CAFFEINE DOES NOT MODULATE NUTRITIVE BLOOD FLOW TO RAT GASTRIC SUBMUCOSA – A MICRODIALYSIS STUDY

Background and Aims: Coff ee irritates the gastric mucosa disrupting its barrier and increasing the risk of peptic ulcers. However, caff eine’s contribution to these eff ects has not yet been elucidated. In this study we looked at the local eff ect of caff eine on the microcirculation and nitric oxide production in rats together with systemic marker of oxidative stress malondialdehyde as possible mechanisms whereby caff eine might participate in mucosal barrier impairment. Materials and Methods: Four groups of rats were anesthetized and administered as a bolus four diff erent intraperitoneal doses of caff eine (0, 1, 10 and 50 mg kg b.wt.). The gastric submucosal microcirculation and nitric oxide production were then recorded for 2.5 hours by in situ microdialysis using the fl ow marker ethanol. At the completion of the experiments, plasma caff eine and malondialdehyde levels as well as morphological mucosal injury were determined. Results: There were no major diff erences in the macroor microscopic pictures of the mucosa among the groups. Local microcirculatory (ethanol out/in ratio) and nitric oxide monitoring failed to demonstrate statistically signifi cant changes as did measurement of plasma malondialdehyde in response to caff eine injections. Conclusions: Caff eine per se seems unlikely to contribute to the gastric mucosal barrier injury associated with coff ee consumption by alterations in nutritive blood fl ow, nitric oxide production or aggravation of systemic oxidative stress. This information is relevant for better understanding of the mechanisms involved in caff eine-mediated infl uences on gastric physiology in relation to the irritant eff ects of coff ee.


INTRODUCTION
Caff eine (contained in coff ee, tea, caff einated beverages, cocoa, chocolate etc.) is the most consumed stimulant drug of abuse worldwide.With regard to coffee, epidemiological data support its irritant role in stomach and esophagus in association with gastroesophageal refl ux leading to enhanced risk of ulcers and cancer in the affl icted areas 1,2 .Recently, this irritant eff ect was also confi rmed in young asymptomatic individuals 3 .The beverage is a complex mixture of possibly antagonistic substances comprising, apart from caff eine, isofl avones, polyphenols, diterpenes etc. exhibiting eff ects according to the type of coff ee bean processing 4,5 .To elucidate the cause of the irritant nature of coff ee to the stomach, focusing on the eff ects of isolated and purifi ed (pharmacologically) active components of coff ee would lead to more information.Of these components caff eine is the most studied.
Maintenance of the gastric mucosal barrier integrity is dependent on the balance between aggressive and protective factors represented by hydrochloric acid on one hand and adequate mucosal blood fl ow with suffi cient mucus production on the other.Caff eine, a known acid secretagogue, has long been suspected of causing mucosal hypoperfusion due to (micro)vascular impairment 6 .This hypothesis was later supported by electron microscopy 7 .More recent observations report suppressed acetylcholine (ACh)-induced mucus production 8 by caff eine and gastric mucosal transmembrane potential diff erence 9 , completing the barrier-braking mosaic picture of caff eine.However, the aforementioned notion has been challenged by experimental observations showing enhanced mucosal blood fl ow by caff eine 10 and a protective infl uence of this drug on mucosal barrier integrity 11 suggesting actually a preventive role for caff eine in gastric mucosal injury 12 .
Caff eine is a methylxanthine with pluripotent and possibly opposing pharmacological actions.It is a nonselective adenosine receptor antagonist, phosphodiesterase inhibitor, ryanodine-sensitive Ca 2+ channel activator and soluble guanylate cyclase (GC) inhibitor.As a consequence, these actions may interfere with nitric oxide (NO) production and/or its second messenger cyclic guanosine monophosphate (cGMP) pathway leading to modulation of a wide spectrum of mucosal barrier-related (patho)physiological eff ects exerted by NO including vascular tone regulation or modulation of oxidative stress level 13 .The literature, to the best of our knowledge, provides rather limited data on caff eine's eff ect on gastric (sub)mucosal microcirculation and has not included parallel monitoring of local NO release (using microdialysis), so far.Similarly, the putative eff ect of caff eine on oxidative stress necessitates clarifi cation.The aim of the present study was twofold: fi rst, to evaluate the possible impact of caff eine on gastric submucosal microcirculation, nitric oxide production and morphology, and second, to measure plasma malondialdehyde (MDA) as a marker of oxidative stress (lipid peroxidation).

Animals
Conventionally bred adult white male Wistar rats (Biotest s.r.o., Konárovice, Czech Republic) weighing 300-400 g, were used.The animals were housed in the animal quarters for at least 7 days prior to experiments under controlled environmental conditions.The rats had free access to standard pellets (ST 1-TOP, Velas, Prague, Czech Republic) except for the 16-18 h before experiments, when they were fasted.Tap water was provided ad libitum until the day of the experiment.All animals received humane care in accordance with the guidelines set by the institutional Animal Use and Care Committee of the Charles University, Prague, Faculty of Medicine in Hradec Králové, Czech Republic.All experimental procedures were approved by the same committee.

Animal preparation
All animals were anesthetized with single i.p. dose of pentobarbital (50 mg kg -1 , Nembutal ® , Abbott Laboratories, North Chicago, USA) and placed in a supine position on an unheated bed.They were kept under general anesthesia by cyclic i.p. administration of Nembutal ® (15 mg kg -1 h -1 ).Body temperature was monitored using a rectal thermometer probe (Ama-digit ad 15 th , Aprecision, Germany) and maintained at 37.5-38.0°C by means of a heating lamp.The trachea was cannulated with a short polyethylene catheter to ensure patent airways.

Substances
For microdialysis, ethanol-enriched normal 0.9 % saline (fi nal concentration 50 mmol l -1 ) 14 was used as a perfusion medium.Caff eine (purchased from Sigma-Aldrich, St. Louis, MO, USA) was dissolved in normal saline to obtain solutions with concentrations 0.5, 5 and 25 mg ml -1 for groups 2, 3 and 4, respectively.

Gastric submucosal microdialysis technique
Modifi ed technique originally described by Kitano et al. 15 was used.Briefl y, following 3-4 cm long midline laparotomy, stomach was exposed.Firstly, respecting the course of blood vessels, a 21 G needle was introduced into the submucosa of gastric fundus with care neither to penetrate through the mucosa into the lumen nor to make an additional opening in the serosa.Secondly, a microdi-alysis probe (MAB 11.8.10 with 6 kDa cut-off polyethylene suplhone membrane, active length 10 mm; outer diameter 0.5 mm; Microbiotech/se AB, Stockholm, Sweden) was cautiously inserted into the preformed tunnel (Fig. 1a).Finally, the probe was fi xed to the serosa with a suture.Continuous microdialysis was eff ectuated by perfusing the catheters with ethanol solution by means of CMA 102 microdialysis pump (CMA Microdialysis AB, Solna, Sweden) at a perfusion rate of 2 μl min -1 .For tissue equilibration, an initial 60 min period (without specimen collection) was allowed which was succeeded by a 30 min period to yield the baseline sample.Thereafter, sampling ensued for the next 2.5 h in 30 min intervals into sterile 200 μl polyethylene microvials.The specimens were aliquoted and stored at -70 °C (for nitric oxide) or -20 °C (for ethanol) until analysis.

Experimental design
The animals were randomly allocated to four groups (6 in each).After the collection of baseline microdialysis sample, the fi rst (sham-operated) group received by intraperitoneal injection an adequate volume of normal saline, whereas the second, third and fourth groups were intraperitoneally administered caff eine solutions with concentrations 0.5, 5 and 25 mg ml -1 (caff eine dose of 1, 10 and 50 mg kg -1 b. wt.) respectively.At the conclusion of experiments, the animals were sacrifi ced by blood withdrawal from abdominal aorta.The obtained plasma was stored at -70 °C for ensuing biochemical determination of caff eine and MDA.After the removal of stomach from the body, the mucosal surface of the glandular part was gently cleaned in cold tap water and closely inspected in search for macroscopic lesions.The tissues surrounding the probe (0.5 × 1.5 cm) with the probe itself in situ were then dissected and the biopsy immersed in 10% formaline for histological examination.

Probe performance stability and in vivo recovery determination for nitrate
The measurement of NO using microdialysis technique was validated in two consecutive steps employing other two groups (A and B) of pentobarbital-anesthetized rats.First, probe performance stability for nitrate was tested continuously for 7 h (measurements during equilibration period inclusive, group A, n = 5) in one experiment based on an assumption of stable NO production throughout the study period.Microdialysis sampling in gastric submucosa was realized in 30 min intervals at a perfusion rate of 2 μl min -1 .As perfusate, normal saline was utilized.Second, in vivo recovery of the same probe type was estimated (group B, n = 3) using zero-net fl ux method originally proposed by Lönnroth et al. 16 .Four perfusion media of increasing concentrations of sodium nitrate in sterile saline were consecutively applied as follows.After the initial tissue equilibration (1 h) with 10 μmol l -1 NaNO 3 , a 30 min sample was collected.The perfusion medium was changed for 15 μmol l -1 nitrate and following 30 min equilibration, another sample harvest (30 min) ensued respecting the probe's lag time (3 min).The experiment Caff eine does not modulate nutritive blood fl ow to rat gastric submucosa -a microdialysis study was completed with 50 and fi nally 120 μmol l -1 nitrate solutions.The results were plotted on a graph and probe recovery was read from the regression equation (slope gradient).Besides microdialysis, these two groups of animals underwent no further experimental treatment.All general steps (anesthesia, surgery, gastric submucosal microdialysis technique including probes but excluding perfusion media, sacrifi ce and analytical techniques) were equal to the experimental groups of the present study.

Nutritive blood fl ow measurement
Using fl ow marker (e.g.ethanol)-enriched perfusion medium, microdialysis off ers opportunity to indirectly assess tissue microcirculation (nutritive blood fl ow) in various organs.The principle of the so called "ethanol dilution technique" is based on the negative correlation between blood perfusion and ethanol effl ux from the probe represented by out/in ratio of ethanol concentrations 14,17,18 .

Biochemical analyses
Total NO was quantitatively determined as the sum of its stable metabolites (nitrite plus nitrate) in an enzymatic colorimetric assay, which involves an enzymatic conversion of nitrate to nitrite by nitrate reductase.Resulting nitrite is detected colorimetrically as a colored azo dye product of the Griess reaction that absorbs visible light at 540 nm.In the experiments commercially available ELISA kits (Assay Designs, Ann Arbor, MI, USA) were utilized only.All procedure steps were performed according to the manufacturer's instructions.The analyses of ethanol were carried out using GC-MS apparatus consisting of Varian 3300 Gas Chromatograph (GC) coupled to Finnigan MAT Magnum Mass Spectrometer (MS; Thermo Fisher Scientifi c, formerly Finnigan).Supelcowax™-10 (30 m x 0.25 mm x 0.25 μm fi lm thickness) GC Capillary column with helium as a carrier gas was employed.Injector and transferline temperatures were set to 230 °C for both.GC oven was programmed as follows: 55 °C, 2 min, 25 °C min -1 to 150 °C, hold for 1 min.The specifi ed MS parameters were 70eV for electron ionization mode (EI) and 209 °C for ion trap temperature.The resulted MS spectra were scanned in 15-70 a.m.u.mass range under 4 microscans s -1 of detection speed regimen.
Plasma caff eine concentrations were determined using a modifi ed HPLC method by Biederbick et al. 19 .Briefl y, caff eine and internal standard hydroxyethyltheophylline (aqueous solution 40 mg l -1 ) from plasma samples (100 μl + 50 μl) were extracted using 2.5 ml of dichlormethane.The organic phase was evaporated under nitrogen at 40 °C and the samples were reconstituted in 150 μl of mobile phase, and 30 μl were injected into the HPLC column.Analysis was performed on a 2695 Waters Separations Module equipped with 996 photodiode array detector and Peltier column-thermostat Jet-Stream (Thermotechnic Products).The mobile phase was made up of an aqueous solution of sodiumdihydrogenphosphate (0.5 mmol l -1 , pH 5) to acetonitrile (ratio 88:12) and was pumped isocratically at a fl ow rate of 0.9 ml min -1 .The temperature of the column was set at 40 °C.The quantifi cation of caff eine was performed at 270 nm.
Plasma MDA was determined as a secondary product of lipid peroxidation in an attempt to evaluate the level of oxidative stress produced by caff eine.The analysis was based on the reaction of MDA with thiobarbituric acid (TBA) producing a red MDA-TBA complex measured photometrically at three distinct wavelengths (485, 532 and 560 nm) and the absorbance corrected according to Allen's formula A corr = A 532 -[(A 560 -A 485 ) × 0.63 + A 485 ] for enhanced specifi city 20 .

Histological analyses
Stomach biopsies were taken to evaluate possible effects of caff eine on tissue morphology at microscopic level and for probe position verifi cation.The samples (0.5 × 1.5 cm) were taken from glandular segment surrounding the probe.The tissues were fi xed in 10 % neutral buff ered formalin and further treated according to standard procedures for hematoxylin-eosin (HE) stain.The sections were evaluated by a blinded professional observer and photographed at 125, 250 and 500-fold magnifi cation.The grading criteria were adopted from Natale et al. and were as follows: grade 0 for normal mucosa, grade I for lysis and segregation of cells on the luminal surface (with intact pit cells), grade II for damage confi ned to gastric pits with detachment of the surface epithelium and grade III Fig. 2. The graph depicts total gastric submucosal nitric oxide (NO) production (presented as a sum of nitrite and nitrate), as monitored in anaesthetized rats by in vivo microdialysis and confi rms its stability throughout the experiment.The arrows specify the time frame of in vivo recovery determination (measured in another study, see Fig. 3).Data are expressed as means ± SEM of 5 measurements.

Fig. 3. In vivo probe calibration ("zero net-fl ux" method).
The Δ microdialysate (dialysate -perfusate) total NO plotted against the perfusate total NO gives a gradient representing the probe's recovery (38 %).The x-intercept corresponds to the concentration of NO in the surrounding extracellular medium (∼ 23 μmol l -1 ).The y-intercept approximates the mean concentration obtained in due course from stability experiment (11 μmol l -1 , Fig. 2).Single results of three animals are displayed as dots.Caff eine does not modulate nutritive blood fl ow to rat gastric submucosa -a microdialysis study which involves injured gastric glands (whole-thickness mucosal necrosis with swelling and possible disconnection of mucosal layers) 21 .

Statistics
Data are expressed as means ± standard error of mean (SEM), unless otherwise noted.For statistical evaluation, normality tests and repeated measures ANOVA were used.The data were processed by the program NCSS 2004.The chosen level of signifi cance was α = 0.05.

Histology
Microdialysis probes were positioned correctly within the submucosal layer of gastric wall without penetration into the organ's lumen.The histological picture was comparable to previous reports 15,22 as indicated in Fig. 1a,  b.Lack of macroscopically measurable whole organ mucosal alterations due to caff eine treatment was mirrored in standard microscopy, which depicted similar grades of tissue damage ranging within grades 0-II around the probe in slides from all experimental groups including sham-operated animals (Fig. 1c-e).The severity of mucosal injury was unrelated to the administered caff eine dose.No grade III lesions were observed.

Microdialysis data
The results of probe performance stability are displayed in Fig. 2. In this experiment, the calculated average outfl ow concentration of total NO (within 5 h-time frame corresponding with the probe calibration study) was 11 μmol l -1 .This is in agreement with the expected value (∼ 9 μmol l -1 ) calculated using extrapolation of data obtained from the probe calibration study, i.e. in case the concentration of total NO in the perfusate was close to zero (Fig. 3).The outcomes of these experiments were consistent with one another indicating stable function of the microdialysis probe and NO production over time.The in vivo recovery of the used probe type under given experimental conditions was determined by the gradient of the regression line and found to be 38 %.The x-intercept corresponds to the concentration of nitrite + nitrate in the surrounding extracellular medium (∼ 23 μmol l -1 , Fig. 3).This level of extracellular NO was not aff ected by the i.p. administration of caff eine (p = 0.9, ANOVA, Fig. 4a).The microcirculation as represented by the ethanol out/in ratio showed a tendency for dose-related alterations after caff eine but the diff erence was not statistically signifi cant (p = 0.1, ANOVA, Fig. 4b).

Plasma analytes
Along with caff eine, plasma MDA was measured as general oxidative stress-related marker at the conclusion of the experiments.Even here, no statistically significant dose-related trend was observed (p = 0.8, ANOVA, Tab. 1).

Eff ect of caff eine on (sub)mucosal morphology
Even though macroscopical observations (e.g. a computerized planimetry) are a "golden standard" in the evaluation of gross gastric mucosal injury for their ability to assess the whole organ, they are inconvenient for detection and scoring of minor (visibly hardly discernible) changes.Since, in accordance with literature 23 , the latter was also the case in our experiment, gastric mucosal impairment was studied microscopically using standard histology.However, owing to the exactingness of the suggested method of whole organ histological evaluation 21 , we decided for the present time to grade the most prominent cellular/tissue changes found in comparable parts of glandular stomachs (in the vicinity of probe implantation site) of experimental groups.Since this time we did not quantify the lesions, the statistical approach was not feasible.Hence, the present study provides only qualitative data on the eff ects of microdialysis and caff eine on rat gastric tissue morphology.

NO measurement using microdialysis
There is very limited data on gastric NO measurement with in vivo microdialysis.Iversen et al. measured nerve-induced release of NO in the wall of rabbit stomach.They utilized diff erent probe types (CMA/10, CMA Medical AB, Stockholm, Sweden, 3 x 0.7 mm dialysis membrane with 20 kDa cut-off ) and perfusion rate (1 μl min -1 ) and estimated interstitial concentrations of nitrite and nitrate to approximate 10 and 70 μmol l -1 , respectively.The calculated in vitro recovery of the probes (for given perfusion rate) equaled 31 -33 % for nitrate and nitrite, respectively.The authors did not determine the recovery in vivo, but assumed that it would be in the range 10-40 % 24 .Suzuki et al. recently employed similar probes to our catheters (MAB 7.8.10 with 15 kDa cut-off dialysis membrane, active length 10 mm; outer diameter 0.5 mm; Microbiotech/se AB, Stockholm, Sweden) and validated them for studying nitrosative chemistry in the lumen of human stomach.The assessed in vitro recovery for nitrite at pH 1.5 and perfusion rate 0.15 ml h -1 was 71 % 25 .Our results of in vivo recovery may be comparable to these studies since the dialysis function of membranes in vivo generally diminishes 16 .The interstitial concentration of total NO found in our study (∼ 23 μmol l -1 ) was lower than reported by Iversen et al. in rabbits.It is a matter of debate to what extent this diff erence may be attributable to species and/or technique used.

Eff ect of caff eine on NO production
Besides vascular tone regulation, nitric oxide released by Ca 2+ -dependent endothelial (e)NOS as well as Ca 2+independent inducible (i)NOS plays a large number of (patho)physiological roles many of which may be related to mucosal defence and injury 13 .It is therefore of interest whether caff eine interferes with the release of this reactive pluripotent radical.Bruce et al. report of a signifi cant decrease in exhaled NO levels in humans one hour after 100-200 mg caff eine intake 26 .The latest fi ndings indicate that caff eine (16 mg kg -1 i.v.) may decrease nitric oxide synthase (NOS) expression in rat skeletal muscles 27 and attenuate glutamate-induced NO synthesis in murine spinal cord in vitro 28 .Moreover, caff eine negates the protective eff ect of ischemic preconditioning, i.e. eliminates reactive hyperemia due to the hypoperfusion-induced accumulation of adenosine and enhanced NO production 29 .In contrast, endothelium of isolated rat aorta responds to caff eine by promotion of NO synthesis 30 .In the present study, we failed to detect signifi cant shifts in NO production up to 2.5 h after the application of increasing doses of caff eine.However, eNOS releases NO in nanomolar quantities, changes we may have been unable to detect with the analytical procedure we used while micromolar amounts of NO are produced by activated iNOS as soon as 2 h after application of inducing substance 31 .Therefore, our results indicate no eff ect of caff eine on Ca 2+ -independent NO production in resting gastric submucosa of anesthetized rats.

Eff ect of caff eine on blood fl ow
Despite the generally recognized constricting role of caff eine in the brain 32 , heart 33 , limb 34 or gut 35 vasculature, the literature is inconsistent as far as gastric (sub)mucosal perfusion is concerned.Understandably, the discrepancy with respect to blood fl ow is refl ected in the contrary eff ects of caff eine on acute induced gastric mucosal injury 36,23,12 .Ozturkcan et al. showed that a single i.p. injection of 7.5 -30 mg kg -1 caff eine leads to elevations in rat gastric mucosal blood fl ow 90 min after drug application 10 .Moreover, Koyama et al. found that ex vivo intraluminal administration of caff eine doses as high as 50-100 mg kg -1 result in a dose-dependent increase in mucosal blood fl ow lasting up to 90 min 12 .Although in the present study some tendencies to decreased nutritive blood fl ow could be observed 30-90 min after caff eine administration, the results show a lack of statistically signifi cant change in this variable throughout the experiment (Fig. 4b).This would be consistent with other studies showing no direct eff ect of caff eine on resting blood fl ow [37][38][39] and/or induced vascular contractility 40 .Along these lines, there is confl ict-  37 .Interestingly, the latter authors ascribe these fi ndings to enhanced endothelial NO synthesis caused by released Ca 2+ from endoplasmic reticulum through activation of ryanodine-sensitive Ca 2+ channels and the suppression of cGMP degradation 30 , while the former investigators account for their observation as the inhibitory eff ect of caffeine on soluble GC with eventual suppression of cGMP formation.Unfortunately, neither group measured real in situ NO production.Nevertheless, these explanatory discrepancies may have some rationale as the diff erential pharmacologic actions of caff eine depend on the plasma caff eine levels 42 .Hence, the fi nal vasoactive action of caffeine may result from a balance between its vasoconstrictive (adenosine-antagonizing) and possibly vasodilating (NO-releasing) ability.

Eff ect of caff eine on oxidative stress
There are data suggesting multifarious mechanisms whereby caff eine might play a role in augmenting oxidative stress 41 .However, the acute unfavorable consequence of caff eine could also follow from its antagonism with adenosine -a substance producing NO and thus preventing mitochondrial oxidant damage in rat cardiomyocytes 43 .In rats that underwent concussive head injury, Al Moutaery et al. demonstrated an increase in the level of infl ammation and oxidative stress (signifi cant increase in neutrophil infi ltration and brain MDA) associated with a dose-dependent increase in mortality of caff eine-pretreated animals 44 .In the present experiment, the plasma concentrations of caff eine were consonant with previous fi ndings 45 and indicate good bioavailability of caff eine after i.p. application.However, 2.5 h after drug administration, only a slight dose-dependent decreasing trend of plasma MDA levels lacking statistical signifi cance was noted.This indicates no acute detrimental eff ect of caff eine on whole body's oxidative stress as far as systemically manifested eff ects are concerned (Tab.1).

CONCLUSIONS
The results of the present study indicate that i.p. administration of caff eine in given dosages does not produce acute macroscopic changes to gastric mucosa and is unlikely to alter gastric submucosal nutritive blood fl ow and nitric oxide production or aggravate systemic oxidative stress level.Additional histomorphometric studies are needed to describe the eff ect of caff eine on gastric mucosa and further attempts need to be made to explore possible mechanisms whereby caff eine might participate in irritant eff ects of coff ee.ABBREVIATIONS cGMP -cyclic guanosine monophosphate GC -guanylate cyclase MDA -malondialdehyde NO -nitric oxide e/iNOS -endothelial/inducible nitric oxide synthase

Fig. 1 .
Fig. 1.Morphological studies.Panel "a" confi rms the position of the microdialysis probe within the submucosal layer of gastric wall (arrow).The catheter is surrounded by noticeable edema and hyperemia with diff use infl ammatory infi ltrate of mild degree penetrating at some places into the mucosa.The probe's tougher inner membrane was slightly damaged during tissue slicing (mu = mucosa, sm = submucosa, mp = muscularis propria).Panel "b" depicts the detail of the probe's outer membrane permeated by polymorphonuclear leukocytes (arrow).Pictures "c-e" show normal gastric mucosa, i. e. grade 0, and typical lesions of grade I and II -detached cells and necrosis of pits (arrows), respectively.Standard hematoxylin-eosine stain, original magnifi cation "a" 125×, "b" 500×, "c-e" 250×.

Fig. 4 .
Fig. 4. Panel "a" presents the total NO production in response to increasing caff eine load.No diff erence was registered within 2.5 h of monitoring.Panel "b" illustrates the time course of microdialysate ethanol out/in ratio (negatively correlates with nutritive blood fl ow) indicating a slight tendency of microcirculation to deteriorate in a dose-dependent manner (particularly at intervals 3 and 5) following the injection of caff eine.However, this trend lacks statistical signifi cance.Results are means ± SEM of 6 measurements.

Table 1 .
41e table summarizes the eff ect of increasing doses of caff eine on plasma caff eine and malondialdehyde (MDA) levels at the conclusion of experiments.In spite of a negative trend, no statistically signifi cant diff erence or association between administered caff eine and plasma MDA was found.Data are displayed as means ± SEM of 6 measurements.Caff eine does not modulate nutritive blood fl ow to rat gastric submucosa -a microdialysis study ing data on the eff ect of caff eine on endothelial function.Papamichael et al. describe acute detrimental eff ects of caff eine up to 90 min after oral intake of 80 mg of caffeine on endothelium-dependent fl ow-mediated dilatation of the brachial artery41whereas Umemura et al. report its favorable role on (acetylcholine-induced) forearm vasodilation one hour after oral administration of 300 mg of caffeine, whereas baseline blood fl ow remained unaltered