Giardiasis in endoscopy patients : A cotnparison of diagnostic techniques

The results of a study comparing techniques for diagnosis of 
giardiasis during endoscopy are presented. Methods of diagnosis included examination 
of duodenal biopsy impression smears, culture of both biopsies and 
aspirates, and examination of preserved aspirate. Giardiasis was diagnosed in 
three of 80 patients; in each case the impression smear was positive. Cultures of 
duodenal biopsies were positive in two patients; in the third patient a biopsy was 
not available. The preserved aspirate was positive in only one patient, while the 
aspirate cultures were never positive. It is proposed that during diagnostic 
endoscopy, routine submission of a duodenal biopsy impression smear from 
patients with unexplained small bowel diarrhea or abdominal discomfort is 
appropriate.

commonly isolated intestinal protozoan parasite m North America ( l ).Infected individuals may be asymptomatic or may exhibit a spectrum of disease severity.T he clinical presentation ranges from a mild, selflimiting enteritis wit h complete recovery, tn a chronic malabsorption syndrome (1,2).T he majority of cases of giardiasis arc diagn osed by routine ova and parasite sc reenrng of stool, a rapid and inexpensive test.A negative result does not in variabl y exclude gia rdiasis as an infected patient does not a lways pass cysts, even when symptomatic (3 ).When deali ng with a patient who has gastroin test inal symptoms suggestive of giard iasis but a negative stool exam, the clinic ian is faced with a dilemma -is additional testing (often expensive and time-consum ing) needed to exclude the possibili ty of giarcl ia l infection ?
Many gastroenterologists include e ndoscopic examination of the small bowel in their work-up of patients with unexplained gastrointestina l d isorders.Thus, to assess techniques for identifying G lamblia infect io n in such patients the authors compared various techniques ava ilable during endoscopy.Duodenal aspirates were obtained during endoscopy and examined fo r trophozoites ( 4 ).Histological examination and impression smears of duodenal biopsies were assessed for trophozoites (5).As well, in accordance with the recent results of Gordts et al (6) , cultures were employed co identify trophozoites from intesti nal aspirates.
Since cultures of duodenal aspirates or biopsies a re conside red co be the most sensitive of these tests (6), they we re compared to the examination of preserved samples of d uodenal aspirates and impression smears of duodenal biopsies.In this pape r the results are presented of a prospective study of giardiasis in 80 patients undergoing endoscopy fo r va rious gastro intestinal symptoms compatible with giardiasis.

PATIENTS AND METHODS
Study description: Eighty patie nts with unexplained diarrhea, abdomina l discomfo rt, abdominal gas, malabsorption or weight loss, who required an upper endoscopy as part of the ir gastrointestinal assessment, were included in the study.All procedures were performed in the Day Care Endoscopy U nit at the Foothills H ospital in C algary, Alberta.Patients undergoing repeat endoscopies for follow-up of gastric ulcers were excluded from the study, as were patien ts with diseases of known peptic and hepatic origin.
The study population consisted o f 36 women and 44 men, aged 23 t o 68 years (mean age 41.3 years).A s a result of an ongoing study of A IDS at the Foo thills Hospital there was a high percentage (1 5% ) of HIV-pos iti ve pa tients.G iardiasis had not been diagnosed in a ny of the patients at the time o f endoscopy, although th e extent of assessment fo r giardiasis was va riable.
The endoscopy procedure required overnight fas ting, followed by sedation with meperidine and benzodiazepine.Endoscopy was carried out in standard fashion using an O lympus panendoscope.The scope was advanced into the second stage of the duodenum where samples were taken for each of the fo ur diagnostic techniq ues under assessment.Twenty millilitres of sterile saline were injected into the biopsy c hannel of the endoscope and 5 to l 0 ml of duodenal aspira te were wi thdrawn by suction into a sterile d isposable suction trap.Since th e saline solution originall y in use contained bacte ricidal components, it was replaced with a preservative-free saline to minimize difficult ies with c ulture of trophozoites.O ne millilitre of the aspirate was removed asepticall y and placed in a tu be of c ulture media (TYI-S-33 media supplemented wit h 10% C LEX, NCTC 109 modified vitamin mix and an t ibiotics ) (6). S ince most a ntibiotics have no effect on the growth of giard ia in culture (7), the following ant ibiotics were added to the culture med ia: penicillin 300 U/mL, streptomycin 300 µ g/mL, genta mycin 40 µ g/mL, tobramycin 100 µ g/mL, a mphotericin 1 µ g/mL.SAF preservati ve (1 5 g sodium aceta te, 20 mL glacial ace tic acid, 40 mL 40% formaldehyde, 925 mL deionized wate r) was added to the remaining aspirate to make a one in four dilution of duodenal fluid to preservative.
Fina lly, three duodenal biopsies were obtained using sta ndard endoscopy biopsy fo rceps.The first biopsy was used to make an imprint smear ensuring tha t th e mucosa!surface was smeared rathe r than touch ed against the glass slide.The slide was a llowed to air dry.T he second biopsy was placed in a tube of culture media and the third biopsy in 10% fo rmalin and sent fo r routine histopathology.The preserved aspirate and impression smear were sent to the microbiology laboratory for standard examination for paras ites, and we re examined within 2 to 4 h by one of three technicians.The two culture tubes were incubated at 37°C until examination.T hey were examined fo r trophozoites under a Zeiss inverted microscope (X 400) within 2 h of collection.The culture tubes were maintained at 37°C for 10 days and examined daily for trophozoites by the same individual.
The study was designed to assess procedures already in practice, so no attempts were made to alter the standa rd protocol for the collection or exa mination of aspi rates a nd biopsies.The only n ew addition to the protocol was the collection of an additional aspira te a nd hiopsy fo r culture.

RESULTS
Gia rdias is was diagnosed in Lhree of 80 patients examined.In each case the hiopsy impression smear was positive.C ult ures of duodenal biopsies were pos itive in two patients.ln the th ird pa ti ent a biopsy fo r culture was not available.The preserved aspirate was positive in only one of the three posi• tive patie nts, while the aspirate cultures were never pos itive.Statistical analysis was not performed because of the low attack rate a nd the sample size.Routine hiscolog1cal examination of duodenal biopsies revealed trophozoites in only one of the three patients.
The results of the study reflected the d ifficulty of obtaining standa rdized samples in a busy referral facility.Stool samples were rarely submitted at the time of endoscopy.This is not an unusual problem considering the preparat ion of patients for endoscopy (laxatives, enemas and/or fasting).In addition, the procedures which required more t ime, such as aspirates, were excluded in several pa tients who d id n ot tolerate the procedure well.The c ul ture procedure required t hawing of the culture •media 10 mins prior to use.In two or three cases this was not done and c ulturing was delayed .Staff changes led to a problem in one case in which culture samples were not collected .U nfo rtuna te ly, this patient was one of the few positive for giardias is.
The first posit ive case was a patient with vague abdominal discomfort and mild duodenitis.Trophozoites were seen in the impression smear, the biop• sy cult ure tube and the biopsy examined histo logicall y.There was no evidence of trophozoites in either the preserved or cultured aspirate.An axen ic in vitro culture of this isolate was established and na med C-1.Subseque nt stool examination revealed large numbers of giardia cysts.The patient was treated with metronidazole (250 mg tid) fo r 10 days and returned one month later with persistent symptoms.Endoscopy was repeated and on ce again trophozoites were seen in the biopsy impression smear and biopsy culture.Both aspirate samples were negative.A second axenic culture was eMablished named C-2.
The second case was an HIV-positive patient with mild diarrhea.Both biopsy samples were positive while the aspirates were negative.The authors were unable to establish an axenic culture.Stool samples were not available at the time of endoscopy, but earlier stool samples had been negative.No trophozoites were observed by histology.A third case of giardiasis was seen m the endoscopy unit during this study.Both impression smear and preserved aspirate were positive, but samples were not available for culture.Histological examination did not reveal trophozoites.
Despite the presence of antibiotics in the culture media, there was a persistent problem with bacterial and fungal overgrowth.This was most noticeable in the aspirate cultures, often within 24 h.Yeast contamination occurred within 48 h in eight of 12 aspirate cultures from HIV-positive patients despite the use of amphotericin.

DISCUSSION
Considering the large numbers of reported cases of giardiasis in Alberta (1613 in 1986 of which 620 were from Calgary) (personal communication), the authors were surprised at the low prevalence (3.75%) of giardiasis in the endoscopy patients with various gastrointestinal problems compatible with giardiasis.Previous studies in the United States and Australia have shown prevalences ranging from 0. 7 to 15.5% (8)(9)(10).Known HIV-positive patients were included in this study because previous results ( 11) indicated that these patients do not have a greater incidence of giardiasis than HIV-positive individuals.
There are several explanations for the low rate of positivity in this study.The majority of cases of giardiasis are acute and diagnosed by stool examinauon, and would therefore not be seen in a referral clinic.There is also an age bias, in that the majority of infections are found in the pediatric population which is seen elsewhere, at the Alberta Children's Hospital.In Alberta almost 50% of reported cases of giardiasis occur in patients under rhe age of 15 years (12).
Nevertheless, the present results indicate that a small percentage of patients undergoing endoscopy are infected with G lamblia, and an appropriate diagnosis may be missed if biopsies are not taken.Prior to this study the technique most commonly used by the endoscopy unit for detection of G lamblia was examination of duodenal aspirates.Other methods were occasionally used, including impression smears and histological examination.The authors showed that two of the three positive cases would not have been detected if duodenal aspirates had been the sole specimens submitted.These results differ from those of other investigators who suggested that examination of duodenal aspirates was as sensitive as mucosa[ impression smears (9, 13 ).Kerlin et al (9) noted in their study that all 21 patients with pos1t1ve impression smears had positive duodenal aspirates.They used a polyethylene catheter to obtain the aspirates, rather than the endoscope the present authors used.Gerdts et al (13) also found that all of their patients positive on impression smear had positive duodenal aspirates, using a pediatric Watson tube modified with a double lumen tube and monitored by radioscopy.
Karnath et al ( 14) compared four methods for detecting giardia, namely: stool exam, duodenal aspirates, jejuna!biopsies and impression smears of jejuna[ mucosa.They acknowledged, as found here, that the impression smear was the most sensitive and easily read test.However, they promoted the use of the aspirate because it was almost as sensitive and less invasive.In their study, 12 patients were positive on smear and biopsy, while 10 were positive on aspirate.The aspirates were obtained using a modifieJ Crosby capsule and fluoroscopy.
There are several reasons that could explain the present low rate of positive aspirates by both direct examination and culture.The collection of duodenal aspirates through the biopsy c hannel of the endoscope involved several steps which might have influenced the sen- Diagnosis of giardiasis during endoscopy sitivity of the test.Obtaining aspirates without saline irrigation was difficult and time-consuming; thus many of the samples obtained had been diluted with saline.As a result the precise amount of intestinal fluid in the sample could not be determined and negative results may have reflected inadequate sampling.The endoscope was cleaned between cases with soap and disinfectants and, although it was rinsed with distilled water prior to use, it is possible that enough disinfectant remained to decrease trophozoite viability.
Jn addition, the aspirates sent to the laboratory for direct examination were preserved, since trophozoites are fragile organisms and may die anc.l break up if not examined immediately.Intestinal fluid is also known to contain components toxic to trophozoites (15).The major difficulty with this approach is that trophozoite motility cannot be used to facilitate their detection.
It would appear that the collection of duodenal aspirates would be more effective if a modified Crosby capsule or feeding tube was used as described earlier (13,14 ), although this would require fluoroscopy which may not be available.The use of a small polyethylene catheter inserted in the suction channel may improve the collection technique.
In contrast to the collection of aspirates, the collection of duodenal biopsies is a straightforward procedure easily accomplished during routine endoscopic examination .The most important requirement of preparing the impression smear is that the mucosa!surface be smeared on the slide.It is not surprising that biopsy cultures had a higher success rate when one considers the biology of the parasite.G lamblia burrows into the mucus lining the epithelium and attaches to the brush border, which protects it from the toxic intestinal contents.When a biopsy is taken the parasites remain in a protected environment, and are more likely to be established in c ulture.
Recently, a sensitive test for detection of giard ia antigen in stool has been developed, which has been shown to detect up to 30% more positive samples than routine ova and parasite screens (16).This test is not designed to replace ova and parasite screens as it is more expensive (approximate ly double the cost}, but it may significantly increase the sensitivity of giardiasis testing.For various reasons, including economics and limited availab ility of the kit, this test is still not in widespread use; neverthe less, it may become a useful tool in the diagnosis of these difficult cases of giardiasis.