Endosonographic possibilities in the pancreatobiliary area

ES is certainly the most accurate technique presently available for visualizing small lesions in the pancreas and (distal) bile duct. However, this technique can at present not be utilized for improving the earlier diagnosis of pancreatic carcinoma but is of great help in the preoperative localization of pancreatic endocrine tumours. ES makes significant contributions to the preoperative loco-regional staging of pancreatic, ampullary and distal biliary malignancies, but it has distinct limitations in large masses and more remote areas (superior mesenteric vein) and in the differentiation between malignant and inflammatory lesions.

Oligoclonality and skewed T cell receptor V beta gene segment expression in in vivo activated human intestinal intraepithelial T lymphocytes

Intraepithelial intestinal T lymphocytes (IEL) bearing alpha beta or gamma delta T cell receptors (TCR) are positioned to serve as a first line of defense against enteric pathogens. To investigate whether intestinal IEL are subject to antigenic selective forces distinct from that influencing (xp T cells in the peripheral blood (PBL), we performed a comparative analysis of V beta gene segment usage in IEL and PBL of immunologically normal donors by quantitative PCR. Primers for 22 different human TCR V beta gene segments of V beta gene segments families were used to analyze the repertoire of TCR beta chain transcripts in colonic IEL (c-IEL), in corresponding colonic lamina propria lymphocytes (c-LPL), and in peripheral blood lymphocytes. In each of the three individuals examined, a limited number (1-4 out of 22) of TCR V beta families predominated and accounted for more than 50% of the total beta chain transcripts from c-IEL, whereas in PBL and c-LPL a more even distribution of V beta gene families was observed. The dominating V beta gene families were V beta 2, V beta 3, V beta 6, V beta 8 and V beta 14. In one individual, V beta 3 comprised more than 40% of the entire repertoire of c-IEL beta chain transcripts. Sequence analysis of the predominant V beta 3 family in this individual revealed identical sequences in 13 of 17 clones analyzed. Human alpha beta TCR+ c-IEL could not be driven to proliferate or exhibit cytotoxic function in vitro however, PCR analysis for detection of lymphokine mRNA revealed constitutive production of several lymphokines known to exert trophic effects on intestinal epithelial cells and pro-inflammatory activities. Taken together, the striking degree of oligoclonality may indicate that the intraepithelial intestinal immune system is targeted to a limited set of hitherto unknown self- or foreign antigens present in the intestine and orchestrates intramucosal inflammatory and regenerative processes.

Concanavalin-A-extractable non-mucous glycoprotein concentrations in gallbladder bile of cholesterol gallstone patients

BACKGROUND

The relationship between protein concentrations and the nucleation activity of bile in cholesterol gallstone patients has already been investigated. Nucleation promoters are mucins and concanavalin A (Con-A)-extractable glycoproteins. Nucleation inhibitors are apolipoproteins. We wanted to investigate whether a change in concentration of apolipoprotein A-I (Apo A-I) or Con-A in the bile of cholesterol stone carriers is dependent on the nucleation time.

METHODS

Total protein was measured by fluorescence photometry, and Con-A-extractable glycoproteins were separated by their affinity to lectins and measured by photometry. Apolipoproteins were measured by radioactive competitive protein binding assay.

RESULTS

The protein concentrations in our bile samples were 2.41 +/- 1.08 mg/ml for the whole group, 2.73 +/- 1.07 mg/ml for a nucleation time less than 3 days, and 2.04 +/- 1.00 for a longer nucleation time. The concentration of the Con-A fraction accounted for 0.289 +/- 0.096 mg/ml, 0.306 +/- 0.081 mg/ml, and 0.274 +/- 0.109, respectively. The Apo A-I concentration was 52 +/- 64 micrograms/ml; 50 +/- 56 micrograms/ml for a nucleation time less than 3 days and 85 +/- 133 micrograms/ml for a longer nucleation time.

CONCLUSIONS

Obviously, individual protein fractions have an effect on the nucleation behaviour of gallbladder bile in cholesterol gallstone patients.

[Carbohydrate substitutes: comparative study of intestinal absorption of fructose, sorbitol and xylitol]

BACKGROUND

The carbohydrate substitutes fructose, sorbitol and xylitol are gaining more and more importance in the production of dietary food. But they can provoke gastrointestinal side-effects. In a randomized double blind study the rate of malabsorption of these sugars was compared and the concomitant symptoms were recorded.

SUBJECTS AND METHODS

25 healthy controls received 25 g of each sugar within 3 consecutive days. The intestinal absorption was determined by H2-exhalation tests and the clinical symptoms were recorded.

RESULTS

The rate of malabsorption was 84% for sorbitol, 36% for fructose and 12% for xylitol (p < 0.01 for sorbitol versus fructose and xylitol). 57% of the participants with pathological H2-test after sorbitol and 56% after fructose reported symptoms, while all of the 3 malabsorbers of xylitol were symptomatic.

CONCLUSIONS

There is an advantage to administering xylitol and fructose with regard to the intestinal absorption and concomitant symptoms as compared with sorbitol. H2-exhalation tests appear to be a reliable diagnostic tool to detect carbohydrate malabsorption and should find broader application in patients suffering from non-specific abdominal complaints.

Functional characterization of a muscarinic receptor stimulating gastrin release from rabbit antral G-cells in primary culture

In previous studies carbachol-induced stimulation of gastrin release from antral G-cells in primary culture suggested the presence of muscarinic acetylcholine receptors on this cell type. Therefore, we attempted to pharmacologically characterize the muscarinic acetylcholine receptor subtype involved. Enzymatically isolated rabbit antral mucosal cells (0.8% G-cells) were separated by counterflow elutriation yielding a fraction (1.7% G-cells) that was placed in culture on collagen-coated well plates. After 24-36 h of culture 13.0 +/- 2.4% of total adherent cells were immunoreactive for gastrin as shown by immunocytochemical staining using the avidin-biotin complex method. In this preparation basal gastrin release ranged from 3.3 +/- 0.3 to 4.1 +/- 0.3% of total cellular content. Maximal gastrin release in response to the acetylcholine receptor agonist carbachol (10(-4) M) or the selective muscarinic receptor agonist arecaidine propargyl ester (10(-4) M) was 8.5 +/- 0.4% and 7.6 +/- 0.4% of total cellular content, respectively. The EC50 values were 3.7 +/- 0.5 x 10(-6) M carbachol and 1.8 +/- 0.4 x 10(-6) M arecaidine propargyl ester. At a concentration of 10(-6) M the non-selective muscarinic receptor antagonist atropine and the muscarinic M3 receptor preferring antagonist hexahydro-sila-difenidol (HHSiD; M3 > or = M1 > M2) completely inhibited gastrin release in response to carbachol (Ki values: 52 x 10(-9) M atropine and 29 x 10(-9) M HHSiD) and arecaidine propargyl ester (Ki values: 11 x 10(-9) M atropine and 13 x 10(-9) M HHSiD).(ABSTRACT TRUNCATED AT 250 WORDS)

Endoscopic lithotripsy of bile duct stones using a new laser with automatic stone recognition

Biliary stones can be removed in 85% to 90% of patients using endoscopic sphincterotomy; in the rest alternative methods are required. Thirty-eight consecutive patients in whom conventional methods had failed underwent laser lithotripsy with a new laser system. A flashlamp-pumped pulsed laser with rhodamine 6G as dye (594 nm) has a tissue-stone recognition system that can identify bile duct stones by analyzing backscattered light and interrupt the pulse in case of tissue contact (Lithognost, Telemit, Munich, Germany). Access of the 0.25- or 0.30-mm-diameter laser fiber to the stones was achieved perorally in 18 patients. In 13 of these cases, eccentrically located stones in the middle or proximal common bile duct were targeted with a 3.4-mm miniscope introduced through a standard duodenoscope. Fluoroscopically guided peroral lithotripsy was performed in 5 patients with stones in the distal common bile duct that could be approached with a standard ERCP catheter. Percutaneous cholangioscopic laser lithotripsy was carried out in 20 patients with stones not amenable to retrograde techniques. The mean number of bile duct stones per patient was 3.6, and the average diameter of the largest stone of each patient was 25 mm (range, 8 to 52 mm). The bile ducts were cleared in all but 1 patient in a mean number of 1.3 sessions lasting 15 to 115 minutes (mean, 60). No laser-related complications were observed. The Lithognost laser was successfully used in 37 of 38 patients referred for the removal of difficult bile duct stones.(ABSTRACT TRUNCATED AT 250 WORDS)

Double blind placebo controlled study on the effect of the nitric oxide donor molsidomin and the 5-HT3 antagonist ondansetron on human esophageal motility

The aim of the present study was to investigate the effects of acute administration of the NO-donor molsidomin (2 mg) and the 5-HT3 antagonist ondansetron (8 mg) on esophageal motility and lower esophageal sphincter pressure (LESP) in 10 healthy volunteers by stationary side hole manometry in a double-blind placebo controlled study design. LESP, contraction amplitudes (5, 10, 15 cm above the LES) and propagation velocity (10-15 cm above the LES) for dry and wet swallows were analysed and additional blood samples were taken for determination of plasma levels of VIP and gastrin. Molsidomin significantly decreased basal LESP from 16.8 +/- 1.6 mmHg to 11.4 +/- 1.0 mmHg, while ondansetron had no influence. Molsidomin also reduced contraction amplitudes of dry swallows (saline 61.9 +/- 7.2 mmHg, molsidomin 40.1 +/- 8.1 mmHg), while it did not influence contraction amplitudes of wet swallows. Ondansetron had no effect on contraction amplitudes. Both substances did not influence propagation velocity of wet swallows, while they reduced propagation velocity of dry swallows significantly (saline 3.9 +/- 0.4 cmls, ondansetron 3.1 +/- 0.2 cm/s, molsidomin 3.1 +/- 0.2 cmls). There were no effects on plasma levels of gastrin or VIP. These data strongly suggest a possible therapeutic role of molsidomin in the treatment of esophageal motility disorders. Effects of ondansetron have to be further evaluated in patients with disturbed esophageal motility.

[Substitution of factor XIII concentrate in treatment refractory ulcerative colitis. A prospective pilot study]

BACKGROUND

Severe acute inflammation in ulcerative colitis is often associated with major intestinal blood loss. Studies on hemostasis present a deficiency of F XIII which is important for clot formation and wound healing.

PATIENTS AND METHODS

A total of ten patients had been treated with 5-aminosalicylic-acid and corticosteroids consequently for three weeks. Thereafter a clinical improvement did not occur, the colitis activity index (CAI) and the endoscopic score (ES) remained elevated (9.9 +/- 1.5 points and 8.9 +/- 2.3 points, respectively). Because of this therapy-resistant active stage of disease in an open and prospective pilot trial F XIII concentrate (1,250 IU, Fibrogammin HS, Behringwerke, Germany) was additionally administered intravenously for ten days.

RESULTS

The additional substitution therapy resulted in a significant improvement of complaints, the stool frequency decreased from 9 +/- 4.1 to 2.4 +/- 1.5 (p < 0.001). Both, the clinical activity index (2.8 +/- 1.6; p < 0.0001, vs day 0) and the endoscopic score (4.4 +/- 2.2; p < 0.005, vs day 0) declined significantly during the F XIII substitution. The F XIII activity was markedly reduced initially (46.1 +/- 17.4%) and showed a significant increase after the substitution (171 +/- 41.7%; p < 0.001).

CONCLUSION

The results may suggest that substitution therapy with F XIII concentrate can be beneficial in patients with therapy-resistant active stage of ulcerative colitis and proven F XIII deficiency. To verify this preliminary results, controlled clinical trials will have to be performed.

Exendin-4 and exendin-(9-39)NH2: agonist and antagonist, respectively, at the rat parietal cell receptor for glucagon-like peptide-1-(7-36)NH2

Exendin-4 is a novel peptide from Heloderma suspectum venom which is 53% homologous with glucagon-like peptide-1 GLP-1-(7-36)NH2, a stimulant of cAMP-dependent H+ production in rat parietal cells. It was the aim of the present study to determine whether this effect of GLP-1-(7-36)NH2 is shared by exendin-4, and whether the responses to either peptide are blocked by exendin-(9-39)NH2, a competitive specific exendin receptor antagonist. In enriched rat parietal cells H+ production was measured indirectly by [14C]aminopyrine accumulation. cAMP production was determined by radioimmunoassay. [125I]GLP-1-(7-36)NH2 was prepared using chloramine T followed by high pressure liquid chromatography (HPLC) purification. Exendin-4 (10(-12) - 10(-8) M) stimulated [14C]aminopyrine accumulation in a concentration-dependent manner (EC50 = 7.6 x 10(-11) M). At the maximally effective concentration (10(-9) M) exendin-4 was as effective as GLP-1-(7-36)NH2 reaching 70-80% of the response to 10(-4) M histamine. Likewise, exendin-4 (10(-11) - 10(-7) M) stimulated parietal cell cAMP production up to 2.8-fold. Maximal stimulation by exendin-4 of [14C]aminopyrine accumulation was not affected by ranitidine (10(-4) M), but was concentration-dependently reduced by exendin-(9-39)NH2 (10(-11) - 10(-7) M). At the maximal concentration, exendin-(9-39)NH2 completely abolished the responses to 10(-9) M exendin-4 and to 10(-9) M GLP-1-(7-36)NH2 while not altering stimulation by 10(-4) M histamine. Binding of [125I]GLP-1-(7-36)NH2 to enriched parietal cells was displaced by exendin-4 (Ki = 4.6 x 10(-10) M) as well as by exendin-(9-39)NH2 (Ki = 4.0 x 10(-9) M).(ABSTRACT TRUNCATED AT 250 WORDS)

Rat parietal cell receptors for GLP-1-(7-36) amide: northern blot, cross-linking, and radioligand binding

The intestinal peptide hormone glucagon-like peptide-1 (GLP-1) (7-36) amide is a potent stimulus of H+ production in isolated rat parietal cells, suggesting the presence of specific GLP-1-receptors on this cell type. Our aim was to characterize these receptors. Enzymatically isolated rat gastric mucosal cells (F0) were fractionated by counterflow elutriation, resulting in five fractions (F1-F5) according to increasing cell diameter and parietal cell content (3, 5, 4, 27, 81%). Additional density gradient centrifugation of F4 yielded enriched chief cells (74%; parietal cells: 1%; F6), whereas density gradient centrifugation of F5 almost purified parietal cells (97%; chief cells: 1%; F7). Northern blot of total cellular RNA from F0-F7 with a probe specific for the GLP-1-(7-36) amide receptor revealed two RNA species of 2.7 and 3.6 kb. These messages were present to some extent in small cells (F1, F2), much more pronounced in F5, abundant in F7, barely detectable in F3 and F4, and absent from F6. Cross-linking of 125I-labeled GLP-1-(7-36) amide to parietal cell membranes revealed a single 59-kDa band that was abolished by unlabeled GLP-1-(7-36) amide. Throughout fractions F1-F7 specific binding of 125I-GLP-1-(7-36) amide was correlated with parietal cell content (r = 0.99; P < 0.01) and H+ production ([14C]aminopyrine accumulation) in response to GLP-1-(7-36) amide or histamine (r = 0.96; P < 0.01). Binding was maximal in purified parietal cells (F7), whereas almost no binding was detectable in enriched chief cells (F6). In F7, Scatchard analysis revealed a single class of high-affinity binding sites (KD = 2.8 +/- 0.6 x 10(-10) M, Bmax = 6.8 +/- 1.4 fmol/10(6) cells, 4,096 +/- 793 receptors/parietal cells). The following half-maximal inhibition values were found for GLP-1-(7-36) amide and (1-37) and (1-36) amide: 6.6 +/- 0.9 x 10(-10), 1.4 +/- 0.7 x 10(-7), and 2.6 +/- 0.4 x 10(-7) M, respectively. Pancreatic glucagon, GLP-2, and oxyntomodulin, products of the proglucagon gene, were 3-4 log units less potent displacers while gastric inhibitory peptide, vasoactive intestinal peptide, and secretin were ineffective. We conclude that rat parietal cells are equipped with specific high-affinity receptors for GLP-1-(7-36) amide, which, in addition, are present in as yet unidentified small cells (F1, F2) but not in chief cells.

RUDER--a prospective, two-year, multicenter study of risk factors for duodenal ulcer relapse during maintenance therapy with ranitidine. RUDER Study Group

In a prospective study of the risk factors for duodenal ulcer relapse during maintenance (150 mg daily) ranitidine therapy, 1899 patients with chronic ulcer disease were recruited to a multicenter, German trial. Healing of all ulcers was confirmed endoscopically; endoscopy was also obligatory after one and two years or if the patients presented in the interim with symptoms of ulcer relapse. By the end of the first year, 247 patients had experienced at least one relapse and, by the end of the second year, 432 patients had relapsed at least once. The crude one- and two-year relapse rates were 13.0% (95% CI 11.5-14.5) and 22.7% (20.9-24.6%), respectively. Univariate analysis indicated that all seven prospectively defined risk factors were associated with an increased two-year relapse rate; of these, duodenal erosions distant from the healed ulcer [odds ratio (95% CI): 2.23 (1.59-3.15); P < 0.0001], smoking, past or present [1.46 (1.12-1.90); P = 0.0050], psychological stress [1.38 (1.09-1.74); P = 0.0085], heavy physical labor [1.45 (1.06-1.98); P = 0.0219], and absence of NSAID intake [1.54 (1.01-2.29); P = 0.0464] were independent risk factors on stepwise logistic regression analysis, whereas persistent symptoms at healing [1.29 (1.03-1.62), P = 0.0310] and frequent prior relapses [1.45 (1.01-2.04); P = 0.0454] were not. Multiple relapses in 107 patients [5.63% (4.60-6.67%)] were associated with duodenal erosions, smoking, stress, and heavy physical labor.(ABSTRACT TRUNCATED AT 250 WORDS)

Release of bombesin-like immunoreactivity from synaptosomal membranes isolated from the rat ileum

In the enteric nervous system, direct effects on peptidergic neurotransmitter release are difficult to assess since the neuronal network predisposes to numerous interactions between the various transmitter systems. The aim of the present study was to examine the release of bombesin-like immunoreactivity from isolated nerve synapses of the enteric nervous system. Enriched synaptosomal fractions were obtained by using homogenized tissue from rat ileum, which was subjected to various steps of differential and sucrose density centrifugation. Specific binding of [3H]saxitoxin served as a marker for neuronal membranes. For comparison, the content of bombesin-like immunoreactivity was determined. Both the enriched synaptosomal fraction (mitochondrial fraction II or P2) and the purified synaptosomal fraction (F2), obtained after discontinuous sucrose density centrifugation, showed substantial enrichment of the neuronal marker [3H]saxitoxin and bombesin-like immunoreactivity. The basal release of bombesin-like immunoreactivity was 52 +/- 17 pg/mg (100%). KCl-evoked depolarization (65 mM) significantly stimulated the release of bombesin-like immunoreactivity to 142.2% (P < 0.05, n = 17). The release was abolished in Ca(2+)-free medium. Stimulation of the release of bombesin-like immunoreactivity was also observed in the presence of the Ca2+ ionophore A-23187 (10(-6) M: 129%, P < 0.05, n = 17), supporting the role of Ca2+ in the release process. Cholinergic stimulation with carbachol elicited a significant dose-dependent release of bombesin-like immunoreactivity (10(-8) M: 106%, 10(-7) M: 175%, P < 0.05, 10(-6) M: 156%, P < 0.05, 10(-5) M: 115%, n = 14), which was reduced by atropine (10(-6) M: 99%, P < 0.01, n = 14). The basal value was 67 +/- 9 pg/mg (100%). The different effects of the muscarinic M1 receptor antagonist pirenzepine, which stimulated release of bombesin-like immunoreactivity in combination with carbachol 10(-6) M (10(-6) M: 123%, n = 10), and of the muscarinic M2 receptor antagonist AFDX 116, which attenuated release of bombesin-like immunoreactivity evoked by carbachol (10(-5) M: 66%, P < 0.01, 10(-6) M: 88%, n = 10), strongly suggest modulation of the release of bombesin-like immunoreactivity at the presynaptic receptor site through an excitatory muscarinic M2 receptor. The basal value was 46 +/- 9 pg/mg (100%). In summary, bombesin-like immunoreactivity can be released from these synaptosomes by both depolarization with KCl in a Ca(2+)-dependent manner and by cholinergic stimulation. The synaptosomes of intrinsic nerves of the gut offer an approach to study the release of neuropeptides and neurotransmitters at the subcellular level independent of the ganglionic network.

pHi and HCO3- dependence of proton extrusion and Cl(-)-base exchange rates in isolated rabbit parietal cells

In many cell types, the regulation of intracellular pH (pHi) is different in the presence vs. absence of HCO3-. We investigated the pHi and HCO3- dependence of proton extrusion and anion exchange rates in isolated rabbit parietal cells loaded with the pH-sensitive dye 2',7'-bis(carboxyethyl)-5(6)-carboxyfluorescein (BCECF). In Cl(-)-depleted parietal cells, the dependence of maximal proton efflux rate on pHi showed a strong inverse correlation but was identical in the presence and absence of HCO3-. Amiloride and Na+ removal inhibited proton extrusion rates to a similar degree with or without HCO3-, whereas 4,4'-diisothiocyanostilbene-2,2'-disulfonic acid (DIDS) had no effect. This suggests that a Na(+)-H+ exchanger is the major acid extrusion ion transporter under these experimental conditions. In Cl(-)-containing cells, there was also some Na(+)-independent, extracellular HCO(3-)- and intracellular Cl(-)-dependent, DIDS-inhibitable pHi recovery from an acid load, most likely due to intracellular Cl(-)-extracellular HCO3- exchange. Recovery from an alkaline load was primarily mediated by anion exchange, and the dependence of maximal anion exchange rates on pHi was very different in the absence and presence of HCO3-. In its absence, maximal anion exchange (Cl(-)-OH-) rates increased slowly over the tested pHi range from 6.4 to 7.8. In the presence of HCO3-, however, there was an S-shaped dependence of maximal flux rates on pHi, with a steep increase in flux rates between 6.8 and 7.5.(ABSTRACT TRUNCATED AT 250 WORDS)

Stimulation of rat parietal cell function by histamine and GLP-1-(7-36) amide is mediated by Gs alpha

It was the aim of the present study to determine in rat parietal cells whether Gs alpha, the stimulatory subunit of adenylate cyclase, mediates adenosine 3',5'-cyclic monophosphate (cAMP)-dependent parietal cell function in response to histamine and glucagon-like peptide 1 (GLP-1)-(7-36) amide. Cytoplasmic membrane from enriched (83 +/- 5%) rat parietal cells were incubated for 30 min with 30 microCi/ml [32P]NAD+ and 40 micrograms/ml preactivated cholera toxin (CT), a pharmacological tool for activation of Gs alpha. Subsequent sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and autoradiography revealed [32P]ADP ribosylation of Gs alpha represented by three proteins with molecular masses ranging from 42 to 45 kDa. In intact parietal cells, CT (10(-12)-10(-8) M) caused marked stimulation of [14C]aminopyrine accumulation and cAMP production confirming the functional importance of Gs alpha in regulation of H+ production. Identical membrane preparations were preincubated (2 h, 4 degrees C) in parallel with and without RM/1, a rabbit polyclonal anti-Gs alpha-antibody. Subsequently, adenylate cyclase was stimulated by histamine, GLP-1-(7-36) amide, CT, or forskolin. At a 1:10 dilution, the antiserum completely abolished adenylate cyclase activity in response to maximal concentrations of histamine, GLP-1-(7-36) amide, and CT while reducing forskolin stimulation by only 22.0 +/- 4.9%. At 1:50, RM/1 reduced responses to histamine, GLP-1-(7-36) amide and CT by 20-30% but failed to inhibit forskolin-stimulated enzyme activity. At 1:100, the antiserum was ineffective versus all stimuli.(ABSTRACT TRUNCATED AT 250 WORDS)

[Chemoimmunotherapy of malignancies of the gastrointestinal tract]

Stimulation of the immune response could be obtained in outpatients with inoperable locally advanced or metastasizing gastrointestinal tumors (n = 56), who were palliatively treated with cyclophosphamide (cy), 350 mg/m2 i.v., thymostimulin, echinacin i.m. and in part additionally with epirubicin, 15 mg/m2 i.v. At present, outpatients with pancreatic (n = 7) or colon carcinomas (n = 1) receive interleukin 2 (IL-2) intralesionally, 2 x 10(5) Cetus units (CU), days 3, 5 and 7 after cy, and by portable continuous infusion systems i.v., 3 x 10(5) CU/day, days 3-8 after cy. Subsets of lymphocytes slightly decreased in the patients' peripheral blood, recall antigen reactivity slightly increased. Patients' mean survival time was 5.7 +/- 1.7 months.

[Globus pharyngis: indications for flexible endoscopy?]

Flexible endoscopy of the upper gastrointestinal tract usually does not form part of the primary diagnostic evaluation of the globus syndrome. In order to find out the value of this investigation in the differential diagnosis of globus sensation, 51 globus patients were examined by flexible endoscopy of the upper gastrointestinal tract. Pathological findings requiring therapy were diagnosed in 70.6% of cases. The most frequent findings were reflux esophagitis (n = 24; 47%) and hiatal hernia (n = 25; 49%). In 16 cases (31.4%) these were accompanied by other pathological lesions. A total of 32 patients (62.7%) suffered from esophageal diseases as sole etiological factors of the globus syndrome, which allowed us to postulate a causative relationship in these cases. Flexible endoscopy, therefore, can contribute significantly to the differential diagnosis of the globus syndrome. The investigator should consider the potentially high diagnostic "yield" of flexible endoscopy for this indication, also.

[Pneumonia due to a rare atypical Mycobacterium in AIDS]

A 38-year-old man, HIV-positive for 6 years, developed fever and cough with deterioration in his general state. Chest radiography demonstrated an infiltration in the left upper lobe and computed tomography showed a septated cavity. Three bronchioalveolar lavages over 4 weeks recovered Klebsiella, Candida, Pseudomonas and Staphylococcus in the lavage fluid. Acid-fast rods were not found in any of the microscopic preparations. His clinical condition and the radiological findings deteriorated despite appropriate antibiotic administration. A further cavity occurred in the right upper lobe and the inflammatory infiltrations extended further. Although no acid-fast organism had been demonstrated, tuberculostatic treatment was begun (daily 300 mg isoniazid, 600 mg rifampicin, 900 mg streptomycin, 2 g pyrazinamide). His general condition and the radiological findings rapidly improved. Four weeks after culturing the lavage fluid atypical Mycobacterium xenopi was isolated. This case illustrates the difficulty of diagnosing an atypical mycobacterial infection. It takes time and effort, but it is of great importance because up to 50% of patients with AIDS contract such infection. Early and appropriate treatment will significantly improve quality of life and life expectancy.