Name: | Description: | Size: | Format: | |
---|---|---|---|---|
5.69 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
Na indução da anestesia antes de um procedimento cirúrgico, o fármaco anestésico é
administrado sempre com a preocupação de obter uma indução rápida. Idealmente, o
médico anestesista deveria ser capaz de avaliar a dose de fármaco necessária para induzir
inconsciência e parar a administração. No entanto, isto nem sempre acontece e existe
sobredosagem.
Anestesia geral representa uma parte complexa da cirurgia, envolvendo relaxamento
muscular, inconsciência e analgesia. Para o efeito utilizam-se diferentes fármacos, entre os
quais o anestésico propofol. Com perda de consciência a infusão é desligada, porém a
concentração continua a aumentar devido a atrasos farmacocinéticos e farmacodinâmicos
inerentes. Existe uma concentração máxima que se verifica após a infusão ser desligada,
pelo que o interesse é parar a administração de propofol no momento exato e encontrar
formas de minimizar o excesso de anestesia, principalmente na fase de indução, para não
provocar sobredosagem.
Analisamos dados clínicos em 217 neurocirurgias (craniotomias, cirurgias coluna/ medula
espinhal) para verificar quais os efeitos (se existirem) desta sobredosagem nos sinais
fisiológicos integrando uma complexidade de covariáveis confundidoras: dados
antropométricos, estado de saúde, tipo cirurgia, parâmetros basais e diferentes protocolos
de indução. O impacto do fármaco anestésico pode avaliar-se pelo seu efeito nos sinais
cerebrais (eletroencefalograma) ou sinais hemodinâmicos (pressão arterial, frequência
cardíaca, débito cardíaco), no entanto uma medida de sobredosagem é mais difícil prever e
quantificar. Regressão linear múltipla foi implementada e testada objetivando uma melhor
abordagem para este problema. Estudamos duas variáveis dependentes para sobredosagem
de anestesia: (I) diferença entre concentração de propofol máxima e concentração de
propofol em LOC (μg/ml); e (II) quantidade de propofol infundido até LOC (mg/kg).
Obtivemos os R "
# de 71,1% e 47,2%, modelos I e II, permitindo avaliar a variabilidade
observada na sobredosagem de propofol. Avaliamos a capacidade de previsão dos
modelos, resultando valores para a REQM de 0,2855 e 0,3609, modelos I e II.
During anesthesia induction before a surgical procedure, the anesthetic drug is administered to ensure a rapid induction. Ideally, the anesthetist should be able to assess the required drug dose to induce unconsciousness and immediately stop administration. However, this does not always happen, leading to overdosing. General anesthesia is a complex part of surgery, divided into muscle relaxation, unconsciousness, and analgesia. Different anesthetic drugs are used for this purpose, including the anesthetic propofol. With the loss of consciousness, induction is halted, but the concentration continues to increase due to inherent delays in its pharmacokinetics and pharmacodynamics. There is a maximum concentration reached after the infusion is stopped, making it crucial to cease propofol administration at the exact moment and find ways to minimize anesthesia excess, especially during the induction phase, to avoid overdosing. We analyzed clinical data from 217 neurosurgeries (craniotomies, spinal cord surgeries) to verify the potential effects of this overdosing on physiological signals, considering a complexity of confounding covariates: anthropometric data, health status, surgery type, baseline parameters, and different induction protocols. The impact of the anesthetic drug can be evaluated through its effects on brain signals (electroencephalogram) or hemodynamic signals (blood pressure, heart rate, cardiac output). However, measuring overdosing is more difficult to predict and quantify. A multiple linear regression was implemented and tested to understand the best approach for this issue. We studied two dependent variables for anesthesia overdose: (I) the difference between the maximum propofol concentration and the propofol concentration at LOC (μg/ml); and (II) the amount of propofol infused until LOC (mg/kg). We obtained values of 71.1% and 47.2%, models I and II, allowing us to assess the observed variability in propofol overdose. We evaluated the predictive capability of the models, resulting in RMSE values of 0.2855 and 0.3609, models I and II.
During anesthesia induction before a surgical procedure, the anesthetic drug is administered to ensure a rapid induction. Ideally, the anesthetist should be able to assess the required drug dose to induce unconsciousness and immediately stop administration. However, this does not always happen, leading to overdosing. General anesthesia is a complex part of surgery, divided into muscle relaxation, unconsciousness, and analgesia. Different anesthetic drugs are used for this purpose, including the anesthetic propofol. With the loss of consciousness, induction is halted, but the concentration continues to increase due to inherent delays in its pharmacokinetics and pharmacodynamics. There is a maximum concentration reached after the infusion is stopped, making it crucial to cease propofol administration at the exact moment and find ways to minimize anesthesia excess, especially during the induction phase, to avoid overdosing. We analyzed clinical data from 217 neurosurgeries (craniotomies, spinal cord surgeries) to verify the potential effects of this overdosing on physiological signals, considering a complexity of confounding covariates: anthropometric data, health status, surgery type, baseline parameters, and different induction protocols. The impact of the anesthetic drug can be evaluated through its effects on brain signals (electroencephalogram) or hemodynamic signals (blood pressure, heart rate, cardiac output). However, measuring overdosing is more difficult to predict and quantify. A multiple linear regression was implemented and tested to understand the best approach for this issue. We studied two dependent variables for anesthesia overdose: (I) the difference between the maximum propofol concentration and the propofol concentration at LOC (μg/ml); and (II) the amount of propofol infused until LOC (mg/kg). We obtained values of 71.1% and 47.2%, models I and II, allowing us to assess the observed variability in propofol overdose. We evaluated the predictive capability of the models, resulting in RMSE values of 0.2855 and 0.3609, models I and II.
Description
Keywords
Estatística univariada Regressão linear múltipla Ajuste global e qualidade do modelo Anestesia geral Propofol Sobredosagem Univariate statistics Multiple linear regression Model fit and quality General anesthesia Overshot
Citation
Carvalho, Lino Manuel Dias - Efeitos de sobredosagem na indução da anestesia: análise multivariada de dados [Em linha]. [S.l.]: [s.n.], 2023. 216 p.
Carvalho, Lino Manuel Dias - Efeitos de sobredosagem na indução da anestesia: análise multivariada de dados [Em linha]. [S.l.]: [s.n.], 2023. 216 p.
Carvalho, Lino Manuel Dias - Efeitos de sobredosagem na indução da anestesia: análise multivariada de dados [Em linha]. [S.l.]: [s.n.], 2023. 216 p.