Sex-Disaggregated Analysis of Central Venous Catheter-Related Bloodstream Infections in Patients with Cancer

Infections are a major cause of morbidity and mortality in patients with cancer [1, 2], with an up to three-fold higher mortality rate compared to the general population [3]. Especially patients with hematological malignancies have a particularly high risk of infections [4‒9]. The causes of infections are manifold, including cell-mediated and humoral immune deficiency, disruption of anatomical barriers, and therapy-related adverse effects such as neutropenia and mucositis or radiation-induced injury [10‒13]. Bloodstream infections (BSIs) are among the most serious infections in patients with cancer, occurring more frequently in patients with hematological malignancies than in those with solid tumors [4, 14‒18]. Of note, men show a higher susceptibility to many infectious diseases, such as BSIs [19‒21], and mortality rates are higher than in women [21, 22].

Central venous catheters (CVCs) are widely used when treating patients with cancer [23], and CVC-related bloodstream infections (CRBSIs) are important health care-associated infections. They also occur more frequently in patients with hematological malignancies than in those with solid tumors [4, 14‒18]. However, little is known about the role of sex differences in BSIs due to CVCs [24, 25], particularly in patients with cancer. Therefore, we here aimed to investigate sex differences in CRBSI in a real-world cohort of patients with cancer.

We used data derived from the Study to Evaluate Central Venous Catheter-related Infections in Hematology and Oncology (SECRECY), a real-world registry on routine CVC surveillance in hematology and oncology [26]. This ongoing multicenter registry was established in March 2013 and initiated in 11 hematology and oncology sites in Germany. Data on nonselected patients with centrally inserted non-tunneled internal jugular, subclavian, or femoral vein CVCs inserted for routine clinical use in adult hematology and oncology patients were collected; peripherally inserted CVCs were not registered. All CVCs were inserted according to local standard operating procedures following current guidelines [23]. More detailed information about the registry and procedures is described elsewhere [27]. CVCs were assessed for CRBSIs and classified according to the 2012 Infectious Diseases Working Party (AGIHO) of the German Society of Hematology and Medical Oncology (DGHO) criteria [28]. In brief, definite CRBSI (dCRBSI) required either a differential time to positivity ≥2 h in a pair of peripheral and central blood cultures or the detection of the same pathogen with the same susceptibility pattern in a peripheral blood culture and at the CVC tip. For probable CRBSI (pCRBSI), one of the following scenarios needed to be fulfilled: (i) detection of the same pathogen typically causing CRBSI in a peripheral and central blood culture without fulfilling criteria of dCRBSI and exclusion of other infection sites; (ii) exit site infection combined with BSI without criteria for dCRBSI. The microbiological analysis was conducted at the local sites and CRBSI assessed by local investigators.

Only internal jugular and subclavian vein CVCs being ≥1 day in situ entered into the registry from March 2013 to March 2024 from 10 active centers were included in the present analysis. Underlying diseases were hematological malignancies, including patients who received autologous or allogeneic hematopoietic stem cell transplantation or chimeric antigen receptor T-cell therapy, and solid tumors. Patients with benign hematological diseases or without cancer were excluded. Only CRBSIs classified as dCRBSI or pCRBSI were considered with the combination of both being summarized as dpCRBSI. Complicated CVC insertions were defined as (i) unusual bleedings; (ii) >2 vein punctures; (iii) Seldinger wire not passable, occurring during CVC insertion. Neutropenia was defined as neutrophil granulocytes <500/µL or white blood cell count <1,000/µL.

The present study focused on the comparison of CRBSI in men and women. It should be noted that in this study, the variable “sex” was determined by the data given in the patient’s health insurance record, representing the “legal gender” of the patient. In this setting, we assume a large overlap with the biologically informed parameter “sex,” but it was not possible to disentangle it fully from the self-assigned parameter “gender” that also includes gender identity and gender roles, as well as possible sex reassignment or intersex conditions in the patient’s history.

CVCs inserted in men and women were matched 1:1 for the underlying disease, the anatomic site of CVC insertion, use of chlorhexidine gluconate gel pad coated CVC dressing, use of antimicrobial coated CVC, complicated CVC insertion, and CVC in situ time by propensity scores (propensity score matching [PSM]) with a standardized mean difference of zero. The endpoints of this analysis were dCRBSI and dpCRBSI rates, as well as dCRBSI and dpCRBSI incidences (incidence density, calculated per 1,000 CVC days [29]). Fisher’s exact test was used to evaluate differences in CRBSI rates, and incidence rate ratios were analyzed using the Mid-P exact test. Continuous variables were compared by using the Mann-Whitney U test, categorical variables by using Fisher exact or Pearson χ² test, and proportions were compared by using the Z score test. Time-to-event analyses (where an event was dCRBSI or dpCRBSI) were performed by using the Kaplan-Meier method and the log-rank test. To identify independent risk factors for dCRBSI or dpCRBSI, multivariable stepwise logistic regression with backward elimination was performed.

Two-sided p values <0.05 were considered as statistically significant. Statistical analyses were performed using SPSS^® version 28 (IBM^®, Armonk, NY, USA) and OpenEpi version 3.01 (Open Source Epidemiologic Statistics for Public Health, Atlanta, GA, USA [30, 31]) statistical software.

From March 2013 to March 2024, a total of 5,391 CVCs were documented within the SECRECY registry. Of these, 5,075 (94.1%) fulfilled the inclusion criteria and were included in the present analysis (Fig. 1; Table 1).

In the entire cohort, 3,012 CVCs (59.3%) were inserted in men and 2,063 (40.7%) in women (p < 0.001). Patients’ characteristics are outlined in Table 1. The median age was 59 years and the vast majority of patients had hematological malignancies (93.2%). The proportion of CVCs inserted during neutropenia was 16.3%, and the proportion removed during neutropenia was 30.0%. Most CVCs (95.1%) were inserted in the internal jugular vein, and chlorhexidine gluconate gel dressing pads were used in 35.7% and antimicrobial coated CVCs in 11.8% of patients. The median CVC in situ time was 14 days.

There was no statistically significant difference in the dCRBSI rate between men and women (157/3,012 [5.2%] vs. 88/2,063 [4.3%]; p = 0.13), but significantly more men had dpCRBSI compared to women (298/3,012 [9.9%] vs. 138/2,063 [6.7%]; p < 0.001). Also, the dCRBSI incidence (3.5 vs. 2.7/1,000 CVC days; p = 0.07) and dpCRBSI incidence (6.6 vs. 4.3/1,000 CVC days; p < 0.001) was higher in men than in women, whereas there was no difference in the time to CRBSI onset (Table 1). However, time-to-event analysis showed statistically significant differences between men and women for cumulative dCRBSI incidence (p = 0.04; online suppl. Fig. S1; for all online suppl. material, see https://doi.org/10.1159/000542535) and for cumulative dpCRBSI incidence (p < 0.001; Fig. 2).

Overall, there was no difference in the distribution of the causative pathogens of dCRBSI and dpCRBSI between men and women (Table 1). The group with dCRBSI (157/245 [64.1%] vs. 88/245 [35.9%]; p < 0.001) and with dpCRBSI (298/436 [68.3%] vs. 138/436 [31.7%]; p < 0.001) included more men than women. However, coagulase-negative staphylococci (CoNS) were more frequently identified as causative pathogen of dCRBSI (111/168 [66.1%] vs. 57/168 [33.9%]; p < 0.001) and dpCRBSI (219/310 [70.6%] vs. 91/310 [29.4%]; p < 0.001) in men than in women. The same is true for other Gram-positive bacteria (33/50 [66.0%]; p = 0.02) and polymicrobial CRBSI (16/21 [76.2%]; p = 0.02). Overall, Gram-positive bacteria were found more frequently as causative pathogens of dCRBSI in men than in women (126/191 [66.0%] vs. 65/191 [34.0%]; p < 0.001), whereas there were no differences in Gram-negative bacteria between both sexes (19/35 [54.3%] vs. 16/35 [45.7%]; p = 0.61). There was also a significant difference in favor of men for Gram-positive bacteria as the causative agent of dpCRBSI (252/360 [70.0%] vs. 108/360 [30.0%]; p < 0.001) with no difference in the proportion of Gram-negative bacteria (23/41 [56.1%] vs. 18/41 [43.9%]; p = 0.43).

As outlined in Figure 1, the PSM cohort comprises 3,024 CVCs, of which 1,512 were inserted in men and 1,512 in women. The parameters age, distribution of underlying diseases, neutropenia, insertion site, use of CVC dressings and coated CVCs, complications during CVC insertion and CVC in situ time, patients’ and CVC characteristics are similar to those of the entire cohort; as in the entire cohort the rate of patients with hematological malignancies was also high at 96.4% (Table 2). With respect to the dCRBSI rate, there was no statistically significant difference between men and women (81/1,512 [5.4%] vs. 62/1,512 [4.1%]; p = 0.12), but more men had dpCRBSI (150/1,512 [9.9%]) than women (101/1,512 [6.7%]; p = 0.002). Also, the dCRBSI incidence was similar between both sexes (3.8 vs. 2.9/1,000 CVC days; p = 0.11), but the incidence of dpCRBSI was higher in men than in women (7.0 vs. 4.7/1,000 CVC days; p = 0.002). The time to CRBSI onset was not different between both groups (Table 2), and the same applies for the cumulative dCRBSI incidence in the time-to-event analysis (p = 0.11; online suppl. Fig. S2). However, the cumulative dpCRBSI incidence was higher in men than in women (p = 0.001; online suppl. Fig. S3). Regarding the distribution of the causative pathogens with regard to sex, no differences were seen with respect to dCRBSI and dpCRBSI (Table 2). However, CoNS were more frequently identified as causative pathogen of dCRBSI in men (60/102 [58.8%]) than in women (42/102 [41.2%]; p = 0.07). The same was true for dpCRBSI and CoNS (112/182 [61.5%] vs. 70/182 [38.5%]; p = 0.002). As in the entire cohort, Gram-positive bacteria in total were more frequently found as causative pathogens of dCRBSI in men than in women (69/116 [59.5%] vs. 47/116 [40.1%]; p = 0.04], while there was no difference in the proportion of Gram-negative bacteria (8/18 [44.4%] vs. 10/18 [55.6%]; p = 0.64). The same was seen for dpCRBSI and Gram-positive bacteria in total (130/212 [61.3%] vs. 82/212 [38.7%]; p < 0.001) and Gram-negative bacteria in total (9/21 [42.9%] vs. 12/21 [57.1%]; p = 0.51). By multivariable regression, neutropenia proved to be an independent risk factor for dCRBSI (Table 3), and both neutropenia and male sex turned out to be a risk factor for dpCRBSI (Table 4).

An important finding from this study of CRBSI in patients with cancer is that men have a higher rate and a higher incidence of dpCRBSI than women. Male sex also proved to be an independent risk factor for dpCRBSI. Furthermore, both dCRBSI rate and dCRBSI incidence were higher in men compared to women even though the difference was not statistically significant due to the small sample size.

Using PSM to reduce confounding variables [32], roughly 40% of CVCs were excluded because a sex-specific matching was not sufficiently possible. However, the characteristics of both cohorts and the findings on CRBSIs were not substantially different, indicating that the data can be considered as reliable.

Data on the comparison of CRBSI between men and women, and especially in those with cancer, are rare. In a study of 64,575 CVCs in 50,254 patients with 1 versus 2 CVCs in use, men tended to be at higher risk of CRBSI in multivariate analysis {risk ratio 1.36 (95% confidence interval [CI]: 0.99–1.89)} [25]. However, only a minority of patients (3.6%) had received chemotherapy, and the authors investigated central line-associated BSIs rather than CRBSIs, potentially overestimating the rate of CVC-derived BSI in patients with cancer [23]. In another large study on 82,225 patients, of which 11.9% had malignancies and 7.5% CVCs, BSIs occurred more often in men than in women (odds ratio [OR] 1.23 [95% CI: 1.11–1.35]), but CRBSI rates were not reported [24].

The reasons for the increased risk of CRBSI in men remain largely unclear. Genetic factors and sex steroid hormones may play a role as well as lifestyle, socioeconomic, and behavioral factors, which all contribute to the so-called sexual dimorphism in bacterial infections. The sexual dimorphism is in part explained by the pro-inflammatory properties of estradiol and by the anti-inflammatory effects of testosterone via the interaction with their specific intracellular receptors [33, 34]. However, it remains unexplained how this may impact the risk of CRBSI. A previous study showed that the risk of CVC tip colonization, an established risk factor for CRBSI [23, 35], is higher in men than in women (hazard ratio 2.04 [95% CI: 1.12–3.85]; p = 0.02) [36]. This finding was confirmed in a recent study on 9,924 CVCs that also showed men to be at higher risk of CVC-related infection than women [37]. Further, men are at higher risk for skin colonization at the CVC insertion site, which is related to an increased CRBSI risk [38]. We agree with previous hypotheses that beards and shaving not only promote the multiplication of pathogens but also result in reduced adherence of CVC dressings with an increased risk of bacterial contamination [38‒40], with skin colonization being associated with a higher risk of CVC-associated BSI (OR: 1.15 [95% CI: 1.02–1.29]; p = 0.03) [39]. Since data on beards/shaving is not being collected in the SECRECY registry, these risk factors cannot be investigated. Unless there were signs of local contamination, inflammation, or detachment, CVC dressings were renewed every 2–3 days for gauze or every 7 days for transparent dressings, respectively, according to guidelines [23]. However, data on actual replacement intervals are not available in the registry. Whether CVC dressings detach more often in men than in women and, as a consequence, need to be renewed more frequently, and whether this actually represents a risk factor for CRBSI, cannot be further analyzed in the present data set. However, these data only seem to apply to jugular vein inserted CVCs: in a subgroup analysis of subclavian vein CVCs with a small number of cases, we found fewer CRBSIs in men than in women (online suppl. Table S1). Moreover, in a prospective cohort study it was shown that male sex was an independent risk factor for skin colonization only in cases of jugular vein insertion (OR: 4.61 [95% CI: 2.14–9.89]) compared to other insertion sites (OR: 0.66 [95% CI: 0.18–2.42]) [38].

While it was shown in previous studies, that men have a higher risk for Gram-positive BSI than women, we show for the first time that men are also at higher risk of CRBSI caused by Gram-positive bacteria. However, given the higher proportion of men in the groups with dCRBSI and dpCRBSI caused by Gram-positive bacteria, the validity of this finding may be considered limited. Also, the present study was not designed to show differences in the sex distribution of CRBSI with regard to different types of bacteria as causative pathogens, which should be analyzed further in future studies. Our finding nevertheless may give a rationale for enhancing measures that specifically address infections caused by skin commensals like chlorhexidine washes [41] in men at high risk of Gram-positive CRBSI.

While there is evidence for an inferior cancer-specific survival and overall survival [42‒45], and for increased mortality rates due to BSIs in men compared to women [20], we are not able to present mortality data since survival data are not recorded within the SECRECY registry. In conclusion, this sex-sensitive, matched analysis showed that men with hematologic malignancies are at higher risk for CRBSI than women. The question whether CVCs should therefore be removed earlier in men should be addressed in future studies.

The authors thank all SECRECY staff for their great support and documentation. SECRECY (Study to Evaluate Central Venous Catheter-related Infections in Hematology and Oncology) was registered in the German Clinical Trial Register (DRKS; No. DRKS00006551; https://drks.de/search/en/trial/DRKS00006551).

The study was approved by the central Ethics Committee (Magdeburg University Hospital, Approval No. 84/14) as well as by respective local Ethics Committees. The need for informed consent was waived by the central Ethics Committee (Magdeburg University Hospital, Approval No. 84/14).

The authors have no conflicts of interest to declare.

This study was not supported by any sponsor or funder.

E.S. was involved in patient management, designed the study, collected, analyzed and interpreted the data, and wrote the manuscript. A.S. and N.G. revised the manuscript. B.B., J.G.-S., O.K., J.L., A.M., J.-H.N., J.N., J.P., M.S.-H., R.S.-B., D.T., P.W., and K.W. was involved in patient management, collected the data, and revised the manuscript. M.V.L.-T. interpreted the data and revised the manuscript. M.H. was involved in patient management, collected and interpreted the data, and wrote the manuscript.

Marie von Lilienfeld-Toal and Marcus Hentrich contributed equally to this work as senior authors.No generative artificial intelligence (AI) and AI-assisted technologies were used in the writing process.

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.