Zakażenie clostridioides difficile (c. diff) – Rokowania, prognozy i postęp choroby

Zakażenie clostridioides difficile (c. diff)
Rokowania, prognozy i postęp choroby

Zakażenie Clostridioides difficile (C. diff) stanowi istotny problem kliniczny, szczególnie w środowisku oddziałów intensywnej terapii (OIT), gdzie częstość występowania wynosi około 0,44%. Choroba charakteryzuje się wysoką śmiertelnością, sięgającą 18,33% w populacji ogólnej, a u pacjentów powyżej 60. roku życia 90-dniowa śmiertelność wzrasta do 28%. Kluczowymi czynnikami prognostycznymi są wiek, choroby współistniejące, podwyższone WBC, niskie stężenie albuminy, podwyższone kreatyniny oraz hospitalizacja na OIT. Szczególnie istotny jest poziom zespołu niewydolności narządowej oceniany skalą SOFA, który koreluje z ryzykiem zgonu i powikłań. Dodatkowo, obecność szczepu PCR-ribotypu 027 oraz niskie wartości CT tgNAAT (≤25) wskazują na cięższy przebieg i wyższe ryzyko śmiertelności. Tradycyjne modele prognostyczne, takie jak ATLAS, Zar czy kryteria IDSA/SHEA, wykazują ograniczoną wartość predykcyjną (maksymalna dodatnia wartość predykcyjna 37,9%, AuROC 0,7-0,8), co podkreśla potrzebę bardziej zaawansowanych narzędzi.

Wprowadzenie do zakażenia Clostridioides difficile (C. diff)
Wskaźniki śmiertelności w zakażeniach C. diff
Czynniki prognostyczne w zakażeniu C. diff

Czynniki demograficzne i kliniczne
Wielowymiarowa ocena zespołu wątłości
Czynniki genetyczne i laboratoryjne

Modele predykcyjne dla zakażeń C. diff

Tradycyjne modele kliniczne
Modele oparte na uczeniu maszynowym

Szczególne sytuacje prognostyczne

Bezobjawowa kolonizacja
Niepowodzenie leczenia
Nawroty zakażenia
Zaburzenia autoimmunologiczne

Wnioski i implikacje kliniczne

Kolejne rozdziały

Wprowadzenie do zakażenia Clostridioides difficile (C. diff)

Zakażenie Clostridioides difficile (C. diff) to jedna z głównych przyczyn biegunki związanej z opieką zdrowotną, charakteryzująca się znaczną śmiertelnością. Zakażenia C. difficile w ostatnich latach wykazują tendencję wzrostową zarówno pod względem liczby, jak i ciężkości przypadków we wszystkich placówkach medycznych, w tym również na oddziałach intensywnej terapii (OIT). Obecna częstość występowania zakażeń C. diff wśród pacjentów OIT szacowana jest na około 0,44%, co ma poważny wpływ na zachorowalność i śmiertelność.¹ Większość zakażeń C. diff ma przebieg łagodny do umiarkowanego i ustępuje po leczeniu, jednak w niektórych przypadkach choroba może rozwijać się gwałtownie i prowadzić do poważnych powikłań.²³

Wskaźniki śmiertelności w zakażeniach C. diff

Zakażenie Clostridioides difficile wiąże się z istotnym ryzykiem śmiertelności, które różni się w zależności od populacji pacjentów i ciężkości choroby. Badania pokazują, że wewnątrzszpitalna śmiertelność związana z zakażeniem C. diff w ogólnej populacji chorych wynosi około 18,33%.⁴ W przypadku pacjentów w wieku powyżej 60 lat, śmiertelność 90-dniowa osiąga nawet 28%, co wskazuje na szczególne ryzyko u osób starszych.⁵

Natomiast w kontekście pacjentów na oddziałach intensywnej terapii, pomimo wyższej surowej śmiertelności, po starannym dostosowaniu do czynników zakłócających, zakażenie C. diff nie jest niezależnie związane ze zwiększoną śmiertelnością i ma jedynie marginalny wpływ na długość pobytu w OIT, jeśli jest wcześnie leczone.⁶

Czynniki prognostyczne w zakażeniu C. diff

Czynniki demograficzne i kliniczne

Liczne badania identyfikują czynniki, które mają istotny wpływ na rokowanie u pacjentów z zakażeniem C. diff:

Wiek – pacjenci w podeszłym wieku mają gorsze rokowanie
Choroby współistniejące – wcześniejsze schorzenia wpływają znacząco na przebieg zakażenia
Liczba białych krwinek – podwyższona wartość WBC jest wskaźnikiem cięższego przebiegu
Stężenie albuminy w surowicy – niskie poziomy albuminy korelują z gorszym rokowaniem
Stężenie kreatyniny w surowicy – podwyższone wartości świadczą o cięższym przebiegu
Przyjęcie na OIT – hospitalizacja na oddziale intensywnej terapii jest niezależnym czynnikiem ryzyka

⁷

Szczególnie istotnym predyktorem śmiertelności jest poziom zespołu niewydolności narządowej, mierzony skalą SOFA (Sequential Organ Failure Assessment). Pacjenci z wyższym wynikiem SOFA w momencie diagnozy mają wyższe ryzyko śmiertelności na OIT lub poważnych powikłań.⁸

Wielowymiarowa ocena zespołu wątłości

Badania wskazują, że wielowymiarowa ocena zespołu wątłości jest najlepszym predyktorem 90-dniowej śmiertelności u starszych pacjentów z zakażeniem C. diff, przewyższając wiek i markery ciężkości choroby. Wśród pacjentów z ciężkim zespołem wątłości aż 51% umiera w ciągu 90 dni. Poziom wątłości przewyższa wiek i ciężkość choroby w prognozowaniu śmiertelności 90-dniowej.⁹¹⁰

Czynniki genetyczne i laboratoryjne

Istotne znaczenie prognostyczne ma również identyfikacja konkretnych szczepów C. difficile. Wykrycie szczepu PCR-ribotypu 027 jest związane z większym ryzykiem śmiertelności. Badania wykazały, że relatywne ryzyko śmiertelności u pacjentów z niskim CT (≤25) w porównaniu do CT>25 wynosi 1,45, a wzrasta do 2,18 u pacjentów z niskim CT i PCR-ribotypem 027.¹¹

Mediana wartości CT tgNAAT (testu amplifikacji kwasów nukleinowych) dla pacjentów, którzy zmarli, była znacząco niższa niż dla tych, którzy przeżyli (25,5 vs 27,5, p = 0,021). Niskie wartości CT tgNAAT (≤25) wskazują na pacjentów z cięższym przebiegiem zakażenia, wyższym ryzykiem śmiertelności i potencjalnie nawrotu choroby.¹²

Modele predykcyjne dla zakażeń C. diff

Tradycyjne modele kliniczne

Pomimo prób stworzenia uniwersalnie akceptowanych klasyfikacji ciężkości choroby, znalezienie zestawu parametrów klinicznych, które mogą poprawnie przewidzieć przebieg i rokowanie zakażenia C. diff u pacjentów w różnych środowiskach klinicznych, pozostaje wyzwaniem.¹³ Istnieje kilka modeli prognostycznych:

ATLAS – model kliniczny oceniający ryzyko powikłań zakażenia C. diff
Skala Zar – narzędzie do przewidywania ciężkości zakażenia
Kryteria ciężkości IDSA/SHEA – wykorzystujące głównie WBC i kreatyniny

¹⁴

Ogólna wartość predykcyjna dla wszystkich tradycyjnych modeli jest jednak niska, z maksymalną dodatnią wartością predykcyjną 37,9% (dla punktu odcięcia ATLAS 9). Chociaż ATLAS, Zar i kryteria ciężkości IDSA wszystkie mieszczą się w zakresie AuROC 0,7-0,8, co mogłoby być uznane za klinicznie użyteczne, AUC może być mylące w przypadkach takich jak zakażenie C. diff, gdzie wyniki są niezrównoważone.¹⁵

ATLAS wykazał znacznie gorsze wyniki (AuROC 0,781 versus 0,718) w analizie post hoc w przewidywaniu ciężkich wyników niezwiązanych z zakażeniem C. diff. Definicja ciężkiego zakażenia według wytycznych IDSA/SHEA wykorzystująca wyłącznie WBC i kreatyniny historycznie była słabym predyktorem wyników.¹⁶

Modele oparte na uczeniu maszynowym

Nowsze podejścia wykorzystują modele uczenia maszynowego (ML) do prognozowania śmiertelności wewnątrzszpitalnej pacjentów z zakażeniem C. diff. Badania wskazują, że modele te przewyższają istniejące skale ciężkości w przewidywaniu wyników śmiertelności.¹⁷

Wszystkie modele ML wykazały odpowiednią dyskryminację (tj. AUROC między 0,69 a 0,72) w przewidywaniu śmiertelności pacjentów. Śmiertelność w grupie wysokiego ryzyka zidentyfikowanej przez modele ML wynosi 57,14%, podczas gdy wskaźnik śmiertelności w grupie niskiego ryzyka wynosi 18,87%.¹⁸¹⁹

Modele ML mogą uwzględniać zmienność w danych laboratoryjnych i wielu chorobach współistniejących w prognozowaniu, czego inne standardowe narzędzia prognostyczne nie są w stanie wykonać. Głównym celem proponowanych modeli ML jest prognozowanie pacjentów z zakażeniem C. diff i wczesne wychwycenie tych, których stan prawdopodobnie się pogorszy.²⁰

Szczególne sytuacje prognostyczne

Bezobjawowa kolonizacja

Około 10 do 20% pacjentów wykazuje bezobjawową kolonizację C. difficile bez objawów choroby. Konsekwencje prognostyczne dla bezobjawowego nosiciela nie są jasne, ale należy zachować czujność ze względu na potencjalne ryzyko rozwoju objawowej choroby.²¹

Niepowodzenie leczenia

Wczesne rozpoznanie niepowodzenia leczenia pozostaje nierozwiązanym problemem klinicznym. W przypadku niepowodzenia leczenia, alternatywne metody obejmują zastąpienie wankomycyny fidaksomycyną, tygecykliną, kombinacją dożylnego metronidazolu i wankomycyny, immunoglobulinami oraz przeszczepieniem mikrobioty kałowej (FMT).²²

Niepowodzenie leczenia jest szczególnie częste u pacjentów OIT ze względu na choroby współistniejące i konieczność kontynuacji leczenia antybiotykami.²³

Nawroty zakażenia

Najczęstszym długoterminowym problemem jest trwające lub nawracające zakażenie C. difficile. Występuje ono, gdy jelito grube ma trudności z całkowitym powrotem do zdrowia. Jelito grube może wolniej dochodzić do zdrowia, jeśli pacjent ma przewlekłe schorzenia lub inne czynniki ryzyka, które czynią go bardziej podatnym na zakażenie C. diff. W takich przypadkach pacjent może być bardziej narażony na cięższe zakażenie lub nawroty wymagające bardziej rozszerzonego leczenia.²⁴

Zaburzenia autoimmunologiczne

Rzadziej niektóre osoby rozwijają zaburzenia autoimmunologiczne po ciężkim zakażeniu. Oznacza to, że ich układ odpornościowy nadal działa tak, jakby miał zakażenie, nawet gdy już go nie ma.²⁵

Wnioski i implikacje kliniczne

Szacowanie ciężkości zakażenia C. diff jest niezbędne dla określenia rokowania i właściwej terapii. Diagnoza i ocena ciężkości oraz progresji choroby są jeszcze bardziej skomplikowane w warunkach OIT i powinny być wspomagane przez kliniczne narzędzia predykcyjne (np. skala ATLAS).²⁶

Aktualne algorytmy diagnostyczne mogą prowadzić do niedoszacowania ciężkości zakażenia C. diff u pacjentów OIT. Działania zapobiegawcze i świadomość czynników ryzyka powinny być priorytetem na każdym OIT. Zespół kliniczny powinien być świadomy indywidualnego profilu ryzyka każdego pacjenta dotyczącego rozwoju zakażenia C. diff podczas pobytu na OIT.²⁷²⁸

Modele uczenia maszynowego wykorzystujące duże zbiory danych stanowią obiecujące podejście do identyfikacji pacjentów wysokiego ryzyka z zakażeniem C. diff. Proponowane modele ML mogą ułatwić wczesne rozpoznanie ciężkości zakażenia C. diff i umożliwić szybką interwencję u pacjentów potrzebujących.²⁹

W przyszłości, aby wzmocnić kolejne iteracje ATLAS lub innych modeli wyłącznie klinicznych, logicznym kolejnym krokiem byłaby ocena nowych biomarkerów, zarówno czynników gospodarza, jak i patogenu.³⁰

Kolejne rozdziały

Zapraszamy do dalszego czytania naszego leksykonu.

Wybierz kolejny rozdział z menu poniżej, aby otworzyć nową podstronę kompedium wiedzy i uzyskać szczegółowe informację o leku, substancji lub chorobie.

Materiały źródłowe

#1 Sleeping with the enemy: Clostridium difficile infection in the intensive care unit | Critical Care | Full Text
https://ccforum.biomedcentral.com/articles/10.1186/s13054-017-1819-6
Over the last years, there was an increase in the number and severity of Clostridium difficile infections (CDI) in all medical settings, including the intensive care unit (ICU). The current prevalence of CDI among ICU patients is estimated at 0.44% and has severe impact on morbidity and mortality. […] Most patients developing CDI in the ICU show a mild to moderate disease course. Nevertheless, difficult-to-treat severe and complicated cases also occur. Treatment failure is particularly frequent in ICU patients due to comorbidities and the necessity of continued antibiotic treatment. […] The main difficulty in finding a universally accepted classification for disease severity consists of determining a set of clinical parameters which can correctly predict the course and prognosis of CDI for patients in different clinical settings. A number of studies have attempted to identify factors that can reliably predict unfavorable outcomes.
#2 C. diff Infection: What It Is, Symptoms & Treatment
https://my.clevelandclinic.org/health/diseases/15548-c-diff-infection
Most C. diff infections are mild and go away with treatment. But the circumstances that cause C. diff infection sometimes allow it to spread very quickly. C. diff infection can be sudden and severe. […] If you have risk factors that make you more vulnerable to C. diff infection, you may be more likely to have a more severe infection or have repeat infections and need more extensive treatment. […] The most common long-term problem is ongoing or repeat infection with C. difficile. This happens when your colon is having trouble recovering completely. Your colon may be slower to recover if: […] More rarely, some people develop autoimmune disorders after a severe infection. This means that their immune systems continue to act as though they have an infection even when they dont anymore.
#3 Sleeping with the enemy: Clostridium difficile infection in the intensive care unit | Critical Care | Full Text
https://ccforum.biomedcentral.com/articles/10.1186/s13054-017-1819-6
Over the last years, there was an increase in the number and severity of Clostridium difficile infections (CDI) in all medical settings, including the intensive care unit (ICU). The current prevalence of CDI among ICU patients is estimated at 0.44% and has severe impact on morbidity and mortality. […] Most patients developing CDI in the ICU show a mild to moderate disease course. Nevertheless, difficult-to-treat severe and complicated cases also occur. Treatment failure is particularly frequent in ICU patients due to comorbidities and the necessity of continued antibiotic treatment. […] The main difficulty in finding a universally accepted classification for disease severity consists of determining a set of clinical parameters which can correctly predict the course and prognosis of CDI for patients in different clinical settings. A number of studies have attempted to identify factors that can reliably predict unfavorable outcomes.
#4 Prediction of in-hospital mortality of Clostriodiodes difficile infection using critical care database: a big data-driven, machine learning approach
https://pmc.ncbi.nlm.nih.gov/articles/PMC8601086/
Clostriodiodes difficile infection (CDI) is a major cause of healthcare-associated diarrhoea with high mortality. […] The mortality rate for CDI in the study cohort was 18.33%. […] Our machine learning derived CDI in-hospital mortality prediction model identified pertinent factors that can assist critical care clinicians in identifying patients at high risk of dying from CDI. […] Machine learning models are developed to predict in-hospital mortality of patients with CDI. […] The proposed machine learning models outperformed existing severity scores in predicting mortality outcomes. […] The proposed models can facilitate early recognition of CDI severity and enable timely intervention to patients in need. […] All ML models had adequate discrimination (ie, AUROC between 0.69 and 0.72) in predicting patient mortality.
#5
https://link.springer.com/article/10.1007/s41999-023-00772-3
Older patients with their first CDI had a high 90-day mortality of 28%. […] Compared with age and CDI severity, the multidimensional frailty assessment is the best predictor of 90-day mortality in older patients with CDI. […] The 90-day mortality among older patients with CDI in a Danish region is 28%. Frailty measured by record-based MPI at discharge outperforms age and disease severity markers in predicting mortality in older patients with CDI. […] The key finding of this population-based study was a 90-day mortality of 28% in patients with CDI and older than 60 years. Among patients with severe frailty 51% died before 90 days. Both patient age, disease severity, and frailty level at discharge predicted mortality. Frailty level outperformed age and severity in predicting mortality at 90 days.
#6 Outcome of ICU patients with Clostridium difficile infection | Critical Care | Full Text
https://ccforum.biomedcentral.com/articles/10.1186/cc11852
If treated early, ICU-acquired CDI is not independently associated with an increased mortality and impacts marginally the ICU length of stay. […] The impact of CDI on mortality was homogeneous across centers. ICU death in patients with CDI infection was associated with a high LOD score (P = 0.01), a high McCabe score (P = 0.02), and with immunosuppression (P = 0.02). […] Despite a significantly higher crude mortality, when using modern statistical models, CDI was not associated with increased mortality, regardless of the control groups, and after careful adjustment on confounding factors of mortality and on other adverse events and nosocomial infections associated with mortality. […] After careful adjustment for confounding variables, CDI is not associated with significant attributable mortality and extra length of stay.
#7 Prediction of in-hospital mortality of Clostriodiodes difficile infection using critical care database: a big data-driven, machine learning approach
https://pmc.ncbi.nlm.nih.gov/articles/PMC8601086/
The mortality of high-risk group is 57.14% while the mortality rate of low-risk group is 18.87%. […] Our ML models could incorporate variability in laboratory data and many comorbidities into prediction, which other standard prognostic tools are unable to perform. […] The finding of serum albumin being a predictor of mortality is concordant with our previous systematic review, which showed that at least five of the 31 articles identified prior comorbidities, age, white blood cell count, serum albumin, serum creatinine and ICU admission as predictors of severity. […] The ultimate aim of our proposed ML model is to prognosticate patients with CDI and to catch those whose conditions are likely to worsen early on. […] In conclusion, by learning from the shortcomings of previous severity models, we have employed a robust and objective ML approach, while capitalising on one of the most extensive ICU databases to develop a CDI severity prediction model.
#8 Sleeping with the enemy: Clostridium difficile infection in the intensive care unit | Critical Care | Full Text
https://ccforum.biomedcentral.com/articles/10.1186/s13054-017-1819-6
The aforementioned report by Bouza et al. confirms that it is reasonable to expect a more severe or complicated course of CDI if the patient has been transferred to the ICU after the initial diagnosis. […] There is evidence that patients with a higher Sequential Organ Failure Assessment (SOFA) score at the time of diagnosis also have a higher risk of ICU mortality or severe complications. […] In conclusion, out of the 2% of ICU patients with CDI, a significant number of cases can be classified mild or moderate. […] Estimating the severity of CDI is essential for prognosis and therapy. Diagnosis and estimation of disease severity and progression are even more complicated in the ICU setting and should be assisted by clinical prediction tools (i.e., ATLAS score). Current diagnostic algorithms may lead to an underestimation of CDI severity in ICU patients.
#9
https://link.springer.com/article/10.1007/s41999-023-00772-3
Older patients with their first CDI had a high 90-day mortality of 28%. […] Compared with age and CDI severity, the multidimensional frailty assessment is the best predictor of 90-day mortality in older patients with CDI. […] The 90-day mortality among older patients with CDI in a Danish region is 28%. Frailty measured by record-based MPI at discharge outperforms age and disease severity markers in predicting mortality in older patients with CDI. […] The key finding of this population-based study was a 90-day mortality of 28% in patients with CDI and older than 60 years. Among patients with severe frailty 51% died before 90 days. Both patient age, disease severity, and frailty level at discharge predicted mortality. Frailty level outperformed age and severity in predicting mortality at 90 days.
#10
https://link.springer.com/article/10.1007/s41999-023-00772-3
The higher predictive value of 90-day mortality for multidimensional frailty when compared to CDI severity may be because the components of the record-based MPI and CDI severity classification differ. […] Disease severity, although outperformed by frailty level, also predicted mortality at 90 days.
#11 The predictive value of quantitative nucleic acid amplification detection of Clostridium difficile toxin gene for faecal sample toxin status and patient outcome | PLOS One
https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0205941
Laboratory diagnosis of Clostridium difficile infection (CDI) remains unsettled, despite updated guidelines. […] We have shown that a low tgNAAT CT value (25) is significantly associated with a toxin positive status, presence of PCR-ribotype 027, and mortality. […] The median tgNAAT CT value for patients who died was significantly lower than for those who survived (25.5 vs 27.5 p = 0.021) with a significant AUROCC for NAAT positives and mortality; 0.572 p = 0.009. […] The relative risk of mortality in patients with a low CT (25) vs a CT25 was 1.45 which increased to 2.18 in those patients who had a low CT and had PCR-ribotype 027. […] Low tgNAAT CT values (25) indicate patients with more severe infection and at higher risk of mortality and possibly recurrence.
#12 The predictive value of quantitative nucleic acid amplification detection of Clostridium difficile toxin gene for faecal sample toxin status and patient outcome | PLOS One
https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0205941
Laboratory diagnosis of Clostridium difficile infection (CDI) remains unsettled, despite updated guidelines. […] We have shown that a low tgNAAT CT value (25) is significantly associated with a toxin positive status, presence of PCR-ribotype 027, and mortality. […] The median tgNAAT CT value for patients who died was significantly lower than for those who survived (25.5 vs 27.5 p = 0.021) with a significant AUROCC for NAAT positives and mortality; 0.572 p = 0.009. […] The relative risk of mortality in patients with a low CT (25) vs a CT25 was 1.45 which increased to 2.18 in those patients who had a low CT and had PCR-ribotype 027. […] Low tgNAAT CT values (25) indicate patients with more severe infection and at higher risk of mortality and possibly recurrence.
#13 Sleeping with the enemy: Clostridium difficile infection in the intensive care unit | Critical Care | Full Text
https://ccforum.biomedcentral.com/articles/10.1186/s13054-017-1819-6
Over the last years, there was an increase in the number and severity of Clostridium difficile infections (CDI) in all medical settings, including the intensive care unit (ICU). The current prevalence of CDI among ICU patients is estimated at 0.44% and has severe impact on morbidity and mortality. […] Most patients developing CDI in the ICU show a mild to moderate disease course. Nevertheless, difficult-to-treat severe and complicated cases also occur. Treatment failure is particularly frequent in ICU patients due to comorbidities and the necessity of continued antibiotic treatment. […] The main difficulty in finding a universally accepted classification for disease severity consists of determining a set of clinical parameters which can correctly predict the course and prognosis of CDI for patients in different clinical settings. A number of studies have attempted to identify factors that can reliably predict unfavorable outcomes.
#14 Validation of Clinical Risk Models for Clostridioides difficile-Attributable Outcomes
https://pmc.ncbi.nlm.nih.gov/articles/PMC9295569/
Clostridioides difficile is the leading health care-associated pathogen, leading to substantial morbidity and mortality; however, there is no widely accepted model to predict C. difficile infection severity. […] The overall predictive value for all models was low, with a maximum positive predictive value of 37.9% (ATLAS cutoff 9). […] No clinical model performed well on external validation, but ATLAS did outperform other models for predicting clinically relevant C. difficile-attributable outcomes at diagnosis. […] While ATLAS, Zar, and the IDSA Severity criteria all fell within an AuROC range 0.7-0.8 that could be considered clinically useful, AUC may be misleading in settings such as CDI where outcomes are unbalanced. […] Therefore, the clinical utility of ATLAS to predict attributable outcomes of infection remains unclear.
#15 Validation of Clinical Risk Models for Clostridioides difficile-Attributable Outcomes
https://pmc.ncbi.nlm.nih.gov/articles/PMC9295569/
Clostridioides difficile is the leading health care-associated pathogen, leading to substantial morbidity and mortality; however, there is no widely accepted model to predict C. difficile infection severity. […] The overall predictive value for all models was low, with a maximum positive predictive value of 37.9% (ATLAS cutoff 9). […] No clinical model performed well on external validation, but ATLAS did outperform other models for predicting clinically relevant C. difficile-attributable outcomes at diagnosis. […] While ATLAS, Zar, and the IDSA Severity criteria all fell within an AuROC range 0.7-0.8 that could be considered clinically useful, AUC may be misleading in settings such as CDI where outcomes are unbalanced. […] Therefore, the clinical utility of ATLAS to predict attributable outcomes of infection remains unclear.
#16 Validation of Clinical Risk Models for Clostridioides difficile-Attributable Outcomes
https://pmc.ncbi.nlm.nih.gov/articles/PMC9295569/
ATLAS performed significantly worse (AuROC 0.781 versus 0.718) in the post hoc analysis to predict severe outcomes not attributable to CDI. […] The IDSA/SHEA Guideline definition for severe infection using WBC and creatinine alone has historically been a poor predictor of outcomes. […] Our data showed that PCR cycle threshold was poorly predictive for severe outcomes, which is in keeping with recent work demonstrating that the immune response, not bacterial burden, mediates severity. […] To augment future iterations of ATLAS or other clinical-only models, evaluating novel biomarkers, either host or pathogen factors, would be a logical next step.
#17 Prediction of in-hospital mortality of Clostriodiodes difficile infection using critical care database: a big data-driven, machine learning approach
https://pmc.ncbi.nlm.nih.gov/articles/PMC8601086/
Clostriodiodes difficile infection (CDI) is a major cause of healthcare-associated diarrhoea with high mortality. […] The mortality rate for CDI in the study cohort was 18.33%. […] Our machine learning derived CDI in-hospital mortality prediction model identified pertinent factors that can assist critical care clinicians in identifying patients at high risk of dying from CDI. […] Machine learning models are developed to predict in-hospital mortality of patients with CDI. […] The proposed machine learning models outperformed existing severity scores in predicting mortality outcomes. […] The proposed models can facilitate early recognition of CDI severity and enable timely intervention to patients in need. […] All ML models had adequate discrimination (ie, AUROC between 0.69 and 0.72) in predicting patient mortality.
#18 Prediction of in-hospital mortality of Clostriodiodes difficile infection using critical care database: a big data-driven, machine learning approach
https://pmc.ncbi.nlm.nih.gov/articles/PMC8601086/
Clostriodiodes difficile infection (CDI) is a major cause of healthcare-associated diarrhoea with high mortality. […] The mortality rate for CDI in the study cohort was 18.33%. […] Our machine learning derived CDI in-hospital mortality prediction model identified pertinent factors that can assist critical care clinicians in identifying patients at high risk of dying from CDI. […] Machine learning models are developed to predict in-hospital mortality of patients with CDI. […] The proposed machine learning models outperformed existing severity scores in predicting mortality outcomes. […] The proposed models can facilitate early recognition of CDI severity and enable timely intervention to patients in need. […] All ML models had adequate discrimination (ie, AUROC between 0.69 and 0.72) in predicting patient mortality.
#19 Prediction of in-hospital mortality of Clostriodiodes difficile infection using critical care database: a big data-driven, machine learning approach
https://pmc.ncbi.nlm.nih.gov/articles/PMC8601086/
The mortality of high-risk group is 57.14% while the mortality rate of low-risk group is 18.87%. […] Our ML models could incorporate variability in laboratory data and many comorbidities into prediction, which other standard prognostic tools are unable to perform. […] The finding of serum albumin being a predictor of mortality is concordant with our previous systematic review, which showed that at least five of the 31 articles identified prior comorbidities, age, white blood cell count, serum albumin, serum creatinine and ICU admission as predictors of severity. […] The ultimate aim of our proposed ML model is to prognosticate patients with CDI and to catch those whose conditions are likely to worsen early on. […] In conclusion, by learning from the shortcomings of previous severity models, we have employed a robust and objective ML approach, while capitalising on one of the most extensive ICU databases to develop a CDI severity prediction model.
#20 Prediction of in-hospital mortality of Clostriodiodes difficile infection using critical care database: a big data-driven, machine learning approach
https://pmc.ncbi.nlm.nih.gov/articles/PMC8601086/
The mortality of high-risk group is 57.14% while the mortality rate of low-risk group is 18.87%. […] Our ML models could incorporate variability in laboratory data and many comorbidities into prediction, which other standard prognostic tools are unable to perform. […] The finding of serum albumin being a predictor of mortality is concordant with our previous systematic review, which showed that at least five of the 31 articles identified prior comorbidities, age, white blood cell count, serum albumin, serum creatinine and ICU admission as predictors of severity. […] The ultimate aim of our proposed ML model is to prognosticate patients with CDI and to catch those whose conditions are likely to worsen early on. […] In conclusion, by learning from the shortcomings of previous severity models, we have employed a robust and objective ML approach, while capitalising on one of the most extensive ICU databases to develop a CDI severity prediction model.
#21 Sleeping with the enemy: Clostridium difficile infection in the intensive care unit | Critical Care | Full Text
https://ccforum.biomedcentral.com/articles/10.1186/s13054-017-1819-6
10 to 20% of patients show an asymptomatic colonization with C. difficile without disease symptoms. The prognostic consequence for the asymptomatic carrier is not clear. […] Early recognition of treatment failure is still an unresolved clinical problem. In the case of treatment failure, alternative treatments include substituting vancomycin with fidaxomycin, tigecycline, a combination of intravenous metronidazole and vancomycin, immunoglobulins, and FMT. […] Preventative measures and an acute awareness of risk factors should be a priority in every ICU. The clinical team should be aware of the individual risk profile of each patient for developing CDI while in the ICU.
#22 Sleeping with the enemy: Clostridium difficile infection in the intensive care unit | Critical Care | Full Text
https://ccforum.biomedcentral.com/articles/10.1186/s13054-017-1819-6
10 to 20% of patients show an asymptomatic colonization with C. difficile without disease symptoms. The prognostic consequence for the asymptomatic carrier is not clear. […] Early recognition of treatment failure is still an unresolved clinical problem. In the case of treatment failure, alternative treatments include substituting vancomycin with fidaxomycin, tigecycline, a combination of intravenous metronidazole and vancomycin, immunoglobulins, and FMT. […] Preventative measures and an acute awareness of risk factors should be a priority in every ICU. The clinical team should be aware of the individual risk profile of each patient for developing CDI while in the ICU.
#23 Sleeping with the enemy: Clostridium difficile infection in the intensive care unit | Critical Care | Full Text
https://ccforum.biomedcentral.com/articles/10.1186/s13054-017-1819-6
Over the last years, there was an increase in the number and severity of Clostridium difficile infections (CDI) in all medical settings, including the intensive care unit (ICU). The current prevalence of CDI among ICU patients is estimated at 0.44% and has severe impact on morbidity and mortality. […] Most patients developing CDI in the ICU show a mild to moderate disease course. Nevertheless, difficult-to-treat severe and complicated cases also occur. Treatment failure is particularly frequent in ICU patients due to comorbidities and the necessity of continued antibiotic treatment. […] The main difficulty in finding a universally accepted classification for disease severity consists of determining a set of clinical parameters which can correctly predict the course and prognosis of CDI for patients in different clinical settings. A number of studies have attempted to identify factors that can reliably predict unfavorable outcomes.
#24 C. diff Infection: What It Is, Symptoms & Treatment
https://my.clevelandclinic.org/health/diseases/15548-c-diff-infection
Most C. diff infections are mild and go away with treatment. But the circumstances that cause C. diff infection sometimes allow it to spread very quickly. C. diff infection can be sudden and severe. […] If you have risk factors that make you more vulnerable to C. diff infection, you may be more likely to have a more severe infection or have repeat infections and need more extensive treatment. […] The most common long-term problem is ongoing or repeat infection with C. difficile. This happens when your colon is having trouble recovering completely. Your colon may be slower to recover if: […] More rarely, some people develop autoimmune disorders after a severe infection. This means that their immune systems continue to act as though they have an infection even when they dont anymore.
#25 C. diff Infection: What It Is, Symptoms & Treatment
https://my.clevelandclinic.org/health/diseases/15548-c-diff-infection
Most C. diff infections are mild and go away with treatment. But the circumstances that cause C. diff infection sometimes allow it to spread very quickly. C. diff infection can be sudden and severe. […] If you have risk factors that make you more vulnerable to C. diff infection, you may be more likely to have a more severe infection or have repeat infections and need more extensive treatment. […] The most common long-term problem is ongoing or repeat infection with C. difficile. This happens when your colon is having trouble recovering completely. Your colon may be slower to recover if: […] More rarely, some people develop autoimmune disorders after a severe infection. This means that their immune systems continue to act as though they have an infection even when they dont anymore.
#26 Sleeping with the enemy: Clostridium difficile infection in the intensive care unit | Critical Care | Full Text
https://ccforum.biomedcentral.com/articles/10.1186/s13054-017-1819-6
The aforementioned report by Bouza et al. confirms that it is reasonable to expect a more severe or complicated course of CDI if the patient has been transferred to the ICU after the initial diagnosis. […] There is evidence that patients with a higher Sequential Organ Failure Assessment (SOFA) score at the time of diagnosis also have a higher risk of ICU mortality or severe complications. […] In conclusion, out of the 2% of ICU patients with CDI, a significant number of cases can be classified mild or moderate. […] Estimating the severity of CDI is essential for prognosis and therapy. Diagnosis and estimation of disease severity and progression are even more complicated in the ICU setting and should be assisted by clinical prediction tools (i.e., ATLAS score). Current diagnostic algorithms may lead to an underestimation of CDI severity in ICU patients.
#27 Sleeping with the enemy: Clostridium difficile infection in the intensive care unit | Critical Care | Full Text
https://ccforum.biomedcentral.com/articles/10.1186/s13054-017-1819-6
The aforementioned report by Bouza et al. confirms that it is reasonable to expect a more severe or complicated course of CDI if the patient has been transferred to the ICU after the initial diagnosis. […] There is evidence that patients with a higher Sequential Organ Failure Assessment (SOFA) score at the time of diagnosis also have a higher risk of ICU mortality or severe complications. […] In conclusion, out of the 2% of ICU patients with CDI, a significant number of cases can be classified mild or moderate. […] Estimating the severity of CDI is essential for prognosis and therapy. Diagnosis and estimation of disease severity and progression are even more complicated in the ICU setting and should be assisted by clinical prediction tools (i.e., ATLAS score). Current diagnostic algorithms may lead to an underestimation of CDI severity in ICU patients.
#28 Sleeping with the enemy: Clostridium difficile infection in the intensive care unit | Critical Care | Full Text
https://ccforum.biomedcentral.com/articles/10.1186/s13054-017-1819-6
10 to 20% of patients show an asymptomatic colonization with C. difficile without disease symptoms. The prognostic consequence for the asymptomatic carrier is not clear. […] Early recognition of treatment failure is still an unresolved clinical problem. In the case of treatment failure, alternative treatments include substituting vancomycin with fidaxomycin, tigecycline, a combination of intravenous metronidazole and vancomycin, immunoglobulins, and FMT. […] Preventative measures and an acute awareness of risk factors should be a priority in every ICU. The clinical team should be aware of the individual risk profile of each patient for developing CDI while in the ICU.
#29 Prediction of in-hospital mortality of Clostriodiodes difficile infection using critical care database: a big data-driven, machine learning approach
https://pmc.ncbi.nlm.nih.gov/articles/PMC8601086/
Clostriodiodes difficile infection (CDI) is a major cause of healthcare-associated diarrhoea with high mortality. […] The mortality rate for CDI in the study cohort was 18.33%. […] Our machine learning derived CDI in-hospital mortality prediction model identified pertinent factors that can assist critical care clinicians in identifying patients at high risk of dying from CDI. […] Machine learning models are developed to predict in-hospital mortality of patients with CDI. […] The proposed machine learning models outperformed existing severity scores in predicting mortality outcomes. […] The proposed models can facilitate early recognition of CDI severity and enable timely intervention to patients in need. […] All ML models had adequate discrimination (ie, AUROC between 0.69 and 0.72) in predicting patient mortality.
#30 Validation of Clinical Risk Models for Clostridioides difficile-Attributable Outcomes
https://pmc.ncbi.nlm.nih.gov/articles/PMC9295569/
ATLAS performed significantly worse (AuROC 0.781 versus 0.718) in the post hoc analysis to predict severe outcomes not attributable to CDI. […] The IDSA/SHEA Guideline definition for severe infection using WBC and creatinine alone has historically been a poor predictor of outcomes. […] Our data showed that PCR cycle threshold was poorly predictive for severe outcomes, which is in keeping with recent work demonstrating that the immune response, not bacterial burden, mediates severity. […] To augment future iterations of ATLAS or other clinical-only models, evaluating novel biomarkers, either host or pathogen factors, would be a logical next step.