Abstract
Background:
Patients with cancer who had sepsis experience a higher mortality risk than noncancer patients. However, it remains unclear how this risk varies with cancer type and metastasis status, sex, age, and infecting microbe.
Methods:
This nationwide cohort study included all noncancer patients and those with cancer ≥18 years hospitalized with sepsis in Norway from 2008 to 2021. Cancer status, sepsis, and hospital mortality were identified from the Norwegian Patient Registry (International Classification of Diseases 10th revision codes). Multivariable regression analyses estimated absolute and relative risks (RR) of hospital mortality across subgroups.
Results:
Of 222,832 hospitalized patients with sepsis, 37,692 (16.9%) had cancer. Hospital mortality was higher in patients with cancer, at 16.9% (males) and 16.2% (females) with nonmetastatic disease, and at 27.1% (males) and 26.2% (females) with metastatic disease. Compared with noncancer patients, adjusted RRs (95% confidence interval) of hospital death were 1.39 (1.34–1.44, males) and 1.63 (1.55–1.71, females) for nonmetastatic cancer and 2.27 (2.18–2.37, males) and 2.75 (2.62–2.89, females) for metastatic cancer. The type of cancer was a key prognostic factor. The association between cancer and mortality was strongest in metastatic patients below 50 years [males 40–49 years: RR, 5.94 (4.43–7.97); females 18–39 years: RR, 8.28 (5.12–13.37)] and in those with gram-negative sepsis.
Conclusions:
The increased hospital mortality in patients with cancer varied with cancer type and the presence of metastasis. The association between cancer and mortality was strongest in females and young adults, as well as in gram-negative sepsis.
Impact:
Knowledge of the specific aspects of sepsis in patients with cancer may improve cancer care and guide future research on targeted sepsis therapies.
Introduction
Sepsis, defined as life-threatening organ failure resulting from a dysregulated immune response to infection, is estimated to cause 20% of global deaths (1, 2). Patients with cancer are particularly vulnerable to severe infections and sepsis due to impaired immune function, both from the malignancy itself and cancer treatment (3). Among patients with cancer who develop sepsis, mortality is consistently higher than in noncancer patients, with reported rates ranging between 20% and 65% (4–7). This excess risk underscores the urgent need to move beyond “one-size-fits-all” management. Nonetheless, despite decades of sepsis research, no specific therapy exists, and current treatment continues to rely on antibiotics and supportive care (8). The marked heterogeneity of sepsis further emphasizes the need for more personalized approaches, similar to oncology, in which patient stratification based on clinical characteristics, biological features, and prognosis is essential (9, 10).
Recent large U.S. database studies confirm the excess risk, with patients with cancer showing higher mortality than noncancer patients. In the National Readmissions Database (2013–2014), the adjusted odds ratio for in-hospital mortality was 1.6 (27.9% vs. 19.5%), and in the National Inpatient Sample (NIS, 2006–2014), the adjusted hazard ratio for 30-day mortality was 1.25 (20.0% vs. 13.9%; refs. 5, 11). Hematologic, respiratory, and abdominal cancers showed the highest mortality. However, the influence of metastatic disease was not examined in these studies, despite evidence from others identifying it as an important risk factor (6, 12, 13). In both U.S. cohorts larger mortality differences were found in younger age groups, but sex was not included as a modifier despite biological evidence that males are at higher risk of cancer, infection, and death (14–17). In addition, NIS data indicated the strongest association with mortality for gram-negative pathogens, which is concerning given the global shift toward more gram-negative infections and increasing antimicrobial resistance (18).
Most other studies comparing noncancer patients and those with cancer focus on small intensive care unit (ICU)–based cohorts although the majority of patients with sepsis receive treatment in general wards (17, 19, 20). Moreover, the U.S. data collected up to 2014 may not be directly transferable to European healthcare systems or contemporary practice. Notably, advancements in cancer diagnosis and treatment, including immunomodulating therapies and more individualized medicine, have altered the side effects of cancer treatment and the infection responses in patients with cancer (21). Taken together, these limitations emphasize the need for updated population-based studies of sepsis outcomes in patients with cancer across all hospital departments. Furthermore, patients with cancer remain underrepresented in clinical sepsis trials, underlining the importance of robust observational data to improve risk stratification in this group (22).
As both cancer and sepsis are age-dependent, the burden of both diseases is expected to increase with demographic changes. The main aim of this study was, therefore, to investigate whether patients with cancer who had sepsis differ from noncancer patients with sepsis in terms of clinical characteristics and hospital mortality. Using national registry data, this unselected prospective cohort study explores the association between cancer and hospital mortality across subgroups of patients with sepsis defined by sex, age, and infecting microbe. By presenting both absolute and relative risk (RR) estimates, we aim to provide a more comprehensive understanding of these associations.
Materials and Methods
Study design and participants
All patients in Norway aged 18 years or older hospitalized with a first episode of sepsis between January 1, 2008, and December 31, 2021, were included in a nationwide, prospective cohort study. Sepsis was defined according to the Angus implementation, as redefined by Rudd and colleagues (2). Eligible patients were identified using International Classification of Diseases, 10th revision (ICD-10) codes. Patients were classified as having explicit sepsis if ICD-10 codes explicitly referenced sepsis (Supplementary Methods). Implicit sepsis was identified based on ICD-10 codes indicating infection combined with acute organ dysfunction. Unless otherwise specified, our analyses include both explicit and implicit sepsis. The codes defining COVID-19–related sepsis were incorporated into the study from February 28, 2020, the date of the first confirmed case in Norway. We used this strategy in the primary and up to 20 secondary coexisting ICD-10 discharge codes (23).
Among these sepsis hospitalizations, patients with cancer were identified by ICD-10 codes from the same hospital stay, ensuring valid data on active cancer and the presence of metastases (Supplementary Methods).
All data were obtained from the Norwegian Patient Registry (NPR) on an individual level using the national identity number. NPR is a national health registry maintained by the Norwegian Directorate of Health. Reporting to NPR is mandatory, leading to minimal missing data and high completeness (24). The diagnostic codes in NPR have been shown to correspond closely with those recorded in the Cancer Registry of Norway, supporting their validity (25). To obtain information on whether the patients were admitted to the ICU or not, we linked NPR with the Norwegian Intensive Care Registry (NIR), applying distributed linkage across registries at the individual level using national identification numbers. NIR covers all intensive care admissions in Norway since May 1, 2014.
Exposure and outcome definition
Information on sex, age, comorbid conditions, cancer, sepsis characteristics, and date of hospitalization was obtained from the NPR database.
The outcome variable was in-hospital, all-cause mortality during the first sepsis admission, based on NPR records. Cancer diagnosis was the primary exposure, with patients grouped by cancer type and presence of metastasis to capture potential heterogeneity in the association with hospital mortality (Supplementary Methods). Patients with multiple active cancers were included in the analyses for each relevant cancer type. Potential confounders included age and comorbidities. Age was categorized into groups (18–39, 40–49, 50–59, 60–69, 70–79, and ≥80 years). Comorbidities were identified by ICD-10 codes from the same hospital stay to ensure clinical relevance and were classified both by number (0, 1, 2, ≥3) and type, including chronic lung disease, cardiovascular disease, diabetes, dementia, chronic kidney disease, chronic liver disease, and chronic immune failure (Supplementary Methods; ref. 26). Our aim was to study cancer diagnosis as a risk factor for death in patients with sepsis, and we therefore considered sepsis characteristics merely as mediators in the association between cancer and sepsis. Characteristics of sepsis included the site of infection, organ system with acute dysfunction, subtypes of explicit sepsis, and the presence of neutropenia. Sites of infection were categorized as respiratory, genitourinary, skin and soft tissue, abdominal, gastrointestinal, postprocedural infections, and other sites [bone, joint, endocarditis/myocarditis, obstetric, ear, mouth, upper airway, central nervous system (CNS), and unknown]. Acute organ dysfunctions were classified both by number and type; types included respiratory, circulatory, renal, hepatic, coagulation, and other (acidosis, unspecific gangrene, CNS, and systemic inflammatory response syndrome of infectious origin with organ dysfunction).
Statistical analysis
Patient characteristics were summarized using the mean and standard deviation (SD) for continuous variables and absolute and relative frequencies for categorical variables. We used generalized linear models with a log link, hereafter referred to as log-binomial models, to estimate RRs, comparing mortality for noncancer patients and those with nonmetastatic and metastatic cancer. Age and the number of comorbidities were considered confounders and included as categorical variables, as these factors may influence immune function, which is important for both cancer and sepsis outcomes. Adjustment for the number of comorbidities was chosen because this approach better accounts for multimorbidity and the interdependency between conditions while still allowing comparison of the overall comorbidity burden between noncancer patients and those with cancer. To investigate whether the association between cancer and mortality risk was different in young patients, we estimated multivariable models with a multiplicative interaction effect and further estimated the relative excess risk due to interaction to assess possible additive interaction effects (using Stata’s reri command). For these analyses, age was dichotomized as <40 years (“young patients with cancer”) and 40+ years, based on the definition of adolescents and young adults with cancer (27). All analyses were stratified by sex, as associations may vary between sexes due to biological differences affecting immune responses, as well as the different distribution of cancer types. To facilitate comparison with previous studies, we also ran a logistic regression analysis for men and women combined. Separate analyses were conducted for all cancers combined and for subtypes, acknowledging the heterogeneity of cancer etiology and treatment. Rare cancer types, such as those affecting other endocrine glands and the gallbladder/bile duct, were excluded in these cancer-specific analyses as the number of cases was too small to allow meaningful estimates. Because chemotherapy disrupts the gut microbiota and antimicrobial resistance is increasing, we conducted exploratory analyses stratified by infecting microbes. As neutropenia is common in patients with cancer but largely absent in noncancer patients, we examined its association with mortality exclusively within the cancer population. To investigate whether the differing distribution of obstetric infections between women with and without cancer influenced the estimates, we conducted a sensitivity analysis by excluding patients with obstetric infections.
In addition to reporting RR estimates, we also calculated absolute risk estimates of hospital death by predicting individual risks from the fitted regression models and averaging over the age and comorbidity distribution of the whole cohort, using Stata’s postestimation “margins” command.
All analyses were conducted using STATA version 18 (RRID: SCR 012763). All P values were calculated with a statistical significance level of <0.05.
Ethics approval and consent to participate
The study received approval from the Regional Committee for Medical and Health Research Ethics (REK) in South-Eastern Norway (2019/42,772) and the Data Access Committee at Nord-Trøndelag Hospital Trust (2021/184). Written patient consent was not required because our study relied on deidentified registry data, in compliance with the Norwegian Act on Medical and Health Research.
Results
Among 12,619,803 adult hospital admissions, 317,705 patients met the criteria for sepsis, of whom 222,832 (120,442 male; 102,390 female) were hospitalized with a first sepsis episode. Among these, 22,653 males and 15,039 females had a cancer diagnosis, and 7,243 male and 5,448 female patients with cancer had metastatic disease (Supplementary Fig. S1). Patients with cancer accounted for 16.9% of all first sepsis hospitalizations in Norwegian hospitals during the study period. The proportion of patients with cancer was higher among male patients with sepsis (18.8%) compared with female patients with sepsis (14.7%).
Clinical characteristics in noncancer patients and those with cancer
The mean age was 70.5 years for males and 72.5 years for females in the noncancer group, 70.7 and 68.4 years in the nonmetastatic cancer group, and 70.3 and 67.0 years in the metastatic cancer group (Table 1). Male and female patients with cancer who had sepsis had fewer comorbid conditions, both in terms of the total number and the prevalence of individual comorbidities, compared with their noncancer counterparts (Table 1).
Table 1.
Characteristics of noncancer patients and those with cancer who had sepsis, Norwegian hospitals, 2008 to 2021.
| | Noncancer | Cancer | ||||
|---|---|---|---|---|---|---|
| Male n = 97,789 |
Female n = 87,351 |
Male n = 22,653 |
Female n = 15,039 |
|||
| Nonmetastatic n = 15,410 |
Metastatic n = 7,243 |
Nonmetastatic n = 9,591 |
Metastatic n = 5,448 |
|||
| Age, years | 70.5 (16.5) | 72.5 (17.8) | 70.7 (13.6) | 70.3 (12.2) | 68.4 (14.4) | 67.0 (12.6) |
| Age range, years | | | | | | |
| 18–39 | 6,340 (6.5%) | 6,561 (7.5%) | 537 (3.5%) | 169 (2.3%) | 438 (4.6%) | 158 (2.9%) |
| 40–59 | 15,336 (15.7%) | 11,155 (12.8%) | 2,129 (13.8%) | 1,009 (13.9%) | 1,825 (19.0%) | 1,271 (23.3%) |
| 60–79 | 43,777 (44.7%) | 32,708 (37.4%) | 8,742 (56.7%) | 4,524 (62.5%) | 5,229 (54.5%) | 3,242 (60.5%) |
| ≥80 | 32,336 (33.1%) | 36,927 (42.3%) | 4,002 (26.0%) | 1,541 (21.3%) | 2,099 (21.9%) | 777 (14.3%) |
| Number/type of comorbidities | | | | | | |
| 0 | 35,728 (36.5%) | 33,422 (38.3%) | 7,630 (49.5%) | 4,150 (57.3%) | 5,560 (57.9%) | 3,572 (65.6%) |
| 1 | 40,937 (41.9%) | 36,942 (42.3%) | 5,575 (36.2%) | 2,448 (33.8%) | 3,037 (31.7%) | 1,494 (27.4%) |
| 2 | 17,671 (18.1%) | 14,688 (16.8%) | 1,896 (12.3%) | 574 (7.9%) | 888 (9.3%) | 341 (6.3%) |
| ≥3 | 3,453 (3.5%) | 2,299 (2.6%) | 309 (2%) | 71 (1%) | 106 (1.1%) | 41 (0.7%) |
| Chronic lung disease | 15,881 (16.2%) | 16,554 (19.0%) | 1,714 (11.1%) | 567 (7.8%) | 1,021 (10.7%) | 429 (7.9%) |
| Cardiovascular disease | 47,614 (48.7%) | 39,501 (45.2%) | 6,098 (39.6%) | 2,410 (33.3%) | 3,033 (31.6%) | 1,408 (25.8%) |
| Diabetes | 12,551 (12.8%) | 8,936 (10.2%) | 1,398 (9.1%) | 567 (7.8%) | 625 (6.5%) | 340 (6.2%) |
| Dementia | 3,675 (3.8%) | 3,955 (4.5%) | 256 (1.7%) | 99 (0.8%) | 115 (1.2%) | 44 (0.8%) |
| Chronic kidney disease | 5,038 (5.2%) | 3,099 (3.4%) | 513 (3.3%) | 168 (2.3%) | 167 (1.7%) | 54 (1%) |
| Chronic liver disease | 591 (0.6%) | 354 (0.4%) | 31 (0.2%) | 6 (0.1%) | 10 (0.1%) | 2 (0.04%) |
| Chronic immune failure | 1,575 (1.6%) | 1,035 (1.2%) | 298 (1.9%) | 41 (0.6%) | 168 (1.8%) | 23 (0.4%) |
| Site of infection | | | | | | |
| Respiratory | 39,054 (39.9%) | 32,520 (37.2%) | 4,478 (29.1%) | 2,119 (29.3%) | 2,411 (25.1%) | 1,300 (23.9%) |
| Genitourinary | 17,874 (18.3%) | 21,258 (24.3%) | 2,325 (15.1%) | 1,120 (15.5%) | 1,382 (14.4%) | 823 (15.1%) |
| Skin and soft tissue | 4,324 (4.4%) | 3,106 (3.6%) | 368 (2.4%) | 142 (2%) | 224 (2.3%) | 102 (1.9%) |
| Abdominal | 4,944 (5.1%) | 4,839 (5.5%) | 903 (5.9%) | 523 (7.2%) | 664 (6.9%) | 467 (8.6%) |
| Gastrointestinal | 5,014 (5.1%) | 4,562 (5.2%) | 527 (3.4%) | 171 (2.4%) | 367 (3.8%) | 170 (3.1%) |
| Following a procedure | 3,756 (3.8%) | 2,393 (2.7%) | 1,145 (7.4%) | 290 (4%) | 473 (4.9%) | 233 (4.3%) |
| Othera | 19,913 (20.4%) | 17,654 (20.2%) | 3,209 (20.8%) | 1,450 (20%) | 2,068 (21.6%) | 1,151 (21.1%) |
| Number of acute organ dysfunctions | | | | | | |
| 0 | 29,564 (30.2%) | 24,362 (27.8%) | 6,035 (39.2%) | 3,301 (45.6%) | 4,162 (43.4%) | 2,775 (50.9%) |
| 1 | 58,876 (60.2%) | 56,164 (64.3%) | 8,144 (52.9%) | 3,503 (48.4%) | 4,771 (49.7%) | 2,350 (43.1%) |
| 2 | 7,569 (7.7%) | 5,590 (6.4%) | 950 (6.2%) | 364 (5%) | 509 (5.3%) | 281 (5.2%) |
| ≥3 | 1,780 (1.9%) | 1,235 (1.5%) | 281 (1.8%) | 75 (1.0%) | 149 (1.6%) | 42 (0.8%) |
| Organ system with acute dysfunction | | | | | | |
| Respiratory | 27,869 (28.5%) | 26,433 (30.3%) | 3,242 (21%) | 1,338 (18.5%) | 1,974 (20.6%) | 1,009 (18.5%) |
| Circulatory | 7,459 (7.6%) | 5,208 (6.0%) | 993 (6.4%) | 436 (6.0%) | 496 (5.2%) | 300 (5.5%) |
| Renal | 32,587 (33.3%) | 26,368 (30.2%) | 3,792 (24.6%) | 1,726 (23.8%) | 1,808 (18.9%) | 959 (17.6%) |
| Hepatic | 1,338 (1.4%) | 1,133 (1.3%) | 206 (1.3%) | 205 (2.8%) | 141 (1.5%) | 186 (3.4%) |
| Coagulation | 1,279 (1.3%) | 999 (1.1%) | 2,003 (13.0%) | 418 (5.8%) | 1,423 (14.8%) | 349 (6.4%) |
| Otherb | 9,253 (9.5%) | 11,163 (12.8%) | 714 (4.6%) | 353 (4.9%) | 446 (4.6%) | 243 (4.5%) |
| Explicit sepsis | 31,643 (32.4%) | 26,368 (30.2%) | 7,563 (49.1%) | 3,650 (50.4%) | 5,077 (52.9%) | 3,079 (56.5%) |
| Unspecified | 11,513 (11.8%) | 9,347 (10.7%) | 3,527 (22.9%) | 1,737 (24.0%) | 2,536 (26.4%) | 1,449 (26.6%) |
| Gram-negativec | 8,223 (8.4%) | 8,185 (9.4%) | 1,814 (11.8%) | 880 (12.2%) | 1,111 (11.6%) | 742 (13.6%) |
| S. aureus | 3,296 (3.4%) | 2,085 (2.4%) | 287 (4.0%) | 516 (3.4%) | 199 (3.7%) | 278 (2.9%) |
| S. pneumoniae | 1,463 (1.5%) | 1,570 (1.8%) | 182 (1.2%) | 73 (1.0%) | 111 (1.2%) | 40 (0.7%) |
| Neutropenia | 543 (0.6%) | 511 (0.6%) | 2,917 (18.9%) | 704 (9.7%) | 2,412 (25.2%) | 719 (13.2%) |
| COVID-19d | 1,783 (13.2%) | 941 (8.9%) | 63 (3.3%) | 16 (1.8%) | 26 (2.4%) | 16 (2.5%) |
| ICU admittancee | 6,957 (12.3%) | 4,744 (9.8%) | 575 (7.0%) | 227 (5.8%) | 305 (6.1%) | 155 (5.3%) |
NOTE: Data are n (%) and mean (SD). Selected ICD-10 codes corresponding to each variable are presented in the Supplementary Methods.
Other infection sites were bone, joint, endocarditis/myocarditis, obstetric, ear, mouth, upper airway, CNS, and unknown.
Other organ systems with acute dysfunctions were acidosis, unspecific gangrene, CNS, and systemic inflammatory response syndrome.
Gram-negative bacteria excluding anaerobic bacteria and Haemophilus influenzae.
Period from February 27, 2021, to December 31, 2021.
Period from May 1, 2014, to December 31, 2021.
Patients with cancer were more likely than noncancer patients to have explicit sepsis (Table 1), with a higher frequency of unspecified sepsis and gram-negative sepsis. For both sexes and in both noncancer patients and those with cancer who had sepsis, the most common site of infection was respiratory, followed by genitourinary. Patients with cancer had fewer acute organ system dysfunctions compared with noncancer patients, both in terms of the number and frequency of individual acute organ dysfunctions, except for hepatic and coagulation. Both male and female patients with nonmetastatic cancer had more than a 10-fold increased prevalence of acute dysfunction in the coagulation system compared with noncancer patients. Neutropenia was observed in 18.9% of male and 25.2% of female patients with nonmetastatic cancer, compared with 9.7% of male and 13.2% of female patients with metastatic cancer. In contrast, only 0.6% of both male and female noncancer patients experienced neutropenia. Patients with cancer were less likely than noncancer patients with sepsis to be diagnosed with COVID-19. Both patients with nonmetastatic and metastatic cancer had a lower frequency of ICU stays compared with noncancer patients of the same sex. Specifically, 12.3% of male noncancer patients were admitted to the ICU, compared with 7.0% of males with nonmetastatic cancer and 5.8% of males with metastatic cancer. Among females, ICU admission occurred in 9.8% of noncancer patients, compared with 6.1% of those with nonmetastatic cancer and 5.3% of those with metastatic disease.
Clinical characteristics across major cancer types
In our analysis of sepsis across five major cancer types (respiratory, colorectal, breast, prostate, and hematologic), respiratory infections were the most common site in all subgroups when disregarding the category “Other infection sites,” except in prostate cancer, in which genitourinary infections predominated (Table 2). Acute respiratory dysfunction was the predominant acute organ dysfunction among male and female patients with respiratory cancer, as well as female patients with breast cancer, whereas acute renal dysfunction was most common in patients with prostate and colorectal cancers. Patients with colorectal cancer had the highest frequency of acute circulatory dysfunctions, whereas those with hematologic cancer had the highest burden of acute coagulation dysfunction. Explicit sepsis was most frequently observed in patients with metastatic colorectal cancer, those with hematologic cancer, and female patients with breast cancer. Neutropenia was most prevalent in patients with hematologic cancer, followed by those with nonmetastatic breast cancer and respiratory cancer (Table 2).
Table 2.
Sepsis characteristics in subgroups of patients with cancer, Norwegian hospitals, 2008 to 2021.
| Males | |||||||
|---|---|---|---|---|---|---|---|
| | Respiratory cancera | Colorectal cancer | Prostate cancer | Hematologic cancerb,c | |||
| Nonmetastatic n = 2,093 |
Metastatic n = 1,277 |
Nonmetastatic n = 1,429 |
Metastatic n = 1,318 |
Nonmetastatic n = 2,709 |
Metastatic n = 1,869 |
Nonmetastatic n = 3,462 |
|
| Site of infection | | | | | | | |
| Respiratory | 1,159 (55.4%) | 680 (53.3%) | 355 (24.8%) | 258 (19.6%) | 639 (23.6%) | 453 (24.2%) | 969 (28.0%) |
| Genitourinary | 129 (6.2%) | 84 (6.6%) | 204(14.3%) | 188 (14.3%) | 891 (32.9%) | 497 (26.6%) | 249 (7.2%) |
| Skin and soft tissue | 22 (1.1%) | 22 (1.7%) | 33 (2.3%) | 26 (2.0%) | 56 (2.1%) | 36 (1.9%) | 123 (3.6%) |
| Abdominal | 41 (2.0%) | 44 (3.5%) | 269 (18.8%) | 165 (12.5%) | 125 (4.6%) | 54 (2.9%) | 90 (2.6%) |
| Gastrointestinal | 48 (2.3%) | 29 (2.3%) | 50 (3.5%) | 29 (2.2%) | 88 (3.3%) | 36 (1.9%) | 152 (4.4%) |
| Infection following a procedure | 33 (1.6%) | 12 (0.9%) | 322 (22.5%) | 112 (8.5%) | 262 (9.7%) | 54 (2.9%) | 81 (2.3%) |
| Otherd | 265 (12.7%) | 182 (14.3%) | 238 (16.7%) | 266 (20.2%) | 503 (18.6%) | 379 (20.3%) | 893 (25.8%) |
| Organ system with acute dysfunction | | | | | | | |
| Respiratory | 975 (46.6%) | 461 (36.1%) | 406 (28.4%) | 203 (15.4%) | 366 (13.5%) | 242 (12.9%) | 490 (14.2%) |
| Circulatory | 139 (6.6%) | 64 (5.0%) | 170 (11.9%) | 97 (7.4%) | 147 (5.4%) | 94 (5.0%) | 170 (4.9%) |
| Renal | 373 (17.8%) | 224 (17.5%) | 505 (35.3%) | 301 (22.8%) | 935 (34.5%) | 520 (27.8%) | 720 (20.8%) |
| Hepatic | 11 (0.5%) | 17 (1.3%) | 11 (0.8%) | 60 (4.6%) | 14 (0.5%) | 18 (1.0%) | 44 (1.3%) |
| Coagulation | 101 (4.8%) | 87 (6.8%) | 28 (2.0%) | 27 (2.1%) | 40 (1.5%) | 116 (6.2%) | 995 (28.7%) |
| Othere | 77 (3.7%) | 58 (4.5%) | 59 (4.1%) | 59 (4.5%) | 202 (7.5%) | 96 (5.1%) | 100 (2.9%) |
| Explicit sepsis | 678 (32.4%) | 517 (40.5%) | 628 (44%) | 760 (57.7%) | 1,223 (45.2%) | 941 (50.4%) | 1,962 (56.7%) |
| Neutropenia | 352 (16.8%) | 195 (15.3%) | 61 (4.3%) | 118 (9.0%) | 60 (2.2%) | 114 (6.1%) | 1,128 (32.6%) |
| ICU admittancef | 93 (4.4%) | 45 (3.5%) | 89 (6.2%) | 48 (3.6%) | 115 (4.3%) | 58 (3.1%) | 113 (3.3%) |
| Females | |||||||
|---|---|---|---|---|---|---|---|
| | Respiratory cancer | Colorectal cancer | Breast cancer | Hematologic cancerb,c | |||
| Nonmetastatic n = 1,560 |
Metastatic n = 1,013 |
Nonmetastatic n = 1,000 |
Metastatic n = 915 |
Nonmetastatic n = 981 |
Metastatic n = 1,015 |
Nonmetastatic n = 2,346 |
|
| Site of infection | | | | | | | |
| Respiratory | 781 (50.1%) | 475 (46.9%) | 205 (20.5%) | 135 (14.8%) | 184 (18.8%) | 246 (24.2%) | 585 (24.9%) |
| Genitourinary | 180 (11.5%) | 100 (9.9%) | 162 (16.2%) | 124 (13.6%) | 128 (13.1%) | 149 (14.7%) | 259 (11.0%) |
| Skin and soft tissue | 19 (1.2%) | 11 (1.1%) | 24 (2.4%) | 18 (2.0%) | 42 (4.3%) | 29 (2.9%) | 59 (2.5%) |
| Abdominal | 43 (2.8%) | 40 (3.9%) | 199 (19.9%) | 102 (11.2%) | 36 (3.7%) | 49 (4.8%) | 62 (2.6%) |
| Gastrointestinal | 54 (3.5%) | 28 (2.8%) | 48 (4.8%) | 35 (3.8%) | 35 (3.6%) | 27 (2.7%) | 112 (4.8%) |
| Infection following a procedure | 24 (1.5%) | 12 (1.2%) | 172 (17.2%) | 75 (8.2%) | 36 (3.7%) | 32 (3.2%) | 42 (1.8%) |
| Otherd | 243 (15.6%) | 195 (19.3%) | 161 (16.1%) | 182 (19.9%) | 198 (20.2%) | 205 (20.2%) | 620 (26.4%) |
| Organ system with acute dysfunction | | | | | | | |
| Respiratory | 730 (46.8%) | 383 (37.8%) | 253 (25.3%) | 115 (12.6%) | 143 (14.6%) | 187 (18.4%) | 289 (12.3%) |
| Circulatory | 82 (5.3%) | 45 (4.4%) | 121 (12.1%) | 65 (7.1%) | 34 (3.5%) | 55 (5.4%) | 81 (3.5%) |
| Renal | 227 (14.6%) | 119 (11.8%) | 280 (28.0%) | 179 (19.6%) | 151 (15.4%) | 137 (13.5%) | 410 (17.5%) |
| Hepatic | 14 (0.9%) | 12 (1.2%) | 23 (2.3%) | 44 (4.8%) | 12 (1.2%) | 44 (4.3%) | 25 (1.1%) |
| Coagulation | 87 (5.6%) | 87 (8.6%) | 23 (2.3%) | 30 (3.3%) | 56 (5.7%) | 73 (7.2%) | 754 (32.1%) |
| Othere | 72 (4.6%) | 44 (4.3%) | 71 (7.1%) | 33 (3.6%) | 43 (4.4%) | 60 (5.9%) | 51 (2.2%) |
| Explicit sepsis | 501 (32.1%) | 428 (42.3%) | 480 (48.0%) | 569 (62.2%) | 619 (63.1%) | 580 (57.1%) | 1,349 (57.5%) |
| Neutropenia | 303 (19.4%) | 224 (22.1%) | 74 (7.4%) | 104 (11.4%) | 304 (31.0%) | 163 (16.1%) | 854 (36.4%) |
| ICU admittancef | 69 (4.4%) | 34 (3.4%) | 41 (4.1%) | 22 (2.4%) | 22 (2.2%) | 20 (2.0%) | 71 (3.0%) |
NOTE: Data are n (%).
Respiratory cancer includes ICD-10 codes C30–C34 and C38.
Hematologic cancer includes ICD-10 codes C81–C86, C96, C90, and C91–C95.
Data on metastatic hematologic cancer are not presented due to a low number of observations.
Other sites of infection were bone, joint, endocarditis/myocarditis, obstetric, ear, mouth, upper airway, CNS, and unknown.
Other organ systems with acute dysfunction were acidosis, unspecific gangrene, CNS dysfunctions, and systemic inflammatory response syndrome in the absence of septic shock.
Period from May 1, 2014, to December 31, 2021.
Hospital mortality in noncancer patients and those with cancer
Patients with cancer were more likely to die during sepsis hospitalization than noncancer patients. For male patients, the crude mortality was 12.7% for noncancer patients, 16.9% for those with nonmetastatic cancer, and 27.1% for those with metastatic cancer. After adjusting for age group and the number of comorbidities, the hospital mortality was 12.5%, 17.4%, and 28.5%, respectively, with an adjusted RR of 1.39 [95% confidence interval (CI), 1.34–1.44] for nonmetastatic and 2.27 (95% CI, 2.18–2.37) for metastatic cancer (Table 3). Among females, the crude hospital mortality was 12.0% in noncancer patients, 16.2% for patients with nonmetastatic cancer, and 26.2% for patients with metastatic cancer. The adjusted mortality in females was 11.6% for noncancer patients, 19.0% for nonmetastatic, and 32.0% for patients with metastatic cancer, corresponding to an adjusted RR of 1.63 (95% CI, 1.55–1.71) for nonmetastatic and 2.75 (95% CI, 2.62–2.89) for metastatic cancer (Table 3). In the combined analysis of males and females, patients with sepsis who had cancer had higher crude hospital mortality than noncancer patients (20.0% vs. 12.3%). After adjustment for sex, age, and number of comorbidities in a logistic regression model, the absolute difference in hospital mortality increased to 10.3% (Supplementary Table S1).
Table 3.
Absolute risk (ARa) and RRa with 95% CI for hospital mortality in patients with sepsis who had cancer versus no cancer by sex and age group, Norwegian hospitals, 2008 to 2021.
| Males | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| | 18–39 years | 40–49 years | 50–59 years | 60–69 years | 70–79 years | ≥80 years | Total population | |||||||
| n/N | AR (95% CI) | n/N | AR (95% CI) | n/N | AR (95% CI) | n/N | AR (95% CI) | n/N | AR (95% CI) | n/N | AR (95% CI) | n/N | AR (95% CI) | |
| Noncancer | 178/6,340 | 2.80% (2.39–3.20) | 195/5,558 | 3.47% (2.99–3.95) | 659/9,778 | 6.62% (6.13–7.10) | 1,669/18,134 | 9.00% (8.59–9.41) | 3,239/25,643 | 12.46% (12.06–12.86) | 6,444/32,336 | 19.85% (19.42–20.28) | 12,384/97,789 | 12.54% (12.34–12.75) |
| Nonmetastatic | 27/537 | 5.15% (3.26–7.04) | 39/572 | 7.38% (5.15–9.61) | 151/1,557 | 10.57% (8.96–12.18) | 582/3,920 | 15.86% (14.66–17.05) | 926/4,822 | 19.92% (18.76–21.00) | 874/4,002 | 22.20% (20.90–23.50) | 2,599/15,410 | 17.44% (16.83–18.04) |
| Metastatic | 15/169 | 9.13% (4.72–13.53) | 46/232 | 21.74% (16.00–27.48) | 208/777 | 28.85% (25.48–32.22) | 594/2,081 | 31.00% (28.85–33.14) | 657/2,443 | 28.95% (27.03–30.87) | 439/1,541 | 29.74% (27.38–32.09) | 1,959/7,243 | 28.48% (27.39–29.56) |
| | n/N | RR (95% CI) | n/N | RR (95% CI) | n/N | RR (95% CI) | n/N | RR (95% CI) | n/N | RR (95% CI) | n/N | RR (95% CI) | n/N | RR (95% CI) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Noncancer | 178/6,340 | 1.00 | 195/5,558 | 1.00 | 659/9,778 | 1.00 | 1,669/18,134 | 1.00 | 3,239/25,643 | 1.00 | 6,444/32,336 | 1.00 | 12,384/97,789 | 1.00 |
| Nonmetastatic | 27/537 | 1.82 (1.22–2.69) | 39/572 | 2.02 (1.45–2.82) | 151/1,557 | 1.53 (1.30–1.82) | 582/3,920 | 1.73 (1.59–1.89) | 926/4,822 | 1.59 (1.49–1.70) | 874/4,002 | 1.13 (1.06–1.20) | 2,599/15,410 | 1.39 (1.34–1.44) |
| Metastatic | 15/169 | 3.23 (1.95–5.35) | 46/232 | 5.94 (4.43–7.97) | 208/777 | 4.23 (3.69–4.86) | 594/2,081 | 3.38 (3.11–3.67) | 657/2,443 | 2.32 (2.15–2.49) | 439/1,541 | 1.52 (1.40–1.65) | 1,959/7,243 | 2.27 (2.18–2.37) |
| P value | | <0.0001 | | <0.0001 | | <0.0001 | | <0.0001 | | <0.0001 | | <0.0001 | | <0.0001 |
| Females | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| | 18–39 years | 40–49 years | 50–59 years | 60–69 years | 70–79 years | >=80 years | Total population | |||||||
| n/N | AR (95% CI) | n/N | AR (95% CI) | n/N | AR (95% CI) | n/N | AR (95% CI) | n/N | AR (95% CI) | n/N | AR (95% CI) | n/N | AR (95% CI) | |
| Noncancer | 92/6,561 | 1.40% (1.12–1.68) | 124/4,153 | 2.93% (2.42–3.44) | 373/7,002 | 5.15% (4.64–5.66) | 1,010/12,954 | 7.66% (7.20–8.11) | 2,193/19,754 | 10.97% (10.53–11.40) | 6,658/36,927 | 17.99% (17.60–18.38) | 10,450/87,351 | 11.64% (11.44–11.85) |
| Nonmetastatic | 21/438 | 4.95% (2.90–7.00) | 42/558 | 8.29% (5.87–10.71) | 123/1,267 | 10.86% (9.03–12.69) | 339/2,496 | 14.32% (12.89–15.75) | 521/2,733 | 19.99% (18.43–21.54) | 512/2,099 | 25.00% (23.11–26.89) | 1,558/9,591 | 18.96% (18.10–19.81) |
| Metastatic | 18/158 | 12.00% (6.77–17.22) | 72/376 | 20.59% (16.25–24.93) | 201/895 | 25.37% (22.21–28.52) | 429/1,633 | 28.06% (25.70–30.42) | 454/1,609 | 30.31% (27.91–32.72) | 253/777 | 33.62% (30.20–37.04) | 1,427/5.448 | 32.02% (30.57–33.46 |
| | n/N | RR (95% CI) | n/N | RR (95% CI) | n/N | RR (95% CI) | n/N | RR (95% CI) | n/N | RR (95% CI) | n/N | RR (95% CI) | n/N | RR (95% CI) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Noncancer | 92/6,561 | 1 | 124/4,153 | 1 | 373/7,002 | 1 | 1,010/12,954 | 1 | 2,193/19,754 | 1 | 6,658/36,927 | 1 | 10,450/87,351 | 1 |
| Nonmetastatic | 21/438 | 3.45 (2.17–5.49) | 42/558 | 2.65 (1.89–3.72) | 123/1,267 | 1.93 (1.59–2.35) | 339/2,496 | 1.87 (1.66–2.09) | 521/2,733 | 1.81 (1.66–1.97) | 512/2,099 | 1.40 (1.29–1.51) | 1,558/9,591 | 1.63 (1.55–1.71) |
| Metastatic | 18/158 | 8.28 (5.12–13.37) | 72/376 | 6.70 (5.11–8.78) | 201/895 | 4.52 (3.86–5.28) | 429/1,633 | 3.66 (3.31–4.05) | 454/1,609 | 2.73 (2.51–2.99) | 253/777 | 1.89 (1.70–2.09) | 1,427/5,448 | 2.75 (2.62–2.89) |
| P value | | <0.0001 | | <0.0001 | | <0.0001 | | <0.0001 | | <0.0001 | | <0.0001 | | <0.0001 |
Abbreviations: n, number of in-hospital deaths; N, number of patients with sepsis.
Interaction on an additive scale (cancer/no cancer and age), males: Relative excess risk due to interaction (RERI; 95% CI) = 1.37 (0.71–2.03); P < 0.0001.
Interaction on an additive scale (metastatic/nonmetastatic and age), males: RERI (95% CI) = 2.35 (0.83–3.87); P = 0.0025.
Interaction on a multiplicative scale (cancer/no cancer and age), males: RR (95% CI) = 0.76 (0.55–1.06); P = 0.109.
Interaction on a multiplicative scale (metastatic/nonmetastatic and age), males: RR (95% CI) = 0.71 (0.42–1.16); P = 0.168.
Interaction on an additive scale (cancer/no cancer and age), females: RERI (95% CI) = 1.37 (0.71–2.03); P < 0.0001.
Interaction on an additive scale (metastatic/nonmetastatic and age), females: RERI (95% CI) = 2.07 (−1.15–5.29); P = 0.208.
Interaction on a multiplicative scale (cancer/no cancer and age), females: RR (95% CI) = 0.37 (0.26–0.53); P < 0.0001.
Interaction on a multiplicative scale (metastatic/nonmetastatic and age), females: RR (95% CI) = 0.31 (0.19–0.49); P < 0.0001.
In the analysis of interaction on an additive scale, age was dichotomized as <40 years (“young patients with cancer”) and 40+ years.
Risk of hospital mortality adjusted for number of comorbidities in age groups and for the number of comorbidities and age group in total in a multivariable log-binomial regression model.
Hospital mortality across cancer types
When assessing the risk of hospital mortality among subgroups of patients with cancer, the highest absolute hospital mortality in male patients with nonmetastatic cancer was observed in those with mesothelioma (41.7%) and CNS cancer (32.4%; Fig. 1A). For female patients with nonmetastatic cancer, the highest absolute hospital mortality was seen in those with CNS cancer (34.7%) and esophageal/stomach cancer (26.5%; Fig. 1B). In patients with metastatic cancer, males with melanoma of the skin exhibited the highest absolute mortality at 42.2%, followed by respiratory cancer (41.0%) and liver cancer (40.7%). Among females with metastatic cancer, respiratory cancer had the highest mortality at 39.3%, followed by esophageal/stomach cancer (37.7%) and thyreoid cancer (36.7%).
Figure 1.
Forest plots illustrating RRs and 95% CIs for hospital mortality in male (A) and female (B) patients with cancer who had sepsis, Norwegian hospitals, 2008–2021. RRs represent the risk of mortality in each cancer type divided by the risk of the remaining sepsis population of the same sex, adjusted for age group and number of comorbidities in a multivariable log-binomial regression model. An RR of 1 indicates no association between cancer type and mortality. Numbers in brackets represent n, the number of patients who died, and N, the number of patients in each cancer group. *, not applicable. The adjusted absolute risks with 95% CIs are provided in the figure.
The strongest associations with hospital mortality were identified in male patients with metastatic melanoma of the skin [RR = 3.34 (95% CI, 2.83–3.94)] and female patients with metastatic respiratory cancer [RR = 3.30 (95% CI, 3.01–3.62); Fig. 1A and B; Supplementary Tables S2A and S2B]. In the large patient groups of metastatic prostate cancer in males and metastatic breast cancer in females, the RRs were 1.50 (95% CI, 1.37–1.64) and 2.90 (95% CI, 2.59–3.24), respectively.
Hospital mortality across age groups
The absolute risk of death in hospital increased with age for both cancer and noncancer patients with sepsis in both sexes (Table 3). The highest mortality was observed in female patients with metastatic cancer aged ≥80 years, in which 33.6% died (95% CI, 30.20–37.04), followed by male patients with metastatic cancer of the same age group, with 29.7% (95% CI, 27.38–32.09) of patients dying. The lowest absolute risk of death was found in female noncancer patients aged 18 to 39 years, with 1.4% of patients dying. Patients with cancer exhibited a higher risk of dying across all age groups and for both sexes although the strength of the association varied. The highest RRs of hospital mortality were observed among female patients with metastatic cancer aged 18 to 39, with an 8.3-fold increase (95% CI, 5.12–13.37), and among male patients with metastatic cancer aged 40 to 49, with a 5.9-fold increase (95% CI, 4.43–7.97), in a multivariable log-binomial regression model adjusted for comorbidity (Table 3). In a sensitivity analysis excluding those with obstetric infections, the RR of hospital mortality for female patients with cancer aged 18 to 39 decreased from 3.45 (95% CI, 2.17–5.49) to 3.32 (95% CI, 2.08–5.28) in the nonmetastatic group and from 8.28 (95% CI, 5.12–13.37) to 7.93 (95% CI, 4.91–12.83) in the metastatic group. In patients aged 40 to 49 years, the RR decreased from 2.65 (1.89–3.72) to 2.64 (1.88–3.70) in the nonmetastatic group and from 6.70 (5.11–8.78) to 6.66 (5.08–8.72) in the metastatic group.
Hospital mortality across infecting microbes
When assessing the association between cancer and hospital mortality in patients with sepsis by infecting microbe, the strongest association was found in patients with gram-negative sepsis, with an RR of 1.89 (95% CI, 1.64–2.18) in males and 2.08 (95% CI, 1.75–2.47) in females with nonmetastatic cancer and an RR of 3.19 (95% CI, 2.74–3.72) in males and 3.23 (95% CI, 2.71–3.84) in females with metastatic cancer. Among those with Staphylococcus aureus sepsis, there was no significant difference in mortality risk in nonmetastatic patients, whereas in patients with metastatic cancer, the risk was 1.8-fold higher for male and 1.6-fold higher for female patients with cancer compared with noncancer patients. Among male patients with Streptococcus pneumoniae sepsis, metastatic cancer was associated with a 2.8-fold higher mortality, whereas no significant association was observed in male patients with nonmetastatic cancer or in female patients with cancer with this infection (Table 4).
Table 4.
Distribution of noncancer patients and those with cancer with subtypes of explicit sepsis and RRa (95% CI) for hospital mortality by sex and infection microbe, Norwegian hospitals, 2008 to 2021.
| Sepsis due to | Noncancer | Nonmetastatic cancer | Metastatic cancer | |||||
|---|---|---|---|---|---|---|---|---|
| n/N | RR (95% CI) | n/N | RR (95% CI) | P value | n/N | RR (95% CI) | P value | |
| Males | | | | | | | | |
| Streptococci, groups A and B | 83/1,036 | 1 | 21/126 | 1.80 (1.16–2.80) | 0.0085 | 10/36 | 3.04 (1.74–5.33) | <0.0001 |
| S. pneumoniae | 153/1,463 | 1 | 23 (182) | 1.05 (0.69–1.58) | 0.824 | 23/73 | 2.76 (1.91–3.97) | <0.0001 |
| Other Streptococci | 128/1,409 | 1 | 27/229 | 1.32 (0.89–1.96) | 0.162 | 22/62 | 2.85 (1.92–4.24) | <0.0001 |
| S. aureus | 496/3,296 | 1 | 82/516 | 1.05 (0.85–1.31) | 0.612 | 80/287 | 1.81 (1.47–2.22) | <0.0001 |
| Other Staphylococci | 74/738 | 1 | 29/262 | 1.27 (0.85–1.90) | 0.238 | 8/87 | 0.98 (0.49–1.96) | 0.950 |
| Anaerobes | 39/442 | 1 | 16/129 | 1.55 (0.92–2.61) | 0.097 | 13/62 | 2.88 (1.59–5.23) | 0.001 |
| Other gram- negative organismsb | 596/8,223 | 1 | 234/1,814 | 1.89 (1.64–2.18) | <0.0001 | 183/880 | 3.19 (2.74–3.72) | <0.0001 |
| Other specified sepsis | 425/3,413 | 1 | 128/836 | 1.27 (1.06–1.52) | 0.010 | 98/416 | 1.98 (1.63–2.41) | <0.0001 |
| Unspecified sepsis | 2,338/11,513 | 1 | 635/3,527 | 1.05 (0.97–1.13) | 0.224 | 506/1,737 | 1.55 (1.43–1.68) | <0.0001 |
| Candida sepsis | 86/316 | 1 | 33/116 | 0.97 (0.69–1.37) | 0.870 | 126/459 | 1.01 (0.53–1.93) | 0.969 |
| Females | | | | | | | | |
| Streptococci, groups A and B | 63/836 | 1 | 6/71 | 1.06 (0.49–2.29) | 0.891 | 9/49 | 2.22 (1.11–4.44) | 0.023 |
| S. pneumoniae | 167/1,570 | 1 | 10/111 | 0.92 (0.50–1.70) | 0.798 | 5/40 | 1.10 (0.49–2.47) | 0.820 |
| Other streptococci | 108/833 | 1 | 12/163 | 0.69 (0.39–1.12) | 0.205 | 16/89 | 1.67 (1.01–2.74) | 0.042 |
| S. aureus | 434/2,085 | 1 | 39/278 | 0.77 (0.58–1.04) | 0.094 | 50/199 | 1.61 (1.25–2.07) | <0.001 |
| Other staphylococci | 53/452 | 1 | 14/141 | 1.24 (0.69–2.23) | 0.468 | 11/69 | 2.18 (1.23–3.87) | 0.0074 |
| Anaerobes | 42/376 | 1 | 18/87 | 1.75 (1.04–2.93) | 0.034 | 9/55 | 1.33 (0.66–2.70) | 0.429 |
| Other gram-negative organismsb | 573/8,185 | 1 | 150/1,111 | 2.08 (1.75–2.47) | <0.0001 | 143/742 | 3.23 (2.71–3.84) | <0.0001 |
| Other specified sepsis | 79/385 | 1 | 77/565 | 1.15 (0.91–1.44) | 0.241 | 79/385 | 1.85 (1.47–2.33) | <0.0001 |
| Unspecified sepsis | 2,113/9,347 | 1 | 439/2,536 | 1.07 (0.98–1.18) | 0.118 | 421/1,449 | 1.07 (0.98–1.18) | 0.118 |
| Candida sepsis | 60/229 | 1 | 22/74 | 1.19 (0.79–1.80) | 0.399 | 14/32 | 1.61 (1.02–2.55) | 0.041 |
Abbreviations: n, number of deaths; N, number of patients with explicit sepsis subtype.
RR of hospital mortality in noncancer patients vs. those with cancer adjusted for age group and comorbidity in the total sepsis population of the same sex in a multivariable log-binomial regression model.
Gram-negative bacteria excluding anaerobic bacteria and Haemophilus influenzae.
Hospital mortality and neutropenia
Among patients with sepsis who had cancer, neutropenic patients had a lower risk of hospital death than non-neutropenic patients. Among patients with nonmetastatic cancer, neutropenia compared with non-neutropenia was associated with an RR of hospital death of 0.52 (95% CI, 0.46–0.59) in males and 0.64 (95% CI, 0.56–0.73) in females. The corresponding RRs for metastatic patients were 0.77 (95% CI, 0.66–0.91) and 0.77 (95% CI, 0.66–0.90) in males and females, respectively (Supplementary Table S3).
Discussion
To our knowledge, this is the first nationwide cohort study to examine the sex-stratified characteristics and hospital mortality differences between noncancer patients and those with cancer who had sepsis across all hospital wards.
In line with previous studies, we found that hospital mortality was significantly higher among patients with cancer. Specifically, mortality was 39% higher in males and 63% higher in females with nonmetastatic cancer and 127% higher in males and 175% higher in females with metastatic cancer. Notably, the association between cancer and hospital mortality was strongest in patients with sepsis younger than 50 years, and cancer type emerged as a key prognostic factor. Additionally, the prognostic impact of cancer varied across sepsis cases caused by different pathogens, with the strongest association observed in gram-negative sepsis. Interestingly, patients with neutropenic cancer who had sepsis exhibited a lower risk of hospital mortality compared with their non-neutropenic counterparts.
Our findings revealed that female patients with sepsis who had cancer had a higher adjusted absolute mortality and a stronger association between cancer diagnosis and hospital mortality than male patients, regardless of metastatic disease. This observation is somewhat unexpected, as men generally exhibit higher mortality rates from both cancer and sepsis (14, 15, 17). Possible explanations may involve differing distributions of cancer types among males and females with sepsis. Furthermore, differences in pharmacokinetics and body composition between the sexes may result in increased toxicity from cancer treatments in females, potentially placing them at greater risk for a more severe progression of sepsis (28). Moreover, the systemic effects of cancer on immune responses may differ between males and females due to genetic, epigenetic, and hormonal variations. This is a complex interplay that requires further investigation (29, 16). Interestingly, we noted that females were less frequently admitted to ICUs compared with males, albeit the difference was modest. Future studies on sepsis should prioritize sex-specific analyses to optimize treatment strategies for women.
The higher mortality risk among patients with metastatic cancer compared with patients with nonmetastatic cancer who had sepsis aligns with findings from previous studies, likely reflecting a greater tumor burden and poorer overall health status (12, 13). Consistent with previous studies, patients with respiratory cancer exhibited particularly high hospital mortality, which may be influenced by airway obstruction leading to severe pneumonia and respiratory failure (5, 7, 11). Our study extends this finding by confirming that it is consistent across both sexes. Furthermore, we found a strong association between metastatic esophageal and stomach cancers and hospital mortality in both sexes, likely due to the anatomic challenges posed by these tumors and the poor nutritional status they often induce (30). Among patients with sepsis who had metastatic melanoma, we found a particularly strong association with mortality in males, which may reflect a poorer cancer prognosis in male patients compared with females with metastatic melanoma. This sexual dimorphism is believed to be influenced by various factors, including differences in mutational burden and a less robust immune response in males, which are also important to sepsis outcomes (14, 16, 31). Additionally, we found a strong association between CNS cancer and mortality in both sexes. These patients are often immunocompromised due to steroid use and are vulnerable to neurologic deterioration triggered by infection. Sepsis increases the permeability of the blood-brain barrier, which can lead to cerebral edema, particularly dangerous in patients with brain tumors in which intracranial pressure is already elevated. Contrary to earlier studies, our analysis did not find higher mortality in patients with sepsis who had hematologic cancers when compared with those with solid tumors, which may reflect advancements in treating hematologic malignancies over the past decade (5, 7). Interestingly, prostate cancer seemed to be associated with decreased mortality. There are few studies on the relationship between prostate cancer and mortality in patients with sepsis. The apparent protective effect observed in our study may reflect characteristics of both the malignancy and the sepsis episode, as well as factors related to diagnostics and treatment. As shown in our data, patients with prostate cancer often present with urinary tract infections, which have been associated with better outcomes than other infection sites (32). In addition, they are less likely to undergo intensive immunosuppressive treatments, which could contribute to better outcomes compared with patients with other malignancies. Importantly, the varying sepsis mortality among patients with differing cancer types may be influenced by a multitude of factors, including patient characteristics, tumor localization, disease severity, and treatment regimens. Previous research has indicated that chemotherapy correlates with higher sepsis mortality, whereas radiotherapy shows no significant impact (33). Variability may also stem from differences in immune responses across cancer types, recognizing that cancer is a systemic disease that affects the immune system far beyond the tumor microenvironment (3). Interestingly, the upregulation of programmed cell death protein 1 (PD-1) and programmed cell death ligand 1 (PD-L1), along with the subsequent “exhaustion” of T cells, as described in cancer progression, has been associated with high sepsis mortality rates in animal models, and clinical phase I studies examining the use of immune checkpoint blockade in patients with sepsis exist (34, 35). Future studies investigating the benefits of immunomodulating therapies in sepsis should be sex-specific and involve patients with cancer. Identifying biomarkers that can accurately detect and quantify immune suppression will be essential to this approach.
In alignment with previous studies, we observed that younger patients with cancer who had sepsis experienced the greatest relative increase in mortality risk compared with noncancer patients of the same age (5, 7). The worse outcomes in these younger patients may be attributed to the aggressive nature of cancers diagnosed at a young age and the more intensive treatments often administered regardless of cancer stage (27). In contrast, the age-related decline in immune function may mitigate the impact of cancer on sepsis mortality in older adults. Our study adds to the existing literature by highlighting that female patients with cancer who had sepsis are at a greater risk of hospital mortality than their male counterparts across all age groups, compared with their respective noncancer controls, with a notable peak in women aged 18 to 39 years.
Consistent with a previous study by Liu and colleagues (11), our study demonstrated that the association between cancer and mortality was particularly pronounced in cases of sepsis caused by gram-negative organisms. Clinical factors, including varying degrees of immunosuppression and sites of infection between noncancer patients and those with cancer may play a significant role. Moreover, a recent study among 4.6 million hospitalized adult patients in the United States found that patients with cancer had a 48% higher risk of extended-spectrum beta-lactamase–producing Enterobacterales, which has emerged as a significant antimicrobial resistance problem in many countries, including Norway (36). Given that chemotherapy disrupts the gut microbiota and that antimicrobial resistance among gram-negative bacteria is increasing, there are growing concerns about the future implications for patients with cancer (37–39).
Among patients with sepsis who had COVID-19, 4.5% were diagnosed with cancer. A population-based study from Norway found similar results as patients with cancer represented 6.5% of all COVID-19 cases and had fatality rates similar to noncancer patients (40). As few patients with sepsis in our study had COVID-19, we believe its impact on our results is minimal.
In the present study, neutropenia was not associated with increased mortality; on the contrary, it seemed to be a protective factor in both patients with nonmetastatic and metastatic cancer of both sexes. Although several previous studies have not demonstrated any clear correlation, a meta-analysis of more than 1,700 critically ill patients with neutropenic cancer hospitalized between 2005 and 2015 found that neutropenia was independently associated with increased mortality, except in those treated with granulocyte colony-stimulating factor (G-CSF; ref. 41). Possible explanations for our findings include adherence to specific treatment guidelines for neutropenic patients that ensure swift intervention (42). Furthermore, a larger proportion of neutropenic patients may be receiving active cancer therapies, which could confer a more favorable cancer prognosis and fewer treatment limitations compared with patients with end-stage disease. The increased use of G-CSF during the last decades, even in advanced cases, may also play a role. Our findings indicate that factors beyond neutropenia drive sepsis mortality in patients with cancer, suggesting that neutropenia alone should not be a determinant for limiting treatment intensity.
Both male and female patients with nonmetastatic cancer exhibited a nearly 10-fold higher risk of acute coagulation system dysfunction compared with noncancer patients. This association is not unexpected given the thrombogenic nature of both cancer and sepsis and is supported by others (43, 44). However, the strength of this association is of significant clinical relevance, emphasizing the need for heightened awareness, early detection, and preventive anticoagulation strategies in this vulnerable group.
The present study has several strengths, including its large sample size and individual patient data from an unselected cohort covering the entire Norwegian population over a 14-year observation period. Another strength of the study is the availability of data enabling analyses within strata of patients with sepsis by sex, age, and clinically important pathogens, thereby providing insights for improved patient stratification. Reporting to NPR is mandatory, leading to minimal missing data and high data completeness (24). According to the Norwegian Directorate of Health’s medical coding guidelines, for a condition to be included in the registry, it must be documented in the patient’s inpatient record that the condition has been treated, examined, and assessed or has had significance for the further treatment. Furthermore, sepsis coding in Norway has undergone strict quality control to ensure uniformity (45).
However, there are also limitations. The validity of administrative data relies both on the accuracy of coding practices and the methodologies used for ICD code extraction and case definition. Consequently, utilizing ICD-10 codes to identify cancer and sepsis may lead to instances of both over- and underdetection. Two prior studies conducted in Germany and the United States, which compared administrative data with clinical data for identifying sepsis, found that administrative data often underdetects cases of sepsis (46, 47). Furthermore, a recent scoping review assessing the accuracy of ICD-coding methods to estimate sepsis epidemiology analyzed 17 studies comparing administrative data with medical chart reviews and concluded that none of the methods used were optimal for identifying sepsis (48). The code-selection strategies used in these studies were either explicit or implicit methods, and they had a median sensitivity of <75% but a median specificity of >85%. In the present study, we chose to identify sepsis using both explicit and implicit sepsis codes. This strategy was designed to capture Norwegian coding practices for sepsis and potentially compensate for the underdetection of sepsis observed with explicit coding strategies and the possible overdetection of sepsis in implicit sepsis cases in which acute organ dysfunction is unrelated to infection (23, 45). We do believe that the latter applies to both patients with cancer and those with other noncommunicable diseases and thereby may give rise to nondifferential misclassification. Changes in the sepsis definition over the study period could also introduce classification uncertainties although the consistent use of the Sepsis-3 definition likely mitigated this effect in our study. By focusing on first-time sepsis admissions, we aimed to minimize competing risks, but this approach may also introduce selection bias by excluding the frailest patients, potentially underestimating sepsis mortality. Moreover, data on ICU admissions were only available for the latter part of the study period (2014–2021). Given advances in cancer care, it is plausible that earlier in the study period, patients with cancer faced more limitations in treatment options due to advanced disease or perceived poor prognosis. Conversely, the progression of personalized cancer therapies may have contributed to a decreased need for intensive care in later years.
Although we believe the NPR coding guidelines minimize the risk of including patients with nonactive cancers, misclassification of cancer diagnoses or the presence of metastasis cannot be entirely ruled out. Any misclassification of patients in long-term remission as having active cancer or underdetection of cancer cases would lead to an underestimation of the association between cancer and mortality.
NPR does not include the cause of death, and distinguishing between cancer- and sepsis-related mortality is inherently complex. In patients with advanced cancer, symptoms often overlap, and both the disease itself and its treatments may mimic sepsis. Additionally, the discontinuation of cancer treatment and reduced functional status due to sepsis can make it challenging to determine which condition is primarily responsible for the patient's deterioration and death, even for the treating clinicians. Mortality rates among patients with cancer may be underestimated, as a greater proportion may have been discharged to hospice care. Importantly, the study did not assess sepsis mortality following hospital discharge.
Although this study provides valuable insights, future research should address the absence of lifestyle data, socioeconomic status, hospital characteristics, performance status, and treatment details to more fully account for these potential confounding factors. Our findings emphasize the need for targeted strategies to reduce mortality in patients with cancer who had sepsis, including preventive measures, early recognition, tailored treatments, and improved management of complications such as organ dysfunction and gram-negative infections. Incorporating patients with cancer in clinical trials is critical for advancing personalized approaches to sepsis care.
The present study demonstrates that patients with cancer who had sepsis continue to experience high hospital mortality compared with other patients with sepsis. The increased mortality varies with the presence of metastases, cancer type, and infecting microbe and is more pronounced in young adults and females. Our findings underscore the need for further research into strategies to reduce mortality in patients with cancer who had sepsis and may inform future research on targeted sepsis therapies in patient subgroups.
Supplementary Material
Supplementary Methods shows ICD-10 codes used to define sepsis, cancer, and comorbidity.
Supplementary Figure 1 shows flowchart of the study population.
Supplementary Table S1 shows Hospital mortality in cancer patients with sepsis by cancer subtype, Norwegian hospitals 2008 to 2021.
Supplementary Table S2A shows Hospital mortality in male cancer patients with sepsis by cancer subtype and metastatic status, Norwegian hospitals 2008 to 2021.Supplementary Table S2B shows Hospital mortality in female cancer patients with sepsis by cancer subtype and metastatic status, Norwegian hospitals 2008 to 2021.
Supplementary Table S3 shows Distribution of cancer patients with neutropenia and Absolute Risks (AR) (95% CI) and Relative Risks (RR) (95% CI) of hospital mortality by sex and the presence of neutropenia. Norwegian hospitals, 2008 to 2021.
Acknowledgments
M. Husby received a PhD grant from Molde University College and Møre og Romsdal Hospital Trust. L.T. Gustad received funding from Nord-Trøndelag Hospital Trust for data extraction and data storage.
Footnotes
Note: Supplementary data for this article are available at Cancer Epidemiology, Biomarkers & Prevention Online (http://cebp.aacrjournals.org/).
Data Availability
We do not have ethical approval to deposit our datasets in publicly available repositories. The data that support the findings of this study are available from Helsedataservice following necessary ethical approvals or from the project owner upon reasonable request.
Authors’ Disclosures
L.T. Gustad reports grants from Nord-Trøndelag Hospital Trust during the conduct of the study. No disclosures were reported by the other authors.
Authors’ Contributions
M. Husby: Conceptualization, formal analysis, funding acquisition, visualization, methodology, writing–original draft, project administration, writing–review and editing. H. Frydenberg: Supervision, writing–review and editing. T.Å. Myklebust: Formal analysis, validation, visualization, methodology, writing–review and editing. N.V. Skei: Data curation, writing–review and editing. E. Solligård: Supervision, writing–review and editing. I. Thune: Methodology, writing–review and editing. L.T. Gustad: Data curation, funding acquisition, project administration, writing–review and editing. A.-S. Furberg: Conceptualization, supervision, funding acquisition, methodology, writing–review and editing.
References
- 1. Singer M, Deutschman CS, Seymour CW, Shankar-Hari M, Annane D, Bauer M, et al. The third international consensus definitions for sepsis and septic shock (Sepsis-3). JAMA 2016;315:801–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Rudd KE, Johnson SC, Agesa KM, Shackelford KA, Tsoi D, Kievlan DR, et al. Global, regional, and national sepsis incidence and mortality, 1990-2017: analysis for the Global Burden of Disease Study. Lancet 2020;395:200–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Hiam-Galvez KJ, Allen BM, Spitzer MH. Systemic immunity in cancer. Nat Rev Cancer 2021;21:345–59. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Lemiale V, Pons S, Mirouse A, Tudesq J-J, Hourmant Y, Mokart D, et al. Sepsis and septic shock in patients with malignancies: a Groupe de Recherche Respiratoire en Réanimation Onco-Hématologique Study. Crit Care Med 2020;48:822–9. [DOI] [PubMed] [Google Scholar]
- 5. Hensley MK, Donnelly JP, Carlton EF, Prescott HC. Epidemiology and outcomes of cancer-related versus non–cancer-related sepsis hospitalizations. Crit Care Med 2019;47:1310–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Vesteinsdottir E, Sigurdsson MI, Gottfredsson M, Blondal A, Karason S. A nationwide study on characteristics and outcome of cancer patients with sepsis requiring intensive care. Acta Oncol 2022;61:946–54. [DOI] [PubMed] [Google Scholar]
- 7. Sharma A, Nguyen P, Taha M, Soubani AO. Sepsis hospitalizations with versus without cancer: epidemiology, outcomes, and trends in nationwide analysis from 2008 to 2017. Am J Clin Oncol 2021;44:505–11. [DOI] [PubMed] [Google Scholar]
- 8. Evans L, Rhodes A, Alhazzani W, Antonelli M, Coopersmith CM, French C, et al. Surviving sepsis campaign: international guidelines for management of sepsis and septic shock 2021. Crit Care Med 2021;49:e1063–43. [DOI] [PubMed] [Google Scholar]
- 9. Wiersinga WJ, van der Poll T. Immunopathophysiology of human sepsis. EBioMedicine 2022;86:104363. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Stanski NL, Wong HR. Prognostic and predictive enrichment in sepsis. Nat Rev Nephrol 2020;16:20–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Liu MA, Bakow BR, Hsu T-C, Chen JY, Su K-Y, Asiedu E-K, et al. Temporal trends in sepsis incidence and mortality in patients with cancer in the US population. Am J Crit Care 2021;30:e71–9. [DOI] [PubMed] [Google Scholar]
- 12. Van de Louw A, Cohrs A, Leslie D. Incidence of sepsis and associated mortality within the first year after cancer diagnosis in middle aged adults: a US population based study. PLoS One 2020;15:e0243449. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Rosolem MM, Rabello LS, Lisboa T, Caruso P, Costa RT, Leal JV, et al. Critically ill patients with cancer and sepsis: clinical course and prognostic factors. J Crit Care 2012;27:301–7. [DOI] [PubMed] [Google Scholar]
- 14. Haupt S, Caramia F, Klein SL, Rubin JB, Haupt Y. Sex disparities matter in cancer development and therapy. Nat Rev Cancer 2021;21:393–407. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Mohus RM, Gustad LT, Furberg A-S, Moen MK, Liyanarachi KV, Askim Å, et al. Explaining sex differences in risk of bloodstream infections using mediation analysis in the population-based HUNT study in Norway. Sci Rep 2022;12:8436. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Klein SL, Flanagan KL. Sex differences in immune responses. Nat Rev Immunol 2016;16:626–38. [DOI] [PubMed] [Google Scholar]
- 17. Cook MB, McGlynn KA, Devesa SS, Freedman ND, Anderson WF. Sex disparities in cancer mortality and survival. Cancer Epidemiol Biomarkers Prev 2011;20:1629–37. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Murray CJL, Ikuta KS, Sharara F, Swetschinski L, Robles Aguilar G, Gray A, et al. Global burden of bacterial antimicrobial resistance in 2019: a systematic analysis. Lancet 2022;399:629–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Skei NV, Nilsen TIL, Mohus RM, Prescott HC, Lydersen S, Solligård E, et al. Trends in mortality after a sepsis hospitalization: a nationwide prospective registry study from 2008 to 2021. Infection 2023;51:1773–86. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Ohbe H, Sasabuchi Y, Doi K, Matsui H, Yasunaga H. Association between levels of intensive care and in-hospital mortality in patients hospitalized for sepsis stratified by sequential organ failure assessment scores. Crit Care Med 2023;51:1138–47. [DOI] [PubMed] [Google Scholar]
- 21. Haanen J, Obeid M, Spain L, Carbonnel F, Wang Y, Robert C, et al. Management of toxicities from immunotherapy: ESMO Clinical Practice Guideline for diagnosis, treatment and follow-up. Ann Oncol 2022;33:1217–38. [DOI] [PubMed] [Google Scholar]
- 22. Nelmes E, Edwards L, Jhanji S, Antcliffe DB, Tatham KC. Patients with cancer and sepsis trials: an unfair representation? Clin Med 2023;23:635–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Skei NV, Damås JK, Gustad LT. Number of ICD-10 diagnosis fields required to capture sepsis in administrative data and truncation bias: a nationwide prospective registry study. PLoS One 2025;20:e0320054. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Bakken IJ, Ariansen AMS, Knudsen GP, Johansen KI, Vollset SE. The Norwegian Patient Registry and the Norwegian Registry for Primary Health Care: research potential of two nationwide health-care registries. Scand J Public Health 2020;48:49–55. [DOI] [PubMed] [Google Scholar]
- 25. Bakken IJ, Gystad SO, Christensen ØOS, Huse UE, Larønningen S, Nygård J, et al. Comparison of data from the Norwegian Patient Register and the Cancer Registry of Norway. Tidsskr Nor Laegeforen 2012;132:1336–40. [DOI] [PubMed] [Google Scholar]
- 26. Stausberg J, Hagn S. New morbidity and comorbidity scores based on the structure of the ICD-10. PLoS One 2015;10:e0143365. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Tricoli JV, Blair DG, Anders CK, Bleyer WA, Boardman LA, Khan J, et al. Biologic and clinical characteristics of adolescent and young adult cancers: acute lymphoblastic leukemia, colorectal cancer, breast cancer, melanoma, and sarcoma. Cancer 2016;122:1017–28. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Unger JM, Vaidya R, Albain KS, LeBlanc M, Minasian LM, Gotay CC, et al. Sex differences in risk of severe adverse events in patients receiving immunotherapy, targeted therapy, or chemotherapy in cancer clinical trials. J Clin Oncol 2022;40:1474–86. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Hoffmann JP, Liu JA, Seddu K, Klein SL. Sex hormone signaling and regulation of immune function. Immunity 2023;56:2472–91. [DOI] [PubMed] [Google Scholar]
- 30. Bossi P, Delrio P, Mascheroni A, Zanetti M. The spectrum of malnutrition/cachexia/sarcopenia in oncology according to different cancer types and settings: a narrative review. Nutrients 2021;13:1980. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Behbahani S, Maddukuri S, Cadwell JB, Lambert WC, Schwartz RA. Gender differences in cutaneous melanoma: demographics, prognostic factors, and survival outcomes. Dermatol Ther 2020;33:e14131. [DOI] [PubMed] [Google Scholar]
- 32. Prest J, Nguyen T, Rajah T, Prest AB, Sathananthan M, Jeganathan N. Sepsis-related mortality rates and trends based on site of infection. Crit Care Explor 2022;4:e0775. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Shvetsov YB, Ogino MH, Glibetic N, Asato CB, Wilkens LR, Le Marchand L, et al. Association of sepsis mortality with specific cancer sites and treatment type: the multiethnic cohort study. J Pers Med 2021;11:146. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Hotchkiss RS, Colston E, Yende S, Crouser ED, Martin GS, Albertson T, et al. Immune checkpoint inhibition in sepsis: a Phase 1b randomized study to evaluate the safety, tolerability, pharmacokinetics, and pharmacodynamics of nivolumab. Intensive Care Med 2019;45:1360–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Watanabe E, Nishida O, Kakihana Y, Odani M, Okamura T, Harada T, et al. Pharmacokinetics, pharmacodynamics, and safety of nivolumab in patients with sepsis-induced immunosuppression: a multicenter, open-label phase 1/2 study. Shock 2020;53:686–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Gupta V, Satlin MJ, Yu K, Martei YM, Sung L, Westblade LF, et al. Burden of antimicrobial resistance in adult hospitalized patients with cancer: a multicenter analysis. Cancer Med 2024;13:e70495. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. GBD 2021 Antimicrobial Resistance Collaborators . Global burden of bacterial antimicrobial resistance 1990–2021: a systematic analysis with forecasts to 2050. Lancet 2024;404:1199–226. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Nanayakkara AK, Boucher HW, Fowler VG Jr, Jezek A, Outterson K, Greenberg DE. Antibiotic resistance in the patient with cancer: escalating challenges and paths forward. CA Cancer J Clin 2021;71:488–504. [DOI] [PubMed] [Google Scholar]
- 39. Aarnoutse R, Ziemons J, Penders J, Rensen SS, de Vos-Geelen J, Smidt ML. The clinical link between human intestinal microbiota and systemic cancer therapy. Int J Mol Sci 2019;20:4145. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Johannesen TB, Smeland S, Aaserud S, Buanes EA, Skog A, Ursin G, et al. COVID-19 in cancer patients, risk factors for disease and adverse outcome, a population-based study from Norway. Front Oncol 2021;11:652535. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Georges Q, Azoulay E, Mokart D, Soares M, Jeon K, Oeyen S, et al. Influence of neutropenia on mortality of critically ill cancer patients: results of a meta-analysis on individual data. Crit Care 2018;22:326. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Kochanek M, Schalk E, von Bergwelt-Baildon M, Beutel G, Buchheidt D, Hentrich M, et al. Management of sepsis in neutropenic cancer patients: 2018 guidelines from the Infectious Diseases Working Party (AGIHO) and Intensive Care Working Party (iCHOP) of the German Society of Hematology and Medical Oncology (DGHO). Ann Hematol 2019;98:1051–69. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Abdol Razak NB, Jones G, Bhandari M, Berndt MC, Metharom P. Cancer-associated thrombosis: an overview of mechanisms, risk factors, and treatment. Cancers (Basel) 2018;10:380. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. Levi M, Schultz M, van der Poll T. Sepsis and thrombosis. Semin Thromb Hemost 2013;39:559–66. [DOI] [PubMed] [Google Scholar]
- 45. Skei NV, Afset JE, Ehrnström B, Nessan BS, Damås JK, Nilsen TIL, et al. Coding practice for sepsis 2008-21. Tidsskr Nor Laegeforen 2024;144:1–14. [DOI] [PubMed] [Google Scholar]
- 46. Schwarzkopf D, Rose N, Fleischmann-Struzek C, Boden B, Dorow H, Edel A, et al. Understanding the biases to sepsis surveillance and quality assurance caused by inaccurate coding in administrative health data. Infection 2024;52:413–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47. Rhee C, Dantes R, Epstein L, Murphy DJ, Seymour CW, Iwashyna TJ, et al. Incidence and trends of sepsis in US hospitals using clinical vs claims data, 2009–2014. JAMA 2017;318:1241–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48. Kumar A, Hammond N, Grattan S, Finfer S, Delaney A. Accuracy of international classification of disease coding methods to estimate sepsis epidemiology: a scoping review. J Intensive Care Med 2024;39:3–11. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Methods shows ICD-10 codes used to define sepsis, cancer, and comorbidity.
Supplementary Figure 1 shows flowchart of the study population.
Supplementary Table S1 shows Hospital mortality in cancer patients with sepsis by cancer subtype, Norwegian hospitals 2008 to 2021.
Supplementary Table S2A shows Hospital mortality in male cancer patients with sepsis by cancer subtype and metastatic status, Norwegian hospitals 2008 to 2021.Supplementary Table S2B shows Hospital mortality in female cancer patients with sepsis by cancer subtype and metastatic status, Norwegian hospitals 2008 to 2021.
Supplementary Table S3 shows Distribution of cancer patients with neutropenia and Absolute Risks (AR) (95% CI) and Relative Risks (RR) (95% CI) of hospital mortality by sex and the presence of neutropenia. Norwegian hospitals, 2008 to 2021.
Data Availability Statement
We do not have ethical approval to deposit our datasets in publicly available repositories. The data that support the findings of this study are available from Helsedataservice following necessary ethical approvals or from the project owner upon reasonable request.