Les complications de l’examen endoscopique de routine du tractus digestif supérieur

Les compl ica t ions de I ' examen endoscop ique de rout ine du t ractus digest i f sup6r ieur

P. M A I N G U E T , Ph. B O H O N C l i n i q u e s U n i v e r s i t a i r e s S a i n t - L u c , B r u x e l l e s ( B e l g i q u e )

Complications in routine endoscopic examination of the upper digestive tract

RI~SUMI~

L'endoscopie diagnostique du tractus digestif sup6rieur, dont la diffusion est consid6rable, jouit d'une r6putation d'inocuit6.

Les complications m6caniques - - perforations cesophagiennes et gastriques - - sont devenues exceptionnelles depuis l'usage d'instruments de fin calibre. En revanche, les complications rares - - publi6es isol6ment - - doivent retenir l'attention : perforations dues ~ la biopsie intestinale, h6matome duod6nal, traumatisme spl6nique, syndrome de Mallory-Weiss post endoscopique.

Les complications infectieuses, moins fr6quentes qu'apr6s CPRE ou coloscopie, sont graves et justifient le strict respect des r~gles de nettoyage et de d6sinfection des instruments et des accessoires.

Les complications ~ haut risque de la s6dation sont fr6quentes et souvent m~connnues chez les sujets fig6s, an6miques ou ob6ses. La surveillance 61ectrocardiographique, oxym6trique, est n6cessaire ; en cas de complication, la disponibilit6 de moyens de r6animation immddiate s'impose.

S U M M A R Y

Diagnostic endoscopy o f the upper digestive tract, which has become considerably widespread, enjoys the reputation o f being harmless.

Mechanical complications - - esophageal and gastric perforations - - have become an exception since the use o f fine calibre instruments. On the other hand, however, rare complications - - published individually - - should be brought to our attention : perforations due to an intestinal biopsy, duodenal hematoma, splenic traumatism, postendoscopic Mallory-Weiss syndrome.

The complications arising from infection, less frequent than after ERCP or colonoscopy, are serious and justify the implementa- tion o f the strict rules for the cleaning and disinfection o f the instruments and accessories.

The high risk complications arising from sedation are frequent and often misunderstood in elderly, anemic or obese patients. Electrocardiographic and oximetric monitoring is necessary and, i f complications arise, the availability o f means for immediate resuscitation is mandatory.

Les complications de l 'endoscopie ceso-gastro- duod6nale ( E O G D ) ~ vis6e diagnostique sont de nature m6canique, infectieuse, ou sont li6es ~t la s6dation, ou r6sultent d 'un d6faut de surveillance per- ou post-endoscopique.

La fr6quence des complications est appr6ci6e sur base d 'enqu~tes ant6rieures h 1980 [49, 50]. Les 6valuations de fr6quence, grosso modo, restent valables pour les complications m6caniques. En revanche, les infections transmissibles par l 'endoscopie ont 6t6 r6cemment l 'objet d 'une attention beaucoup plus stricte, dict6e par de nou- velles cat6gories de patients explor6s ( immunod6-

prim6s, candidats aux greffes d 'organes, S IDA) , par des moyens plus s61ectifs d'identification des patients contamin6s, et enfin par une maitrise meilleure d 'appareils immergeables, soumis ~ des modes de net toyage et de d6sinfection standar- dis6s.

Les complications de la s6dation, longtemps sous-estim6es, sont mieux connues et plus ais6- ment pr6venues grfice ~t la surveillance continue per-endoscopique (61ectrocardiographie, oxym6trie, surveillance tensionnelle) de patients soumis ~t des m6dicaments s6datifs, morphiniques ou analg6si- ques [9, 22].

Tir6s ~ part : pr p. MAINGUET, Universit6 Catholique de Louvain, Cliniques Universitaires Saint-Luc, Service de Gastroent6rologie, avenue Hippocrate 10, 1200 Bruxelles (Belgium).

Mots-cl~s : complications, endoscopie du tractus digestif sup6rieur. Key-words : complications, upper G.I. tract endoscopy.

Acta Endoscopica Volume 24 - N ~ 3 - 1994 213

L E S C O M P L I C A T I O N S M I ~ C A N I Q U E S

Perforation oesophagienne

La fr6quence globale des complications de l'endoscopie digestive haute 6tait de 1,32/1000 selon une enqu6te de I 'A.G.A. (1976) [49]. La fr6quence des perforations 0esophagiennes 6tait de 0 , 1 % en 1980, ce qui plaqait cet accident grave au premier rang des complications. I1 faut rappeler qu'en d6pit d'un diagnostic plus pr6coce, le taux de mortalit6 des perforations spontan6es et instrumentales de l'0esophage thoracique reste 61ev6 : de 64 % de d6c~s en 1970 dans la s6rie d 'Abbott [1], il passe ~ 19 % dans la s6rie du GEEMO en 1987 (Groupe d 'E tude Europ6en des Maladies de l'(Esophage) [30]; de plus, un diagnostic retard6 n6cessite habituellement l'0esophagectomie [51].

La large diffusion des fibroscopes de fin calibre et des instruments de vid6o-endoscopie a r6duit la fr6quence de cet accident, si l'on excepte les complications cons6cutives gi un acte th6rapeutique (dilatation, traitement des tumeurs par laser, mise en place de proth6ses).

Les remarques suivantes s'imposent : - - L a perforation au cours d'une 0esophago-

scopie diagnostique survient le plus souvent au niveau de l'0esophage cervical: d6chirure d'un sinus piriforme, d'une face post6rieure de l'0eso- phage cervical bombant sous une saillie vert6brale cervicale ost6ophytique. Dans la majorit6 des cas, l'endoscopiste a 6t6 surpris par l'absence de sensation de rupture, et il faut souligner que l'cesopha- gogramme avec produit de contraste soluble ne r6v61e la perforation que dans la moiti6 des cas [50]. Des perforations cesophagiennes sur biopsie de 16sions ulc6reuses ou tumorales n'ont plus 6t6 rapport6es depuis 1964 [52].

- - L'impaction du fibroscope. Les endoscopes flexibles de fin calibre peuvent,

en cours d'introduction, subir un retournement en r6tro-vision et immobiliser l'extr6mit6 distale de l'instrument dans la lumi6re cesophagienne. Lors- que l'impaction est cervicale, le retrait progressif est possible mais il ne peut 6tre poursuivi en cas de r6sistance. Il est alors pr6f6rable d'utiliser un second endoscope et de pousser les deux instruments dans la cavit6 gastrique afin de permettre le redressement du premier et le retrait sans dom- mage pour la paroi cesophagienne [42]. L'impaction peut 6galement se produire au niveau d'une hernie hiatale lorsque l'appareil est introduit en r6trovision dans la poche herniaire. Le d6gage- ment de l'appareil impact6 s'op~re de la m6me faqon qu'au niveau de l'cesophage.

Perforation gastrointestinale

La fr6quence cumul6e des h6morragies et perforations compliquant l 'endoscopie haute de routine varie de 0,008 ~ 0,03 % [37, 52, 65].

La perforation instrumentale proprement dite est donc exceptionnelle et se situe, d'habitude au

niveau de la face ant6rieure de l'estomac, notam- ment sur une vaste 16sion ulc6rative de sa portion crhniale [50]. Quelques perforations de l'intestin gr61e [41] ou du caecum [59] ont 6t6 rapport6es. Ces perforations r6sultent d'une insuffiation excessive sur 16sion gr61e ou colique pr6-existante (diverticulose j6junale, st6nose tumorale du gr61e proximal [3], ou encore un ad6nocarcinome du caecum [59]. En pratique, ces accidents peuvent ~tre 6vit6s en contr61ant la dur6e et l'importance de l'insuffiation, 6ventuellement les r6actions d'un patient vigile. Enfin, la d6couverte d e 16sions gas- troduod6nales multiples 6voquant le lymphome, doit 6videmment faire suspecter l'existence de 16sions intestinales distales h haut risque de perforation.

Le remplacement de la biopsie du gr61e h la sonde par les biopsies perendoscopiques [8, 29, 47, 48, 63, 64] dans le diagnostic des affections dif- fuses de la muqueuse j6junale expose h u n faible risque de perforation.

Toutefois, Scott et al. (1993) ont rapport6 2 cas de perforation duod6nale et j6junale apr~s biopsies ~ la pince chez les malades coeliaques [64].

En cas de plaintes abdominales, pr6coces ou tardives, une radiographie d 'abdomen sans pr6pa- ration doit 6tre effectu6e.

Pneumopathies par fausse ddglutition

Relativement fr6quent en unit6s de soins inten- sifs ou chez les patients explor6s pour h6morragie digestive (scl6rose de varices, etc.), l'accident est plus exceptionnel en endoscopie digestive haute de routine.

Parmi les facteurs pr6disposants, figurent : - - l a stase 0esophagienne (st6nose organique,

achalasie avec m6ga-oesophage) ou gastrique ; - - u n reflux gastro-oesophagien important chez

un malade soumis h une s6dation prononc6e ou prolong6e.

L'endoscopie doit de toute faqon 6tre effectu6e en disposant d'une seconde source d'aspiration avec sonde de fin calibre permettant l'aspiration bronchique imm6diate.

L E S C O M P L I C A T I O N S R A R E S

La pr6sente revue n'en dresse pas le bilan exhaustif. Certaines complications rares doivent cependant retenir l 'attention :

- - Le gonf lement des glandes salivaires et paro- tides, quoique rarement signal6, n'est pas exceptionnel apr~s E O G D [53]. Trois m6canismes pathog6niques ont 6t6 propos6s: r6tention sali- vaire de s6cr6tions 6paissies, quinte de toux suivie de congestion veineuse du parenchyme glandu- laire, stimulation para-sympathique exag6r6e et vaso-dilatation en cours d 'EOGD. L'anomalie est spontan6ment r6gressive.

214 Volume 24 - N ~ 3 - 1994 Acta Endoscopica

- - L e traumatisme spl~nique compliqu6 d'h6- morragie sous-capsulaire ou intra-abdominale est une complication connue de la coloscopie [14, 27, 32, 44, 60, 66] ou plus rarement de la CPRE [56, 69]. Lewis (1991) a rapport6 un cas de d6capsula- tion spl6nique apr6s E O G D , dont les suites furent control6es par le traitement chirurgical [45].

- - U n cas d 'hdmatome sous-sdreux du duode- n u m distal apr~s E O G D a 6t6 r6cemment rapport6 par Pa/ikk6nen et al. [57].

- - Penston J.G. et al. (1992) ont relev6 7 cas de syndrome de Mallory-Weiss apr6s EOGD dans une s6rie de 10 000 examens cons6cutifs effectu6s en 6 ans [58]. Les auteurs soulignent la plus grande fr6quence de cette complication chez des femmes hg6es porteuses d'une hernie hiatale et l'absence d'6ructation en salve au cours de l'examen. La fr6quence de cette complication dans leur s6rie est comparable ~ celle de publications ant6- rieures [13, 31].

L E S C O M P L I C A T I O N S I N F E C T I E U S E S

Les infections transmises par les endoscopes sont peu fr6quentes lorsque les pr6cautions de nettoyage et de d6sinfection sont correctement appliqu6es. N6anmoins, leurs manifestations peuvent 6tre retard6es et par cons6quent, m6connues.

Les endoscopies digestives hautes provoquent un pourcentage non n6gligeable de bact6ri6mies, variant d'apr6s les s6ries de 0 ~ 10 %, selon Essioux et Vergeau [28], dont la revue porte sur 12 articles publi6s de 1971 ~t 1984. En moyenne, leur fr6quence serait de 4,65 % apr6s EOGD avec biopsies [4, 11].

Dans la pratique de la CPRE, le risque est surtout 1i6 ~ une d6contamination insuffisante [18, 26]; exceptionnellement, les mesures recomman- d6es (nettoyage complet, d6contamination, s6chage et stockage appropri6s) peuvent rester inefficaces et imposer le remplacement de l'appareil [62].

En revanche, les complications infectieuses avec manifestations cliniques sont plus rares et, d'apr6s la litt6rature, leur fr6quence serait inf6rieure 0 , 1 % [19, 55, 65].

Au cours des quinze h vingt derni6res ann6es, des cas d'infection transmises par endoscopie ont 6t6 rapport6s. La majorit6 d'entre elles concernent les transmissions d'infections bact6riennes, en par- ticulier Pseudomonas Aeruginosa et Salmonella SP. Les contaminations par Pseudomonas Aeruginosa (PA) concernent en majorit6 des explorations par CPRE [2, 17, 18, 24, 26]. En tout, 45 cas ont 6t6 rapport6s dont 4 sont d6c6d6s.

Le risque, quoique minime, existe en EOGD et une septic6mie h PA a 6t6 rapport6e apr6s endoscopie chez un patient atteint de leuc6mie aigu6 [34].

Acta Endoscopica

De nombreux exemples de septic6mies et m6me d'6pid6mies h Salmonella ont 6t6 d6crits apr~s diverses explorations endoscopiques : gastroscopies, CPRE, sigmoidoscopie et coloscopie [6, 12, 15, 21, 25, 38, 39, 54, 61, 70]. Sur 72 cas rapport6s, 38 avaient une expression clinique et 1 cas est d6c6d6 [25]. Ici encore, le mode de d6contamination 6tait souvent d6fectueux.

Un autre risque d'infection m6connu concerne Helicobacter pylori. Un exemple indiscutable d'infection crois6e par 0eso-gastro-duod6noscopie a 6t6 fourni par la sErie de Langenberg et al. (1990) [43]. Graham D. et al. [33] ont confirm6 le r61e de l'endoscopie dans la contamination iatrog~ne par H pylori avec primo-infection et achlorhydrie 6pid6mique.

Gullini et al. [35], dans une 6tude prospective, ont montr6 qu'avant lavage et d6sinfection, H pylori a 6t6 retrouv6 au niveau de 33,3 % des fibroscopes et 11,1% des pinces h biopsie. Cette donn6e confirme la valeur pr6ventive des proc6d6s de nettoyage et d6sinfection dans le risque de transmission de H pylori d'individu h individu.

Jusqu'~ pr6sent, aucune observation documentEe de transmission du virus HIV par endoscopie n'a 6t6 rapport6e par les auteurs qui ont 6valu6 ce risque [16, 36].

N6anmoins, le d6calage entre la contamination et la positivit6 de la s6rologie ou les manifestations d'immuno-d6ficience, expose au risque de sous-estimer cette contamination.

La virulence du virus HIV est abolie apr~s exposition aux d6sinfectants de routine lorsque des concentrations inf6rieures aux standards sont utili- s6es. Des endoscopes contamin6s par des taux 61ev6s de virus HIV, apr~s trempage dans la glutaraldehyde pendant 30 minutes, ne contiennent plus de formes viables du virus en fin de d6sinfection. Plusieurs 6tudes prospectives ont d6montr6 qu'un temps de nettoyage de 6 ~ 8 minutes suffit 6carter ce risque.

Le virus de l'h6patite B a une virulence sup6- rieure au virus HIV lorsqu'il est inject6 par voie parent6rale. En d6pit de la large diffusion de l 'endoscopie, un seul cas document6 de contamination endoscopique a 6t6 rapport6 [10]. Le temps de d6sinfection souhait6 est de 30 minutes. Une 6tude prospective de Lok et al. [46] a montr6 que les patients examin6s apr6s un patient porteur du virus HB V conservent tous une s6rologie n6gative six mois apr~s l'examen lorsqu'ils ont 6t6 explor6s avec des appareils correctement d6sinfect6s.

Un cas de contamination par le virus C apr6s CPRE a 6t6 r6cemment rapport6 [67].

C O M P L I C A T I O N S D E L A S E D A T I O N

Le choix de la s6dation au cours de l'endoscopie digestive haute de routine d6pend de facteurs

Volume 24 - N ~ 3 - 1994 215

psychologiques, d'habitudes d'6cole et de la compliance 6ventuelle des malades vis-a-vis d'examens de contr61e.

Trinchet et al. [68] ont montr6 qu'un mois apr~s une endoscopie haute sans pr6m6dication, 5 % des patients refusent l'6ventualit6 d'une nouvelle exploration. La r6alisation d'une endoscopie sous s6dation pose un probl~me commun h t o u s l e s patients, quel que soit le mode de s6dation uti- lis6 : existence d'une contrainte concernant l'inter- diction de la conduite automobile et perte d'une journ6e d'activit6 en raison de troubles post- endoscopiques (diminution du degr6 de vigilance, amn6sie, etc.).

- - La s6dation 16g~re est obtenue par administration intra-musculaire de Midazolam: celle-ci n6cessite une surveillance per- et post-endoscopique par un personnel entrain6, la mise en place d'une voie d'acc~s veineux, et la possibilit6 d'administrer de l'oxyg~ne h la demande.

- - L a s6dation profonde - - neuroleptanalg6sie ou anesth6sie g6n6rale - - n6cessite la pr6sence d'un m6decin anesth6siste-r6animateur et la surveillance 61ectrocardiographique, oxym6trique et tensionnelle du patient pendant l 'examen et au cours de la p6riode de r6veil.

Plusieurs 6tudes ont montr6 une d6saturation art6rielle en oxyg~ne en cours d'examen refl6tant une hypoxie temporaire. Celle-ci s'observe dans environ 15 % des cas au cours de l'endoscopie de routine [5] ; elle est plus fr6quente et plus s6v~re chez les patients subissant une coloscopie totale [20]. Dhariwal et al. (1992) [23], dans une 6tude prospective de 82 patients soumis ~ une EOGD (dur6e moyenne : 8,5 + 0,42 min) sous Midazo- lam IV (dose moyenne : 6,3 + 0,15 mg), ont enre- gistr6 une d6saturation > 4 % pendant 15,68 % de la dur6e d'examen. Parmi les facteurs de pr6- disposition ~ une d6saturation perendoscopique, figurent : l'gtge (> 65 ans), l'an6mie (h6moglobine < 10/dl) et l'ob6sit6 (indice corporel > 28).

En effet, l'hypox6mie s6v~re entra~ne un trouble du rythme cardiaque, des modifications i sch6mi-

ques qui peuvent conduire h l'arr6t cardiaque. Le risque d'apn6e et d'arr6t cardiaque n'est pas n6gli- geable, surtout apr~s 60 ans : Iber F.L. et al. rapportent 10 cas dans une s6rie de 10 000 endoscopies cons6cutives. Quatre sont survenus lors de l'administration du m6dicament, les autres apr~s la fin de l'examen [40].

L'6tude de Bell (1990) [7] souligne 6galement l'importance du risque: plus de 50 % des d6c~s associ6s ~ l'endoscopie digestive haute r6sultent de probl~mes cardio-pulmonaires.

Par cons6quent, l'administration de s6datifs (Midazolam) ou de morphiniques, doit 6tre adap- t6e h la lourdeur de l'acte technique, h l'hge des patients et au contexte clinique (cirrhose, m6dica- ments consomm6s, etc.), et 6tre surveill6e par 61ectrocardiographie et oxym6trie sous le contr61e d'un anesth6siste.

CONCL USIONS

Les complications m6caniques de I 'EOGD sont devenues rares grace h l'emploi de vid6o-endoscopes ou d'instruments de fin calibre. Une attention particuli~re doit 6tre accord6e h des complications exceptionnelles - - publi6es cas par cas - - justifiant parfois une sanction chirurgicale rapide (h6matome duod6nal - - d6chirure de la capsule spl6nique). Les endoscopistes doivent 6tre incit6s publier de semblables observations, ce qui contri- bue h la r6ussite de l'ensemble de la pratique endoscopique.

Les deux probl~mes majeurs restent l'infection et les accidents de la s6dation. Dans les deux cas, la pr6vention est non seulement possible mais ses r~gles sont correctement d6finies. Elles alourdis- sent la routine quotidienne, mais la s6curit6 des malades est h ce prix, d'autant que I 'EOGD est largement pratiqu6e en dehors de centres hospita- liers.

R~F~RENCES

1. ABOTF O.A. , MANSOUR K.A., LOGAN W.D., HACHTER C.R., SYMBAS F.N. - - Atraumatic so-called ~, Spontaneous ,~ rupture of the esophagus. J. Thorac. Car- diovasc. Surg., 1970, 59, 67-83.

2. ALLEN J.L., ALLEN M.O., OLSON M.M., GERDING D.N., SHANHOLTZER C.J., MEIER PB et al. - - Pseu- domonas infection of the biliary system resulting from use of a contaminated endoscope. Gastroenterology, 1975, 69, 507-510.

3. ASTON N.O., HOUGHTON A.D. - - Jejunal perforation after esophagogastroscopy. Gastrointestinal Endoscopy, 1989, 35, 586-587.

4. BALTCH A.L., BURHUC T., A G R A W A L A. et al. - - Bacteremia after upper gastrointestinal endoscopy. Arch. Int. Med., 1977, 137, 594-597.

5. BARKIN J.S., KRIEGER B., BLINDER M., BOSCH- BLINDER L., GOLDBERG R.I., PHILLIPS R.S. - - Oxygen desaturation and changes in breathing pattern in patients undergoing colonoscopy and gastroscopy. Gastroin- testinal Endoscopy, 1989, 35, 526-530.

6. BEECHAM J.H., COHEN M.L., PARKIN W.E. - - Sal- monella typhimurium, transmission by fiberoptic upper gastrointestinal endoscopy. JAMA, 1979, 241, 1013-1015.

7. BELL G.D. - - Review article, premedication and intrave- nous sedation for upper gastrointestinal endoscopy. Ali- ment. Pharmacol. Therap., 1990, 4, 103-122.

8. BELLASSAI J., BEKRI L., HAOT J., MAINGUET P. - - Evaluation de la validit6 des biopsies endoscopiques dans les syndromes de malabsorption de l'adulte. Acta Endosc., 1982, 12, 307-316.


9 . B I G A R D M . A . , H A B E R E R J . P . - - Tube digestif et pan- cr~as. Editorial. Techniques anesth6siques au cours des endoscopies digestives. Gastroenterol. Clin. Biol., 1989, 13, 651-653.

10. BIRNIE G.G. , QUIGLEY E.M., CLEMENTS G.B., FOLLET E.A.C. , WATKINSON G. - - Endoscopic transmission of hepatitis B virus. Gut, 1983, 24, 171-174.

11. BOTOMAN V.A., SURAWITZ. - - Bacteremia with gastrointestinal endoscopic procedures. Gastrointest. Endosc., 1986, 32, 342-346.

12. B U R K H A R D T F . - - Salmonellenubertragung durch Gas- troskopie. Hosp. Hygiene, 1976, 11, 337-339.

13. CAPRON J.P. - - Mallory-Weiss syndrome produced by upper endoscopy (Letter). Gastroenterology, 1982, 83, 522- 523.

14. CASTELLI M. - - Splenic rupture, an unusual late complication of colonoscopy. Can. Med. Assoc. J., 1986, 134, 916-917.

15. CHMEL H . , A R M S T R O N G D . - - Salmonella typhi by fiberoptic endoscopy. Lancet, 1977, 2, 134.

16. CLASSEN M., DANCYGIER H., G U R T L E R I., DEIN- H A R D T F . - - Risk of transmitting HIV by endoscopes (letter). Endoscopy, 1988, 20, 128.

17. CLASSEN D.C., JACOBSON J.A., BURKE J.P., JACOBSON J . T . , E V A N S R . S . - - Serious pseudomonas infections associated with endoscopic retrograde cholangiopancreatography. Am. J. Med., 1988, 84, 590-596.

18. CRYAN E.M., FALKINER F.R., MULVIHILL T.E., KEANE C.T., KEELING P . W . - - Pseudomonas aeruginosa cross-infection following endoscopic retrograde cholangiopancreatography. J. Hosp. Infect., 1984, 5, 371-376.

19. DALY J.P., MOLINIE C., MISSONNIER G., ESSIOUX H., SALIOU P., DUROSOIR J.L., CRISTAU P., LA- V E R D A N T C.H. - - Le risque infectieux en endoscopie digestive. Etude prospective et r6sultats d 'une enqu6te na- tionale. Med. Armies, 1979, 7, 799-805.

20. DARK D.S., CAMPBELL D.R., WESSELIUS L.J. - - Arterial oxygen desaturation during gastrointestinal endoscopy. The Am. J. of Gastroent., 1990, vol. 85, 10, 1317- 1321.

21. DEAN A.G. - - Transmission of Salmonella typhi by fiberoptic endoscopy. Lancet, 1977, 2, 134.

22. DEVALOIS B., S A U T E R E A U D., DESPORT J.C., D U P U Y J.F., SARDIN B., CLAUDE R., PILLEGRAND B. - - S6dation en endoscopic digestive. Gastroenterol. Clin. Biol., 1989, 13, 679-686.

23. D H A R I W A L A., PLEVRIS J.N., LO N.T.C., FINLAY- SON N.D.C., H E A D I N G R.C., HAYES P.C. - - Age, anemia, and obesity-associated oxygen desaturation during upper gastrointestinal endoscopy. Gastroint. Endosc., 1992, vol. 38, 6, 684-688.

24. DOHERTY D.E. , FALKO J.M., LEFKOWITZ N., ROGERS J . , F R O M K E S R . J . - - Pseudomonas aeruginosa sepsis following endoscopic retrograde cholangiopancreatography (ERCP). Dig. Dis. Sci., 1982, 27, 169-170.

25. DWYER D.M., KLEIN E.G. , ISTRE G.R. , ROBINSON M.G., N E U M A N N D.A. , McCOY G.A. - - Salmonella newport infections transmitted by fiberoptic colonoscopy. Gastrointest. Endosc., 1987, 33, 84-87.

26. E A R N S H A W J.J., CLARK A.W., THOM B.T. - - Out- break of Pseudomonas aeruginosa following endoscopic retrograde cholangiopancreatography. J. Hosp. Infect., 1985, 6, 95-97.

27. ELLIS W.R., HARRISON J.M., WILLIAMS R.S. - - Rupture of spleen at coloscopy. Br. Med. J., 1979, 1, 307- 308.

28. ESSIOUX H., V E R G E A U B. - - Probl~mes infectieux pos6s par l 'endoscopie digestive. 1" pattie, incidence de l'infection. Lettre Infectiol., 1987, 2, 143-153.

29. GAY G., DELMOq'TE S., DE G R E Z Th. - - L'endoscopie de l 'intestin gr61e en 1992, est-ce la fin du tunnel ? Acta Endosc., 1992, 22, 129-140.

30. G A Y E T B. - - Perforations de I'~esophage (rapport du Groupe d'dtude europ6en des maladies de l'~esophage). Actualit~s Digestives, 1987, 9, 52-53.

31. GILBERT D.A., SILVERSTEIN F.E., TEDESCO F.J. - - National ASGE survey on upper gastrointestinal bleeding, complications of endoscopy. Dig. Dis. Sci., 1981, 26 (Suppl. July), 55s-59s.

32. GORES P.F., SIMSO L.A. - - Splenic injury during colonoscopy. Arch. Surg., 1989, 124, 1342.

33. G R A H A M D.Y., ALPERT L.C., SMITH J.L., YOSHI- M U R A H.H. - - Iatrogenic Campylobacter pylori infection is a cause of epidemic achlorhydria. Am. J. Gastroenterol., 1988, 83, 974-980.

34. G R E E N E W.H., MOODY M., HARTHLEY R. et al. - - Esophagoscopy as a source of Pseudomonas Aeruginosa sepsis in patients with acute leukemia, the need for sterili- zation of endoscopes. Gastroenterology, 1974, 67, 912-919.

35. GULLINI S., BASSO O., CABERLETFI I., VINCENZI M., BOCCIA S., MACARIO F., CANTARINI D. and ALVISI V. - - Can Campylobacter pylori be transmitted by fiberscope ? Elsevier Science Publishers (B.V) (Biomedical Division). Gastroduodenai pathology and Campylobacter pylori. F. M6graud and H. Lamouliatte, editors, 1989, pp. 435-437.

36. HANSON P..I., G O R D., .IEFFRIES D.J., COLLINS J.V. - - Elimination of high titre HIV from fiberoptic endoscopes. Gut, 1990, 31, 657-659.

37. HART R., CLASSEN M. - - Complications of diagnostic endoscopy. Endoscopy, 1990, 22, 229-233.

38. HAWKEY P.M., DAVIES A.J., VIANT A.C. , LUSH C . J . , M O R T E N S E N N . J . - - Contamination of endoscopes by Salmonella species. J. Hosp. Infect., 1981, 2, 373-376.

39. H O L M B E R G S.D., OSTERHOLM M.T., SENGER K.A., CPOHEN M.L. - - Drug-resistant Salmonella from animals fed antimicrobials. N. Engl. J. Med., 1984, 311, 617-622.

40. IBER F.L., LIVAK A., KRUSS D.M. - - Apnea and cardiopuimonary arrest during and after endoscopy. J. Clin. Gastroenterol., 1992, 14 (2), 109-113.

41. JEFFREY G.P. , REED W.D. - - Non-instrumental small bowel perforation following upper gastrointestinal endoscopy. Gastrointestinal endoscopy, t989, vol. 35, 6, 585- 586.

42. KAVIN H . , S C H N E I D E R J . - - Impaction of fibre-optic gastroscope in the esophagus, an unusual complication of gastroscopy. S. Aft. Med. J., 1970, 44, 478-479.

43. L A N G E N B E R G W., RAUWS E.A., O U D B I E R J.H., TYTGAT G . N . J . - - Patient-to-patient transmission of Campylobacter pylori infection by fiberoptic gastroduode- noscopy and biopsy. J. Infect. Dis., 1990, 161, 507-511.

44. LEVINE E, WETZEL L . H . - - Splenic trauma during colonoscopy. A.J.R., 1987, 149, 939-940.

45. LEWIS Fr. W., MOLOO N., STIEGMANN G.V., GOFF J.S. - - Splenic injury complicating therapeutic upper gastrointestinal endoscopy and ERCP. Gastrointest. Endosc., 1991, 37, 632-633.

46. LOK A.S.F., LAI CHING-LUNG, HUI WAI-MO, MAT- THEW M.T.N.G. , PUI-CHEE W.U., SHIU-KUM LAM and ELSIE KY LEUNG. - - Liver and biliary. Absence of transmission of hepatitis B by fiberoptic upper gastrointestinal endoscopy. Journal of Gastroenterology and Hepatolo- gy, 1987, 2, 175-180.

47. M A I N G U E T P., DEBONGNIE J.C., HAOT J. - - La j6junoscopie dans la surveillance du malade coeliaque trait6. Acta Endosc., 1989, 19, 95-101.

48. M A I N G U E T P., D E G R E Z Th., J O U R E T A., H A O T J. - - La maladie coeliaque de l'adulte, aspects cliniques - - r61e de l 'endoscopie. Acta Gastro-Enterot. Belg., 1992, LV, 181-189.

49. MANDELSTAM P., S U G A W A C., SILVIS S.E., NEBEL O.T., R O G E R S B.H.G. Complications associated with esophagogastroduodenoscopy and with esophageal dilation. Gastrointest. Endosc., 1976, 23, 16-19.

Acta Endoscopica Volume 24- N ~ 3 - 1994 217

50. MEYERS M ; A . , G H A H R E M A N I G . G . - - Complications of fiberoptic endoscopy. I. Esophagoscopy and gastroscopy. Radiology, 1975, 155, 293-300.

51. MICHEL Ch., LEFRANGOIS C., APOIL B., FOUR- NIER L., SEGOL Ph., G I G N O U X M. - - Les perforations spontan6es et instrumentales de l'~esophage thoracique (non n6oplasiques, non caustiques). A propos de 20 cas. Ann. Chir., 1991, 45, n ~ 7, 570-576.

52. MILLER G. - - Komplikationen bei der Endoskopie des oberen Gastrointestinaltraktes. Leber. Magen. Darm., 1987, 5, 299-304.

53. NIJHAWAN S., RAI R.R. - - Parotid swelling after upper gastrointestinal endoscopy. Gastrointest. Endosc., 1992, 38, 1, 94.

54. O 'CONNOR B.H., BENNETF J.R., A L E X A N D E R J.G. et al. - - Salmonellosis infection transmitted by fiberoptic endoscopes. Lancet, 1982, 2, 864-866.

55. OI I., M O H U Y A M A M., T A K E M O T O T. - - Complica- tions of duodenoscopy and retrograde catheterization of the biliary and pancreatic ducts. In Third World Congress of Gastrointestinal Endoscopy, Mexico 1974.

56. ONG E., BOHMLER U., WURBS D. - - Splenic injury as a complication of endoscopy, two case reports and a literature review. Endoscopy, 1991, 23, 302-304.

57. PA,~KKONEN M., HEIKKINEN M. - - Subserosal hematoma of the duodenum as a complication of diagnostic esophagogastroduodenoscopy. Endoscopy, 1993, 25 (3), 253.

58. PENSTON J.G., BOYD E.J.S., WORMSLEY K.G. - - Mallory-Weiss tears occurring during endoscopy, a report of seven cases. Endoscopy, 1992, 24, 262-265.

59. REX D.K., HAWES R.H., G O U L E T R.J. - - Cecal perforation after upper endoscopy and pyloric dilation. Gas- trointestinal Endoscopy, 1989, 34, 490-491.

60. ROCKEY D.C., W E B E R J.R., W R I G H T T.L., WALL S.D. - - Splenic injury following colonoscopy. Gastrointest. Endosc., 1990, 36, 306-309.

61. SCHLIESSLER K.H., R O Z E N D A A L B., TAAL C., MEAWlSSEN S.G.M. - - Outbreak of Salmonella agona infection after upper intestinal fiberoptic endoscopy. Lan- cet, 1980, 2, 1246.

62. SCHOUTENS-SERRUYS E., ROST F., DEPRE G., CREMER M . , L O R I E R S M . - - The significance of bacterial contamination of fiberoptic endoscopes (letter). J. Hosp. Infect., 1981, 2, 392-394.

63. SCOTT B.B., JENKINS D. - - Endoscopic small bowel biopsy. Gastrointest. Endosc., 1981, 27, 162-167.

64. SCOTT B. et al. - - Perforation from endoscopic small bowel biopsy. Gut, 1993, 34 (1), 134-135.

65. SILVIS S.E., NEBEL O., ROGERS G. et al. - - Endoscopic complications. JAMA, 1976, 9, 928-930.

66. TELMOS A.J. , MITTAL V.K. - - Splenic rupture following colonoscopy. JAMA, 1978, 237, 2718.

67. TENNENBAUM R., C O L A R D E L L E P., CHOCHON M., MAISONNEUVE P., JEAN F., ANDRIEU J. - - H6patite C apr~s cholangiographie r6trograde (lettre). Gas- troenterol. Clin. Biol., 1993, 17, 763.

68. TRINCHET J.C., B E A U G R A N D M., BARS L., HUET B., FERRIER J.P. - - La pr6m6dication par le Diazepam am61iore l'acceptabilit6 imm6diate et h long terme de la fibroscopie digestive haute. Une 6tude en double aveugle chez 81 malades. Gastroenterol. Clin. Biol., 1983, 9, 61A (res).

69. TRONSDEN E., ROSSELAND A.R. , MOER A., SOL- HEIM K . - - Rupture of the spleen following endoscopic retrograde cholangiopancreatography (ERCP). Acta Chir. Scand., 1989, 155, 75-76.

70. TUFFNELL P.G. - - Salmonella infections transmitted by a gastroscope. Can. J. Public. Health., 1976, 67, 141-142.

The complications arising from esogastroduode- nal endoscopy (OGDE), intended at establishing a diagnosis, are of a mechanical or of an infectious nature or related to a per - - or post - - endoscopy lack of surveillance.

The frequency o f complications has been assessed on the basis o f surveys anterior to 1980 [49, 50]. The evaluation of the frequency remains, by and large, valid in as far as mechanical complications are concerned. On the other hand, however, infections which can be transmitted by endoscopy, have recently come under very much stricter scrutiny, dictated by the new categories o f patients examined (immunodepressed, candidates for organ trans- plants, AIDS), by the more selective means of identifying contaminated patients and finally by a better control o f the equipment which can be soaked and undergo standardized cleaning and disinfection procedures.

Complications arising from sedation, overlooked for a long time, are now better known since the continuous per-endoscopy monitoring (electrocar- diography, oximetry, blood pressure) o f patients subjected to sedative, morphinic or analgesic medication.

M E C H A N I C A L C O M P L I C A T I O N S

P e r f o r a t i o n o f t h e E s o p h a g u s

The overall frequency o f complications arising from endoscopy o f the upper digestive tract was 1.32 per 1000 according to the A .G.A. survey (1976) [49]. The frequency o f esophageal perforations was 0 . 1 % in 1980, which means that this

serious accident takes first place among the complications. It is worth recalling that, despite a much earlier diagnosis, the mortality rate due to instrumental and spontaneous perforations of the thoracic esophagus remains high : f rom 64 % o f deaths in the 1970 series of Abbott [1] , it fell to 19 % in the G E E M O (European Study Group of Diseases of the Esophagus) series [30J; furthermore, a late diagnosis usually means an esophagectomy is necessary [51] .

The widespread availability o f fine calibre fiberscopes and video-endoscopic instruments has considerably reduced the frequency o f this type of accident, if one excludes complications resulting from therapeutic treatment (dilatation, laser-treatment o f tumours, endoprosthesis setting etc.).


The following points must be stressed : - - The perforation during a diagnostic endoscopy

occurs most frequently in the region of the cervical esophagus : tearing o f a piriform sinus or of the posterior side o f the cervical esophagus, bulging out beneath the protrusion o f an osteophytic cervical vertebra. In the majority o f cases, the endoscopist is surprised by the absence of any sensation of rupture, and it should be stressed that an esophago- gram with soluble contrast compounds reveals a perforation in only about half o f the cases [50]. Esophageal perforations during biopsy of ulcerous or tumoral lesions have not been reported since 1964 [52].

- - lmpaction o f the fiberscope.

With thin calibre, flexible endoscopes, they are liable to flip back on themselves into retrovision during their insertion, and wedge the distal extre- mity in the lumen o f the esophagus. When the impaction is in the cervical area, the progressive withdrawal is possible but it should not be conti- nued if any resistance is encountered. In that case, it is preferable to use a second endoscope and push both instruments into the gastric cavity, so that the first can straighten out, and then withdraw it without causing any damage to the walls o f the esophagus (42). Impaction can also happen in a hiatal hernia, when the instrument is inserted in retrovision into the hernia. The impacted instrument is extricated in the same manner as for the esophagus.

Gastrointestinal Perforation

The cumulative frequency of hemorrhages and perforations which give rise to complications in routine upper digestive tract endoscopy varies from 0.008 to 0.03 % [37, 52, 65]. So, the actual instrumental perforation itself is therefore rather exceptional and usually occurs in the area of the anterior part of the stomach, particularly on a extended ulcerative lesion in the cranial region [50]. Some perforations o f the small intestine [41] or the caecum [59] have been reported. These perforations are the consequence o f excessive insuffiation on a pre-existing lesion o f the small intestine or the colon [3] or again, an adenocarcinoma o f the caecum [59]. In practice, these accidents can be avoi- ded by controlling the duration and the extent of the insufflation, and, eventually, watching the reac- tions of a lightly anesthetized patient. Finally, if multiple gastroduodenal lesions indicating a lym- phoma are found, this should evidently arouse sus- picion to the presence o f intestinal lesions with a high risk of perforation.

Replacing the biopsy o f the small intestine with a probe, by a perendoscopic biopsy [8, 29, 47, 48, 63, 64] in the diagnosis o f diffuse affections of the jejunal mucosa offers a very small risk of perforation.

Scott et al. (1993) have reported, however, 2 cases of duodenal and jejunal perforation following

a biopsy with forceps in patients with coeliac di- sease [64].

When confronted with abdominal pain, whether early or late, a radiography o f the abdomen without preparation should be carried out.

P n e u m o p a t h y re lated to inha la t ion

Although this is relatively frequent in intensive care units or in patients being examined for a digestive hemorrhage (sclerosis o f varices, etc.), this accident is somewhat exceptional in routine endoscopy of the upper digestive tract.

Among the factors which predispose to this accident, are :

- - esophageal (organic stenosis, achalasia with a megaesophagus) or gastric stasis and, finally, the considerable gastro-esophageal refluxes in patients subjected to prolonged or marked sedation. The endoscopy should in any case be carried out while disposing of a second aspiration unit, with a very fine calibre probe allowing immediate bronchial aspiration.

RARE COMPLICATIONS

This review cannot draw up an exhaustive list, but some rare complications, however, need to be pointed out :

- - T h e swelling of the salivary and parotid glands is not exceptional following OGDE although it is rarely reported [53]. Three pathogenic mecha- nisms have been proposed : salivary gland retention of thickened secretions, fits o f coughing followed by venous congestion o f the glandular parenchyma, and excessive parasympathetic stimulation during O G D E and vasodilatation. This anomaly regresses spontaneously.

- - Splenic trauma complicated by a subcapsular or an intra-abdominal hemorrhage is a well-known complication of fibercolonoscopy [14, 27, 32, 44, 60, 66] or, more rarely, o f endoscopic retrograd cholangio-pancreatography (ERCP) [56, 59]. Lewis (1991) has reported a case o f splenic decapsulation following OGDE, which was brought under control by surgical treatment [45].

A case o f a subserous membrane hematoma of the distal duodenum, following OGDE, has been recently reported by Pi~iikkOnen et al. [57].

Penson J.G. et al. (1992) have reported 7 cases of Mallory - Weiss syndrome following OGDE in a series of 10,000 consecutive examination carried out over 6 years [58]. The authors underlined the highest frequency o f this complication in elderly women with a hiatal hernia and the absence o f eructation salvos during the examination. The frequency o f this complication in their series is comparable to that o f previous publications.


COMPLICATIONS DUE TO INFECTION

The infections transmitted by endoscopes are not frequent providing the cleaning and disinfection procedures have been correctly carried out. Never- theless, infections can appear later and consequently go unrecorded.

Endoscopies of the upper digestive tract can pro- voke a not insignificant percentage o f bacteremias varying from 0 to 10 %, depending on the series, according to Essioux and Vergeau [28], who revie- wed 12 articles published between 1971 and 1984. Indeed, their mean frequency could be 4.64 %, following O G D E with biopsy [4, 11].

In the practice o f ERCP, the risk is particularly due to an insufficient decontamination [18, 26] and, exceptionally, the recommended procedures (tho- rough cleaning, decontamination, drying and appropriate storage) can be ineffective and the equipment must then be replaced [62].

On the other hand, the complications due to infections with clinical symptoms are much rarer and, according to the literature, their frequency is less than 0 .1% [19, 55, 65].

During the past fifteen to twenty years, several cases of infection transmitted by endoscopy have been reported. The majority concern the transmission of bacterial infections and, particularly, Pseu- domonas Aeruginosa and Salmonella sp. The contaminations by Pseudomonas Aeruginosa (PA) are mostly related to ERCP examinations [2, 17, 18, 24, 26]. All in all, 45 cases have been reported of which 4 resulted in death. The risk, although mini- mal, exists during O G D E and a PA' septicemia has been reported following endoscopy in a patient suffering from acute leukemia [34].

Many examples o f septicemia and even Salmo- nella epidemics have been described following various endoscopic examinations : gastroscopies, ERCP, sigmoidoscopy and colonoscopy [6, 12, 15, 21, 25, 38, 39, 54, 61, 70]. Of the 72 cases which have been reported, 38 had clinical symptoms and 1 case resulted in death [25]. Here again, the decontamination procedure was often defective.

Another unrecognized risk o f infection concerns Helicobacter pylori. An indisputable example of cross infection by esogastroduodenoscopy has been provided by the series of Langenberg et al. (1990) [43]. D. Graham et al. [33] have confirmed the role played by endoscopy in the iatrogenic contamination by H. pylori, with primary infection and epidemic achlorhydria.

In a prospective study, Gullini et al. [35] have shown that before cleaning and disinfection, H pylori was found on 33.3 % of fiberscopes and 11.1% of biopsy forceps. This data confirms the merit o f cleaning and disinfection procedures in preventing the risk o f H. pylori transmission from one individual to another.

Up to now, no documented observation of the transmission o f the H I V virus by endoscopy has been reported by the authors who have assessed this risk [16, 36]. Nevertheless, the time-lag between infection by the H I V virus and the appearance of seropositivity or o f immuno-deficiency symptoms, creates a risk o f underestimating this contamination.

The virulence o f the H I V virus is destroyed after being exposed to routine disinfectants even when their concentration is lower than the recommended standards. Endoscopes contaminated with a high level of HIV virus and soaked for 30 minutes in glutaraldehyde did not have any viable forms o f the virus remaining after the disinfection. Several prospective studies have demonstrated that 6 to 8 minutes of cleaning are sufficient to rule out this risk.

The hepatitis B virus has a higher virulence than the HIV virus when it is injected by the parenteral route. Despite the widespread use of endoscopy, only one documented case o f endoscopic contamination has been reported [10]. The recommended disinfection time is 30 minutes. A prospective study undertaken by Lok et al. [46] has shown that the patients examined after a patient carrying the Hepa- titis B virus, retain a negative serology six months after the examination providing the equipment had been properly disinfected.

A case of contamination by the C virus following ERCP has been recently reported [67].

COMPLICATIONS DUE TO SEDATION

The choice o f sedation during endoscopy of the upper digestive tract depends on psychological factors, the routines o f the medical school and the eventual compliance of the patients towards checkups.

Trinchet et al. [68] have shown that one month after an endoscopy o f the upper digestive tract without premedication, 5 % of the patients refused the eventuality o f a new examination. Carrying out an endoscopy under sedation poses a common pro- blem to all the patients regardless of the form of sedation employed :

- - t h e restrictions with regard to driving a car and losing a day because o f post-endoscopic indis- position (diminished degree of wakefulness, amne- sia, etc.).

- - The light sedation is obtained by intra-muscu- lar administration o f Midazolam : this requires the per- and postendoscopic surveillance by well-trai- ned staff, the placing o f a venous catheter and the possibility to administer oxygen on demand.

- - T h e deep sedation - - neuroleptanalgesia or general anesthesia - - requires the presence of an anesthesiologist and electrocardiographic, oximetric and blood pressure monitoring of the patient during the examination as well as during recovery.


Several studies have revealed arterial oxygen desaturation during the examination, indicating tem- porary hypoxia. This is observed in about 15 % of cases during a routine endoscopy [5] ; it is more frequent and more severe in patients undergoing a total colonoscopy [20]. In a prospective study of 82 patients who underwent an O G D E (mean duration : 8.5 +_ 0.42 min.), under Midazolam IV anesthesia (mean dose : 6.3 +__ 0.15 rag), Dhariwal et al. (1992) [23] recorded a desaturation > 4 % during 15.68 % o f the time o f the examination. A m o n g the factors which predispose to a per endoscopic desaturation are the fol lowing: age (> 65 years), anemia (hemoglobin < lO/dl) and obesity (body index > 28).

Indeed, severe hypoxemia causes cardiac rhythm disorders and ischemic changes which can lead to cardiac arrest. The risk o f apnea and cardiac arrest is not negligeable, especially in patients over the age o f 60 : F.L. lber et aL reported 10 cases in a series of 10,000 consecutive endoscopies. Four occurred during the administration o f medication and the other six, at the end o f the examination. The study by Bell (1990) [7] also underlines the importance of the risk : more than 50 % of the deaths associated with endoscopy o f the upper digestive tract are the result o f cardiopulmonary problems.

Consequently, administering sedatives (Midazo- lain) or morphinics must be adapted to the gravity o f the technical procedure, the age of the patients and the clinical context (cirrhosis, medication absorbed, etc.) and must be supervised by an anesthesiologist, with electrocardiographic and oximetric monitoring.

C O N C L U S I O N

Mechanical complications o f O G D E have now become much rarer due to the use o f video-endoscopes or fine calibre instruments. Special attention should be paid to exceptional complications published case by case - - which sometimes justify a rapid surgical intervention (duodenal hematoma, tearing of the splenic capsule). Endoscopists should be encouraged to publish such observations which would contribute to the general success of endoscopic practice.

The two major problems which remain are infection and sedation accidents. In both cases, preven- tion is not only possible but the rules are now correctly defined. They may burden the daily routine but the safety of the patients is at stake, all the more that O G D E is widely practiced outside hospi- tal units.


Documents

Les complications de l’examen endoscopique de routine du tractus digestif supérieur