Yoğun bakım hastalarında, yatış günü platelet, CRP ve albümin düzeyleri ile sağkalım ilişkisinin araştırılması
Küçük Resim Yok
Tarih
2018
Yazarlar
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
Trakya Üniversitesi
Erişim Hakkı
info:eu-repo/semantics/openAccess
Özet
Yoğun bakım hastalarında mortaliteyi tahmin edecek parametreler tedavinin gerekliliğini, yerini, hızını belirlediğinden önemlidirler. Kullanılan skorlama sistemleri karmaşık olması nedeniyle pratik kullanımları sınırlıdır. Biz çalışmamızda CRP, Alb, PLT ve yaş parametrelerinin değerlerini kullanarak mortalite riskini belirlemeyi amaçladık T.Ü. Tıp Fakültesi İç Hastalıkları AD Yoğun Bakım BD kliniğinde, 1 Ocak -31 Aralık 2013 arasında yatarak tedavi 18 yaş ve üzeri, incelenen verilerine sağlıklı olarak ulaşılabilen tüm hastalar dahil edildi. Bu özellikler dışındaki hastalar ile aynı hastaya ait mükerrer yatışlar çalışma dışı bırakıldı. Hastaların yaş ve cinsiyet bilgileri, sağ-kalım durumları, PLT (K/uL), CRP (mg/dl) ve Alb (gr/dl) düzeyleri ortak bir veri tabanında toplandı. Veriler, sağ-kalım sonucuna etkileri açısından istatistiksel olarak incelendi. Çalışmamızda ex grubun yaş ortalaması(69,7±15,6).Sağ grubun yaş ortalaması (65,3±17,8) (p=0,021) olarak hesaplandı. Çalışmamız sonuçlarının değerlendirmesinde dikkat çeken bir bulgu da, ortalama yaş 65'ten 70'lere çıktığında hayatta kalma oranının belirgin biçimde düştüğüdür. Çok değişkenli lojistik regresyon analizi yaş artışının mortalite riskini arttırdığını gösterdi (odds ratio=1,02; p=0,027). Çalışmamızda Ex grubu platelet ortalaması (199,7±133,7 K/uL), Sağ grubu platelet ortalamasına (272,1±137,3 K/uL) göre belirgin olarak düşük bulundu (p<0,001). Uygulanan ROC analizinde platelet cut-off değeri ?194 K/uL alındığında %58,7 sensitivite ve %69,9 spesifite ile mortaliteyi işaret ediyordu (AUC=0,660; p<0,001). Ayrıca çok değişkenli lojistik regresyon analizi Platelet ?194 K/uL değeri için mortalite riskinin odds ratio=3,32 kat arttırdığı gösterdi. Ex grubu albümin ortalaması (2,9±0,7 gr/dl), Sağ grubu albümin ortalamasına göre (3,3±0,6 gr/dl) (p<0,001) anlamlı düzeyde düşük bulundu. albümin cut-off değeri ?2,8 gr/dlolarak alındığında; %44,0 sensitivite ve %81,3 spesifite ile (AUC=0,654) (p<0,001) hasta mortalitesine işaret etti. Ayrıca, çok değişkenli lojistik regresyon analizi de albümin düzeyi ?2,8 gr/dl için mortalite riskinin odds ratio=2,51 kat arttığını gösterdi. da Ex grubu CRP ortalamasının (13,4±11,9 mg/dl) Sağ grubu CRP ortalamasından (8,3±9,5 mg/dl) anlamlı olarak yüksek bulundu. (p<0,001). Ayrıca, CRP cut-off değeri >3,34 mg/dl alındığında; %76,3 sensitivite ve %47,8 spesifite ile hasta mortalitesine işaret etti AUC=0,647 (p<0,001). Çalışmamızın en çarpıcı bulgusu olarak; hastaların yaş, PLT, albümin ve CRP düzeylerine bağlı olarak yoğun bakımda ölme olasılığını hesaplayan denklem, lojistik regresyon fonksiyonu yardımıyla aşağıdaki gibi bulundu; (1,12354 (0,01616*Yaş) (0,00322*Platelet) (0,53455*Albümin) (0,02357*CRP)) 1 exp Sonuç olarak; yoğun bakıma yeni yatmış hastalarda yaş yüksekliği, CRP yüksekliği,Albumin düşüklüğü,platelet düşüklüğü yüksek mortalite riski ile ilişkili bulunmuşolup,bu parametrelerin hepsinin birarada olması durumunda mortalite riski daha da anlamlı olarak artmaktadır. Bu parametrelerinin birlikte kullanılarak bir formülasyon ile yoğun bakıma kabul edilen hastalarda yüzde kaç olasılıkla mortalite gelişeceğini tahmin eden bir yazılım geliştirilmiştir. Anahtar kelimeler: Yoğun bakım, skorlama sistemleri, albümin, CRP, platelet,mortalite
The parameters to predict mortality in intensive care patients are important because they determine the necessity, location and speed of treatment. Due to their complexity, the practical use of the scoring systems is limited. In our study, we aimed to determine the mortality risk by using the values of C-reactive protein (CRP), albumen (Alb), platelet (PLT) and age parameters. All patients who were inpatient treatment in the intensive care unit of Department of Internal Medicine in Trakya University between 1 January and 31 December 2013, aged 18 yr. and whom we were able to access all the necessary data were included in the study. Recurrent hospitalizations of the same patient and patients who didn't have the characteristics given above were excluded from the study. During the study, age and gender information, survival, PLT, CRP (mg / dl) and Alb (gr / dl) levels were collected in a common database. All the data were analysed statistically in terms of their effects on survival. The average age of the group of death in our study (69.7 ± 15.6) and the average age of the survival group was calculated as (65,3 ± 17,8) (p = 0,021). The remarkable finding of our study is that, the significantly decrease of survival rate when the average age increased from 65 to 70. Multivariate logistic regression analysis showed that increased age increased the risk of mortality. (odds ratio=1,02; p=0,027).In our study, the platelet average of the death group (199,7 ± 133,7 K / uL) was significantly lower than the group of survivors (272,1 ± 137,3 K / uL) (p <0,001). In the ROC analysis, when platelet cut-off value is taken as ?194 K / uL, death was measured with sensitivity of 58.7% and specificity of 69.9%. (AUC = 0.660; p <0.001). In addition, multivariate logistic regression analysis showed that the mortality risk for Platelet ?194 K / uL increased odds ratio = 3.32 times. The average level of albumin in the death group (2.9 ± 0.7 gr / dl) was significantly lower than the group of survivors (3.3 ± 0.6 gr / dl) (p <0.001). Albumin indicated patient mortality with sensitivity of 44.0% and specificity of 81.3% (AUC = 0.654) (p <0.001) when the cut-off value is taken as ?2,8 gr / dl. Furthermore, multivariate logistic regression analysis showed that the mortality risk for albumin level ?2,8 gr / dl increased by odds ratio = 2,51. The median CRP level was significantly higher in the of the deceased group (13.4 ± 11.9 mg / dl) than the survivors group (8.3 ± 9.5 mg / dl) (P <0.001). When CRP cut-off value is taken as >3,34 mg/dl; in the study the patient mortality measured with 76.3% sensitivity and 47.8%specificity AUC = 0.647 (p <0.001). As the most striking finding of our study; The equation which was found with the help of logistic regression function and which calculates the probability of decease of patients in intensive care unit depending on their age, PLT, albumin and CRP levels can be seen below; As a result; The elevated age, CRP elevation, low albumin, and low platelet levels were associated with high mortality risk for patient who newly admitted to intensive care unit. If all of these parameters are combined, the risk of mortality even increases significantly. A software has been developed with a formulation that uses these parameters to estimate the mortality rate of patients admitted to intensive care. Key Words: Intensive care, scoring systems, albumen, C-reactive protein, platelet,mortality
The parameters to predict mortality in intensive care patients are important because they determine the necessity, location and speed of treatment. Due to their complexity, the practical use of the scoring systems is limited. In our study, we aimed to determine the mortality risk by using the values of C-reactive protein (CRP), albumen (Alb), platelet (PLT) and age parameters. All patients who were inpatient treatment in the intensive care unit of Department of Internal Medicine in Trakya University between 1 January and 31 December 2013, aged 18 yr. and whom we were able to access all the necessary data were included in the study. Recurrent hospitalizations of the same patient and patients who didn't have the characteristics given above were excluded from the study. During the study, age and gender information, survival, PLT, CRP (mg / dl) and Alb (gr / dl) levels were collected in a common database. All the data were analysed statistically in terms of their effects on survival. The average age of the group of death in our study (69.7 ± 15.6) and the average age of the survival group was calculated as (65,3 ± 17,8) (p = 0,021). The remarkable finding of our study is that, the significantly decrease of survival rate when the average age increased from 65 to 70. Multivariate logistic regression analysis showed that increased age increased the risk of mortality. (odds ratio=1,02; p=0,027).In our study, the platelet average of the death group (199,7 ± 133,7 K / uL) was significantly lower than the group of survivors (272,1 ± 137,3 K / uL) (p <0,001). In the ROC analysis, when platelet cut-off value is taken as ?194 K / uL, death was measured with sensitivity of 58.7% and specificity of 69.9%. (AUC = 0.660; p <0.001). In addition, multivariate logistic regression analysis showed that the mortality risk for Platelet ?194 K / uL increased odds ratio = 3.32 times. The average level of albumin in the death group (2.9 ± 0.7 gr / dl) was significantly lower than the group of survivors (3.3 ± 0.6 gr / dl) (p <0.001). Albumin indicated patient mortality with sensitivity of 44.0% and specificity of 81.3% (AUC = 0.654) (p <0.001) when the cut-off value is taken as ?2,8 gr / dl. Furthermore, multivariate logistic regression analysis showed that the mortality risk for albumin level ?2,8 gr / dl increased by odds ratio = 2,51. The median CRP level was significantly higher in the of the deceased group (13.4 ± 11.9 mg / dl) than the survivors group (8.3 ± 9.5 mg / dl) (P <0.001). When CRP cut-off value is taken as >3,34 mg/dl; in the study the patient mortality measured with 76.3% sensitivity and 47.8%specificity AUC = 0.647 (p <0.001). As the most striking finding of our study; The equation which was found with the help of logistic regression function and which calculates the probability of decease of patients in intensive care unit depending on their age, PLT, albumin and CRP levels can be seen below; As a result; The elevated age, CRP elevation, low albumin, and low platelet levels were associated with high mortality risk for patient who newly admitted to intensive care unit. If all of these parameters are combined, the risk of mortality even increases significantly. A software has been developed with a formulation that uses these parameters to estimate the mortality rate of patients admitted to intensive care. Key Words: Intensive care, scoring systems, albumen, C-reactive protein, platelet,mortality
Açıklama
Tıpta Uzmanlık
Anahtar Kelimeler
Biyoistatistik, Biostatistics
