Bloodstream infections and other frequently encountered infections in clinical practice, such as pneumonia and complicated urinary tract infections (UTIs), caused by Carbapenem-resistant Enterobacterales (CRE) are associated with increased morbidity and mortality (Wright et al., 2021). More recently, global dissemination of CRE has occurred at an alarmingly rapid pace (Logan and Weinstein, 2017). Internationally, the focus has shifted toward early detection and infection control (Logan and Weinstein, 2017; Patel and Bonomo, 2013). CRE have become endemic in some regions of Europe, Asia, South America, and Africa, with high-risk patients, such as solid organ transplant (SOT) recipients, bone marrow transplant (BMT) recipients, and patients with hematological malignancies, being particularly vulnerable to infection due to frequent therapy with antibiotics and prolonged hospital stays (Pouch and Satlin, 2017; Nordmann et al., 2011).

Solid organ transplant recipients, BMT recipients, and patients with hematological malignancies have experienced increasing morbidity due to infections caused by multidrug-resistant CRE (Kerneis and Lucet, 2019; Patel et al., 2008). Approximately 1.8–2.0% of BMT recipients experienced incident bloodstream infections caused by CRE, and the 30-day mortality rates ranged from 52 to 63% following bacteremia in patients with hematological malignancies (Pouch and Satlin, 2017; Yang et al., 2020). Approximately one-quarter (range of 16–24%) of all patients with CRE bacteremia are patients with hematological malignancies (Pouch and Satlin, 2017).

The incidence of CRE infection among SOT recipients ranges from 3.0 to 23%, and overall mortality rates associated with CRE infection range from 24 to 70% (Freire et al., 2019; Aguado et al., 2018; Gutiérrez-Gutiérrez et al., 2017; Tzouvelekis et al., 2012). Although a recent matched cohort study of patients with CRE found that SOT patients with CRE did not have worse outcomes than non-transplant patients with CRE, 90-day mortality remained substantial among the SOT patients. The study concluded that reducing hospital stay to prevent CRE infection should be a goal for transplant programs (Boutzoukas et al., 2025). Overall, data on patients with CRE infections and immunosuppression remain limited.

A previous study conducted in Saudi Arabia on hospitalized patients showed that OXA-48 and NDM carbapenemases are the most common enzymes found in CRE infections caused by Klebsiella pneumoniae (Alraddadi et al., 2022). Both this multicenter prospective study and a more recent retrospective cohort study conducted in the region provided specific details regarding CRE colonization and infection in high-risk patients (Alraddadi et al., 2022; Alraddadi et al., 2024). The current study aimed to further describe the epidemiology and outcomes associated with CRE infection in high-risk SOT patients, BMT recipients, and patients with hematological malignancies in Saudi Arabia.

The study population for this retrospective study comprised 155 patients with CRE infections, of whom 118 (76.1%) had received a solid organ transplant (SOT) and 37 (23.9%) had a bone marrow transplant (BMT) or had a hematological malignancy (Table 1). There were 64 (54.2%) patients who received kidney transplants, 42 (35.6%) patients who received liver transplants, one (0.9%) patient who received a heart transplant, 13 (11.0%) patients who received a lung transplant, and two (1.7%) patients who received a small bowel/intestine transplant. In addition, four patients received both kidney and liver transplants. The median [interquartile range (IQR)] age was 53.5 (38.1–64.4) years, and 56.8% of the population was male. The most common comorbidities among the patients were diabetes mellitus (41.9%) and hypertension (40.0%) (Table 1).

While 16.8% of the patients had documented CRE colonization within the last 90 days, almost one-third of the patients (32.3%) had a UTI (Table 1). The median (IQR) Charlson Comorbidity Index was 3 (2–6). Among the 72 (46.5%) patients with CRE bacteremia and a Pitt bacteremia score, the median score was 2.0 (0.0–5.5). More than three-quarters of the isolated organisms (76.8%) were K. pneumoniae, and the bla_OXA-48 (57.4%) gene was the most common molecular type, followed by bla_NDM (32.9%).

Following the hospital protocol for treatment, definitive therapy for CRE infection included ceftazidime/avibactam for 67.7% of the patients, while 27.7% received aztreonam. The majority of the patients with the bla_OXA-48 gene received ceftazidime/avibactam, while those with the bla_NDM gene were treated with a combination of ceftazidime/avibactam and aztreonam (Table 1). Among the 61 patients with available sensitivity data for ceftazidime/avibactam and the blaOXA-48 molecular type, 17 (27.9%) showed resistance to this treatment regimen, while 44 patients (72.1%) had susceptible isolates. However, when analyzing the subgroup of 43 patients with only the blaOXA-48 gene, only two patients (4.7%) demonstrated resistance to ceftazidime/avibactam. The median (IQR) MIC value for ceftazidime/avibactam for patients with NDM and patients with OXA-48 was 16 (1–16).

There were 64 patients (41.3%) admitted to the intensive care unit (ICU) during their hospital stay, and over one-third of all patients (36.1%) experienced acute kidney injury (Table 2). Just over one-third (34.2%) of patients died within 90 days of their first CRE culture, with 67.9% of these deaths being attributable to CRE (Table 2). Attributable mortality was defined as death directly linked to CRE infection, including cases with no other clear cause of death or cases where the patient had persistent infection without clinical improvement until death.

The association between patient type and mortality was significant in this patient population, where 34.0% of the patients who died were BMT recipients compared to 18.6% of survivors who were BMT recipients (p = 0.03) (Table 3). Among the patients who died, 35.9% had hospital-acquired pneumonia/ventilator-associated pneumonia (HAP/VAP), compared to 10.8% of the patients who survived with this infection (p < 0.001). The association between the type of carbapenemase found in the CRE isolates from the patients’ infections and mortality was also significant. Among the deaths, 75.5% (n = 53) occurred in the patients with isolates harboring blaNDM, blaOXA-48, or both blaNDM and blaOXA-48. While 62.3% of the patients who died developed CRE bacteremia, 38.2% of survivors were bacteremic (p = 0.003) (Table 3).

Predicting mortality using a multivariable logistic regression model indicated that the BMT recipients and patients with hematological malignancies were 2.84 times more likely to die within 90 days of their first positive culture compared to the SOT recipients [adjusted odds ratio (AOR) = 2.84, 95% confidence interval (CI) = 1.01–8.01, p = 0.049] (Table 4). The participants who developed HAP/VAP were also significantly more likely to experience 90-day all-cause mortality (AOR = 3.70, 95% CI = 1.33–10.30, p = 0.012). Compared to the patients with the blaNDM gene, the patients with an infection harboring the blaOXA-48 gene were 3.97 times more likely to die within 90 days of the first culture (AOR = 3.97, 95%CI = 1.04–15.15, p = 0.044). Having CRE bacteremia and acute kidney injury were also associated with an increased likelihood of experiencing 90-day all-cause mortality (Table 4). A multivariable model analyzing the patients with CRE bacteremia (n = 72) revealed that the BMT recipients (n = 29 of the 72, 40.3%) had an increased likelihood of experiencing 90-day mortality compared to the SOT recipients (AOR = 3.71, 95% CI = 1.05–13.14, p = 0.042). In addition, for every one-unit increase in the Pitt bacteremia score, the patients were 1.32 times more likely to die within 90 days of their first positive culture (p = 0.005) (Table 5).

This study reveals that mortality is alarmingly high among immunosuppressed SOT patients and BMT recipients, with more than one-third of the population dying within 90 days of their first positive culture and just over two-thirds of these deaths being attributable to CRE infection. Furthermore, BMT recipients, patients with hematological malignancies, and patients infected with CRE carrying the blaOXA-48 gene are at increased risk of 90-day mortality. Given the high prevalence of blaOXA-48 CRE infections in the region, this is particularly concerning due to the heightened risk these high-risk patients face of developing infections, compounded by recurrent antibiotic use and prolonged hospital stays. Despite the availability of effective antibiotics for CRE, it is alarming that mortality in this patient population remains high. This is likely due to delays in initiating the appropriate antibiotic therapy or the development of resistance. These factors highlight the urgent need for rapid diagnostic tools with quick turnaround times to detect CRE and the current unmet needs regarding the limited treatment options for CRE.

Due to sample size considerations, the SOT patients and the BMT recipients were combined in the analyses. Overall, 29.7% of the SOT patients died compared to 48.6% of the BMT recipients, which was not unexpected considering the severity of illness among many BMT recipients. While more than three-quarters of all patients had a K. pneumoniae infection, 37.8% of these patients died, compared to 22.2% of the patients with an E. coli infection. Running the multivariable model with only the SOT patients revealed similar findings, except that the association with 90-day mortality was no longer significant for the patients with the blaOXA-48 isolate or for the bacteremic patients. When the lung, heart, and small bowel SOT patients (n = 16) were removed from the SOT model, the results were similar to the original model, except that the association between blaOXA-48 isolates and mortality was no longer significant.

A matched case–control study compared patients with carbapenem-resistant K. pneumoniae to patients with carbapenem-susceptible K. pneumoniae infections as controls (Patel et al., 2008). Carbapenem-resistant K. pneumoniae infection was found to be associated with recent organ or stem-cell transplantation, and the cases were more likely than the controls to die during hospitalization and from the infection (Patel et al., 2008). SOT patients have increased susceptibility to infections related to long-term immunosuppression. Although there are specific guidelines, it is still challenging to determine how this condition can be managed in recipients with CRE infections (Pérez-Nadales et al., 2022).

Carbapenem resistance has been shown to be a threat to allograft and patient survival, with limited antibiotics available for treatment (Patel et al., 2010). Among SOT recipients in CRE-endemic areas, 3.0–10% develop infections correlated with the transplanted organ (Satlin et al., 2014). Active surveillance is needed to monitor and prevent infection in immunocompromised patients, especially considering that empirical therapy is not active against CRE and its identification can take 2–4 days (Satlin et al., 2014; Blaschke et al., 2012). In addition, several studies have revealed high in-hospital mortality rates for CRE bacteremia among greatly immunocompromised BMT recipients and patients with hematological malignancies, with the majority of deaths being attributable to the infection (Satlin et al., 2014; Satlin et al., 2013; Muchtar et al., 2012; Snitkin et al., 2012). Analyses from a study indicated that considering genomic and epidemiological data simultaneously may assist in controlling CRE hospital-acquired infections (Snitkin et al., 2012). More information is needed regarding factors that may be specific to immunocompromised patients to facilitate CRE prevention in hospitals (Satlin et al., 2014).

The results from our study revealed that a number of factors were independently associated with 90-day all-cause mortality, including the type of patient (BMT compared to SOT), bacteremia, hospital-acquired pneumonia/ventilator-associated pneumonia (HAP/VAP), CRE infection carrying blaOXA-48, and acute kidney injury. A multivariable model for the bacteremic patients (n = 72) indicated that being a BMT patient (reference SOT) and a higher Pitt bacteremia score were significant predictors of mortality.

A recent study with some similar findings reported that 30-day mortality of CRE infection was independently associated with higher Simplified Acute Physiology Score II, sepsis at the time of diagnosis, pneumonia, monotherapy, and inappropriate empiric antibiotic therapy (So-Ngern et al., 2023). Additional studies have shown that factors associated with 14-, 28-, and 30-day mortality, infection-related mortality, and overall hospital mortality in patients with CRE infections include chronic renal failure, high APACHE II score, presentation with septic shock and severe sepsis, neutropenia, dialysis, age over 60 years, ultimately fatal disease, and rapidly fatal underlying diseases (Li et al., 2019; Tumbarello et al., 2015; de Maio Carrilho et al., 2016; Daikos et al., 2014). In our study, we were not able to determine why patients with infections caused by CRE-producing OXA-48 were more likely to die compared to those infected with other carbapenemase-producing strains. It is not clear whether this is due to the development of further antimicrobial resistance or the associations with other types of resistance genes. The most plausible explanation is the presence of unmeasured confounders not captured in our analysis, most likely related to the timing of antibiotic initiation and the appropriateness of the empirical regimen for this specific carbapenemase enzyme.

The current study benefited from the inclusion of a decent sample size of immunocompromised patients with molecular testing for CRE infection, confirming previous findings in the region regarding molecular identification for CRE infection. These studies also showed that a positive CRE screening test increased the likelihood of 30-day mortality in high-risk patients (Alraddadi et al., 2024). To date, this is the largest reported cohort of post-transplant patients with CRE in the region. In addition, our study confirmed that factors such as pneumonia and bacteremia (n = 72) are independently associated with mortality in patients with CRE, in line with findings from previous studies. Furthermore, we demonstrated that the BMT recipients, compared to the SOT patients, and patients with molecular identification of blaOXA-48 in their CRE infections were at greater risk for 90-day all-cause mortality in Saudi Arabia.

Although we were able to collect data on several potential predictor variables associated with CRE infection and mortality, the retrospective nature of our study and reliance on medical records may have limited our ability to measure or account for certain relevant factors in the analyses. It is possible that relevant clinical differences exist between BMT recipients and patients with hematological malignancies, although these patients were combined in the analysis due to sample size constraints. In addition, sequencing was not performed in this study, which may have predicted outcomes from CRE infection more accurately. The absence of sequencing may have limited our ability to detect virulence factors or resistance mechanisms that could have contributed to the observed increase in mortality. In conclusion, our study confirmed previous findings regarding the high mortality rate associated with CRE infection in immunocompromised populations. Future research should focus on developing effective strategies to care for immunocompromised patients with prolonged hospital stays to reduce the risk of acquiring CRE infections.