CC BY
13
II. ХИРУРГИЯ
ESTIMATION OF RISK FACTORS AND FORECAST OF BLEEDING RELAPSES FROM ESOPHAGEAL VARICES IN PATIENTS WITH PORTAL HYPERTENSION AT LONG-TERM TREATMENT
МРНТИ 76.29.34
Mamedov A.Ya.
Scientific Center of Surgery named after academician M.A. Topchibashev, Baku, Azerbaijan
ABOUT THE AUTHORS
Mamedov. A. Ya. - Scientific Center of Surgery named after academician M.A. Topchibashev, Baku, Azerbaijan. email: akifym@gmail.com
Abstract
The aim of the study: to develop prognostic criteria for assessing the severity of PH, as well as for conducting stage-by-stage endoscopic prophylaxis of bleeding from esophageal and gastric varices in the long-term treatment
Materials and methods. This study is based on the research of the results of endoscopic treatment of 157 patients with EGV admitted to the Scientific Center for Surgery named after M.A. Topchibashev for the period from 2009-2019. Among these patients with PH, 111 (70.7%) were male, 46 (29.3%) were female. The age of patients varied from 31 to 83 years (average 46.3 y 1.7 years). The age of 77.7% (122 patients) of the studied was 31-60 years, which indicates the prevalence of the working age persons among the studied patients.
Conclusion. In the literature there are various mathematical models for predicting the postoperative stage of the disease. It should be noted that today there is no ideal model for predicting the relapse of bleeding. Since a significant part of the models was developed before the widespread use of endoscopic intervention in clinical practice.
¥зак мерзiмдi емдеу барысында портальдык гипертензиясы бар наукастардыц ецеш кектамырларыньщ варикозды кецеюшщ нэтижесжде кан кетудщ кайталануын болжау жэне кауш-катер факторларын багалау
Keywords
portal hypertension, bleeding, varicose veins, endoscopic hemostasis, prognosis
Мамедов А.Я.
Академик М.А. Топчибашев атында?ы Рылыми Хирургия Орталь™, Баку, Эзiрбайжан Ацдатпа
Мацсаты: узак мерз'шд'! емдеу барысында ещеш пен асказан кектамырларыньщ варикозды кенеюiнiн нэтижес'мде кан кетудщ кезендк эндоскопиялык профилактикасын журпзу максатында ПГ ауырлык дэрежеан багалауга арналган болжамалы критерийлердi дайындау.
Материалдары жэне эдютер'1. Бул зерттеу жумысы 2009-2019 жж. аралыгында М.А. Топчибашев атындагы Яылыми хирургия орталыгына ещеш пен асказан кектамырларыныщ варикозды кещею! деген диагнозбен тускен 157 наукасты эндоскопиялык емдеу барысындагы нэтижелерд'! зерт-теуге непзделген. ПГ-ы бар наукастардыщ 111 (70,7%) - ер адамдар, 46 (29,3%) - эйел адамдар. Наукастардыщ жастары 31-ден 83 жаска деШн турленед'! (орта есеппен 46,3±1,7 жас). Зерттеуге алынган наукастардыщ 77,7%-ныщ (122 наукастыщ) жасы 31-60 жасты курайды, бул зерттеуге алынган тулгалардыщ арасында ещбекке кдб'тетт'! жастагы адамдардын басым екенШ керсетедi.
Тужырым. Эдебиеттерде осы сыркаттыщ операциядан кейШп кезендерiн болжауга арналган турл'! математикалык модельдер бар. Алайда бYгiнгi тащда кайта кан кету рецидивн болжауга арналган мШЫз модель жок екенШ айтып етуiмiз керек. вйткеш модельдщ непзп белг клиникалык практика-да кещ колданыс тапканеа дейн дайындалгандыктан, ^рп уакытта тузетулер мен тиЫ езгертулер енпзуд'! кажет етед'!.
АВТОРЛАР ТУРАЛЫ
Мамедов А.Я. - Академик М.А. Топчибашев arbiHMaFbi Гылыми Хирургия Орталь^ы, Баку, дзiрбайжан. email: akifym@gmail.com
Туйш сездер
портальдьщ гипертензия, кан кету, коктамырдыц варикозды кецеюi, эндоскопиялык гемостаз, болжау
Оценка факторов риска и прогнозирование рецидивов кровотечения из ва-рикозно расширенных вен пищевода у больных с портальной гипертензией в отдаленные сроки лечения
ОБ АВТОРАХ
Мамедов А.Я. -
Научный Центр Хирургии имени академика М.А. Топчибашева, Баку, Азербайджан. email: akifym@gmail.com
Ключевые слова
портальная гипертензия, кровотечение, варикозное расширение вен, эндоскопический гемостаз, прогноз
Мамедов А.Я.
Научный Центр Хирургии имени академика М.А. Топчибашева, Баку, Азербайджан
Аннотация
Цель: разработать прогностические критерии оценки тяжести ПГ для поэтапной эндоскопической профилактики кровотечений из варикозно расширенных вен пищевода и желудка при длительном лечении.
Материалы и методы. Данное исследование основано на изучении результатов эндоскопического лечения 157 пациентов с расширением вен пищевода и желудка, поступивших в Научный центр хирургии им. М. А. Топчибашева за период 2009-2019 гг. Среди этих больных ПГ 111 (70,7%) были мужчины, 46 (29,3%) - женщины. Возраст пациентов варьировал от 31 до 83лет (в среднем 46,3± 1,7 года). Средний возраст 77,7% (122 пациента) обследованных больных составил 31-60 лет, что свидетельствует о преобладании среди обследованных лиц трудоспособного возраста.
Вывод. В литературе имеются различные математические модели для прогнозирования послеоперационной стадии заболевания. Следует отметить, что на сегодняшний день не существует идеальной модели для прогнозирования рецидива кровотечения, поскольку значительная часть моделей была разработана до широкого применения эндоскопических вмешательств в клинической практике.
Relevance
Observations of recent years show an increase in the number of patients with portal hypertension (PH). Many authors believe that the main reason for this is the growth of factors leading ultimately to liver cirrhosis (LC) [1,2].
The development of PH in 40-60% of cases is accompanied by bleeding from varicose dilated veins of the esophagus and stomach (VDVES (ESPHAGEAL AND GASTRIC VARICES (EGV)?)), which in some cases leads to death [3,4].
In recent years, in many countries of the world, including Azerbaijan, endoscopic methods of hemo-stasis are widely used to eliminate this complication - endoscopic ligation (EL), sclerotherapy (ES), as well as other methods [3,5,6]. However, it should be noted that in the analysis of literature data, the timing of these methods, indications for the choice of endoscopic hemostasis, depending on the severity of the disease, as well as comparative results in the long term are not described at the proper level.
The data show that the prevention and treatment of bleeding from EGV in patients with PH remains an urgent problem in modern medicine. We believe that in order to improve the results of complex treatment, it is necessary to evaluate prognostic criteria, and on their basis to improve methods for choosing the treatment of bleeding from EGV, as well as staged prevention of bleeding recurrence in patients with PH.
The aim of the study
To develop prognostic criteria for assessing the severity of PH, as well as for conducting stage-by-stage endoscopic prophylaxis of bleeding from esophageal and gastric varices in the long-term treatment
Materials and methods
This study is based on the research of the results of endoscopic treatment of 157 patients with EGV admitted to the Scientific Center for Surgery named after M.A. Topchibashev for the period from 2009-2019. Among these patients with PH, 111 (70.7%) were male, 46 (29.3%) were female. The age of patients varied from 31 to 83 years (average 46.3±1.7 years). The age of 77.7% (122 patients) of the studied was 31-60 years, which indicates the prevalence of the working age persons among the studied patients.
The diagnosis of PH and the presence of EGV was made on the basis of generally accepted diagnostic methods.
I degree of expansion was noted in 32 (20.4%), II degree in 51 (32.5%), III degree in 68 (43.3%), and IV degree in 6 (3.8%) cases.
By studying localization, it was found that in 86 (54.8%) patients EGV was located in the lower third of the esophagus, in 9 (5.7%) patients in the stomach, and in 62 (39.5%) it was cardioesophageal.
Types of endoscopic interventions Quantity
Abs. %
Endoscopic ligation of varicose veins of the esophagus 120 76,4
Endoscopic sclerotherapy of varicose veins of the esophagus 35 22,3
Endoscopic ligation of varicose veins of the stomach 16 10,2
Endoscopic sclerotherapy for varicose veins of the stomach 17 10,8
Obstruction of varicose veins of the bottom of the stomach with histoacrylic glue 6 3,8
The degree of expansion of varicose veins ( x, ) With relapse (n=18) Without relapse (n=86) Confidence in the difference
abs. % abs. %
I-II degree 2 11,11 50 58,14 p<0,001
III-IV degree 16 88,89 36 41,86 p<0,001
Total 18 100 86 100
Table 1.
Endoscopic interventions on esophageal and gastric varices in patients with portal hypertension
Table 2.
The degree of expansion of varicose veins
The degree of expansion of varicose veins ( x, ) Number of observations Probabilities
With relapse Without relapse P( x[l A1) P( X1 a2)
I-II degree 2 50 0,111 0,581
III-IV degree 16 36 0,889 0,419
Total 18 86 1 1
Table 3.
Diagnostic table
In the studied patients, PH in 109 (69.4%) cases was the result of liver cirrhosis (LC), and in 48 (30.6%) cases - the result of extrahepatic causes (portal vein thrombosis), portal cavernous transformation, portal vein thrombosis.
To conduct a comparative analysis of the treatment results, we divided the observed groups into 2 groups: group I included 81 patients who were admitted to the hospital urgently with bleeding from EGV; group II included 76 (48.4%) patients who were admitted as planned (without bleeding) for endoscopic diagnosis and treatment.
The severity of the underlying disease in 83 (52.9%) cases was complicated by the presence of concomitant diseases, among which the prevailing were cardiovascular as well as endocrine diseases.
Table 1 shows the types of endoscopic interventions performed to the studied patients.
In the postoperative period (12 months), 104 patients were examined - 18 with relapse of repeated bleeding and 86 without relapse.
According to the data before and postoperative examinations, the following significant risk factors for repeated bleeding were determined: the degree of varicose vein dilatation, the presence of vascu-lopathy and gastropathy, classification according to Child Pugh, thrombocytopenia, bad habits, portal hypertension >20 mmHg.
To predict relapse of bleeding, we used the well-known Bayes formula, which is successfully used to predict various phenomena [7].
P(Ak )P(xj / Ak) (1)
P(Ak / xj) = j-—
Kk 2P(Ak)P(xj /Ak)
k
where P(Ak) - is the a priori probability of the
state A, k=1,2...../; 2 PA ) = 1;
P(xij) - a priori probability of gradation xij ,
/=1,2.....r; /=1,2,... m;
r - is the total number of features, m - is the number of values of each feature.
In this study, two variants of the condition were considered (k=2): A1 - there is a relapse of repeated bleeding, A2 - there is no relapse.
According to the above data, the a priori probabilities of states are calculated: P(A1)= 0,17; P(A2)=0,83. The frequency of occurrence was determined and diagnostic tables were compiled for each risk factor.
It should be noted that among the studied parameters, the degree of expansion of varicose veins in the prognostic aspect (to study the likelihood of bleeding) is one of the most basic. Studies have shown that the presence of III and IV degree of expansion of the veins of the esophagus and stomach are high risk factors, which dictates the need for a planned procedure for staged prophylaxis of bleeding with endoscopic methods.
However, it should be noted that in patients with PH in the presence of I-II degree of EGV, vasculopa-thy and gastropathy, the risk of bleeding increases several times.
The reason for this is arrosive changes in the mucosa of the esophagus and stomach, as well as degenerative changes in the wall of varicose veins.
When conducting staged prophylaxis of bleeding recurrence in the long-term treatment period, it is necessary to study the severity of the underly-
Table 4.
The presence of vas-culopathy and gastropathy
Vasculopathy and gastropathy ( x2 ) With relapse (n=18) Without relapse (n=86) Confidence in the difference
abs. % abs. %
Yes 16 88,89 39 45,35 p<0,001
No 2 11,11 47 54,65 p<0,001
Total 18 100 86 100
Table 5.
Diagnostic table
Vasculopathy and gastropathy ( x\ ) Number of observations Probabilities
With relapse Without relapse P( x[ / a,) P(x[ /A2)
Yes 16 39 0,889 0,453
No 2 47 0,111 0,547
Total 18 86 1 1
Table 6.
Child Pugh classification
Child Pugh classification ( x^ ) With relapse (n=18) Without relapse (n=86) Confidence in the difference
abs. % abs. %
Class A 4 22,22 53 61,63 p<0,001
Class B 9 50,00 26 30,23 p>0,05
Class C 5 27,78 7 8,14 p<0,05
Total 18 100 86 100
Table 7
Diagnostic table
Child Pugh classification ( x3 ) Number of observations Probabilities
With relapse Without relapse With relapse Without relapse
Class A 4 53 0,222 0,616
Class B 9 26 0,500 0,302
Class C 5 7 0,278 0,081
Total 18 86 1 1
Table 8.
The presence of thrombocytopenia
Thrombocytopenia ( x; ) With relapse (n=18) Without relapse (n=86) Confidence in the difference
abs. % abs. %
Yes 13 72,22 41 47,67 p>0,05
No 5 27,78 45 52,33 p<0,05
Total 18 100 86 100
Table 9.
Diagnostic table
Thrombocytopenia ( x; ) Number of observations Probabilities
With relapse Without relapse With relapse Without relapse
Yes 13 41 0,722 0,477
No 5 45 0,278 0,523
Total 18 86 1 1
Table 10.
Bad habits
Bad habits ( x'5 ) With relapse (n=18) Without relapse (n=86) Confidence in the difference
abs. % abs. %
Yes 8 44,44 21 24,42 p<0,05
No 10 55,56 65 75,58 p<0,005
Total 18 100 86 100
Table 11.
Diagnostic table
Number of observations Probabilities
Bad habits ( x'5 ) With relapse Without relapse With relapse Without relapse
Yes 8 21 0,444 0,244
No 10 65 0,556 0,756
Total 18 86 1 1
ing disease. So, in the case of PH, the cause of which was the LC, it is necessary at the stages of treatment to study the severity according to Child Pugh.
The diagnostic table shown in table 7 indicates the high prognostic significance of this parameter for choosing the time for prophylactic endoscopic interventions for EGV.
Among the risk factors for bleeding from EGV, trobmocytopenia is limited not only by a high prognostic value.
This parameter is an indicator of the functional state of the liver, the compensatory capabilities of the body, the presence of congenital hematological disorders. It is necessary for the choice of complex treatment methods, since the methods of endo-scopic and drug hemostasis for bleeding from EGV in most cases were ineffective and fatal. This indicates the need for a thorough staged prevention of bleeding in these patients.
Another informative risk factor for the development of bleeding from EGV in the prognostic plan is the presence of a patient's history of bad habits. The latter include nutritional factors: the use of alcohol, nicotine, an irrational lifestyle, etc.
Table 11 indicates the significance and likelihood of the risk of developing bleeding from EGV in patients with unhealthy habits.
One of the important prognostic factors for the development of bleeding in the observed patients is hypertension in the portal vein. The study of the portal gradient in the dynamics of treatment is necessary for the adoption of appropriate preventive and therapeutic measures.
The diagnostic table shown in table 13 indicates the prognostic value of this indicator for studying the likelihood of developing bleeding from EGV.
Thus, all indicators for the application of formula (1) are defined:
In the future, given the above tables, according to the Bayes formula, we can calculate the probability of the outcome depending on the signs:
(2)
(3)
P(A1 ! x[, x'2,x'3, xf,x'stxl6) = 6
0.17 H Pix'j/A,)
__i^i_
6 6 0.17 -ft P(xj/A1) + 0.83 -H P(x'./A2)
j=i
P(A2 / x'1,x'2,x'3,x'i,x's,x'6) =
6
0.83 ]J P(xj/A2)
__y-i_
6 6 0.17 P(x'. / A^ + 0.83 - J} P{x'. / A2
j-l i-l
where xj - gradations of a sign. Values of P( xj / Aj) and P( x'j / A2) are calculated from the above data.
Examp/e 1. Patient A.B.A., born in 1953, case history No. 22794, was admitted to the clinic with a diagnosis of "Portal hypertension complicated by bleeding from varicose veins of the esophagus." The degree of expansion of varicose veins - IV degree, vasculopathy and gastropathy are recorded, classification according to Child Pugh - class B, thrombocytopenia - was observed, has bad habits, hypertension in the portal vein is >20 mmHg.
Given the data in table 14, using formulas 2
and 3, we obtain that, in the presence of these
signsP(A1 /xj,x2,x3,xj,x'5,x'6) = 0,752 ,
P(A2 / xj, x2, x3, xj, x5, x'6) = 0,248 . Thus, the probability of relapse of bleeding in this patient, according to the developed prognostic model, is 75.2%. In a repeated examination of the patient 12 months after the operation, a relapse of repeated bleeding was recorded, which corresponds to our calculations.
Examp/e 2. Patient A.S.B., born in 1967, case history No. 18887, was admitted to the clinic with a diagnosis of "Portal hypertension complicated by bleeding from varicose veins of the esophagus." The degree of expansion of varicose veins - III degree, vasculopathy and gastropathy are recorded, classi-
Hypertension in the portal vein ( x'6 ) With relapse (n=18) Without relapse (n=86) Confidence in the difference
abs. % abs. %
>20 mmHg 15 83,3 38 44,2 p<0,001
<20 mmHg 3 16,7 48 55,8 p<0,001
Total 18 100 86 100
Table 12.
Hypertension in the portal vein >20 mmHg
Hypertension in the portal vein ( x'6 ) Number of observations Probabilities
With relapse Without relapse With relapse Without relapse
>20 mmHg 15 38 0,833 0,442
<20 mmHg 3 48 0,167 0,558
Total 18 86 1,0 1,0
Table 13.
Diagnostic table
Table 14.
Table 15.
Prognosis model feasibility table
Sign (x.) Gradations of a sign: xj P(xj / A,) P(xj / A2)
1 The degree of expansion of varicose veins III-IV degree 0,889 0,419
I-II degree 0,111 0,581
2 Vasculopathy and gastropathy Yes 0,889 0,453
No 0,111 0,547
Class A 0,222 0,616
3 Child Pugh classification Class B 0,500 0,302
Class C 0,278 0,081
4 Thrombocytopenia Yes 0,722 0,477
No 0,278 0,523
5 Bad habits Yes 0,444 0,244
No 0,556 0,756
6 Portal vein hypertension >20 mmHg 0,833 0,442
<20 mmHg 0,167 0,558
Relapse observation result Relapse prognosis Total
Expected Not expected
Observed 16 2 18
Not observed 2 31 33
Total 18 33 51
fication according to Child Pugh - class A, thrombocytopenia - was observed, does not have bad habits, hypertension in the portal vein is <20 mmHg.
Given the data in table 14, using formulas 2 and 3, we obtain that, in the presence of these signs P(A1 /x\,x2,x3,xJ,x'5,x'6) = 0,124 ,
P(A2 / x1, x2, x3, xJ, x5, x6) = 0,876 . Thus, the probability of relapse of bleeding in this patient, according to the developed prognostic model, is 12.4%. In fact, upon repeated examination of the patient 12 months after surgery, relapse of repeated bleeding was not observed.
To assess the effectiveness of the prognostic model, we examined the performance of all 18 patients with relapse of bleeding and 33 without relapse (51 in total) after surgery.
According to empirical data, a model feasibility table was compiled (Table 15) [8].
According to the table calculated performance indicators:
Sensitivity Model: SE = 0.889 Model specificity: SP = 0.939 The overall skill of the model: P = 0.922 Model accuracy criterion: Q = 0.828
As shown, the effectiveness of the predictive model using the Bayes formula has a level in the range of 82-92%.
Conclusion
In the literature there are various mathematical models for predicting the postoperative stage of the disease. It should be noted that today there is no ideal model for predicting the relapse of bleeding. Since a significant part of the models was developed before the widespread use of endoscopic intervention in clinical practice, there is currently a need for improvements and appropriate amendments.
Our model, which includes a number of basic preoperative parameters, improves the prediction of relapse recurrence.
Findings
The prognostic model developed by us (using the Bayes formula) allows us to timely assess the real state of EGV in patients with PH, take step-by-step preventive measures to prevent the development of bleeding, and reduce mortality and thereby improve treatment outcomes.
References
1. Vinokurov M.M., Jakovleva Z.A., Buldakova L.V., Timofeeva M.S. Varikoznoe rasshirenie ven pish-hevoda i zheludka pri portal'noj gipertenzii. Jen-doskopicheskie metody ostanovki i profilaktiki krovotechenij // Fundamental'nye issledovanija, 2013, № 7-2, s.281-285.
2. Dzidzava I.I. Otdalennye rezul'taty hirurgicheskoj korrekcii portal'noj gipertenzii i prognosticheskie faktory vyzhivaemosti u bol'nyh cirrozom pecheni: Avtoref. dis. ... d-ra med. nauk. SPb., 2010, 46 s.
3. Wang C., Han J., Xiao L. et al. Efficacy of vaso-pressin/terlipressin and somatostatin/octreotide
for the prevention of early variceal rebleeding after the initial control of bleeding: a systematic review and meta-analysis. // Hepatol. Int., 2015,9, p. 120-129.
4. Kicenko E.A., Anisimov A.Ju., Andreev A.I. Sovre-mennoe sostojanie problemy krovotechenij iz varikozno-rasshirennyh ven pishhevoda i zheludka // Vestnik sovremennoj klinicheskoj mediciny, 2014, № 5, s. 89-98.
5. Kurbonov H.H Jendoskopicheskoe ligirovanie varikozno-rasshirennyh ven pishhevoda (VRVP) u bol'nyh cirrozom pecheni s ugrozoj pishhevodno-
go krovotechenija // Problemy gastrojenterologii, 2014, №1, s.48-51
6. Tripathi D., Stanley A., Hayes P. et al. U.K. guidelines on the management of variceal haemorrhage in cirrhotic patients. // Gut., 2015, 64, p. 1680-1704.
7. Rustamov Ja.I., Jefendiev G.D.. Postroenie prog-nosticheskoj modeli bezotkaznoj raboty subarte-zianskih skvazhin. // Vestnik HNTU, Herson, 2015, № 3 (54), s. 179-183
8. Junkerov V.I., Grigor'ev S.G. Matematiko-statis-ticheskaja obrabotka dannyh medicinskih issledo-vanij. SPb.: VMedA, 2002, 266 s.
