecting renal artery stenosis.²⁹ On the other hand, it is costly, and the gadolinium contrast solution used is nephrotoxic, though not as toxic as the contrast used in computed tomographic angiography.

As previously stated, these imaging studies should be used only if corrective measures will be undertaken if clinically significant renal artery stenosis is found. Even in such cases, revascularization may not be curative in all cases. Its effectiveness has been compared with that of medical management alone in a number of studies.^30,31 A meta-analysis³² of six key trials involving more than 1,200 patients showed no difference in systolic and diastolic blood pressures and other clinical outcomes, including all-cause mortality, between the two treatment groups over a 29-month follow-up period.

Hence, although we advise that causes of secondary hypertension be considered in cases of inappropriate hypertension, aggressive management must be pursued on a case-by-case basis.

CASE 2: DRUG ADVERSE EFFECTS

A 75-year-old Hispanic woman with a history of treated breast cancer was recently diagnosed with hypertension. Her blood pressure is controlled on amlodipine (Norvasc) 10 mg daily, and her blood pressure today is 128/80 mm Hg. Her only complaint during this office visit is some swelling of her ankles.

Edema and dihydropyridine calcium channel blockers

Like all drugs, antihypertensive medications come with their own set of adverse effects. These are more common as people age—hence the importance of identifying and effectively managing them in the elderly population.

Calcium channel blockers, especially the dihydropyridines—ie, nifedipine (Adalat), amlodipine, felodipine (Plendil), and isradipine (DynaCirc)—are known to cause peripheral vasodilation. Peripheral edema is a common dose-related effect in people on these drugs. In one study, median leg weight increased by 80 g after amlodipine 5 mg was given for 4 weeks, and by another 68 g on a 10-mg dose.³³

Ankle swelling, encountered more in women, can be very bothersome. The swelling is related to hyperfiltration of fluid into the interstitial space secondary to intracapillary hypertension. Calcium channel blockers predominantly cause arteriolar dilation by paralyzing the precapillary sphincter, thereby elevating intracapillary pressure.

Traditionally, physicians have lowered the dose of the calcium channel blocker, switched to another drug, or added a diuretic to alleviate the swelling. However, giving a diuretic for edema induced by a calcium channel blocker may not relieve the edema.³⁴

Peripheral edema is much less encountered when a calcium channel blocker is given with an inhibitor of the renin-angiotensin system.³⁵ A meta-analysis concluded that the incidence of peripheral edema was lowered by 38% with such a combination. The same study found angiotensin-converting enzyme (ACE) inhibitors significantly more efficacious for this effect than angiotensin receptor blockers (ARBs).³⁵

ACE inhibitors and ARBs are known to cause venodilation, thereby lowering intra-capillary pressure. It is probable that this effect helps remove the extra fluid sequestered in the capillary bed by the arteriolar dilation from the calcium channel blocker.

Pedal edema associated with use of a calcium channel blocker occurs much more commonly in the elderly than in the young. It is clearly dose-dependent, and the incidence peaks after 6 months of therapy. In the patient described above, adding a low dose of an ACE inhibitor or an ARB (if the patient is ACE inhibitor-intolerant) should relieve the swelling.

Hyponatremia and diuretics

Electrolyte imbalances are another common problem encountered in the elderly. Even though for years attention has been directed to the potassium level, hyponatremia has been equally associated with adverse effects in the elderly, such as an increased risk of fractures as shown in the Rotterdam study.³⁶

In 180 hypertensive inpatients, mean age 76.4, Sharabi et al³⁷ found the incidence of hyponatremia to be three times higher in women than in men (odds ratio 3.10, 95% confidence interval 2.07 to 4.67). Patients were 10 times more likely to be affected after age 65 and 14 times more likely after age 75. Most of the patients affected (74.5%) used a thiazide-type diuretic. Even though in many of the patients diuretics were used for more than 1 year before hyponatremia developed, susceptible patients—such as the frail elderly—can develop severe hyponatremia within days of starting to use a thiazide.³⁸

Severe hyponatremia is potentially life-threatening. Most cases are caused by thiazide rather than loop diuretics.³⁸ Thiazides inhibit electrolyte transport at the cortical diluting sites. As they decrease the glomerular filtration rate acutely, they increase proximal water reabsorption (making the plasma hypotonic), reducing water delivery distally. Loop diuretics, on the other hand, have their main effect at the thick ascending limb, reducing the osmolality at the medullary interstitium and not affecting proximal water reabsorption. Additionally, loop diuretics have a shorter half-life than thiazides, which makes hyponatremia more likely to happen with thiazides.

In patients who develop hyponatremia secondary to diuretic use, appropriate treatment includes stopping the medication, restricting water intake, and repleting electrolyte stores.³⁸ As with any cause of chronic hyponatremia, correction must be cautiously monitored and not hastily done.

Therefore, we advise adding a thiazide diuric with caution in the elderly, and we advise avoiding thiazides in patients with high water or alcohol intake.

CASE 3: DEMENTIA AND HYPERTENSION

A 74-year-old man with long-standing hypertension, gout, and chronic obstructive pulmonary disease was recently diagnosed with Alzheimer dementia. He takes enalapril (Vasotec) 10 mg daily for his blood pressure. His blood pressure is 130/78 mm Hg.

Dementia is one of the most important and common neurologic disorders in the elderly. With the rise in average life expectancy, its magnitude has grown to cause a substantial emotional and economic burden on society and health care.

Midlife hypertension has been demonstrated to be an important modifiable risk factor for late-life cognitive decline,³⁹ mild cognitive impairment,⁴⁰ and dementia of all causes.⁴¹ It has been suggested that hypertension might be part of the pathogenesis of dementia, and targeting high blood pressure might prevent its onset.

Moreover, a significant reduction in both Alzheimer and vascular dementia was demonstrated (risk reduction 55%) with the use of a long-acting dihydropyridine calcium channel blocker (nitrendipine) in the Syst-Eur study.⁴² However, data from studies such as Systolic Hypertension in the Elderly Program (SHEP) and the HYVET substudy of cognitive function assessement⁴³ showed no difference in dementia between placebo and active therapy with chlorthalidone (Hygroton) (in SHEP) or indapamide (Lozol) (in the HYVET substudy).

Disorders of calcium homeostasis are associated with the brain’s aging process. Probably, the neuroprotective effect of nitrendipine seen in Syst-Eur was due to its ability to affect this process, independent of its blood pressure-lowering effect.

In another prospective study, people over 60 years of age who complained of subjective memory loss showed a significant and positive association between memory scores and the use of calcium channel blockers (+0.14 ± 0.09 in users vs −0.12 ± 0.06 in nonusers; P = .016) independently of age, sex, white matter hyperintensities, and carotid wall cross-sectional area, all of which were associated with worse memory scores.⁴⁴

Drugs that block the renin-angiotensin system have also been proposed to delay the onset and slow the progression of dementia.⁴⁵ A small randomized, controlled trial suggested that centrally active ACE inhibitors—those that cross the blood-brain barrier, such as captopril (Capoten), lisinopril (Prinivil), ramipril (Altace), and fosinopril (Monopril)—slow cognitive decline in Alzheimer dementia more than non-centrally active ACE inhibitors or calcium channel blockers.⁴⁶

Sink et al⁴⁷ examined data from participants in the Cognition Substudy of the Cardiovascular Health Study⁴⁸ on the effect of ACE inhibitors on cognitive decline. ACE inhibitors, as a class, showed no benefit in reducing the risk of dementia compared with other antihypertensive drug classes. However, centrally active ACE inhibitors, compared with other medications, were associated with a 65% reduction in cognitive decline per year of drug exposure (P = .01). Non-centrally active ACE inhibitors worsened cognitive decline.

It appears that the brain’s renin-angiotensin system plays a role in the pathogenesis of dementia. Indeed, ACE has been shown to degrade amyloid-beta protein, and its level was increased in brain tissue of Alzheimer patients postmortem.⁴⁹

The relationship between blood pressure and cognitive function appears to be curvilinear, so that low blood pressure in late life is also associated with dementia and Alzheimer dementia.⁵⁰ In 5,816 patients age 65 and older, Morris et al⁵¹ calculated the percentile scores of four cognitive tests according to the level of blood pressure. Patients with systolic blood pressures of 100 mm Hg, 120 mm Hg, and 180 mm Hg scored lower on the Mini Mental State Examination than those in the 140 to 160 mm Hg range. Patients with diastolic blood pressures between 80 and 90 mm Hg appeared to have the best cognitive function. This further emphasizes that blood pressure control must be pursued in the very elderly, albeit less aggressively. The MIND substudy of the SPRINT trial⁹ is likely to shed more light on this relationship.

When needed for optimal blood pressure control in a hypertensive patient at risk of dementia, a calcium channel blocker of the dihydropyridine type or a centrally active ACE inhibitor, or both, is preferable.

CASE 4: LABILE HYPERTENSION

A 74-year-old man with hypertension and diabetes mellitus comes to see you in the office. On physical examination, his blood pressure is 175/110 mm Hg. His blood pressure during his last visit 3 months ago was 120/75. He brings a log with him that shows random fluctuations in his blood pressure readings. He takes hydrochlorothiazide 25 mg daily for his blood pressure.

Hypertension in some patients continuously fluctuates between low and high levels. A study in Canada found that up to 15% of all adult hypertensive patients might have labile hypertension.⁵² In the presence of a normal average blood pressure, visit-to-visit blood pressure variability is usually considered a trivial matter. However, some but not all studies have shown that such visit-to-visit variability in blood pressure is an independent predictor of future cardiovascular events in hypertensive patients, independent of the mean systolic blood pressure.^52–54

Blood pressure fluctuates from heartbeat to heartbeat, from morning to night, from winter to summer, and from sitting to standing, and it is prone to increase with exertion, stress, and other factors. But excessive fluctuations in the elderly are most likely the result of excessive stiffness of the arterial tree and a decrease in the windkessel (cushioning) function of the aorta. As a consequence, even small-volume changes in the intravascular system can trigger large blood pressure fluctuations.

There is some evidence that antihypertensive drug classes may differ in their effects on visit-to-visit blood pressure variability. In a 2010 study comparing the effects of different antihypertensive drugs on blood pressure variation, calcium channel blockers and non-loop diuretics were associated with less variation in systolic blood pressure, and calcium channel blockers reduced it the most.⁵⁵

In the patient described above, switching to a low-dose calcium channel blocker with a thorough follow-up is a reasonable plan.

CASE 5: ORTHOSTATIC HYPOTENSION

A 73-year-old woman with long-standing hypertension complains of some dizziness, especially when getting out of bed in the morning. On physical examination, her blood pressure is 134/100 mm Hg sitting and 115/90 standing. She takes amlodipine 10 mg daily, enalapril 10 mg daily, and chlorthalidone 25 mg daily. Chlorthalidone had been added on her last visit 1 month before.

As a result of the increase in the number of elderly patients with hypertension and guidelines recommending better control in this age group, the number of elderly patients on anti-hypertensive drugs has risen significantly. At the same time, the elderly have increasingly presented with adverse effects of treatment.

Orthostatic hypotension is a drop in systolic pressure of 20 mm Hg or more or a drop in diastolic pressure of 10 mm Hg or more on standing, with or without symptoms. These are caused by cerebral hypoperfusion and include dizziness, lightheadedness, generalized weakness, visual blurring, and, in severe cases, syncope.

Alpha-blockers and nitrates have been most commonly implicated in causing orthostatic hypotension, due to venous pooling. Clearly, not all antihypertensive drugs are equal with regard to their venodilatory effects. Thiazide diuretics, by causing fluid volume depletion, and beta-blockers, by interfering with compensatory cardioacceleration with upright posture, are also commonly involved in causing an excessive blood pressure drop with standing.

Systolic orthostatic hypotension has been shown to be a significant and independent predictor of cardiovascular morbidity and death.⁵⁶ Moreover, syncope and subsequent falls are an important cause of injury and death in the elderly.⁵⁷ The clinical combination of hypertension and orthostatic hypotension is, therefore, especially challenging. A compromise between accepting a higher cardiovascular risk at either end of the spectrum with an added higher risk for fall at the lower end has to be made.

To prevent orthostatic hypotension in the elderly, it is important to avoid prescribing high-risk drugs. When starting antihypertensive therapy, a low dose should be used, and the dose should be titrated upward very slowly and cautiously. If orthostatic hypotension is suggested by the history or by the orthostatic test, which is warranted in all elderly hypertensive patients before starting or significantly altering therapy, the potential culprit drug should be withdrawn and the patient reassessed. Improved hydration, elevating the head of the bed, and taking the antihypertensive drug at night are ways to improve symptoms, but these remain largely unproven.

In this patient, the newly added chlorthalidone was stopped, and her symptoms improved.

PSEUDOHYPERTENSION

Since hypertension is defined by a numerical value, it is prudent that this value be accurate. Treating a falsely high reading or leaving a falsely low reading untreated will predispose the elderly patient to increased risk either way. One rare condition in the elderly that can give a falsely high blood pressure reading is pseudohypertension.

Pseudohypertension is a condition in which indirect blood pressure measured by the cuff method overestimates the true intra-arterial blood pressure due to marked underlying arteriosclerosis. The Osler maneuver can be used to differentiate true hypertension from pseudohypertension.⁵⁸ This is performed by palpating the pulseless radial or brachial artery distal to the inflated cuff. If the artery is palpable despite being pulseless, the patient is said to be “Osler-positive” and likely has pseudohypertension.⁵⁸

Pseudohypertension should be suspected if the patient has orthostatic hypotension despite normal blood pressure sitting and standing. Also, elevated blood pressure without appropriate target organ disease should raise the suspicion of pseudohypertension. Apart from the Osler maneuver, measuring the intraarterial pressure can confirm this diagnosis.

The management of hypertension has advanced significantly in the last few decades. But the race for more effective means to control this epidemic and its associated complications is far from won. A high percentage of patients in the United States have hypertension that is uncontrolled. Most of these belong to the most rapidly growing demographic group in the United States, ie, the elderly.

It is estimated that more than 70% of medical practice will be directed to geriatric needs in the coming years. It is therefore very important for clinicians to be comfortable with managing hypertension in the elderly.

A GROWING PROBLEM IN AN AGING POPULATION

Between 1980 and 2009, the US population age 65 and older increased from 25.6 million to 39.6 million, of which 42% are men and 58% women.¹ This number is expected to reach 75 million by the year 2040. People over 85 years of age are the fastest growing subset of the US population.² As many as 50% of people who were born recently in countries such as the United States, the United Kingdom, France, Denmark, and Japan will live to celebrate their 100th birthday.³

According to the Framingham Heart Study, by age 60 approximately 60% of the population develops hypertension, and by 70 years about 65% of men and about 75% of women have the disease. In the same study, 90% of those who were normotensive at age 55 went on to develop hypertension. The elderly also are more likely to suffer from the complications of hypertension and are more likely to have uncontrolled disease.

Compared with younger patients with similar blood pressure, elderly hypertensive patients have lower cardiac output, higher peripheral resistance, wider pulse pressure, lower intravascular volume, and lower renal blood flow.⁴ These age-related pathophysiologic differences must be considered when treating antihypertension in the elderly.

IS TREATING THE ELDERLY BENEFICIAL?

Most elderly hypertensive patients have multiple comorbidities, which tremendously affect the management of their hypertension. They are also more likely than younger patients to have resistant hypertension and to need multiple drugs to control their blood pressure. In the process, these frail patients are exposed to a host of drug-related adverse effects. Thus, it is relevant to question the net benefit of treatment in this age group.

Many studies have indeed shown that treating hypertension reduces the risk of stroke and other adverse cardiovascular events. A decade ago, Staessen et al,⁵ in a meta-analysis of more than 15,000 patients between ages 62 and 76, showed that treating isolated systolic hypertension substantially reduced morbidity and mortality rates. Moreover, a 2011 meta-analysis of randomized controlled trials in hypertensive patients age 75 and over also concluded that treatment reduced cardiovascular morbidity and mortality rates and the incidence of heart failure, even though the total mortality rate was not affected.⁶

Opinion on treating the very elderly (≥ 80 years of age) was divided until the results of the Hypertension in the Very Elderly trial (HYVET)⁷ came out in 2008. This study documented major benefits of treatment in the very elderly age group as well.

The consensus, therefore, is that it is appropriate, even imperative, to treat elderly hypertensive patients (with some cautions—see the sections that follow).

GOAL OF TREATMENT IN THE ELDERLY

Targets for blood pressure management have been based primarily on observational data in middle-aged patients. There is no such thing as an ideal blood pressure that has been derived from randomized controlled trials for any population, let alone the elderly. The generally recommended blood pressure goal of 140/90 mm Hg for elderly hypertensive patients is based on expert opinion.

Moreover, it is unclear if the same target should apply to octogenarians. According to a 2011 American College of Cardiology/American Heart Association (ACC/AHA) expert consensus report,⁸ an achieved systolic blood pressure of 140 to 145 mm Hg, if tolerated, can be acceptable in this age group.

An orthostatic decline in blood pressure accompanies advanced age and is an inevitable adverse effect of some antihypertensive drugs. Accordingly, systolic blood pressure lower than 130 and diastolic blood pressure lower than 70 mm Hg are best avoided in octogenarians.⁸ Therefore, when hypertension is complicated by coexisting conditions that require a specific blood pressure goal, it would seem reasonable to not pursue the lower target as aggressively in octogenarians as in elderly patients under age 80.

Having stated the limitations in the quality of data at hand—largely observational—it is relevant to mention the Systolic Blood Pressure Intervention trial (SPRINT).⁹ This ongoing randomized, multicenter trial, launched by the National Institutes of Health, is assessing whether maintaining blood pressure levels lower than current recommendations further reduces the risk of cardiovascular and kidney diseases or, in the SPRINT-MIND substudy, of age-related cognitive decline, regardless of the type of antihypertensive drug taken. Initially planning to enroll close to 10,000 participants over the age of 55 without specifying any agegroup ranges, the investigators later decided to conduct a substudy called SPRINT Senior that will enroll about 1,750 participants over the age of 75 to determine whether a lower blood pressure range will have the same beneficial effects in older adults.

Given the limitations in the quality and applicability of published data (coming from small, nonrandomized studies with no long-term follow-up), SPRINT is expected to provide the evidence needed to support standard vs aggressive hypertension control among the elderly. The trial is projected to run until late 2018.

MANAGEMENT APPROACH IN THE ELDERLY

Blood pressure should be recorded in both arms before a diagnosis is made. In a number of patients, particularly the elderly, there are significant differences in blood pressure readings between the two arms. The higher reading should be relied on and the corresponding arm used for subsequent measurements.

Lifestyle interventions

Similar to the approach in younger patients with hypertension, lifestyle interventions are the first step to managing high blood pressure in the elderly. The diet and exercise interventions in the Dietary Approaches to Stop Hypertension (DASH) trial have both been shown to lower blood pressure.^10,11

Restricting sodium intake has been shown to lower blood pressure more in older adults than in younger adults. In the DASH trial,¹² systolic blood pressure decreased by 8.1 mm Hg with sodium restriction in hypertensive patients age 55 to 76 years, compared with 4.8 mm Hg for adults aged 23 to 41 years. In the Trial of Nonpharmacologic Interventions in the Elderly (TONE),¹³ in people ages 60 to 80 who were randomized to reduce their salt intake, urinary sodium excretion was 40 mmol/day lower and blood pressure was 4.3/2.0 mm Hg lower than in a group that received usual care. Accordingly, reducing salt intake is particularly valuable for blood pressure management in the salt-sensitive elderly.¹⁴

Drug therapy

The hypertension pandemic has driven extensive pharmaceutical research, and new drugs continue to be introduced. The major classes of drugs commonly used for treating hypertension are diuretics, calcium channel blockers, and renin-angiotensin system blockers. Each class has specific benefits and adverse-effect profiles.

It is appropriate to start antihypertensive drug therapy with the lowest dose and to monitor for adverse effects, including orthostatic hypotension. The choice of drug should be guided by the patient’s comorbid conditions (Table 1) and the other drugs the patient is taking.¹⁵ If the blood pressure response is inadequate, a second drug from a different class should be added. In the same manner, a third drug from a different class should be added if the blood pressure remains outside the optimal range on two drugs.

The average elderly American is on more than six medications.¹⁶ Some of these are for high blood pressure, but others interact with antihypertensive drugs (Table 2), and some, including nonsteroidal anti-inflammatory drugs (NSAIDs) and steroids, directly affect blood pressure. Therefore, the drug regimen of an elderly hypertensive patient should be reviewed carefully at every visit. The Screening Tool of Older Person’s Prescriptions (STOPP), a list of 65 rules relating to the most common and most potentially dangerous instances of inappropriate prescribing and overprescribing in the elderly,¹⁷ has been found to be a reliable tool in this regard, with a kappa-coefficient of 0.75. Together with the Screening Tool to Alert Doctors to Right [ie, Appropriate, Indicated] Treatment (START),¹⁷ which lists 22 evidence-based prescribing indicators for common conditions in the elderly, these criteria provide clinicians with an easy screening tool to combat polypharmacy.

Given the multitude of factors that go into deciding on a specific management strategy in the elderly, it is not possible to discuss individualized care in all patients in the scope of one paper. Below, we present several case scenarios that internists commonly encounter, and suggest ways to approach each.

CASE 1: SECONDARY HYPERTENSION

A 69-year-old obese man who has hypertension of recent onset, long-standing gastroesophageal reflux disease, and benign prostatic hypertrophy comes to your office, accompanied by his wife. He has never had hypertension before. His body mass index is 34 kg/m². On physical examination, his blood pressure is 180/112 mm Hg.

We start with this case to emphasize the importance of considering causes of secondary hypertension in all patients with the disease (Table 3).¹⁸ Further workup should be pursued in those who appear to have “inappropriate” hypertension. This could present as refractory hypertension, abrupt-onset hypertension, hypertension that is first diagnosed before age 20 or after age 60, or loss of control over previously well-controlled blood pressure. Secondary hypertension must always be considered when the history or physical examination suggests a possible cause.

Renal artery stenosis increases in incidence with age. Its prevalence is reported to be as high as 50% in elderly patients with other signs of atherosclerosis such as widespread peripheral artery disease.¹⁹

Obstructive sleep apnea also commonly coexists with hypertension and its prevalence also increases with age. In addition, elderly patients with obstructive sleep apnea have a higher incidence of cardiovascular complications, including hypertension, than middle-aged people.²⁰ Numerous studies have found that the severity of obstructive sleep apnea corresponds with the likelihood of systemic hypertension.^21–23 Randomized trials and meta-analyses have also concluded that effective treatment with continuous positive airway pressure reduces systemic blood pressure,^24–27 although by less than antihypertensive medications do.

A causal relationship between obstructive sleep apnea and hypertension has not been established with certainty. It is recommended, however, that patients with resistant hypertension be screened for obstructive sleep apnea as a possible cause of their disease.

Other causes of secondary hypertension to keep in mind when evaluating patients who have inappropriate hypertension include thyroid disorders, alcohol and tobacco use, and chronic steroid or NSAID use. Pheochromocytoma and adrenal adenoma, though possible, are less prevalent in the elderly.

Case continued

Physical examination in the above patient revealed an epigastric systolic-diastolic bruit, a sign that, although not sensitive, is highly specific for renal artery stenosis, raising the suspicion of this condition. Duplex ultrasonography of the renal arteries confirmed this suspicion. The patient underwent angiography and revascularization, resulting in a distinct fall in, but not normalization of, his blood pressure.

Though we do not intend to detail the diagnostic approaches and treatments for the various causes of secondary hypertension, we need to briefly mention those for renal artery stenosis.

According to the 2006 ACC/AHA guidelines on peripheral artery disease,²⁸ testing for renal artery stenosis is indicated only if a subsequent corrective procedure is a viable option.

Renal arteriography remains the gold standard for diagnosing renal artery stenosis. However, noninvasive imaging has largely replaced it.

Duplex Doppler ultrasonography, compared with angiography, has a sensitivity of 84% to 98% depending on operator experience, and a specificity of 62% to 99% for detecting renal artery stenosis.²⁹ Some of its limiting factors are the time needed to do the study, its steep learning curve and operator-dependence, and interference with the results by body fat and intestinal gas.

Computed tomographic angiography has a sensitivity of over 90% for detecting renal artery stenosis, and its specificity has been shown to be as high as 99% in certain studies.²⁹ Use of contrast can be a limiting factor in some clinical settings.

Magnetic resonance angiography also offers a sensitivity of 90% to 100% and specificities of 76% to 94% for detecting renal artery stenosis.²⁹ On the other hand, it is costly, and the gadolinium contrast solution used is nephrotoxic, though not as toxic as the contrast used in computed tomographic angiography.

As previously stated, these imaging studies should be used only if corrective measures will be undertaken if clinically significant renal artery stenosis is found. Even in such cases, revascularization may not be curative in all cases. Its effectiveness has been compared with that of medical management alone in a number of studies.^30,31 A meta-analysis³² of six key trials involving more than 1,200 patients showed no difference in systolic and diastolic blood pressures and other clinical outcomes, including all-cause mortality, between the two treatment groups over a 29-month follow-up period.

Hence, although we advise that causes of secondary hypertension be considered in cases of inappropriate hypertension, aggressive management must be pursued on a case-by-case basis.

CASE 2: DRUG ADVERSE EFFECTS

A 75-year-old Hispanic woman with a history of treated breast cancer was recently diagnosed with hypertension. Her blood pressure is controlled on amlodipine (Norvasc) 10 mg daily, and her blood pressure today is 128/80 mm Hg. Her only complaint during this office visit is some swelling of her ankles.

Edema and dihydropyridine calcium channel blockers

Like all drugs, antihypertensive medications come with their own set of adverse effects. These are more common as people age—hence the importance of identifying and effectively managing them in the elderly population.

Calcium channel blockers, especially the dihydropyridines—ie, nifedipine (Adalat), amlodipine, felodipine (Plendil), and isradipine (DynaCirc)—are known to cause peripheral vasodilation. Peripheral edema is a common dose-related effect in people on these drugs. In one study, median leg weight increased by 80 g after amlodipine 5 mg was given for 4 weeks, and by another 68 g on a 10-mg dose.³³

Ankle swelling, encountered more in women, can be very bothersome. The swelling is related to hyperfiltration of fluid into the interstitial space secondary to intracapillary hypertension. Calcium channel blockers predominantly cause arteriolar dilation by paralyzing the precapillary sphincter, thereby elevating intracapillary pressure.

Traditionally, physicians have lowered the dose of the calcium channel blocker, switched to another drug, or added a diuretic to alleviate the swelling. However, giving a diuretic for edema induced by a calcium channel blocker may not relieve the edema.³⁴

Peripheral edema is much less encountered when a calcium channel blocker is given with an inhibitor of the renin-angiotensin system.³⁵ A meta-analysis concluded that the incidence of peripheral edema was lowered by 38% with such a combination. The same study found angiotensin-converting enzyme (ACE) inhibitors significantly more efficacious for this effect than angiotensin receptor blockers (ARBs).³⁵

ACE inhibitors and ARBs are known to cause venodilation, thereby lowering intra-capillary pressure. It is probable that this effect helps remove the extra fluid sequestered in the capillary bed by the arteriolar dilation from the calcium channel blocker.

Pedal edema associated with use of a calcium channel blocker occurs much more commonly in the elderly than in the young. It is clearly dose-dependent, and the incidence peaks after 6 months of therapy. In the patient described above, adding a low dose of an ACE inhibitor or an ARB (if the patient is ACE inhibitor-intolerant) should relieve the swelling.

Hyponatremia and diuretics

Electrolyte imbalances are another common problem encountered in the elderly. Even though for years attention has been directed to the potassium level, hyponatremia has been equally associated with adverse effects in the elderly, such as an increased risk of fractures as shown in the Rotterdam study.³⁶

In 180 hypertensive inpatients, mean age 76.4, Sharabi et al³⁷ found the incidence of hyponatremia to be three times higher in women than in men (odds ratio 3.10, 95% confidence interval 2.07 to 4.67). Patients were 10 times more likely to be affected after age 65 and 14 times more likely after age 75. Most of the patients affected (74.5%) used a thiazide-type diuretic. Even though in many of the patients diuretics were used for more than 1 year before hyponatremia developed, susceptible patients—such as the frail elderly—can develop severe hyponatremia within days of starting to use a thiazide.³⁸

Severe hyponatremia is potentially life-threatening. Most cases are caused by thiazide rather than loop diuretics.³⁸ Thiazides inhibit electrolyte transport at the cortical diluting sites. As they decrease the glomerular filtration rate acutely, they increase proximal water reabsorption (making the plasma hypotonic), reducing water delivery distally. Loop diuretics, on the other hand, have their main effect at the thick ascending limb, reducing the osmolality at the medullary interstitium and not affecting proximal water reabsorption. Additionally, loop diuretics have a shorter half-life than thiazides, which makes hyponatremia more likely to happen with thiazides.

In patients who develop hyponatremia secondary to diuretic use, appropriate treatment includes stopping the medication, restricting water intake, and repleting electrolyte stores.³⁸ As with any cause of chronic hyponatremia, correction must be cautiously monitored and not hastily done.

Therefore, we advise adding a thiazide diuric with caution in the elderly, and we advise avoiding thiazides in patients with high water or alcohol intake.

CASE 3: DEMENTIA AND HYPERTENSION

A 74-year-old man with long-standing hypertension, gout, and chronic obstructive pulmonary disease was recently diagnosed with Alzheimer dementia. He takes enalapril (Vasotec) 10 mg daily for his blood pressure. His blood pressure is 130/78 mm Hg.

Dementia is one of the most important and common neurologic disorders in the elderly. With the rise in average life expectancy, its magnitude has grown to cause a substantial emotional and economic burden on society and health care.

Midlife hypertension has been demonstrated to be an important modifiable risk factor for late-life cognitive decline,³⁹ mild cognitive impairment,⁴⁰ and dementia of all causes.⁴¹ It has been suggested that hypertension might be part of the pathogenesis of dementia, and targeting high blood pressure might prevent its onset.

Moreover, a significant reduction in both Alzheimer and vascular dementia was demonstrated (risk reduction 55%) with the use of a long-acting dihydropyridine calcium channel blocker (nitrendipine) in the Syst-Eur study.⁴² However, data from studies such as Systolic Hypertension in the Elderly Program (SHEP) and the HYVET substudy of cognitive function assessement⁴³ showed no difference in dementia between placebo and active therapy with chlorthalidone (Hygroton) (in SHEP) or indapamide (Lozol) (in the HYVET substudy).

Disorders of calcium homeostasis are associated with the brain’s aging process. Probably, the neuroprotective effect of nitrendipine seen in Syst-Eur was due to its ability to affect this process, independent of its blood pressure-lowering effect.

In another prospective study, people over 60 years of age who complained of subjective memory loss showed a significant and positive association between memory scores and the use of calcium channel blockers (+0.14 ± 0.09 in users vs −0.12 ± 0.06 in nonusers; P = .016) independently of age, sex, white matter hyperintensities, and carotid wall cross-sectional area, all of which were associated with worse memory scores.⁴⁴

Drugs that block the renin-angiotensin system have also been proposed to delay the onset and slow the progression of dementia.⁴⁵ A small randomized, controlled trial suggested that centrally active ACE inhibitors—those that cross the blood-brain barrier, such as captopril (Capoten), lisinopril (Prinivil), ramipril (Altace), and fosinopril (Monopril)—slow cognitive decline in Alzheimer dementia more than non-centrally active ACE inhibitors or calcium channel blockers.⁴⁶

Sink et al⁴⁷ examined data from participants in the Cognition Substudy of the Cardiovascular Health Study⁴⁸ on the effect of ACE inhibitors on cognitive decline. ACE inhibitors, as a class, showed no benefit in reducing the risk of dementia compared with other antihypertensive drug classes. However, centrally active ACE inhibitors, compared with other medications, were associated with a 65% reduction in cognitive decline per year of drug exposure (P = .01). Non-centrally active ACE inhibitors worsened cognitive decline.

It appears that the brain’s renin-angiotensin system plays a role in the pathogenesis of dementia. Indeed, ACE has been shown to degrade amyloid-beta protein, and its level was increased in brain tissue of Alzheimer patients postmortem.⁴⁹

The relationship between blood pressure and cognitive function appears to be curvilinear, so that low blood pressure in late life is also associated with dementia and Alzheimer dementia.⁵⁰ In 5,816 patients age 65 and older, Morris et al⁵¹ calculated the percentile scores of four cognitive tests according to the level of blood pressure. Patients with systolic blood pressures of 100 mm Hg, 120 mm Hg, and 180 mm Hg scored lower on the Mini Mental State Examination than those in the 140 to 160 mm Hg range. Patients with diastolic blood pressures between 80 and 90 mm Hg appeared to have the best cognitive function. This further emphasizes that blood pressure control must be pursued in the very elderly, albeit less aggressively. The MIND substudy of the SPRINT trial⁹ is likely to shed more light on this relationship.

When needed for optimal blood pressure control in a hypertensive patient at risk of dementia, a calcium channel blocker of the dihydropyridine type or a centrally active ACE inhibitor, or both, is preferable.

CASE 4: LABILE HYPERTENSION

A 74-year-old man with hypertension and diabetes mellitus comes to see you in the office. On physical examination, his blood pressure is 175/110 mm Hg. His blood pressure during his last visit 3 months ago was 120/75. He brings a log with him that shows random fluctuations in his blood pressure readings. He takes hydrochlorothiazide 25 mg daily for his blood pressure.

Hypertension in some patients continuously fluctuates between low and high levels. A study in Canada found that up to 15% of all adult hypertensive patients might have labile hypertension.⁵² In the presence of a normal average blood pressure, visit-to-visit blood pressure variability is usually considered a trivial matter. However, some but not all studies have shown that such visit-to-visit variability in blood pressure is an independent predictor of future cardiovascular events in hypertensive patients, independent of the mean systolic blood pressure.^52–54

Blood pressure fluctuates from heartbeat to heartbeat, from morning to night, from winter to summer, and from sitting to standing, and it is prone to increase with exertion, stress, and other factors. But excessive fluctuations in the elderly are most likely the result of excessive stiffness of the arterial tree and a decrease in the windkessel (cushioning) function of the aorta. As a consequence, even small-volume changes in the intravascular system can trigger large blood pressure fluctuations.

There is some evidence that antihypertensive drug classes may differ in their effects on visit-to-visit blood pressure variability. In a 2010 study comparing the effects of different antihypertensive drugs on blood pressure variation, calcium channel blockers and non-loop diuretics were associated with less variation in systolic blood pressure, and calcium channel blockers reduced it the most.⁵⁵

In the patient described above, switching to a low-dose calcium channel blocker with a thorough follow-up is a reasonable plan.

CASE 5: ORTHOSTATIC HYPOTENSION

A 73-year-old woman with long-standing hypertension complains of some dizziness, especially when getting out of bed in the morning. On physical examination, her blood pressure is 134/100 mm Hg sitting and 115/90 standing. She takes amlodipine 10 mg daily, enalapril 10 mg daily, and chlorthalidone 25 mg daily. Chlorthalidone had been added on her last visit 1 month before.

As a result of the increase in the number of elderly patients with hypertension and guidelines recommending better control in this age group, the number of elderly patients on anti-hypertensive drugs has risen significantly. At the same time, the elderly have increasingly presented with adverse effects of treatment.

Orthostatic hypotension is a drop in systolic pressure of 20 mm Hg or more or a drop in diastolic pressure of 10 mm Hg or more on standing, with or without symptoms. These are caused by cerebral hypoperfusion and include dizziness, lightheadedness, generalized weakness, visual blurring, and, in severe cases, syncope.

Alpha-blockers and nitrates have been most commonly implicated in causing orthostatic hypotension, due to venous pooling. Clearly, not all antihypertensive drugs are equal with regard to their venodilatory effects. Thiazide diuretics, by causing fluid volume depletion, and beta-blockers, by interfering with compensatory cardioacceleration with upright posture, are also commonly involved in causing an excessive blood pressure drop with standing.

Systolic orthostatic hypotension has been shown to be a significant and independent predictor of cardiovascular morbidity and death.⁵⁶ Moreover, syncope and subsequent falls are an important cause of injury and death in the elderly.⁵⁷ The clinical combination of hypertension and orthostatic hypotension is, therefore, especially challenging. A compromise between accepting a higher cardiovascular risk at either end of the spectrum with an added higher risk for fall at the lower end has to be made.

To prevent orthostatic hypotension in the elderly, it is important to avoid prescribing high-risk drugs. When starting antihypertensive therapy, a low dose should be used, and the dose should be titrated upward very slowly and cautiously. If orthostatic hypotension is suggested by the history or by the orthostatic test, which is warranted in all elderly hypertensive patients before starting or significantly altering therapy, the potential culprit drug should be withdrawn and the patient reassessed. Improved hydration, elevating the head of the bed, and taking the antihypertensive drug at night are ways to improve symptoms, but these remain largely unproven.

In this patient, the newly added chlorthalidone was stopped, and her symptoms improved.

PSEUDOHYPERTENSION

Since hypertension is defined by a numerical value, it is prudent that this value be accurate. Treating a falsely high reading or leaving a falsely low reading untreated will predispose the elderly patient to increased risk either way. One rare condition in the elderly that can give a falsely high blood pressure reading is pseudohypertension.

Pseudohypertension is a condition in which indirect blood pressure measured by the cuff method overestimates the true intra-arterial blood pressure due to marked underlying arteriosclerosis. The Osler maneuver can be used to differentiate true hypertension from pseudohypertension.⁵⁸ This is performed by palpating the pulseless radial or brachial artery distal to the inflated cuff. If the artery is palpable despite being pulseless, the patient is said to be “Osler-positive” and likely has pseudohypertension.⁵⁸

Pseudohypertension should be suspected if the patient has orthostatic hypotension despite normal blood pressure sitting and standing. Also, elevated blood pressure without appropriate target organ disease should raise the suspicion of pseudohypertension. Apart from the Osler maneuver, measuring the intraarterial pressure can confirm this diagnosis.

The management of hypertension has advanced significantly in the last few decades. But the race for more effective means to control this epidemic and its associated complications is far from won. A high percentage of patients in the United States have hypertension that is uncontrolled. Most of these belong to the most rapidly growing demographic group in the United States, ie, the elderly.

It is estimated that more than 70% of medical practice will be directed to geriatric needs in the coming years. It is therefore very important for clinicians to be comfortable with managing hypertension in the elderly.

A GROWING PROBLEM IN AN AGING POPULATION

Between 1980 and 2009, the US population age 65 and older increased from 25.6 million to 39.6 million, of which 42% are men and 58% women.¹ This number is expected to reach 75 million by the year 2040. People over 85 years of age are the fastest growing subset of the US population.² As many as 50% of people who were born recently in countries such as the United States, the United Kingdom, France, Denmark, and Japan will live to celebrate their 100th birthday.³

According to the Framingham Heart Study, by age 60 approximately 60% of the population develops hypertension, and by 70 years about 65% of men and about 75% of women have the disease. In the same study, 90% of those who were normotensive at age 55 went on to develop hypertension. The elderly also are more likely to suffer from the complications of hypertension and are more likely to have uncontrolled disease.

Compared with younger patients with similar blood pressure, elderly hypertensive patients have lower cardiac output, higher peripheral resistance, wider pulse pressure, lower intravascular volume, and lower renal blood flow.⁴ These age-related pathophysiologic differences must be considered when treating antihypertension in the elderly.

IS TREATING THE ELDERLY BENEFICIAL?

Most elderly hypertensive patients have multiple comorbidities, which tremendously affect the management of their hypertension. They are also more likely than younger patients to have resistant hypertension and to need multiple drugs to control their blood pressure. In the process, these frail patients are exposed to a host of drug-related adverse effects. Thus, it is relevant to question the net benefit of treatment in this age group.

Many studies have indeed shown that treating hypertension reduces the risk of stroke and other adverse cardiovascular events. A decade ago, Staessen et al,⁵ in a meta-analysis of more than 15,000 patients between ages 62 and 76, showed that treating isolated systolic hypertension substantially reduced morbidity and mortality rates. Moreover, a 2011 meta-analysis of randomized controlled trials in hypertensive patients age 75 and over also concluded that treatment reduced cardiovascular morbidity and mortality rates and the incidence of heart failure, even though the total mortality rate was not affected.⁶

Opinion on treating the very elderly (≥ 80 years of age) was divided until the results of the Hypertension in the Very Elderly trial (HYVET)⁷ came out in 2008. This study documented major benefits of treatment in the very elderly age group as well.

The consensus, therefore, is that it is appropriate, even imperative, to treat elderly hypertensive patients (with some cautions—see the sections that follow).

GOAL OF TREATMENT IN THE ELDERLY

Targets for blood pressure management have been based primarily on observational data in middle-aged patients. There is no such thing as an ideal blood pressure that has been derived from randomized controlled trials for any population, let alone the elderly. The generally recommended blood pressure goal of 140/90 mm Hg for elderly hypertensive patients is based on expert opinion.

Moreover, it is unclear if the same target should apply to octogenarians. According to a 2011 American College of Cardiology/American Heart Association (ACC/AHA) expert consensus report,⁸ an achieved systolic blood pressure of 140 to 145 mm Hg, if tolerated, can be acceptable in this age group.

An orthostatic decline in blood pressure accompanies advanced age and is an inevitable adverse effect of some antihypertensive drugs. Accordingly, systolic blood pressure lower than 130 and diastolic blood pressure lower than 70 mm Hg are best avoided in octogenarians.⁸ Therefore, when hypertension is complicated by coexisting conditions that require a specific blood pressure goal, it would seem reasonable to not pursue the lower target as aggressively in octogenarians as in elderly patients under age 80.

Having stated the limitations in the quality of data at hand—largely observational—it is relevant to mention the Systolic Blood Pressure Intervention trial (SPRINT).⁹ This ongoing randomized, multicenter trial, launched by the National Institutes of Health, is assessing whether maintaining blood pressure levels lower than current recommendations further reduces the risk of cardiovascular and kidney diseases or, in the SPRINT-MIND substudy, of age-related cognitive decline, regardless of the type of antihypertensive drug taken. Initially planning to enroll close to 10,000 participants over the age of 55 without specifying any agegroup ranges, the investigators later decided to conduct a substudy called SPRINT Senior that will enroll about 1,750 participants over the age of 75 to determine whether a lower blood pressure range will have the same beneficial effects in older adults.

Given the limitations in the quality and applicability of published data (coming from small, nonrandomized studies with no long-term follow-up), SPRINT is expected to provide the evidence needed to support standard vs aggressive hypertension control among the elderly. The trial is projected to run until late 2018.

MANAGEMENT APPROACH IN THE ELDERLY

Blood pressure should be recorded in both arms before a diagnosis is made. In a number of patients, particularly the elderly, there are significant differences in blood pressure readings between the two arms. The higher reading should be relied on and the corresponding arm used for subsequent measurements.

Lifestyle interventions

Similar to the approach in younger patients with hypertension, lifestyle interventions are the first step to managing high blood pressure in the elderly. The diet and exercise interventions in the Dietary Approaches to Stop Hypertension (DASH) trial have both been shown to lower blood pressure.^10,11

Restricting sodium intake has been shown to lower blood pressure more in older adults than in younger adults. In the DASH trial,¹² systolic blood pressure decreased by 8.1 mm Hg with sodium restriction in hypertensive patients age 55 to 76 years, compared with 4.8 mm Hg for adults aged 23 to 41 years. In the Trial of Nonpharmacologic Interventions in the Elderly (TONE),¹³ in people ages 60 to 80 who were randomized to reduce their salt intake, urinary sodium excretion was 40 mmol/day lower and blood pressure was 4.3/2.0 mm Hg lower than in a group that received usual care. Accordingly, reducing salt intake is particularly valuable for blood pressure management in the salt-sensitive elderly.¹⁴

Drug therapy

The hypertension pandemic has driven extensive pharmaceutical research, and new drugs continue to be introduced. The major classes of drugs commonly used for treating hypertension are diuretics, calcium channel blockers, and renin-angiotensin system blockers. Each class has specific benefits and adverse-effect profiles.

It is appropriate to start antihypertensive drug therapy with the lowest dose and to monitor for adverse effects, including orthostatic hypotension. The choice of drug should be guided by the patient’s comorbid conditions (Table 1) and the other drugs the patient is taking.¹⁵ If the blood pressure response is inadequate, a second drug from a different class should be added. In the same manner, a third drug from a different class should be added if the blood pressure remains outside the optimal range on two drugs.

The average elderly American is on more than six medications.¹⁶ Some of these are for high blood pressure, but others interact with antihypertensive drugs (Table 2), and some, including nonsteroidal anti-inflammatory drugs (NSAIDs) and steroids, directly affect blood pressure. Therefore, the drug regimen of an elderly hypertensive patient should be reviewed carefully at every visit. The Screening Tool of Older Person’s Prescriptions (STOPP), a list of 65 rules relating to the most common and most potentially dangerous instances of inappropriate prescribing and overprescribing in the elderly,¹⁷ has been found to be a reliable tool in this regard, with a kappa-coefficient of 0.75. Together with the Screening Tool to Alert Doctors to Right [ie, Appropriate, Indicated] Treatment (START),¹⁷ which lists 22 evidence-based prescribing indicators for common conditions in the elderly, these criteria provide clinicians with an easy screening tool to combat polypharmacy.

Given the multitude of factors that go into deciding on a specific management strategy in the elderly, it is not possible to discuss individualized care in all patients in the scope of one paper. Below, we present several case scenarios that internists commonly encounter, and suggest ways to approach each.

CASE 1: SECONDARY HYPERTENSION

A 69-year-old obese man who has hypertension of recent onset, long-standing gastroesophageal reflux disease, and benign prostatic hypertrophy comes to your office, accompanied by his wife. He has never had hypertension before. His body mass index is 34 kg/m². On physical examination, his blood pressure is 180/112 mm Hg.

We start with this case to emphasize the importance of considering causes of secondary hypertension in all patients with the disease (Table 3).¹⁸ Further workup should be pursued in those who appear to have “inappropriate” hypertension. This could present as refractory hypertension, abrupt-onset hypertension, hypertension that is first diagnosed before age 20 or after age 60, or loss of control over previously well-controlled blood pressure. Secondary hypertension must always be considered when the history or physical examination suggests a possible cause.

Renal artery stenosis increases in incidence with age. Its prevalence is reported to be as high as 50% in elderly patients with other signs of atherosclerosis such as widespread peripheral artery disease.¹⁹

Obstructive sleep apnea also commonly coexists with hypertension and its prevalence also increases with age. In addition, elderly patients with obstructive sleep apnea have a higher incidence of cardiovascular complications, including hypertension, than middle-aged people.²⁰ Numerous studies have found that the severity of obstructive sleep apnea corresponds with the likelihood of systemic hypertension.^21–23 Randomized trials and meta-analyses have also concluded that effective treatment with continuous positive airway pressure reduces systemic blood pressure,^24–27 although by less than antihypertensive medications do.

A causal relationship between obstructive sleep apnea and hypertension has not been established with certainty. It is recommended, however, that patients with resistant hypertension be screened for obstructive sleep apnea as a possible cause of their disease.

Other causes of secondary hypertension to keep in mind when evaluating patients who have inappropriate hypertension include thyroid disorders, alcohol and tobacco use, and chronic steroid or NSAID use. Pheochromocytoma and adrenal adenoma, though possible, are less prevalent in the elderly.

Case continued

Physical examination in the above patient revealed an epigastric systolic-diastolic bruit, a sign that, although not sensitive, is highly specific for renal artery stenosis, raising the suspicion of this condition. Duplex ultrasonography of the renal arteries confirmed this suspicion. The patient underwent angiography and revascularization, resulting in a distinct fall in, but not normalization of, his blood pressure.

Though we do not intend to detail the diagnostic approaches and treatments for the various causes of secondary hypertension, we need to briefly mention those for renal artery stenosis.

According to the 2006 ACC/AHA guidelines on peripheral artery disease,²⁸ testing for renal artery stenosis is indicated only if a subsequent corrective procedure is a viable option.

Renal arteriography remains the gold standard for diagnosing renal artery stenosis. However, noninvasive imaging has largely replaced it.

Duplex Doppler ultrasonography, compared with angiography, has a sensitivity of 84% to 98% depending on operator experience, and a specificity of 62% to 99% for detecting renal artery stenosis.²⁹ Some of its limiting factors are the time needed to do the study, its steep learning curve and operator-dependence, and interference with the results by body fat and intestinal gas.

Computed tomographic angiography has a sensitivity of over 90% for detecting renal artery stenosis, and its specificity has been shown to be as high as 99% in certain studies.²⁹ Use of contrast can be a limiting factor in some clinical settings.

Magnetic resonance angiography also offers a sensitivity of 90% to 100% and specificities of 76% to 94% for detecting renal artery stenosis.²⁹ On the other hand, it is costly, and the gadolinium contrast solution used is nephrotoxic, though not as toxic as the contrast used in computed tomographic angiography.

As previously stated, these imaging studies should be used only if corrective measures will be undertaken if clinically significant renal artery stenosis is found. Even in such cases, revascularization may not be curative in all cases. Its effectiveness has been compared with that of medical management alone in a number of studies.^30,31 A meta-analysis³² of six key trials involving more than 1,200 patients showed no difference in systolic and diastolic blood pressures and other clinical outcomes, including all-cause mortality, between the two treatment groups over a 29-month follow-up period.

Hence, although we advise that causes of secondary hypertension be considered in cases of inappropriate hypertension, aggressive management must be pursued on a case-by-case basis.

CASE 2: DRUG ADVERSE EFFECTS

A 75-year-old Hispanic woman with a history of treated breast cancer was recently diagnosed with hypertension. Her blood pressure is controlled on amlodipine (Norvasc) 10 mg daily, and her blood pressure today is 128/80 mm Hg. Her only complaint during this office visit is some swelling of her ankles.

Edema and dihydropyridine calcium channel blockers

Like all drugs, antihypertensive medications come with their own set of adverse effects. These are more common as people age—hence the importance of identifying and effectively managing them in the elderly population.

Calcium channel blockers, especially the dihydropyridines—ie, nifedipine (Adalat), amlodipine, felodipine (Plendil), and isradipine (DynaCirc)—are known to cause peripheral vasodilation. Peripheral edema is a common dose-related effect in people on these drugs. In one study, median leg weight increased by 80 g after amlodipine 5 mg was given for 4 weeks, and by another 68 g on a 10-mg dose.³³

Ankle swelling, encountered more in women, can be very bothersome. The swelling is related to hyperfiltration of fluid into the interstitial space secondary to intracapillary hypertension. Calcium channel blockers predominantly cause arteriolar dilation by paralyzing the precapillary sphincter, thereby elevating intracapillary pressure.

Traditionally, physicians have lowered the dose of the calcium channel blocker, switched to another drug, or added a diuretic to alleviate the swelling. However, giving a diuretic for edema induced by a calcium channel blocker may not relieve the edema.³⁴

Peripheral edema is much less encountered when a calcium channel blocker is given with an inhibitor of the renin-angiotensin system.³⁵ A meta-analysis concluded that the incidence of peripheral edema was lowered by 38% with such a combination. The same study found angiotensin-converting enzyme (ACE) inhibitors significantly more efficacious for this effect than angiotensin receptor blockers (ARBs).³⁵

ACE inhibitors and ARBs are known to cause venodilation, thereby lowering intra-capillary pressure. It is probable that this effect helps remove the extra fluid sequestered in the capillary bed by the arteriolar dilation from the calcium channel blocker.

Pedal edema associated with use of a calcium channel blocker occurs much more commonly in the elderly than in the young. It is clearly dose-dependent, and the incidence peaks after 6 months of therapy. In the patient described above, adding a low dose of an ACE inhibitor or an ARB (if the patient is ACE inhibitor-intolerant) should relieve the swelling.

Hyponatremia and diuretics

Electrolyte imbalances are another common problem encountered in the elderly. Even though for years attention has been directed to the potassium level, hyponatremia has been equally associated with adverse effects in the elderly, such as an increased risk of fractures as shown in the Rotterdam study.³⁶

In 180 hypertensive inpatients, mean age 76.4, Sharabi et al³⁷ found the incidence of hyponatremia to be three times higher in women than in men (odds ratio 3.10, 95% confidence interval 2.07 to 4.67). Patients were 10 times more likely to be affected after age 65 and 14 times more likely after age 75. Most of the patients affected (74.5%) used a thiazide-type diuretic. Even though in many of the patients diuretics were used for more than 1 year before hyponatremia developed, susceptible patients—such as the frail elderly—can develop severe hyponatremia within days of starting to use a thiazide.³⁸

Severe hyponatremia is potentially life-threatening. Most cases are caused by thiazide rather than loop diuretics.³⁸ Thiazides inhibit electrolyte transport at the cortical diluting sites. As they decrease the glomerular filtration rate acutely, they increase proximal water reabsorption (making the plasma hypotonic), reducing water delivery distally. Loop diuretics, on the other hand, have their main effect at the thick ascending limb, reducing the osmolality at the medullary interstitium and not affecting proximal water reabsorption. Additionally, loop diuretics have a shorter half-life than thiazides, which makes hyponatremia more likely to happen with thiazides.

In patients who develop hyponatremia secondary to diuretic use, appropriate treatment includes stopping the medication, restricting water intake, and repleting electrolyte stores.³⁸ As with any cause of chronic hyponatremia, correction must be cautiously monitored and not hastily done.

Therefore, we advise adding a thiazide diuric with caution in the elderly, and we advise avoiding thiazides in patients with high water or alcohol intake.

CASE 3: DEMENTIA AND HYPERTENSION

A 74-year-old man with long-standing hypertension, gout, and chronic obstructive pulmonary disease was recently diagnosed with Alzheimer dementia. He takes enalapril (Vasotec) 10 mg daily for his blood pressure. His blood pressure is 130/78 mm Hg.

Dementia is one of the most important and common neurologic disorders in the elderly. With the rise in average life expectancy, its magnitude has grown to cause a substantial emotional and economic burden on society and health care.

Midlife hypertension has been demonstrated to be an important modifiable risk factor for late-life cognitive decline,³⁹ mild cognitive impairment,⁴⁰ and dementia of all causes.⁴¹ It has been suggested that hypertension might be part of the pathogenesis of dementia, and targeting high blood pressure might prevent its onset.

Moreover, a significant reduction in both Alzheimer and vascular dementia was demonstrated (risk reduction 55%) with the use of a long-acting dihydropyridine calcium channel blocker (nitrendipine) in the Syst-Eur study.⁴² However, data from studies such as Systolic Hypertension in the Elderly Program (SHEP) and the HYVET substudy of cognitive function assessement⁴³ showed no difference in dementia between placebo and active therapy with chlorthalidone (Hygroton) (in SHEP) or indapamide (Lozol) (in the HYVET substudy).

Disorders of calcium homeostasis are associated with the brain’s aging process. Probably, the neuroprotective effect of nitrendipine seen in Syst-Eur was due to its ability to affect this process, independent of its blood pressure-lowering effect.

In another prospective study, people over 60 years of age who complained of subjective memory loss showed a significant and positive association between memory scores and the use of calcium channel blockers (+0.14 ± 0.09 in users vs −0.12 ± 0.06 in nonusers; P = .016) independently of age, sex, white matter hyperintensities, and carotid wall cross-sectional area, all of which were associated with worse memory scores.⁴⁴

Drugs that block the renin-angiotensin system have also been proposed to delay the onset and slow the progression of dementia.⁴⁵ A small randomized, controlled trial suggested that centrally active ACE inhibitors—those that cross the blood-brain barrier, such as captopril (Capoten), lisinopril (Prinivil), ramipril (Altace), and fosinopril (Monopril)—slow cognitive decline in Alzheimer dementia more than non-centrally active ACE inhibitors or calcium channel blockers.⁴⁶

Sink et al⁴⁷ examined data from participants in the Cognition Substudy of the Cardiovascular Health Study⁴⁸ on the effect of ACE inhibitors on cognitive decline. ACE inhibitors, as a class, showed no benefit in reducing the risk of dementia compared with other antihypertensive drug classes. However, centrally active ACE inhibitors, compared with other medications, were associated with a 65% reduction in cognitive decline per year of drug exposure (P = .01). Non-centrally active ACE inhibitors worsened cognitive decline.

It appears that the brain’s renin-angiotensin system plays a role in the pathogenesis of dementia. Indeed, ACE has been shown to degrade amyloid-beta protein, and its level was increased in brain tissue of Alzheimer patients postmortem.⁴⁹

The relationship between blood pressure and cognitive function appears to be curvilinear, so that low blood pressure in late life is also associated with dementia and Alzheimer dementia.⁵⁰ In 5,816 patients age 65 and older, Morris et al⁵¹ calculated the percentile scores of four cognitive tests according to the level of blood pressure. Patients with systolic blood pressures of 100 mm Hg, 120 mm Hg, and 180 mm Hg scored lower on the Mini Mental State Examination than those in the 140 to 160 mm Hg range. Patients with diastolic blood pressures between 80 and 90 mm Hg appeared to have the best cognitive function. This further emphasizes that blood pressure control must be pursued in the very elderly, albeit less aggressively. The MIND substudy of the SPRINT trial⁹ is likely to shed more light on this relationship.

When needed for optimal blood pressure control in a hypertensive patient at risk of dementia, a calcium channel blocker of the dihydropyridine type or a centrally active ACE inhibitor, or both, is preferable.

CASE 4: LABILE HYPERTENSION

A 74-year-old man with hypertension and diabetes mellitus comes to see you in the office. On physical examination, his blood pressure is 175/110 mm Hg. His blood pressure during his last visit 3 months ago was 120/75. He brings a log with him that shows random fluctuations in his blood pressure readings. He takes hydrochlorothiazide 25 mg daily for his blood pressure.

Hypertension in some patients continuously fluctuates between low and high levels. A study in Canada found that up to 15% of all adult hypertensive patients might have labile hypertension.⁵² In the presence of a normal average blood pressure, visit-to-visit blood pressure variability is usually considered a trivial matter. However, some but not all studies have shown that such visit-to-visit variability in blood pressure is an independent predictor of future cardiovascular events in hypertensive patients, independent of the mean systolic blood pressure.^52–54

Blood pressure fluctuates from heartbeat to heartbeat, from morning to night, from winter to summer, and from sitting to standing, and it is prone to increase with exertion, stress, and other factors. But excessive fluctuations in the elderly are most likely the result of excessive stiffness of the arterial tree and a decrease in the windkessel (cushioning) function of the aorta. As a consequence, even small-volume changes in the intravascular system can trigger large blood pressure fluctuations.

There is some evidence that antihypertensive drug classes may differ in their effects on visit-to-visit blood pressure variability. In a 2010 study comparing the effects of different antihypertensive drugs on blood pressure variation, calcium channel blockers and non-loop diuretics were associated with less variation in systolic blood pressure, and calcium channel blockers reduced it the most.⁵⁵

In the patient described above, switching to a low-dose calcium channel blocker with a thorough follow-up is a reasonable plan.

CASE 5: ORTHOSTATIC HYPOTENSION

A 73-year-old woman with long-standing hypertension complains of some dizziness, especially when getting out of bed in the morning. On physical examination, her blood pressure is 134/100 mm Hg sitting and 115/90 standing. She takes amlodipine 10 mg daily, enalapril 10 mg daily, and chlorthalidone 25 mg daily. Chlorthalidone had been added on her last visit 1 month before.

As a result of the increase in the number of elderly patients with hypertension and guidelines recommending better control in this age group, the number of elderly patients on anti-hypertensive drugs has risen significantly. At the same time, the elderly have increasingly presented with adverse effects of treatment.

Orthostatic hypotension is a drop in systolic pressure of 20 mm Hg or more or a drop in diastolic pressure of 10 mm Hg or more on standing, with or without symptoms. These are caused by cerebral hypoperfusion and include dizziness, lightheadedness, generalized weakness, visual blurring, and, in severe cases, syncope.

Alpha-blockers and nitrates have been most commonly implicated in causing orthostatic hypotension, due to venous pooling. Clearly, not all antihypertensive drugs are equal with regard to their venodilatory effects. Thiazide diuretics, by causing fluid volume depletion, and beta-blockers, by interfering with compensatory cardioacceleration with upright posture, are also commonly involved in causing an excessive blood pressure drop with standing.

Systolic orthostatic hypotension has been shown to be a significant and independent predictor of cardiovascular morbidity and death.⁵⁶ Moreover, syncope and subsequent falls are an important cause of injury and death in the elderly.⁵⁷ The clinical combination of hypertension and orthostatic hypotension is, therefore, especially challenging. A compromise between accepting a higher cardiovascular risk at either end of the spectrum with an added higher risk for fall at the lower end has to be made.

To prevent orthostatic hypotension in the elderly, it is important to avoid prescribing high-risk drugs. When starting antihypertensive therapy, a low dose should be used, and the dose should be titrated upward very slowly and cautiously. If orthostatic hypotension is suggested by the history or by the orthostatic test, which is warranted in all elderly hypertensive patients before starting or significantly altering therapy, the potential culprit drug should be withdrawn and the patient reassessed. Improved hydration, elevating the head of the bed, and taking the antihypertensive drug at night are ways to improve symptoms, but these remain largely unproven.

In this patient, the newly added chlorthalidone was stopped, and her symptoms improved.

PSEUDOHYPERTENSION

Since hypertension is defined by a numerical value, it is prudent that this value be accurate. Treating a falsely high reading or leaving a falsely low reading untreated will predispose the elderly patient to increased risk either way. One rare condition in the elderly that can give a falsely high blood pressure reading is pseudohypertension.

Pseudohypertension is a condition in which indirect blood pressure measured by the cuff method overestimates the true intra-arterial blood pressure due to marked underlying arteriosclerosis. The Osler maneuver can be used to differentiate true hypertension from pseudohypertension.⁵⁸ This is performed by palpating the pulseless radial or brachial artery distal to the inflated cuff. If the artery is palpable despite being pulseless, the patient is said to be “Osler-positive” and likely has pseudohypertension.⁵⁸

Pseudohypertension should be suspected if the patient has orthostatic hypotension despite normal blood pressure sitting and standing. Also, elevated blood pressure without appropriate target organ disease should raise the suspicion of pseudohypertension. Apart from the Osler maneuver, measuring the intraarterial pressure can confirm this diagnosis.

A 64-year-old man with chronic lymphocytic leukemia (CLL), Rai stage IV, was admitted to the hospital to undergo his first cycle of chemotherapy with fludarabine, cyclophosphamide, and rituximab. The physical examination at this time was normal except for splenomegaly and painless bilateral inguinal lymphadenopathy. His laboratory results on admission are shown in Table 1.

On the first day after chemotherapy, laboratory testing revealed an elevated plasma potassium level of 7.7 mmol/L. The specimen was drawn into a BD Vacutainer plasma separator tube with lithium-heparin additive (Becton, Dickinson, and Company, Franklin Lake, NJ) and analyzed on a Unicel DXC 800 chemistry analyzer (Beckman Coulter, Inc, Brea, CA).

1. Which electrocardiographic feature is not associated with hyperkalemia?

• Peaked P waves
• Prolonged PR interval
• Shortened QT interval
• Widened QRS
• Asystole

EVALUATING CARDIAC TOXICITY FROM HYPERKALEMIA

Hyperkalemia is not associated with peaked P waves, but rather with a reduction in the size of the P waves.

Hyperkalemia, defined as a plasma potassium concentration above 5.5 mmol/L, occurs as a result either of a release or a shift of intracellular potassium into the intravascular space or of decreased renal excretion. The earliest changes noted on electrocardiography are peaking and narrowing of T waves, followed by shortening of the QT interval. As hyperkalemia progresses, electrocardiography may show bradycardia, absent P waves, and PR interval prolongation, including second- or third-degree atrioventricular block.

At a plasma potassium concentration greater than 7 mmol/L, there may be a junctional escape rhythm, a sine wave pattern with widening of the QRS complex merging with T waves, ventricular fibrillation, or asystole. However, even with high potassium levels, electrocardiographic changes may be absent.¹

The patient denied fatigue, muscle weakness, or palpitations. Electrocardiography did not show peaked T waves, shortened QT intervals, decreased P waves, prolonged PR interval, or widening of the QRS interval.

2. Which is an appropriate intervention for hyperkalemia with cardiac toxicity?

• Membrane stabilization with calcium gluconate
• Shifting potassium into the cells with a beta-adrenergic agonist given by nebulization
• Shifting potassium into the cells with insulin and dextrose
• Removal of potassium with sodium polystyrene sulfonate (Kayexalate)
• Removal of potassium with dialysis

In the setting of cardiac toxicity, the management of hyperkalemia involves stabilizing the cardiac muscle membrane, shifting potassium into the cells, and removing potassium from the body. It is important to do all three interventions when cardiac toxicity is present to provide sustained therapeutic benefit.²

MANAGING HYPERKALEMIA

General principles

When potassium levels are above 6 mmol/L, electrocardiography should always be done. Immediately stop potassium supplementation and any drugs that can cause hyperkalemia, such as potassium-sparing diuretics, angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, nonsteroidal anti-inflammatory drugs, and trimethoprim-sulfamethoxazole (Bactrim). If the potassium level is greater than 6.5 mmol/L, the patient should be placed on telemetric monitoring, and the potassium level should be measured often.

Hyperkalemia with cardiac toxicity

Membrane stabilization involves intravenous infusion of 10 mL of 10% calcium gluconate. The onset of action is 1 to 3 minutes, and the duration of action is 30 to 60 minutes.

Shifting of potassium is done either with insulin or with a beta-adrenergic agonist nebulizer. With insulin, 10 U of regular insulin is given intravenously along with 50 mL of 50% dextrose; the onset of action is 20 minutes, and the duration is 4 to 6 hours. The dose of beta-adrenergic agonist depends on the type used; the onset of action is 20 minutes, with a duration of 2 to 4 hours.

Removal of potassium is achieved either by drug therapy or by dialysis. Sodium or calcium polystyrene sulfonate is given by mouth, 15 g every 6 hours, or 30 to 60 g by retention enema; the onset of action is 1 to 2 hours, with a duration of 4 to 6 hours. If dialysis is used, 2 to 3 hours is recommended; the onset of action is immediate and lasts for the duration of the dialysis session.

CASE CONTINUED

The patient was moved to a telemetry unit. A repeat plasma potassium measurement after 30 minutes confirmed hyperkalemia (Table 1). The specimen was transported to the laboratory via a pneumatic tube system, centrifuged at 3,300 rpm for 10 minutes, and analyzed on a Beckman Unicel DXC 800 chemistry analyzer. The time from specimen collection to analysis was approximately 60 minutes.

He was treated with oral sodium polystyrene sulfonate because no electrocardiographic changes were observed. Subsequent plasma potassium levels drawn, collected, and analyzed with the same technique described above were persistently high. Repeated electrocardiography continued to show no changes related to hyperkalemia.

DIAGNOSING THE CAUSE OF THE APPARENT HYPERKALEMIA

3. Which is the most likely cause of hyperkalemia in this patient?

• Acute renal failure
• Tumor lysis syndrome
• Hemolysis
• Reverse pseudohyperkalemia
• Pseudohyperkalemia

DIFFERENTIAL DIAGNOSIS

The patient had normal levels of blood urea nitrogen and creatinine and adequate urine output, thus ruling out acute renal failure. Hyperuricemia, hyperphosphatemia, and hypocalcemia were not found, thus ruling out tumor lysis syndrome.

In vitro hemolysis is assessed by visual inspection showing a pink or red hue to serum or plasma, or by hemolysis index calculation using spectrophotometric measurements. Factors associated with in vitro hemolysis include vein fragility, the phlebotomist’s skill and technique, and transportation of the specimen, including duration, mode, and temperature. The plasma potassium level was repeatedly measured from a lithium-heparin tube, thus minimizing the possibility of laboratory error. No evidence of hemolysis was observed during the phlebotomy, transportation, or specimen analysis.

Serum and plasma potassium levels were simultaneously measured to test for pseudohyperkalemia, a falsely elevated serum potassium concentration caused by the release of platelet potassium during clot formation or after venipuncture. Contributing factors include the prolonged use of a tourniquet, hemolysis, and marked leukocytosis or thrombocytosis. A serum-to-plasma potassium gradient greater than 0.4 mmol/L is diagnostic of pseudohyperkalemia.

Reverse pseudohyperkalemia, a falsely high potassium level in plasma samples, is defined as a serum-to-plasma potassium gradient less than 0.4 mmol/L (Table 2). This was the most likely cause of the hyperkalemia in our patient.

On the second day after chemotherapy, two blood samples were collected simultaneously—one into a lithium heparin BD Vacutainer plasma separator tube, and the other into a plain, red-top BD Vacutainer serum collection tube without heparin. The specimens were transported to the laboratory by pneumatic tube and were centrifuged at 3,300 rpm for 10 minutes. The specimens were analyzed simultaneously 20 minutes after collection on a Unicel DXC 800 chemistry analyzer. The analysis revealed a serum-to-plasma potassium gradient of −7.1 mmol/L (serum potassium 3.6 mmol/L and plasma potassium 10.7 mmol/L).³ Repeated potassium measurements drawn in similar fashion after 1 hour continued to show a markedly elevated potassium concentration compared with the serum concentration (Table 1).

Serum and plasma samples were again measured simultaneously 24 hours after detecting hyperkalemia to evaluate for pseudohyperkalemia in this patient. In another published report, serum and plasma measurements were obtained 1 week after observing hyperkalemia.⁴ Blood gas analysis was done the same day in two other reported cases.^4,5

Of note, further review of our patient’s medical history noted hyperkalemia at the time he was diagnosed with CLL. At that time, his plasma potassium level was 7.5 mmol/L, a repeated plasma potassium level was 6.9 mmol/L, and a subsequent blood gas analysis—done 30 minutes after the repeated plasma potassium measurement using a Rapidlab analyzer (Siemens Healthcare Diagnostics, Washington, DC)—showed a potassium level of 3.4 mmol/L. The phenomenon of reverse pseudohyperkalemia was not recognized at that time.

The true mechanism of reverse pseudohyperkalemia has not yet been established. Even minor leakage of intracellular potassium from leukemic cells can have a major effect on the extracellular potassium level. Mechanical stressors in the form of pneumatic tube transport and specimen sampling into vacuum tubes have been implicated as causes of this artifact.^5,6 Another possible mechanism is heparin-induced lysis of leukocytes in the setting of hematologic malignancy during laboratory processing.^4,7,8

LESSONS LEARNED

In patients with hematologic proliferative disorders who develop hyperkalemia in the absence of electrocardiographic changes and an obvious cause of increased potassium levels (eg, acute renal failure, tumor lysis syndrome), we should entertain the possibility of hemolysis, laboratory error, pseudohyperkalemia, and reverse pseudohyperkalemia. The potassium level should be remeasured to rule out laboratory error and hemolysis. In patients with marked leukocytosis or thrombocytosis, simultaneous measurement of serum and plasma potassium levels helps diagnose pseudohyperkalemia and reverse pseudohyperkalemia. Also, prompt blood gas analysis can help identify spurious hyperkalemia.

A 64-year-old man with chronic lymphocytic leukemia (CLL), Rai stage IV, was admitted to the hospital to undergo his first cycle of chemotherapy with fludarabine, cyclophosphamide, and rituximab. The physical examination at this time was normal except for splenomegaly and painless bilateral inguinal lymphadenopathy. His laboratory results on admission are shown in Table 1.

On the first day after chemotherapy, laboratory testing revealed an elevated plasma potassium level of 7.7 mmol/L. The specimen was drawn into a BD Vacutainer plasma separator tube with lithium-heparin additive (Becton, Dickinson, and Company, Franklin Lake, NJ) and analyzed on a Unicel DXC 800 chemistry analyzer (Beckman Coulter, Inc, Brea, CA).

1. Which electrocardiographic feature is not associated with hyperkalemia?

• Peaked P waves
• Prolonged PR interval
• Shortened QT interval
• Widened QRS
• Asystole

EVALUATING CARDIAC TOXICITY FROM HYPERKALEMIA

Hyperkalemia is not associated with peaked P waves, but rather with a reduction in the size of the P waves.

Hyperkalemia, defined as a plasma potassium concentration above 5.5 mmol/L, occurs as a result either of a release or a shift of intracellular potassium into the intravascular space or of decreased renal excretion. The earliest changes noted on electrocardiography are peaking and narrowing of T waves, followed by shortening of the QT interval. As hyperkalemia progresses, electrocardiography may show bradycardia, absent P waves, and PR interval prolongation, including second- or third-degree atrioventricular block.

At a plasma potassium concentration greater than 7 mmol/L, there may be a junctional escape rhythm, a sine wave pattern with widening of the QRS complex merging with T waves, ventricular fibrillation, or asystole. However, even with high potassium levels, electrocardiographic changes may be absent.¹

The patient denied fatigue, muscle weakness, or palpitations. Electrocardiography did not show peaked T waves, shortened QT intervals, decreased P waves, prolonged PR interval, or widening of the QRS interval.

2. Which is an appropriate intervention for hyperkalemia with cardiac toxicity?

• Membrane stabilization with calcium gluconate
• Shifting potassium into the cells with a beta-adrenergic agonist given by nebulization
• Shifting potassium into the cells with insulin and dextrose
• Removal of potassium with sodium polystyrene sulfonate (Kayexalate)
• Removal of potassium with dialysis

In the setting of cardiac toxicity, the management of hyperkalemia involves stabilizing the cardiac muscle membrane, shifting potassium into the cells, and removing potassium from the body. It is important to do all three interventions when cardiac toxicity is present to provide sustained therapeutic benefit.²

MANAGING HYPERKALEMIA

General principles

When potassium levels are above 6 mmol/L, electrocardiography should always be done. Immediately stop potassium supplementation and any drugs that can cause hyperkalemia, such as potassium-sparing diuretics, angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, nonsteroidal anti-inflammatory drugs, and trimethoprim-sulfamethoxazole (Bactrim). If the potassium level is greater than 6.5 mmol/L, the patient should be placed on telemetric monitoring, and the potassium level should be measured often.

Hyperkalemia with cardiac toxicity

Membrane stabilization involves intravenous infusion of 10 mL of 10% calcium gluconate. The onset of action is 1 to 3 minutes, and the duration of action is 30 to 60 minutes.

Shifting of potassium is done either with insulin or with a beta-adrenergic agonist nebulizer. With insulin, 10 U of regular insulin is given intravenously along with 50 mL of 50% dextrose; the onset of action is 20 minutes, and the duration is 4 to 6 hours. The dose of beta-adrenergic agonist depends on the type used; the onset of action is 20 minutes, with a duration of 2 to 4 hours.

Removal of potassium is achieved either by drug therapy or by dialysis. Sodium or calcium polystyrene sulfonate is given by mouth, 15 g every 6 hours, or 30 to 60 g by retention enema; the onset of action is 1 to 2 hours, with a duration of 4 to 6 hours. If dialysis is used, 2 to 3 hours is recommended; the onset of action is immediate and lasts for the duration of the dialysis session.

CASE CONTINUED

The patient was moved to a telemetry unit. A repeat plasma potassium measurement after 30 minutes confirmed hyperkalemia (Table 1). The specimen was transported to the laboratory via a pneumatic tube system, centrifuged at 3,300 rpm for 10 minutes, and analyzed on a Beckman Unicel DXC 800 chemistry analyzer. The time from specimen collection to analysis was approximately 60 minutes.

He was treated with oral sodium polystyrene sulfonate because no electrocardiographic changes were observed. Subsequent plasma potassium levels drawn, collected, and analyzed with the same technique described above were persistently high. Repeated electrocardiography continued to show no changes related to hyperkalemia.

DIAGNOSING THE CAUSE OF THE APPARENT HYPERKALEMIA

3. Which is the most likely cause of hyperkalemia in this patient?

• Acute renal failure
• Tumor lysis syndrome
• Hemolysis
• Reverse pseudohyperkalemia
• Pseudohyperkalemia

DIFFERENTIAL DIAGNOSIS

The patient had normal levels of blood urea nitrogen and creatinine and adequate urine output, thus ruling out acute renal failure. Hyperuricemia, hyperphosphatemia, and hypocalcemia were not found, thus ruling out tumor lysis syndrome.

In vitro hemolysis is assessed by visual inspection showing a pink or red hue to serum or plasma, or by hemolysis index calculation using spectrophotometric measurements. Factors associated with in vitro hemolysis include vein fragility, the phlebotomist’s skill and technique, and transportation of the specimen, including duration, mode, and temperature. The plasma potassium level was repeatedly measured from a lithium-heparin tube, thus minimizing the possibility of laboratory error. No evidence of hemolysis was observed during the phlebotomy, transportation, or specimen analysis.

Serum and plasma potassium levels were simultaneously measured to test for pseudohyperkalemia, a falsely elevated serum potassium concentration caused by the release of platelet potassium during clot formation or after venipuncture. Contributing factors include the prolonged use of a tourniquet, hemolysis, and marked leukocytosis or thrombocytosis. A serum-to-plasma potassium gradient greater than 0.4 mmol/L is diagnostic of pseudohyperkalemia.

Reverse pseudohyperkalemia, a falsely high potassium level in plasma samples, is defined as a serum-to-plasma potassium gradient less than 0.4 mmol/L (Table 2). This was the most likely cause of the hyperkalemia in our patient.

On the second day after chemotherapy, two blood samples were collected simultaneously—one into a lithium heparin BD Vacutainer plasma separator tube, and the other into a plain, red-top BD Vacutainer serum collection tube without heparin. The specimens were transported to the laboratory by pneumatic tube and were centrifuged at 3,300 rpm for 10 minutes. The specimens were analyzed simultaneously 20 minutes after collection on a Unicel DXC 800 chemistry analyzer. The analysis revealed a serum-to-plasma potassium gradient of −7.1 mmol/L (serum potassium 3.6 mmol/L and plasma potassium 10.7 mmol/L).³ Repeated potassium measurements drawn in similar fashion after 1 hour continued to show a markedly elevated potassium concentration compared with the serum concentration (Table 1).

Serum and plasma samples were again measured simultaneously 24 hours after detecting hyperkalemia to evaluate for pseudohyperkalemia in this patient. In another published report, serum and plasma measurements were obtained 1 week after observing hyperkalemia.⁴ Blood gas analysis was done the same day in two other reported cases.^4,5

Of note, further review of our patient’s medical history noted hyperkalemia at the time he was diagnosed with CLL. At that time, his plasma potassium level was 7.5 mmol/L, a repeated plasma potassium level was 6.9 mmol/L, and a subsequent blood gas analysis—done 30 minutes after the repeated plasma potassium measurement using a Rapidlab analyzer (Siemens Healthcare Diagnostics, Washington, DC)—showed a potassium level of 3.4 mmol/L. The phenomenon of reverse pseudohyperkalemia was not recognized at that time.

The true mechanism of reverse pseudohyperkalemia has not yet been established. Even minor leakage of intracellular potassium from leukemic cells can have a major effect on the extracellular potassium level. Mechanical stressors in the form of pneumatic tube transport and specimen sampling into vacuum tubes have been implicated as causes of this artifact.^5,6 Another possible mechanism is heparin-induced lysis of leukocytes in the setting of hematologic malignancy during laboratory processing.^4,7,8

LESSONS LEARNED

In patients with hematologic proliferative disorders who develop hyperkalemia in the absence of electrocardiographic changes and an obvious cause of increased potassium levels (eg, acute renal failure, tumor lysis syndrome), we should entertain the possibility of hemolysis, laboratory error, pseudohyperkalemia, and reverse pseudohyperkalemia. The potassium level should be remeasured to rule out laboratory error and hemolysis. In patients with marked leukocytosis or thrombocytosis, simultaneous measurement of serum and plasma potassium levels helps diagnose pseudohyperkalemia and reverse pseudohyperkalemia. Also, prompt blood gas analysis can help identify spurious hyperkalemia.

A 64-year-old man with chronic lymphocytic leukemia (CLL), Rai stage IV, was admitted to the hospital to undergo his first cycle of chemotherapy with fludarabine, cyclophosphamide, and rituximab. The physical examination at this time was normal except for splenomegaly and painless bilateral inguinal lymphadenopathy. His laboratory results on admission are shown in Table 1.

On the first day after chemotherapy, laboratory testing revealed an elevated plasma potassium level of 7.7 mmol/L. The specimen was drawn into a BD Vacutainer plasma separator tube with lithium-heparin additive (Becton, Dickinson, and Company, Franklin Lake, NJ) and analyzed on a Unicel DXC 800 chemistry analyzer (Beckman Coulter, Inc, Brea, CA).

1. Which electrocardiographic feature is not associated with hyperkalemia?

• Peaked P waves
• Prolonged PR interval
• Shortened QT interval
• Widened QRS
• Asystole

EVALUATING CARDIAC TOXICITY FROM HYPERKALEMIA

Hyperkalemia is not associated with peaked P waves, but rather with a reduction in the size of the P waves.

Hyperkalemia, defined as a plasma potassium concentration above 5.5 mmol/L, occurs as a result either of a release or a shift of intracellular potassium into the intravascular space or of decreased renal excretion. The earliest changes noted on electrocardiography are peaking and narrowing of T waves, followed by shortening of the QT interval. As hyperkalemia progresses, electrocardiography may show bradycardia, absent P waves, and PR interval prolongation, including second- or third-degree atrioventricular block.

At a plasma potassium concentration greater than 7 mmol/L, there may be a junctional escape rhythm, a sine wave pattern with widening of the QRS complex merging with T waves, ventricular fibrillation, or asystole. However, even with high potassium levels, electrocardiographic changes may be absent.¹

The patient denied fatigue, muscle weakness, or palpitations. Electrocardiography did not show peaked T waves, shortened QT intervals, decreased P waves, prolonged PR interval, or widening of the QRS interval.

2. Which is an appropriate intervention for hyperkalemia with cardiac toxicity?

• Membrane stabilization with calcium gluconate
• Shifting potassium into the cells with a beta-adrenergic agonist given by nebulization
• Shifting potassium into the cells with insulin and dextrose
• Removal of potassium with sodium polystyrene sulfonate (Kayexalate)
• Removal of potassium with dialysis

In the setting of cardiac toxicity, the management of hyperkalemia involves stabilizing the cardiac muscle membrane, shifting potassium into the cells, and removing potassium from the body. It is important to do all three interventions when cardiac toxicity is present to provide sustained therapeutic benefit.²

MANAGING HYPERKALEMIA

General principles

When potassium levels are above 6 mmol/L, electrocardiography should always be done. Immediately stop potassium supplementation and any drugs that can cause hyperkalemia, such as potassium-sparing diuretics, angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, nonsteroidal anti-inflammatory drugs, and trimethoprim-sulfamethoxazole (Bactrim). If the potassium level is greater than 6.5 mmol/L, the patient should be placed on telemetric monitoring, and the potassium level should be measured often.

Hyperkalemia with cardiac toxicity

Membrane stabilization involves intravenous infusion of 10 mL of 10% calcium gluconate. The onset of action is 1 to 3 minutes, and the duration of action is 30 to 60 minutes.

Shifting of potassium is done either with insulin or with a beta-adrenergic agonist nebulizer. With insulin, 10 U of regular insulin is given intravenously along with 50 mL of 50% dextrose; the onset of action is 20 minutes, and the duration is 4 to 6 hours. The dose of beta-adrenergic agonist depends on the type used; the onset of action is 20 minutes, with a duration of 2 to 4 hours.

Removal of potassium is achieved either by drug therapy or by dialysis. Sodium or calcium polystyrene sulfonate is given by mouth, 15 g every 6 hours, or 30 to 60 g by retention enema; the onset of action is 1 to 2 hours, with a duration of 4 to 6 hours. If dialysis is used, 2 to 3 hours is recommended; the onset of action is immediate and lasts for the duration of the dialysis session.

CASE CONTINUED

The patient was moved to a telemetry unit. A repeat plasma potassium measurement after 30 minutes confirmed hyperkalemia (Table 1). The specimen was transported to the laboratory via a pneumatic tube system, centrifuged at 3,300 rpm for 10 minutes, and analyzed on a Beckman Unicel DXC 800 chemistry analyzer. The time from specimen collection to analysis was approximately 60 minutes.

He was treated with oral sodium polystyrene sulfonate because no electrocardiographic changes were observed. Subsequent plasma potassium levels drawn, collected, and analyzed with the same technique described above were persistently high. Repeated electrocardiography continued to show no changes related to hyperkalemia.

DIAGNOSING THE CAUSE OF THE APPARENT HYPERKALEMIA

3. Which is the most likely cause of hyperkalemia in this patient?

• Acute renal failure
• Tumor lysis syndrome
• Hemolysis
• Reverse pseudohyperkalemia
• Pseudohyperkalemia

DIFFERENTIAL DIAGNOSIS

The patient had normal levels of blood urea nitrogen and creatinine and adequate urine output, thus ruling out acute renal failure. Hyperuricemia, hyperphosphatemia, and hypocalcemia were not found, thus ruling out tumor lysis syndrome.

In vitro hemolysis is assessed by visual inspection showing a pink or red hue to serum or plasma, or by hemolysis index calculation using spectrophotometric measurements. Factors associated with in vitro hemolysis include vein fragility, the phlebotomist’s skill and technique, and transportation of the specimen, including duration, mode, and temperature. The plasma potassium level was repeatedly measured from a lithium-heparin tube, thus minimizing the possibility of laboratory error. No evidence of hemolysis was observed during the phlebotomy, transportation, or specimen analysis.

Serum and plasma potassium levels were simultaneously measured to test for pseudohyperkalemia, a falsely elevated serum potassium concentration caused by the release of platelet potassium during clot formation or after venipuncture. Contributing factors include the prolonged use of a tourniquet, hemolysis, and marked leukocytosis or thrombocytosis. A serum-to-plasma potassium gradient greater than 0.4 mmol/L is diagnostic of pseudohyperkalemia.

Reverse pseudohyperkalemia, a falsely high potassium level in plasma samples, is defined as a serum-to-plasma potassium gradient less than 0.4 mmol/L (Table 2). This was the most likely cause of the hyperkalemia in our patient.

On the second day after chemotherapy, two blood samples were collected simultaneously—one into a lithium heparin BD Vacutainer plasma separator tube, and the other into a plain, red-top BD Vacutainer serum collection tube without heparin. The specimens were transported to the laboratory by pneumatic tube and were centrifuged at 3,300 rpm for 10 minutes. The specimens were analyzed simultaneously 20 minutes after collection on a Unicel DXC 800 chemistry analyzer. The analysis revealed a serum-to-plasma potassium gradient of −7.1 mmol/L (serum potassium 3.6 mmol/L and plasma potassium 10.7 mmol/L).³ Repeated potassium measurements drawn in similar fashion after 1 hour continued to show a markedly elevated potassium concentration compared with the serum concentration (Table 1).

Serum and plasma samples were again measured simultaneously 24 hours after detecting hyperkalemia to evaluate for pseudohyperkalemia in this patient. In another published report, serum and plasma measurements were obtained 1 week after observing hyperkalemia.⁴ Blood gas analysis was done the same day in two other reported cases.^4,5

Of note, further review of our patient’s medical history noted hyperkalemia at the time he was diagnosed with CLL. At that time, his plasma potassium level was 7.5 mmol/L, a repeated plasma potassium level was 6.9 mmol/L, and a subsequent blood gas analysis—done 30 minutes after the repeated plasma potassium measurement using a Rapidlab analyzer (Siemens Healthcare Diagnostics, Washington, DC)—showed a potassium level of 3.4 mmol/L. The phenomenon of reverse pseudohyperkalemia was not recognized at that time.

The true mechanism of reverse pseudohyperkalemia has not yet been established. Even minor leakage of intracellular potassium from leukemic cells can have a major effect on the extracellular potassium level. Mechanical stressors in the form of pneumatic tube transport and specimen sampling into vacuum tubes have been implicated as causes of this artifact.^5,6 Another possible mechanism is heparin-induced lysis of leukocytes in the setting of hematologic malignancy during laboratory processing.^4,7,8

LESSONS LEARNED

In patients with hematologic proliferative disorders who develop hyperkalemia in the absence of electrocardiographic changes and an obvious cause of increased potassium levels (eg, acute renal failure, tumor lysis syndrome), we should entertain the possibility of hemolysis, laboratory error, pseudohyperkalemia, and reverse pseudohyperkalemia. The potassium level should be remeasured to rule out laboratory error and hemolysis. In patients with marked leukocytosis or thrombocytosis, simultaneous measurement of serum and plasma potassium levels helps diagnose pseudohyperkalemia and reverse pseudohyperkalemia. Also, prompt blood gas analysis can help identify spurious hyperkalemia.

A 60-year-old man with hepatocellular carcinoma was admitted to the hospital with pulmonary emboli secondary to inferior vena caval thrombosis that extended to the right atrium.

He became hypotensive on the second day, with a heart rate of 124 per minute, respiratory rate 44 per minute, pulse oxygen saturation 79% on room air, and systolic blood pressure 70 mm Hg. Physical examination revealed abdominal ecchymoses resembling the Cullen sign and flank ecchymoses resembling the Grey Turner sign (Figures 1 and 2).

He was given a bolus of normal saline followed by infusion of fresh frozen plasma and packed red blood cells. His lactate level was 17.52 mg/dL (reference range 0.1–2.2). His hemoglobin and hematocrit decreased precipitously—the hemoglobin from 10.5 g/dL to 5.9 (reference range 14.0–17.5), and the hematocrit from 30.3% to 17.8% (reference range 41–50). He was transferred to the intensive care unit. A do-not-resuscitate order was instituted, and he died 12 hours later.

CONDITIONS RESULTING IN THE CULLEN AND GREY TURNER SIGNS

The Cullen sign, a bluish discoloration of the periumbilical skin, was originally described in 1918 by the gynecologist Thomas Cullen, MD, in a patient with a ruptured ectopic pregnancy.¹ The Grey Turner sign, an ecchymotic discoloration of the lateral abdominal wall or flank, was first reported in 1920 by a surgeon, Dr. George Grey Turner, in a patient with acute pancreatitis.²

The signs occur in about 1% of patients with acute pancreatitis and predict a poor prognosis, with a reported death rate of 37%.3

The appearance of ecchymoses in the periumbilical area or flank has been taught as a hallmark of acute pancreatitis.⁴ However, the original patient described with the Cullen sign did not have pancreatitis,^4,5 and it has been reported with many other conditions, including ruptured aortic aneurysm, splenic rupture, and rectus sheath hematoma, as a complication of anticoagulation or perforated duodenal ulcer, and, as in our patient, as a manifestation of liver disease.

How they occur

The common pathway leading to the occurrence of these subcutaneous ecchymoses is retroperitoneal bleeding followed by tracking of blood from the retroperitoneum through a defect in the transversalis fascia to the abdominal wall musculature and then to the periumbilical subcutaneous tissue. In the Cullen sign, blood diffuses from the retroperitoneum along the gastrohepatic and falciform ligaments to the umbilicus. In the Grey Turner sign, blood diffuses from the posterior pararenal space to the lateral edge of the quadratus lumborum muscle.^6–8 Blood from a retroperitoneal hemorrhage may also diffuse and pool at the inguinal ligament (Fox sign) or at the scrotum (Bryant sign).⁵

The time to the appearance of the Cullen or the Grey Turner sign is thought to be at least 24 hours after the onset of retroperitoneal bleeding and averages about 3 days after the onset of pancreatitis.³

A 60-year-old man with hepatocellular carcinoma was admitted to the hospital with pulmonary emboli secondary to inferior vena caval thrombosis that extended to the right atrium.

He became hypotensive on the second day, with a heart rate of 124 per minute, respiratory rate 44 per minute, pulse oxygen saturation 79% on room air, and systolic blood pressure 70 mm Hg. Physical examination revealed abdominal ecchymoses resembling the Cullen sign and flank ecchymoses resembling the Grey Turner sign (Figures 1 and 2).

He was given a bolus of normal saline followed by infusion of fresh frozen plasma and packed red blood cells. His lactate level was 17.52 mg/dL (reference range 0.1–2.2). His hemoglobin and hematocrit decreased precipitously—the hemoglobin from 10.5 g/dL to 5.9 (reference range 14.0–17.5), and the hematocrit from 30.3% to 17.8% (reference range 41–50). He was transferred to the intensive care unit. A do-not-resuscitate order was instituted, and he died 12 hours later.

CONDITIONS RESULTING IN THE CULLEN AND GREY TURNER SIGNS

The Cullen sign, a bluish discoloration of the periumbilical skin, was originally described in 1918 by the gynecologist Thomas Cullen, MD, in a patient with a ruptured ectopic pregnancy.¹ The Grey Turner sign, an ecchymotic discoloration of the lateral abdominal wall or flank, was first reported in 1920 by a surgeon, Dr. George Grey Turner, in a patient with acute pancreatitis.²

The signs occur in about 1% of patients with acute pancreatitis and predict a poor prognosis, with a reported death rate of 37%.3

The appearance of ecchymoses in the periumbilical area or flank has been taught as a hallmark of acute pancreatitis.⁴ However, the original patient described with the Cullen sign did not have pancreatitis,^4,5 and it has been reported with many other conditions, including ruptured aortic aneurysm, splenic rupture, and rectus sheath hematoma, as a complication of anticoagulation or perforated duodenal ulcer, and, as in our patient, as a manifestation of liver disease.

How they occur

The common pathway leading to the occurrence of these subcutaneous ecchymoses is retroperitoneal bleeding followed by tracking of blood from the retroperitoneum through a defect in the transversalis fascia to the abdominal wall musculature and then to the periumbilical subcutaneous tissue. In the Cullen sign, blood diffuses from the retroperitoneum along the gastrohepatic and falciform ligaments to the umbilicus. In the Grey Turner sign, blood diffuses from the posterior pararenal space to the lateral edge of the quadratus lumborum muscle.^6–8 Blood from a retroperitoneal hemorrhage may also diffuse and pool at the inguinal ligament (Fox sign) or at the scrotum (Bryant sign).⁵

The time to the appearance of the Cullen or the Grey Turner sign is thought to be at least 24 hours after the onset of retroperitoneal bleeding and averages about 3 days after the onset of pancreatitis.³

A 60-year-old man with hepatocellular carcinoma was admitted to the hospital with pulmonary emboli secondary to inferior vena caval thrombosis that extended to the right atrium.

He became hypotensive on the second day, with a heart rate of 124 per minute, respiratory rate 44 per minute, pulse oxygen saturation 79% on room air, and systolic blood pressure 70 mm Hg. Physical examination revealed abdominal ecchymoses resembling the Cullen sign and flank ecchymoses resembling the Grey Turner sign (Figures 1 and 2).

He was given a bolus of normal saline followed by infusion of fresh frozen plasma and packed red blood cells. His lactate level was 17.52 mg/dL (reference range 0.1–2.2). His hemoglobin and hematocrit decreased precipitously—the hemoglobin from 10.5 g/dL to 5.9 (reference range 14.0–17.5), and the hematocrit from 30.3% to 17.8% (reference range 41–50). He was transferred to the intensive care unit. A do-not-resuscitate order was instituted, and he died 12 hours later.

CONDITIONS RESULTING IN THE CULLEN AND GREY TURNER SIGNS

The Cullen sign, a bluish discoloration of the periumbilical skin, was originally described in 1918 by the gynecologist Thomas Cullen, MD, in a patient with a ruptured ectopic pregnancy.¹ The Grey Turner sign, an ecchymotic discoloration of the lateral abdominal wall or flank, was first reported in 1920 by a surgeon, Dr. George Grey Turner, in a patient with acute pancreatitis.²

The signs occur in about 1% of patients with acute pancreatitis and predict a poor prognosis, with a reported death rate of 37%.3

The appearance of ecchymoses in the periumbilical area or flank has been taught as a hallmark of acute pancreatitis.⁴ However, the original patient described with the Cullen sign did not have pancreatitis,^4,5 and it has been reported with many other conditions, including ruptured aortic aneurysm, splenic rupture, and rectus sheath hematoma, as a complication of anticoagulation or perforated duodenal ulcer, and, as in our patient, as a manifestation of liver disease.

How they occur

The common pathway leading to the occurrence of these subcutaneous ecchymoses is retroperitoneal bleeding followed by tracking of blood from the retroperitoneum through a defect in the transversalis fascia to the abdominal wall musculature and then to the periumbilical subcutaneous tissue. In the Cullen sign, blood diffuses from the retroperitoneum along the gastrohepatic and falciform ligaments to the umbilicus. In the Grey Turner sign, blood diffuses from the posterior pararenal space to the lateral edge of the quadratus lumborum muscle.^6–8 Blood from a retroperitoneal hemorrhage may also diffuse and pool at the inguinal ligament (Fox sign) or at the scrotum (Bryant sign).⁵

The time to the appearance of the Cullen or the Grey Turner sign is thought to be at least 24 hours after the onset of retroperitoneal bleeding and averages about 3 days after the onset of pancreatitis.³

Q: Which would be the most appropriate diagnosis?

• Pericarditis
• Acute inferior and right ventricular myocardial infarction
• Anterior and inferior myocardial infarction
• None of the above

A: The correct answer is acute inferior and right ventricular myocardial infarction.

Her electrocardiogram showed sinus rhythm and inferior ST-segment elevation myocardial infarction (STEMI) evidenced by ST-segment elevation in leads II, III, and aVF. Hemodynamic instability or ST-segment elevation of more than 1 mm in lead V₁ raises the suspicion of right ventricular myocardial infarction. In such patients, the American Heart Association guidelines recommend electrocardiography with right-sided precordial leads.¹

A 1-mm ST-segment elevation in the right precordial lead V₄R is one of the most predictive electrocardiographic findings in right ventricular infarction.² The electrocardiographic changes in this type of myocardial infarction may be transient and resolve within 10 hours in up to 48% of cases.³

Echocardiography can also be used to confirm the possibility of right ventricular infarction.

Q: Which clinical condition can occur as a complication of right ventricular myocardial infarction?

• Profound hypotension after nitrate administration
• High-degree heart block
• Atrial fibrillation
• All of the above

A: All of the conditions can occur.

Right ventricular involvement is very common, noted in up to 50% of patients with acute inferior STEMI in postmortem studies.⁴ However, hemodynamically significant right ventricular dysfunction is much less common.

Intravenous volume loading with normal saline is one of the first steps in the management of hypotension associated with right ventricular infarction. Patients with significant bradycardia or a high degree of atrioventricular block may require pacing. Early reperfusion should be achieved, if possible. Heightened suspicion is critical to the early diagnosis of this condition, since the prognosis is much worse than for isolated inferior STEMI.⁴

Our patient was found to have right coronary artery disease requiring percutaneous coronary intervention.

Q: Which would be the most appropriate diagnosis?

• Pericarditis
• Acute inferior and right ventricular myocardial infarction
• Anterior and inferior myocardial infarction
• None of the above

A: The correct answer is acute inferior and right ventricular myocardial infarction.

Her electrocardiogram showed sinus rhythm and inferior ST-segment elevation myocardial infarction (STEMI) evidenced by ST-segment elevation in leads II, III, and aVF. Hemodynamic instability or ST-segment elevation of more than 1 mm in lead V₁ raises the suspicion of right ventricular myocardial infarction. In such patients, the American Heart Association guidelines recommend electrocardiography with right-sided precordial leads.¹

A 1-mm ST-segment elevation in the right precordial lead V₄R is one of the most predictive electrocardiographic findings in right ventricular infarction.² The electrocardiographic changes in this type of myocardial infarction may be transient and resolve within 10 hours in up to 48% of cases.³

Echocardiography can also be used to confirm the possibility of right ventricular infarction.

Q: Which clinical condition can occur as a complication of right ventricular myocardial infarction?

• Profound hypotension after nitrate administration
• High-degree heart block
• Atrial fibrillation
• All of the above

A: All of the conditions can occur.

Right ventricular involvement is very common, noted in up to 50% of patients with acute inferior STEMI in postmortem studies.⁴ However, hemodynamically significant right ventricular dysfunction is much less common.

Intravenous volume loading with normal saline is one of the first steps in the management of hypotension associated with right ventricular infarction. Patients with significant bradycardia or a high degree of atrioventricular block may require pacing. Early reperfusion should be achieved, if possible. Heightened suspicion is critical to the early diagnosis of this condition, since the prognosis is much worse than for isolated inferior STEMI.⁴

Our patient was found to have right coronary artery disease requiring percutaneous coronary intervention.

Q: Which would be the most appropriate diagnosis?

• Pericarditis
• Acute inferior and right ventricular myocardial infarction
• Anterior and inferior myocardial infarction
• None of the above

A: The correct answer is acute inferior and right ventricular myocardial infarction.

Her electrocardiogram showed sinus rhythm and inferior ST-segment elevation myocardial infarction (STEMI) evidenced by ST-segment elevation in leads II, III, and aVF. Hemodynamic instability or ST-segment elevation of more than 1 mm in lead V₁ raises the suspicion of right ventricular myocardial infarction. In such patients, the American Heart Association guidelines recommend electrocardiography with right-sided precordial leads.¹

A 1-mm ST-segment elevation in the right precordial lead V₄R is one of the most predictive electrocardiographic findings in right ventricular infarction.² The electrocardiographic changes in this type of myocardial infarction may be transient and resolve within 10 hours in up to 48% of cases.³

Echocardiography can also be used to confirm the possibility of right ventricular infarction.

Q: Which clinical condition can occur as a complication of right ventricular myocardial infarction?

• Profound hypotension after nitrate administration
• High-degree heart block
• Atrial fibrillation
• All of the above

A: All of the conditions can occur.

Right ventricular involvement is very common, noted in up to 50% of patients with acute inferior STEMI in postmortem studies.⁴ However, hemodynamically significant right ventricular dysfunction is much less common.

Intravenous volume loading with normal saline is one of the first steps in the management of hypotension associated with right ventricular infarction. Patients with significant bradycardia or a high degree of atrioventricular block may require pacing. Early reperfusion should be achieved, if possible. Heightened suspicion is critical to the early diagnosis of this condition, since the prognosis is much worse than for isolated inferior STEMI.⁴

Our patient was found to have right coronary artery disease requiring percutaneous coronary intervention.

Most of us have not spent the past 25 years on the front line continuously managing HIV-infected patients, but I am sure that at various points in our lives we all have been touched by the AIDS epidemic. Whether comforting a woman with knee pain in the office who is crying over the impending death of her son who lives in a group home for men with AIDS, diagnosing immune thrombocytopenia in a college student only to realize it is the seminal manifestation of his HIV infection, pleading unsuccessfully with several neurosurgeons to get one to perform a brain biopsy on an “enhancing ring lesion” in a young gay opera singer, or being part of a team caring for a gouty patient with AIDS and hepatitis C who had just undergone a successful liver transplantation, we all have our stories with resultant reflections on the era of medicine in which we practice.

In July 2012, the New York Times described the new home test for HIV infection as part of “the normalization of a disease once seen as a mark of shame.”¹ As with home pregnancy testing, people can now self-manage their need to know about what is going on in their body. But HIV goes so much deeper than this: it has been and remains a metaphor for and a reflection of many of the social issues that permeate our current political and social environment.

The politics and the social reactions to testing for HIV over the years since the virus was recognized in 1983–1984 is stuff for sociopsychologic treatises. Antibody testing was available in 1985, but in the absence of treatment, to test was simply to deliver a death sentence. Plus, with a diagnosis of AIDS, there would be no dental care, no insurance, no renting of an apartment, and perhaps no job. For some, family ties would be broken as closet doors would be thrown open, revealing a now unrecognized visage wearing the “mark of shame.” Some gay advocates rallied hard against testing, since anonymity and social protection for the infected could not be assured, a pragmatic response to blatant discrimination. In 1987, the first home test for HIV was in development, but—no surprise—there was no need for it.

As early treatments such as zidovudine (AZT) appeared and the value of specific antibiotic prophylaxis was demonstrated, there was some initial hope for treatment, and thus testing made medical sense. The size of the population infected (we were looking at the tip of the iceberg) was also being realized, so testing appealed to the social consciousness—try to limit infection. But discrimination wasn’t gone, and the politics of the time couldn’t quite handle all of the implications of a rapidly growing epidemic. America wasn’t ready for clean-needle-exchange programs, promotion of condom use, or open discussion of gay lifestyles. The Reagan White House was initially dead silent, except for proposing to limit entrance of potentially infected immigrants and promoting abstinence as the ideal protective approach.

Social righteousness took some hold, and protection of patient anonymity and autonomy became of paramount importance. But unintended consequences turned out to include limitation of testing: laws were written to require that HIV testing be accompanied by “appropriate,” stringently defined counseling, something that wasn’t always feasible. Patients needed to sign a release to be tested (“opt in”); many just said no. This tied the hands of physicians, so we developed work-arounds: we checked lymphocyte counts and CD4 counts to help us take care of patients too afraid to let us test for HIV directly.

Finally, in 2006, as therapies began to become increasingly effective and more data started to accumulate regarding the benefits of early antiretroviral therapy, the US Centers for Disease Control and Prevention recommended routine testing for all patients entering most acute health care facilities, unless they would actively decline (“opt out”). We have still not hit full stride in implementing universal testing for HIV. Nor have we hit our stride on fully accepting all demographic segments of the population. In some communities, HIV infection is still equivalent to the scarlet letter of Hester Prynne, not just because of the disease itself but because of the lifestyle it implies. Legislating laboratory testing practices cannot change all social attitudes. But maybe, hopefully, it is another step.

Dr. Christine Koval in this issue of the Journal discusses the practical use of the newly approved home HIV test. It is a short article, but it took a very, very long time for social and political forces to be modestly aligned sufficiently for there to be anything to write about. Since perhaps 18% of HIV-infected Americans are unaware of their infection, maybe some TV ads for this test, wedged between the ads for treating erectile dysfunction, can indeed bring (as the New York Times described) further “normalization” to the approach to managing HIV-infected patients.

Most of us have not spent the past 25 years on the front line continuously managing HIV-infected patients, but I am sure that at various points in our lives we all have been touched by the AIDS epidemic. Whether comforting a woman with knee pain in the office who is crying over the impending death of her son who lives in a group home for men with AIDS, diagnosing immune thrombocytopenia in a college student only to realize it is the seminal manifestation of his HIV infection, pleading unsuccessfully with several neurosurgeons to get one to perform a brain biopsy on an “enhancing ring lesion” in a young gay opera singer, or being part of a team caring for a gouty patient with AIDS and hepatitis C who had just undergone a successful liver transplantation, we all have our stories with resultant reflections on the era of medicine in which we practice.

In July 2012, the New York Times described the new home test for HIV infection as part of “the normalization of a disease once seen as a mark of shame.”¹ As with home pregnancy testing, people can now self-manage their need to know about what is going on in their body. But HIV goes so much deeper than this: it has been and remains a metaphor for and a reflection of many of the social issues that permeate our current political and social environment.

The politics and the social reactions to testing for HIV over the years since the virus was recognized in 1983–1984 is stuff for sociopsychologic treatises. Antibody testing was available in 1985, but in the absence of treatment, to test was simply to deliver a death sentence. Plus, with a diagnosis of AIDS, there would be no dental care, no insurance, no renting of an apartment, and perhaps no job. For some, family ties would be broken as closet doors would be thrown open, revealing a now unrecognized visage wearing the “mark of shame.” Some gay advocates rallied hard against testing, since anonymity and social protection for the infected could not be assured, a pragmatic response to blatant discrimination. In 1987, the first home test for HIV was in development, but—no surprise—there was no need for it.

As early treatments such as zidovudine (AZT) appeared and the value of specific antibiotic prophylaxis was demonstrated, there was some initial hope for treatment, and thus testing made medical sense. The size of the population infected (we were looking at the tip of the iceberg) was also being realized, so testing appealed to the social consciousness—try to limit infection. But discrimination wasn’t gone, and the politics of the time couldn’t quite handle all of the implications of a rapidly growing epidemic. America wasn’t ready for clean-needle-exchange programs, promotion of condom use, or open discussion of gay lifestyles. The Reagan White House was initially dead silent, except for proposing to limit entrance of potentially infected immigrants and promoting abstinence as the ideal protective approach.

Social righteousness took some hold, and protection of patient anonymity and autonomy became of paramount importance. But unintended consequences turned out to include limitation of testing: laws were written to require that HIV testing be accompanied by “appropriate,” stringently defined counseling, something that wasn’t always feasible. Patients needed to sign a release to be tested (“opt in”); many just said no. This tied the hands of physicians, so we developed work-arounds: we checked lymphocyte counts and CD4 counts to help us take care of patients too afraid to let us test for HIV directly.

Finally, in 2006, as therapies began to become increasingly effective and more data started to accumulate regarding the benefits of early antiretroviral therapy, the US Centers for Disease Control and Prevention recommended routine testing for all patients entering most acute health care facilities, unless they would actively decline (“opt out”). We have still not hit full stride in implementing universal testing for HIV. Nor have we hit our stride on fully accepting all demographic segments of the population. In some communities, HIV infection is still equivalent to the scarlet letter of Hester Prynne, not just because of the disease itself but because of the lifestyle it implies. Legislating laboratory testing practices cannot change all social attitudes. But maybe, hopefully, it is another step.

Dr. Christine Koval in this issue of the Journal discusses the practical use of the newly approved home HIV test. It is a short article, but it took a very, very long time for social and political forces to be modestly aligned sufficiently for there to be anything to write about. Since perhaps 18% of HIV-infected Americans are unaware of their infection, maybe some TV ads for this test, wedged between the ads for treating erectile dysfunction, can indeed bring (as the New York Times described) further “normalization” to the approach to managing HIV-infected patients.

Most of us have not spent the past 25 years on the front line continuously managing HIV-infected patients, but I am sure that at various points in our lives we all have been touched by the AIDS epidemic. Whether comforting a woman with knee pain in the office who is crying over the impending death of her son who lives in a group home for men with AIDS, diagnosing immune thrombocytopenia in a college student only to realize it is the seminal manifestation of his HIV infection, pleading unsuccessfully with several neurosurgeons to get one to perform a brain biopsy on an “enhancing ring lesion” in a young gay opera singer, or being part of a team caring for a gouty patient with AIDS and hepatitis C who had just undergone a successful liver transplantation, we all have our stories with resultant reflections on the era of medicine in which we practice.

In July 2012, the New York Times described the new home test for HIV infection as part of “the normalization of a disease once seen as a mark of shame.”¹ As with home pregnancy testing, people can now self-manage their need to know about what is going on in their body. But HIV goes so much deeper than this: it has been and remains a metaphor for and a reflection of many of the social issues that permeate our current political and social environment.

The politics and the social reactions to testing for HIV over the years since the virus was recognized in 1983–1984 is stuff for sociopsychologic treatises. Antibody testing was available in 1985, but in the absence of treatment, to test was simply to deliver a death sentence. Plus, with a diagnosis of AIDS, there would be no dental care, no insurance, no renting of an apartment, and perhaps no job. For some, family ties would be broken as closet doors would be thrown open, revealing a now unrecognized visage wearing the “mark of shame.” Some gay advocates rallied hard against testing, since anonymity and social protection for the infected could not be assured, a pragmatic response to blatant discrimination. In 1987, the first home test for HIV was in development, but—no surprise—there was no need for it.

As early treatments such as zidovudine (AZT) appeared and the value of specific antibiotic prophylaxis was demonstrated, there was some initial hope for treatment, and thus testing made medical sense. The size of the population infected (we were looking at the tip of the iceberg) was also being realized, so testing appealed to the social consciousness—try to limit infection. But discrimination wasn’t gone, and the politics of the time couldn’t quite handle all of the implications of a rapidly growing epidemic. America wasn’t ready for clean-needle-exchange programs, promotion of condom use, or open discussion of gay lifestyles. The Reagan White House was initially dead silent, except for proposing to limit entrance of potentially infected immigrants and promoting abstinence as the ideal protective approach.

Social righteousness took some hold, and protection of patient anonymity and autonomy became of paramount importance. But unintended consequences turned out to include limitation of testing: laws were written to require that HIV testing be accompanied by “appropriate,” stringently defined counseling, something that wasn’t always feasible. Patients needed to sign a release to be tested (“opt in”); many just said no. This tied the hands of physicians, so we developed work-arounds: we checked lymphocyte counts and CD4 counts to help us take care of patients too afraid to let us test for HIV directly.

Finally, in 2006, as therapies began to become increasingly effective and more data started to accumulate regarding the benefits of early antiretroviral therapy, the US Centers for Disease Control and Prevention recommended routine testing for all patients entering most acute health care facilities, unless they would actively decline (“opt out”). We have still not hit full stride in implementing universal testing for HIV. Nor have we hit our stride on fully accepting all demographic segments of the population. In some communities, HIV infection is still equivalent to the scarlet letter of Hester Prynne, not just because of the disease itself but because of the lifestyle it implies. Legislating laboratory testing practices cannot change all social attitudes. But maybe, hopefully, it is another step.

Dr. Christine Koval in this issue of the Journal discusses the practical use of the newly approved home HIV test. It is a short article, but it took a very, very long time for social and political forces to be modestly aligned sufficiently for there to be anything to write about. Since perhaps 18% of HIV-infected Americans are unaware of their infection, maybe some TV ads for this test, wedged between the ads for treating erectile dysfunction, can indeed bring (as the New York Times described) further “normalization” to the approach to managing HIV-infected patients.