<<

Gut 1996; 38: 11-14 11 Comparison of effects of calcium carbasalate and on gastroduodenal mucosal damage in

human volunteers Gut: first published as 10.1136/gut.38.1.11 on 1 January 1996. Downloaded from

F E Murray, N Hudson, J C Atherton, A T Cole, F Scheck, C J Hawkey

Abstract substantial morbidity and mortality from Calcium carbasalate is a therapeutically . Effervescent calcium active salicylate which seems to cause less carbasalate (ECC) is a calcium urea chelate of gastroduodenal mucosal damage than two aspirin molecules, formed by replacing the aspirin in laboratory animals. This endo- two molecules of water in crystalline calcium scopist-blinded, randomised, cross over acetyl salicylate by a molecule of urea. It forms trial aimed to compare acute gastric a stable and very soluble complex. In solution, mucosal damage in 20 healthy volunteers the complex immediately releases aspirin in the treated with acetyl salicylic (ASA) form of acetyl salicylate ion. When adminis- (650 mg three times daily) and efferves- tered orally, the acetyl salicylate is readily cent calcium carbasalate (ECC) (826.8 mg absorbed and has similar , , three times daily) bioequivalent to 650 mg and anti-inflammatory properties to aspirin. ASA over a five day period. Endoscopy Studies using intramucosal potential difference was performed immediately before treat- and gastrointestinal bleeding have suggested ment and on day 5 of each treatment. that ECC may be less damaging to the human Serum salicylate, B2, and gastrointestinal tract. 14 gastric mucosal E2 (PGE2) We now report a direct comparison of concentrations were measured after gastroduodenal injury, assessed endoscopi- endoscopy. ECC caused fewer gastric cally, caused by plain ASA for five days and mucosal erosions than ASA. The total effervescent calcium carbasalate (826.8 mg number of gastric erosions was 23.8 (16.1) three times daily, equivalent to 650 mg ASA in the ASA treated subjects compared three times daily for five days). with 9.1 (8.7) in ECC treated subjects (p=0.004). Differences between ASA and http://gut.bmj.com/ ECC were significant for both the gastric Patients and methods antrum and body, and for both haemor- This was a randomised, endoscopist-blind, rhagic and non-haemorrhagic erosions. cross over trial comparing the effects of effer- The mean gastric body Lanza score for vescent calcium carbasalate (ECC) and plain mucosal damage was lower after ECC ASA on gastroduodenal mucosal injury in than ASA (p=0.003). The visual analogue healthy volunteers. Healthy volunteers of score for gastric body damage was lower either sex, aged between 18 and 45, within on September 25, 2021 by guest. Protected copyright. for ECC (16.9 mm (15.9)) than for ASA 15% of their desirable weight (according to (32.7 mm (20.8)) p=0.008. Serum salicy- Metropolitan Life tables), were recruited. late concentrations were similar after After giving written informed consent, subjects both preparations (ASA: 66 (23) mgIl, were screened to enter this study. versus ECC: 58 (17) mgll, NS). Serum Subjects underwent a medical examination was similarly reduced with haematological and biochemical screen- using both preparations - 97.2 (3.5)% inhi- ing and a screening endoscopy before the start bition with ASA, 95.2 (5.5)% inhibition of the study. The main exclusion criteria with calcium carbasalate (NS). Suppres- included , peptic ulcer disease, aspirin sion of gastric mucosal PGE2 synthesis intolerance, and the presence of more than was similar with both preparations (ASA: three erosions at screening endoscopy. 83.4 (17.1)%; ECC 84.3 (12.9)%; NS). It is Volunteers were instructed to refrain from Division of concluded that ECC causes significantly (except oral contraceptives), , Gastroenterology, Department of less gastroduodenal mucosal damage than and spicy food during the study. They were Therapeutics, ASA administered at bioequivalent doses randomised to receive either 325 mg plain University Hospital, as judged by serum salicylate, serum ASA (Aspirin, ) or ECC containing 413 Nottingham NG7 2UH and mucosal values. F E Murray thromboxane, PGE2 mg acetyl salicylate (Redupsan, UPSA, Rueil N Hudson ECC may therefore be a less harmful Malmaison, France) first. During the active J C Atherton alternative treatment to plain ASA. treatment period, subjects were instructed to A T Cole 38: take two tablets 20 minutes before each meal F Scheck (Gut 1996; 11-14) C J Hawkey (breakfast, lunch, and dinner) for four days Keywords: calcium carbasalate, acetyl , and one dose on the fifth day. There was a Correspondence to: gastric mucosal damage. 16 to Professor C J Hawkey, wash out period of 23 days between Division of Gastroenterology, active treatments. University Hospital, Nottingham NG7 2UH. Before treatment and on day 5, subjects Accepted for publication Acetyl salicylic acid (ASA) causes acute underwent evaluation ofsymptoms and adverse 12 July 1995 gastroduodenal mucosal injury, resulting in events, endoscopic evaluation, and had blood 12 Murray, Hudson, Atherton, Cole, Scheck, Hawkey

TABLE I Modified Lanza samples taken from serum salicylate and significant. The results were expressed as mean scale thromboxane B2 (TxB2) measurement. (SD) (as well as mean and interquartile range Score Endoscopic appearances Endoscopy was performed without sedation. (IQR) for efficacy criteria) for quantitative The number of erosions (both haemorrhagic parameters and frequencies and percentages 0 Normal mucosa and non-haemorrhagic), intramucosal haemor- for qualitative parameters. Data regarding the

1 Mucosal erythema Gut: first published as 10.1136/gut.38.1.11 on 1 January 1996. Downloaded from 1 1-5 Erosions rhages, and superficial and deep ulcers in the Lanza scale are expressed as median (IQR). 3 6-10 Erosions 4 > 10 Erosions or an oesophagus, , and duodenum were The two randomised groups were compared ulcer counted. These data were used to derive a for demographic parameters using the modified Lanza score of mucosal damage Pearson's x2 test for qualitative parameters and (Table I). A visual analogue score (100 mm) of the Student's t test for quantitative parameters. mucosal damage was also used to evaluate The efficacy analysis was performed in three gastric antrum, body, and duodenal damage. steps. Firstly, the presence of residual effects Endoscopic biopsy was performed in the gastric was investigated using the paired sample t test body to determine PGE2 synthesis. After with- for quantitative parameters and the initial drawal of the endoscope, blood was taken for assessments at the first visit and the third visit estimation of serum salicylate by high perfor- (baseline evaluations before each sequence mance liquid chromatography (HPLC). Serum under active treatment). Secondly, within TxB2 and mucosal (PGE2) treatment effects were tested using the same were assayed as previously described.5 6 Briefly, tests to compare initial and final assessment mucosal biopsy specimens were washed in 0. 15 within each treatment, pooling the subjects of ml ofTris-saline and vortexed in 500 ,ul ofTris- the two sequences. Thirdly, the two treatments saline for one minute at room temperature. The were compared using a two way analysis of biopsy specimens were then removed and variance model (testing treatments effect, washed and the supematant stored at -70°C sequence effect and sequence by treatment until assayed. PGE2 was measured in unex- ) for quantitative parameters and a tracted supernatant using a validated specific modified x2 test (Cochran Mantel-Haenszel radioimmunoassay (RIA), and expressed as test) to compare the frequencies of qualitative pg/mg wet weight ofgastric mucosa. The serum parameters adjusting for the sequence. TxB2 concentration was measured in separated serum using RIA. Results

STATISTICAL METHODS PATIENT POPULATION From previous data, we estimated an average Twenty two volunteers were screened to enter total number of erosions with aspirin of 11.2 the study. One subject was not included in the

and aimed for the power to detect reduction of study because ofthe presence of a baseline bio- http://gut.bmj.com/ mucosal damage by two erosions (29% reduc- chemical abnormality (raised serum creati- tion a=0 05, 1-p=09). As a result of sample nine) and did not receive either treatment. One size calculation in log transformed values this other subject who had been randomised to led to 20 subjects being recruited to complete plain ASA dropped out before the end of the the study. first treatment period because she was unwill- Statistical analysis was done using the SAS ing to undergo further endoscopy. She did not

and the Statexact packages. All tests were two complain of any symptom while taking the on September 25, 2021 by guest. Protected copyright. tailed; p values below 0.05 were considered treatment and reported no adverse event. Thus 20 subjects (16 men and 4 women, TABLE II Comparison ofgastroduodenal damage associated withfive days' treatment with aged 23.95 (3.6) years (range 19-35), and acetyl salicylic acid (ASA) or effervescent calcium carbasatate (ECC) (values mean weight 73 kg (8)) completed the entire study (SD)) and were evaluated. Eleven subjects received Total no oferosions ASA as the first treatment and ECC as the ECC p Value second treatment; nine subjects received ECC ASA as the first treatment and ASA as the second Gastric body: Total erosions and ulcers 12-4 (9-7) 5-0 (7-3) 0-005 treatment. Haemorrhagic erosions 11-8 (9 6) 4-6 (6-9) 0.004 Non-haemorrhagic erosions 0-7 (1-5) 0-4 (1-0) Lanza score (IQR) 3 (2-4) 2 (0-2) 0-003 VAS (mm) 32-7 (20-8) 16-9 (15-9) 0-008 COMPARISON OF EFFECTS OF ASA AND ECC ON Gastric antrum: GASTRIC MUCOSAL DAMAGE Total erosions and ulcers 11-4 (9-4) 4-1 (4-1) 0-002 Haemorrhagic erosions 5-7 (6-9) 2-2 (2.7) 0-006 There was no difference in the number of Non-haemorrhagic erosions 5-5 (5-8) 1-9 (3-0) 0-01 erosions at baseline in the two groups. There Lanza score 3-5 (3-4) 2 (0-5-2) 0-0006 VAS (mm) 36-9 (19.3) 17-9 (15-7) 0-002 was no carry over from one treatment period to Duodenal bulb: the next. No effect on mucosa in Total erosions and ulcers 3-6 (4-9) 2-1 (4-4) NS oesophageal Non-haemorrhagic erosions 0-4 (0-8) 0-4 (1-4) NS terms of erythema or erosions was observed. Haemorrhagic erosions 3-2 (4-7) 1-6 (4-2) NS As previously observed, ASA caused a highly Lanza score (IQR) 1-5 (1-2-5) 0-5 (0-2) NS VAS (mm) 21-3 (23-9) 12-7 (22-9) NS significant and noticeable increase in the total Gastric body+antrum (pooled): number of erosions and ulcers Total erosions and ulcers 23-8 (16-1) 9-1 (8-7) 0-004 gastric (haemor- Haemorrhagic erosions 17-5 (14-3) 6-8 (7-2) 0-007 rhagic and non-haemorrhagic) in the stomach Non-harmorrhagic 6-2 (6-3) 2-3 (3-4) 0-02 as a from 0 to 23.8 Gastroduodenal (pooled): whole, (16.1) (p<0-005) Total erosions and ulcers 23-7 (17-9) 11-2 (10-0) 0-009 (Table II). This increase was seen in both the Haemorrhagic erosions 17-8 (14-2) 7-2 (7-4) 0-008 0 to 12-4 and the 9.4 4-0 0-04 body (from (9.7) (p=0.0001) Non-haemorrhagic erosions (9-4) (5-5) antrum (from 0.1 to 11*4 (9-4)) (p=O00001) VAS=visual analogue scale. (Table II). For ECC, the total number of Calcium carbasalate and aspirin in gastroduodenal mucosal damage 13

TABLE III Comparison ofeffects ofacetyl salicylic acid (ASA) and effervescent calcium were pooled there was a significant difference carbasatate (ECC) on prostaglandin E2 (PGE2) and thromboxane (TxBJ concentrations in favour of ECC for the total number of ASA ECC p Value erosions (p=0.01), haemorrhagic erosions (p= 0.008), non-haemorrhagic erosions (p=004), Day 5 serum salicylate (mg/l) 66 (23) 58 (17) NS

Day 5 serum TxB2 (pg/mi) 357 (376) 220 (55) NS and the total number of erosions and ulcers Gut: first published as 10.1136/gut.38.1.11 on 1 January 1996. Downloaded from Inhibition of serum TxB2 97/2% 95-2% NS (p=0009). One subject developed a super- Inhibition of gastric mucosal PGE2 83-4 (17 1)% 84-3 (12-9)% NS ficial ulcer in the duodenal bulb while being treated with ASA and ECC but this was not gastric erosions and ulcers increased from 0 to associated with Hpylori. 9.1 (8.7) for the stomach as a whole (p=0O00001). These increases were seen in both body (0 to 5.0 (7.3), p=0-0001) and the EFFECTS OF ASA AND ECC ON SERUM antrum (0 to 4X1) (4-1), p=0-0001) (Table II). SALICYLATE, TXB2 AND PGE2 VALUES The total number of erosions was significantly The mean serum salicylate concentration lower with ECC than ASA whether the whole measured on day 5 with each treatment was stomach (p=0.004), the body (p=0.005), or available in 19 subjects with ASA and 18 sub- the antrum (p=0.002) was considered (Table jects with ECC. Serum salicylate values II). Most of the gastric erosions observed were measured after treatment with ASA and ECC haemorrhagic. There were significantly fewer were similar (66 (23) mg/l versus 58 (17) mg/l, haemorrhagic erosions with ECC than with NS): Table III. There was a significant sup- ASA in the antrum (p=0 006), in the body pression in the mean serum TxB2 level after (p=0.004), and in the stomach as a whole five days treatment with both ASA (30873.75 (p=0.007). Antral ulcers were observed in (31686.05) pg/ml to 375-25 (37642); three subjects (one deep, two superficial) while p=0002) and ECC (24238.67 (3327099) to receiving ASA; all had a negative CLO test 22040 (55.26) pg/ml; p=0014); the percent- (Delta West, Perth, Western Australia) ages of suppression from day 1 were 97.2% suggesting that they were Helicobacter pylori and 95-2% respectively. There was no signifi- negative. No gastric ulcers were seen in the cant difference between the two treatments for subjects while receiving ECC. this criterion (Table III). Both treatments also When data were translated into Lanza scores caused significantly reduced gastric mucosal for the stomach, there were significant differ- synthesis of PGE2 (Table III). ences between the two treatments (p=0.003 for gastric body and 0.0006 for antrum) (Table II). Visual analogue scores for gastric body and ADVERSE EVENTS antrum mucosal damage were also significantly Eleven subjects reported at least one adverse lower after treatment with ECC (16.9 (15.9) event while taking ASA, compared with nine http://gut.bmj.com/ and 17.9 (15.7) respectively) than with ASA subjects while taking ECC. Ten subjects taking (32.7 (20.8) and 36.9 (19-3) respectively); ASA experienced compared with p=0O008 and p=0.002 respectively. seven taking ECC. Nausea was reported in three subjects on ASA and in two on ECC. Heartburn was reported in two subjects on EFFECTS OF ASA AND ECC ON DUODENAL ASA and in one on ECC. Vomiting was noted MUCOSA in one subject with each , and anorexia in on September 25, 2021 by guest. Protected copyright. The damage noted in the duodenal bulb was one subject while taking ASA. Adverse reac- less severe than in the stomach. ASA increased tions were graded as mild except for two the total number of erosions in the duodenal patients on ASA and one on ECC in whom bulb from 0.05 (0) to 3'6 (4.9) (p=0-001). they were considered moderate. There was no This was accompanied by a deterioration in significant difference between the two drugs in the Lanza score from 0 to 1-5 (IQR 1-2.5) this regard. Three subjects receiving ASA (p=00003) and the visual analogue score developed gastric antral ulceration. In all cases, from 0-65 to 21.30 (23.9) mm (p=0.001) ulceration healed spontaneously after stopping (Table II). The numbers of both non-haemor- treatment. No gastric ulcers developed in rhagic erosions (0.04 (0.8)) (p=0.002) and subjects receiving ECC. One subject devel- haemorrhagic erosions (0 to 3.2 (4.7)) oped a superficial duodenal bulb ulceration (p=004) were increased. with both treatments. The administration of ECC induced a small but significant (p=005) increase in the total number of erosions from 0 to 2.1 (4.4), a Discussion worsening of the Lanza score (0 to 0 5 (IQR In this study, ASA and ECC had similar 0-2) (p=0.002), and an increase in the activity as judged by similar serum salicylate number of non-haemorrhagic erosions in the concentrations and inhibition of serum TxB2 duodenum bulb (Table II). With ECC there and gastric mucosal PGE2. Despite this, ECC was no significant increase in haemorrhagic caused significantly less acute gastric mucosal erosions in the duodenal bulb, and there was injury in subjects compared with ASA. In addi- no significant deterioration in any criteria in tion, three subjects developed gastric ulcera- the second part of the duodenum (data not tion while taking ASA compared with none shown). Differences in mucosal damage taking ECC. These differences persisted when between ASA and ECC in the duodenum did subgroups with erosions classified according to not achieve statistical significance. whether they were of haemorrhagic or non- When whole stomach and duodenum data haemorrhagic appearance and whether they 14 Murray, Hudson, Atherton, Cole, Scheck, Hawkey

were seen in both the body and the antrum of Whether the reduction in mucosal damage stomach. There were relatively few erosions in observed is a result of buffering of salicylate by the duodenum but there was a consistent trend urea in ECC or of cytoprotective action of towards less damage when subjects received ECC is not known. Other possibilities are that

ECC compared with ASA. aspirin is released from the complex only very Gut: first published as 10.1136/gut.38.1.11 on 1 January 1996. Downloaded from These data suggest that ASA and ECC are slowly so that exposure of the gastric mucosa is therapeutically equivalent but that ECC causes effectively limited. However, ECC caused less less acute gastric injury as assessed by the mucosal damage in the presence of similar serum salicylate concentration, inhibition of levels of inhibition of mucosal PGE2 synthesis serum TxB2 values, and inhibition of mucosal with each compound. PGE2 synthesis. If the differences were to In summary, ECC in doses therapeutically persist with longer term ingestion, this would equivalent to ASA in terms of inhibition of suggest that ECC is less ulcerogenic with serum and mucosal synthesis, chronic use. During the study no gastric ulcers caused less gastric mucosal damage than ASA. developed in subjects taking ECC, compared with three ulcers with ASA. Our study evalu- ated relatively high doses ofboth drugs. ASA is 1 Florent C, Flourie B, Pfeiffer A, Bernier JJ. Mesure de al difference de potentiel transepitheliate gastrique. Effects used at lower doses for cardiovascular prophy- des agents agressifs (aspirine, alcool) et des pansements laxis but it is known that even doses as low as gastriques. Gastroenterol Clin Biol 1985; 9: 58-64. 2 Muir A, Cossar I. Aspirin and gastric bleeding: further 75 mg/d are associated with easily detectable studies of calcium aspirin. Am J Dig Dis 1961; 6: 115-25. gastric mucosal injury. If the differences 3 Pearson RN, Holt PR, Watson RM, Keating RP. Aspirin and gastrointestinal bleeding. Am J Med 1961; 31: 259-65. between ASA and ECC were also seen at lower 4 Leonard JR, Levy G. Reduction or prevention of aspirin- doses, ECC might prove to be a more appro- induced occult gastrointestinal blood loss in man. Clin Pharmacol Therap 1969; 10: 571-5. priate agent for cardiovascular prophylaxis. 5 Hawkey CJ. Synthesis of prostaglandin E2, thromboxane B2, Data emerging from epidemiological studies and prostaglandin catabolism in and gastric ulcer. Gut 1986; 27: 1484-92. suggest a 40 to 80% increase in the incidence 6 Hawkey CJ, Hawthorne AB, Hudson N, Cole AT, Mahida of bleeding peptic ulcer in patients taking YR, Daneshmend TK. Separation of the impairment of haemostasis by aspirin from mucosal injury in the human aspirin as cardiovascular prophylaxis compared stomach. Clin Sci 1991; 81: 565-73. with placebo.7 7 Steering Committee ofthe Physicians Health Study Research Group. Final report of the aspirin component of the Our data do not identify the mechanism by outgoing physicians health study. N Engl Med 1989; 321: which ECC is less injurious than ASA. 129-35. http://gut.bmj.com/ on September 25, 2021 by guest. Protected copyright.