You are on page 1of 9

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/7426126

Management of Diabetes During Acute Stroke and Inpatient Stroke


Rehabilitation

Article  in  Archives of Physical Medicine and Rehabilitation · January 2006


DOI: 10.1016/j.apmr.2005.07.306 · Source: PubMed

CITATIONS READS
18 905

5 authors, including:

Felicia Hill-Briggs Robert Samuel Mayer


Johns Hopkins Medicine Johns Hopkins Medicine
109 PUBLICATIONS   2,870 CITATIONS    25 PUBLICATIONS   461 CITATIONS   

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Felicia Hill-Briggs on 23 January 2019.

The user has requested enhancement of the downloaded file.


2377

CLINICAL MANAGEMENT REVIEW

Management of Diabetes During Acute Stroke and Inpatient


Stroke Rehabilitation
Sherita Hill Golden, MD, MHS, Felicia Hill-Briggs, PhD, Kathleen Williams, MSN, APN, Karen Stolka, RD,
R. Samuel Mayer, MD
ABSTRACT. Golden SH, Hill-Briggs F, Williams K, Stolka IABETES, A COMMON COMORBID condition among
K, Mayer RS. Management of diabetes during acute stroke and
inpatient stroke rehabilitation. Arch Phys Med Rehabil 2005;
D patients in the acute rehabilitation setting, results in poorer
outcomes during hospitalization. Hyperglycemia among dia-
86:2377-84. betic inpatients is associated with an increased risk of infec-
tious complications,1-3 pulmonary and vascular complications,4
Objectives: To summarize evidence on the impact of hy- and in-hospital mortality.5 In stroke patients, the problem of
perglycemia on stroke outcomes and to present therapy algo- poor diabetes-related outcomes deserves particular attention.
rithms for inpatient management in diabetic stroke patients. Prevalence of diabetes in acute stroke patients is reported to be
Data Sources: Guidelines for inpatient management of di- as high as 24% to 38%.6,7 An estimated 66% of stroke survi-
abetes were reviewed and extracted from a technical review vors meet criteria for diabetes mellitus or impaired glucose
and recommendations from 2 national diabetes and endocrine tolerance at 3 months poststroke.6
organizations. MEDLINE database searches were conducted Diabetes management in stroke patients can be particularly
using key words: stroke, diabetes, hyperglycemia, hypoglyce- difficult. Impairments such as dysphagia, neuropathy, hemipa-
mia, inpatient, hospitalized, treatment, outcomes, disability, resis and ataxia, visual impairment, cognition, and mood im-
self-management, and education. pairment have significant consequences for diabetes care and
Study Selection: Studies were selected that specifically ad- control. The inpatient physiatrist is often challenged by these
dressed the impact of the following in stroke patients: hyper- issues, and, therefore, should be updated with the latest litera-
glycemia and diabetes on rehabilitation outcomes, management ture and recommendations on diabetic management in this
strategies for hyperglycemia and diabetes, and strategies for special population. While the literature on tight glycemic con-
facilitating diabetes self-management. trol in the inpatient setting has been recently summarized in an
Data Extraction: Two authors independently extracted data extensive technical review by the American Diabetes Associ-
and management practices from selected articles and published ation (ADA),8 our goal is to highlight and summarize the
practice guidelines. management issues that affect acute stroke and poststroke
Data Synthesis: Diabetes is prevalent in stroke patients and rehabilitation patients and to provide a succinct resource for
results in poorer inpatient hospital and rehabilitation outcomes. physiatrists. The purposes of this clinical management review
Management of diabetes in stroke patients is further compli- are: (1) to summarize literature pertaining to diabetes- and
cated by impairments in mobility and vision, necessitating hyperglycemia-related outcomes in acute inpatient stroke and
accommodation strategies and tools for self-management. Op- rehabilitation patients; (2) to describe interventions to improve
timal management of hyperglycemia using insulin or oral hy- glycemic control and outcomes in hospitalized patients; (3) to
poglycemic agents results in reduced morbidity and mortality present algorithms for inpatient insulin and oral hypoglycemic
among diabetic inpatients. therapy, based on clinical practice guidelines; (4) to discuss
Conclusions: To achieve inpatient glycemic management special rehabilitation considerations and specific recommenda-
targets, use of clinical management algorithms, self-manage- tions for managing diabetes in patients with stroke and disabil-
ment tools, and systems approaches such as diabetes manage- ity (eg, dysphagia, hemiparesis and ataxia, visual and cognitive
ment teams are useful. impairment); and (5) to present areas of future research in this
Key Words: Diabetes mellitus; Hyperglycemia; Rehabilita- patient population.
tion; Stroke.
© 2005 by the American Congress of Rehabilitation Medi- METHODS
cine and the American Academy of Physical Medicine and The ADA technical review and 2005 Clinical Practice Rec-
Rehabilitation ommendations on management of diabetes during hospitaliza-
tion and the American Association of Clinical Endocrinologists
(AACE) guidelines for inpatient management of diabetes were
From the Departments of Medicine (Golden, Hill-Briggs, Williams), Physical reviewed and extracted. Reference lists from these documents
Medicine and Rehabilitation (Hill-Briggs, Mayer), and Nutrition (Stolka), Johns yielded additional studies. A series of MEDLINE database
Hopkins University School of Medicine, Baltimore, MD; and the Department of searches were conducted using key words including stroke,
Epidemiology (Golden), Johns Hopkins University Bloomberg School of Public
Health Baltimore, MD.
diabetes, hyperglycemia, hypoglycemia, rehabilitation, inpa-
Supported by the Robert Wood Johnson Minority Medical Faculty Development tient, hospitalized, treatment, outcomes, disability, self-man-
Program Award, the National Institute of Diabetes, Digestive, and Kidney Diseases agement, and education. Searches were limited to English-
(grant no. 3 U01 DK48485-10S1), and the National Heart, Lung, and Blood Institute language, human studies. Expert practices by a university
(grant no. 1 K01 HL076644-01).
No commercial party having a direct financial interest in the results of the research
hospital-affiliated inpatient diabetes management service were
supporting this article has or will confer a benefit upon the author(s) or upon any incorporated into the treatment algorithms presented.
organization with which the author(s) is/are associated. Thirty studies were selected that specifically addressed the
Reprint requests to Sherita Hill Golden, MD, MHS, Johns Hopkins University impact of hyperglycemia and diabetes on short-term and reha-
School of Medicine, Division of Endocrinology and Metabolism, 2024 E Monument
St, Ste 2-600, Baltimore, MD 21205, e-mail: sahill@jhmi.edu.
bilitation outcomes in stroke patients, management strategies
0003-9993/05/8612-9849$30.00/0 for hyperglycemia and diabetes in stroke patients, and strate-
doi:10.1016/j.apmr.2005.07.306 gies for facilitating diabetes self-management in patients with

Arch Phys Med Rehabil Vol 86, December 2005


2378 HYPERGLYCEMIA IN REHABILITATION, Golden

stroke and disability. Two authors independently extracted data glycemia, defined as a blood sugar of 130mg/dL or higher, had
and management practices from the selected articles and the a longer LOS and greater inpatient charges than those who
published practice guidelines. were normoglycemic.15
To our knowledge, the Copenhagen Stroke Study10 is the
RESULTS only study that has evaluated the effects of diabetes on reha-
bilitation outcomes. In this study, while clinical outcomes were
Stroke, Hyperglycemia, and Diabetes comparable in patients with and without diabetes, those with
There are several proposed mechanisms linking hyperglyce- diabetes recovered more slowly, despite having similar Scan-
mia to poor clinical outcomes, including adverse effects on dinavian Stroke Scale scores as the nondiabetic patients at
immune, endothelial, and cardiovascular function; increased baseline and at the end of rehabilitation. This study did not
thrombosis, inflammation, and oxidative stress; and adverse specifically examine the relation of glucose control on admis-
effects on the brain, particularly following brain ischemia.8 As sion or during rehabilitation to clinical outcomes or length of
recently reviewed by Clement et al,8 hyperglycemia has harm- rehabilitation.
ful effects on the brain that worsen stroke outcomes, including
enhanced neuronal damage following brain ischemia, espe- Interventions to Improve Glycemic Control and
cially in the ischemic penumbra, and increased acidosis and Outcomes in Hospitalized Patients
lactate levels, which predict clinical outcomes and infarct size
Three clinical trials have shown that intensive glycemic control
in stroke. Additional harmful effects at the cellular level in-
in the acute hospital setting improves clinical outcomes in indi-
clude DNA fragmentation, disruption of the blood-brain bar-
viduals with diabetes. In the Diabetes Mellitus, Insulin Glucose
rier, increased deposition of ␤-amyloid precursor protein, and
Infusion in Acute Myocardial Infarction study,20 patients with
increased superoxide levels.8
diabetes and acute myocardial infarction randomized to intensive
Diabetes is associated with poorer prognosis following
insulin therapy followed by a multishot regimen for 3 months
stroke, including higher mortality rate during hospitaliza-
experienced a 29% reduction in 1-year mortality compared with
tion9,10; higher incidence of urinary retention and urinary tract
individuals receiving conventional therapy. Van den Berghe et
infection during rehabilitation11; and poorer functional out-
al21 demonstrated that patients admitted to the surgical intensive
comes, including worse neurologic recovery,12 slowed rate of
care unit who received intensive insulin therapy via insulin infu-
physical recovery, and reduced rate and magnitude of cognitive
sion to maintain a blood glucose of 80 to 110mg/dL had a
recovery.13 Several studies have shown that the degree of
significant reduction in morbidity and mortality compared with
hyperglycemia is directly related to outcomes in the acute
patients receiving less intensive insulin therapy. Finally, a recent
stroke setting. Capes et al14 performed a meta-analysis of 26
study22 in diabetic patients undergoing coronary artery bypass
studies and found that for both ischemic and hemorrhagic
surgery showed reduced atrial fibrillation, a shorter postoperative
strokes, an admission glucose from 108 to 144mg/dL was
LOS, and improved long-term postsurgical follow-up in patients
associated with increased in-hospital and 30-day mortality.
treated with glucose insulin infusion perioperatively compared
Hyperglycemia, defined as an admission glucose of 130mg/dL
with standard therapy. Despite data demonstrating association of
or higher, was also found to independently predict 30-day and
hyperglycemia with poorer outcomes following stroke,17,19 there
1-year mortality in a study of 656 acute stroke patients, 52% of
have been no clinical trials examining the impact of intensive
whom had diabetes.15 Similarly, individuals with stress hyper-
glycemic control in the acute stroke setting on clinical outcomes.
glycemia, newly diagnosed diabetes, and previously diagnosed
The Glucose Insulin in Stroke Trial (GIST)23 examined the safety
diabetes had a higher mortality than normoglycemic patients in
of glucose-insulin-potassium infusion to lower glucose levels to
another study, despite the type and site of the strokes being
72 to 126mg/dL and found that this is safe and feasible; however,
similar among the groups.16
the GIST report did not report data on functional recovery.
Hyperglycemia has also been associated with poorer func-
Two studies24,25 evaluating interventions by a diabetes team
tional outcomes following stroke. Compared with individuals
have shown significant reductions in LOS. In 1 retrospective
with a mean admission glucose of 144mg/dL, those with an
study,25 the patients who were seen by the diabetes team had a
admission glucose of 160mg/dL had less functional improve-
significantly shorter LOS than patients managed by their inter-
ment following stroke, and for every 100mg/dL increase in
nist alone or by an individual endocrine consultant. In the
admission glucose there was a 24% reduction in the odds of
second study,24 which was a clinical trial, patients randomized
neurologic improvement.17 This study did not perform sub-
to the diabetes team had a significantly shorter LOS and a
group analyses in patients with and without diabetes. Similarly,
lower 1-month readmission rate.
Pulsinelli et al12 found that stroke-related deficits were more
severe among individuals with an admission glucose of more
than 120mg/dL, and in a subgroup of patients who were fol- Guidelines for Inpatient Clinical Management of
lowed prospectively, only 43% of those with admission glu- Hyperglycemia
coses of more than 120mg/dL were able to return to work The ADA has recently published guidelines for the manage-
compared with 76% with glucoses below this level. Admission ment of hyperglycemia in the hospital in a thorough technical
glucose and/or a history of diabetes has also been shown to review,8 which was recently summarized in the 2005 Clinical
independently predict hemorrhage as a complication of throm- Practice Recommendations.26 The guidelines of the ADA and
bolytic therapy18 and in an imaging study, doubling of the the AACE are summarized in appendix 1. These are the rec-
admission glucose from 90 to 180mg/dL reduced the penum- ommended glycemic targets for all inpatients and are not
bral salvage, increased the final infarct size, and resulted in specific to stroke or rehabilitation patients.
greater lactate production in the area of the infarct.19 These Insulin therapy in the hospital setting. To adequately con-
changes were independent of baseline stroke severity, lesion trol blood sugars, patients require both basal and nutritional
size, and diabetes status. insulin. Basal insulin is the amount of exogenous insulin nec-
In addition to being associated with poor clinical outcomes, essary to prevent unchecked gluconeogenesis and ketosis and
hyperglycemia is associated with increased length of stay nutritional insulin is the amount of insulin necessary to cover
(LOS).5 In a study of acute stroke patients, those with hyper- discrete meals or other forms of nutrition (ie, intravenous

Arch Phys Med Rehabil Vol 86, December 2005


HYPERGLYCEMIA IN REHABILITATION, Golden 2379

Table 1: Pharmacokinetics of Various Insulins


Insulins Onset of Action Peak of Action Duration of Action

Rapid acting
Lispro/aspart 5⫺15min 1⫺2h 3⫺5h
Regular 30⫺60min 2⫺4h 6⫺8h
Intermediate acting
Neutral protamine Hagedorn 1⫺3h 5⫺7h 13⫺18h
Long acting
Ultralente 2⫺4h 8⫺14h 18⫺30h
Glargine Within 4h Peakless ⬎24h

dextrose, total parenteral nutrition [TPN], enteral feedings, are not modified during hospitalization to meet a patient’s
nutritional supplements).8 Table 1 summarizes the various changing insulin requirements. Queale et al27 found that the
types of insulins as well as their pharmacokinetics, which is risk of hyperglycemia among medical patients was 3-fold
important to understand in selecting insulins for basal and higher when sliding scale insulin regimens were used alone,27
nutritional coverage. In determining the appropriate dose of highlighting the importance of their use in conjunction with
insulin, the total daily dosage of insulin should be calculated basal insulin.
based on the patient’s body weight. Forty to 50% of this total Finally, there are conditions in which glucose is optimally
daily dose should be administered as basal insulin and the managed with a continuous insulin infusion in the form of a
remaining daily dosage should be divided to cover meals/ regular insulin drip. These conditions include the following:
nutritional intake.8 A suggested algorithm is outlined in appen- diabetic ketoacidosis and nonketotic hyperosmolar state, peri-
dix 2. operative care, organ transplantation, myocardial infarction and
Patients with insulin deficiency require some form of basal cardiogenic shock, stroke, high-dose glucocorticoid use, NPO
insulin administration at all times, even when they are nothing status in T1D, critical care illness, and during dose-finding
by mouth (NPO), to prevent the development of diabetic ke- strategy prior to initiation, or reinitiation of subcutaneous
toacidosis. Sliding scale insulin coverage is often inadequate in insulin.
these patients (see below) and failure to provide stable, basal Oral hypoglycemic therapy in the hospital setting. The
insulin can result in significant morbidity and potential mor- classes of oral agents and their mechanisms and onset of action
tality. These patients include those with known type 1 diabetes are summarized in table 2. The use of oral agents in the
mellitus (T1D) and those with a history of a pancreatectomy or inpatient setting has not been well studied. Each class of oral
pancreatic dysfunction, wide fluctuations in blood sugar levels, hypoglycemic agent has restrictions that may limit its use in
diabetic ketoacidosis, insulin use for greater than 5 years, hospitalized patients.8 Sulfonylureas are long-acting, which
and/or a history of diabetes for greater than 10 years. increases the risk of hypoglycemia, they do not allow rapid
Correction dose or supplemental insulin, also frequently dose adjustment, and they have varied durations of action in
referred to as sliding scale insulin, is used to treat hyperglyce- different individuals. Metformin is contraindicated in patients
mia before or between meals and to correct hyperglycemia in with congestive heart failure, hypoperfusion, renal insuffi-
NPO patients or in those receiving basal insulin but not eating ciency (creatinine ⬎1.4mg/dL in women, ⬎1.5mg/dL in men),
scheduled meals.8 The following correction doses are recom- old age (⬎80y), and conditions of hypoxia (ie, chronic obstruc-
mended based on the preprandial blood sugars: 1U per tive pulmonary disease). In addition, metformin should not be
50mg/dL increment over 180mg/dL in T1D and 1U per used in renal transplant recipients due to the increased risk of
30mg/dL increment over 180mg/dL in type 2 diabetes mellitus renal insufficiency in this setting. Patients with macroalbumin-
(T2D). While supplemental insulin is useful in tightening gly- uria (spot urine microalbumin ⱖ300mg/g of creatinine) should
cemic control in the inpatient setting, traditional sliding scales have a creatinine clearance checked, regardless of age, to
are dangerous when not used with basal insulin and when they ensure that it is greater than 60mL·min⫺1·1.73m⫺2. The thia-

Table 2: Mechanism of Action of Oral Agents


Class of Oral Agents Agents Mechanism of Action

Sulfonylureas Glipizide Insulin secretagogues


Glyburide
Glimepiride
Metglitinides Nateglinide
Postprandial insulin secretagogues
Repaglinide
Biguanides Metformin Insulin sensitizer (liver)
Thiazolidinediones Pioglitazone Insulin sensitizer (skeletal muscle)
Rosiglitazone
␣-glucosidase inhibitors Acarbose Delay glucose absorption by antagonizing enzymes that
hydrolyze starches to oligosaccharides (pancreatic
␣-amylase) and oligo-, tri-, and disaccharides to
glucose (intestinal ␣-glucosidases)
Miglitol

Arch Phys Med Rehabil Vol 86, December 2005


2380 HYPERGLYCEMIA IN REHABILITATION, Golden

Fig 1. Algorithm for initiation of oral hypoglycemic therapy. Abbreviations: ALT, alanine transaminase; AST, aspartate transaminase; bid,
twice daily; BMI, body mass index.© 2004 by the Endocrine Society. Adapted with permission.50

zolidinediones may precipitate congestive heart failure due to hypoglycemic coma. Moreover, during acute hypoglycemia,
increased intravascular volume, are contraindicated in patients patients demonstrate impairments in cognition, mood, and be-
with liver disease, and have a slow onset of action (ie, several havior, and function that may not fully recover until 90 min-
weeks). utes32 to as much as 1 day33 after restoration of blood glucose
In patients who are newly diagnosed with T2D during their to normal. Thus, hypoglycemia should be avoided in poststroke
hospitalization, it may be necessary to initiate therapy with oral patients. Although the overall frequency of hypoglycemia in
agents in preparation for hospital discharge. A suggested algo- this population is unknown, treatment of hyperglycemia in
rithm, used by our inpatient diabetes management service, is acute stroke with glucose-potassium-insulin infusions resulted
outlined in figure 1. Finally, patients well controlled with oral in hypoglycemia in 16% of patients in an initial study23 and in
hypoglycemic agents in the outpatient setting may require 9% of patients in a larger, follow-up study.34
supplemental insulin while hospitalized. Current guidelines Risk factors for development of hypoglycemia include sudden
recommend short-acting supplemental insulin as needed prior NPO status or reduction in oral intake; discontinuation of enteral
to meals (1U per 30mg/dL increment over 180mg/dL).8 If there feedings, TPN, or intravenous dextrose; administration of premeal
is significant fasting hyperglycemia with an elevated glycosy- insulin without ingestion of the meal; unexpected transport from
lated hemoglobin (Hb A1c) level, consideration can be given to the nursing unit after rapid-acting insulin is given; and reduction in
adding a shot of 10 to 20U of bedtime insulin, in the form of the dose of corticosteroids.8 Prevention of hypoglycemia in these
neutral protamine Hagedorn (NPH) insulin or glargine, to the settings requires open communication between nursing and phy-
oral agents. Initiating insulin therapy in the hospital allows sician staff and recognition of when to administer intravenous
time for diabetic teaching so that the patient is comfortable dextrose in certain situations (eg, prior to transport of patient from
with insulin administration prior to discharge. the unit after rapid-acting insulin is given).
Prevention and management of hypoglycemia. Hypogly-
cemia is a side effect of insulin therapy in the hospital and is Impact of Stroke-Related Impairments on Diabetic
usually avoidable. Most studies28-30 showing that hypoglyce- Management: Special Considerations—Dysphagia
mia leads to brain injury independent of ischemia have been Dysphagia leads to dietary restrictions; however, it is possi-
conducted in animals; however, a recent study31 in humans ble to provide a pureed or soft diet that also limits fat and
demonstrated axonal injury postmortem in 13 patients with carbohydrate load. Often, institutions will serve mashed pota-

Arch Phys Med Rehabil Vol 86, December 2005


HYPERGLYCEMIA IN REHABILITATION, Golden 2381

toes and gravy for a pureed diet. This is high in both fats and the meal, once it is clear how much actual carbohydrate the
carbohydrates. Healthier alternatives may include low fat pro- patient consumed.
teins like tuna salad, ground turkey, or scrambled eggs. Most
vegetables can be pureed as well. Low-sugar puddings and Impact of Stroke-Related Impairments on Diabetic
yogurts are widely available. Management: Special Considerations—Hemiparesis and
Fluid restrictions require particularly close monitoring in
patients with diabetes, who are more prone to dehydration and Ataxia and Visual Impairment Impact on
renal insufficiency. Patients who require thickened liquids may Self-Management
require intravenous or percutaneous endoscopic gastrostomy Diabetes self-management education is an essential compo-
(PEG) tube supplementation of fluids if they cannot maintain nent of the care of patients with diabetes prior to discharge.8 In
adequate oral hydration. Chemistry panels should be monitored its position statement on diabetes education for people with
frequently, and daily intake and output should also be tracked. disabilities, The American Association of Diabetes Educators36
In patients with underlying renal insufficiency who are on emphasizes that disability does not necessarily preclude pa-
long-term management with thickened enteral supplementa- tients from managing diabetes effectively; therefore, persons
tion, scheduled hydration via PEG tube should be considered. with disabilities should be provided with reasonable accommo-
In an algorithm suggested by Kerr et al,35 they recommend dations, including appropriate, adaptive diabetes self-manage-
administering 100 to 200mL of water over 2 hours via the PEG ment education and methods.
tube after each bolus feeding to maintain adequate hydration. Hemiparesis and ataxia. Hand function may be impaired
In addition to adjusting hydration requirements, insulin ther- either from hemiparesis or ataxia. This makes self-administra-
apy may need to be adjusted during trials of thickened liquids. tion of insulin difficult. For these patients devices are available
The total carbohydrate content of the meal determines the that secure the insulin vial to allow syringe loading with 1
amount of insulin coverage; therefore, the carbohydrate content hand. This type of device has also been described as feasible to
of a thickening agent is considered along with the carbohydrate construct and assemble during inpatient rehabilitation, gener-
content of the liquid. ally by occupational therapists, and taught by occupational
Patients who are unable to swallow at all and require enteral therapists and nursing staff.37-39 Using these devices, hemiple-
feedings pose a particular challenge. However, in some ways, gic patients have been reported to become independent in
these patients may be easier to manage, as their intake is highly accurate insulin self-administration within a few days of in-
regulated and the exact caloric intake is stable from day to day. struction39,40 and to maintain accurate use over succeeding
When tube feedings are cycled or bolused, the timing of months, with significant improvement in Hb A1c.40
nutritional insulin will need to be adjusted to cover these Limb stabilization devices are available to hold the hand in
enteral cycles. Bolus feedings more approximate natural eating place during testing of blood glucose. Large-button meters,
patterns and can be covered with an injection of short-acting self-activated meters, meters that utilize auto disks rather than
insulin prior to each bolus. Night-time cycled feedings will individual test strips, and meters that do not require a test strip
require higher dose basal insulin coverage overnight but lower can accommodate for impaired manual dexterity.
morning doses. We found 1 study35 that has attempted to create Visual impairment. Visual problems may arise from dip-
an algorithm for determining the insulin dosage necessary to lopia, hemianopsia, or diabetic retinopathy.41 Adaptive blood
cover PEG feedings. In a study by Kerr,35 after patients are glucose monitoring and insulin injection equipment are readily
transferred to their dedicated stroke unit and PEG feedings are available. Rehabilitation providers need to be aware of these
commenced, patients are started on day 1 on a 22-hour con- adaptations. Recommended insulin self-administration accom-
tinuous feed and their glycemic control is managed with intra- modations for visual impairment include: devices that guide the
venous insulin, with blood glucose readings being checked needle accurately into the insulin vial, syringe-loading magni-
every 2 hours. This is maintained for 24 to 48 hours to deter- fier devices for reading the calibrations on the needle; devices
mine the patient’s total daily insulin requirement. Patients are with an auditory signal for each increment of insulin being
then changed to three 6-hourly feeds and are treated with basal drawn into the needle; and premixed and predrawn insulin
insulin in the form of NPH at 2200 and prandial insulin prior devices and pens.
to each feed.35 The insulin dose is based on the requirements Adaptations to blood glucose monitoring systems for pa-
in the preceding 24 to 48 hours. Finally, patients are transi- tients with visual impairment include tactually identifiable con-
tioned to three 4-hourly feeds with similar insulin coverage. trol buttons, large print displays, and talking monitors for
Using this method in 332 consecutive patients, they achieved a step-by-step instruction and reading of results. Methods for
mean fasting blood glucose of 158mg/dL with stable insulin guiding placement of blood on the test strip include curved test
requirements on day 1 (22-h feed; 43⫾5.4U/h) and on day 7 strips to guide the finger to the target and auditory signals that
(bolus feeds; 47⫾4.4U/h).35 While they did not perform this indicate that a sufficient amount of blood has been placed on
method for patients maintained only on overnight tube feed- the test strip. A review and product evaluation of accessibility
ings, presumably determining the insulin requirement on over- of 17 blood glucose monitoring systems, with product evalua-
night tube feedings by maintaining an insulin drip during the tion of available features for visual impairment, has been
initial feeding period might provide a means for determining published by Uslan et al.42
insulin requirements in this setting as well. Cognitive and mood impairment. Cognitive and commu-
Finally, for patients who are not consuming full meals by nication deficits can have a significant impact on diabetes
mouth and are instead “grazing” on frequent, small meals, we self-management in stroke patients. Speech therapists and re-
would recommend a regimen of carbohydrate counting to de- habilitation psychologists must assess the degree of deficits and
termine the amount of nutritional insulin required. In this the potential for new learning. In some cases, the stroke patient
manner, patients will receive enough coverage to prevent hy- may no longer be capable of self-management, and family
perglycemia. In this setting, the rapid-acting analogs, such as training will be needed.
lispro or aspart, are preferred, as their half-lives are short and Clinically significant depressive disorder is present in an
they are associated with less risk of hypoglycemia. Because of estimated 15% of persons with diabetes and is associated with
their rapid onset of action, they can also be administered after poorer glycemic control and nonadherence to diabetes self-

Arch Phys Med Rehabil Vol 86, December 2005


2382 HYPERGLYCEMIA IN REHABILITATION, Golden

care.43,44 Poststroke depression occurs in up to two thirds of lifespan who are at low risk of developing long-term hyper-
patients and is also associated with poorer stroke recovery.45,46 glycemia-related complications of diabetes, and patients who
The resulting high risk of depression in patients with diabetes lack adequate out-of-hospital support to assist with in-home
following stroke necessitates evaluation and appropriate man- diabetes monitoring and management, if needed. In these set-
agement. Choice of effective treatments, including cognitive- tings, discharging patients with prefilled insulin syringes and/or
behavioral psychotherapy, antidepressant medications, psycho- insulin pens of premixed insulin (70/30) may be preferable.
therapy, or a combination may be determined based on While this type of regimen limits the ability to individually
depression severity and symptom profile. When antidepres- adjust the nutritional and basal insulin, it may provide a safer
sants are used, selection of agent is based on factors including alternative.
coexisting medical conditions, drug interactions, and side ef-
fects. It is important to consider side effects of tricyclics prior Areas of Future Research
to prescribing them for the patient with diabetes, including
potential for weight gain, potential for hyperglycemia unrelated Our literature review has identified several areas of limita-
to weight gain (nortiptyline), anticholinergic side effects, and tion that need to be addressed in the future by well-conducted
orthostatic hypotension.44 Selective serotonin reuptake inhibi- cohort studies and clinical trials of diabetes management in the
tors have not been associated with weight gain or hyperglyce- acute and rehabilitation phases of stroke recovery. Clinical
mia, and fluoxetine has been reported to improve glycemic trials are needed to study the use of intensive insulin therapy in
control47; however side effects of gastrointestinal distress, and the acute stroke setting and inpatient rehabilitation setting to
sexual dysfunction, which may complicate diabetes-related determine more systematically the extent to which better gly-
gastric and sexual complications, must be considered.44 cemic control in patients with diabetes may be associated with
improvement in clinical outcomes (eg, medical morbidity and
DISCUSSION mortality during hospitalization, stroke recurrence, develop-
ment of further diabetes complications postdischarge), rate of
The current literature supports an association between hy- functional recovery, and of rehabilitation LOS. Development
perglycemia and poor clinical outcomes in the acute stroke of additional methods are required for determining the amount
setting. The literature is less clear about the role of tight and type of insulin therapy needed to cover tube feedings in
glycemic control in the poststroke rehabilitation setting in poststroke patients. Finally, well-done observational studies of
improving long-term clinical outcomes. The inpatient rehabil- stroke and rehabilitation patients with diabetes are required to
itation setting, however, represents a window of opportunity to determine the frequency of and risk factors for hypoglycemia
tighten glycemic control and to initiate diabetes self-manage- in this population.
ment education prior to discharge so that patients can maintain
tighter control after discharge. While there are not specific
inpatient rehabilitation data to support stringent glycemic tar- CONCLUSIONS
gets following discharge from rehabilitation in poststroke pa- Intensive inpatient control is important in improving post-
tients, we know that poor glycemic control is associated with stroke outcomes and preventing complications in the inpatient
an increased risk of stroke. In our recent meta-analysis of 3 rehabilitation setting. Adaptive devices enable patients to
community-based cohort studies,48 there was a 17% increased achieve this, even those with physical limitations. Glucose
risk of stroke for every 1-percentage point increase in glycosy- management can be challenging in these complex patients;
lated hemoglobin (a measure on long-term glycemic control). therefore, the physiatrist should enlist the help of an endocri-
There are circumstances, however, where glycemic targets nologist and/or diabetes management team. In the future, op-
may need to be relaxed at the time of discharge to prevent timal inpatient management of diabetes will likely require more
complications from intensive therapy, namely, hypoglycemia. systems approaches, including diabetes management teams and
These include patients with a poor prognosis and shortened development of clinical management algorithms.8

APPENDIX 1: INPATIENT DIABETES MANAGEMENT CLINICAL PRACTICE GUIDELINES

ADA8 AACE49

Target plasma glucose levels: Intensive care unit:


⬍110mg/dL preprandial 110mg/dL
⬍180mg/dL peak postprandial
Intensive insulin therapy: Noncritical care units:
Maintain blood glucose 80⫺110mg/dL to reduce 110mg/dL preprandial
morbidity and mortality in surgical intensive care unit 180mg/dL maximal glucose
Intravenous insulin infusion: Values ⬎180mg/dL are an indication to monitor glucose
Post myocardial infarction (followed by multidose levels more frequently to determine the direction
injection) of glucose trend and determine the need for more
Major surgery, hemodynamic instability, nothing by intensive intervention
mouth status, poorly controlled patients with
fluctuating sugars, severe insulin deficiency or
resistance

Arch Phys Med Rehabil Vol 86, December 2005


HYPERGLYCEMIA IN REHABILITATION, Golden 2383

APPENDIX 2: ALGORITHM FOR DETERMINING A PATIENT’S INSULIN REQUIREMENT8

Calculate the total daily dose (TDD) of insulin—multiply body weight in kilograms by:
● 0.5 to 0.7U if patient has type 1 diabetes mellitus
● 0.4 to 1.0U (or more) if patient has type 2 diabetes mellitus
Determine the basal insulin requirement:
● 40% to 50% of TDD
● Administer once daily in the morning or at bedtime using a long-acting insulin analog (glargine) or administer using
an intermediate-acting insulin (neutral protamine Hagedorn) twice daily or at bedtime
Determine nutritional insulin requirement:
● TDD minus basal insulin dose
● Split the dose to cover meals
If using lispro or aspart, divide dose into 3 injections (prior to breakfast, lunch, and dinner)
If using regular insulin divide into 2 injections prior to breakfast and dinner

15. Williams LS, Rotich J, Qi R, et al. Effects of admission hyper-


References glycemia on mortality and costs in acute ischemic stroke. Neurol-
1. Fietsam R Jr, Bassett J, Glover JL. Complications of coronary ogy 2002;59:67-71.
artery surgery in diabetic patients. Am Surg 1991;57:551-7. 16. Kiers L, Davis SM, Larkins R, et al. Stroke topography and
2. Zerr KJ, Furnary AP, Grunkemeier GL, Bookin S, Kanhere V, outcome in relation to hyperglycaemia and diabetes. J Neurol
Starr A. Glucose control lowers the risk of wound infection in Neurosurg Psychiatry 1992;55:263-70.
diabetics after open heart operations. Ann Thorac Surg 1997;63: 17. Bruno A, Biller J, Adams HP Jr, et al. Acute blood glucose level
356-61. and outcome from ischemic stroke. Trial of ORG 10172 in Acute
3. Golden SH, Peart-Vigilance C, Kao WH, Brancati FL. Perioper- Stroke Treatment (TOAST) Investigators. Neurology 1999;52:
ative glycemic control and the risk of infectious complications in 280-4.
a cohort of adults with diabetes. Diabetes Care 1999;22:1408-14. 18. Demchuk AM, Morgenstern LB, Krieger DW, et al. Serum glu-
4. Hjortrup A, Sorensen C, Dyremose E, Hjortso NC, Kehlet H. cose level and diabetes predict tissue plasminogen activator-
Influence of diabetes mellitus on operative risk. Br J Surg 1985; related intracerebral hemorrhage in acute ischemic stroke. Stroke
72:783-5. 1999;30:34-9.
5. Umpierrez GE, Isaacs SD, Bazargan N, You X, Thaler LM, 19. Parsons MW, Barber PA, Desmond PM, et al. Acute hyperglyce-
Kitabchi AE. Hyperglycemia: an independent marker of in- mia adversely affects stroke outcome: a magnetic resonance im-
hospital mortality in patients with undiagnosed diabetes. J Clin aging and spectroscopy study. Ann Neurol 2002;52:20-8.
Endocrinol Metab 2002;87:978-82. 20. Malmberg K. Prospective randomised study of intensive insulin
6. Gray CS, Scott JF, French JM, Alberti KG, O’Connell JE. Prev- treatment on long term survival after acute myocardial infarction
alence and prediction of unrecognised diabetes mellitus and im- in patients with diabetes mellitus. DIGAMI (Diabetes Mellitus,
paired glucose tolerance following acute stroke. Age Ageing Insulin Glucose Infusion in Acute Myocardial Infarction) Study
2004;33:71-7. Group. BMJ 1997;314:1512-5.
7. Teasell R, Foley N, Doherty T, Finestone H. Clinical character- 21. van den Berghe G, Wouters P, Weekers F, et al. Intensive insulin
istics of patients with brainstem strokes admitted to a rehabilita- therapy in the critically ill patients. N Engl J Med 2001;345:
tion unit. Arch Phys Med Rehabil 2002;83:1013-6. 1359-67.
8. Clement S, Braithwaite SS, Magee MF, et al. Management of 22. Lazar HL, Chipkin SR, Fitzgerald CA, Bao Y, Cabral H, Apstein
diabetes and hyperglycemia in hospitals. Diabetes Care 2004;27: CS. Tight glycemic control in diabetic coronary artery bypass
553-91. graft patients improves perioperative outcomes and decreases
9. Lithner F, Asplund K, Eriksson S, Hagg E, Strand T, Wester PO. recurrent ischemic events. Circulation 2004;109:1497-502.
Clinical characteristics in diabetic stroke patients. Diabetes Metab 23. Scott JF, Robinson GM, French JM, O’Connell JE, Alberti KG,
1988;14:15-9. Gray CS. Glucose potassium insulin infusions in the treatment of
10. Jorgensen H, Nakayama H, Raaschou HO, Olsen TS. Stroke in acute stroke patients with mild to moderate hyperglycemia: the
patients with diabetes. The Copenhagen Stroke Study. Stroke Glucose Insulin in Stroke Trial (GIST). Stroke 1999;30:793-9.
1994;25:1977-84. 24. Koproski J, Pretto Z, Poretsky L. Effects of an intervention by a
diabetes team in hospitalized patients with diabetes. Diabetes Care
11. Kong KH, Young S. Incidence and outcome of poststroke urinary
1997;20:1553-5.
retention: a prospective study. Arch Phys Med Rehabil 2000;81:
25. Levetan CS, Salas JR, Wilets IF, Zumoff B. Impact of endocrine
1464-7.
and diabetes team consultation on hospital length of stay for
12. Pulsinelli WA, Levy DE, Sigsbee B, Scherer P, Plum F. Increased patients with diabetes. Am J Med 1995;99:22-8.
damage after ischemic stroke in patients with hyperglycemia with 26. American Diabetes Association. Standards of medical care in
or without established diabetes mellitus. Am J Med 1983;74: diabetes. Diabetes Care 2005;28:S4-36.
540-4. 27. Queale WS, Seidler AJ, Brancati FL. Glycemic control and sliding
13. Desmond DW, Moroney JT, Sano M, Stern Y. Recovery of scale insulin use in medical inpatients with diabetes mellitus. Arch
cognitive function after stroke. Stroke 1996;27:1798-803. Intern Med 1997;157:545-52.
14. Capes SE, Hunt D, Malmberg K, Pathak P, Gerstein HC. Stress 28. Agardh CD, Kalimo H, Olsson Y, Siesjo BK. Hypoglycemic brain
hyperglycemia and prognosis of stroke in nondiabetic and diabetic injury. I. Metabolic and light microscopic findings in rat cerebral
patients: a systematic overview. Stroke 2001;32:2426-32. cortex during profound insulin-induced hypoglycemia and in the

Arch Phys Med Rehabil Vol 86, December 2005


2384 HYPERGLYCEMIA IN REHABILITATION, Golden

recovery period following glucose administration. Acta Neuropathol a patient with hemiplegia. Diabetes Technol Ther 2004;6:
(Berl) 1980;50:31-41. 505-9.
29. Auer RN, Olsson Y, Siesjo BK. Hypoglycemic brain injury in the 41. Park WL, Mayer RS, Moghimi C, Park JM, Deremeik JT. Rehabil-
rat. Correlation of density of brain damage with the EEG isoelec- itation of hospital patients with visual impairments and disabilities
tric time: a quantitative study. Diabetes 1984;33:1090-8. from systemic illness. Arch Phys Med Rehabil 2005;86:79-81.
30. Siemkowicz E, Hansen AJ. Clinical restitution following cerebral 42. Uslan MM, Eghtesadi K, Burton D. Accessibility of blood glucose
ischemia in hypo-, normo- and hyperglycemic rats. Acta Neurol monitoring systems for blind and visually impaired people. Dia-
Scand 1978;58:1-8. betes Technol Ther 2003;5:439-48.
31. Dolinak D, Smith C, Graham DI. Hypoglycaemia is a cause of 43. Anderson RJ, Freedland KE, Clouse RE, Lustman PJ. The prev-
axonal injury. Neuropathol Appl Neurobiol 2000;26:448-53. alence of comorbid depression in adults with diabetes: a meta-
32. Frier BM. Hypoglycaemia and cognitive function in diabetes. Int analysis. Diabetes Care 2001;24:1069-78.
J Clin Pract Suppl 2001;30-7. 44. Lustman PJ, Singh PK, Clouse RE. Recognizing and managing
33. Strachan MW, Deary IJ, Ewing FM, Frier BM. Recovery of depression in patients with diabetes. In: Anderson BJ, Rubin RR,
cognitive function and mood after severe hypoglycemia in adults editors. Practical psychology for diabetes clinicians. 2nd ed. Al-
with insulin-treated diabetes. Diabetes Care 2000;23:305-12. exandria: American Diabetes Association; 2002. p 229-38.
34. Gray CS, Hildreth AJ, Alberti GK, O’Connell JE. Poststroke 45. Verdelho A, Henon H, Lebert F, Pasquier F, Leys D. Depressive
hyperglycemia: natural history and immediate management. symptoms after stroke and relationship with dementia: a three-
Stroke 2004;35:122-6. year follow-up study. Neurology 2004;62:905-11.
35. Kerr D, Hamilton P, Cavan DA. Preventing glycaemic excursions 46. Singh A, Black SE, Herrmann N, et al. Functional and neuroana-
in diabetic patients requiring percutaneous endoscopic gastros- tomic correlations in poststroke depression: the Sunnybrook
tomy (PEG) feeding after a stroke. Diabet Med 2002;19:1006-8. Stroke Study. Stroke 2000;31:637-44.
36. American Association of Diabetes Educators. The core curriculum 47. Goodnick PJ. Use of antidepressants in treatment of comorbid
for diabetes education. 5th ed. Chicago: AADE; 2003. diabetes mellitus and depression as well as in diabetic neuropathy.
37. Holliman K. Another device for one-handed insulin management. Ann Clin Psychiatry 2001;13:31-41.
Am J Occup Ther 1979;33:393. 48. Selvin E, Marinopoulos S, Berkenblit G, et al. Meta-analysis:
38. Keon HM, Hanna AK. Self-administration of insulin by a hemi- glycosylated hemoglobin and cardiovascular disease in diabetes
plegic individual [letter]. Diabetes Care 1980;3:705. mellitus. Ann Intern Med 2004;141:421-31.
39. Harvey RL, Stachowski KM, Dewulf SK. Independent insulin 49. Garber AJ, Moghissi ES, Bransome ED Jr, et al. American Col-
administration by the hemiplegic patient: stabilization of an lege of Endocrinology position statement on inpatient diabetes and
insulin pen with a new device. Arch Phys Med Rehabil 1992; metabolic control. Endocrine Pract 2004;10:77-82.
73:779-81. 50. Golden SH. Treatment of the metabolic syndrome. In: Clinical
40. Sohmiya M, Kanazawa I, Inomata N, et al. A new device to endocrinology update 2004 syllabus; 2004 Oct 3-6; Baltimore
introduce self-injection of insulin by his non-dominant hand in (MD). p 42. Chevy Chase: Endocrine Society Pr; 2004.

Arch Phys Med Rehabil Vol 86, December 2005

View publication stats

You might also like