Clinical Observation of the Temporal Association between Crack Cocaine and

Duodenal Ulcer Perforation Rashmi Sharma, MD, Claude H. Organ, Jr., MD, Elsa R. Hirvela, MD,

Vernon J. Henderson, MD, Oakland, California

HYPOTHESIS: To determine if a cause-effect rela- tionship exists between crack cocaine use and duodenal ulcer perforation (DUP).

PATIENTS AND METHODS: A retrospective study was conducted of all patients undergoing emergency surgical management for peptic ulcer disease over a 6-year period at a large inner-city munici- pal teaching hospital. The hospital records of 78 consecutive patients presenting with complica- tions of peptic ulcer disease between April 1990 and April 1996 were reviewed. Group A (n = 24) consisted of patients with confirmation of crack cocaine usage within 8 hours of clinical presenta- tion; group B (n = 54) consisted of patients with no antecedent history of crack cocaine use. De- mographic data, timing of drug use, clinical pre- sentation, laboratory and radiographic findings, toxicology screening, operative findings, and postoperative course were compared between the two groups.

RESULTS: Both groups revealed a similar gender distribution, tobacco use, prior peptic ulcer symptoms, and laboratory findings. Group A pa- tients were younger (t test, P = 0.01) and more likely to present with perforation, whereas pa- tients in group B presented with a combination of symptoms (chi square, P = 0.03). Duodenal ulcer perforation was present in 75% of patients in group A compared with 4.6% of patients in group B (chi square, P = 0.04). Group B patients had a significantly longer hospital stay compared with those in group A (t test, P = 0.01). Both crack co- caine and alcohol are independent predictors of duodenal ulcer perforation.

CONCLUSIONS: Patients with recent use of crack cocaine and/or alcohol are more likely to present with duodenal perforations. Although a temporal association between crack cocaine use and duo- denal ulcer perforation was demonstrated, this study does not confirm a cause-effect relation-

-

ship. A prospective cohort study is needed to clarify the pathogenesis of this potential cause- effect relationship. Am J Surg. 1997;174:629- 633. 0 1997 by Excerpta Medica, Inc.

C rack-cocaine first appeared as a drug of abuse during the mid 198Os, and its use has increased dramati- cally during the past decade. Crack cocaine is a

leading cause of drug-related deaths in the United States.’ Cocaine was responsible for 50% of all drug-related deaths reported to the Drug Abuse Warning Network.’ The 1990 National Household Drug Survey reported that 11% of American teenagers admitted to occasional use of cocaine.3 The rapid absorption of crack cocaine, when smoked, is facilitated by the vast pulmonary absorptive surface.4 The effects of crack cocaine are therefore immediate and in- tense. A significant body of literature has developed around the medical complications of crack use.5-7 An increasing number of gastroduodenal ulcer perforations have been de- scribed in temporal association with crack cocaine utiliza- tion.s-I0 This study was designed to describe the temporal relationship between crack cocaine use and duodenal ulcer perforation (DUE’). Clinical differences were examined be- tween patients who used crack cocaine (group A) and pa- tients with no history of crack cocaine use (group B).

PATIENTS AND METHODS A retrospective analysis was performed on a consecutive

group of patients who underwent operative intervention at the Alameda County Medical Center (ACMC) for com- plications of peptic ulcer disease between April 1990 and April 1996 (n = 78). Hospital records were reviewed for age, gender, clinical presentation, history of tobacco and alcohol use, prior history of peptic ulcer symptoms, drug abuse, laboratory and radiography findings, and operative management. History of drug abuse was documented as part of the social history on admission history and physical ex- amination. Since routine toxicology screening cannot be performed at our hospital without the patient’s consent, drug testing was performed only on patients who gave in-

~ formed consent. Comparisons between groups were made _. . . _

From the Department of Surgery, University of California, Davis- East Bay, and the Alameda County Medical Center, Oakland, Cal- ifornia.

Requests for reprints should be addressed to Claude H. Organ, Jr., MD, University of California, Davis-East Bay, Department of Surgery, 1411 East 31”’ Street, Oakland, California 94602.

Presented at the 49th Annual Meeting of the Southwestern Sur- gical Congress, Ranch0 Mirage, California, April 13-l 6, 1997.

by t test for continuous data and by chi-square analysis tor categorical data. Differences of I’ less than 0.05 were con- sidered significant.

The variables found to be significant by univariate analysis were subjected to multivariate analysis to determine which variables represent independent associations with DUP. A forward step-wise logistic regression analysis was performed to determine those factors, which independently predict the

0 1997 by Excerpta Medica, Inc. 0002-961 O/97/$1 7.00 629 All rights reserved. PII SOOO2-9610(97)00215-8

1 TEMPORAL ASSOCIATION OF CRACK C

T likelihood of DUP. All statistical analysis was performed using SPSS for windows, version 7.5 (SPSS Inc, Chicago, Illinois). A P value of <O.lO was used as a criterion for entrance of variables into the regression equation, and a P value of >0.20 was utilized as a criterion for exclusion of variables from the regression equation.

RESULTS Seventy-eight patients met the inclusion criteria for this

review (Table I). Patients were divided into groups as fol- lows: group A (n = 24), patients who had smoked crack cocaine within 8 hours of presentation; and group B (n = 54), patients with no antecedent history of crack cocaine use. All patients who consented to drug testing in group A were positive for cocaine in urine toxicology (n = 4). No patients in group B tested positive for cocaine in urine tox- icology (n = 2). The average age of group A patients was lower than that of group B patients (38 years versus 49 years, t test, P = 0.01). Males comprised 70% and 79% of patients in groups A and B, respectively (chi square P = 0.4). To- bacco use was high in both groups (A = 83%, B = 70%, chi square P = 0.3) as was alcohol use (A = 92%, B = 63%, chi square P = 0.03). Fewer patients in group A had a prior peptic ulcer disease history than patients in group B (42% versus 57%). Perforation was the major presenting sign in group A (92%) whereas patients in group B pre- sented with a variety of signs (perforation 63%, bleeding 22%, and obstruction 15%). This difference was statistically significant (chi square P = 0.03). There were no significant intergroup differences in the amylase values or leukocytosis. Pneumoperitoneum was more common in group A than in group B (75% versus 52%, chi square P = 0.03).

Duodenal perforation was present in 75% of patients in group A (18 of 24) and in 46% of patients in group B (25 of 54). This difference was statistically significant (chi square P = 0.04). There was a significant difference in the type of repair between these groups. Most patients in group A (70%) underwent omental patch closure alone whereas 50% of patients in group B underwent definitive ulcer pro- cedures. Group B patients had a significantly increased hos- pital stay compared with patients in group A (12 versus 7 days; t test P = 0.01).

Variables of age, alcohol use, and crack cocaine use were subjected to logistic regression analysis. Both crack and al- cohol use were found to be statistically significant and in- dependent predictors of duodenal ulcer perforation.

COMMENTS This is the largest reported series of patients who smoked

crack cocaine and presented with gastroduodenal perfora- tion. In this series, patients with a recent history of crack cocaine use were younger, more likely to have a history of alcohol use, and more likely to present primarily with per- foration than drug-free patients. Patients who had smoked crack cocaine were found to have perforations primarily in the duodenum; most of these patients had omental patch closure and had shorter hospital stay when compared with patients with no history of crack cocaine use. By logistic regression analysis, both crack and alcohol have been found to be independent predictors of duodenal ulcer perforation.

A summary of data from three other institutions reporting patients who smoked crack cocaine and presented with gas-

630 THE AMERICAN JOURNAL OF SURGERY@ VOLUME 174

-

i I

TABLE I

Summary of Data from Alameda County Medical Center

Group A Group B (n = 24) (n = 54) P Value

Average age (years) 38 49 0.01 Gender (m:f) 19:5 38:16 NS Tobaccouse 20/24 (83%) 38/54 (70%) NS Alcohol use 22124 (92%) 34/54 (63%) 0.03 Prior peptic ulcer

symptoms lo/24 (42%) 31/54 (57%) NS Pneumoperitoneum 18124 (75%) 28/54 (52%) 0.03 Duodenal perforation 18/24 (75%) 25/54 (46%) 0.04 Patch closure 17/24 (70%) 27/54 (50%) NS Length of hospital stay

(days) 7 12 0.01

MS = not significant.

troduodenal perforation is listed in Table II. At each in- stitution, these patients were younger than the drug-free patients requiring closure of a perforated peptic ulcer. At most institutions, patients with a recent history of crack cocaine use did not present with leukocytosis and had no prior history of peptic ulcer disease. Pneumoperitoneum was a common finding at most institutions. In all series the ma- jority of patients with a history of crack cocaine use had perforations in the duodenum. At three out of four insti- tutions, the majority of these patients underwent omento- pexy with or without omental patch closure. Contrary to previous reports, we found no gender-related differences in our patients, perhaps because of our larger sample size and the prevalence of crack cocaine use by females in our catch- ment area.

The pathogenesis of these perforations continues to re- main unknown. A strong temporal relationship exists be- tween the use of crack cocaine and duodenal ulcer perfo- ration. In our series of patients, both crack cocaine and alcohol were found to be independent predictors of duo- denal ulcer disease. Existing literature about the effects of alcohol on peptic ulcer disease is mixed; most authors sug- gest that it predisposes to both gastric and duodenal ulcers while others claim that it has no effect on duodenal ulcer disease. Whether alcohol or crack cocaine or the combi- nation of two was primarily responsible for duodenal ulcer perforations is not answered by this study.

Cocaine causes both adrenergic-dependent and adrener- @c-independent vasoconstriction of vascular smooth mus- cle. i1-i3 In a neonatal swine model, Hebra et ali4 have dem- onstrated that cocaine causes significant and sustained increases in systemic vascular resistance (SVR) and mes- enteric vascular resistance (MVR); they further demon- strated that high doses of cocaine produce a higher MVR than do lower doses. Cocaine administration may increase arterial thrombi by several mechanisms.15-‘7 It is likely that cocaine-induced gastroduodenal perforations could occur secondary to focal tissue ischemia induced by vasoconstric- tion as well as in-situ mesenteric thrombi.” Moldau and Fischman” have shown that cocaine causes significant in- creases in the concentration of ACTH and corticosterone in rats. Iversen and Salt” have observed that certain ste- roids cause dose-dependent inhibition of catecholamine up-

IECEMBER 1997

Figure 1. A comparison of the time, in minutes, to achieve peak physiologic and subjective changes in human volunteers given co- caine by various routes of administration. Adapted with permis- sion from Warner.

Figure 2. Hydrolysis of cocaine to ecgonine methyl ester, benzoylecgonine, and ecgonine. Adapted with permission from Shuster.23

take, potentiating the vasoconstrictive properties of cocaine and cause further ischemia. This supports the hypothesis that cocaine could likely cause perforations of the duode- num directly via vasoconstriction and thrombus formation and indirectly through stimulation of the hypothalamic-pi- tuitary-adrenal axis.

How does crack cocaine differ from other forms of co- caine? Why have we observed an association between crack cocaine and duodenal perforation and failed to observe this finding with other forms of cocaine? The answer may lie in the pharmacology of cocaine. The development of peak co- caine concentration depends primarily upon the route of administration of the drug. Smoking crack produces higher peak levels of cocaine than intranasal or oral administration (Figure 1 ).* There are reports that crack cocaine produces higher mean peak values of heart rate acceleration and sys- tolic and diastolic blood pressures over intravenous injec- tion.22 Biodegradation of cocaine involves both enzymatic and nonenzymatic hydrolysis to yield ecgonine methyl ester and benzoylecgonine respectively (Figure 2).23,24 A notable differences between crack and other forms of cocaine is the production of pyrolyzates: benzoic acid, methylecgonidine, and methyl-4-(3-pyridine) butyrate. In the pyrolysis of

THE AMERICAN JOURNAL OF SURGERY@ VOLUME 174 DECEMBER 1997 631

EMPORAL ASSOCIATION OF CRACK COCAINE AND ULCER PERFORATION/SHARMA ET AL

f!l Benzoic acid q Methylecgonidine

260 300 350 400 650

Temperature (“C)

Figure 3. Effect of temperature on the volatilization and pyrolysis of cocaine during a 1 O-minute period at an air flow rate of 400 mU min. Adapted with permission from Martin and Lue.*5

crack as temperature increases, the percent of pyrolyzates also increases (Figure 3).z5 The pharmacological activity of these compounds is unknown but may be important in ex- plaining the physiologic effects of crack cocaine.

CONCLUSION Surgeons are aware of an increase in duodenal perforations

associated with crack cocaine use. This interesting obser- vation served as the impetus for investigating the relation- ship between crack cocaine use and duodenal ulcer perfo- ration. Patients with a history of recent use of crack cocaine or alcohol are more likely to present with duodenal perfo- rations when compared with patients with no such history. This study confirms a temporal association between crack cocaine use and duodenal perforations. However, a causal relationship cannot be established. A prospective cohort study involving larger number of patients and identification of other important variables is needed to firmly establish the link between crack cocaine and duodenal ulcer perfo- ration. Future studies are necessary to clarify biochemical and neurotransmitter effects of crack cocaine on the gut mucosa. Furthermore, the effects of pyrolyzates should be examined in an experimental model as a potential contrib- utor to crack cocaine-induced duodenal perforation.

REFERENCES 1. Mittlemon RE. Wetli CV. Death caused bv recreational cocaine use: an update. ./AMA. 1984;252:1889-1893.

2. Data from the Drug Abuse Warning Network (DAWN} Annual Data 1989. Washington, DC: US Department of Health and Hu- man Services, Series 1, Number 9, DHHS Publication No. (ADM) 90-7017, 1990. 3. National Institute on Drug Abuse: Main findings 1990. Washing- ton, DC: US Department of Health and Human Services, DHHS Publication No. (ADM) 91-1788, 1991. 4. Warner EA. Cocaine abuse. Ann Intern Med. 1993;119:226-235. 5. Gawin F, Ellinwood EH. Cocaine and other stimulants: action, abuse and treatment. NEJM. 1988;318:1173-1182, 6. Cregler L, Mark H. Medical complications of cocaine abuse. NEJM. 1986;315:1495-1500. 7. Brody S, Slovis C, Wrenn K. Cocaine-related medical problems: consecutive series of 233 patients. Am J Med. 1990;88:325-33 1. 8. Abramson D, Aertler J, Lewis T, &al J. Crack-related perforated gastropyloric ulcer. ] Clin Gastroenterof. 1991;13:17-19. 9. Kram H, Hardin E, Clark S, Shoemaker W. Perforated ulcers related to smoking “crack” cocaine. Am Surgeon. 1992;58:293-294. 10. Lee H, LaMaute H. Acute gastroduodenal perforations associ- ated with use of crack. Ann Surg. 1989;211:14-17. 11. Ritchie JM, Greene NM. Local anesthetics. In: Goodman LS, Gilman AC, Rail TW, Murad F, eds. The Pharmacological Basis of Therapeutics. 7’h ed. New York: Macmillan; 1985:309-310. 12. Webb RC, Vanhoutee PM. Cocame-induced release of norad- renaline in rat tail artery. ] Phurm Pharmacol. 1982;34:134-136. 13. Rongione AJ, Sleg PC, Gal D, Isner JM. Cocaine causes en- dothelium-induced vasoconstriction of vascular smooth muscle. Cir- culation. 1988;78(suppl 11):436. 14. Hebra A, Braun M, McGechin K, et al. Systemic and mesen- teric vascular effects of platelet-activating factor and cocame. Am Surgeon. 1993;59:50-59. 15. Schnetzer GW. Platelets and thrombogenesis. Current con- cepts. Am Heart .l. 1972;83:552-564. 16. Togna G, Tempesta E, Togna AR, et al. Platelet responsiveness and biosynthesis of thrombaxane and prostacyclin in response to in vitro cocaine treatment. Huemostasis. 1985;15:100-107. 17. Eichhorn EJ, Demian SE, Willard JE, et al. Cocaine use reduces prostacylin production in rabbit aorta. J Am Coil Cardiol. 1991;17:349A. 18. Levine SR, Welch KM. Cocaine and stroke. Stroke. 1988;19:779-783. 19. Moldau RL, Fischman AJ. Cocaine induced secretion of ACTH, beta-endorphine, and corticosterone. Peptides. 1987;8:819- 822. 20. lversen LL, Salt PJ. Inhibition of catecholamine uptake by ste, raids in the isolated rat heart. Br I Pharmacol. 1970;40:528-530. 21. Jones RT. The pharmacology of cocaine smoking in humans. NZDA Res Monograph. 1990;99:30-41. 22. Perez-Reyes M, Guiseppi D, Ondrusek G, et al. Free-base co- caine smoking. C&n Phurmacoi Ther. 1982;32:459-465. 23. Shuster L. Pharmacokinetics, metabolism and disposition of co- caine. In: Lakoskt J, Gallaway M, White F, eds. Cocaine. Boca Ra- ton, FL: CRC Press; 1992:1-14. 24. Jeffcoat AR, Perez-Reyes M, Hill JM, et al. Cocaine disposition in humans after intravenous injection, nasal insufflation and smok- ing. Drug Memb Dispos. 1989;17:153-159. 25. Marcin B, Lue L. Pyrolysis and volatilization of c0caine.f A&t Toxicol. 1989;13:158-164.

Caine within 8 hours and those who had not used it at all? Is that a routine on your admission forms? How did you manage to get such a high incidence of screening tests for cocaine unless you were specifically looking pro- spectively for something? Did nobody fall into an inter- mediate zone (ie, used it 12 hours ago rather than 8 as opposed to none) !

DISCUSSION John B. Cone, MD (Little Rock, Arkansas): I think it is

important that we have a better understanding of what crack cocaine does to our surgical patients considering the widespread use and abuse that we are seeing these days.

I do have some questions. First of all, methodologically, how were you able in a retrospective study to so nicely divide the patients into those who had used crack co-

632 THE AMERICAN JOURNAL OF SURGERYB VOLUME 174 DECEMBER 1997

ITEMPORAL ASSOCIATION OF CRACK COCAINE AND ULCER PERFORATION/SHARMA ET AL

The next question is: What was your rationale for simple patch closure, given the fact that what we presume to be the risk factor (the crack cocaine) is likely to still be present once they are discharged from the hospital? What was your treatment algorithm beyond that? Hz blockers? Treatment for H p$ori?

David Feliciano, MD (Atlanta, Georgia): Our group has treated over 60 of these in the last 3 years. In the past year we have been doing biopsies for Helicobacter. I think one of the things you might consider, rather than worrying so much about the specific chemical effects of the cocaine, is to see how many of your patients have concomitant Heli- cobacter. They may simply need treatment for Helicobacter postoperatively in addition to counseling for their cocaine abuse.

Evan Kokoska, MD (St. Louis, Missouri): Instillation of ethanol is a common laboratory model for gastrointestinal mucosal injury. Your group A had an increased incidence of ethanol abuse as well. Were you able with multivariant

analysis to determine that crack cocaine was more signifi- cant than ethanol for predisposing the gastric mucosa per- foration?

CLOSING Rashmi Sharma, MD: We believe that perforations in

patients who had smoked crack cocain is an ischemic pro- cess and not a traditional ulcer-related pathogenesis. There- fore, acid-reducing operative procedures are not recom- mended in patients with a history of crack cocaine and duodenal ulcer perforation. Furthermore, Hz blockers are not recommended for these groups of patients since it has not been shown that the cause of their perforation is due to high acid production.

This was a retrospective study and hospital records were reviewed. Under social history, patients were asked if they had any history of crack cocaine use. Patients with no his- tory of crack cocaine use or no history of cocaine use were placed in group B. We have not looked at H pylon‘ in our patients.

THE AMERICAN JOURNAL OF SURGERY@ VOLUME 174 DECEMBER 1997 633