- Email: [email protected]
Gastric safety and entericcoated aspirin
THE LANCET
Sensitivity plate of P aeruginosa
reproduced with other polymyxins, chlorhexidine, and, to a lesser extent, vancomycin. Because all these agents act on the cell membrane we thought that some slight damage to an otherwise abnormal or impermeable cell membrane might be necessary to allow passage of nutrients into the cell. We were unable to find any evidence of previous polymyxin treatment, but because the patient had had an indwelling urinary catheter for some time it is possible that chlorhexidine bladder washouts had been used, leading to selection of the organism we isolated. I do not think that in our case we had witnessed the emergence of a superbug. On the contrary, it was unable to exist without the presence of a low concentration of a rarely used antibiotic. We have not isolated antibiotic-dependent P aeruginosa since that time. Laurence Hayeh Department of Pathology, Torbay Hospital, Torquay TQ2 7AA. UK
1 Farrag N, Eltringham I, Liddy H. Vancomycin-dependent Enterucuccars faecalis. Laizcet 1996; 348: 1581-82.
SIR-such isolates as Farrag and colleagues report’ might be more widespread or develop more frequently than recent publications s ~ g g e s t . We ~~’ have isolated a vancomycin-resistant and teicoplanin-resistant nutritionallydependent Enterococcus spp from the urine of a 54-year-old woman who presented with acute lymphocytic leukaemia and pneumonia in September, 1996. She was given cotrimoxazole 960 mg thrice weekly as prophylaxis against Pneumocystis carinii infection. While neutropenic during induction chemotherapy, this patient had three episodes of fever and infection: a Hickman catheter-tunnel infection and two Staphylococcus epidemidis central-line infections. She was treated with gentamicin, pipericillinitazobactam, and teicoplanin. Three courses of teicoplanin were given for a total of 32 days. During her second course of
430
teicoplanin, vancomycin-resistant and teicoplanin-resistant enterococci were isolated from urine and she was treated with nitrofurantoin. 3 weeks later, while receiving her third course of teicoplanin, urine yielded vancomycindependent and teicoplanin-dependent enterococci. Nitrofurantoin therapy was instituted and teicoplanin discontinued. The patient’s condition was complicated by the presence of urinary tract infection with Escherichia coli 4 days before and 5 days after the recovery of the glycopeptide-dependent enterococci. Subsequently E coli was also isolated from her blood. The glycopeptide-dependent enterococcal isolate failed to grow on blood agar, cysteine lactose electrolytedeficient agar, diagnostic sensitivity test agar, or McConkey agar unless these media were supplemented with vancomycin, teicoplanin, or the patient’s urine. New York City medium (which contains the antimicrobial agents vancomycin, colistin, trimethoprim, and amphotericin) supported the growth of this isolate. A previous study of glycopeptidedependent enterococci reported an isolate that was dependent on vancomycin only.* The present isolate was resistant to both vancomycin and teicoplanin, suggesting a VanA phen~type.~ This isolate has not been characterised fully, but we concur with Farrag and colleagues’ concern that current routine microbiological practice may fail to detect h c h nutritionally-dependent isolates, especially from blood specimens. These specimens can pose a particular problem because, unlike urine specimens that may provide the factor(s) necessary to support growth of glycopeptide-dependent isolates, most blood-culture protocols require specimen dilution in the blood-culture medium, a procedure that could result in the presence of insufficient glycopeptide to allow growth of such dependent isolates. *Angela S Rossney, Samuei I McConkey, Conor T Keane Department of Clinical Microbiology,
St James’s Hospital, Dublin 8, Ireland 1 Farrag N, Eltringham I, Liddy H. Vancomycin-dependent Enterucuccus faecalis. Lancet 1996; 348: 1581-82. 2 Fraimow HS, Jungkind DL, Lander DW, Urinary tract infection Delso DR, Dean JL. with an Enterucoccusfaecalis isolate that requires vancomycin for growth. A n n Int Meed 1994; 121: 22-26. 3 Rosato A, Pierre J, Billot-IUein D, Buu-Hoi A, Gutmann L. Inducible and constitutive expression of resistance to glycopeptides and vancomycin dependence in glycopeptide-resistant Enterucuccus avium. Antimicrub Agents Chemuther 1995; 39: 830-33.
Gastric safety and entericcoated aspirin SIR-I
1 Kelly JP,Kaufman DW, Jurgelon JN, Sheehn J, Koff RS, Shipiro S. Risk of aspirin-associated major upper
Vol349
*
February 8, 1997
THE LANCET
2
3
4
5
gastrointestinal bleeding with enteric or buffered product. Lancet 1996; 348: 1314-1 6 , Petroski D. Endoscopic comparison of three aspirin preparations and placebo: low dose prophylactic aspirin effects on gastroduodenal mucosa. 57th annual scientific meeting, American College of Gastroenterology, 1992: poster 142. Petroski D. Endoscopic comparison of three aspirin preparations and placebo. Clin Ther 1993; 15: 314-20. Petroski D. Endoscopic comparison of various preparations: gastric mucosal adaptability to aspirin restudied. Curr Ther Res 1989; 45: 945-54. Savon J, Allen ML, DiMarino AJ. Gastrointestinal blood loss with low-dose (325 mg) plain and enteric-coated aspirin administration. A m J Gastroenterol 1995; 90: 581-85.
Authors’ reply SIR-In response to Petroski’s comments, we did not suggest that enteric-coated or buffered aspirin carries a higher risk of upper gastrointestinal bleeding than plain aspirin. Nor did we include individuals who had already had an episode of upper gastrointestinal bleeding. We disagree that the number of cases (550) was small: ours is one of the largest published epidemiological studies of upper gastrointestinal bleeding. Use of high doses of enteric-coated aspirin was infrequent in our study population, and we lacked the data to evaluate its association with upper gastrointestinal bleeding at doses greater than 325 mg. The potential biases Petroski mentions were considered in our report, as were disparities between our findings and those of those of endoscopic studies. Since ours is the first observational study, we emphasise, as we state in the report, that our results need to be confirmed by others.
due to inhibition of gastric cycloxygenase (with blockade of gastric prostaglandin production) after the drug is absorbed, irrespective of route of administration or formulation given. Prophylaxis against aspirinassociated gastroduodenal damage, even when enteric-coated preparations are used, is necessary. H,-receptor antagonists can prevent duodenal (but not gastric) lesions by non-steroidal anti-inflammatory drugs, but specific data concerning aspirin are not available. By contrast, co-prescription of misoprostol can protect both stomach and duodenum from the harmful effects of NSAIDs. The protective effect of misoprostol against low-dose aspirin has been described by UI< investigators: 100 pg daily misoprostol was better than placebo in preventing gastric haemorrhagic lesions induced by a 4-week course with aspirin 300 mg daily.5These data suggest that high-risk patients with strong indications for antiplatelet treatment with low-dose aspirin may benefit from prophylactic intake of a very low dose of the prostaglandin analogue. Mario Guslandi Gastroenterology Unit, S Raffaele Hospital, 20132 Milano. Italy
1 Kelly JP, Kaufman DW, Jurgelon JM,
2
3
*Judith Parsells Kelly, David Kaufman, Samuel Shapiro Boston University School of Medicine, Slone Epidemiology Unit, Brookline, MA 02146, USA
SIR-Kelly and colleagues’ report that low doses of enteric-coated or buffered aspirin can promote upper gastrointestinal bleeding to the same extent as plain aspirin. Although gastric damage by aspirin is not prevented by buffering, the comparative safety of enteric-coated preparations has been controversial because of endoscopic studies reporting fewer gastric erosions and less bleeding than with regular aspirin. The enteric-coated drug may be less harmful to the stomach during very short-term administration since there may be little or no contact with the gastric mucosa, but untoward gastric effects are likely after repeated admini~tration?~ Aspirin is thought to injure the gastric mucosa through local and systemic effects, the latter being
-
Vo1349 February 8, 1997
4
5
Sheehan J, Koff RS, Shapiro S. Risk of aspirin-associated major upper gastrointestinal bleeding with enteric coated or buffered product. Lancet 1996; 348: 1413-1 6. Jaszewski R. Frequency of gastroduodenal lesions in asymptomatic patients on chronic aspirin or nonsteroidal antiinflammatory drug therapy. J Clin Gastroenterul 1990; 12: 10-12. Savon Jj,Allen ML, di Marino AJ, Hermann GA, I h m RP.Gastrointestinal blood loss with low dose (350 mg) plain and enteric-coated aspirin administration. A m J Gastruenterul1995; 90: 581-85. Guslandi M. Gastric toxicity of anti-platelet therapy with low-dose aspirin. Drugs 1997; 53: 1-5. Goddard AF, Donnelly MT, Filipowicz B, Morant SV, Shield MJ, Hawkey CJ. Low dose misoprostol as prophylaxis against low dose aspirin-induced gastroduodenal mucosal injury. Gut 1996; 39: A33.
SIR-with prolonged daily use of aspirin, gastrointestinal loss of iron is likely, even when bleeding is so occult as to be detectable only by sensitive methods. A perspective on aspirininduced iron loss is gained by comparison with menstruation, which has undeniable impact on iron status. Iron lost from regular aspirin use can cumulatively exceed usual menstrual losses. Averaged over the month, typical menstrual blood losses amount to 1-2 mL per day. One aspirin (325 mg) tablet can cause 1-10 mL of
bleeding from the stomach, with much individual variati0n.l Adults without chronic blood loss store 300-1000 mg of iron, and blood contains about 0.5 mg of iron per mL. With aspirinassociated iron losses of 0.5-5 mg per day, iron stores could be exhausted in months, depending on the initial size of the iron load and the rate of blood loss. Induction of iron depletion may be an important second mode of action of aspirin used long-term, rather than an undesirable side-effect. Loss of stored iron is an increasingly likely mechanism of protection against heart diseaseZand ~ a n c e r .For ~ the protective action of aspirin against cancer, the most plausible previously proposed mechanisms remain essentially speculative.4Prolonged use of aspirin is thought necessary for protection against cancer, and there is an apparently increasing effectiveness against heart disease with duration of treatment. These patterns are consistent with an iron-loss mechanism since they suggest a cumulative protective effect (progressive lessening of iron stores over time). The benefits of daily aspirin Symmons5 discusses in his Nov 23 Commentary were all associated with treatments that do not eliminate gastrointestinal bleeding. Is a daily aspirin beneficial in spite of, or because of gastrointestinal blood loss? Studies are uninformative on this fundamental question. The clinical trials were inescapably a test of the global effects of aspirin. What is the empirical basis for the conclusion that benefit is only a direct result of intravascular platelet inhibition and not a consequence of occult bleeding? Efforts to minimise bleeding are founded not on empirically based conclusions, but on the assumption that aspirin-associated bleeding is an undesirable side-effect under all circumstances. A new aspirin designed to be harmless to the stomach may turn out to be a less effective aspirin. Without new trials of efficacy there is a danger that ineffectiveness might never be detected if so-called safe versions of aspirin are prescribed for prevention of heart disease and cancer. Until the contributions of all possible mechanisms are known, those with aspirin-induced iron depletion should be cautioned against supplemental iron. Jerome L Sullivan Pathology and Laboratory Medicine Service, Veterans Affairs Medical Center, Charleston, SC 29401. USA
1 Graham DY, Smith JL. Aspirin and the stomach. Ann Intern Med 1986; 104: 390-98. 2 Sullivan JL. Iron ns cholesterol. Perspectives on the iron and heart disease debate. J Clin Epzdenziol 1996; 49: 1345-52.
43 1
Sensitivity plate of P aeruginosa
reproduced with other polymyxins, chlorhexidine, and, to a lesser extent, vancomycin. Because all these agents act on the cell membrane we thought that some slight damage to an otherwise abnormal or impermeable cell membrane might be necessary to allow passage of nutrients into the cell. We were unable to find any evidence of previous polymyxin treatment, but because the patient had had an indwelling urinary catheter for some time it is possible that chlorhexidine bladder washouts had been used, leading to selection of the organism we isolated. I do not think that in our case we had witnessed the emergence of a superbug. On the contrary, it was unable to exist without the presence of a low concentration of a rarely used antibiotic. We have not isolated antibiotic-dependent P aeruginosa since that time. Laurence Hayeh Department of Pathology, Torbay Hospital, Torquay TQ2 7AA. UK
1 Farrag N, Eltringham I, Liddy H. Vancomycin-dependent Enterucuccars faecalis. Laizcet 1996; 348: 1581-82.
SIR-such isolates as Farrag and colleagues report’ might be more widespread or develop more frequently than recent publications s ~ g g e s t . We ~~’ have isolated a vancomycin-resistant and teicoplanin-resistant nutritionallydependent Enterococcus spp from the urine of a 54-year-old woman who presented with acute lymphocytic leukaemia and pneumonia in September, 1996. She was given cotrimoxazole 960 mg thrice weekly as prophylaxis against Pneumocystis carinii infection. While neutropenic during induction chemotherapy, this patient had three episodes of fever and infection: a Hickman catheter-tunnel infection and two Staphylococcus epidemidis central-line infections. She was treated with gentamicin, pipericillinitazobactam, and teicoplanin. Three courses of teicoplanin were given for a total of 32 days. During her second course of
430
teicoplanin, vancomycin-resistant and teicoplanin-resistant enterococci were isolated from urine and she was treated with nitrofurantoin. 3 weeks later, while receiving her third course of teicoplanin, urine yielded vancomycindependent and teicoplanin-dependent enterococci. Nitrofurantoin therapy was instituted and teicoplanin discontinued. The patient’s condition was complicated by the presence of urinary tract infection with Escherichia coli 4 days before and 5 days after the recovery of the glycopeptide-dependent enterococci. Subsequently E coli was also isolated from her blood. The glycopeptide-dependent enterococcal isolate failed to grow on blood agar, cysteine lactose electrolytedeficient agar, diagnostic sensitivity test agar, or McConkey agar unless these media were supplemented with vancomycin, teicoplanin, or the patient’s urine. New York City medium (which contains the antimicrobial agents vancomycin, colistin, trimethoprim, and amphotericin) supported the growth of this isolate. A previous study of glycopeptidedependent enterococci reported an isolate that was dependent on vancomycin only.* The present isolate was resistant to both vancomycin and teicoplanin, suggesting a VanA phen~type.~ This isolate has not been characterised fully, but we concur with Farrag and colleagues’ concern that current routine microbiological practice may fail to detect h c h nutritionally-dependent isolates, especially from blood specimens. These specimens can pose a particular problem because, unlike urine specimens that may provide the factor(s) necessary to support growth of glycopeptide-dependent isolates, most blood-culture protocols require specimen dilution in the blood-culture medium, a procedure that could result in the presence of insufficient glycopeptide to allow growth of such dependent isolates. *Angela S Rossney, Samuei I McConkey, Conor T Keane Department of Clinical Microbiology,
St James’s Hospital, Dublin 8, Ireland 1 Farrag N, Eltringham I, Liddy H. Vancomycin-dependent Enterucuccus faecalis. Lancet 1996; 348: 1581-82. 2 Fraimow HS, Jungkind DL, Lander DW, Urinary tract infection Delso DR, Dean JL. with an Enterucoccusfaecalis isolate that requires vancomycin for growth. A n n Int Meed 1994; 121: 22-26. 3 Rosato A, Pierre J, Billot-IUein D, Buu-Hoi A, Gutmann L. Inducible and constitutive expression of resistance to glycopeptides and vancomycin dependence in glycopeptide-resistant Enterucuccus avium. Antimicrub Agents Chemuther 1995; 39: 830-33.
Gastric safety and entericcoated aspirin SIR-I
1 Kelly JP,Kaufman DW, Jurgelon JN, Sheehn J, Koff RS, Shipiro S. Risk of aspirin-associated major upper
Vol349
*
February 8, 1997
THE LANCET
2
3
4
5
gastrointestinal bleeding with enteric or buffered product. Lancet 1996; 348: 1314-1 6 , Petroski D. Endoscopic comparison of three aspirin preparations and placebo: low dose prophylactic aspirin effects on gastroduodenal mucosa. 57th annual scientific meeting, American College of Gastroenterology, 1992: poster 142. Petroski D. Endoscopic comparison of three aspirin preparations and placebo. Clin Ther 1993; 15: 314-20. Petroski D. Endoscopic comparison of various preparations: gastric mucosal adaptability to aspirin restudied. Curr Ther Res 1989; 45: 945-54. Savon J, Allen ML, DiMarino AJ. Gastrointestinal blood loss with low-dose (325 mg) plain and enteric-coated aspirin administration. A m J Gastroenterol 1995; 90: 581-85.
Authors’ reply SIR-In response to Petroski’s comments, we did not suggest that enteric-coated or buffered aspirin carries a higher risk of upper gastrointestinal bleeding than plain aspirin. Nor did we include individuals who had already had an episode of upper gastrointestinal bleeding. We disagree that the number of cases (550) was small: ours is one of the largest published epidemiological studies of upper gastrointestinal bleeding. Use of high doses of enteric-coated aspirin was infrequent in our study population, and we lacked the data to evaluate its association with upper gastrointestinal bleeding at doses greater than 325 mg. The potential biases Petroski mentions were considered in our report, as were disparities between our findings and those of those of endoscopic studies. Since ours is the first observational study, we emphasise, as we state in the report, that our results need to be confirmed by others.
due to inhibition of gastric cycloxygenase (with blockade of gastric prostaglandin production) after the drug is absorbed, irrespective of route of administration or formulation given. Prophylaxis against aspirinassociated gastroduodenal damage, even when enteric-coated preparations are used, is necessary. H,-receptor antagonists can prevent duodenal (but not gastric) lesions by non-steroidal anti-inflammatory drugs, but specific data concerning aspirin are not available. By contrast, co-prescription of misoprostol can protect both stomach and duodenum from the harmful effects of NSAIDs. The protective effect of misoprostol against low-dose aspirin has been described by UI< investigators: 100 pg daily misoprostol was better than placebo in preventing gastric haemorrhagic lesions induced by a 4-week course with aspirin 300 mg daily.5These data suggest that high-risk patients with strong indications for antiplatelet treatment with low-dose aspirin may benefit from prophylactic intake of a very low dose of the prostaglandin analogue. Mario Guslandi Gastroenterology Unit, S Raffaele Hospital, 20132 Milano. Italy
1 Kelly JP, Kaufman DW, Jurgelon JM,
2
3
*Judith Parsells Kelly, David Kaufman, Samuel Shapiro Boston University School of Medicine, Slone Epidemiology Unit, Brookline, MA 02146, USA
SIR-Kelly and colleagues’ report that low doses of enteric-coated or buffered aspirin can promote upper gastrointestinal bleeding to the same extent as plain aspirin. Although gastric damage by aspirin is not prevented by buffering, the comparative safety of enteric-coated preparations has been controversial because of endoscopic studies reporting fewer gastric erosions and less bleeding than with regular aspirin. The enteric-coated drug may be less harmful to the stomach during very short-term administration since there may be little or no contact with the gastric mucosa, but untoward gastric effects are likely after repeated admini~tration?~ Aspirin is thought to injure the gastric mucosa through local and systemic effects, the latter being
-
Vo1349 February 8, 1997
4
5
Sheehan J, Koff RS, Shapiro S. Risk of aspirin-associated major upper gastrointestinal bleeding with enteric coated or buffered product. Lancet 1996; 348: 1413-1 6. Jaszewski R. Frequency of gastroduodenal lesions in asymptomatic patients on chronic aspirin or nonsteroidal antiinflammatory drug therapy. J Clin Gastroenterul 1990; 12: 10-12. Savon Jj,Allen ML, di Marino AJ, Hermann GA, I h m RP.Gastrointestinal blood loss with low dose (350 mg) plain and enteric-coated aspirin administration. A m J Gastruenterul1995; 90: 581-85. Guslandi M. Gastric toxicity of anti-platelet therapy with low-dose aspirin. Drugs 1997; 53: 1-5. Goddard AF, Donnelly MT, Filipowicz B, Morant SV, Shield MJ, Hawkey CJ. Low dose misoprostol as prophylaxis against low dose aspirin-induced gastroduodenal mucosal injury. Gut 1996; 39: A33.
SIR-with prolonged daily use of aspirin, gastrointestinal loss of iron is likely, even when bleeding is so occult as to be detectable only by sensitive methods. A perspective on aspirininduced iron loss is gained by comparison with menstruation, which has undeniable impact on iron status. Iron lost from regular aspirin use can cumulatively exceed usual menstrual losses. Averaged over the month, typical menstrual blood losses amount to 1-2 mL per day. One aspirin (325 mg) tablet can cause 1-10 mL of
bleeding from the stomach, with much individual variati0n.l Adults without chronic blood loss store 300-1000 mg of iron, and blood contains about 0.5 mg of iron per mL. With aspirinassociated iron losses of 0.5-5 mg per day, iron stores could be exhausted in months, depending on the initial size of the iron load and the rate of blood loss. Induction of iron depletion may be an important second mode of action of aspirin used long-term, rather than an undesirable side-effect. Loss of stored iron is an increasingly likely mechanism of protection against heart diseaseZand ~ a n c e r .For ~ the protective action of aspirin against cancer, the most plausible previously proposed mechanisms remain essentially speculative.4Prolonged use of aspirin is thought necessary for protection against cancer, and there is an apparently increasing effectiveness against heart disease with duration of treatment. These patterns are consistent with an iron-loss mechanism since they suggest a cumulative protective effect (progressive lessening of iron stores over time). The benefits of daily aspirin Symmons5 discusses in his Nov 23 Commentary were all associated with treatments that do not eliminate gastrointestinal bleeding. Is a daily aspirin beneficial in spite of, or because of gastrointestinal blood loss? Studies are uninformative on this fundamental question. The clinical trials were inescapably a test of the global effects of aspirin. What is the empirical basis for the conclusion that benefit is only a direct result of intravascular platelet inhibition and not a consequence of occult bleeding? Efforts to minimise bleeding are founded not on empirically based conclusions, but on the assumption that aspirin-associated bleeding is an undesirable side-effect under all circumstances. A new aspirin designed to be harmless to the stomach may turn out to be a less effective aspirin. Without new trials of efficacy there is a danger that ineffectiveness might never be detected if so-called safe versions of aspirin are prescribed for prevention of heart disease and cancer. Until the contributions of all possible mechanisms are known, those with aspirin-induced iron depletion should be cautioned against supplemental iron. Jerome L Sullivan Pathology and Laboratory Medicine Service, Veterans Affairs Medical Center, Charleston, SC 29401. USA
1 Graham DY, Smith JL. Aspirin and the stomach. Ann Intern Med 1986; 104: 390-98. 2 Sullivan JL. Iron ns cholesterol. Perspectives on the iron and heart disease debate. J Clin Epzdenziol 1996; 49: 1345-52.
43 1