Risk factors for rebleeding

INTRODUCTION

Acute episodes of bleeding from the esophagus as well as from stomach and duodenum are emergency conditions in clinical medicine, with the incidence from 50 to 150/100000 per year (1, 2). The rate of rebleeding after initial hemostasis, especially mortality, seems not to have been significantly changed during the last decades in spite of using modern diagnostic and therapeutic methods. Important studies indicate that about 4 to 14% of patients are dying because of upper gastrointestinal bleedings (UGIB) or its complications (3-5). Having taken into consideration the fact that the population is generally becoming older which is the most emphasized risk factor causing mortality after bleeding or rebleeding from the upper parts of the digestive system, so it is considered that there is discrete decreasing mortality rate (6,7). Rebleeding after initial successfully treated bleeding lesion is considered the main risk factor for death, and, in spite of applied hemostasis method, it happens in about 10 to 30% patients (8). The generally accepted viewpoint considers the risk of rebleeding and mortality to be connected with many independent, but related factors.

Up to 80% of upper gastrointestinal bleedings are “self-limited”, which means that they stop spontaneously either one of the endoscopic hemostatic procedure was carried out or it was not (9). If we analyze only the data from history and clinical parameters of the disease, it is practically impossible to define the group of patients whose bleeding is not going to stop spontaneously. American National Institutes of Health (NIH) held Concensus Conference in 1989. where the endoscopic therapy of bleeding ulcer was marked as the method of choice with “excellent efficiency and acceptable low level of the related complications” (10).

MATERIAL AND METHODS

This study included 315 patients who had the signs of bleeding from the upper parts of gastrointestinal system and were admitted to the Emergency Centre of the Clinical Centre of Serbia (EC) in Belgrade in the period from January 1998 to March 1999. We included in the study only the patients with successfully completed endoscopic injection hemostasis of bleeding lesions during the initial endoscopy. The patients who had acute bleeding from esophageal varices were not included in the study. The history of disease was taken, clinical and laboratory tests were performed after the admittion to the EC.

We analyzed gender and age, the clinical presentation of acute bleeding, the data about the use of non-steroidal anti-inflammatory drugs and acetylsalicylic acid and the data about coexisting diseases (comorbidity). The state of shock was especially taken into consideration among clinical parametes. It was arbitrary taken that those patients who had systolic pressure lower than 100 mmHg and heart frequency more than 100/min were in the state of shock. The laboratory parameters included the tests of hemoglobin level (lower borderline was taken as 10 g/l).

The upper endoscopy was performed within the first 12 hours from the admittion in a standard way with the pharingeal anesthesia. The endoscopy was always performed by an experienced endoscopist who immediately and independently decided about the necessity and the type of the endoscopic sclerotherapy that was to be applied. Intervention was always started with the anaemization of the bleeding lesion surroundings using dilute epinephrin injection (1:10000 solution in 0, 9% NaCL). After that, the sclerosating agent was injected directly into the lesion. We used absolute alcohol (96% ethanol) as a sclerosating agent. Localization and the type of lesion were recorded during the endoscopy as well as the type of the bleeding stigmata. We marked following seven localizations: esophagus, stomach fornix, corpus, and antrum, duodenal bulbus anterior and posterior wall and the postoperative anasthomosis zone (not depending on the type of the operation). Bleeding lesions were divided into five types: ulcer, malignancy, Mallory-Weiss’ syndrome, erosive disease of the upper parts of the digestive system and “the other lesions”. Vascular malformations and Dieulafoy’s lesions were considered to be in the category of “the others”. The bleeding stigmata were grouped according to modified Forrest’s classification (11). According to the study design, only the bleeding lesions that required endoscopic injection hemostasis were classified: adherent clot in the ulcer base, non-bleeding visible vessel, oozing of blood and spurting arterial bleeding. Rebleeding episodes were registered. Rebleeding is defined as a new episode of bleeding during the hospitalization, which was manifested as recurent haematemesis, the appearance of fresh blood in the nasogastric aspirate or the necessity of operation because of continuous bleeding. Mortality was defined as death during the hospitalization (12). The surgucal therapy and the duration of hospital stay were noted for all patients as well.

RESULTS

The study included 315 patients: 210 (66, 7%) male and 105 (33,3) % female. The characteristics of the patients with the number of rebleedings and mortality after initial successfully applied hemostasis during the hospitalization are summarized in Table 1.

Rebleeding was registered on 53 (16, 8%) patients, while 11(3, 5%) patients died during the hospitalization. Surgery was performed on 37 (11, 7%) patients. Most of the patients were operated because of uncontrolled rebleedings. The two left patients (5, 4%) were operated due to stomach cancer in the attempt of its radical treatment. As to the statistical data, the operation did not increase mortality rate while compared with non-operated patients (p = 0,127). Every parameter was analyzed individually in relation to rebleeding and mortality. The age of the patients does not influence neither on rebleeding nor on mortality (p > 0, 05). Gender, hemoglobin level at the

moment of admittance to the hospital, treatment with ulcerous drugs, and clinical presentation at the moment of bleeding did not influence significantly on rebleeding or mortality. Statistically more significant rate of rebleeding is noted on patients who were initially in shock compared to the patients who were not hypotensive and did not have tachicardia at the moment of the admittion to the hospital ( p = 0,006). It is statistically very significant that patients who died were initially in the state of shock, compared to the patients who survived (p = 0, 0001). The type of bleeding stagmata had statistically significant influence on rebleeding (p = 0,018), but not on mortality (p = 0,310) (Figure 1).

Coexisting illnesses influenced most significantly both the occurrence of rebleeding (p = 0, 0001) and mortality during the hospitalization (p = 0, 0001) (Figure 2). Statistically very significant was the rebleeding episode which increased the rate of mortality among treated patients (p = 0, 0001).

It was also analyzed the number of days spent in hospital (Table 2). That analysis showed that the average duration of hospital stay for non-operated patients, for patients without rebleeding and for patients with favourable outcome, was 6 to 7 days.

We defined the significant parameters for patients who had rebleeding or died during the hospitalization. In the context of all parameters, we analyzed individual influence of each parameter related to assigned variable by the logistic regression. First, the dependent variable was related to the occurrence of rebleeding, in the second case it was related to mortality. The state of shock at the moment of admittance, the type of the stigmata of bleeding and coexisting diseases contributed significantly to the occurrence of rebleeding after initial successful endoscopic hemostasis. The analysis of the same parameters related to mortality showed statistical significance for: the occurrence of rebleeding, the state of shock at the moment of hospitalization and the degree of coexisting illnesses.

DISCUSSION

Cook et al. introduced meta-analysis of rebleeding and mortality on 2412 patients and concluded that there was no statistically significant difference depending on applied endoscopic techniques for controlling the bleeding in regard to rebleeding rate. Statistically, mortality was reduced significantly only in the group treated by laser (13).

Lin et al. pointed out that the combination of contact-thermal with injection method was the most efficient either in the prevention of peptic ulcer rebleeding or in decreasing the need for blood transfusions (14).

The analyses of the variance showed that rebleeding increases mortality rate 5 to 16 times, which is the first step towards death. Modern techniques of endoscopic hemostasis make bleeding stop successfully in majority of cases and consequently decrease the rebleeding rate significantly. However, it does not give the expected mortality rate decrease among these patients.

Two studies point out that the decrease of rebleeding rate is followed by adequate decrease of mortality when laser hemostasis is applied (13, 15). Most authors conclude that rebleeding is a significant factor that leads to mortality after acute bleeding from the upper parts of the digestive system, but it is not the only one factor that influences significantly on the final result (15). Today it is not in debate should we apply initial endoscopic hemostasis, but it is very important how to decide what optimal treatment should be used in every individual patient after successfully applied endoscopic hemostasis (8). The final goal of a great number of studies is to select the patients with risk factors for rebleeding and mortality, but there is still no final answer (8). At the same time revealing the patients without risk factors for rebleeding, or with just a little risk, is a significant goal in attempt to optimize the treatment, choose the way of ambulatory treatment and reduce the number of hospitalizations and treatment expenses.

In our group, total number of rebleeding was 53 (16, 8%), the number of patients who died was 11 (3, 5%). These results are in accordance with the most of the published ones, which included applying injection hemostasis in patients with UGIB (4, 13, 17, and 19). Almost all the studies introduce a greater number of men than women. In most frequently quoted Rockall’s study, which analyzed risk factors for rebleeding and mortality, men had higher risk for rebleeding, while women had higher mortality rate (15). Majority of authors find the old age to be one of the statistically significant factors for increasing the rate of rebleeding and mortality after acute bleeding (8, 15, and 16). Most of our patients were from 60 to 80 years old (47, 6%). No one patient younger than 40 died. The results of our study prove that the rate of rebleeding and mortality is increasing with older age but these differences are not found to be statistically significant.

There were 21, 6% of our patients in the state of shock, 27, 9 % of whom rebleeded and 13, 2% of them died. In their study, Brullet et al. (17) define the state of shock to be one of the factors that contribute to the failure of injection hemostasis of bleeding stomach ulcer. Hypotension is also emphasized to be a bad prognostic factor in the study about hemathemesis and bleeding of Schiler et al. (16) from “the pre-endoscopic era”.

Peura et al. found hemoglobin level to be lower than 10 g/l in 60, 9% of the patients with the episode of UIGB. Similar to our results, Pimpl (19) and Katchinski (4) do not consider the hemoglobin level to be prognostically significant factor for rebleeding or mortality. In our study the hemoglobin level did not differ significantly among the rebleeding patients, the patients who died or the rest who were tested.

According to our data, the rebleeding was approximately the same in the group of patients who were treated with ulcerous drugs or who did not have such treatment (16, 9% and 16, 7%, respectively). Similar data were found for mortality. Laslo et al. published a great multicentric study about clinical aspects of upper gastrointestinal bleeding, related to previous treatment with ulcerous drugs. The main conclusion of this study is the fact that there are no important differences in clinical presentation of significant UGIB, in spite of taking drugs previously (20).

A great number of studies underline the influence of coexisting diseases over the outcome of these patients (8, 19, 21, and 22). Most of these studies do not separate coexisting diseases, but use the universal scale for their “common” estimation. ICED scale (Index of Co-Existing Disease, National Auxiliary Publication Service, USA) is accepted by most of the studies (23). Our study shows clearly that the presence and severity of the coexisting diseases contribute significantly to the occurrence of rebleeding and mortality (p = 0, 0001). So in the group of patients with severe, life threatening associated diseases, half of them had the episode of rebleeding, while almost half of this half (42,8%) died.

Bour in his study about endoscopic “interobserver agreement” made an interesting point in describing and naming the identical bleeding stigmata at peptic ulcer. A very high level of observer agreement is described related to spurting arterial bleeding, a satisfactory level of agreement is attained in regard to adherent clot and very low level of agreement was reached in regard to non-bleeding visible vessel and oozing of blood. This group of authors makes a conclusion that the limitation of number of stigmata classes into several categories and adequate endoscopist’s training could lead to better precision in the description of these entities (23). In our study, the episodes of rebleeding and mortality are connected most frequently with the most serious stigmata of bleeding, so called active spurting arterial bleeding. Contrary to this, the patients with an adherent blood clot at the bottom of the lesion had the least complications.

The most frequent rebleedings were registered from the stomach fundus (21, 1%), while the most risky zone for bleeding was the posterior wall of duodenal bulbus. At the same time, unexpectedly this location takes only the third place related to the percentage of rebleeding, after the stomach fundus and stomach corpus. The patients with ulcer, malignant tumors and arterio-venous malformation of the blood vessels had the greatest number of rebleeding episodes. The patients with rebleedings from malignant tumors and angiodisplasia had the highest risk of dying. As to localizations related to death, the most risky zones were stomach corpus and stomach fundus (8% of patients died per each localization). Rockall et al. point out that right diagnosis (the cause of bleeding) influences significantly over the outcome of the patient. They also emphasize that the worst thing is the presence of malignancy, and the best prognostic factor is Mallory-Weiss’ syndrome (15).

We did not include surgical treatment in the analysis of different parameters that influenced over the bleeding episode and/or mortality among these patients. We consider that the operation result and postoperative period depend on a number of facts that are not necessarily connected with the upper gastrointestinal bleeding itself. Most authors use similar procedure (8, 15, 19, 22, 24, and 26). The percentage of the operated patients due to bleeding varies in different series. Peura et al. analyzed 1235 patients, among which 7% were operated (18). Rockall quotes the operative rate of five, 7% (27), while Everett shows that 8% of patients from the high-risk group were operated. It is vivid that the percentage of the operated patients in all above mentioned studies is lower than 11, 1% which we present in our study. The better efficiency in all those studies is probably due to applying more efficient initial hemostasis by using combined contact-thermal and injection method.

The need and duration of hospitalization for all the bleeding patients is a controversial question. In their review about ulcer and non-variceal bleeding and appropriate measures in the treatment of such patients, Rollhauser and Fleischer (28) quote that the average duration of hospital stay is two, 9 to 4, 6 days. The authors conclude that shorter stay at hospital is clearly connected with performing the endoscopy shortly after admission and with proper identification of type of the bleeding lesion. Rockall et al. published a study about a possible selection of the patients for earlier dismissal from hospital according to the risk of rebleeding, underlining that the average duration of hospital stay is about 8.6 days, and the duration is half less with lower risk (21). Our patients, with favourable outcome and without rebleeding, had been at hospital for 6-7 days, which is an unjust long period of time.

The above-mentioned way of analyzing patients, according to the risk of rebleeding and mortality by identifying the risk factors, could be a significant element in estimation of the condition of every individual patient and it could influence the final decision connected with the further treatment. On the other hand, it could be the contribution to the more efficient standardization of medical procedures and clinical studies on therapeutic endoscopy, in estimation of existing algorithms for diagnosis and treatment and in optimization of hospital stay duration for these patients.

1. Cutler JA: Upper gastrointestinal bleeding. Nature and magnitude of the problem in the US. Dig Dis Sci 1981; 26(suppl.7):90-6S.

2. Johnstone SJ: Epidemiology and course of gastrointestinal haemorrhage in northeast Scotland. BMJ 1973; 3: 655-60.

3. Branicki FJ, Coleman SY, Fok PJ et al: Bleeding peptic ulcer: a prospective evaluation of risk factors for rebleeding and mortality. World J Surg 1990; 14:262-9.

4. Katschinski B, Logan R, Davies J et al: Prognostic Factors in Upper Gastrointestinal Bleeding. Dig Dis Sci 1994; 39:706-712.

5. Vreeburg EM, Snel P, de Bruijne JW, Bartelsman JF, Rauws EA, Tytgat GN: Acute upper gastrointestinal bleeding in the Amsterdam area: incidence, diagnosis, and clinical outcome. Am J Gastroenterol 1997; 92: 236-43.

6. Gilbert DA: Epidemiology of upper gastrointestinal bleeding. Gastrointest Endosc 1990; 36:S8.

7. Kang JY, Piper DW. Improvement in mortality rates in bleeding peptic ulcer disease. Med J Aust 1980; 1: 213.

8. Saeed Z, Ramirez FC, Hepps KS et al. Prospective validation of the Baylor bleeding score for predicting the likelihood of rebleeding after endoscopic hemostasis of peptic ulcers. Gastrointest Endosc 1995; 41:561-5.

9. Fleisher D. Etiology and prevalence of severe persistent upper gastrointestinal bleeding. gastroenterology 1983; 84: 538.

10. Proceedings of the Consensus Conference on Therapeutic Endoscopy in Bleeding Ulcers. Gastrointest Endosc 1990; 36 (suppl 1).

11. Forrest JAH. Endoscopy of upper gastrointestinal bleeding. Lancet 1974; 2: 394-7.

12. Vreeburg EM, Terwee CB, Snel P et al. Validation of the Rockall risk scoring system in upper gastrointestinal bleeding. Gut 1999; 44:331-335.

13. Cook DJ, Guyatt GH, Salena BJ, Laine LA. Endoscopic Therapy for Acute Nonvariceal Upper Gastrointestinal Hemorrhage: A Meta-Analysis. Gastroenterology 1992;102:139-148.

14. Tseng GY. A prospective randomized trial comparing injection of epinephrine alone, bipolar electrocoagulation alone and the combination of injection of epinephrine with bipolar electrocoagulation for the arrest of peptic ulcer bleeding. DDW, Orlando, USA, 1999.

15. Rockall TA, Logan RF, Devlin HB et al. Risk assessment after acute upper gastrointestinal haemorrhage. Gut 1996; 38: 316-321.

16. Schiller KFR, Truelove SC, Gwyn Williams D: Haematemesis and Melaena, with Special reference to Factors Influencing the Outcome. BMJ 1970;2: 7-14.

17. Brullet E, Campo R, Calvet X et al. Factors related to the failure of endoscopic injection therapy for bleeding gastric ulcer. Gut 1996; 39: 155-158.

18. Peura DA, Lanza FL, Gostout CJ et al. The American College of Gastroenterology Bleeding Registry: Preliminary Findings. Am J gastroenterol 1997; 92(6):924-928.

19. Pimpl W, Boeckl O, Waclawiczek HW, Heinerman M. Estimation of the Mortality Rate of patients with Severe Gastroduodenal Hemorrhage with the Aid of a new Scoring System. Endoscopy 1987; 19: 101-106.

20. Laslo A, Kelly JP, Kaufman DE et al. Clinical Aspects of Upper Gastrointestinal Bleeding Associated With the Use of Nonsteroidal Anti-inflammatory Drugs. Am J Gastroenterol 1998; 93: 721-725.

21. Rockall TA, Logan RFA, Devlin HB, Northfield TC. Selection of patients for early discharge or outpatient care after acute upper gastrointestinal haemorrhage. Lancet 1996; 347: 1138-40.

22. Saeed ZA, Winchester CB, Michaletz PA et al. A Scoring System to Predict Rebleeding after Endoscopic Therapy of Nonvariceal Upper Gastrointestinal Hemorrhage, with a Comparison of Heat Probe and Ethanol Injection. Am J Gastroenterol 1993; 88:1842-1849.

23. Greenfield S, Apolone G, McNeil BJ, Cleary PD. The importance of co-existing disease in the occurrence of postoperative complications and one-year recovery in patients undergoing total hip-replacement. med Care 1993; 31:141-54.

24. Provenzale D, Sandler RS, Wood DR et al. Development of a Scoring System to Predict Mortality from Upper Gastrointestinal Bleeding. Am J Med Sci 1987; 294: 26-32.

25. Bour B, Person B, Cales P et al. Interobserver agreement on endoscopic diagnosis of bleeding peptic ulcers. Gastrointest Endosc 1997; 46: 27-32.

26. Jones SGW, Davies R, Epworth I et al. Use of Rockall score to assess the effectiveness of an upper gastrointestinal bleeding unit. Gut 1996; 39 (Suppl 1): A4.

27. Rockall TA, Logan RFA, Devlin HB, Northfield TC. Influencing the practice and outcome in acute upper gastrointestinal haemorrhage. Gut 1997; 41:606-611.

28. Rollhauser C, Fleischer DE: Ulcers and Nonvariceal Bleeding. Endoscopy 1999; 31: 256-8.

29. Popović N, Bulajić M, Glišić M et al. Endoscopic injection hemostasis for non-variceal bleeding of stomach and duodenum. 6th World Congress of Endoscopic Surgery, Rome, Italy, June, 1998.

30- Barkun A, Bardou M, Marshall JK. Consensus recommendations for managing patients with nonvariceal upper gastrointestinal bleeding. Ann Intern Med 2003; 139:843.

31- Courtney AE, Mitchell RMS, Rocke L, Johnston BT. proposed risk stratification in upper gastrointestinal haemorrhage: Is hospitalisation essential? Emerg Med J 2004; 21:39-40.

SAŽETAK

UVOD: Endoskopska injekciona hemostaza je efikasan metod postizanja hemostaze kod krvarećih lezija gornjeg dela digestivnog trakta. Rekrvarenje se ipak dešava kod 10 do 30% pacijenata, a 4 do 14% tih pacijenata umre od posledica krvarenja. CILJ rada je da se odrede faktori rizika za rekrvarenje i smrt posle inicijalno uspešne endoskopske injekcione hemostaze. MATERIJAL I METODE: Prospektivna studija je uključila 315 pacijenata koji su primljeni u bolnicu zbog epizode akutnog nevariksnog krvarenja iz gornjih delova digestivnog sistema. Svi pacijenti su u toku prvih 12 sati od prijema u bolnicu endoskopirani i krvareća lezija je tom prilikom uspešno hemostazirana. Registrovane su epizode rekrvarenja i smrtni ishod u toku inicijalne hospitalizacije. Praćeni su sledeći parametri: starost, pol, uzimanje ulcerogenih lekova, šokno stanje, stigmata krvarenja, lokalizacija krvareće lezije, udružene bolesti i rekrvarenje. REZULTATI I ZAKLJUČAK: Rekrvarenje se desilo kod 58 pacijenata (16.8%) i bilo je statistički značajno povezano sa šoknim stanjem u trenutku prijema u bolnicu, stgmatima krvarenja i udruženim bolestima. Umrlo je 11 pacijenata (3.5%). Šok, rekrvarenje i udružene bolesti su identifikovani kao nezavisni, statistički značajni prediktori smrtnog ishoda. Napred navedeni način analiziranja pacijenata prema riziku od rekrvarenja i smrtnog ishoda i definisanje rizičnih faktora može da predstavlja zanačajan elemenat u proceni stanja svakog individualnog pacijenta i da kliničaru olakša donošenje odluke o daljem tretmanu posle inicijalne endoskopske injekcione hemostaze.

UVOD

Akutna krvarenja iz jednjaka, želuca i dvanaestopalačnog creva predstavljaju urgentno stanje u kliničkoj medicini, sa incidencom od 50 do 150 na 100.000 stanovnika za godinu dana (1,2). Bez obzira na uvođenje savremenih dijagnostičkih i terapijskih metoda, stopa rekrvarenja posle inicijalne hemostaze, a posebno smrtnost pacijenata, naizgled se nije značajno izmenila u poslednjih nekoliko decenija. Velike studije ukazuju da oko 4 do 14% pacijenata umire od posledica ili komplikacija krvarenja iz gornjih delova gastrointestinalnog trakta (3-5). Ako se uzme u obzir da je populacija generalno starija, i da je starija životna dob jedan od često isticanih faktora rizika za smrtni ishod posle krvarenja ili rekrvarenja iz gornjih delova digestivnog traktai, možemo smatrati da realno postoji diskretan trend poboljšanja stope mortaliteta (6,7). Rekrvarenje (ponovno krvarenje), posle inicijalno uspešno tretirane krvare}e lezije, smatra se glavnim faktorom rizika za smrtni ishod i javlja se u oko 10 do 30% pacijenata bez obzira na primenjeni metod hemostaze (8). Uopšteno je prihvaćeno stanovište da je rizik od rekrvarenja i fatalnog ishoda povezan sa više međusobno nezavisnih faktora.

I do 80% krvarenja iz gornjih delova digestivnog trakta je “self-limited” (samo-ograničavajuće), odnosno prestaje spontano, bez obzira da li je učinjena neka od endoskopskih hemostaznih procedura (9). Na osnovu anamnestičkih i kliničkih parametara teško je identifikovati, grupu pacijenata čije krvarenje neće spontano prestati. Američki National Institutes of Health (NIH) održao je 1989. godine Konsenzus koferenciju na kojoj je endoskopska terapija krvarećih ulkusa označena metodom izbora uz “izvanrednu efikasnost i prihvatljivo nizak nivo pratećih komplikacija (10).

MATERIJAL I METODE

U studiju je uključeno 315 pacijenata koji su zbog znakova krvarenja iz gornjih delova gastrintestinalnog trakta primljeni u Urgentni centar Kliničkog centra Srbije (UCKCS) u Beogradu u periodu od januara 1998. do marta 1999. godine. U studiju su uključeni samo pacijenti kojima je u toku inicijalne endoskopije uspešno učinjena endoskopska injekciona hemostaza krvareće lezije. U studiju nisu uključeni pacijenti koji su imali akutno krvarenje iz varikoziteta jednjaka. Pacijenti su po prijemu u UCKCS ispitani anamnestički, klinički i laboratorijski. Od anamnestičkih podataka analizirani su pol i starost pacijenata, način kliničke prezentacije akutnog krvarenja, podaci o uzimanju nesteroidnih antireumatika i acetil-salicilne kiseline i postojanje udruženih bolesti. Od kliničkih parametara posebna pažnja je usmerena na postojanje šoknog stanja. Arbitrarno je uzeto da su u šoku pacijenti koji imaju sistolni pritisak niži od 100 mmHg, a frekvenciju pulsa veću od 100/min. Od laboratorijskih parametara je ispitivan nivo hemoglobina, a kao granična vrednost je uzet nivo od 10 g/l. Gornja endoskopija je rađena na standardan način sa faringealnom anestezijom. Endoskopija je uvek rađena od strane iskusnog endoskopiste koji je na licu mesta samostalno i nezavisno odlučivao o potrebi i vrsti endoskopske skleroterapije. Intervencija je uvek započinjana anemizacijom zone ubrizgavanjem epinefrina (rastvor sa 0.9% NaCl u odnosu 1:10.000) u okolinu krvareće lezije. Potom je injektirano i sklerozantno sredstvo u samu leziju. Kao sklerozantno sredstvo je korišćen apsolutni-96% etanol. Ispiranjem lezije fiziološkim rastvorom verifikovano je postizanje konačne hemostaze krvareće lezije. U toku endoskopije registrovani su lokalizacija i vrsta lezije, kao i postojanje stigmata krvarenja. Kao mesto lezije označeno je sedam lokalizacija: jednjak, forniks želuca, korpus želuca, antrum želuca, prednji zid bulbusa duodenuma, zadnji zid bulbusa duodenuma i zona postoperativne anastomoze bez obzira na vrstu predhodne operacije. Krvareće lezije su podeljene u 5 grupa: ulkusi, maligniteti, Mallory-Weiss-ov sindrom, erozivna bolest gornjih delova digestivnog trakta-a i “ostale” lezije. U kategoriju “ostalih” lezija ubrojene su vaskularne malformacije i Dieulafoy-ova lezija. Stigmata krvarenja su grupisana prema modifikovanoj Forrest-ovoj klasifikaciji (11), pri čemu su, zbog dizajna studije, klasifikovane samo krvareće lezije koje su zahtevale endoskopsku injekcionu hemostazu: adherentni koagulum u dnu lezije, vidljiv krvni sud u dnu lezije, aktivno slivajuće krvarenje iz lezije i aktivno špricajuće krvavljenje iz lezije. Registrovane su epizode ponovnog krvarenja (rekrvarenja). Rekrvarenje je definisano kao nova epizoda krvarenja za vreme hospitalizacije koja se manifestvovala kao ponovna hematemeza, pojava sveže krvi u nazogastričnom aspiratu ili potreba za hiruškom intervencijom zbog nastavljanja krvarenja. Mortalitet je definisan kao smrtni ishod u toku perioda hospitalizacije (12). Registrovana je i operativna terapija i dužina hospitalizacije svih pacijenata.

REZULTATI

Studijom je obuhvaćeno 315 pacijenata: 210 (66.7%) muškaraca i 105 (33.3%) žena. Tabela 1 zbirno prikazuje karakteristike pacijenata sa osvrtom na ponovljeno krvarenje i smrtni ishod u toku hospitalizacije posle inicijalno uspešno izvedene hemostaze.

Rekrvarenje je registrovano kod 53 (16.8%) pacijenta, dok je 11 (3.5%) pacijenata umrlo u toku hospitalizacije. Operativno je lečeno 37 (11.7%) pacijenata. Većina pacijenata, 35 (94.6%) je operisana zbog nekontrolisanog rerkrvarenja. Preostala 2 pacijenta (5.4%) su operisani u pokušaju radikalnog lečenja tumora želuca. Operacija statistički značajno nije uticala na povećanu stopu smrtnosti u odnosu na neoperisane pacijente (p=0.127). Pojedinačno je analiziran svaki parametar u odnosu na rekrvarenje i smrtni ishod. Starost pacijenata ne utiče na rekrvarenje kao i na smrtni ishod (p>0.05). Pol pacijenta, nivo hemoglobina na prijemu, uzimanje ulcerogenih lekova i klinička prezentacija u trenutku krvarenja takodje statistički nisu značajno

uticali na ponovljeno krvarenje i smrtni ishod pacijenata. Statistički značajno veća stopa rekrvarenja je zapažena kod pacijenata koji su inicijalno bili u šoku u odnosu na pacijente koji nisu bili hipotenzivni i tahikardični na prijemu (p=0.006). Inicijalno stanje šoka je visoko statistički signifikantno bilo češće kod pacijenata koji su umrli u odnosu na grupu pacijenata koja je preživela (p=0.0001). Vrsta stigmata krvarenja je statistički značajno uticala na pojavu recidiva krvavljenja (p=0.018), ali ne i na smrtni ishod pacijenata (p=0.310). (Grafikon 1).

Udružene bolesti su visoko signifikantno uticale i na pojavu rekrvarenja (p=0.0001) i na smrtni ishod pacijenata u toku hospitalizacije (p=0.0001) (Grafikon 2). Epizoda rekrvarenja je statistički visoko signifikantno uticala na smrtni ishod analiziranih pacijenata (p=0.0001).

Analiza broja dana provedenih u bolnici (Tabela 2) pokazuje da je prosečna dužina hospitalizacije neoperisanih bolesnika, bez rekrvarenja i sa povoljnim ishodom bila 6 do 7 dana.

U studiji smo definisali parametre značajne za pacijenate koji su rekrvarili ili umrli u toku hospitalizacije. Logističkom regresijom smo analizirali pojedinačni uticaj svakog parametra u kontekstu svih parametara u odnosu na zadatu varijablu. Zavisna varijabla se u prvom slučaju odnosila na pojavu rekrvarenja, a u drugom na letalni ishod pacijenta. Prisustvo šoka na prijemu, vrsta stigmata krvarenja i udružene bolesti statistički su značajno doprinosili pojavi rekrvarenja posle inicijalne uspešne endoskopske hemostaze krvareće lezije. Analizom istih parametara u odnosu na letalni ishod pacijenata zapažena je statistička značajnost za: pojavu rekrvarenja, postojanje šoka na prijemu i stanje udruženih bolesti.

DISKUSIJA

Cook i saradnici prikazuju meta-analizu rekrvarenja i smrtnosti kod 2412 pacijenata i zaključuju da nije postojala statistički značajna razlika u pogledu stope rekrvarenja u zavisnosti od primenjene endoskopske tehnike za zaustavljanje krvarenja. Smrtnost pacijenata je bila statistički značajno redukovana samo u grupi koja je tretirana laserskom metodom (13). Lin i saradnici ukazuju da je kombinacija kontaktno-termalnog i injekcionog metoda najefikasnija u prevenciji rekrvarenja peptičkih ulkusa kao i u smanjenju potrebe za transfuzijama krvi kod pacijenata (14). Analize varijanse su ukazale da pojava rekrvavljenja povećava stopu smrtnosti za 5 do 16 puta što predstavlja uvod u smrtni ishod pacijenata. Savremene tehnike endoskopske hemostaze omogućavaju uspešno zaustavljanje krvarenja u visokom procentu slučajeva uz sledstveni, značajan pad stope rekrvarenja. Međutim, to ne vodi očekivanom snižavanju stope mortaliteta ovih pacijenata. Dve studije ukazuju da je sniženje stope rekrvarenja praćeno adekvatnim smanjenjem mortaliteta uz korišćenje laserske hemostaze (13,15). Većina autora zaključuje da rekrvarenje predstavlja značajan faktor koji determiniše smrtni ishod posle akutnog krvarenja iz gornjih delova digestivnog sistema, ali da nije i jedini faktor koji značajno utiče na finalni ishod (15). Danas se više ne postavlja pitanje da li endoskopski načiniti inicijalnu hemostazu, već koji optimalni tretman pojedinačnog pacijenta, nakon postignute hemostaza sprovesti (8). Mogućnost selektiranja pacijenata koji će rekrvariti i grupe pacijenata sa visokim rizikom za smrtni ishod predstavlja krajnji cilj niza studija, ali i dalje bez konačnog odgovora (8). U isto vreme otkrivanje pacijenata koji nemaju rizik ili imaju mali rizik za rekrvarenje predstavlja značajan cilj u pokušaju optimalizacije tretmana, ambulantnog lečenja i smanjivanja troškova hospitalizacije i lečenja.

U našoj grupi ukupno je rekrvarilo 53 (16.8%) pacijenta, a preminulo je 11 (3.5%) pacijenata. Ovi rezultati su u skladu sa većinom objavljenih rezultata u literaturi za primenjeni način injekcione hemostaze krvarećih lezija gornjih delova gastrointestinalnog trakta (4,13,17,19). Gotovo sve studije prikazuju nešto veći broj muškaraca nego žena sa akutnim krvarenjem iz gornjih delova digestivnog sistema. U veoma citiranoj Rockall-ovoj studiji koja se bavi analizom faktora koji značajno utiču na rekrvarenje i smrtni ishod, muški pol predstavlja veći rizik za rekrvarenja, dok je veća učestalost umiranja kod žena.(15) Većina autora navodi da je starija životna dob jedan od statistički značajnih faktora koji doprinose povećanoj stopi rekrvarenja i mortaliteta posle akutnog krvarenja iz gornjih delova digestivnog sistema (8,15,16). Najveći broj naših pacijenata je bio u starosnoj grupi izmedju 60 i 80 godina (47.6%). Nijedan pacijent mladji od 40 godina nije preminuo. Rezultati naše studije pokazuju da postoji blagi trend povećanja stepena rekrvarenja i smrtnosti sa povećanjem starosti pacijenata. Međutim, ove razlike nisu statistički značajne. U šoku je bilo 21.6% naših pacijenata, od kojih je 27.9% rekrvarilo, a 13.2% pacijenata je umrlo. Studija Brullet-a i saradnika (17) određuje šokno stanje pacijenta kao jedan od faktora koji značajno doprinosi neuspehu injekcione hemostaze krvarećih želudačnih ulkusa. Hipotenzija kao loš prognostički faktor apostrofirano je i u studiji Schiler-a i saradnika (16) iz “preendoskopske ere” o hematemezi i meleni, sa posebnim osvrtom na faktore koji utiču na ishod pacijenata. Nivo hemoglobina ispod 10 g/l Peura i sardnici (18) pronalaze kod 60.9% pacijenata koji su imali epizodu akutnog krvarenja iz gornjih delova digestivnog sistema. Slično našim nalazima, Pimpl (19) i Katchinski (4) ne označavaju nivo hemoglobina kao prognostički značajan faktor za rekrvarenje ili smrt pacijenta. U našoj studiji, nivo hemoglobina se nije statistički značajno razlikovao kod pacijenata koji su rekrvavili i/ili umrli u odnosu na ostale ispitanike. Prema našim podacima, rekrvarenje je približno podjednako zastupljeno u grupama pacijenata na terapiji ulcerogenim lekovima, ili bez terapije (16.9% i 16.7%, ponaosob). Slične podatke nalazimo i za smrtni ishod. Laslo i saradnici objavljuju veliku multicentričnu studiju o kliničkim aspektima krvarenja iz gornjih delova digestivnog sistema u odnosu na prethodno konzumiranje ulcerogenih lekova. Glavni zaključak ove studije je da ne postoje statistički značajne razlike u kliničkoj prezentaciji značajnih krvarenja iz gornjih delova digestivnog sistema, bez obzira na konzumiranje ovih lekova (20). Čitav niz studija ističe u prvi plan značaj udruženih bolesti za ishod ovih pacijenata (8,19,21,22). Većina ovih studija ne razvrstava udružene bolesti, već koristi univerzalnu skalu za “opštu” procenu udruženih bolesti. Korišćena je od mnogih prihvaćena ICED skala (Index of Co-Existing Disease, National Auxiliary Publication Service) (23). Naša studija jasno pokazuje da postojanje težih udruženih bolesti statistički značajno doprinosi pojavi rekrvarenja i letalnog ishoda (p=0.0001). Tako je u grupi pacijenata koji su imali teške, neposredno po život opasne udružene bolesti, polovina pacijenata imala epizodu rekrvarenja, dok je gotovo polovina (42.8%) pacijenata umrlo. Interesantno je zapažanje Boura u studiji o “saglasnosti endoskopista” (“interobserver agreement”) u označavanju istih stigmata krvavljenja kod peptičkog ulkusa. Vrlo visoki stepen saglasnosti je opisan kada je u pitanju slivajuće krvarenje, zadovoljavajući stepen saglasnosti je postignut za postojanje adherentnog koaguluma i vrlo nizak stepen saglasnosti kada je u pitanju špricajuće krvarenje i postojanje vidljivog krvnog suda u dnu lezije. Ova grupa autora zaključuje da limitiranje klasa stigmata u nekoliko kategorija i adekvatan trening endoskopista doprinose većoj preciznosti u opisivanju ovih lezija (25). Epizode rekrvarenja i letalni ishod u našoj studiji je najčešće povezan sa najozbiljnijom stigmatom krvarenja, tj. aktivnim špricajućim krvarenjem. Nasuprot tome, najmanje komplikacija se javljalo kod pacijenata koji su imali adherentni ugrušak u dnu krvareće lezije. Vrsta stigmata krvarenja je statistički značajno uticala na pojavu rekrvarenja, ali ne i na eventualni letalni ishod pacijenta. Najčešća rekrvarenja su registrovana iz lezija fundusa želuca (23.1%), dok se najrizičnija zona za krvarenje, zadnji zid bulbusa duodenuma, nalazila na trećem mestu po procentu rekrvarenja, iza fundusa i korpusa želuca. Najviše epizoda rekrvarenja je kod pacijenata koji su imali ulkuse, maligne tumore i arterio-venske malformacije krvnih sudova. Najveći rizik od umiranja su imali pacijenti koji su krvarili iz malignih tumora i angiodisplazija, a po lokalizaciji najrizičnije zone su korpus i fundus želuca (umrlo je po oko 8% pacijenata sa pomenutim lokalizacijama). Rockall i saradnici ukazuju da dijagnoza (uzrok krvarenja) statistički značajno utiče na ishod pacijenata, pri čemu ističu postojanje maligniteta kao najnepovollniji, a Mallory-Weiss-ov sindrom ka najbolji prognostički faktor (15). Pri analizi različitih parametara koji utiču na epizodu rekrvarenja i/ili smrtni ishod, operativno lečenje nismo uključili u analizu jer smatramo da je ishod operacije i postopeartivni tok uslovljen nizom činilaca koji ne moraju biti direktno povezani sa samim krvarenjem iz gornjih delova digestivnog sistema. Sličan postupak primenjuje većina autora (8,15,19,22,24,26). Procenat operisanih pacijenata zbog akutnog krvarenja iz gornjih delova digestivnog sistema varira u različitim serijama. Peura i saradnici prikazuju 1235 pacijenata od kojih je 7.1% operisano (18). Rockall navodi da je operativna stopa oko 5.7% (27), dok Everett prikazuje da je u grupi visokorizičnih pacijenata 8.8% operisano. Uočljivo je da je procenat operisanih pacijenata u svim opisanim studijama manji od 11.1% koji saopštavamo u našoj studiji. Rezultati pomenutih stidija su najverovatnije posledica efikasnije inicijalne hemostaze krvarećih lezija zbog korišćenja kombinovanog kontaktno-termalnog i injekcionog metoda u većini citiranih studija. Dužina hospitalizacije i potreba za hospitalizacijom svih pacijenata koji su akutno krvarili iz gornjih delova digestivnog sistema je kontroverzno pitanje. Rollhauser i Fleischer (28) u revijskom radu o ulkusima i nevariksnom krvarenju i korišćenju odgovarajućih smernica u tretmanu ovakvih pacijenata navode da je prosečna dužina hospitalizacije ovih pacijenata 2.9 do 4.6 dana u zavisnosti od stepena poštovanja napred navedenih smernica. Autori zaključuju da je kraći boravak u bolnici jasno korelirao sa ranije urađenom endoskopijom i identifikacijom stigmata i vrste krvareće lezije. Rockall i saradnici objavljuju studiju o mogućnosti selektiranja pacijenata za rani otpust iz bolnice u zavisnosti od vrednosti skora rizika za ponovno krvarenje, navodeći da je prosečno vreme provedeno u bolnici oko 8.6 dana, a da je broj dana za grupe sa malim rizikom gotovo upola manji.(21) Naši pacijenti koji su imali povoljan ishod lečenja i nisu rekrvarili provodili su u bolnici 6 do 7 dana što smatramo neopravdano dugim periodom.

Navedeni način analiziranja pacijenata prema riziku od rekrvarenja i smrtnog ishoda i definisanje rizičnih faktora može da predstavlja značajan elemenat u proceni stanja svakog individualnog pacijenta i može da doprinese konačnoj odluci kliničara o daljem tretmanu. S druge strane, ovo može biti doprinos efikasnijoj standardizaciji medicinskih procedura i kliničkih studija o terapijskoj endoskopiji, proceni postojećih algoritama u dijagnostikovanju, lečenju i optimalizaciji dužine hospitalizacije ovakvih pacijenata.