Objectives: To analyse clinical, echocardiographic, and prognostic characteristics of Staphylococcus aureus infective endocarditis (IE) compared with endocarditis caused by other pathogens.
Design: Cohort study.
Methods: 194 consecutive patients with definite IE according to the Duke criteria prospectively examined by transthoracic and transoesophageal echocardiography were enrolled. Patients without identified microorganisms were excluded. The S aureus IE group (n = 61) was compared with the group with IE caused by other pathogens (n = 133).
Results: Compared with IE caused by other pathogens, S aureus IE was characterised by severe co-morbidity, a shorter duration of symptoms before diagnosis, and a higher prevalence of right sided IE, cutaneous portal of entry, and history of renal failure. Severe sepsis, major neurological events, and multiple organ failure were more frequent during the acute phase in S aureus IE. In-hospital mortality (34% v 10%, p < 0.001) was higher in patients with S aureus IE and the 36 month actuarial survival rate was lower in S aureus IE than in IE caused by other pathogens (47% v 68%, p = 0.002). Multivariate analyses identified S aureus infection as a predictive factor for in-hospital mortality and for overall mortality.
Conclusions:S aureus IE compared with IE caused by other pathogens occurs in a more debilitated clinical setting and is characterised by a higher prevalence of severe sepsis, major neurological events, and multiple organ failure leading to higher mortality.
- Staphylococcus aureus
- clinical feature
Statistics from Altmetric.com
Staphylococcus aureus and streptococci are the main microorganisms responsible for infective endocarditis (IE).1,2S aureus IE is a malignant disease, with a constantly increasing frequency estimated between 17–40% of cases in recent series.1–8 Few studies have been specifically devoted to S aureus IE,9–14 have not compared S aureus IE with IE caused by other microorganisms, and have not systematically used transoesophageal echocardiography (TOE). The objective of this single centre study was to analyse the clinical, echocardiographic, and prognostic characteristics of S aureus IE and to compare them with the characteristics of IE caused by other pathogens.
Between January 1990 and December 2000, 243 consecutive patients with definite IE according to the Duke criteria15,16 were referred to our echocardiographic laboratory. The Duke criteria were applied retrospectively to patients hospitalised before publication of the Duke criteria. All patients were examined by transthoracic echocardiography (TTE) and TOE. Of the 243 patients, 194 patients with a positive blood culture were prospectively enrolled. Patients without identified microorganisms (n = 49) were excluded. In 61 patients (25.2%) S aureus was isolated, 15 patients (6%) had Staphylococcus epidermidis IE, 85 patients (35%) had streptococcal IE, 21 patients (8.7%) had enterococcal IE, and 12 patients (4.9%) had IE caused by uncommon microorganisms. The results of blood cultures were provided by the bacteriology laboratories and comprised identification of bacteria and antibiotic susceptibility testing. Patients were divided into two groups according to the microorganisms isolated: S aureus IE group (n = 61) and other pathogens IE group (n = 133).
Age, sex, presence of co-morbidity (history of diabetes, cancer, haematological malignancy, cirrhosis, renal failure, dialysis, heart failure, or coronary artery disease), hypertension, valvar heart disease, valve prosthesis, cardiac surgery, and the presence of an intravascular device (venous catheter, pacemaker, or dialysis device) were analysed. A co-morbidity index taking into account the patient’s age and history was calculated.1,17
The following acute clinical events present on admission or occurring during hospitalisation were recorded: heart failure, neurological event, peripheral embolism, severe sepsis, multiple organ failure, and hospitalisation in the intensive care unit. The duration of symptoms before admission (interval between the presumed onset of symptoms and the date of admission for IE) and the portal of entry of the infection were investigated. Embolic events were diagnosed based on clinical signs and data derived from a non-invasive procedure (cerebral and thoraco-abdominal computed tomography recorded in 70% of patients).18 A major neurological event was defined as the development of an ischaemic stroke with hemiplegia, haemorrhagic stroke, cerebral abscess, features of encephalopathy, or coma.19 A minor neurological event was defined as a transient ischaemic attack or cerebral embolic accident with no serious clinical signs.19 Severe sepsis was defined as a systemic inflammatory syndrome secondary to an infectious process leading to organ dysfunction or signs of hypoperfusion or hypotension.20 Multiple organ failure was defined as dysfunction of at least two organs during hospitalisation for IE.
Echocardiography was performed with a Hewlett Packard Sonos phased array (HP 1000, HP 2500 or HP 5500; Hewlett Packard, Andover, Massachusetts, USA) ultrasound machine with a 2.5 MHz transthoracic transducer and a 5 MHz transoesophageal transducer. TTE was systematically completed by TOE. All echocardiographic studies were performed by standard techniques and by experienced echocardiographers during the acute phase of IE without any complications. The presence, size, and location of valvar vegetation or perivalvar abscess were evaluated. Standard definitions were used for vegetations, abscesses, and other cardiac infective lesions.15,21,22 All TOE recordings were reviewed by an experienced echocardiographer to measure in various planes the maximum length of vegetations18 and the maximum surface area of abscesses.23 The mobility of vegetations was graded on a scale of 1 to 4 with severe mobility corresponding to grade 4.24 Valvar regurgitation was quantified by Doppler echocardiography by standard method.25
Follow up data included surgical treatment and death occurring during hospitalisation or follow up. Early surgery was defined as surgery performed during hospitalisation for management of IE.5,26 In-hospital mortality was defined as death occurring during the hospitalisation for IE.5,26 Overall mortality refers to death occurring during hospitalisation and follow up. Late mortality excluded death occurring during hospitalisation. All patients completed follow up with a mean follow up duration of 30 months.
Data were statistically analysed with SPSS 9.0 software (SPSS Inc, Chicago, Illinois, USA). Quantitative variables were expressed as the mean (SD). The two groups were compared by Student’s t test or χ2 test. The cumulative probability of survival was estimated by the Kaplan-Meier actuarial method at one month intervals and reported as mean (SE) estimated survival. The log rank test was used to determine any significant differences. To evaluate the impact of S aureus IE on survival, multivariate models incorporating age, sex, co-morbidity index (taking the patient’s age and history into account), usual prognostic factors (prosthetic valve, heart failure, major neurological events, renal failure, abscess, vegetation size), and early surgery were studied. A multivariate logistic regression model was used to determine in-hospital mortality and a Cox multivariate model for overall mortality. A probability value of p < 0.05 was considered significant.
Of the 194 patients studied (140 men and 54 women, mean (SD) age 59 (15) years), 61 patients had S aureus IE and 133 patients had IE caused by other microorganisms. Twenty four per cent of patients of our series were referred from another hospital. IE affected a native valve in 89% of patients and a valve prosthesis in 11% of patients (six mechanical prosthesis, 15 biological prosthesis). A vegetation was visualised in 98% of cases and a perivalvar abscess in 22%. Early surgery was performed on 37% of patients (33% of patients with S aureus IE and 39% of patients in IE caused by other pathogens, p = 0.33). The indications for early surgery were heart failure in 18 patients (25%), embolic risk in 16 (22%), uncontrolled infection in 15 (21%), severe regurgitation in six (8%), abscess in three (4%), and multiple indications in 14 (19%). In the 72 patients who underwent early surgery, 24 had mechanical valves, 22 had biological valves, 14 had mitral valve or tricuspid valve repair, six had pacemaker line extraction, four had multiple valve surgery, one had a homograft, and one had aortic valve vegetectomy.
The mean duration of antibiotic treatment before surgery was 10 (8) days in the S aureus IE group and 15 (16) days in other pathogens group (p = 0.30).
Comparison of clinical characteristics
Tables 1 and 2 show the results of univariate analysis comparing clinical setting and clinical characteristics during hospitalisation of S aureus IE versus IE caused by other pathogens.
Mean age, sex ratio, prevalence of history of hypertension, history of heart failure, history of coronary disease, cancer, alcoholism, cirrhosis, frequency of prosthetic valve IE, aortic valve IE, acute heart failure, embolic event, minor neurological event, and of early surgery were comparable between the two groups (tables 1 and 2). Compared with IE caused by other pathogens, S aureus IE was characterised by severe co-morbidity (p = 0.026), a shorter mean duration of symptoms before admission (p < 0.001), a higher prevalence of history of renal failure (p < 0.001) and dialysis (p = 0.005), right sided IE (p = 0.009), a cutaneous portal of entry (p < 0.001), severe sepsis (p < 0.001), admission to the intensive care unit (p < 0.001), multiple organ failure (p = 0.001), and major neurological events (p = 0.04) (tables 1 and 2).
Table 3 compares echocardiographic characteristics. The presence and size of abscess, mean vegetation size, and severe mobile vegetations were comparable in the two groups. Severe valvar regurgitation and multiple valve endocarditis were less frequent in the S aureus IE group.
In-hospital mortality was 17% in the total population. Compared with IE caused by other pathogens, the in-hospital mortality rate was higher in patients with S aureus IE (34% v 10%, p < 0.001). Table 4 reports the clinical complications of patients who died in hospital. Severe sepsis was more frequent in patients who died during the hospital phase of IE in the S aureus IE group. Among the 61 patients with S aureus IE, 21 died during the acute phase of IE. The causes of death in the S aureus IE group were multiorgan failure in 12 patients, neurological event in two, sudden death in three, heart failure in one, operative death in two, and tamponade in one. The causes of death among the 13 patients who died in the other pathogens group during the acute phase of IE were multiorgan failure in five patients, neurological event in two, sudden death in three, and operative death in three. The mean duration of antibiotic treatment before death was 22 (19) days in the S aureus IE group versus 22 (18) days (p = 0.95).
The overall 36 month mortality rate was 35%. By multivariate analyses S aureus infection was identified as a powerful predictive factor for in-hospital mortality and for overall mortality (tables 5 and 6). The 36 month actuarial overall survival rate was significantly lower in the S aureus IE group than in the group with IE caused by other pathogens (47% v 68%, p = 0.002)) fig 1).
Thirty four of the 160 survivors after hospitalisation died after discharge during the 36 month follow up. The causes of late death were heart failure in seven patients, neurological event in four, cancers in seven, sudden death in two, myocardial infarction in one, operative death in one patient who underwent cardiac surgery after discharge at four months, and unknown in 12. The 36 month late survival rate after discharge was similar in the two groups (72% v 75%, p = 0.95; fig 2).
Our study of 194 patients with definite IE according to the Duke criteria15,16 shows that S aureus IE compared with IE caused by other pathogens is characterised by a shorter duration of symptoms before diagnosis and by a higher prevalence of co-morbidity, history of renal failure, right sided IE, cutaneous portal of entry, severe sepsis, major neurological events, and multiple organ failure. S aureus IE is more severe in terms of morbidity and mortality. To our knowledge, clinical and morphological characteristics have not been compared between S aureus IE and IE caused by other pathogens in terms of the Duke criteria.
S aureus IE has been reported with increasing frequency during the past decades3,6,13 and has emerged as a dominant cause of IE.6,27 In our study S aureus was responsible for 25% of all cases of IE. S aureus IE usually occurs in a debilitated clinical setting: chronic renal failure, haemodialysis, diabetes, alcoholism, cancer, haematological malignancy, immunodepression, and drug addiction.6,12,28,29 In our study, severe co-morbidity was associated with S aureus IE. The association between haemodialysis, renal failure, and S aureus IE was confirmed in our study. This susceptibility may be caused in part by an underlying disease and its treatment (immunosuppression, intravenous catheter). S aureus is a bacterium that easily adheres to inert structures. Intravascular device associated S aureus IE has recently been recognised as an emerging problem by other investigators.13,30 In our series, the presumed source of infection was an intravascular device in 17% of patients with S aureus IE.29 Right sided IE was more frequent in the S aureus IE group.1,13 Not surprisingly, the cutaneous portal of entry was identified as a major source of infection in the S aureus IE group in our study. Only three patients in our study had a history of injecting drug use.
S aureus is a malignant disease known to be responsible for severe sepsis.31,32 Accordingly, admission to the intensive care unit, development of severe sepsis, and multiple organ failure were significantly more frequent in the S aureus IE group in our study. This virulence has been ascribed to a variety of complex factors, which include its capsule and cell wall, the production of extracellular enzymes and toxins that promote tissue invasion, its capacity to persist intracellularly in phagocytes, and its potential to acquire resistance to antimicrobials. Clinical signs of S aureus IE appear and evolve rapidly, leading to hospital admission3,4,9,14 as illustrated by the shorter mean duration of symptoms in the S aureus IE group in our study. Neurological complications are sometimes the first presenting signs of IE. Symptoms may range from simple confusion to unexplained coma.19,33 Previous series have noted a high incidence of neurological events in S aureus IE.19,34,35 Accordingly, a high frequency of 18% major neurological events in the S aureus IE group compared with 8% in IE caused by other pathogens was observed in our series. Only 11% of embolic events occurred after initiation of antibiotics. Because not all patients underwent cerebral and thoraco-abdominal computed tomography, the true incidence of embolic events in the current study may have been underestimated. The frequency of heart failure, one of the main complications of IE, was comparable in the two groups. Surgery was performed according to established guidelines, predominantly in the presence of a complication such as recurrent embolism, heart failure, or evidence of perivalvar extension.36,37 Because of the virulence of S aureus and the severity of the clinical features, some authors recommend aggressive management with very broad indications for early surgery performed in 33% of cases in the current series.3,38 However, surgery in the S aureus group is sometimes not performed because of a prohibitive operative risk related to the debilitated clinical setting associated with extremely severe clinical features and multiorgan failure (six patients in our series). In our study, as in other series reported in the literature,1,39–41 early surgery defined as surgery performed during hospitalisation for management of IE5,26 was also frequent in the group with IE caused by other pathogens. The benefit of early surgery in patients with S aureus IE was not addressed in the current study, which has the disadvantages of observational studies of consecutive patients where the decision to operate or not was not randomised but based on the clinical judgement of the physician or surgical team. Thus, further studies are needed to evaluate whether earlier surgical intervention for selected patients will improve the outcome of S aureus IE.37
S aureus IE is associated with high morbidity and mortality. S aureus IE compared with IE caused by other pathogens occurs in a more debilitated clinical setting. The prognosis of S aureus IE is therefore very serious, with a more severe prognosis than with IE caused by other pathogens and with a high in-hospital mortality of 34% in our series, which ranges between 30–46% according to various authors,1–3,31,42 even reaching 71% in a study published in 1986.9 The present study showed that in IE, independent of age, sex, co-morbidity index or usual prognostic factors, S aureus infection causes an excess risk of in-hospital mortality and of overall mortality. Because S aureus right sided IE is a different disease, we performed a separate prognostic multivariate analysis excluding S aureus right sided IE. We obtained similar results to the analysis including all patients. Patients with S aureus prosthetic IE are a very high risk subgroup. In the current study we had only six patients with a prosthetic valve in the S aureus IE group. Therefore, we could not specifically analyse S aureus prosthetic IE and compare it with non-S aureus IE. We cannot exclude the possibility that our patients may constitute a selected cohort from a referral centre with more severe illness than the average population with IE. Thus, the differences in mortality and morbidity may to some extent be a result of referral bias.
The excess mortality associated with S aureus IE in the current study, essentially related to the severity of sepsis and the particularly high risk clinical setting, mainly occurs during the hospital phase as indicated by actuarial survival curves, which tend to become parallel in the two groups after discharge from hospital (fig 1). Accordingly, S aureus infection was not a predictive factor of late mortality. Rapid management is therefore essential with a need for early surgery in selected patients.3,36,38