ORIGINAL Gender Based Analysis of a Population Series of Patients Hospitalized with Infective Endocarditis in Portugal – How do Women and Men Compare?

Background: The impact of gender on the outcome of patients hospitalized with infective endocarditis (IE) is not fully understood. Objective: To verify the association between gender and the clinical profile of patients hospitalized with IE, treatment strategies, and clinical outcomes. Methods: This is a retrospective nationwide study of patients hospitalized with IE, based on hospital admissions between 2010 and 2018 in Portugal. Descriptive statistics were used to present variables. An inferential analysis was performed using multiple logistic regression. A 95% confidence interval and a 5% significance level were considered. Results: In total, 3266 (43.1%) women and 4308 (56.9%) men were hospitalized with IE. The women were older (76 vs 69 years old, p<0.001), more frequently presented arterial hypertension (39.8% vs 35.4%, p<0.001) and atrial fibrillation (29.5% vs 21.2%, p<0.001), and had less cardiovascular comorbidities. Acute heart failure was more common in women (32.9 vs 26.9%, p<0.001) and acute renal failure (13.6% vs 11.7%, p<0.001) and sepsis (12.1% vs 9.1%, p<0.001), in men. Women were less likely to undergo cardiac surgery (OR 0.48 – 95%CI 0.40–0.57, p<0.001) and had a higher postoperative mortality (OR 1.84, 95% CI 1.19–2.84, p=0.006). In-hospital mortality rates were comparable between genders (20.3% vs 19.6%, p=0.45). Conclusions: Women were less likely to undergo cardiac surgery when hospitalized with IE, and the female gender was a predictor factor for postoperative mortality. Overall, in-hospital mortality was not influenced by gender. Further research is necessary to fully clarify the impact of gender on IE management and outcomes.

(up to twenty). The hospital discharge report used the International Classification of Diseases (ICD)-9 until 2016, and ICD-10 from then onwards. All hospitalizations of patients with an IE diagnosis at discharge (ICD-9-CM codes 421.0, 421.1, 421.9, and 424.9; ICD-10-CM I33.0, I33.9, I38, and I39) between January 1, 2010 and December 31, 2018 were considered. Each patient and hospitalization episode were linked to the first institution of hospitalization.

Variables
For each index IE hospitalization associated with valvular surgery, we identified the year of hospitalization, date of surgery (when available), presence of a cardiothoracic unit at the first hospital of admission, clinical information (gender, age, year of discharge, length of hospital stay), cardiovascular history and comorbidities (diabetes mellitus, non-rheumatic valve disease, rheumatic valve disease, affected cardiac valve, arterial hypertension, chronic kidney disease, chronic coronary artery disease, cancer, human immunodeficiency virus [HIV] infection, cardiac devices and heart valve prostheses, atrial fibrillation, chronic liver disease, chronic obstructive pulmonary disease, use of opioid drugs, congenital heart disease), microorganisms (Staphylococcus, Staphylococcus aureus, Streptococcus, Enterococcus, Gram-negative bacteria, anaerobes, fungi, Brucella), complications compatible with IE (heart failure, embolic stroke, ischemic stroke, transient ischemic attack, septic shock, splenic abscess, acute renal failure, acute coronary syndrome, central nervous system abscess, or meningitis), cardiac surgery, and in-hospital death using the ICD-9 and ICD-10 codes -Supplement Table S1.

Definition of postoperative and in-hospital mortality
Postoperative mortality was defined as all-cause deaths that occurred during the index hospitalization in patients who underwent cardiac surgery.
In-hospital mortality was defined as all-cause deaths that occurred during the index hospitalization in all patients hospitalized with IE.

Statistical analysis
Continuous variables were presented as means ± standard deviations if following a normal distribution; otherwise, medians and interquartile ranges were displayed. Categorical variables were expressed as frequencies and percentages. For the bivariate analysis, the comparison between continuous variables was performed through recognized the need for reducing the gender gap in clinical care and research. 5 Various publications have highlighted the impact of gender on several fields of cardiology such as coronary artery disease [6][7][8] or heart failure, 9,10 as well as in general medicine involving sepsis 11 or pneumonia. 9 Infective endocarditis (IE) constitutes a rare condition with significant morbidity and mortality. Demographic changes, with an aging population associated with an upward trend in the implantation of prosthetic cardiac valves and cardiac devices, have justified a similar trend in the incidence of this pathology. [12][13][14] The impact of gender in the clinical profile and outcome of patients with IE has been seldom approached. A higher incidence among men has been noted 15,16 in international registries. Moreover, a recent systematic review of observational series from Portugal 17 concluded that a higher prevalence of men was noted in all studies and this predominance was also noted in all surgical series (men constituted more than two-thirds of all operated patients). Still, scarce and conflicting evidence persists regarding the influence of gender on the access to cardiac surgery and fatal outcome, [18][19][20][21][22][23][24] being mostly based on single-center cohorts. Populational studies addressing this issue are rare, 23 yet crucial to avoid selection bias in observational studies performed in tertiary centers.
Therefore, using population-based data, the authors sought to explore the association of gender and the clinical profile of patients hospitalized with IE, access to surgical interventions, and clinical outcomes.

Study design
This is a nationwide cross-sectional study using inpatient discharge data from all public hospitals of the Portuguese National Health System (NHS) considering patients admitted with IE. A comparative analysis based on gender (women and men) was performed.
Our data source and study population were described elsewhere. 25 In brief, our data comprised hospital discharge reports including clinical information (gender, age, geographical region, hospital, date of admission and discharge, length of hospital stay, destination -home, unknown, another acute hospital, a palliative care institution or outpatient clinic, discharge against medical advice, deceased), a clinical diagnosis list (one primary diagnosis and up to nineteen secondary diagnoses), and procedures To assess the factors associated with in-hospital surgical intervention and in-hospital mortality, inferential analysis was performed using multiple logistic regression (a generalized linear model using binomial distribution for the error and the logit link function). The stepwise (forward) method, based on Akaike information criteria minimization, was used for the selection of variables included in the model. The adjusted odds ratio, as well as the 95% confidence interval (95% CI), were estimated for each variable included in the regression model.
All tests were 2-tailed. The level of significance was set to α=0.05.
Data were analyzed using IBM SPSS Statistics for Windows version 24 (IBM Corp., Armonk, NY, USA).

Main clinical features of the overall cohort
During the study period, 3266 (43.1%) women and 4308 (56.9%) men were hospitalized with IE -see Table 1.
The length of hospital stay was superior in men. Women were on average older than men and presented higher prevalence of arterial hypertension (39.8% vs 35.5%) and atrial fibrillation (29.5% vs 21.2%). In contrast, men presented higher rates of non-rheumatic valve disease, coronary artery disease, HIV, and chronic liver disease. Staphylococcus and specifically Staphylococcus aureus were the most prevalent. Streptococcus and Enterococcus were more commonly identified in men; Gram-negative agents were more frequent in women. Regarding valve disease, left heart valve involvement was more frequent in men.
In-hospital complications were similar between the two groups except for heart failure, which was more common in women, and hemorrhagic stroke, acute renal failure, and sepsis, which were more prevalent in men.
Men underwent surgery more often. Finally, the inhospital fatal outcome was analogous between men and women. A bivariate analysis of exclusively medically managed patients revealed that women proportionally had a lower in-hospital mortality (19.4% vs 21.6%, p=0.026). Table 2 shows a stepwise logistic regression approach to assess the in-hospital surgical management of patients hospitalized with IE. Women were less likely to undergo cardiac surgery during hospitalization (OR 0.48 -95% CI 0.40-0.57, p<0.001).

In-hospital surgical cohort
In total, 232 (24.7%) women and 705 (75.3%) men underwent cardiac surgery during the index hospitalization for IE -Supplement Table S2.
Both groups were quite similar regarding comorbid conditions. A higher rate of atrial fibrillation was noted in women and a higher rate of malignancy , in men. Aortic valve and right heart valve disease were more prevalent in men. Regarding in-hospital complications, a higher rate of ischemic stroke was noted in men. Overall, postoperative mortality was significantly higher in women.

Postoperative mortality
Independent predictors of postoperative mortality are shown on Table 3. Female gender was an independent predictor of fatal outcome (OR 1.8, 95% CI 1.20-2.84, p=0.002). Other independent prognostic factors of inhospital mortality in patients with IE subjected to surgery were previous coronary intervention, chronic kidney or liver disease, Staphylococcus or Streptococcus infection, acute renal failure, and sepsis. Younger patients were less likely to have a fatal outcome.

Discussion
In a contemporary populational cohort of patients hospitalized with IE in Portugal and after controlling for several risk factors, important differences were noted between women and men. Men were more prevalent, being younger and with a higher rate of comorbidities. On the other hand, women were less likely to undergo cardiac surgery during the incident hospitalization for IE, with a higher post-operative mortality rate. Even so, overall in-hospital mortality was comparable among men and women.
Our data showed a 1.3 man/woman ratio, which is in accordance with a higher prevalence of IE in men in populational studies. [12][13][14] The reasons for this remain speculative, but factors such as a higher valve disease prevalence in men and different hormone profiles have been previously mentioned. 19 Women were older but men presented a higher burden of comorbid conditions such as coronary artery disease, cardiac valve disease, liver failure, or malignancy. Infectious agents such as Staphylococcus, Streptococcus, or Enterococcus were more common in men whereas women presented a higher rate of Gram-negative agents. In-hospital complications were analogous, with women presenting a higher rate of acute heart failure while men had a higher rate of acute renal failure and sepsis. All these aspects were noted in previous gender-based studies. 18,21,23 The male/female ratio increased to 3.0 in the surgical subgroup. Indeed, women were less likely to undergo cardiac surgery, which is a common finding with other studies. 18,19,21,23,26 Nonetheless, general indications for performing cardiac surgery in the context of IE 27 include acute heart failure, uncontrolled infection, and embolism. In this cohort, the profile of in-hospitalrelated complications was comparable between men and women, the latter group presenting a higher incidence of heart failure; this fails to explain the gender disparity in the access to surgical treatment among patients with IE. Additionally, men presented a higher incidence of hemorrhagic stroke, which normally delays and sometimes excludes cardiac interventions. Physician awareness 28 and other comorbidities not taken into account in this study such as frailty score, 29 dementia, 30 or neurologic sequelae could also have contributed individually to the decision to perform cardiac surgery and should be considered. Age, female gender, and endocarditis are variables included in cardiac surgery risk stratification scores such as EUROSCORE 31 or the STS 32 and the fact that women were older could have precluded cardiac surgery.
Additionally, women who underwent surgery in our study had a higher rate of fatal outcome when compared to men and constituted an independent risk factor for postoperative mortality. Demographics, comorbidities, infectious agents, and in-hospital-related complications in this surgical cohort were similar between men and women. This real-world data validates the increased surgical risk in women with IE, which is already considered in the above-mentioned risk scores. The higher surgical susceptibility of women was also found by Curlier et al. 23 and in a recent meta-analysis by Varela et al. 33 Weber et al. 22 concluded that a higher postoperative mortality rate in women was due to the presence of more comorbidities and perioperative risk factors, which was apparently not the case in the current study. This higher mortality after surgery in the female sex was also noted in other fields of cardiac surgery, such as after coronary artery bypass graft surgery. 34 Older age, a lower body surface area, 35 a higher incidence of heart failure, and referral bias 21 could also partially justify this higher surgical susceptibility.  Finally, regarding in-hospital mortality, no significant differences were noted between men and women. This is similar to findings from Curlier et al. 23 It would be critical to analyze the causes of death in these patients. Actually, a recent study based on national death certificate data in Portugal 36 concluded that, in a cohort whose basic cause of death was IE, women presented a 19% higher death rate than men. Conversely, Thuny et al. 37 concluded that a high burden of comorbidities and aortic valve involvement may explain an increased death rate among women.
In our study, women differed from men in significant aspects of clinical presentation and treatment options. They were older and less likely to undergo cardiac surgery when hospitalized with IE, despite presenting less comorbid conditions and a comparable complication rate. Higher postoperative mortality, but not overall in-hospital mortality, was observed. Further research is needed to understand reasons that can explain individualized management strategies and the impact of surgery on survival in women hospitalized with IE.
Our study has several limitations. First, ICD-9 and ICD-10 codes were used to identify individual cases of IE and clinical variables on an administrative database. The authors were unable to consult the patient's electronic database to confirm the diagnosis and to identify other important variables. Therefore, diagnosis or coding errors could have occurred. Second, the authors were unable to trace patients beyond the index hospitalization and cardiac surgery could have been performed after discharge. This could lead to an underestimation of the number of cardiac surgeries performed in this cohort. Therefore, only in-hospital surgical management during the index hospitalization was assessed. Third, this is a populational study; management strategies by individual physicians and the patients' personal treatment options could thus eventually lead to a referral bias towards cardiac surgery. Fourth, variables such as microbiological subgroups, source of bacteremia (community vs healthcare-related), presence of dementia, frailty score, or a socioeconomic situation that could have further explained our findings were not available for analysis.

Conclusions
In Portugal between 2010 and 2018, hospitalization with IE occurred less frequently in women, who were older and presented fewer comorbidities. A comparable rate of in-hospital complications was observed, with a higher prevalence of acute heart failure. Regarding surgical treatment, women were less likely to undergo cardiac surgery and female gender was an independent predictor of postoperative mortality. Further research is warranted to understand the reasons behind the influence of gender on treatments and outcomes of this disease. The individualized management of IE is frequently challenging, and the influence of gender should be carefully considered together with comorbidities and complications to improve the outcome of these patients.

Data availability statement
Data supporting the findings of this study are available from ACSS but restrictions apply to the availability of these data, which were used under license for the current study, thus not being publicly available. Data are however available from the authors upon reasonable request and only after permission by ACSS.

Potential Conflict of Interest
No potential conflict of interest relevant to this article was reported.

Sources of Funding
There were no external funding sources for this study.

Study Association
This study is not associated with any thesis or dissertation work.