The spectrum of acute bacterial meningitis in elderly patients

Background We conducted a prospective, observational study in Barcelona to determine the epidemiology, clinical features, and outcome of elderly patients with acute bacterial meningitis (ABM) compared with younger adults. Methods During 1982–2010, all patients with ABM were prospectively evaluated. There were two groups: I (15–64 years) and II (≥ 65 years). All patients underwent clinical examination on admission and at discharge following a predefined protocol. Results We evaluated 635 episodes of ABM. The incidence was 4.03/100,000 (Group I) and 7.40 /100,000 inhabitants/year (Group II) (RR = 1.84; 95%CI: 1.56–2.17, P < 0.0001). Elderly patients had co-morbid conditions more frequently (P < 0.0001) and more frequently lacked fever (P = 0.0625), neck stiffness (P < 0.0001) and skin rash (P < 0.0001), but had an altered level of consciousness more often (P < 0.0001). The interval admission-start of antibiotic therapy was longer for elderly patients (P < 0.0001). Meningococcal meningitis was less frequent in elderly patients (P < 0.0001), whereas listerial (P = 0.0196), gram-negative bacillary (P = 0.0065), and meningitis of unknown origin (P = 0.0076) were more frequent. Elderly patients had a higher number of neurologic (P = 0.0009) and extra-neurologic complications (P < 0.0001). The overall mortality ratio was higher in elderly patients (P < 0.0001). Conclusions Elderly people are at higher risk of having ABM than younger adults. ABM in the elderly presents with co-morbid conditions, is clinically subtler, has a longer interval admission-antibiotic therapy, and has non-meningococcal etiology. It is associated with an earlier and higher mortality rate than in younger patients.


Background
During the past years we have witnessed significant changes in the epidemiology of ABM. The most important change is the dramatic decline of incidence of meningitis due to Haemophilus influenzae serotype b (Hib) due to infant vaccination [1,2], and the striking decrease of early-onset disease caused by group B Streptococcus because of use of intra-partum antibiotic prophylaxis [3].
The incidence and pattern of many infectious diseases vary over time. These changes are due to many factors, but among them modifications in the population structure stand out. Population structure in developed countries is continuously changing, and this change is mainly driven by two factors; first, the continuing aging of native population, and second the incoming migratory fluxes from developing countries. However, most likely the first one is still predominant and, therefore, elderly people are one of the fastest-growing segments in most western societies. The incidence of meningeal pathogen infections is highly influenced by the age of the patient, and ABM is considered a disease of extreme ages of life [4].
The aim of our study was to determine the spectrum of ABM in elderly patients over the last 30 years, compared to a younger population from the same geographical area, both recruited from a community university hospital in Barcelona (Spain).

Study population
We used data from a large, prospective, single-hospital cohort of patients with ABM enrolled over a 30-year period at the Hospital de la Santa Creu i Sant Pau (Barcelona, Spain). Our institution is a 620-bed tertiary university hospital providing care for an urban area with a population of 441,392 inhabitants. From 1982 to 2010 all consecutive adults with ABM diagnosed at our Hospital were prospectively identified and followed. The study was approved by the Ethics Committee of the Hospital de la Santa Creu I Sant Pau.

Diagnosis of acute bacterial meningitis
The diagnosis of ABM caused by a specific bacterial pathogen was based on the presence of consistent clinical findings (sudden onset of headache, fever, nausea, vomiting, neck stiffness and/or altered mental status) and one of the following: a positive cerebrospinal fluid (CSF) culture; or a negative CSF culture with a finding of pleocytosis and at least one of the following: a positive CSF antigen test, a positive blood culture, or identification of gram-negative diplococci on Gram stain of CSF in patients with a petechial or purpuric rash and a fulminant course. These last cases were considered to be caused by Neisseria meningitidis [5]. Episodes of ABM without etiologic diagnosis were included if patients had a consistent clinical picture and CSF cyto-biochemical findings consistent with purulent meningitis (pleocytosis of more than 5 polymorphonuclear leukocytes/mm 3 , hypoglycorrachia defined as a CSF/blood glucose ratio < 0.50 and high CSF protein defined as more than 45 mg/dl) despite negative blood and negative CSF Gram stain and cultures [6,7]. Patients with viral, fungal, or mycobacterial meningitis or a history of neurosurgical procedures or traumatic head/spinal injury were excluded.
From each episode, the following data were collected: sex, age, underlying conditions, co-morbidity and co-morbidity scoring according to the Charlson index, clinical characteristics, laboratory and microbiologic results (isolates and susceptibility tests), complications, treatment, and outcome. Data were collected during the index hospitalization, and each patient was seen on admission and followed after discharge by one of the authors (PD or VP).

Microbiologic methods
Isolates were obtained from routine cultures and were identified using standard methods [8]. The disc diffusion susceptibility test was performed according to Clinical Laboratory Standards Institute (CLSI) guidelines [9], using commercially available discs (Bio-Rad ™ , Marnes La Coquette, France). MICs were determined using the broth micro dilution method according to the CLSI guidelines [10] using commercial panels (Sensitre, ™ Trek diagnostic systems, West Sussex, England) or Etest ™ (AB Biodisk, Solna, Sweden) according to the recommendations of the manufacturer. Resistance breakpoints for penicillin were MIC ≥ 0.12 mg/L for pneumococcus and meningococcus [11].

Definitions
A subject aged > 14 years was considered an adult and when aged ≥ 65 years, elderly. Pre-admission receipt of adequate antibiotic therapy has been previously defined [12]. The interval from onset of symptoms and signs of ABM until admission was the interval symptomsadmission (ISA), whereas that from admission to the first dose of antibiotic for ABM treatment was the interval admission-therapy (IAT). Co-morbid conditions were considered to be present if the patient had a confirmed diagnosis of one or more of the following: cancer, systemic vasculitides, cirrhosis of the liver, diabetes, chronic renal failure, human immunodeficiency virus infection, heart failure, and chronic pulmonary disease [12]. Alcohol abuse was defined when alcohol intake was > 40 g daily [13]. Co-morbidity was graded according to the Charlson Co-morbidity Scale [14].
Extra-meningeal infection was considered when a patient had a focal infection distant from the Central Nervous System and the same pathogenic bacterium was isolated from the primary focus or from blood culture [15]. Appropriate antibiotic therapy was defined as administration of one or more antimicrobial agents shown to be effective against the meningeal pathogen, which achieve adequate CSF concentrations, commenced on admission or before neurological or systemic deterioration in inpatients [16]. Impaired mental status, seizures and focal neurologic signs after the first dose of antibiotics were considered neurological complications. The development of septic shock, acute respiratory failure, renal failure, and/or consumption coagulopathy was considered as extra-neurological complications if related to ABM and apparent on admission or shortly afterwards [17][18][19][20][21][22].
Nosocomial meningitis was defined as that developing more than 48 hours after admission or within one week after discharge from the hospital [23]. Coma has been defined elsewhere [17]. Inpatient deaths were considered due to ABM when meningitis was the underlying and immediate cause of death. Causes of death were classified in two categories: neurological causes or systemic causes [18,23]. ABM was not considered to be the underlying cause of death if a disease process unrelated to meningitis began > 24 h after meningitis resolution and if it initiated the train of morbid events leading to death [23]. Sequelae were defined as previously outlined [12].

Statistical analysis
Continuous variables were compared using Student's t-test or the Mann-Whitney U test when appropriate. To compare the incidence rates and to estimate the rate ratio plus the 95% CI, we used a Poisson model. We analyzed categorical data using the χ 2 test or Fisher's exact test when indicated. Survival curves in younger patients and in those older than 65 years were compared by means of the logrank test. Statistical analyses were performed using Statistical Product and Service Solutions (SPSS) software version 16 (SPSS Inc., Chicago, IL).
higher probability, although not statistically different, of having received out-of-hospital antibiotic therapy. An extra-meningeal focus of infection was found more frequently in patients from group II, and this was especially true for pneumonia and infra-diaphragmatic foci of infection (RR = 1.66; 95% CI: 1.20-2.29, P = 0.0126). Elderly patients more frequently lacked fever (although the difference did not reach statistical significance), neck stiffness and skin rash, but had an altered level of consciousness more often (Table 2). Despite being brought to the hospital with an interval similar to that of younger patients, elderly patients suffered a greater delay from arrival to the hospital to starting antibiotic therapy.

Evolving features and outcome
Elderly patients developed more frequently coma, and cranial nerve palsies of new appearance. Patients from group II also developed acute respiratory failure and acute renal failure more frequently (Table 4). Almost 90% of patients in each group received adequate empiric antibiotic therapy. The overall mortality ratio was significantly higher for patients in group II (29.8% vs. 11.7%) ( Table 4). This higher mortality rate was caused by a higher mortality rate of pneumococcal and of meningitis of unknown origin. Survival curves for both groups of patients showed an earlier probability of death for elderly patients (Figure 1). When death directly attributable to ABM was considered, the probability of dying was again higher for elderly patients. Development of complications conveyed higher rates of mortality in both groups, without statistically significant differences between them (Table 5). There were no differences between groups with respect to postmeningitic sequelae.

Discussion
Our study shows that elderly patients are at much higher risk of developing ABM than younger adults. We also observed a change in the etiology of the disease, Neisseria meningitidis ranked first for younger patients, whereas it was surpassed by pneumococcal, listerial, and meningitis of unknown origin in the elderly. Another important, although expectable, change is the increased prevalence of co-morbid conditions among elderly people. This may have an impact, not only on the etiology of ABM, but also on the rate of complications and eventually on mortality. Other differing findings for elderly people include a higher chance of having received out-of-hospital antibiotic therapy, more frequently lacking fever, neck stiffness and skin rash, developing neurologic and extra-neurologic complications more frequently, and having an increased risk of death because of ABM. In summary, the linkage and interaction of higher co-morbidity, different etiologic spectrum, more pre-admission antibiotic therapy and subtler clinical manifestations may contribute to explain most of the differences between elderly and younger patients.  However, our work has inherent limitations. First, our study is based on a single hospital and this may imply that the results may only be applicable to places with a population structure and local ecology similar to ours. However, changes in the spectrum of ABM similar to those observed in our study have been described, especially those referring to the epidemiology of meningococcal disease [2,[24][25][26]. Moreover, the pattern of population change documented in Spain, has also been reported in other Western, mainly European countries [27]. Therefore, the population scenario may resemble that described in the present work. Our study spans over a long period of time; during 30 years it is obvious that the treatment and management of critically ill patients and of ABM have changed a lot. Although the case record forms were standardized and remained the same throughout the study, they may reflect differing sensitivities among treating physicians and in the end are a reflection of the "real world". An important part of this real world is meningitis of unknown origin which is the most diagnosis-challenging ABM. We had strict criteria for this diagnosis, and when reasonable doubts arose, the case was not included in our series.
Changes in population structure contribute to changes in the spectrum of ABM. During the study period, Barcelona population has ranged from 1,754,900  [28]. Therefore, a decrease in the population more prone to meningococcal meningitis was observed together with an increase in the population more susceptible to having pneumococcal and listerial meningitis. Surprisingly, meningitis of unknown origin also significantly increased in group II, which may be explained by a higher prevalence of out-of-hospital antibiotic therapy among elderly people. The negative impact of pre-admission antibiotic therapy on the positivity of microbiological tests is well known [12,29]. Although N. meninigitidis is not a common cause of ABM in elderly patients, our epidemiological environment has for many years been one of high endemicity of meningococcal disease [24,30,31]. Moreover, the incidence of disease due to serogroup C meningococci increased in Spain from the mid-nineties [32]. This change of pattern led to the carrying out of a mass vaccination campaign against serogroup C in Spain and to inclusion of the meningococcal C conjugate vaccine in the routine vaccination schedule [33]. Since group C meningococci have a predilection for causing disease in older adults [12], the vaccination campaign may also have contributed to the decreased incidence of meningococcal meningitis among the elderly as our results show. This can be explained by the increase in herd immunity [34].
With respect to clinical findings, although elderly patients were brought to the hospital not later than younger ones, the time before first antibiotic dose was significantly longer. There are a cluster of circumstances that may explain this finding; elderly patients had taken out-ofhospital antibiotic therapy more frequently, and this, together with being older, makes the symptoms and signs of ABM subtler, as exemplified by the less frequent presence of symptoms and/or signs of ABM and skin lesions [15,29,[35][36][37][38], which may eventually lead to a diagnostic delay. But, above all, elderly patients had a cerebral CT performed more frequently, and this delayed the spinal tap and consequently the start of antibiotic therapy. Delay  in starting antibiotic therapy has been advocated as a main factor in ABM-driven mortality [36][37][38][39][40][41][42]. Therefore, a high index of suspicion of ABM is needed in elderly patients with fever and CNS dysfunction to rule out meningeal infection eventually leading to a prompt spinal tap and early starting of empiric antibiotic therapy. There were no differences between both groups in terms of CSF parameters except for a higher CSF protein content for group II patients, which is most likely due to the predominance of pneumococcal, listerial and gramnegative bacillary meningitis in this group. We know that these meningitis have a higher CSF protein content than meningococcal which predominates among group I patients (data not shown). Underlying immune suppression may cause changes in CSF parameters; and we found like Erdem et al. [37] lower CSF counts in immunosuppressed patients but not a different CSF/blood glucose ratio. In addition, the diagnostic yield of CSF Gram-stained smear was inferior in elderly patients, but we must take into account that pre-admission antibiotic therapy had been taken by 37% of patients, and we know that this is the most frequent cause of meningitis with negative CSF Gram-stained smears [35,43,44].
Both neurologic and extra-neurologic complications appeared more frequently in elderly patients. This may be explained by two facts; first, the physiological reserve and the ability of many organs and systems to respond to insults decrease with age, and second, in elderly patients, the baseline organ function may be compromised because of prior co-morbid conditions. Therefore, ABM and/or its associated systemic consequences may further damage an already dysfunctional organ, eventually leading to its failure. However, once present, they usually convey high mortality rates both in young and elder patients. The excess mortality of ABM in elderly people is a usual finding in related literature [37,38,44,45]. In fact, this difference in mortality rate is mainly driven by an excess mortality in pneumococcal and meningitis of unknown origin.

Conclusions
In summary, elderly people are at higher risk of having ABM than younger adults, ABM appears more frequently in patients with co-morbid conditions, is associated with a longer interval from admission to antibiotic therapy, is caused by S. pneumoniae, and patients develop neurologic and extra-neurologic complications more frequently. ABM in elderly patients is associated with an earlier and higher mortality rate than in younger ones.