Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Attributable mortality due to nosocomial sepsis in Brazilian hospitals: a case–control study

Attributable mortality due to nosocomial sepsis in Brazilian hospitals: a case–control study Background Nosocomial sepsis is a major healthcare issue, but there are few data on estimates of its attributable mortality. We aimed to estimate attributable mortality fraction (AF) due to nosocomial sepsis. Methods Matched 1:1 case–control study in 37 hospitals in Brazil. Hospitalized patients in participating hospitals were included. Cases were hospital non‑survivors and controls were hospital survivors, which were matched by admission type and date of discharge. Exposure was defined as occurrence of nosocomial sepsis, defined as antibiotic prescription plus presence of organ dysfunction attributed to sepsis without an alternative reason for organ failure; alternative definitions were explored. Main outcome measurement was nosocomial sepsis‑attributable fractions, esti‑ mated using inversed‑ weight probabilities methods using generalized mixed model considering time‑ dependency of sepsis occurrence. Results 3588 patients from 37 hospitals were included. Mean age was 63 years and 48.8% were female at birth. 470 sepsis episodes occurred in 388 patients (311 in cases and 77 in control group), with pneumonia being the most common source of infection (44.3%). Average AF for sepsis mortality was 0.076 (95% CI 0.068–0.084) for medical admissions; 0.043 (95% CI 0.032–0.055) for elective surgical admissions; and 0.036 (95% CI 0.017–0.055) for emergency surgeries. In a time‑ dependent analysis, AF for sepsis rose linearly for medical admissions, reaching close to 0.12 on day 28; AF plateaued earlier for other admission types (0.04 for elective surgery and 0.07 for urgent surgery). Alterna‑ tive sepsis definitions yield different estimates. *Correspondence: Fernando G. Zampieri fzampieri@hcor.com.br Full list of author information is available at the end of the article © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. Zampieri et al. Annals of Intensive Care (2023) 13:32 Page 2 of 13 Conclusion The impact of nosocomial sepsis on outcome is more pronounced in medical admissions and tends to increase over time. The results, however, are sensitive to sepsis definitions. Keywords Sepsis, Attributable mortality, Epidemiology Introduction Additional file  1: for details). The protocol was approved Sepsis is a major healthcare issue that may account for at the ethics committee of the coordinating site and at more than 11 million yearly deaths worldwide [1]. While all other including sites (and is available with Additional most cases are community-acquired, nosocomial sepsis file  2). Due to the retrospective nature of the analysis, is an important source of burden for healthcare systems. consent was waived. First patient included was admitted Nosocomial sepsis has been repeatedly associated with to the hospital in May 2018 and last patient included was increase in costs, hospital length-of-stay and mortality admitted to the hospital in January 2020. [2, 3], with most reports focusing on the consequences of sepsis in the population of critically ill patients [4–8], Objective including those who were first admitted due to commu - To estimate attributable mortality fraction due to noso- nity-acquired sepsis [3, 6]. Attributable mortality frac- comial sepsis in adult patients. tion is defined as the proportion of deaths that occur related to an exposure, that is, the proportion of deaths Case and control definitions and matching that would not have occurred if the exposure did not take Cases were defined as hospital deaths and controls as place. The attributable mortality fraction of ICU-acquired patients who were discharged alive. After local ethics infections in critically ill patients has been suggested to committee approval, sites were instructed to obtain a be of close to 11% among patients admitted with sepsis list of the most recent 50 adult patients who died dur- and 21% among patients without sepsis [6], with values ing hospitalization. These patients were paired to the as high as 35% being reported for specific infections [5]. closest temporal patient discharged alive who had the There are few reports on occurrence and impact of same admission type (medical, elective surgical or urgent nosocomial infections after hospital admission in a surgical) as the case patient. A margin of 30  days was broader population [8]. In a recent meta-analysis, only allowed for matching. Matching was manually made; eight studies reported hospital wide data on nosocomial for each case, locally trained personal obtained a list for sepsis, but no study reported an estimate of attributable all patients discharge alive at the same day and checked mortality fraction of nosocomial sepsis [8]. Obtaining a for matching, if unsuccessful more medical records for a good estimate of attributable mortality fraction is impor- wider margin were obtained until the closest discharge tant because it is a major driver of public health deci- alive was obtained. For example, a patient who was sions [9]. Considering that overall hospital mortality is admitted for a medical reason and died was paired with low, a case–control design may provide a good estimate the closest medical admission patient discharged alive of attributable mortality fraction and inform healthcare from the same institution. No other matching method policy makers. was performed. Elective surgical admissions were defined We conducted a case–control study aiming to estimate as admissions due to scheduled surgery. Emergency sur- attributable mortality of nosocomial sepsis in Brazilian gical admissions were defined as those whose surgery hospitals. We hypothesized that nosocomial infections was indicated in the first 24 h after admission. would represent an important burden with high attribut- able mortality fraction. Due to lack of consensus on the Sample size calculation operational definitions of nosocomial sepsis, we have also Sample size was calculated using Fleiss, Tytun and stressed data with different definitions and compared our Ury method to estimate a difference in proportions estimates. [10]. Assuming a prevalence of sepsis in 15% in control patients, 1500 patients per group were required to have a 90% power to detect an odds ratio of at least 1.4 for Methods the association between nosocomial sepsis and mortal- Design: observational, case–control study ity. Sample size was increased to 1800 patients per group Setting after the first protocol amendment before data were col - Thirty-eight hospitals in Brazil. Hospitals in Brazil lected due to an increase in fund availability from the that had at least 100 beds and that had a critical care sponsor; this yields a similar power for an odds ratio as unit were eligible for participating in this study (see Zampier i et al. Annals of Intensive Care (2023) 13:32 Page 3 of 13 low as 1.35. We specified that each participating hos - Statistical analysis pital would include 100 patients (50 cases and 50 con- Descriptive statistics were used for univariate analyses. trols); therefore, the increase in sample size was made by The primary objective was to estimate attributable frac - increasing the number of participating hospitals from 30 tion of mortality due to nosocomial sepsis. We assumed to 38, to account for possible drop-offs and incomplete that in-hospital mortality would be a sufficiently rare pairs. event so that odds ratio estimates obtained from a case control design would be reliable estimators of relative risk for calculation of the nosocomial sepsis mortality Data collection attributable fraction [13–15]. Overall hospital data were collected in a structured case The model adjustment was defined a priori, before report form. A local data collector at each site was trained data collection, according to the protocol available in by the sponsor (HCor Research Institute) before data col- Additional file  2; adjustment variables were elected due lection. Patient data included demographic information, to clinical relevant. The following approach was used comorbidities (using Charlson Comorbidity Index and for attributable fraction calculation for each admission Modified Frailty Index) [11, 12], reason for admission and type: the inverted probability weights (IPW) for each daily data (from admission to 28  days) on suspicion of patient were calculated, representing the cumulative infection, antibiotic use, occurrence of organ failure (see risk of the patient acquiring sepsis during hospitaliza- below), and occurrence of other non-infectious clinically tion, under a multivariable logistic regression analysis relevant events. The suspected infection was diagnosed at including baseline age, Charlson Comorbidity Index, local physician’s discretion and collected from the charts. and a time-dependent variable for the occurrence of Clinically relevant events were by the steering commit- clinically relevant events. We estimated the association tee on agreement, and were defined as occurrence of between sepsis occurrence on hospital mortality within stroke, unstable angina or acute myocardial infarction, 28  days through a mixed logistic regression model severe acute hypertension (hypertension that demanded weighted by IPW. The model included the partici - medical intervention), fall, seizure, pulmonary embolism, pant center as a random intercept, and age, Charlson bronchospasm, lower or upper intestinal track bleeding, Comorbidity Index, infection at admission (to account and need for surgical procedure. for community-acquired sepsis at admission), IPW, and the accumulated dependent time variables of sep- Sepsis definition sis occurrence and clinically relevant events with their Sepsis was defined as suspected infection requiring interaction with time (modeled with a third-degree pol- antibiotic use plus at least one organ dysfunction in the ynomial). From the daily odds ratios estimated by the absence of other clinically relevant events occurring in model, the Miettinen formula was used to calculate the the same day. The following criteria were used for diagno - attributable mortality of sepsis [16]: sis acute organ failure: systolic blood pressure < 90 mmHg and/or mean arterial pressure < 65 mmHg and/or drop in OR − 1 P ∗ , systolic blood pressure > 40  mmHg; arterial partial pres- i OR sure to inspired oxygen fraction ratio below 300 or need for supplementary oxygen to maintain peripheral oxygen i = day, saturation above 90%; abnormal mental status; increase P = proportion of patients that had sepsis among non − survivors. in serum creatinine to values of at least 2  mg/dL and/or The attributable fraction assessed this way therefore urinary output < 0.5 mL/kg/h for 2 h; total bilirubin con- represents the proportion of deaths occurring up to centration above 2 mg/dL; platelet count < 100,000 units/ that day given that the patient had sepsis before that mm ; abnormal coagulation in the absence of anticoagu- given day; that is, how the percentage of deaths that lant use (international normalized ratio above 1.5 or acti- would not have occurred if the patient did not have vated partial thromboplastin time above 60 s). If a patient sepsis up to that moment. We also present the mar- developed the before-mentioned criteria up to the sec- ginal effect of having one sepsis episode on the whole ond day of hospital admission, the infection was consid- period, with an estimate of average attributable fraction ered as not nosocomial and was therefore not considered obtained from the model. All analyses were performed for attributable fraction estimates; similarly, patients that using the R 4.1.1 software [17]. IPW was estimated were admitted with community-acquired infection and using the IPW package [18]. The delta method was used that died without having a second event within hospital to calculate the 95% confidence intervals of the odds admission (that is, without having nosocomial sepsis), ratio and attributable mortality. were not considered in nosocomial sepsis-attributable fraction calculation. Zampieri et al. Annals of Intensive Care (2023) 13:32 Page 4 of 13 Sensitivity analysis Results Different definitions of sepsis were explored (Table  1); Hospital and patient features the same analysis was performed using different defi- A total of 1794 pairs of cases and controls from 37 hospi- nitions as sensitivity analyses. An internal consensus tals were included in the analysis (one additional hospi- was created within the steering committee on prob- tal agreed to participate but did not collect data or filled ability degrees of sepsis according to combinations of forms and was excluded from the analysis). Hospital fea- use of antibiotics, organ failure, results of cultures, and tures are shown in Additional file  1: Table S1. Six hospi- occurrence of clinically relevant events (Table 1; Addi- tals did not complete the full sample of 100 patients; in 4 tional file  1: Fig. S1). Two ad hoc sensitivity analyses cases because it was impossible to find pairs in the pre- were made, one estimating the attributable fraction specified timeframe. Baseline patient features are shown considering definitive, very probable, and probable in Table  2. Average population age was 63  years (stand- sepsis, and a second sensitivity analysis considering ard deviation of 18.6), and 48.8% were female at birth. only definitive and very probable sepsis. This approach Median Charlson Comorbidity Index was 2 (interquar- was used to assess whether different definitions would tile range from 0 to 4). One-fifth (22%) of all included result in different estimates of attributable fraction. patients had a recorded recent (in the previous month) One additional post hoc analysis based on increase in hospital admission before the current hospital admis- Sequential Organ Failure Assessment score (SOFA) sion, and 23.7% already needed assistance for daily liv- as per Sepsis 3 also performed [19, 20]; in this analy- ing activities. Median hospital length-of-stay was 8  days sis, patients were considered to have nosocomial sep- (interquartile range 4–14  days). Admission with acute sis if their SOFA score increased at least two points infection occurred in 39.5% of non-survivors and 30.0% over hospital admission SOFA score values and a new of survivors. Patient locale or status over time is shown antibiotic was started, regardless of other informa- in Additional file  1: Fig. S2. Patient features according to tion on clinically relevant events. A sensitivity analy- development or not of nosocomial sepsis are shown in sis using the main definition but excluding patients Additional file 1: Table S2. admitted with infection was also performed. Finally, A total of 388 patients (311 cases and 77 control, com- two additional sensitivity analysis were conducted for prising 10.8% of all included patients) had 470 nosoco- non-surgical patients adding main diagnosis category mial sepsis episodes (387 in cases and 83 in controls). (neurological, cardiovascular, infection, renal, abdomi- Days of antimicrobial therapy were 603 and 437 per nal, or others) as predictors. 1000 patients/day for cases and controls, respectively. Table 1 Sepsis definitions used for analysis Antibiotic Organ failure Cultures Other clinically Sepsis source Part of Part of prescribed relevant event information alternative alternative available definition 1 definition 2 Primary definition Main sepsis definition Yes Yes Positive or negative No Yes – – Alternative definition Definitive Yes Yes Positive No Yes Yes Yes Very probable, either: (1) or Yes Yes Positive Yes Yes Yes Yes (2) or Yes Yes Negative No Yes Yes Yes (3) Yes Yes Negative Yes Yes Yes Yes Probable No Yes NA No No Yes No Possible No Yes NA Yes No No No SOFA increase definition based on Sepsis 3 (post hoc) SOFA definition Yes Yes – Variable No No (“Sepsis 3”) NA data not collected Data on cultures were only collected for patients that received antibiotics. SOFA definition was performed post hoc. Any patients that had an increase in SOFA score of at least 2 points over baseline (hospital admission) SOFA and that either were started antibiotics or received new antibiotics were considered as septic Zampier i et al. Annals of Intensive Care (2023) 13:32 Page 5 of 13 Table 2 Patient features, resource use and occurrence of clinically relevant events in cases and controls Non-survivors Survivors p value (n = 1794) (n = 1794) Age, mean (SD) 68.5 (17.2) 57.5 (18.4) < 0.001 Sex at birth, n (%) 0.095 Female 850 (47.4%) 901 (50.2%) Male 944 (52.6%) 893 (49.8%) Charlson Comorbidity Index, median [IQR] 2 [1–6] 1 [0–3] < 0.001 Modified Frailty Index, median [IQR] 2 [1–3] 1 [0–2] < 0.001 Previous hospitalization (last month), n (%) 471 (26.3%) 319 (17.8%) < 0.001 Pneumonia on past year, n (%) 175 (9.8%) 79 (4.4%) < 0.001 Episode of mental confusion on past year, n (%) 238 (13.3%) 116 (6.5%) < 0.001 Previously on hospice/long‑term facility/homecare, n (%) 93 (5.2%) 39 (2.2%) < 0.001 Dependency for daily living activities, n (%) 620 (34.6%) 232 (12.9%) < 0.001 Known comorbidities at admission, n (%) Dementia 196 (10.9%) 58 (3.2%) < 0.001 Transitory Ischemic Attack 18 (1.0%) 18 (1.0%) 1.00 Stroke 174 (9.7%) 83 (4.6%) < 0.001 Previous myocardial infarction 137 (7.6%) 110 (6.1%) 0.086 Angina/coronary stent 137 (7.6%) 129 (7.2%) 0.656 Heart failure 221 (12.3%) 144 (8.0%) < 0.001 Hypertension 912 (50.8%) 767 (42.8%) < 0.001 Diabetes, uncomplicated 473 (26.4%) 403 (22.5%) 0.007 Diabetes, complicated 136 (7.6%) 119 (6.6%) 0.299 Rheumatologic disease 66 (3.7%) 78 (4.3%) 0.349 Acquired immunodeficiency syndrome 50 (2.8%) 43 (2.4%) 0.529 Cirrhosis 65 (3.6%) 37 (2.1%) 0.006 Cancer 726 (40.5%) 453 (25.3%) < 0.001 Hospital admission Admission type Medical 1524 (84.9%) 1524 (84.9%) – Elective surgery 149 (8.3%) 149 (8.3%) – Urgent surgery/trauma 121 (6.7%) 121 (6.7%) – Relevant diagnosis at admission Infection 709 (39.5%) 538 (30.0%) < 0.001 Respiratory diagnosis Asthma 11 (0.6%) 19 (1.1%) 0.199 Chronic pulmonary obstructive disease 87 (4.8%) 53 (3.0%) 0.004 Other chronic lung disease 19 (1.1%) 16 (0.9%) 0.735 Cardiac diseases ST‑ elevation myocardial infarction 31 (1.7%) 38 (2.1%) 0.466 Non‑ST ‑ elevation myocardial infarction 33 (1.8%) 28 (1.6%) 0.606 Unstable angina 18 (1.0%) 31 (1.7%) 0.083 Angina, unspecified 12 (0.7%) 9 (0.5%) 0.663 Uncompensated heart failure 137 (7.6%) 87 (4.8%) 0.001 Deep vein thrombosis 48 (2.7%) 32 (1.8%) 0.089 Pulmonary thromboembolism 27 (1.5%) 27 (1.5%) 1.00 Neurological diseases Ischemic stroke 83 (4.6%) 58 (3.2%) 0.039 Hemorrhagic stroke 18 (1.0%) 10 (0.6%) 0.183 Transient ischemic attack 3 (0.2%) 12 (0.7%) 0.035 Subarachnoid hemorrhage 17 (0.9%) 8 (0.4%) 0.107 Zampieri et al. Annals of Intensive Care (2023) 13:32 Page 6 of 13 Table 2 (continued) Non-survivors Survivors p value (n = 1794) (n = 1794) Polyradiculopathy/myasthenia 3 (0.2%) 3 (0.2%) 1.00 Renal diseases Acute, non‑related to cirrhosis 91 (5.1%) 39 (2.2%) < 0.001 Chronic, not on dialysis 75 (4.2%) 43 (2.4%) 0.004 Chronic, needing dialysis 31 (1.7%) 40 (2.2%) 0.338 Abdominal diseases Uncompensated cirrhosis 36 (2.0%) 20 (1.1%) 0.042 Digestive bleeding 12 (0.7%) 7 (0.4%) 0.358 Spontaneous bacterial peritonitis 2 (0.1%) 0 (0.0%) 0.500 Hepatorenal syndrome 3 (0.2%) 0 (0.0%) 0.250 Acute pancreatitis 24 (1.3%) 27 (1.5%) 0.778 Uncompensated diabetes 70 (3.9%) 61 (3.4%) 0.477 Admission for diagnostic procedures 535 (29.8%) 515 (28.7%) 0.463 Other 648 (36.1%) 640 (35.7%) 0.808 Resource use Intensive care unit admission, n (%) 853 (47.5%) 340 (19%) < 0.001 Hospital length‑ of‑stay, median [IQR] 9 [5–19] 6 [4–11] < 0.001 Antibiotic days of therapy, days per 1000 patients/day 603 437 < 0.001 Days using antibiotics, median [IQR] 5.5 [1.3–11] 1 [0–6] < 0.001 Events during hospitalization Sepsis (main definition) 525 (29.3%) 232 (12.9%) Up to 2 days from admission 383 (21.3%) 198 (11%) < 0.001 After 2 days from admission 311 (17.3%) 77 (4.3%) < 0.001 Stroke 65 (3.6%) 28 (1.6%) < 0.001 b,c Coronary syndr ome 80 (4.5%) 42 (2.3%) 0.001 Acute severe hypertensive episode 228 (12.7%) 133 (7.4%) < 0.001 Fall 63 (3.5%) 20 (1.1%) < 0.001 Seizure 93 (5.2%) 34 (1.9%) < 0.001 Pulmonary thromboembolism 40 (2.2%) 11 (0.6%) < 0.001 Bronchospasm 118 (6.6%) 27 (1.5%) < 0.001 Digestive bleeding 122 (6.8%) 27 (1.5%) < 0.001 Severe pain 138 (7.7%) 76 (4.2%) < 0.001 Matching variable ST-elevation myocardial infarction Non-ST-elevation myocardial infarction Pain episode that required more than 2 rescues or new diagnostic procedure Pneumonia was the most common source (Additional Additional file  1: Table  S5 (74 patients with any positive file  1: Table  S3, Fig. S3). The most common organ dys - culture). function at sepsis diagnosis was abnormal mental status (Fig.  1). More than half of the patients had more than Attributable mortality fraction one of the specified organ dysfunctions during their Daily odds ratio for mortality obtained from the model first sepsis episode (Additional file  1: Fig. S4). Sepsis was and its respective attributable fraction are shown in more frequently diagnosed in ICU, followed by ward Fig.  2. The reported odds ratio is the effect size of dying (Additional file  1: Fig. S5). There were 573 positive cul - up to a specific day (x-axis) given the patient had infec - tures in the population (Additional file  1: Table S4, strati- tion in the preceding days (up to hospital admission). fied for suspected source); a list of pathogens in patients Attributable fraction is interpreted as percentage of with sepsis according to the main definition is shown in deaths occurring up to that day given that the patient Zampier i et al. Annals of Intensive Care (2023) 13:32 Page 7 of 13 A − All episodes of sepsis 27 27 27 14 14 12 12 11 11 6 66 6 Abnormal Mental Status 175 Hypotension Coagulopathy 120 Hypoxemia 119 High Lactate 114 Kidney Injury 41 Hepatic 300 200 100 0 Set size B − Pneumonia 88 8 8 3 33 33 3 3 3 128 Abnormal Mental Status 72 Hypotension 61 Hypoxemia Coagulopathy 51 High Lactate Kidney Injury 10 Hepatic 150 100 50 0 Set size C − Not pneumonia 8 8 66 6 5 5 4 4 4 3 3 135 Abnormal Mental Status 103 Hypotension 92 Coagulopathy 68 High Lactate 65 Kidney Injury 59 Hypoxemia Hepatic 200 150 100 50 0 Set size Fig. 1 Organ failure at diagnosis in A all septic patients, B pneumonia, C not pneumonia. Each dot represents intersections, with the number of cases shown in bars previously had sepsis. For medical admissions, attribut- urgent surgery patients, with peak values around 0.04– able fraction rose after the sixth day of admission, and 0.07, and a less linear increase over time. reached close to 0.12 at 28  days. The effect of nosoco - Overall marginal odds ratio for mortality and attrib- mial sepsis was less pronounced on elective surgery and utable fraction for patients with any sepsis episode were Intersection size Intersection size Intersection size Zampieri et al. Annals of Intensive Care (2023) 13:32 Page 8 of 13 Medical (n = 3048) Elective surgery (n = 298) Emergency Surgery / Trauma (n = 242) 8 8 8 7 7 7 6 6 6 5 5 5 4 4 4 3 3 3 2 2 2 1 1 1 0 0 0 3 45678 910111213141516171819202122232425262728 3 45678 910111213141516171819202122232425262728 3456789 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 Day Day Day 0.14 0.14 0.14 0.12 0.12 0.12 0.10 0.10 0.10 0.08 0.08 0.08 0.06 0.06 0.06 0.04 0.04 0.04 0.02 0.02 0.02 0.00 0.00 0.00 −0.02 −0.02 −0.02 −0.04 −0.04 −0.04 3 45678 910111213141516171819202122232425262728 3 45678 910111213141516171819202122232425262728 3456789 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 Day Day Day Sepsis 75 144 192 231 260 277 296 304 311 320 326 331 335 340 345 348 352 353 355 358 359 361 361 361 362 362 Sepsis 6 7 9 10 10 10 10 10 11 11 11 11 11 12 12 12 12 12 13 13 13 13 Sepsis 5 6 8 8 8 10 10 12 12 12 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 Death 231 356 450 533 617 713 789 861 909 955 994 1027 1052 1089 1112 1142 1162 1190 1216 1235 1258 1278 1293 1305 1315 1339 Death 29 33 43 50 59 62 68 72 74 81 83 86 91 95 102 105 108 111 114 115 116 118 Death 28 44 53 57 62 65 70 74 75 80 83 85 89 90 92 96 97 99 100 101 101 104 105 106 107 107 3 45678 910111213141516171819202122232425262728 3 45678 910111213141516171819202122232425262728 3456789 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 Day Day Day Fig. 2 Distribution of odds ratio (upper row) and AF (lower row) according to admission type (columns). The odds ratio should be interpreted as the odds ratio of dying up to a specific day (x‑axis) given that the patient has acquired sepsis totalizing 1121 septic episodes (945 on non-survivors 1.73 (95% CI 1.60; 1.87), attributable fraction = 0.076 and 176 in survivors (Additional file  1: Fig. S8 shows (95% CI 0.068; 0.084); 2.75 (95% CI 1.47; 4.03), attrib- infection sources and organ dysfunctions). Marginal utable fraction = 0.043 (95% CI 0.032; 0.055); and 1.75 odds ratio for mortality and attributable fraction under (95% CI 1.06; 2.43), attributable fraction = 0.036 (95% this definition were 1.60 (95% CI 1.51; 1.70), attributable CI 0.017; 0.055]), for medical, elective surgery and fraction = 0.141 (95% CI 0.128; 0.155); 3.61 (95% CI 2.90; emergency surgery patients, respectively. 4.32, attributable fraction = 0.392 (95% CI 0.363; 0.422); and 2.57 (95% CI 2.04; 3.10), attributable fraction = 0.251 Alternative analysis 1: definitive, very probable (95% CI 0.218; 0.284), for medical, elective surgical and and probable sepsis emergency surgery groups. Results for odds ratio and Using this broader definition, a total of 1129 patients had attributable fraction over time are shown in Additional nosocomial sepsis during hospital stay. A total of 1387 file 1: Fig. S9. septic episodes were recorded (1058 on non-survivors and 329 in survivors). 74 patients had definitive sepsis, Post hoc analysis: SOFA definition (based on Sepsis 3) 850 very probable sepsis and 205 probable sepsis (see 729 patients had nosocomial sepsis as by an increase in Additional file  1: Fig. S6 for details on infection source of SOFA score of at least two points over baseline (869 sep- organ dysfunction). The overall marginal odds ratio for tic episodes: 782 on non-survivors and 87 in survivors). mortality and attributable fraction for patients with one Marginal odds ratio for mortality and attributable frac- episode of definite, very probable, or probable sepsis were tion under this definition were 2.25 (95% CI 2.10; 2.39), 1.32 (95% CI 1.25; 1.39), attributable fraction = 0.101 attributable fraction = 0.176 (95% CI 0.167; 0.185); 3.52 (95% CI 0.083; 0.118); 2.85 (95% CI 2.32; 3.39), attrib- (95% CI 2.82; 4.22, attributable fraction = 0.358 (95% CI utable fraction = 0.374 (95% CI 0.337; 0.412); and 2.21 0.330; 0.386); and 2.93 (95% CI 2.30; 3.56), attributable (95% CI 1.77; 2.64), attributable fraction = 0.230 (95% CI fraction = 0.246 (95% CI 0.219; 0.274), for medical, elec- 0.192; 0.268) for medical, elective surgery and emergency tive surgical and emergency surgery groups. Results for surgery groups. Results over time are shown Additional odds ratio and attributable fraction over time are shown file  1: Fig. S7; this definition resulted in an attributable in Additional file 1: Fig. S10. fraction of up to 0.25 for medical admissions, and higher values for elective and urgent surgery than the main defi - Post hoc analysis: exclusion of patients admitted nition used (peaking over 0.45 at 28 days for elective sur- with infection under the main definition gery), with a more linear ascend over time. A total of 3007 patients were considered in this analy- sis—1411 non-survivors and 1596 survivors—with a Alternative analysis 2: definitive and very probable sepsis total of 177 and 37 episodes of sepsis, respectively. Odds A total of 924 patients had nosocomial sepsis as defined ratio for mortality and attributable fraction were 2.21 using the definitive and very probable sepsis criteria, Attributable fraction Odds Ratio Attributable fraction Odds Ratio Attributable fraction Odds Ratio Zampier i et al. Annals of Intensive Care (2023) 13:32 Page 9 of 13 Under this definition, attributable fraction was higher for (95% CI 1.95; 2.47), attributable fraction = 0.055 (95% medical admissions than for elective surgical and emer- CI 0.05; 0.061); 2.84 (95% CI 1.35; 4.31, attributable frac- gency surgical patients. Attributable fraction over time tion = 0.034 (95% CI 0.024; 0.043); and 1.71 (95% CI 0.94; linearly increased up to 28 days for medical patients, but 2.48), attributable fraction = 0.035 (95% CI 0.013; 0.057), not for elective and emergency surgery admissions where for medical, elective surgical and emergency surgery a plateau was observed. Pneumonia was the most com- groups (Additional file 1: Fig. S11). mon infection source, and abnormal mental status was the commonest organ dysfunction observed. Antibiotic Post hoc analysis: additional adjustment for main reason use was high in both groups. for admission using main definition and the SOFA Different sepsis definitions will inevitably be associ - definition for non-surgical patients ated with different prevalence and effect sizes for mor - Adding further adjustment according to main admission tality, with consequential direct impact on attributable category for non-surgical patients analyses yield an aver- mortality fraction. We also explored additional defini - age odds ratio for mortality of 1.64 (95% CI 1.51; 1.76) tions of sepsis according to key features including posi- with an average attributable fraction of 0.070 (95% CI tive cultures, occurrence of clinically relevant events, of 0.061; 0.0782); time-dependent effects are shown in and antibiotic use. The fact that clinically relevant events, Additional file  1: Fig. S12. The only admission type that obtained directly from chart review, was considered was consistently associated with an attributable fraction before attributing organ failure to infection may enhance above 0.10 was admission due to infection (Additional the capability of measuring the effect associated directly file  1: Fig. S13), reinforcing the importance of adjustment with infection. Indeed, organ failure is associated with for infection in the primary analysis. For the increase a myriad of clinical conditions [21]; this creates a situa- SOFA definition (Sepsis 3), adding admission category tion where antibiotics are prescribed due to new organ resulted in average odds ratio was 2.39 (95% CI 2.24; failure even if coexisting events that may be responsible 2.54) and average attributable fraction was 0.184 (95% CI for organ failure occur simultaneously. Our main defini - 0.176; 0.192). tion was stricter than other sepsis definitions by limiting sepsis diagnosis to the absence of coexisting events that Comparison of the definitions could cause organ failure at the same day [20]. The two Results for the comparison of the odds ratio and attribut- alternative definitions were more comprehensive, as seen able fraction for medical admission are shown in Fig.  3. by the higher number of cases reported: the first being Odds ratio for mortality increased over time for all defi - broader than the second. The first alternative attributed nitions; however, due to differences in prevalence of new organ dysfunctions in the absence of clinically rel- events, the attributable fraction was lower for the main evant events as septic events even in the absence of anti- definition when compared to both alternative definitions. biotic use (“any new organ dysfunction in a hospitalized Comparison for the definitions for elective surgery and patient is sepsis until proven otherwise”), while the sec- emergency surgery patients is shown in Additional file  1: ond did not include such patients. The resulting attrib - Fig. S14 and S15; the main definition provided the lower utable fractions reflected the changes in both effect size AF for surgical patients. The post hoc analysis based (odds ratio) and prevalence; despite being associated with on SOFA score increase as per Sepsis 3 suggestion pro- higher odds ratios for mortality, the main definition had vided resulted in estimates for odds ratio and attributable the lowest attributable fraction. Differences in attribut - fraction that largely followed the results of the second able fraction among definitions were specially pressing alternative definition. Exclusion of patients with known in surgical patients. A definition based on increase in sepsis at admission reduced the attributable fraction of SOFA provided results similar to the second alternative nosocomial sepsis due to decrease in number of events definition. When baseline sepsis patients were excluded, (Additional file  1: Fig. S11). Adding further adjustments the attributable fraction was markedly reduced (which for admission type also did not change estimates for non- is expected since infection is an important risk factor surgical patients (Additional file 1: Table S6). for a secondary insult), but the odds ratio for mortality remained high. Discussion These findings have several important implications. In this case–control study including 1794 pairs of patients First, even under the strictest definition sepsis-attribut - from 37 Brazilian hospitals, we found that nosocomial able fraction was still very important, reaching around sepsis, defined by an acute nosocomial infection with 0.12 for prolonged medical admissions; the longer the organ failure in the absence or other clinically relevant medical patient remained in the hospital, the highest the events, was an important contributor to hospital mor- odds ratio for mortality associated with a septic episode. tality with a significant attributable mortality fraction. Zampieri et al. Annals of Intensive Care (2023) 13:32 Page 10 of 13 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 345678 910111213141516171819202122232425262728 Day Main Definition Alternative 1 Alternative 2 0.3 0.2 0.1 0.0 −0.1 345678 910111213141516171819202122232425262728 Day 145 270 363 444 498 547 583 603 627 651 670 687 699 709 716 726 733 739 745 750 753 759 760 761 763 767 Alternative 2 191 344 462 553 626 692 741 764 797 830 851 868 881 892 897 908 913 918 926 931 935 941 942 943 944 948 Alternative 1 75 144 192 231 260 277 296 304 311 320 326 331 335 340 345 348 352 353 355 358 359 361 361 361 362 362 Main Definition 231 356 450 533 617 713 789 861 909 955 994 1027 1052 1089 1112 1142 1162 1190 1216 1235 1258 1278 1293 1305 1315 1339 Death 345678 910111213141516171819202122232425262728 Day Fig. 3 Comparison of the three definitions for medical admissions. Upper panel: the daily odds ratio for mortality in medical patients according to the definition used. Bottom panel: the respective AF for each definition. Note that a higher OR does not equal higher AF due to changes in prevalence This value is aligned with other estimated of attribut - (emergency surgical patients) to 0.076 (medical patients) able fraction in more severe patients and may represent a is sufficiently small so that an intervention may reduce reasonable starting point for quality improvement initia- nosocomial infections without easily noticeable effects in tives [5]. These results, under the main definition, should mortality, unless sample size is very large, under this defi - probably be seen as a minimum value for sepsis-attrib- nition. Since infection burden is not exclusively related to utable fraction in hospitalized patients. Albeit impor- mortality, involving costs, long-term outcomes, quality tant, an average attributable fraction ranging from 0.036 of life, among others, understanding sepsis-attributable Attributable fraction Odds Ratio Zampier i et al. Annals of Intensive Care (2023) 13:32 Page 11 of 13 fraction may avoid over-simplistic conclusions, such as also limited by the assumption that odds ratio and rela- interpreting that an intervention that reduces infection tive risk will be similar when outcomes are infrequent. occurrence but not mortality prevented only non-fatal This is, in fact, the de facto approach made by several infections or even that some infections may not be asso- statistical packages that estimate the attributable frac- ciated with higher mortality at all. tion [24]. Our model adjusted for time-dependency of The variability observed by tweaking sepsis definitions covariates but did not consider further daily informa- should also be seen an alert that it is hard to isolate the tion besides diagnosis of infection and clinically relevant impact of a single event within the intricate path of a hos- events. The main adjustment model was defined a priori, pital stay, specifically for surgical patients where different according to a stablished protocol. All variable selection broader definitions provided results strikingly different approaches may be subject to criticism, but we refrained from medical patients. It is conceivable that “true” sep- from using variables that were associated with either sep- sis-attributable fraction may be somewhere in between sis or mortality as suggested [25]. Adding admission cat- the main definition and the second alternative defini - egories to the main analysis did not change the estimated tion. Far from suggesting that nosocomial sepsis is not of AF significantly. Other approaches could be employed an issue, our results highlight that even when considering to estimate attributable fraction [26]. The list of clinically known factors for poor hospital outcome such as age and relevant events is not exhaustive and was defined by the comorbidities, sepsis could be directly responsible from steering committee during protocol discussions before something between 7.6% and 14.1% of all hospital death data were collected but is somewhat arbitrary. Finally, for medically admitted patients. For surgical patients, our results reflect the Brazilian panorama; it is uncertain the margins are even wider depending on the definition whether they are transposable to other settings. used, peaking over 40%. Attributable fraction is a rela- tive measurement, and not a direct estimate of burden. Conclusion Low-middle income countries are suggested to be more Nosocomial sepsis is an important contributor to hos- affected by nosocomial sepsis, which may result in a pital mortality. The impact of nosocomial sepsis on out - higher numeric burden of deaths in this population. come is more pronounced in medical admissions and We hope that this manuscript fosters the discussion on tends to increase over time. Different sepsis definitions whether sepsis would benefit from a more nuanced diag - led to important changes on attributable fraction. nosis approach where probability categories are used to tailor diagnosis and treatment (as is the case of aspergil- Supplementary Information losis, where possible/probable categories have been in The online version contains supplementary material available at https:// doi. use but may also be applicable to other medical condi-org/ 10. 1186/ s13613‑ 023‑ 01123‑y. tions) [22, 23]. Despite over 30 years of controversy, all Additional file 1: Table S1. Hospital Features. Table S2. Comparison sepsis definitions are binary, that is, they do not consider between patients that develop and did not develop nosocomial sepsis the uncertainty that permeates clinical decision-making, according to the main definition. Table S3. Sepsis episodes. Figure S1. focusing more on severity of illness than in the very prob- Venn diagram for the sepsis definitions used in the manuscript. Figure S2. Daily patient location/status over time for controls (top) and cases ability that the findings are due to active infection. (bottom). Figure S3. Number of nosocomial sepsis episodes according to Our work has several limitations. As with any case infection source. Figure S4. Number of organ dysfunctions at presenta‑ control trial, selecting appropriate controls is challeng- tion for (A) first episode of sepsis, and (B) all sepsis episodes. Figure S5. Patient locale at sepsis diagnosis. Top: Stratified according to outcome; ing. We used the closest temporal admission discharged Bottom: All patients. Table S4. Pathogens isolated from cultures according alive with the same admission type as this seemed a good to suspected site; more than one pathogen was possible for each patient. compromise between feasibility and adequacy. As can CRBI: Catheter related bloodstream infections. Note that pathogens could be isolated in any culture collected from the patient within be seen in Table  1, the number of possible clinical con- the 48h timeframe. The local diagnosis was considered as reference; ditions is very large, and if matching criteria were too therefore the isolated pathogen could not be considered the culprit for strict, we would have ended with issues in obtaining the infection. Table S5. Positive cultures for patients that had one septic episode according to main definition. CRBI: Catheter related bloodstream proper controls; we tried to overcome this by adjusting infections. Note that pathogens could be isolated in any culture col‑ for several relevant confounders, including age, comor- lected from the patient within the 48h timeframe. Same as in Table S3, bidities, etc. Most hospitals in this study still used non- the final diagnosis was made by the site. Figure S6. (A) Infection source considering first alternative definition. The number of patients that did not electronic (paper) healthcare records, thereby making receive antibiotic reflects patients that developed new organ failure in the triaging of possible controls challenging. Restricting our absence of any other clinically relevant event and were considered as pos‑ study to only hospitals with electronic healthcare records sibly septic under this definition. (B) Organ dysfunction at presentation for the alternative definition 1 analysis. Figure S7. Distribution of odds ratio would induce another source of bias, since these hospi- (upper row) and AF (lower row) according to admission type (columns) tals would inherently have more resources. We estimated for the first alternative definition considering the effects of definitive, very attributable fraction from a case–control study, which is probably, and probable sepsis. Figure S8. Sepsis sources (A) and organ Zampieri et al. Annals of Intensive Care (2023) 13:32 Page 12 of 13 Consent for publication dysfunction at presentation for the second alternative definition analysis. Not applicable. Figure S9. Distribution of odds ratio (upper row) and PAF (lower row) according to admission type (columns) for the second alternative defini‑ Competing interests tion considering the effects of definitive and very probably sepsis. Figure The authors declare no relevant competing interest. S10. Distribution of odds ratio (upper row) and PAF (lower row) according to admission type (columns) for the post hoc definition based on SOFA Author details score. Figure S11. Distribution of odds ratio (upper row) and PAF (lower 1 HCor Research Institute, Rua Desembargador Eliseu Guilherme, 200, 8th Floor, row) according to admission type (columns) after excluding patients with 2 São Paulo, Brazil. Department of Critical Care Medicine, Faculty of Medi‑ infection at baseline. Figure S12. Comparison of the three definitions for cine and Dentistry, University of Alberta, 2‑124E Clinical Sciences Building, elective surgery admissions. Upper panel: The daily odds ratio for mortality 3 8440‑112 St NW, Edmonton, AB T6G2B7, Canada. Intensive Care Unit, Emer‑ in elective surgery patients according to the definition used. Bottom gency Medicine Discipline, Hospital das Clínicas da Faculdade de Medicina panel: The respective AF for each definition. Note that a higher OR does 4 da Universidade de São Paulo, São Paulo, Brazil. Intensive Care Unit, Hospital not equal higher AF since prevalence of events also changes. Figure S13. 5 Sírio‑Libanês, São Paulo, SP, Brazil. Unidade de Terapia Intensiva, Santa Casa de Comparison of the three definitions for emergency surgery admissions. 6 Misericórdia de Porto Alegre, Porto Alegre, RS, Brazil. Department of Critical Upper panel: The daily odds ratio for mortality in medical patients accord‑ 7 Care Medicine, Hospital Israelita Albert Einstein, Sao Paulo, Brazil. Austral‑ ing to the definition used. Bottom panel: The respective AF for each defini‑ ian and New Zealand Intensive Care Research Centre (ANZIC‑RC), School tion. Note that a higher OR does not equal higher AF since prevalence of of Public Health and Preventive Medicine, Monash University, Melbourne, events also changes. 8 Australia. Department of Intensive Care, Austin Hospital, Melbourne, Aus‑ tralia. Intensive Care Unit, AC Camargo Cancer Center, São Paulo, SP, Brazil. Additional file 2: IMPACTO ‑MAPA study. 10 11 Hospital Maternidade São Vicente de Paulo, Barbalha, CE, Brazil. Hospital Federal dos Servidores do Estado, Rio de Janeiro, RJ, Brazil. Hospital da Luz, Acknowledgements São Paulo, SP, Brazil. BP‑A Beneficência Portuguesa de São Paulo, Sao Paulo, 14 15 The authors would like to thank all local data collectors for their work. SP, Brazil. Hospital Paulistano, São Paulo, SP, Brazil. Instituto de Cardiologia Collaborators: Barbara Macedo, Fabio S Coutinho (Hospital São Paulo – do Distrito Federal, Brasilia, DF, Brazil. Hospital Nereu Ramos, Florianópolis, 17 18 UNIFESP); Jussara A Arraes (Hospital Maternidade São Vicente de Paulo); SC, Brazil. Hospital Baía Sul, Florianópolis, SC, Brazil. Hospital das Clínicas Viviane S N Xavier (Hospital Federal dos Servidores do Estado); Eliana V N da Faculdade de Medicina de Ribeirão Preto, Ribeirão Preto, SP, Brazil. Santa Martins (Hospital da Luz); Juliana Chaves Coelho (BP ‑ A Beneficência Portu‑ Casa de Misericórdia Belo Horizonte, Belo Horizonte, MG, Brazil. Hospital guesa de São Paulo); Silvana S Santos (AC Camargo Cancer Center); Andreia São José, Criciúma, SC, Brazil. Hospital Dona Helena, Joinville, SC, Brazil. 22 23 Pardini (Hospital Israelita Albert Einstein); Cassio Luis Zandonai (Hospital Nereu Hospital de Amor‑Fundação PIO XII, Barretos, SP, Brazil. Centro Hospitalar Ramos); Julia B de Carvalho (Hospital das Clínicas da Faculdade de Medicina Unimed, Joinville, SC, Brazil. Hospital Maternidade São José, Colatina, ES, 25 26 de Ribeirão Preto); Isabela O B Louredo (Santa Casa de Misericórdia Belo Brazil. Hospital Unimed Vitória, Vitória, ES, Brazil. Hospital Distrital Evandro Horizonte); Renata C Gonçalves (Hospital São José); Micheli C Arruda (Hospital Ayres de Moura Antônio Bezerra, Fortaleza, CE, Brazil. Hospital e Maternidade Dona Helena); Mariana Regina da Cunha (Hospital de Amor ‑ Fundação PIO Sepaco, Sao Paulo, SP, Brazil. Hospital Japones Santa Cruz, Sao Paulo, SP, Bra‑ 29 30 XII); Mariana Bonomini F de Almeida (Hospital Baía Sul); Juliano Ramos (Centro zil. Hospital da Cidade, Salvador, BA, Brazil. Santa Casa de Misericórdia de Hospitalar Unimed); Bruna M Binda (Hospital Maternidade São José); Priscila São João Del Rei, Belo Horizonte, MG, Brazil. Hospital Ana Nery, Salvador, BA, 32 33 L S Almeida (Hospital Unimed Vitória); Marcia Maria R de Oliveira (Hospital Brazil. Fundação São Francisco de Assis, Belo Horizonte, MG, Brazil. Hospital Distrital Evandro Ayres de Moura Antônio Bezerra); Luciana S de Mattos Regional Dr. Clodolfo Rodrigues de Melo, Maceio, AL, Brazil. Hospital Erasto (Hospital da Cidade); Samara G da Silva (Santa Casa de Misericórdia de São Gaertner, Curitiba, PR, Brazil. Hospital das Clínicas da Universidade Federal de João Del Rei); Daniela C Dorta (Hospital Ana Nery); Martha Hadrich (Santa Goiás, Goiânia, GO, Brazil. Hospital Evangélico de Cachoeiro de Itapemirim, Casa de Misericórdia de Porto Alegre); Fernanda A F Gonçalves (Hospital das Cachoeiro de Itapemirim, ES, Brazil. Hospital Universitário Regional do Norte Clínicas da Universidade Federal de Goiás); ); Kaytiussia R de Sena (Instituto de do Paraná, Londrina, PR, Brazil. Department of Anesthesiology, Pain and Criti‑ Cardiologia do Distrito Federal); Pamella M dos Prazeres (Hospital Evangélico cal Care‑Hospital São Paulo, Escola Paulista de Medicina, Universidade Federal de Cachoeiro de Itapemirim); Josiane Festti (Hospital Universitário Regional do de Sao Paulo, Sao Paulo, SP, Brazil. Norte do Paraná). Received: 9 January 2023 Accepted: 24 March 2023 Author contributions FGZ, ABC, LUT, TCL, AS‑N, LCPA, APNJr, LPD and FRM contributed to the study conception and design. Several authors contributed to data collection (except FGZ, ABC, AS‑N, LCPA, RNS, LPD). FGZ, RHNS and LPD analyzed data. FGZ wrote the first draft of the manuscript and all authors revised it. All authors read and References approved the final manuscript. 1. Rudd KE, Johnson SC, Agesa KM, Shackelford KA, et al. Global, regional, and national sepsis incidence and mortality, 1990–2017: analysis for the Funding Global Burden of Disease Study. Lancet. 2020;395:200–11. The study was funded the Brazilian Ministry of Health through the PROADI‑ 2. Zhao GJ, Li D, Zhao Q, et al. Incidence, risk factors and impact on out‑ SUS (Programa de Desenvolvimento Institucional do Sistema Único de Saúde). comes of secondary infection in patients with septic shock: an 8‑ year retrospective study. Sci Rep. 2016;6:38361. Availability of data and materials 3. Denstaedt SJ, Singer BH, Standiford TJ. Sepsis and nosocomial infection: De‑identified data, data dictionary, and analyses scripts used are available patient characteristics, mechanisms, and modulation. Front Immunol. upon reasonable request and after approval of the steering committee of the 2018;9:2446. proposal and after agreement of Brazilian Ministry of Health for data sharing. 4. Digiovine B, Chenoweth C, Watts C, Higgins M. The attributable mortality and costs of primary nosocomial bloodstream infections in the intensive care unit. Am J Respir Crit Care Med. 1999;160:976–81. Declarations 5. Pittet D, Tarara D, Wenzel RP. Nosocomial bloodstream infection in critically ill patients. Excess length of stay, extra costs, and attributable Ethics approval and consent to participate mortality. JAMA. 1994;271:1598–601. The protocol was approved at the ethics committee of the coordinating site 6. van Vught LA, Klein Klouwenberg PM, Spitoni C, et al. Incidence, risk and at all other including sites (and is available with Additional file 2). Due to factors, and attributable mortality of secondary infections in the intensive the retrospective nature of the analysis, consent was waived. care unit after admission for sepsis. JAMA. 2016;315:1469–79. 7. Machado FR, Cavalcanti AB, Bozza FA, et al. The epidemiology of sepsis in Brazilian intensive care units (the Sepsis PREvalence Assessment Zampier i et al. Annals of Intensive Care (2023) 13:32 Page 13 of 13 Database, SPREAD): an observational study. Lancet Infect Dis. 2017;17:1180–9. 8. Markwart R, Saito H, Harder T, et al. Epidemiology and burden of sepsis acquired in hospitals and intensive care units: a systematic review and meta‑analysis. Intensive Care Med. 2020;46:1536–51. 9. von Cube M, Timsit JF, Schumacher M, Motschall E, Schumacher M. Quantification and interpretation of attributable mortality in core clinical infectious disease journals. Lancet Infect Dis. 2020;20:e299–306. 10. Fleiss JL, Tytun A, Ury HK. A simple approximation for calculating sample sizes for comparing independent proportions. Biometrics. 1980;36:343–6. 11. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of clas‑ sifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40:373–83. 12. Zampieri FG, Iwashyna TJ, Viglianti EM, et al. Association of frailty with short‑term outcomes, organ support and resource use in critically ill patients. Intensive Care Med. 2018;44:1512–20. 13. Labrecque JA, Hunink MMG, Ikram MA, Ikram MK. Do case–control stud‑ ies always estimate odds ratios? Am J Epidemiol. 2021;190:318–21. 14. Zhao W, Chen YQ, Hsu L. On estimation of time‑ dependent attribut‑ able fraction from population‑based case–control studies. Biometrics. 2017;73:866–75. 15. Bruzzi P, Green SB, Byar DP, Brinton LA, Schairer C. Estimating the popula‑ tion attributable risk for multiple risk factors using case–control data. Am J Epidemiol. 1985;122:904–14. 16. Miettinen OS. Proportion of disease caused or prevented by a given exposure, trait or intervention. Am J Epidemiol. 1974;99:325–32. 17. R Core Team (2021). R: a language and environment for statistical com‑ puting. R Foundation for Statistical Computing, Vienna, Austria. https:// www.R‑ proje ct. org/. 18. Van der Wal WM, Geskus RB. IPW: an R package for inverse probability weighting. J Stat Softw. 2011;43:1–3. 19. Vincent JL, Moreno R, Takala J, Willatts S, DeMendonça A, Bruining H, Reinhart CK, Suter PM, Thijs LG. The SOFA (Sepsis‑related Organ Failure Assessment) score to describe organ dysfunction/failure. On behalf of the Working Group on Sepsis‑Related Problems of the European Society of Intensive Care Medicine. Intensive Care Med. 1996;199622(7):707–10. https:// doi. org/ 10. 1007/ BF017 09751. 20. Singer M, Deutschman CS, Seymour CW, et al. The third international consensus definitions for sepsis and septic shock (Sepsis‑3). JAMA. 2016;315:801–10. 21. Pedersen PB, Hrobjartsson A, Nielsen DL, Henriksen DP, Brabrand M, Las‑ sen AT. Prevalence and prognosis of acutely ill patients with organ failure at arrival to hospital: a systematic review. PLoS ONE. 2018;13: e0206610. 22. Donnelly JP, Chen SC, Kauffman CA, et al. Revision and update of the con‑ sensus definitions of invasive fungal disease from the European organiza‑ tion for research and treatment of cancer and the mycoses study group education and research consortium. Clin Infect Dis. 2020;71:1367–76. 23. Tresker S. A typology of clinical conditions. Stud Hist Philos Biol Biomed Sci. 2020;83: 101291. 24. Dahlqwist E, Sjolander A. AF: Model‑Based Estimation of Confounder ‑ Adjusted Attributable Fractions. R package version 0.1.5, 2019. Available on https:// CRAN.R‑ proje ct. org/ packa ge= AF . Accessed 15 July 2022. 25. Ferenci T. Variable selection should be blinded to the outcome. Int J Epidemiol. 2017;46(3):1077–9. https:// doi. org/ 10. 1093/ ije/ dyx048. (PMID: 28402483). 26. Mishra S, Baral SD. Rethinking the population attributable fraction for infectious diseases. Lancet Infect Dis. 2020;20:155–7. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub‑ lished maps and institutional affiliations. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Annals of Intensive Care Springer Journals

Attributable mortality due to nosocomial sepsis in Brazilian hospitals: a case–control study

Download PDF
13 pages

Loading next page...
 
/lp/springer-journals/attributable-mortality-due-to-nosocomial-sepsis-in-brazilian-hospitals-rvTuO0frXa

References (32)

Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2023
eISSN
2110-5820
DOI
10.1186/s13613-023-01123-y
Publisher site
See Article on Publisher Site

Abstract

Background Nosocomial sepsis is a major healthcare issue, but there are few data on estimates of its attributable mortality. We aimed to estimate attributable mortality fraction (AF) due to nosocomial sepsis. Methods Matched 1:1 case–control study in 37 hospitals in Brazil. Hospitalized patients in participating hospitals were included. Cases were hospital non‑survivors and controls were hospital survivors, which were matched by admission type and date of discharge. Exposure was defined as occurrence of nosocomial sepsis, defined as antibiotic prescription plus presence of organ dysfunction attributed to sepsis without an alternative reason for organ failure; alternative definitions were explored. Main outcome measurement was nosocomial sepsis‑attributable fractions, esti‑ mated using inversed‑ weight probabilities methods using generalized mixed model considering time‑ dependency of sepsis occurrence. Results 3588 patients from 37 hospitals were included. Mean age was 63 years and 48.8% were female at birth. 470 sepsis episodes occurred in 388 patients (311 in cases and 77 in control group), with pneumonia being the most common source of infection (44.3%). Average AF for sepsis mortality was 0.076 (95% CI 0.068–0.084) for medical admissions; 0.043 (95% CI 0.032–0.055) for elective surgical admissions; and 0.036 (95% CI 0.017–0.055) for emergency surgeries. In a time‑ dependent analysis, AF for sepsis rose linearly for medical admissions, reaching close to 0.12 on day 28; AF plateaued earlier for other admission types (0.04 for elective surgery and 0.07 for urgent surgery). Alterna‑ tive sepsis definitions yield different estimates. *Correspondence: Fernando G. Zampieri fzampieri@hcor.com.br Full list of author information is available at the end of the article © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. Zampieri et al. Annals of Intensive Care (2023) 13:32 Page 2 of 13 Conclusion The impact of nosocomial sepsis on outcome is more pronounced in medical admissions and tends to increase over time. The results, however, are sensitive to sepsis definitions. Keywords Sepsis, Attributable mortality, Epidemiology Introduction Additional file  1: for details). The protocol was approved Sepsis is a major healthcare issue that may account for at the ethics committee of the coordinating site and at more than 11 million yearly deaths worldwide [1]. While all other including sites (and is available with Additional most cases are community-acquired, nosocomial sepsis file  2). Due to the retrospective nature of the analysis, is an important source of burden for healthcare systems. consent was waived. First patient included was admitted Nosocomial sepsis has been repeatedly associated with to the hospital in May 2018 and last patient included was increase in costs, hospital length-of-stay and mortality admitted to the hospital in January 2020. [2, 3], with most reports focusing on the consequences of sepsis in the population of critically ill patients [4–8], Objective including those who were first admitted due to commu - To estimate attributable mortality fraction due to noso- nity-acquired sepsis [3, 6]. Attributable mortality frac- comial sepsis in adult patients. tion is defined as the proportion of deaths that occur related to an exposure, that is, the proportion of deaths Case and control definitions and matching that would not have occurred if the exposure did not take Cases were defined as hospital deaths and controls as place. The attributable mortality fraction of ICU-acquired patients who were discharged alive. After local ethics infections in critically ill patients has been suggested to committee approval, sites were instructed to obtain a be of close to 11% among patients admitted with sepsis list of the most recent 50 adult patients who died dur- and 21% among patients without sepsis [6], with values ing hospitalization. These patients were paired to the as high as 35% being reported for specific infections [5]. closest temporal patient discharged alive who had the There are few reports on occurrence and impact of same admission type (medical, elective surgical or urgent nosocomial infections after hospital admission in a surgical) as the case patient. A margin of 30  days was broader population [8]. In a recent meta-analysis, only allowed for matching. Matching was manually made; eight studies reported hospital wide data on nosocomial for each case, locally trained personal obtained a list for sepsis, but no study reported an estimate of attributable all patients discharge alive at the same day and checked mortality fraction of nosocomial sepsis [8]. Obtaining a for matching, if unsuccessful more medical records for a good estimate of attributable mortality fraction is impor- wider margin were obtained until the closest discharge tant because it is a major driver of public health deci- alive was obtained. For example, a patient who was sions [9]. Considering that overall hospital mortality is admitted for a medical reason and died was paired with low, a case–control design may provide a good estimate the closest medical admission patient discharged alive of attributable mortality fraction and inform healthcare from the same institution. No other matching method policy makers. was performed. Elective surgical admissions were defined We conducted a case–control study aiming to estimate as admissions due to scheduled surgery. Emergency sur- attributable mortality of nosocomial sepsis in Brazilian gical admissions were defined as those whose surgery hospitals. We hypothesized that nosocomial infections was indicated in the first 24 h after admission. would represent an important burden with high attribut- able mortality fraction. Due to lack of consensus on the Sample size calculation operational definitions of nosocomial sepsis, we have also Sample size was calculated using Fleiss, Tytun and stressed data with different definitions and compared our Ury method to estimate a difference in proportions estimates. [10]. Assuming a prevalence of sepsis in 15% in control patients, 1500 patients per group were required to have a 90% power to detect an odds ratio of at least 1.4 for Methods the association between nosocomial sepsis and mortal- Design: observational, case–control study ity. Sample size was increased to 1800 patients per group Setting after the first protocol amendment before data were col - Thirty-eight hospitals in Brazil. Hospitals in Brazil lected due to an increase in fund availability from the that had at least 100 beds and that had a critical care sponsor; this yields a similar power for an odds ratio as unit were eligible for participating in this study (see Zampier i et al. Annals of Intensive Care (2023) 13:32 Page 3 of 13 low as 1.35. We specified that each participating hos - Statistical analysis pital would include 100 patients (50 cases and 50 con- Descriptive statistics were used for univariate analyses. trols); therefore, the increase in sample size was made by The primary objective was to estimate attributable frac - increasing the number of participating hospitals from 30 tion of mortality due to nosocomial sepsis. We assumed to 38, to account for possible drop-offs and incomplete that in-hospital mortality would be a sufficiently rare pairs. event so that odds ratio estimates obtained from a case control design would be reliable estimators of relative risk for calculation of the nosocomial sepsis mortality Data collection attributable fraction [13–15]. Overall hospital data were collected in a structured case The model adjustment was defined a priori, before report form. A local data collector at each site was trained data collection, according to the protocol available in by the sponsor (HCor Research Institute) before data col- Additional file  2; adjustment variables were elected due lection. Patient data included demographic information, to clinical relevant. The following approach was used comorbidities (using Charlson Comorbidity Index and for attributable fraction calculation for each admission Modified Frailty Index) [11, 12], reason for admission and type: the inverted probability weights (IPW) for each daily data (from admission to 28  days) on suspicion of patient were calculated, representing the cumulative infection, antibiotic use, occurrence of organ failure (see risk of the patient acquiring sepsis during hospitaliza- below), and occurrence of other non-infectious clinically tion, under a multivariable logistic regression analysis relevant events. The suspected infection was diagnosed at including baseline age, Charlson Comorbidity Index, local physician’s discretion and collected from the charts. and a time-dependent variable for the occurrence of Clinically relevant events were by the steering commit- clinically relevant events. We estimated the association tee on agreement, and were defined as occurrence of between sepsis occurrence on hospital mortality within stroke, unstable angina or acute myocardial infarction, 28  days through a mixed logistic regression model severe acute hypertension (hypertension that demanded weighted by IPW. The model included the partici - medical intervention), fall, seizure, pulmonary embolism, pant center as a random intercept, and age, Charlson bronchospasm, lower or upper intestinal track bleeding, Comorbidity Index, infection at admission (to account and need for surgical procedure. for community-acquired sepsis at admission), IPW, and the accumulated dependent time variables of sep- Sepsis definition sis occurrence and clinically relevant events with their Sepsis was defined as suspected infection requiring interaction with time (modeled with a third-degree pol- antibiotic use plus at least one organ dysfunction in the ynomial). From the daily odds ratios estimated by the absence of other clinically relevant events occurring in model, the Miettinen formula was used to calculate the the same day. The following criteria were used for diagno - attributable mortality of sepsis [16]: sis acute organ failure: systolic blood pressure < 90 mmHg and/or mean arterial pressure < 65 mmHg and/or drop in OR − 1 P ∗ , systolic blood pressure > 40  mmHg; arterial partial pres- i OR sure to inspired oxygen fraction ratio below 300 or need for supplementary oxygen to maintain peripheral oxygen i = day, saturation above 90%; abnormal mental status; increase P = proportion of patients that had sepsis among non − survivors. in serum creatinine to values of at least 2  mg/dL and/or The attributable fraction assessed this way therefore urinary output < 0.5 mL/kg/h for 2 h; total bilirubin con- represents the proportion of deaths occurring up to centration above 2 mg/dL; platelet count < 100,000 units/ that day given that the patient had sepsis before that mm ; abnormal coagulation in the absence of anticoagu- given day; that is, how the percentage of deaths that lant use (international normalized ratio above 1.5 or acti- would not have occurred if the patient did not have vated partial thromboplastin time above 60 s). If a patient sepsis up to that moment. We also present the mar- developed the before-mentioned criteria up to the sec- ginal effect of having one sepsis episode on the whole ond day of hospital admission, the infection was consid- period, with an estimate of average attributable fraction ered as not nosocomial and was therefore not considered obtained from the model. All analyses were performed for attributable fraction estimates; similarly, patients that using the R 4.1.1 software [17]. IPW was estimated were admitted with community-acquired infection and using the IPW package [18]. The delta method was used that died without having a second event within hospital to calculate the 95% confidence intervals of the odds admission (that is, without having nosocomial sepsis), ratio and attributable mortality. were not considered in nosocomial sepsis-attributable fraction calculation. Zampieri et al. Annals of Intensive Care (2023) 13:32 Page 4 of 13 Sensitivity analysis Results Different definitions of sepsis were explored (Table  1); Hospital and patient features the same analysis was performed using different defi- A total of 1794 pairs of cases and controls from 37 hospi- nitions as sensitivity analyses. An internal consensus tals were included in the analysis (one additional hospi- was created within the steering committee on prob- tal agreed to participate but did not collect data or filled ability degrees of sepsis according to combinations of forms and was excluded from the analysis). Hospital fea- use of antibiotics, organ failure, results of cultures, and tures are shown in Additional file  1: Table S1. Six hospi- occurrence of clinically relevant events (Table 1; Addi- tals did not complete the full sample of 100 patients; in 4 tional file  1: Fig. S1). Two ad hoc sensitivity analyses cases because it was impossible to find pairs in the pre- were made, one estimating the attributable fraction specified timeframe. Baseline patient features are shown considering definitive, very probable, and probable in Table  2. Average population age was 63  years (stand- sepsis, and a second sensitivity analysis considering ard deviation of 18.6), and 48.8% were female at birth. only definitive and very probable sepsis. This approach Median Charlson Comorbidity Index was 2 (interquar- was used to assess whether different definitions would tile range from 0 to 4). One-fifth (22%) of all included result in different estimates of attributable fraction. patients had a recorded recent (in the previous month) One additional post hoc analysis based on increase in hospital admission before the current hospital admis- Sequential Organ Failure Assessment score (SOFA) sion, and 23.7% already needed assistance for daily liv- as per Sepsis 3 also performed [19, 20]; in this analy- ing activities. Median hospital length-of-stay was 8  days sis, patients were considered to have nosocomial sep- (interquartile range 4–14  days). Admission with acute sis if their SOFA score increased at least two points infection occurred in 39.5% of non-survivors and 30.0% over hospital admission SOFA score values and a new of survivors. Patient locale or status over time is shown antibiotic was started, regardless of other informa- in Additional file  1: Fig. S2. Patient features according to tion on clinically relevant events. A sensitivity analy- development or not of nosocomial sepsis are shown in sis using the main definition but excluding patients Additional file 1: Table S2. admitted with infection was also performed. Finally, A total of 388 patients (311 cases and 77 control, com- two additional sensitivity analysis were conducted for prising 10.8% of all included patients) had 470 nosoco- non-surgical patients adding main diagnosis category mial sepsis episodes (387 in cases and 83 in controls). (neurological, cardiovascular, infection, renal, abdomi- Days of antimicrobial therapy were 603 and 437 per nal, or others) as predictors. 1000 patients/day for cases and controls, respectively. Table 1 Sepsis definitions used for analysis Antibiotic Organ failure Cultures Other clinically Sepsis source Part of Part of prescribed relevant event information alternative alternative available definition 1 definition 2 Primary definition Main sepsis definition Yes Yes Positive or negative No Yes – – Alternative definition Definitive Yes Yes Positive No Yes Yes Yes Very probable, either: (1) or Yes Yes Positive Yes Yes Yes Yes (2) or Yes Yes Negative No Yes Yes Yes (3) Yes Yes Negative Yes Yes Yes Yes Probable No Yes NA No No Yes No Possible No Yes NA Yes No No No SOFA increase definition based on Sepsis 3 (post hoc) SOFA definition Yes Yes – Variable No No (“Sepsis 3”) NA data not collected Data on cultures were only collected for patients that received antibiotics. SOFA definition was performed post hoc. Any patients that had an increase in SOFA score of at least 2 points over baseline (hospital admission) SOFA and that either were started antibiotics or received new antibiotics were considered as septic Zampier i et al. Annals of Intensive Care (2023) 13:32 Page 5 of 13 Table 2 Patient features, resource use and occurrence of clinically relevant events in cases and controls Non-survivors Survivors p value (n = 1794) (n = 1794) Age, mean (SD) 68.5 (17.2) 57.5 (18.4) < 0.001 Sex at birth, n (%) 0.095 Female 850 (47.4%) 901 (50.2%) Male 944 (52.6%) 893 (49.8%) Charlson Comorbidity Index, median [IQR] 2 [1–6] 1 [0–3] < 0.001 Modified Frailty Index, median [IQR] 2 [1–3] 1 [0–2] < 0.001 Previous hospitalization (last month), n (%) 471 (26.3%) 319 (17.8%) < 0.001 Pneumonia on past year, n (%) 175 (9.8%) 79 (4.4%) < 0.001 Episode of mental confusion on past year, n (%) 238 (13.3%) 116 (6.5%) < 0.001 Previously on hospice/long‑term facility/homecare, n (%) 93 (5.2%) 39 (2.2%) < 0.001 Dependency for daily living activities, n (%) 620 (34.6%) 232 (12.9%) < 0.001 Known comorbidities at admission, n (%) Dementia 196 (10.9%) 58 (3.2%) < 0.001 Transitory Ischemic Attack 18 (1.0%) 18 (1.0%) 1.00 Stroke 174 (9.7%) 83 (4.6%) < 0.001 Previous myocardial infarction 137 (7.6%) 110 (6.1%) 0.086 Angina/coronary stent 137 (7.6%) 129 (7.2%) 0.656 Heart failure 221 (12.3%) 144 (8.0%) < 0.001 Hypertension 912 (50.8%) 767 (42.8%) < 0.001 Diabetes, uncomplicated 473 (26.4%) 403 (22.5%) 0.007 Diabetes, complicated 136 (7.6%) 119 (6.6%) 0.299 Rheumatologic disease 66 (3.7%) 78 (4.3%) 0.349 Acquired immunodeficiency syndrome 50 (2.8%) 43 (2.4%) 0.529 Cirrhosis 65 (3.6%) 37 (2.1%) 0.006 Cancer 726 (40.5%) 453 (25.3%) < 0.001 Hospital admission Admission type Medical 1524 (84.9%) 1524 (84.9%) – Elective surgery 149 (8.3%) 149 (8.3%) – Urgent surgery/trauma 121 (6.7%) 121 (6.7%) – Relevant diagnosis at admission Infection 709 (39.5%) 538 (30.0%) < 0.001 Respiratory diagnosis Asthma 11 (0.6%) 19 (1.1%) 0.199 Chronic pulmonary obstructive disease 87 (4.8%) 53 (3.0%) 0.004 Other chronic lung disease 19 (1.1%) 16 (0.9%) 0.735 Cardiac diseases ST‑ elevation myocardial infarction 31 (1.7%) 38 (2.1%) 0.466 Non‑ST ‑ elevation myocardial infarction 33 (1.8%) 28 (1.6%) 0.606 Unstable angina 18 (1.0%) 31 (1.7%) 0.083 Angina, unspecified 12 (0.7%) 9 (0.5%) 0.663 Uncompensated heart failure 137 (7.6%) 87 (4.8%) 0.001 Deep vein thrombosis 48 (2.7%) 32 (1.8%) 0.089 Pulmonary thromboembolism 27 (1.5%) 27 (1.5%) 1.00 Neurological diseases Ischemic stroke 83 (4.6%) 58 (3.2%) 0.039 Hemorrhagic stroke 18 (1.0%) 10 (0.6%) 0.183 Transient ischemic attack 3 (0.2%) 12 (0.7%) 0.035 Subarachnoid hemorrhage 17 (0.9%) 8 (0.4%) 0.107 Zampieri et al. Annals of Intensive Care (2023) 13:32 Page 6 of 13 Table 2 (continued) Non-survivors Survivors p value (n = 1794) (n = 1794) Polyradiculopathy/myasthenia 3 (0.2%) 3 (0.2%) 1.00 Renal diseases Acute, non‑related to cirrhosis 91 (5.1%) 39 (2.2%) < 0.001 Chronic, not on dialysis 75 (4.2%) 43 (2.4%) 0.004 Chronic, needing dialysis 31 (1.7%) 40 (2.2%) 0.338 Abdominal diseases Uncompensated cirrhosis 36 (2.0%) 20 (1.1%) 0.042 Digestive bleeding 12 (0.7%) 7 (0.4%) 0.358 Spontaneous bacterial peritonitis 2 (0.1%) 0 (0.0%) 0.500 Hepatorenal syndrome 3 (0.2%) 0 (0.0%) 0.250 Acute pancreatitis 24 (1.3%) 27 (1.5%) 0.778 Uncompensated diabetes 70 (3.9%) 61 (3.4%) 0.477 Admission for diagnostic procedures 535 (29.8%) 515 (28.7%) 0.463 Other 648 (36.1%) 640 (35.7%) 0.808 Resource use Intensive care unit admission, n (%) 853 (47.5%) 340 (19%) < 0.001 Hospital length‑ of‑stay, median [IQR] 9 [5–19] 6 [4–11] < 0.001 Antibiotic days of therapy, days per 1000 patients/day 603 437 < 0.001 Days using antibiotics, median [IQR] 5.5 [1.3–11] 1 [0–6] < 0.001 Events during hospitalization Sepsis (main definition) 525 (29.3%) 232 (12.9%) Up to 2 days from admission 383 (21.3%) 198 (11%) < 0.001 After 2 days from admission 311 (17.3%) 77 (4.3%) < 0.001 Stroke 65 (3.6%) 28 (1.6%) < 0.001 b,c Coronary syndr ome 80 (4.5%) 42 (2.3%) 0.001 Acute severe hypertensive episode 228 (12.7%) 133 (7.4%) < 0.001 Fall 63 (3.5%) 20 (1.1%) < 0.001 Seizure 93 (5.2%) 34 (1.9%) < 0.001 Pulmonary thromboembolism 40 (2.2%) 11 (0.6%) < 0.001 Bronchospasm 118 (6.6%) 27 (1.5%) < 0.001 Digestive bleeding 122 (6.8%) 27 (1.5%) < 0.001 Severe pain 138 (7.7%) 76 (4.2%) < 0.001 Matching variable ST-elevation myocardial infarction Non-ST-elevation myocardial infarction Pain episode that required more than 2 rescues or new diagnostic procedure Pneumonia was the most common source (Additional Additional file  1: Table  S5 (74 patients with any positive file  1: Table  S3, Fig. S3). The most common organ dys - culture). function at sepsis diagnosis was abnormal mental status (Fig.  1). More than half of the patients had more than Attributable mortality fraction one of the specified organ dysfunctions during their Daily odds ratio for mortality obtained from the model first sepsis episode (Additional file  1: Fig. S4). Sepsis was and its respective attributable fraction are shown in more frequently diagnosed in ICU, followed by ward Fig.  2. The reported odds ratio is the effect size of dying (Additional file  1: Fig. S5). There were 573 positive cul - up to a specific day (x-axis) given the patient had infec - tures in the population (Additional file  1: Table S4, strati- tion in the preceding days (up to hospital admission). fied for suspected source); a list of pathogens in patients Attributable fraction is interpreted as percentage of with sepsis according to the main definition is shown in deaths occurring up to that day given that the patient Zampier i et al. Annals of Intensive Care (2023) 13:32 Page 7 of 13 A − All episodes of sepsis 27 27 27 14 14 12 12 11 11 6 66 6 Abnormal Mental Status 175 Hypotension Coagulopathy 120 Hypoxemia 119 High Lactate 114 Kidney Injury 41 Hepatic 300 200 100 0 Set size B − Pneumonia 88 8 8 3 33 33 3 3 3 128 Abnormal Mental Status 72 Hypotension 61 Hypoxemia Coagulopathy 51 High Lactate Kidney Injury 10 Hepatic 150 100 50 0 Set size C − Not pneumonia 8 8 66 6 5 5 4 4 4 3 3 135 Abnormal Mental Status 103 Hypotension 92 Coagulopathy 68 High Lactate 65 Kidney Injury 59 Hypoxemia Hepatic 200 150 100 50 0 Set size Fig. 1 Organ failure at diagnosis in A all septic patients, B pneumonia, C not pneumonia. Each dot represents intersections, with the number of cases shown in bars previously had sepsis. For medical admissions, attribut- urgent surgery patients, with peak values around 0.04– able fraction rose after the sixth day of admission, and 0.07, and a less linear increase over time. reached close to 0.12 at 28  days. The effect of nosoco - Overall marginal odds ratio for mortality and attrib- mial sepsis was less pronounced on elective surgery and utable fraction for patients with any sepsis episode were Intersection size Intersection size Intersection size Zampieri et al. Annals of Intensive Care (2023) 13:32 Page 8 of 13 Medical (n = 3048) Elective surgery (n = 298) Emergency Surgery / Trauma (n = 242) 8 8 8 7 7 7 6 6 6 5 5 5 4 4 4 3 3 3 2 2 2 1 1 1 0 0 0 3 45678 910111213141516171819202122232425262728 3 45678 910111213141516171819202122232425262728 3456789 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 Day Day Day 0.14 0.14 0.14 0.12 0.12 0.12 0.10 0.10 0.10 0.08 0.08 0.08 0.06 0.06 0.06 0.04 0.04 0.04 0.02 0.02 0.02 0.00 0.00 0.00 −0.02 −0.02 −0.02 −0.04 −0.04 −0.04 3 45678 910111213141516171819202122232425262728 3 45678 910111213141516171819202122232425262728 3456789 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 Day Day Day Sepsis 75 144 192 231 260 277 296 304 311 320 326 331 335 340 345 348 352 353 355 358 359 361 361 361 362 362 Sepsis 6 7 9 10 10 10 10 10 11 11 11 11 11 12 12 12 12 12 13 13 13 13 Sepsis 5 6 8 8 8 10 10 12 12 12 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 Death 231 356 450 533 617 713 789 861 909 955 994 1027 1052 1089 1112 1142 1162 1190 1216 1235 1258 1278 1293 1305 1315 1339 Death 29 33 43 50 59 62 68 72 74 81 83 86 91 95 102 105 108 111 114 115 116 118 Death 28 44 53 57 62 65 70 74 75 80 83 85 89 90 92 96 97 99 100 101 101 104 105 106 107 107 3 45678 910111213141516171819202122232425262728 3 45678 910111213141516171819202122232425262728 3456789 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 Day Day Day Fig. 2 Distribution of odds ratio (upper row) and AF (lower row) according to admission type (columns). The odds ratio should be interpreted as the odds ratio of dying up to a specific day (x‑axis) given that the patient has acquired sepsis totalizing 1121 septic episodes (945 on non-survivors 1.73 (95% CI 1.60; 1.87), attributable fraction = 0.076 and 176 in survivors (Additional file  1: Fig. S8 shows (95% CI 0.068; 0.084); 2.75 (95% CI 1.47; 4.03), attrib- infection sources and organ dysfunctions). Marginal utable fraction = 0.043 (95% CI 0.032; 0.055); and 1.75 odds ratio for mortality and attributable fraction under (95% CI 1.06; 2.43), attributable fraction = 0.036 (95% this definition were 1.60 (95% CI 1.51; 1.70), attributable CI 0.017; 0.055]), for medical, elective surgery and fraction = 0.141 (95% CI 0.128; 0.155); 3.61 (95% CI 2.90; emergency surgery patients, respectively. 4.32, attributable fraction = 0.392 (95% CI 0.363; 0.422); and 2.57 (95% CI 2.04; 3.10), attributable fraction = 0.251 Alternative analysis 1: definitive, very probable (95% CI 0.218; 0.284), for medical, elective surgical and and probable sepsis emergency surgery groups. Results for odds ratio and Using this broader definition, a total of 1129 patients had attributable fraction over time are shown in Additional nosocomial sepsis during hospital stay. A total of 1387 file 1: Fig. S9. septic episodes were recorded (1058 on non-survivors and 329 in survivors). 74 patients had definitive sepsis, Post hoc analysis: SOFA definition (based on Sepsis 3) 850 very probable sepsis and 205 probable sepsis (see 729 patients had nosocomial sepsis as by an increase in Additional file  1: Fig. S6 for details on infection source of SOFA score of at least two points over baseline (869 sep- organ dysfunction). The overall marginal odds ratio for tic episodes: 782 on non-survivors and 87 in survivors). mortality and attributable fraction for patients with one Marginal odds ratio for mortality and attributable frac- episode of definite, very probable, or probable sepsis were tion under this definition were 2.25 (95% CI 2.10; 2.39), 1.32 (95% CI 1.25; 1.39), attributable fraction = 0.101 attributable fraction = 0.176 (95% CI 0.167; 0.185); 3.52 (95% CI 0.083; 0.118); 2.85 (95% CI 2.32; 3.39), attrib- (95% CI 2.82; 4.22, attributable fraction = 0.358 (95% CI utable fraction = 0.374 (95% CI 0.337; 0.412); and 2.21 0.330; 0.386); and 2.93 (95% CI 2.30; 3.56), attributable (95% CI 1.77; 2.64), attributable fraction = 0.230 (95% CI fraction = 0.246 (95% CI 0.219; 0.274), for medical, elec- 0.192; 0.268) for medical, elective surgery and emergency tive surgical and emergency surgery groups. Results for surgery groups. Results over time are shown Additional odds ratio and attributable fraction over time are shown file  1: Fig. S7; this definition resulted in an attributable in Additional file 1: Fig. S10. fraction of up to 0.25 for medical admissions, and higher values for elective and urgent surgery than the main defi - Post hoc analysis: exclusion of patients admitted nition used (peaking over 0.45 at 28 days for elective sur- with infection under the main definition gery), with a more linear ascend over time. A total of 3007 patients were considered in this analy- sis—1411 non-survivors and 1596 survivors—with a Alternative analysis 2: definitive and very probable sepsis total of 177 and 37 episodes of sepsis, respectively. Odds A total of 924 patients had nosocomial sepsis as defined ratio for mortality and attributable fraction were 2.21 using the definitive and very probable sepsis criteria, Attributable fraction Odds Ratio Attributable fraction Odds Ratio Attributable fraction Odds Ratio Zampier i et al. Annals of Intensive Care (2023) 13:32 Page 9 of 13 Under this definition, attributable fraction was higher for (95% CI 1.95; 2.47), attributable fraction = 0.055 (95% medical admissions than for elective surgical and emer- CI 0.05; 0.061); 2.84 (95% CI 1.35; 4.31, attributable frac- gency surgical patients. Attributable fraction over time tion = 0.034 (95% CI 0.024; 0.043); and 1.71 (95% CI 0.94; linearly increased up to 28 days for medical patients, but 2.48), attributable fraction = 0.035 (95% CI 0.013; 0.057), not for elective and emergency surgery admissions where for medical, elective surgical and emergency surgery a plateau was observed. Pneumonia was the most com- groups (Additional file 1: Fig. S11). mon infection source, and abnormal mental status was the commonest organ dysfunction observed. Antibiotic Post hoc analysis: additional adjustment for main reason use was high in both groups. for admission using main definition and the SOFA Different sepsis definitions will inevitably be associ - definition for non-surgical patients ated with different prevalence and effect sizes for mor - Adding further adjustment according to main admission tality, with consequential direct impact on attributable category for non-surgical patients analyses yield an aver- mortality fraction. We also explored additional defini - age odds ratio for mortality of 1.64 (95% CI 1.51; 1.76) tions of sepsis according to key features including posi- with an average attributable fraction of 0.070 (95% CI tive cultures, occurrence of clinically relevant events, of 0.061; 0.0782); time-dependent effects are shown in and antibiotic use. The fact that clinically relevant events, Additional file  1: Fig. S12. The only admission type that obtained directly from chart review, was considered was consistently associated with an attributable fraction before attributing organ failure to infection may enhance above 0.10 was admission due to infection (Additional the capability of measuring the effect associated directly file  1: Fig. S13), reinforcing the importance of adjustment with infection. Indeed, organ failure is associated with for infection in the primary analysis. For the increase a myriad of clinical conditions [21]; this creates a situa- SOFA definition (Sepsis 3), adding admission category tion where antibiotics are prescribed due to new organ resulted in average odds ratio was 2.39 (95% CI 2.24; failure even if coexisting events that may be responsible 2.54) and average attributable fraction was 0.184 (95% CI for organ failure occur simultaneously. Our main defini - 0.176; 0.192). tion was stricter than other sepsis definitions by limiting sepsis diagnosis to the absence of coexisting events that Comparison of the definitions could cause organ failure at the same day [20]. The two Results for the comparison of the odds ratio and attribut- alternative definitions were more comprehensive, as seen able fraction for medical admission are shown in Fig.  3. by the higher number of cases reported: the first being Odds ratio for mortality increased over time for all defi - broader than the second. The first alternative attributed nitions; however, due to differences in prevalence of new organ dysfunctions in the absence of clinically rel- events, the attributable fraction was lower for the main evant events as septic events even in the absence of anti- definition when compared to both alternative definitions. biotic use (“any new organ dysfunction in a hospitalized Comparison for the definitions for elective surgery and patient is sepsis until proven otherwise”), while the sec- emergency surgery patients is shown in Additional file  1: ond did not include such patients. The resulting attrib - Fig. S14 and S15; the main definition provided the lower utable fractions reflected the changes in both effect size AF for surgical patients. The post hoc analysis based (odds ratio) and prevalence; despite being associated with on SOFA score increase as per Sepsis 3 suggestion pro- higher odds ratios for mortality, the main definition had vided resulted in estimates for odds ratio and attributable the lowest attributable fraction. Differences in attribut - fraction that largely followed the results of the second able fraction among definitions were specially pressing alternative definition. Exclusion of patients with known in surgical patients. A definition based on increase in sepsis at admission reduced the attributable fraction of SOFA provided results similar to the second alternative nosocomial sepsis due to decrease in number of events definition. When baseline sepsis patients were excluded, (Additional file  1: Fig. S11). Adding further adjustments the attributable fraction was markedly reduced (which for admission type also did not change estimates for non- is expected since infection is an important risk factor surgical patients (Additional file 1: Table S6). for a secondary insult), but the odds ratio for mortality remained high. Discussion These findings have several important implications. In this case–control study including 1794 pairs of patients First, even under the strictest definition sepsis-attribut - from 37 Brazilian hospitals, we found that nosocomial able fraction was still very important, reaching around sepsis, defined by an acute nosocomial infection with 0.12 for prolonged medical admissions; the longer the organ failure in the absence or other clinically relevant medical patient remained in the hospital, the highest the events, was an important contributor to hospital mor- odds ratio for mortality associated with a septic episode. tality with a significant attributable mortality fraction. Zampieri et al. Annals of Intensive Care (2023) 13:32 Page 10 of 13 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 345678 910111213141516171819202122232425262728 Day Main Definition Alternative 1 Alternative 2 0.3 0.2 0.1 0.0 −0.1 345678 910111213141516171819202122232425262728 Day 145 270 363 444 498 547 583 603 627 651 670 687 699 709 716 726 733 739 745 750 753 759 760 761 763 767 Alternative 2 191 344 462 553 626 692 741 764 797 830 851 868 881 892 897 908 913 918 926 931 935 941 942 943 944 948 Alternative 1 75 144 192 231 260 277 296 304 311 320 326 331 335 340 345 348 352 353 355 358 359 361 361 361 362 362 Main Definition 231 356 450 533 617 713 789 861 909 955 994 1027 1052 1089 1112 1142 1162 1190 1216 1235 1258 1278 1293 1305 1315 1339 Death 345678 910111213141516171819202122232425262728 Day Fig. 3 Comparison of the three definitions for medical admissions. Upper panel: the daily odds ratio for mortality in medical patients according to the definition used. Bottom panel: the respective AF for each definition. Note that a higher OR does not equal higher AF due to changes in prevalence This value is aligned with other estimated of attribut - (emergency surgical patients) to 0.076 (medical patients) able fraction in more severe patients and may represent a is sufficiently small so that an intervention may reduce reasonable starting point for quality improvement initia- nosocomial infections without easily noticeable effects in tives [5]. These results, under the main definition, should mortality, unless sample size is very large, under this defi - probably be seen as a minimum value for sepsis-attrib- nition. Since infection burden is not exclusively related to utable fraction in hospitalized patients. Albeit impor- mortality, involving costs, long-term outcomes, quality tant, an average attributable fraction ranging from 0.036 of life, among others, understanding sepsis-attributable Attributable fraction Odds Ratio Zampier i et al. Annals of Intensive Care (2023) 13:32 Page 11 of 13 fraction may avoid over-simplistic conclusions, such as also limited by the assumption that odds ratio and rela- interpreting that an intervention that reduces infection tive risk will be similar when outcomes are infrequent. occurrence but not mortality prevented only non-fatal This is, in fact, the de facto approach made by several infections or even that some infections may not be asso- statistical packages that estimate the attributable frac- ciated with higher mortality at all. tion [24]. Our model adjusted for time-dependency of The variability observed by tweaking sepsis definitions covariates but did not consider further daily informa- should also be seen an alert that it is hard to isolate the tion besides diagnosis of infection and clinically relevant impact of a single event within the intricate path of a hos- events. The main adjustment model was defined a priori, pital stay, specifically for surgical patients where different according to a stablished protocol. All variable selection broader definitions provided results strikingly different approaches may be subject to criticism, but we refrained from medical patients. It is conceivable that “true” sep- from using variables that were associated with either sep- sis-attributable fraction may be somewhere in between sis or mortality as suggested [25]. Adding admission cat- the main definition and the second alternative defini - egories to the main analysis did not change the estimated tion. Far from suggesting that nosocomial sepsis is not of AF significantly. Other approaches could be employed an issue, our results highlight that even when considering to estimate attributable fraction [26]. The list of clinically known factors for poor hospital outcome such as age and relevant events is not exhaustive and was defined by the comorbidities, sepsis could be directly responsible from steering committee during protocol discussions before something between 7.6% and 14.1% of all hospital death data were collected but is somewhat arbitrary. Finally, for medically admitted patients. For surgical patients, our results reflect the Brazilian panorama; it is uncertain the margins are even wider depending on the definition whether they are transposable to other settings. used, peaking over 40%. Attributable fraction is a rela- tive measurement, and not a direct estimate of burden. Conclusion Low-middle income countries are suggested to be more Nosocomial sepsis is an important contributor to hos- affected by nosocomial sepsis, which may result in a pital mortality. The impact of nosocomial sepsis on out - higher numeric burden of deaths in this population. come is more pronounced in medical admissions and We hope that this manuscript fosters the discussion on tends to increase over time. Different sepsis definitions whether sepsis would benefit from a more nuanced diag - led to important changes on attributable fraction. nosis approach where probability categories are used to tailor diagnosis and treatment (as is the case of aspergil- Supplementary Information losis, where possible/probable categories have been in The online version contains supplementary material available at https:// doi. use but may also be applicable to other medical condi-org/ 10. 1186/ s13613‑ 023‑ 01123‑y. tions) [22, 23]. Despite over 30 years of controversy, all Additional file 1: Table S1. Hospital Features. Table S2. Comparison sepsis definitions are binary, that is, they do not consider between patients that develop and did not develop nosocomial sepsis the uncertainty that permeates clinical decision-making, according to the main definition. Table S3. Sepsis episodes. Figure S1. focusing more on severity of illness than in the very prob- Venn diagram for the sepsis definitions used in the manuscript. Figure S2. Daily patient location/status over time for controls (top) and cases ability that the findings are due to active infection. (bottom). Figure S3. Number of nosocomial sepsis episodes according to Our work has several limitations. As with any case infection source. Figure S4. Number of organ dysfunctions at presenta‑ control trial, selecting appropriate controls is challeng- tion for (A) first episode of sepsis, and (B) all sepsis episodes. Figure S5. Patient locale at sepsis diagnosis. Top: Stratified according to outcome; ing. We used the closest temporal admission discharged Bottom: All patients. Table S4. Pathogens isolated from cultures according alive with the same admission type as this seemed a good to suspected site; more than one pathogen was possible for each patient. compromise between feasibility and adequacy. As can CRBI: Catheter related bloodstream infections. Note that pathogens could be isolated in any culture collected from the patient within be seen in Table  1, the number of possible clinical con- the 48h timeframe. The local diagnosis was considered as reference; ditions is very large, and if matching criteria were too therefore the isolated pathogen could not be considered the culprit for strict, we would have ended with issues in obtaining the infection. Table S5. Positive cultures for patients that had one septic episode according to main definition. CRBI: Catheter related bloodstream proper controls; we tried to overcome this by adjusting infections. Note that pathogens could be isolated in any culture col‑ for several relevant confounders, including age, comor- lected from the patient within the 48h timeframe. Same as in Table S3, bidities, etc. Most hospitals in this study still used non- the final diagnosis was made by the site. Figure S6. (A) Infection source considering first alternative definition. The number of patients that did not electronic (paper) healthcare records, thereby making receive antibiotic reflects patients that developed new organ failure in the triaging of possible controls challenging. Restricting our absence of any other clinically relevant event and were considered as pos‑ study to only hospitals with electronic healthcare records sibly septic under this definition. (B) Organ dysfunction at presentation for the alternative definition 1 analysis. Figure S7. Distribution of odds ratio would induce another source of bias, since these hospi- (upper row) and AF (lower row) according to admission type (columns) tals would inherently have more resources. We estimated for the first alternative definition considering the effects of definitive, very attributable fraction from a case–control study, which is probably, and probable sepsis. Figure S8. Sepsis sources (A) and organ Zampieri et al. Annals of Intensive Care (2023) 13:32 Page 12 of 13 Consent for publication dysfunction at presentation for the second alternative definition analysis. Not applicable. Figure S9. Distribution of odds ratio (upper row) and PAF (lower row) according to admission type (columns) for the second alternative defini‑ Competing interests tion considering the effects of definitive and very probably sepsis. Figure The authors declare no relevant competing interest. S10. Distribution of odds ratio (upper row) and PAF (lower row) according to admission type (columns) for the post hoc definition based on SOFA Author details score. Figure S11. Distribution of odds ratio (upper row) and PAF (lower 1 HCor Research Institute, Rua Desembargador Eliseu Guilherme, 200, 8th Floor, row) according to admission type (columns) after excluding patients with 2 São Paulo, Brazil. Department of Critical Care Medicine, Faculty of Medi‑ infection at baseline. Figure S12. Comparison of the three definitions for cine and Dentistry, University of Alberta, 2‑124E Clinical Sciences Building, elective surgery admissions. Upper panel: The daily odds ratio for mortality 3 8440‑112 St NW, Edmonton, AB T6G2B7, Canada. Intensive Care Unit, Emer‑ in elective surgery patients according to the definition used. Bottom gency Medicine Discipline, Hospital das Clínicas da Faculdade de Medicina panel: The respective AF for each definition. Note that a higher OR does 4 da Universidade de São Paulo, São Paulo, Brazil. Intensive Care Unit, Hospital not equal higher AF since prevalence of events also changes. Figure S13. 5 Sírio‑Libanês, São Paulo, SP, Brazil. Unidade de Terapia Intensiva, Santa Casa de Comparison of the three definitions for emergency surgery admissions. 6 Misericórdia de Porto Alegre, Porto Alegre, RS, Brazil. Department of Critical Upper panel: The daily odds ratio for mortality in medical patients accord‑ 7 Care Medicine, Hospital Israelita Albert Einstein, Sao Paulo, Brazil. Austral‑ ing to the definition used. Bottom panel: The respective AF for each defini‑ ian and New Zealand Intensive Care Research Centre (ANZIC‑RC), School tion. Note that a higher OR does not equal higher AF since prevalence of of Public Health and Preventive Medicine, Monash University, Melbourne, events also changes. 8 Australia. Department of Intensive Care, Austin Hospital, Melbourne, Aus‑ tralia. Intensive Care Unit, AC Camargo Cancer Center, São Paulo, SP, Brazil. Additional file 2: IMPACTO ‑MAPA study. 10 11 Hospital Maternidade São Vicente de Paulo, Barbalha, CE, Brazil. Hospital Federal dos Servidores do Estado, Rio de Janeiro, RJ, Brazil. Hospital da Luz, Acknowledgements São Paulo, SP, Brazil. BP‑A Beneficência Portuguesa de São Paulo, Sao Paulo, 14 15 The authors would like to thank all local data collectors for their work. SP, Brazil. Hospital Paulistano, São Paulo, SP, Brazil. Instituto de Cardiologia Collaborators: Barbara Macedo, Fabio S Coutinho (Hospital São Paulo – do Distrito Federal, Brasilia, DF, Brazil. Hospital Nereu Ramos, Florianópolis, 17 18 UNIFESP); Jussara A Arraes (Hospital Maternidade São Vicente de Paulo); SC, Brazil. Hospital Baía Sul, Florianópolis, SC, Brazil. Hospital das Clínicas Viviane S N Xavier (Hospital Federal dos Servidores do Estado); Eliana V N da Faculdade de Medicina de Ribeirão Preto, Ribeirão Preto, SP, Brazil. Santa Martins (Hospital da Luz); Juliana Chaves Coelho (BP ‑ A Beneficência Portu‑ Casa de Misericórdia Belo Horizonte, Belo Horizonte, MG, Brazil. Hospital guesa de São Paulo); Silvana S Santos (AC Camargo Cancer Center); Andreia São José, Criciúma, SC, Brazil. Hospital Dona Helena, Joinville, SC, Brazil. 22 23 Pardini (Hospital Israelita Albert Einstein); Cassio Luis Zandonai (Hospital Nereu Hospital de Amor‑Fundação PIO XII, Barretos, SP, Brazil. Centro Hospitalar Ramos); Julia B de Carvalho (Hospital das Clínicas da Faculdade de Medicina Unimed, Joinville, SC, Brazil. Hospital Maternidade São José, Colatina, ES, 25 26 de Ribeirão Preto); Isabela O B Louredo (Santa Casa de Misericórdia Belo Brazil. Hospital Unimed Vitória, Vitória, ES, Brazil. Hospital Distrital Evandro Horizonte); Renata C Gonçalves (Hospital São José); Micheli C Arruda (Hospital Ayres de Moura Antônio Bezerra, Fortaleza, CE, Brazil. Hospital e Maternidade Dona Helena); Mariana Regina da Cunha (Hospital de Amor ‑ Fundação PIO Sepaco, Sao Paulo, SP, Brazil. Hospital Japones Santa Cruz, Sao Paulo, SP, Bra‑ 29 30 XII); Mariana Bonomini F de Almeida (Hospital Baía Sul); Juliano Ramos (Centro zil. Hospital da Cidade, Salvador, BA, Brazil. Santa Casa de Misericórdia de Hospitalar Unimed); Bruna M Binda (Hospital Maternidade São José); Priscila São João Del Rei, Belo Horizonte, MG, Brazil. Hospital Ana Nery, Salvador, BA, 32 33 L S Almeida (Hospital Unimed Vitória); Marcia Maria R de Oliveira (Hospital Brazil. Fundação São Francisco de Assis, Belo Horizonte, MG, Brazil. Hospital Distrital Evandro Ayres de Moura Antônio Bezerra); Luciana S de Mattos Regional Dr. Clodolfo Rodrigues de Melo, Maceio, AL, Brazil. Hospital Erasto (Hospital da Cidade); Samara G da Silva (Santa Casa de Misericórdia de São Gaertner, Curitiba, PR, Brazil. Hospital das Clínicas da Universidade Federal de João Del Rei); Daniela C Dorta (Hospital Ana Nery); Martha Hadrich (Santa Goiás, Goiânia, GO, Brazil. Hospital Evangélico de Cachoeiro de Itapemirim, Casa de Misericórdia de Porto Alegre); Fernanda A F Gonçalves (Hospital das Cachoeiro de Itapemirim, ES, Brazil. Hospital Universitário Regional do Norte Clínicas da Universidade Federal de Goiás); ); Kaytiussia R de Sena (Instituto de do Paraná, Londrina, PR, Brazil. Department of Anesthesiology, Pain and Criti‑ Cardiologia do Distrito Federal); Pamella M dos Prazeres (Hospital Evangélico cal Care‑Hospital São Paulo, Escola Paulista de Medicina, Universidade Federal de Cachoeiro de Itapemirim); Josiane Festti (Hospital Universitário Regional do de Sao Paulo, Sao Paulo, SP, Brazil. Norte do Paraná). Received: 9 January 2023 Accepted: 24 March 2023 Author contributions FGZ, ABC, LUT, TCL, AS‑N, LCPA, APNJr, LPD and FRM contributed to the study conception and design. Several authors contributed to data collection (except FGZ, ABC, AS‑N, LCPA, RNS, LPD). FGZ, RHNS and LPD analyzed data. FGZ wrote the first draft of the manuscript and all authors revised it. All authors read and References approved the final manuscript. 1. Rudd KE, Johnson SC, Agesa KM, Shackelford KA, et al. Global, regional, and national sepsis incidence and mortality, 1990–2017: analysis for the Funding Global Burden of Disease Study. Lancet. 2020;395:200–11. The study was funded the Brazilian Ministry of Health through the PROADI‑ 2. Zhao GJ, Li D, Zhao Q, et al. Incidence, risk factors and impact on out‑ SUS (Programa de Desenvolvimento Institucional do Sistema Único de Saúde). comes of secondary infection in patients with septic shock: an 8‑ year retrospective study. Sci Rep. 2016;6:38361. Availability of data and materials 3. Denstaedt SJ, Singer BH, Standiford TJ. Sepsis and nosocomial infection: De‑identified data, data dictionary, and analyses scripts used are available patient characteristics, mechanisms, and modulation. Front Immunol. upon reasonable request and after approval of the steering committee of the 2018;9:2446. proposal and after agreement of Brazilian Ministry of Health for data sharing. 4. Digiovine B, Chenoweth C, Watts C, Higgins M. The attributable mortality and costs of primary nosocomial bloodstream infections in the intensive care unit. Am J Respir Crit Care Med. 1999;160:976–81. Declarations 5. Pittet D, Tarara D, Wenzel RP. Nosocomial bloodstream infection in critically ill patients. Excess length of stay, extra costs, and attributable Ethics approval and consent to participate mortality. JAMA. 1994;271:1598–601. The protocol was approved at the ethics committee of the coordinating site 6. van Vught LA, Klein Klouwenberg PM, Spitoni C, et al. Incidence, risk and at all other including sites (and is available with Additional file 2). Due to factors, and attributable mortality of secondary infections in the intensive the retrospective nature of the analysis, consent was waived. care unit after admission for sepsis. JAMA. 2016;315:1469–79. 7. Machado FR, Cavalcanti AB, Bozza FA, et al. The epidemiology of sepsis in Brazilian intensive care units (the Sepsis PREvalence Assessment Zampier i et al. Annals of Intensive Care (2023) 13:32 Page 13 of 13 Database, SPREAD): an observational study. Lancet Infect Dis. 2017;17:1180–9. 8. Markwart R, Saito H, Harder T, et al. Epidemiology and burden of sepsis acquired in hospitals and intensive care units: a systematic review and meta‑analysis. Intensive Care Med. 2020;46:1536–51. 9. von Cube M, Timsit JF, Schumacher M, Motschall E, Schumacher M. Quantification and interpretation of attributable mortality in core clinical infectious disease journals. Lancet Infect Dis. 2020;20:e299–306. 10. Fleiss JL, Tytun A, Ury HK. A simple approximation for calculating sample sizes for comparing independent proportions. Biometrics. 1980;36:343–6. 11. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of clas‑ sifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40:373–83. 12. Zampieri FG, Iwashyna TJ, Viglianti EM, et al. Association of frailty with short‑term outcomes, organ support and resource use in critically ill patients. Intensive Care Med. 2018;44:1512–20. 13. Labrecque JA, Hunink MMG, Ikram MA, Ikram MK. Do case–control stud‑ ies always estimate odds ratios? Am J Epidemiol. 2021;190:318–21. 14. Zhao W, Chen YQ, Hsu L. On estimation of time‑ dependent attribut‑ able fraction from population‑based case–control studies. Biometrics. 2017;73:866–75. 15. Bruzzi P, Green SB, Byar DP, Brinton LA, Schairer C. Estimating the popula‑ tion attributable risk for multiple risk factors using case–control data. Am J Epidemiol. 1985;122:904–14. 16. Miettinen OS. Proportion of disease caused or prevented by a given exposure, trait or intervention. Am J Epidemiol. 1974;99:325–32. 17. R Core Team (2021). R: a language and environment for statistical com‑ puting. R Foundation for Statistical Computing, Vienna, Austria. https:// www.R‑ proje ct. org/. 18. Van der Wal WM, Geskus RB. IPW: an R package for inverse probability weighting. J Stat Softw. 2011;43:1–3. 19. Vincent JL, Moreno R, Takala J, Willatts S, DeMendonça A, Bruining H, Reinhart CK, Suter PM, Thijs LG. The SOFA (Sepsis‑related Organ Failure Assessment) score to describe organ dysfunction/failure. On behalf of the Working Group on Sepsis‑Related Problems of the European Society of Intensive Care Medicine. Intensive Care Med. 1996;199622(7):707–10. https:// doi. org/ 10. 1007/ BF017 09751. 20. Singer M, Deutschman CS, Seymour CW, et al. The third international consensus definitions for sepsis and septic shock (Sepsis‑3). JAMA. 2016;315:801–10. 21. Pedersen PB, Hrobjartsson A, Nielsen DL, Henriksen DP, Brabrand M, Las‑ sen AT. Prevalence and prognosis of acutely ill patients with organ failure at arrival to hospital: a systematic review. PLoS ONE. 2018;13: e0206610. 22. Donnelly JP, Chen SC, Kauffman CA, et al. Revision and update of the con‑ sensus definitions of invasive fungal disease from the European organiza‑ tion for research and treatment of cancer and the mycoses study group education and research consortium. Clin Infect Dis. 2020;71:1367–76. 23. Tresker S. A typology of clinical conditions. Stud Hist Philos Biol Biomed Sci. 2020;83: 101291. 24. Dahlqwist E, Sjolander A. AF: Model‑Based Estimation of Confounder ‑ Adjusted Attributable Fractions. R package version 0.1.5, 2019. Available on https:// CRAN.R‑ proje ct. org/ packa ge= AF . Accessed 15 July 2022. 25. Ferenci T. Variable selection should be blinded to the outcome. Int J Epidemiol. 2017;46(3):1077–9. https:// doi. org/ 10. 1093/ ije/ dyx048. (PMID: 28402483). 26. Mishra S, Baral SD. Rethinking the population attributable fraction for infectious diseases. Lancet Infect Dis. 2020;20:155–7. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub‑ lished maps and institutional affiliations.

Journal

Annals of Intensive Care – Springer Journals

Published: Apr 26, 2023

Keywords: Sepsis; Attributable mortality; Epidemiology

There are no references for this article.