Up to 95% of patients have at least one relevant complication within the first three months after stroke.1 Complications impair neurological outcomes2–4 and approximately one-third of patients with ischaemic stroke die during hospitalisation due to one or more complications.5 To minimise the impact of stroke-associated adverse events, patients should be treated in specialised units where complications are detected earlier compared with general units. In addition, treatment in stroke units improves survival by preventing life-threatening situations.3
However, even in specialised stroke units, stroke-associated infections remain one of the major complications in acute stroke, with frequencies between 21 and 65%.6 The incidence of infection among stroke patients is thus significantly higher than the general prevalence of hospital-acquired infection, which ranges from 6 to 9% in all hospitalised patients.7,8
Bacterial pneumonia and urinary tract infection (UTI) are the predominant infections in acute stroke patients.2,9 The incidence of UTI in acute stroke ranges between 6 and 27%, whereas the frequency of stroke-associated pneumonia (SAP) lies between 5 and 22%.6,10–12 This compares with an average rate of pneumonia of only 3.5% in non-stroke patients treated in a geriatric hospital.13 In the general population, the risk of UTI is between 3 and 10% per day of catheterisation.14,15
Diagnosis of Post-stroke Infections
Post-stroke Pneumonia
Traditionally, the diagnosis of pneumonia requires a combination of clinical assessment, radiological imaging and appropriate microbiological tests; however, a reliable diagnosis of pneumonia remains a medical challenge, even in stroke patients. Unspecific symptoms of respiratory infection are fever or hypothermia, cough, purulent secretion, dyspnoea, myalgia, arthralgia, headache and delirium; therefore, clinical examinations alone do not improve diagnostic properties. Radiological examination and microbiological tests are mandatory to establish the diagnosis of classic (bacterial) pneumonia. This is reasonable because plain chest radiography is an inexpensive test and an important initial examination in all patients with suspected pneumonia.16 However, diagnostic properties are limited even in immobilised stroke patients.17,18
Computed tomography (CT) is a valuable adjunct in negative or non-diagnostic chest radiography, unresolved pneumonias and when complications are suspected.19 However, even in current Centers for Disease Control and Prevention (CDC) guidelines, chest radiography remains the central diagnostic tool.
In stroke patients, we recommend the use of the criteria for a ‘clinically defined pneumonia’ (see Table 1) in which diagnosis is possible without further microbiological tests.
Urinary Tract Infection
It is important to differentiate between UTI and asymptomatic bacteriuria. Asymptomatic bacteriuria (see Table 2) is a common phenomenon in populations with structural or functional abnormalities of the genito-urinary tract, but even healthy individuals frequently have positive urine cultures. Asymptomatic bacteriuria is seldom associated with adverse outcomes.
Symptomatic UTI has been defined by the CDC, as depicted in Table 3. One should note that in stroke patients clinical signs of UTI (e.g. dysuria or urgency) are of only minor diagnostic value, as stroke patients are often unconscious or dysphasic. Thus, laboratory and microbiological testing are of enhanced relevance. Interestingly, diagnosis of UTI can also be a decision of the treating physicians.
Prediction of Post-stroke Infections
Early identification of patients at high risk of post-stroke infection may promote and justify intensive monitoring and tailored anti-infective treatment.
Pneumonia
Post-stroke pneumonia is usually explained as a result of aspiration due to neurological deficits, such as impaired level of consciousness, disturbed protective reflexes20 or dysphagia.21 Additional risk factors for the SAP have been identified: stroke severity, stroke subtype, lesion size, mechanical ventilation, age, gender and history of diabetes.21–24 A study by Walter et al. underlined the pivotal role of post-stroke dysphagia with a 10-fold relative risk of post-stroke pneumonia.23 The first published scoring system to predict post-stroke pneumonia was developed including National Institutes of Health Stroke Scale (NIHSS), age, sex, mechanical ventilation and dysphagia.25 Sellars et al. established a scoring system by identifying age ≥65 years, dysarthria, modified Rankin Scale (mRS) ≥4, Abbreviated Mental Test (AMT) <8 and dysphagia (failed water swallowing test) as independent clinical risk factors for SAP in 412 stroke patients. The presence of two or more of these risk factors carried 90.9% sensitivity and 75.6% specificity for the development of pneumonia in this study.26 However, both scores need further refinement and prospective validation in large multicentre trials.
The impact of aspiration on the risk of post-stroke pneumonia is indisputable. However, aspiration alone is not sufficient to explain the high incidence of pneumonia in acute stroke, as about 50% of healthy subjects also aspirate pharyngeal secrets every night to a similar extent to stroke patients without developing pneumonia.27,28 Clinical and experimental evidence suggests that acute ischaemic central nervous system (CNS) injury is associated with temporary immunodeficiency and that an impaired antibacterial host defence is an important factor for the increased susceptibility to infection after CNS injury.6,10–12,29
In experimental stroke models, cerebral ischaemia induces a rapid suppression of cellular immune responses in lymphatic organs – in the lung as well as in blood. In experimental stroke, these changes in immunity precede the development of bacterial infection in the lung.30–33 Impaired early lymphocyte responses, in particular reduced interferon gamma (IFN-γ) production by natural killer (NK) and T cells, appear to be an essential stroke-induced defect in the antibacterial defence as prevention of lymphocyte apoptosis by caspase inhibitors,34 adoptive transfer of IFN-γ-producing lymphocytes (i.e. T and NK cells) or early treatment with recombinant IFN-γ inhibit pneumonia after experimental stroke.30,35 Stroke-induced immunodepression is long-lasting and facilitates the development of severe pneumonia after aspiration of an otherwise harmless small dose of Streptococcus pneumoniae even 14 days after experimental stroke.35
Recent evidence from clinical studies36–41 indicates that down-regulation of systemic cellular immune responses, including rapid decrease in peripheral blood lymphocyte counts and functional deactivation of monocytes and T-helper type 1 cells, also occurs in stroke patients. Additionally, signs of immunodepression are more prominent in patients who develop infectious complications. Collectively, these clinical data corroborate experimental findings that changes in immune responsiveness after stroke occur before the onset of infectious complications and indicate that the extent of stroke-induced immunodepression correlates with the risk of infectious complication.
Recently, we and others described immune parameters significantly associated with infectious complications after stroke.36,38,42,43 For example, patients with infection had significantly lower levels of the major histocompatibility class II (MHCII) molecule human leukocyte antigen-DR (HLA-DR) on monocytes at days one, three and eight than non-infected patients. Moreover, reduced monocytic HLA-DR expression at day one was a strong independent predictor of subsequent post-stroke infection.36,41 Thus, immunological or infection parameters may be helpful predictors of post-stroke pneumonia. The current concept of post-stroke infections is shown in Figure 1.
Urinary Tract Infection
Several parameters are known to increase the risk of UTI after stroke including female sex, age, dependency before stroke, stroke severity (measured by NIHSS), poor cognitive function and catheterisation.22,44 Catheterisation is a well-described risk factor for healthcare-associated UTI. Their inappropriate use may be more common in stroke patients, thereby further increasing the risk of UTI. Usually, catheterisation is a consequence of urinary dysfunction and retention, occurring in 29–58% of stroke patients.45 Urine-storage disorder due to bladder hyper-reflexia seems to be more common after stroke.46 Risk factors for bladder dysfunction include large infarcts and cortical involvement. In addition, aphasia, cognitive impairment and severe functional impairment are independently associated with bladder dysfunction.47
Only a few studies have focused on the use of Foley catheters in patients with acute stroke; therefore, frequency of use remains unclear. However, clinical characteristics of stroke patients may promote catheter placement. Disturbance of consciousness and dysphasia impaired the ability to communicate the need to urinate. In addition, the high incidence of bladder dysfunction will increase the probability of being catheterised. Immobilisation due to motor dysfunction will complicate transfer to the toilet and even hamper the use of helpful devices, such as urinals or bedpans. In addition, the above-mentioned immunodepressive state induced by the acute CNS lesion may lead not only to chest infections but also to UTI.
Prevention of Post-stroke Infections
Pneumonia
As dysphagia is the most important risk factor for post-stroke pneumonia, several approaches aim at minimising the risk of aspiration. One well-known drawback is the lack of reliable diagnostic screening methods to identify patients at high risk of aspiration. Clinical bedside assessments have been shown to miss up to 40% of those patients who aspirate (‘silent aspiration’).47–49 Videofluoroscopy (VFS) can be considered the gold standard in diagnosing dysphagia with aspirations (silent or not) because of its ability to study the entire process of swallowing.49 Nevertheless, this examination requires patient co-operation and a sitting posture, so it cannot be proposed for all patients in the early period after stroke.
However, screening for dysphagia is effective to prevent post-stroke pneumonia. Using a formal dysphagia screening protocol decreases the risk of pneumonia in patients hospitalised for ischaemic stroke by three-fold.50 In order to reduce the risk of aspiration, dysphagia is usually managed by placement of a nasogastric tube. Percutaneous endoscopic gastrostomy (PEG) is applied to patients who remain unable to swallow in the long term. Although a nasogastric tube is easy to insert, it is uncomfortable and can be easily dislodged, leading to treatment failure and aspiration. PEG placement is an invasive procedure and can be complicated by peritonitis or bowel perforation.51 The Feed Or Ordinary Diet (FOOD) trials demonstrated that early enteral tube feeding might reduce case fatality compared with no tube feeding for more than seven days after stroke onset. However, the overall increase in survival was associated with an increased proportion of survivors with poor outcome. Compared with nasogastric tube feeding, early PEG tube feeding might be associated with poorer outcome and should be avoided in stroke patients. Furthermore, there were no significant differences between groups in the frequency of pneumonia, suggesting that early tube feeding does not prevent or favour infection rates.52
Another practical approach to prevent post-stroke pneumonia is anti-infective prophylaxis. In a rodent stroke model, moxifloxacin given immediately or 12 hours after middle cerebral artery occlusion prevented infection and fever, significantly reduced mortality and improved neurological outcome.53 Over the last few years, three heterogeneous randomised clinical trials have tested the efficacy of antibiotic prophylaxis in acute stroke to prevent post-stroke infection.36,54,55 Differences in inclusion criteria, infection criteria, type of intervention and outcome measures may have led to conflicting results. Although a meta-analysis demonstrated that antibacterial prophylaxis reduced the occurrence of post-stroke infection from 38 to 24% (odds ratio [OR] 0.44, 95% confidence interval [CI] 0.23–0.86),56 the beneficial effect on stroke outcome remains an open matter and has to be evaluated in future trials.
Urinary Tract Infection
As the use of Foley catheters increases the risk of UTI after stroke, the limited use of catheterisation may prevent UTI after stroke. Educational interventions can influence the inappropriate use of Foley catheters57,58 and thus dramatically reduce the frequency of UTI.58 Antiseptic-coated catheters have also been studied in the prevention of UTI. Only small clinical trials have shown that silver-oxide- coated catheters may reduce asymptomatic bacteriuria. However, these trials did not evaluate the efficacy of silver-oxide-coated catheters in symptomatic UTI.59 Antibiotic-impregnated catheters or condom catheters may reduce the frequency of asymptomatic bacteriuria.59,60 Thus, the general use of antiseptic-coated or condom catheters can not be recommended.
Antimicrobial prophylaxis may reduce the risk of stroke-associated UTI. However, the current urological guidelines strongly discourage prophylactic antibacterial treatment due to the lack of convincing evidence that it prevents bacteriuria and the risk of adverse events and antimicrobial resistance.61
Treatment of Post-stroke Infections
Current guidelines for stroke treatment recommend early diagnosis and antibacterial treatment in cases of suspected infection.62,63 Basically, anti-infective treatment of post-stroke infection follows current treatment guidelines of hospital-acquired (bacterial) pneumonia (HAP) or UTI, respectively. After initiating empirical antibacterial treatment, the isolation of pathogens from relevant species should be the main diagnostic focus. Thereafter, treatment should be adapted to results of resistance testing.
Pneumonia
Post-stroke pneumonia usually occurs within the first three to five days after hospitalisation36,54,55 and thus can be considered as early-onset HAP. Early-onset HAP is primarily attributed to Gram-negative bacteria, such as Haemophilus influenzae, and Gram-positive bacteria such as methicillin (meticillin)-sensitive Staphylococcus aureus (MSSA) and S. pneumoniae. Late-onset nosocomial pneumonia is usually attributed to higher-level antibiotic-resistant Gram-negative bacteria (e.g. Pseudomonas aeruginosa, Acinetobacter spp.) and Gram-positive bacteria (e.g. methicillin-resistant S. aureus [MRSA]). This classification leads to different strategies for empirical antimicrobial treatment: monotherapy with narrow-spectrum antibiotics for the treatment of early-onset pneumonia but broadspectrum therapy for Pseudomonas spp. or MRSA with late-onset infection. However, this dichotomised concept is currently under discussion.64,65 Importantly, empirical antibacterial treatment should follow local information about antimicrobial resistance. It is strongly recommended to stratify the risk of higher-level resistant bacteria using the following risk factors: inhabitant of healthcare facilities, medical co-morbidity (chronic heart disease, cirrhosis, renal insufficiency) and preceding anti-infective treatment.
Patients with low risk of higher-level resistant bacteria should be treated with antibacterial monotherapy. Aminopenicillin/β-lactamase inhibitors (BI) (e.g. amoxicilline/sulbactam), cephalosporins group II/III (e.g. cefuroxime/ceftriaxone) or fluorochinolones (e.g. moxifloxacin/ levofloxacin) are recommended in these cases.
High-risk patients should be treated with a combination of cephalosporines group IIIb (e.g. ceftazidime) and aminoglycosides (e.g. gentamicine). Three to five days after clinical improvement, the antibacterial treatment should be terminated. In total, the treatment period should not exceed 10 to 14 days (see Table 4).
Urinary Tract Infection
The most common pathogen in UTI remains Escherichia coli (55–80%). In about 5–10% of cases, other Enterobacteriaceae, such as Proteus mirabilis and Klebsiella spp., can be isolated. Occasionally, Staphylococcus saprophyticus are isolated.66–68 Knowledge of the antimicrobial susceptibility profile of uropathogens causing uncomplicated UTI in the community should guide therapeutic decisions. However, the resistance pattern of E. coli strains and other pathogens causing UTI may vary considerably between European regions and countries, so that no general recommendations are suitable throughout Europe.
The following antimicrobials can be considered (see Table 5): trimethoprim-sulfamethoxazole (TMP-SMX), fluoroquinolones (e.g. ciprofloxacin, levofloxacin,) and β-lactams (e.g. ampicillin/sulbactam, ceftriaxone). Treatment duration should not exceed seven days.
Conclusion and Perspectives
Even in dedicated stroke units with regular quality management protocols, post-stroke infections remain relevant complications that can be prevented and treated. Although their effect on long-term outcome remains an open issue, it is highly probable that post-stroke infections have a negative impact on mortality and neurological status. In the next few years, several predictive tools (e.g. clinical scores, blood-borne biomarkers) will be developed and validated. Thus, patients at risk will be identified early and reliably. Additionally, pre-emptive and biomarker-guided anti-infective treatment approaches will be tested to improve neurological outcome after stroke. ■