Incidence of and Risk Factors for Medical Complications During Stroke Rehabilitation
Background and Purpose—The aims of this study were to examine the frequency, types, and clinical factors associated with medical complications that occur during inpatient rehabilitation and to identify risk factors for complications that require a transfer to an acute care facility.
Methods—A cohort of 1029 patients consecutively admitted for inpatient stroke rehabilitation was studied. Demographic and stroke information, impairment, preexisting medical conditions, and admission laboratory abnormalities were recorded. Medical complications, defined as new or exacerbated medical problems, were documented for each patient. Complications that required transfer off rehabilitation were noted. Univariate and multiple logistic regression analyses were used to determine factors that were associated with risk of medical complications and risk of transfer off rehabilitation.
Results—Seventy-five percent of patients experienced ≥1 medical complication during rehabilitation. Significant factors for the development of any medical complication included greater neurological deficit (odds ratio [OR], 4.10; confidence interval [CI], 1.88 to 8.91), hypoalbuminemia (OR, 1.71; 95% CI, 1.15 to 2.52), and history of hypertension (OR, 1.81; 95% CI, 1.27 to 2.59). Nineteen percent of patients had a medical complication that required transfer to an acute care facility. Significant factors for transfers were elevated admission white blood cell counts (OR, 1.92; 95% CI, 1.32 to 2.79), low admission hemoglobin levels (OR, 1.89; 95% CI, 1.32 to 2.68), greater neurological deficit (OR, 2.46; 95% CI, 1.37 to 4.39), and a history of cardiac arrhythmia (OR, 1.79; 95% CI, 1.18 to 2.67).
Conclusions—Medical complications are common among patients undergoing stroke rehabilitation. A significant number of these medical complications may require a transfer to an acute facility.
The prevention, recognition, and management of medical conditions after stroke constitute important activities during and after comprehensive inpatient rehabilitation, but systematic investigation of both the incidence of and risk factors for these medical problems has been limited.
The frequency of medical complications during inpatient rehabilitation among patients with stroke is difficult to determine, but estimates have been reported to range between 48%1 and 96%,2 depending on criteria for defining complications, method of investigation, and specific patient group studied. The most common types of medical conditions that occur are urinary tract infections, venous thrombi, pneumonias, joint and soft tissue pain, sepsis, and falls.1 2 3 4 5 6 7 8
Identifying clinical factors that are associated with increased risk of experiencing these complications is valuable to facilitate the implementation of appropriate prevention and management interventions. Risk factors for medical complications in stroke rehabilitation have not been well studied to date, but limited and somewhat contradictory evidence suggests that they include advancing age, severity of neurological deficits and disability caused by the stroke, and certain preexisting medical conditions.1 2 3 4 6 7 8 9 10
The purposes of the present study were to (1) examine prospectively the frequencies and specific types of medical complications that occur during inpatient rehabilitation, (2) determine which medical complications during rehabilitation result in a transfer to an acute care facility, and (3) determine risk factors for these problems in a large and heterogeneous sample of stroke patients.
Subjects and Methods
A total of 1029 patients with a primary diagnosis of stroke who were consecutively admitted between December 1993 and August 1997 to a freestanding urban academic specialty rehabilitation hospital for initial comprehensive acute inpatient stroke rehabilitation were studied. Included were patients who were ≥18 years of age, whose stroke occurred within the prior 3 months, whose primary reason for admission to the rehabilitation program was disability resulting from stroke, and who stayed in inpatient rehabilitation for ≥3 days. Stroke, defined as an acute event of cerebrovascular origin causing focal or global neurological dysfunction lasting >24 hours, was confirmed by both clinical and radiographic means.
On admission to the rehabilitation facility, age, sex, ethnicity, and marital status were documented for each patient, as were type, depth, and side of the cerebral lesion. Type of stroke was defined as either hemorrhagic or ischemic. Hemorrhagic strokes included both intracerebral and subarachnoid hemorrhages. Depth of stroke was defined as either cortical or subcortical. Brain stem strokes were included in the subcortical group. Side of lesion was defined as right, left, or bilateral. Duration between stroke onset and rehabilitation admission and rehabilitation hospital length of stay was recorded for each patient.
Each patient was rated on admission to rehabilitation by an attending physician using the National Institutes of Health Stroke Scale (NIHSS), a well-validated instrument used to assess impairment levels in 15 neurological functions frequently affected by stroke.11 12 13 Although this scale has been used extensively to assess efficacy of pharmacological interventions in acute stroke management trials,14 15 it has rarely and only recently been applied in the stroke rehabilitation setting.16 17 18 In this study, raw NIHSS scores were categorized into 4 groups according to severity of neurological deficit: mild (0 to 5), moderate (6 to 10), moderate to severe (11 to 15), and severe (≥16) impairment.
Using acute hospital medical records and rehabilitation admission patient interview, physicians identified for each patient those preexisting medical conditions that were present before the stroke; they also identified medical complications that occurred during the acute poststroke hospital stay before the rehabilitation admission using a standardized list of 124 medical diagnoses. Statistical analyses were limited to the 26 medical conditions that were present in >5% of the sample.
An experienced research assistant reviewed each patient’s medical record and documented values for admission body temperature, initial systolic and diastolic blood pressures, and 7 medical laboratory tests performed on most patients at rehabilitation admission. Seven laboratory tests assessed level of hemoglobin, creatinine, serum albumin, serum sodium, potassium, and bicarbonate, and white blood cell count.
The investigators continuously and prospectively monitored the occurrence of medical complications during rehabilitation. A standardized medical complication event guide was used to record these complications, which were defined as new or exacerbated medical problems documented by appropriate history, physical examination, laboratory, radiographic, and/or other clinical studies that generated additional physician evaluation, a change in medication, or additional medical interventions. These additional medical interventions occurred either at the rehabilitation facility or, less commonly, during transfer to an acute care facility. Those medical complications that caused a transfer to an acute care facility were noted.
These 83 possible complications included but were not limited to pulmonary, cardiac, gastrointestinal, genitourinary, skin, musculoskeletal/rheumatologic, endocrine, neurological, vascular, hematologic/oncological, trauma, and psychiatric problems. Depression was defined as dysphoria requiring either psychotherapy or pharmacological intervention. Specific and consistent established criteria for each medical problem were applied to the assessments to determine whether a medical complication had occurred. Recurrent events, such as adjustments of antihypertensive medication or insulin dosages, were counted as a single medical management problem. The occurrence of any medical complication in rehabilitation and transfers to an acute care facility were considered as single dichotomous (presence/absence) variables.
A research assistant reviewed each patient’s medical record for the occurrence of any of these medical complications. To maximize the consistency and uniformity with which the criteria for complications were applied, only one research assistant reviewed all medical records for complications. For situations in which the occurrence or specific types of complications were in doubt, a physician reviewed the complications.
The frequencies of each of the preexisting medical conditions and laboratory abnormalities were tabulated. The incidence of medical complications that developed during rehabilitation and the incidence of those medical complications that resulted in a transfer to an acute facility were determined. Many patients had >1 complication. Because of the diverse nature and incidence of medical complications occurring during rehabilitation, the presence or absence of any medical complication and the presence or absence of any transfer were used in subsequent analyses.
When appropriate, univariate statistics (χ2 and t tests) were used to determine the relationships between each of the demographic and stroke characteristics, neurological impairment data, preexisting medical conditions, and admission laboratory abnormalities with the occurrence of any medical complication and the occurrence of any transfer to an acute care facility.
Multiple logistic regression analyses were conducted with the occurrence of any medical complication during rehabilitation and the occurrence of any transfer as the dependent variables. Independent variables consisted of only those factors that were significantly correlated with the dependent variables on univariate analysis.
Variables with a value of P<0.01 on univariate analysis were entered into the model. This relatively rigorous significance level of 0.01 was selected as the criterion for inclusion to account for the effect of random events occurring in a large sample or in a series of tests using multiple variables. The level of significance for independent factors on multivariate was set at 0.01. All analyses were performed with SPSS version 8.0 for Windows.
Additional analysis was conducted to determine the frequency of medical complications and the specific complications that were most common for each of the 4 impairment severity groups.
A description of the demographic and stroke characteristics of the 1029 patients in the study sample is provided in Table 1⇓. Strokes were predominantly ischemic in origin, and they more commonly involved the left hemisphere and cortical region. The distribution of the 4 neurological impairment severity levels on admission to rehabilitation, as assessed by the raw NIHSS score, was as follows: 282 patients (28%) with mild impairment, 390 patients (38%) with moderate impairment, 213 patients (21%) with moderate to severe impairment, and 131 patients (13%) with severe impairment. Mean onset-to-admission interval was 17.4±14.9 days. Mean length of rehabilitation stay was 28.0±13.8 days.
Frequencies of preexisting medical conditions among the patients are displayed in Table 2⇓, which demonstrates that hypertension (74%), history of smoking (53%), coronary artery disease (29%), diabetes mellitus (28%), and presence of an indwelling urethral catheter (24%) were the most common preexisting conditions among the patients in the sample. The rate of abnormal laboratory measures was 60.2% for hypoalbuminemia (n=709), 52.6% (n=1009) for low levels of hemoglobin, 19.7% for elevated white blood cell count (n=1003), 14.5% for bicarbonate (n=1005), 11.6% for sodium (n=1014), 11.0% for potassium (n=1014), and 9.1% for elevated creatinine (n=1009).
Of the 1029 total patients, 773 patients (75%) experienced ≥1 medical complication during rehabilitation. The types and frequencies of the 27 most common or most clinically significant medical complications that occurred during rehabilitation are listed in Table 3⇓. This table shows that urinary tract infection (31%), joint and soft tissue pain (14%), depression (13%), falls (11%), and acute blood pressure elevations (9%) were the most common medical problems occurring in these patients. There were 3 deaths.
One hundred ninety-seven patients (19%) had a complication that required a transfer to an acute care facility for diagnosis and/or treatment. Table 3⇑ shows that deep venous thrombosis, pneumonia, angina, and atrial arrhythmias resulted in the majority of transfers to an acute hospital.
Table 4⇓ reports those demographic and stroke characteristics, neurological impairment data, preexisting medical conditions, laboratory abnormalities, and other factors that were significantly related to the occurrence of medical complications during rehabilitation and the risk of a complication requiring a transfer. Thirteen factors were found to be significantly related to risk of complication, of which the strongest were greater neurological deficit level as measured by NIHSS; feeding tube; indwelling urethral catheter; hypoalbuminemia; tracheostomy; history of a pressure ulcer; elevated white blood cell count; abnormal serum electrolyte levels; history of hypertension; history of renal failure; elevated serum creatinine levels; anemia; and longer onset-to-admission interval.
Eight factors were significantly related to risk of a complication requiring a transfer to an acute facility. These were greater neurological impairment; presence of a feeding tube; elevated white blood cell count; abnormal admission hemoglobin levels; presence of an indwelling urethral catheter; history of pneumonia; and a history of a cardiac arrhythmia.
The results of the multivariate logistic regression analysis with the occurrence of any medical complication during rehabilitation as the dependent variable are shown in Table 5⇓. Only those factors significantly related to complication occurrence through univariate analysis were entered into the multivariate analysis. Only 3 of the factors were independently related to the occurrence of a medical complication: more severe neurological deficits, hypoalbuminemia, and a history of hypertension.
Results of the multivariate logistic regression analysis with the occurrence of a transfer during rehabilitation as the dependent variable are shown in Table 6⇓. Four factors were found to be independent factors for the occurrence of a transfer during rehabilitation to an acute care facility. These were an elevated white blood cell count and low hemoglobin levels on admission to rehabilitation, greater impairment, and a history of a cardiac arrhythmia.
Because neurological impairment level was the strongest factor predicting the occurrence of medical complications, additional analysis was conducted to determine the frequency of medical complications and the specific complications that were most common for each of the 4 impairment severity groups. Table 7⇓ illustrates that the likelihood of complications increases as severity of neurological deficit worsens. Urinary tract infection was the most common complication for all 4 severity groups; both joint/soft tissue pain and depression also were common, but the relative frequencies of each medical condition varied considerably with the severity of the impairment level. For example, patients with the most severe impairment level were found to be ≈4 times more likely to have urinary tract infections as were patients with the least level of impairment. Also, the more severe, medically intensive, or life-threatening complications were considerably more common among patients with the most severe neurological deficits. These also are the types of complications that are likely to interrupt rehabilitation by requiring a transfer to an acute care hospital.
In addition to the residual neurological deficits that result from a stroke, a variety of medical conditions often occur, many of which can hinder rehabilitation efforts and ultimately impede the stroke survivor’s ability to resume an active lifestyle. This study confirms that patients who sustain a stroke have a high frequency of preexisting medical conditions.9 10 19 Interestingly, the 4 preexisting conditions that were determined to be the most common among the patients in this study also are important risk factors for stroke (hypertension, smoking history, coronary artery disease, and diabetes mellitus).20 21 22 The relatively large numbers of indwelling urethral catheters and feeding tubes also are noteworthy because of their potential to contribute directly to, or to be otherwise associated with, increased utilization of resources during stroke rehabilitation.18 Likewise, the finding that nearly two thirds of those who were tested for serum albumin level had hypoalbuminemia also is significant, given the past association of low serum albumin levels and medical conditions during rehabilitation.5
It also is striking to note that 3 of every 4 patients admitted to an inpatient stroke rehabilitation program might experience a medical complication and that many of these patients experience >1 medical condition. Although many of these problems may not be life threatening (such as most urinary tract infections, joint/soft tissue pain, and depression), some complications can be serious (such as pneumonia and venous thromboembolism) and may require a transfer to an acute care facility for treatment. Nineteen percent of patients in this study had complications that required a transfer. This underscores the common clinical observation that patients who undergo rehabilitation often have significant complex medical issues.
Specific characteristics were associated with a greater likelihood that a medical condition would complicate the rehabilitation program and that a complication would require a transfer to an acute care facility. Of the 45 potential factors evaluated, only 3 were associated with the occurrence of any complication, and 4 factors were associated with the occurrence of a complication that required a transfer to an acute care facility after adjustment for other factors.
The severity of the neurological deficit on admission to rehabilitation was the strongest predictor of both any medical complications and of complications that required a transfer. Patients with the greatest degree of impairment experienced complications most frequently. This finding is consistent with the observations of Dromerick and Reding,2 who used a 3-level impairment scale to show that severity of stroke was related to the likelihood and number of complications in 100 stroke rehabilitation patients. In another study of 245 stroke patients, Kalra and colleagues7 reported that complications were seen most frequently among patients with severe neurological deficits and poor prognosis based on the Orpington Prognostic Score. These results reflect the increased overall complexity of patients with more severe impairments.
Patients with the most severe neurological impairments experienced the most serious complications, such as deep vein thrombosis, pneumonia, pressure ulcers, and gastrointestinal hemorrhage. These were the types of complications that frequently required a transfer to an acute care facility. Once again, this finding reflects the close association between severity of stroke and severity of medical illness. The frequency of specific complications as a function of neurological impairment level has not been observed in the past, and knowledge of this association may be valuable in assisting clinicians in targeting preventive and treatment methods for specific patients.
Risk factors for medical complications that have been reported in the past include implementation of the acute care prospective payment system1 ; severity of stroke1 2 ; admission disability level2 ; length of rehabilitation stay2 ; low serum albumin level5 ; prestroke disability8 ; location of stroke in the anterior cerebral circulation region8 ; and urinary incontinence.8 Factors that have been found to be unrelated to complications during stroke rehabilitation include interval between stroke onset and rehabilitation admission2 and type of stroke (ischemic versus hemorrhagic).2 Advanced age was found to be predictive of complications by Davenport et al8 but unrelated to complications by Dromerick and Reding.2 Findings of the present study confirmed many of these results and extended available information on significant markers for medical illness.
Few studies have examined the risk factors for acute transfers during rehabilitation. The decision to transfer a patient to an acute care facility depends on a number of circumstances that may vary widely among rehabilitation programs. These include the type and severity of the complication, the ability of the rehabilitation nursing and medical staff to manage the complication, and the patient’s ability to continue to participate in the rehabilitation program while undergoing treatment for the complication. One study examined risk factors for acute care transfers among patients in a traumatic brain injury rehabilitation program and found that both a history of pneumonia and a history of recent surgery were associated with risk of acute care transfers. Low levels of hemoglobin, although not statistically significant, showed a trend toward association in that study.23
Even though all medical complications may be preventable to some extent, implementation of certain interventions during acute care and rehabilitation may help to reduce the frequencies of certain specific complications. Examples include the early identification of dysphagia to prevent aspiration pneumonia, initiation of anticoagulation therapy to prevent thromboembolism, judicious use of indwelling catheters to prevent urinary tract infections, and frequent body positioning to prevent pressure ulcers. Because certain medical complications that become manifest during rehabilitation originate during the acute poststroke phase, some of these clinical interventions are most effective when started immediately after the stroke.
Complications were recorded only during the inpatient rehabilitation stay, which was of variable duration across the sample. Although this was the most clinically feasible method to examine complication rates and the most clinically useful for attempting to study and enhance in-hospital care, it presents several conceptual problems. In particular, the frequency of medical complications may vary, depending on the period of the observation. It is possible that severe strokes contribute to an increased risk of medical complications and that these complications might prolong inpatient stays. However, it might be possible that longer stays simply increase the opportunity to observe medical complications.
Generalizability of the results of this investigation may be limited by the fact that only one center was studied. However, a multicenter study of medical complications in 311 stroke patients found similar types and frequencies of complications.24 In addition, the sample in this study is the largest group ever used for which medical complications were studied systematically.
Despite these limitations, the potential clinical implications of the study are clear. Knowledge of the severity of the stroke, together with information derived from the medical history and easily obtained laboratory tests, can assist clinicians in identifying patients at higher risk for complications that would interrupt the rehabilitation hospitalization. It also can facilitate cost-effective allocation of medical resources. In particular, patients who are identified as being at high risk for medical complications requiring a transfer to an acute facility might benefit from less intensive rehabilitation until their medical status has stabilized. These patients might require more frequent monitoring of their clinical status during a more intensive rehabilitation program.
This research was supported by the US Department of Education, National Institute on Disability and Rehabilitation Research, grants H133B30024 and H133B980021; through the Rehabilitation Research and Training Center on Enhancing Quality of Life of Stroke Survivors; and by the Rehabilitation Institute of Chicago.
Portions of this work were presented at the North American Stroke Association Annual Scientific Meeting, October 17, 1997, Montreal, Quebec, Canada.
- Received June 26, 2000.
- Revision received November 1, 2000.
- Accepted November 1, 2000.
- Copyright © 2001 by American Heart Association
Dobkin BH. Neuromedical complications in stroke patients transferred for rehabilitation before and after diagnostic related groups. J Neuro Rehab. 1987;1:3–7.
Dromerick A, Reding M. Medical and neurological complications during inpatient stroke rehabilitation. Stroke. 1994;25:358–361.
Roth EJ. Medical complications encountered in stroke rehabilitation. Phys Med Rehabil Clin North Am. 1991;2(3):563–577.
Kalra L, Yu G, Wilson K, Roots P. Medical complications during stroke rehabilitation. Stroke. 1995;26:990–994.
Davenport RJ, Dennis MS, Wellwood I, Warlow CP. Complications after acute stroke. Stroke. 1996;27:415–420.
Roth EJ, Green D. Cardiac complications during inpatient stroke rehabilitation. Top Stroke Rehabil. 1996;3:86–92.
Roth EJ, Mueller K, Green D. Stroke rehabilitation outcome: impact of coronary artery disease. Stroke. 1988;19:42–47.
Brott T, Adams HP, Olinger CP, Marler JR, Barsan WG, Biller J, Spilker, Holleran R, Eberele R, Hertzberg V, Rorick M, Moomaw CJ, Walker M. Measurements of acute cerebral infarction: a clinical examination scale. Stroke. 1989;20:864–870.
Muir KW, Weir CJ, Murray GD, Povey C, Lees KR. Comparison of neurological scales and scoring systems for acute stroke prognosis. Stroke. 1996;27:1817–1820.
Brott TG, Haley EC, Levy DE, Barsan W, Broderick J, Sheppard GL, Spilker J, Kongable GL, Massey S, Reed R, Marler JR. Urgent therapy for stroke, part I: pilot study of tissue plasminogen activator administered within 90 minutes. Stroke. 1992;23:632–640.
Wityk RJ, Pessin MS, Kaplan RF, Caplan LR. Serial assessment of acute stroke using the NIH stroke scale. Stroke. 1994;25:362–365.
Heinemann AW, Harvey RL, McGuire JR, Ingberman D, Lovell L, Semik P, Roth EJ. Measurement properties of the NIH stroke scale during acute rehabilitation. Stroke. 1997;28:1174–1180.
Dyken ML, Wolf PA, Barnett HJM, Bergan JJ, Hass WK, Kannel WB, Kuller L, Kurtzke JF, Sundt TM. Risk factors in stroke: a statement for physicians by the Subcommittee on Risk Factors and Stroke Council. Stroke. 1984;15:1105–1111.
Shinton R, Beevers G. Meta-analysis of relation between cigarette smoking and stroke. BMJ. 1989;298:789–794.
Wolf PA, D’Agostino RB, Belanger AJ, Kannel WB. Probability of stroke: a risk profile from the Framingham study. Stroke. 1991;22:312–318.
Langhorne P, Stott DJ, Roberston L, MacDonald J, Jones L, McAlpine C, Dick F, Taylor GS, Murray G. Medical complications after stroke: a multicenter study. Stroke. 2000;31:1223–1229.