Propantheline Bromide in Massive Upper Gastrointestinal Haemorrhage

The effectiveness of the anticholinenergic, propantheline bromide, administered parenterally in patients with upper gastrwintestinal bleeding has been investigated in a double-blind study. The basic material consisted of 99 patients. No differences were found between the group receiving propantbeline bromide and the group receiving placebo in regard to clinical factors such as duration of intensive care, total hospital stay and surgical frequency. However, 21 patients under 50 years of age showed a significantly lower blood transfusion requirement with the use of prop-antheline bromide.


INTRODUCTION
The treatment of massive upper gastro-intestinal (G-I) bleeding is a current clinical problem. The borderline between medical and surgical treatment and the surgical method of choice has aroused great interest. An active attitude towards the surgical therapy at a relatively early stage seems to be a trend which can favourably affect the high mortality rate of these patients-especially the old ones (3,7). In emergency operations, however, the mortality is considerably higher than in elective surgery. Older patients with haemorrhagic gastritis especially have a high mortality in emergency surgery (9).
In some places it has been the practice to give anticholinergics to patients with upper G-I bleeding in order to inhibit the production of acid and hence the irritation of the gastric mucous membrane. However, the clinical effect of anticholinergic treatment in acute upper G-I bleeding has not been documented.
The purpose of this investigation was to evaluate the possible effects of an anticholinergic, propantheline bromide (Ercotina@; ERCO Lakemedel Patients with esophageal varices were omitted, but otherwise no further selection was made. All patients were treated primarily in the intensive care unit. Except for other treatment (see below) they received either propantheline bromide (PB) in doses of 180 mg per 24 hours intravenously, or placebo (P). 84 patients were divided into groups of PB (44 cases) and P (40 cases) in a doubleblind manner. Besides these, 15 consecutive patients received PB and this group neither diagnostically nor agewise differed from the others and received the same treatment. They were included in the schedule, the staff not knowing whether or not the ampoules contained propantheline bromide or placebo. The PB group thus consisted of a total of 59 cases (average age 64), 45 males (average age 62) and 16 females (average age 68). The P group comprised 40 cases (average age 61), 30 males and 10 females with the same average age (Fig. I). The diagnostic division into PB and P groups with the age distribution is shown in Table 1. One patient in the P group with the diagnosis of gastric cancer has been omitted.
Complicating illnesses such as heart-lung disease, diabetes, malignancy, and severe invalidism were found in 25 patients (42%) of the PB group and in 21 patients (52%) of the P group. Shock or preshock was found in the case histories of 37 PB patients (63%) and in 17 P patients (43%). On admission the Hb-value was 8.7 g per 100 ml. (S.D.=2.4) in both groups. G-I bleeding.
patients received phytomenadione intravenously (Konak-ion@; F. Hoffmann LaRoche & Co. AG, Basel, Switzerland) 20 mgx2 and ascorbic acid (ACO Lakemedel AB, Solna, Sweden) 2 g per 24 hours. A nasal gastric tube was inserted. X-rays were taken of the esophagus and stomach during the first 24 hours of care in all patients, when their condition so permitted. Emergency gastroscopy was performed on a smaller number of patients when the X-ray diagnosis was uncertain or negative and there was no need for emergency surgery.
The following indications for surgery were established (8): I. For patientsX0 years of age or with complicating illness 1. Primary shock controlled but repeated preshock or shock.
2. Primary shock controlled but continued bleeding which required more than I200 ml of blood during the following 24 hours.

.
No shock, but patient required 1600 ml of blood or

METHODS
Immediately upon admission to the emergency room the patients were treated with dextran 70 (Macrodex"; Pharmacia AB, Uppsala, Sweden). Blood samples were taken for blood typing, Hb, haematocrit, electrolyte balance, coagulation time, and routine liver tests. The patients were then conveyed to the intensive care unit. As soon as blood was available, transfusions in the form of whole blood or erythrocyte concentrate were given in order to control the shock as soon as possible, and to gradually obtain a Hb-value of around 10 g per 100 ml or a haematocrit of about 30%. Haemoglobin, haematocrit, and thrombocyte determinations were done a t least daily. A central venous catheter was inserted in 31 patients of the PB group and 17 of the P group. The central venous pressure was measured at least every 30 minutes during the acute phase of the haemomhage. After catheterization, pulmonary X-ray was performed. The arterial blood gases were determined. A complete ECG was recorded and the patients were kept continuously on ECG monitoring. PB or P were given in an invertose~lectrolyte solution. All more in 24 hours.
11. For patients40 years of age and otherwise essentially well 1. The bleeding had not stopped within 36 hours and the patient had received more than I 6 0 0 ml of blood in 24 hours.
2. Patient had required 2000 ml of blood within 24 hours or less and the bleeding had not stopped.

111.
Primary shock could not be controlled.
The following indications for surgery were followed (8): A. For patients over 60 years of age or with a com-

RESULTS
Of the 59 patients in the PB group, 10 (17%) died, and of the 40 in the P group, 4 (10%) died. All patients who did not survive were 65 years of age or older. In the PB group there were 35 within this age range while there were 21 in the P group. Thus the mortality in the former group was 29% and in the latter 19%. Among the dead, 7 of the PB group had severe complicating illness and 4 of the P group. All 10 deceased in the PB group had a history of shock, while this was the case in only 2 of the 4 deaths in the P group. Severe surgical complications occurred in 5 of the dead in the PB group but not in any of those in the P group (Table 11).
Emergency surgery was performed on 25 patients 0-9 days after admission to the hospital and all except 3 were operated upon within the first 48 hours. No patient died on the day of surgery. 1-7 days after surgery, 7 patients died and 1 after 18 days. In the PB group 5 deaths occurred out of 16 operated upon (37%) compared with 2 out of 9 in the P group (22%). Among the patients undergoing elective surgery there were no deaths. The average age for emergency surgery was 66 in the Pk group and 63 in the P group. Those undergoing elective surgery were somewhat younger, on average 60 years in the PB group and 57 in the P group.
Emergency surgery was performed with vag-otomy+pyloroplasty in 10 cases in the former group and in 6 cases in the latter, while gastric resection ad modum Billroth I1 was done in 6 PB patients and in 3 P patients. Later elective surgery was performed on 7 patients in the PB group and 8 patients in the P group (Table 111).
The stay in the intensive care unit for the nonsurgical cases was the same for both groups-25.6 hours on average for the PB group and 25.9 hours for the P group.
The amount of blood transfused during the first 24 hours was recorded. The total amount was also calculated for the non-surgical cases and for the preoperative period of the surgical cases. The average number of transfusions within the PB and P   groups in relation to diagnosis are shown in Table  IV and for surgical or non-surgical therapy in Table  V. Hb on admission was on an average 8.7 g per 100 ml in both groups. The lowest recorded Hb-values were, on average, 7.6 g per 100 ml in the PB group and 7.3 g per 100 ml in the P group. The highest recorded Hb was the same in both groups, namely 12.2 g per 100 ml. A tendency towards a smaller transfusion requirement was noticed in the PB group compared with the P group. However, statistical significance was not reached. Therefore, the severity of haemorrhage was evaluated, considering both the transfusion requirement and the Hbvalues recorded during the same time. Different age groups were analysed in this manner. In order to make adequate comparisons between the PB and P groups the amount of blood that was needed to maintain the same Hb-value during the first 24 hours was calculated. This calculation could be done as both groups were treated identically with regard to blood volume replacement. The patient's blood volume that was needed for the calculation was obtained from Nadler et al. (10). In Table VI 21 patients, 50 years and younger, are presented according to age, Hb-value on admission, lowest Hbvalue recorded and Hb-value after the first 24 hours of treatment. The amount of blood given during the first 24 hours is presented and also the calculated blood requirement for maintaining a constant Hbvalue during the same time. The PB and P groups were comparable with regard to age, Hb-value on admission and blood given (P>O.OS). Neither the frequency of patients with a history of shock nor their total hospital stay differed. However, the blood requirement for maintaining the same Hbvalue during the first 24 hours was significantly lower among patients in the PB group compared with those in the P group (P<O.OOl).
In Table VII the same variables as in the preced-I ing table are shown for the 12 patients 50-60 years of age. No differences were observed in any of the variables between the PB and P groups.

DISCUSSION
Massive upper G-I haemorrhage was especially observed among elderly persons and, among these, males seemed the most predisposed. The average age of the patients in this material was 63 years, without any great difference between the PB and P groups. In comparison with the extensive material ofBoulos et al.  (7), the average age in this investigation was about 10 years higher. However, this result is in agreement with a Swedish study on 200 patients with a similar age distribution (4). The average age was higher among the patients with gastric ulcer than those with duodenal ulcer and this agrees with the above-mentioned Swedish study.  Patients with duodenal ulcer represent 46% in both the PB and P groups. The number of duodenal ulcer patients was smaller in this investigation than in the studies of Boulos et al.  (7). In those materials the percentage of patients with duodenal ulcer was 70%, 55%, and 52% respectively. The reason for the smaller number of patients with duodenal ulcer in this series is probably the greater proportion of older patients.
Of the 14 patients who died in connection with the actual haemorrhage, 10 were PB cases and 4 were P cases. The causes of death in these patients are shown in Table 111. The primary cause of death was: septicaemia in 3 cases, cardiopulmonary complications in 5 cases, surgical complications in 5 cases, and the haemorrhage was judged as being the direct cause in 2 cases. In these latter 2 patients, surgery was indicated but they were judged, inoperable. Furthermore, it is considered in this study that the PB therapy could be excluded as having no relation to the cause of death in any of the 10 deaths occurring in the PB group.
The essential variables used for evaluation of the effect of PB were the severity of the haemorrhage as indicated by the transfusion requirement and the Hb values. The average value for the amount of blood transfused tends to be lower in the PB group than in the P group, and this tendency is especially noticeable among the patients with gastric and duodenal ulcer (Table IV). However, this trend for the amount of blood transfused was not statistically significant (Student's f-test).
Surgical vagotomy is a recognized operative procedure in a bleeding ulcer (5). Analogously, an effect by pharmacological vagotomy might also be expected. This should thus be observed in patients with a high acid production, i.e. in the younger patients (1). Furthermore, bleeding in older patients often occurs because of hardened, arteriosclerotic vessels, and in these cases one probably should not expect any great effect from anticholinergic therapy. As 66% of the patients were over 60 years of age in this investigation, it was decided to examine more particularly the records of those patients under 50 and 60 years of age in regard to the severity of the haemorrhage during the treatment. This was determined as the transfusion requirement during the first 24 hours to maintain a constant Hb (Tables VI and VII). The transfusion requirement in patients under 50 years of age was significantly lower in the PB group than in the P group (P<O.OOl) ( Table VI). However, no significant difference was observed between the 12 patients in the age group 51 to 60 (Table VII). Although this material is small, it seems that PB gives a transfusion-saving effect in the younger patients with upper G-I bleeding. The study indicates that the upper age limit for a therapeutic effect is about SO years. The asymmetry which was earlier suggested as unfavourable in regard to frequency of shock, complicating illness, and mortality in both groups has lost its significance in the younger patients. There now appears an equal numerical division between the PB and P groups. The smaller transfusion requirement for the patients less than SO years of age has not resulted in any differences in the necessary duration of intensive care, the fre-I n the patients over SO there was no evidence that PB has any significant effect. The blood-saving effect which PB seemed to give in the patients below SO years of age would probably in a larger study show itself in a decreased need for surgery. In those who did not require surgery the lower transfusion need is in itself a positive sign. Therefore, it is thought that propantheline bromide therapy may be indicated in patients below SO years of age with severe upper G-I haemorrhage.