- Email: [email protected]
Video-capsule endoscopy gastric and small bowel transit time and completeness of the examination in patients with diabetes mellitus
Digestive and Liver Disease 39 (2007) 575–580
Digestive Endoscopy
Video-capsule endoscopy gastric and small bowel transit time and completeness of the examination in patients with diabetes mellitus K. Triantafyllou a,∗ , C. Kalantzis b , A.A. Papadopoulos a , P. Apostolopoulos b , T. Rokkas c , N. Kalantzis b , S.D. Ladas a a
Hepato-Gastroenterology Unit, 2nd Department of Internal Medicine, Medical School, Athens University, “Attikon” University General Hospital, 1 Rimini Street, 124 62 Haidari, Athens, Greece b Gastroenterology Department, NIMTS Hospital, 10, Monis Petraki Street, 11521 Athens, Greece c Gastroenterology Clinic, “Henry Dunant” Hospital, 107 Messogeion Avenue, 11526 Athens, Greece Received 20 October 2006; accepted 30 January 2007 Available online 12 April 2007
Abstract Background. Small intestine video-capsule endoscopy examination in patients with diabetes mellitus may be incomplete due to delayed gastric emptying. Aim. To measure video-capsule endoscopy gastric and small bowel transit time and to assess the completeness of the examination in diabetes mellitus patients. Methods. In this retrospective, case-control study, we examined capsule endoscopy videos from 29 consecutive diabetes mellitus patients. Fifty-eight matched for sex, type of preparation, age and reason for referral non-diabetic controls were selected from our video-capsule endoscopy database. Two independent experienced investigators measured transit times and assessed examinations’ completeness. Results. Video-capsule endoscopy gastric transit time was significantly longer in diabetes mellitus (87, 1–478 min) compared to nondiabetic patients (24, 4–108 min, p < 0.001). The caecum was visualized in 20/29 (69%) diabetes mellitus and 52/58 (89.6%) non-diabetic controls (p = 0.02). In 16 diabetes mellitus patients that video-capsule endoscopy reached the caecum, small bowel transit time was significantly shorter (261.2 ± 55.5 min) compared to their 32 non-diabetic matched controls (302 ± 62.7 min, p = 0.03). Conclusions. Patients with type 2 diabetes mellitus have prolonged video-capsule endoscopy gastric transit time compared to non-diabetic patients. Prospective studies are required to complete our understanding of video-capsule endoscopy transit times in the setting of diabetes mellitus. © 2007 Editrice Gastroenterologica Italiana S.r.l. Published by Elsevier Ltd. All rights reserved. Keywords: Diabetes mellitus; Gastric transit time; Small bowel transit time; Video-capsule endoscopy
1. Introduction Video-capsule endoscopy (VCE) is a widespread noninvasive modality to diagnose small intestinal diseases [1,2]. However, incomplete examination of the small intestine by capsule endoscopy occurs in about 15–20% of the patients [1–4]. A crucial factor for incomplete examinations is the long retention of the capsule in the stomach [5] and, there∗
Corresponding author. Tel.: +30 210 5832552; fax: +30 210 5326422. E-mail address: [email protected] (K. Triantafyllou).
fore, inadequate time for the capsule to visualize the small intestine, because the life span of capsule battery is 8 h. It is known that patients with diabetes mellitus (DM) have delayed gastric emptying and this may result in a higher percentage of incomplete VCE examinations. Studies using radionuclide techniques have shown that gastric emptying of both solid and liquid meals is abnormally slow in 30–50% patients with longstanding type 1 [6,7] or type 2 DM [6,8]. Therefore, long retention of the VCE in the stomach of DM patients undergoing investigation of the small bowel may limit the diagnostic yield of the examination.
1590-8658/$30 © 2007 Editrice Gastroenterologica Italiana S.r.l. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.dld.2007.01.024
576
K. Triantafyllou et al. / Digestive and Liver Disease 39 (2007) 575–580
The aim of our study was to measure both gastric transit time (GTT) and small bowel transit time (SBTT) of the videocapsule and to assess the completeness of examination in DM patients.
nal mucosa; 3 = poor: large volume of residual fluids or food residues, which disturbed visualization and interpretation [9]. 2.2. Statistical analysis
2. Methods This retrospective, case-controlled study included the videos of 29 consecutive DM outpatients who underwent capsule endoscopy examination of the small bowel in NIMTS, Attikon and Henry Dunant Hospitals in Athens, Greece, between June 2002 and April 2006. All cases had type 2 DM for 8 (5–12) years. Two of the patients used insulin, while the rest of them were treated with antidiabetic oral medications. No type 1 DM patient was detected in our database. Each diabetic patient was matched with two non-diabetic (ND) outpatient controls. Therefore, 58 matched controls in order for sex, type of preparation, age and reason for referral were selected from the VCE examinations of our institutions, performed at the same period. Matching has been performed by a blinded to the aim of the study person. Before the VCE study, each patient received bowel preparation according to the preference of the participating department: either only clear liquids diet (CL) the day before the examination (Henry Dunant Hospital) or CL plus a purge with 2 L of polyethelene glycol (PEG) solution (NIMTS Hospital) or CL plus a purge with 45 mL of sodium phosphate (PS) (Attikon University General Hospital). Neither case nor control had dyspeptic symptoms or known gastroparesis and none of them had suspected obstruction or any other reason for prolonged capsule transit times. All patients had at least once a negative upper and lower gastrointestinal endoscopy before referred for VCE examination.
In our series, mean VCE GTT in ND patients is 30 min without normal distribution [10]. We estimated that group sample sizes of 25 diabetics and 50 controls achieve 80% power to detect a difference of 60 min between the null hypothesis (both group VCE GTT means are 30 min) and the alternative hypothesis (the mean of VCE GTT of DM patients is 90 min) with unknown and unequal group standard deviations and with a significance level (alpha) of 0.05, using a two-sided two-sample t-test. Results in the text and in the tables are presented as absolute and value percent for qualitative data. Qualitative data were assessed by non-parametric tests, as appropriate. Agreement between categorical measurements was assessed by Kappa statistics. Distribution of quantitative data was assessed by Kolmogorov-Smirnov test. Normally distributed quantitative data are presented as mean value (± standard deviation) and were assessed by Student’s two-sided t-test. Non-normally distributed quantitative data are presented as median value with ranges and were assessed by nonparametric test. Correlations were assessed by regression analysis. A p-value of <0.05 indicated statistical significance. Comparison data in figures are presented as box-andwhisker plots. The box includes 50% of the results falling between 25th and 75th percentile, the median value is represented as a horizontal line inside the box and outliers are shown as open circles and extreme values as asterisks.
3. Results 2.1. Study design Eligible for evaluation studies were coded and submitted to two independent experienced viewers, for video review. Both had an experience of at least 50 capsule endoscopy studies. All videos reviewed with the PillCam SB® capsule endoscopy system (Given Imaging Ltd., Israel). Images viewed by the Rapid Reader (version 3.1) using dual view mode at maximal speed (40 frames/s). Reviewers were blinded to the patient’s status (case vs. control), preparation, indication and the diagnosis of each VCE study. Each viewer assessed VCE GTT, VCE SBTT and the overall quality of preparation independently. VCE GTT was defined as the time taken from the first gastric image to the first duodenal image. VCE SBTT was defined as the time taken from the first duodenal image to the first caecal image for cases in which the capsule reached the caecum. The overall quality of preparation for each study case was determined using a three-step scale: 1 = good: excellent visibility of the mucosa; 2 = moderate: small to moderate volume of residual fluid which did not impair visualization of the small intesti-
Demographic characteristics, indication for VCE study and type of bowel preparation in DM patients and ND controls are shown in Table 1. There was perfect matching regarding sex and bowel preparation, while there were no significant differences between cases and controls concerning mean age (p = 0.79) and indications for VCE imaging (X2 = 5, d.f. = 4, p = 0.28). The commonest indication for both groups was iron deficiency anaemia (41.4% for DM and 46.5% for ND patients, respectively). There was no difference in the two groups regarding the diagnostic yield of the VCE (X2 = 6.8, d.f. = 5, p = 0.24) (Table 2). 3.1. Quality of bowel preparation and interobserver agreement Interobserver agreement of the quality of bowel preparation was excellent (κ = 0.84, p < 0.001) (Table 2) and agreement of caecum visualization was 100%. There was also excellent correlation between the two investigators for both VCE GTT (r = 1, p < 0.001) and SBTT (r = 0.98,
K. Triantafyllou et al. / Digestive and Liver Disease 39 (2007) 575–580
577
Table 1 Patients’ demographic characteristics, indications for video-capsule endoscopy study and type of preparation DM (n = 29)
ND (n = 58)
Sex, n (%) Male Female
12 (41.4) 17 (58.6)
24 (41.4) 34 (58.6)
Age, years (mean ± S.D.)
62.6 ± 12.0
61.9 ± 11.6
Type of bowel preparation, n (%) CL PS PEG
5 (17.2) 6 (20.7) 18 (62.1)
10 (17.2) 12 (20.7) 36 (62.1)
Indication for capsule endoscopy study, n (%) Iron deficiency anaemia Recurrent melena of unknown aetiology Chronic diarrhoea of obscure aetiology Recurrent abdominal pain Miscellaneous disease
12 (41.4) 5 (17.25) 6 (20.7) 3 (10.3) 3 (10.3)
27 (46.5) 7 (12.1) 12 (20.7) 7 (12.1) 5 (8.6)
p-value 1
0.79 1
0.96
DM: diabetes mellitus patients; ND: non-diabetic controls; CL: clear liquids for 12 h, followed by an 8 h overnight fast, prior to testing; PS: CL + 45 mL of sodium phosphate (Fleet Phosphosoda® , Botania, Greece) with water in the evening before the procedure, followed by an 8 h overnight fast, prior to testing; PEG: CL + 2 L polyethylene glycol solution (Clean Prep PEG+E® , Kite Hellas, Greece) in the evening before the examination, followed by an 8 h overnight fast, prior to testing.
p < 0.001) estimations. Therefore, we analysed the mean of the two values given by the investigators for each transit time. The overall quality of small intestinal mucosal visibility was not significantly different among cases and controls (X2 = 4.1, d.f. = 2, p = 0.13) (Table 2). Two of the DM patients but none of the controls had a poor quality of small bowel preparation. Both cases had very long GTT (420 min and 478 min, respectively); the video-capsule did not reach the caecum and no firm diagnosis was made in either of them.
3.2. VCE gastric and small bowel transit times and rate of complete examination GTT of the VCE was significantly longer in DM patients (87, 1–478 min) as compared to controls (24, 4–108 min, Z = −4.97, p < 0.001) (Table 2, Fig. 1). There was no correlation between duration of DM and VCE GTT (F = 0.87, p = 0.36). The rate of caecum visualization (complete examination) was significantly lower in DM patients (20/29, 69%) than
Table 2 Overall quality of preparation, transit times, caecum visualization and diagnosis in cases and controls DM (n = 29)
ND (n = 58)
Quality of preparation, n (%) First viewer Good Moderate Poor
16 (55.2) 11 (37.9) 2 (6.9)
34 (58.6) 24 (41.4) 0
Quality of preparation, n (%) Second viewer Good Moderate Poor
15 (51.8) 12 (41.4) 2 (6.9)
32 (55.2) 26 (44.8) 0
Finding, n (%) Normal Angiodysplasia Erosions Celiac disease Miscellaneous No diagnosisa
14 (48.3) 5 (17.2) 4 (13.9) 0 3 (10.3) 3 (10.3)
15 (25.9) 12 (20.7) 17 (29.3) 3 (5.2) 7 (12.1) 4 (6.8)
VCE GTT, min (median, range) Caecum reached, n (%) VCE SBTT, min (mean ± S.D.)b
87, 1–478 20 (69) 261.2 ± 55.5
24, 4–108 52 (89.6) 302.0 ± 62.7
p-value 0.13
0.13
0.24
<0.001 0.016 0.032
DM: diabetes mellitus patients; ND: non-diabetic controls; VCE: video-capsule endoscopy; GTT: gastric transit time; SBTT: small bowel transit time. a Due to active bleeding or inadequate bowel preparation. b Only for cases (n = 16) and matched controls (n = 32) that the capsule reached the caecum.
578
K. Triantafyllou et al. / Digestive and Liver Disease 39 (2007) 575–580
Fig. 1. Video-capsule endoscopy gastric transit time was significantly longer in DM patients as compared to controls (p < 0.001). DM: diabetes mellitus patients, ND: non-diabetic controls. Data are presented as box-and-whisker plot. The box includes 50% of the results falling between 25th and 75th percentile, the median value is represented as a horizontal line inside the box, outliers are shown as open circles and extreme values as asterisks.
Fig. 3. Video-capsule endoscopy small bowel transit time was significantly shorter in DM patients as compared to controls (p = 0.032). DM: diabetes mellitus patients; ND: non-diabetic controls. Data are presented as box-andwhisker plot. The box includes 50% of the results falling between 25th and 75th percentile, the median value is represented as a horizontal line inside the box, outliers are shown as open circles and extreme values as asterisks.
in ND controls (52/58, 89.6%; X2 = 5.8, d.f. = 1, p = 0.016) (Table 2). No relation between duration of DM and visualization of the ceacum (Z = −0.95, p = 0.34) has been detected. In DM patients, the completion of VCE study was significantly associated with shorter VCE GTT (caecum reached: 68.25, 1–221 min; caecum not reached: 225, 70–478 min; Z = −3.06, p = 0.002) and the rate of complete VCE studies decreased progressively with increasing VCE GTT. However, in controls no relation was detected between the length of VCE GTT and the completion of the examination (caecum reached: 24, 4–108 min; caecum not reached: 31.75, 23–55 min; Z = −1.53, p = 0.12) (Fig. 2).
VCE SBTT was assessed in 16 cases that the capsule reached the caecum and their 32 matched controls, in order to keep the matching. VCE SBTT among DM patients was significantly shorter (261.2 ± 55.5 min) compared to ND controls (302.0 ± 62.7 min, p = 0.032) (Table 2, Fig. 3).
Fig. 2. Rate of complete examinations in relation to video-capsule gastric transit time in diabetic patients and controls. The number of cases that capsule reached the caecum/all cases at the respective video-capsule gastric transit time range is presented within the bars of the histogram. VCE: video-capsule endoscopy; GTT: gastric transit time; DM: diabetes mellitus patients; ND: non-diabetic controls.
4. Discussion Wireless VCE is a new method, superior to any known to date radiological and endoscopy investigations of the small bowel [11], to visualize the small intestinal mucosa and diagnose mucosal, vascular or tumour-like diseases. However, there are two main causes of incomplete examination, notably delayed gastric emptying [12], or even retention [5,13], of the capsule in the stomach and inadequate visualization of the small intestinal mucosa due to food residue of green viscous mucus in the terminal ileum [14]. DM patients frequently have delayed gastric emptying [6–8] and even gastroparesis [15]. The main pathogenetic factors in DM delayed gastric emptying are vagal autonomic neuropathy, interstitial cells of Cajal pathology and hyperglycaemia [15]. Until today, there is no study designed to evaluate the impact of DM on the completion rate of small bowel VCE. Selby [12], using logistic regression analysis, showed that several factors including DM had no effect on the completion of VCE small bowel visualization. However, the main aim of Selby’s study was the effect of metoclopramide and not DM on the completeness of VCE studies. On the contrary, our data clearly show that VCE GTT in patients with longstanding type 2 DM is at about 60 min longer compared to that
K. Triantafyllou et al. / Digestive and Liver Disease 39 (2007) 575–580
of ND controls. This results to significantly fewer complete small bowel examinations, since the caecum was visualized in less than 70% of the diabetics compared to 90% of the controls. Furthermore, in DM patients, we have also shown that the longer the VCE GTT is, the fewer will be the complete examinations. When VCE GTT was longer than 90 min, 50% of the capsules reached the caecum, and this percentage decreased progressively with increasing VCE GTT. No capsule reached the caecum when VCE GTT was longer than 240 min. Our findings have clinical and health-economical implications on the use of VCE in diabetics. Capsule endoscopy is an expensive (each capsule costs about 600D in Europe) and time-consuming examination. Since completion rate of the examination is reduced by 20% in DM, every effort should be made to increase the percentage of complete VCE studies. DM patients may benefit from early identification of prolonged retention of the capsule in the stomach by real-time VCE images and endoscopic placement of the capsule in the duodenum. Moreover, DM patients may actually benefit from systematic use of medications aiming to accelerate gastric emptying before and/or during the examination, although the use of prokinetics is controversial in VCE [3,12,16] and the efficacy of currently available prokinetics is not satisfactory in diabetic gastroparesis [17]. Both these approaches should be tested prospectively. A limitation of our study is that it is not prospective. However, matching was excellent, our sample was homogenous (type 2 DM patients) and we carefully excluded known causes of delayed transit times other than DM, such as gastric surgery and Crohn’s disease with small intestinal strictures. In addition, two independent investigators with excellent interobserver agreement evaluated all capsule videos blindly [18]. Another possible limitation is the mixture of bowel preparations used in the study. We have shown that 2 L of PEG solution or 45 mL of PS the evening before the study do not significantly affect VCE gastric or SBTTs in NDs, as compared with liquid diet alone [10]. Despite that DM patients were matched with controls, one cannot exclude the possibility that the effects of the three preparations could be different in DM as compared to controls. A prospective randomized study could resolve this uncertainty. Due to the retrospective design of our study, we have no data on the presence and extend of autonomic neuropathy and on glycaemic control. The lack of these data is not expected to influence our result interpretation. There are reports that gastric emptying correlates with the presence of autonomic neuropathy [19–21], but others have shown that this association is very weak [22,23] and although gastric emptying varies inversely with increasing glycaemia [8] in type 2 diabetics, a modest improvement in glycaemic control has no significant effect on gastric emptying in these patients [24]. Moreover, changes in blood glucose level within the physiological post-prandial range also affect gastric emptying [25], thus making the control for glycaemia very difficult during VCE examinations.
579
Our study has not been powered to detect differences regarding VCE SBTT between DM patients and ND controls. However, our finding that in the cases the capsule reached the caecum, DM patient’s VCE SBTT was faster than that of controls cannot be adequately explained and warrants further investigation. This could be explained by the selection of patients in whom VCE SBTT was complete. Since VCE GTT was longer in these patients, then VCE SBTT must have been faster to allow the capsule to visualize the ceacum. In conclusion, this is the first published report that in DM patients video-capsule GTT is significantly prolonged resulting in fewer complete small bowel investigations. Modalities, such as new powerful prokinetic medications or real-time VCE with endoscopic placement of the capsule in the duodenum, in case the capsule remains in the stomach, should be studied prospectively to examine whether they could enhance video-capsule small bowel endoscopy completeness rate in DM patients.
Practice points • During small bowel VCE studies in type 2 DM patients, the capsule remains in their stomach significantly longer than in controls. • The prolongation of VCE GTT in patients with type 2 DM results in fewer complete small bowel investigations.
Research agenda • In patients with type 2 DM, the effect of bowel preparation on the completeness rate of small bowel VCE study is unknown. • New powerful prokinetic medications should be studied prospectively to examine whether they could enhance video-capsule small bowel endoscopy completeness rate in type 2 DM patients. • Real-time VCE with endoscopic placement of the capsule in the duodenum should also be studied prospectively, as an alternative, in case the capsule remains in the stomach for long in type 2 DM patients.
Conflict of interest statement None declared.
580
K. Triantafyllou et al. / Digestive and Liver Disease 39 (2007) 575–580
References [1] Swain P, Fritscher-Ravens A. Role of video endoscopy in managing small bowel disease. Gut 2004;53:1866–75. [2] Rey J-F, Ladas S, Alhassani A, Kuznetsov K, ESGE Guidelines Committee. European Society of Gastrointestinal Endoscopy (ESGE). Video capsule endoscopy: update to guidelines (May 2006). Endoscopy 2006;38:1047–53. [3] Caddy G, Moran L, Chong A, Miller A, Taylor A, Desmond P. The effect of erythromycin on video capsule endoscopy intestinal transit time. Gastrointest Endosc 2006;63:262–6. [4] Sturniolo GC, Di Leo V, Vettorato MG, De Boni M, Lamboglia F, De Bona M, et al. Small bowel exploration by wireless capsule endoscopy: results from 314 procedures. Am J Med 2006;119:341–7. [5] Ben-Soussan E, Savoye G, Antonietti M, Ramirez S, Lerebours E, Ducrotte P. Factors that affect gastric passage of video capsule. Gastrointest Endosc 2005;62:785–90. [6] Rabine JC, Barnett JL. Management of patients with gastroparesis. J Clin Gastroenterol 2001;32:11–21. [7] Jones KL, Russo A, Stevens JE, Wishart JM, Berry MK, Horowitz M. Predictors of delayed gastric emptying in diabetes. Diab Care 2001;24:1264–9. [8] Horowitz M, Harding PE, Maddox AF, Wishart JM, Akkermans LM, Chatterton BE, et al. Gastric and oesophageal emptying in patients with type 2 (non-insulin dependent) diabetes mellitus. Diabetologia 1989;32:151–9. [9] Niv Y, Niv G, Wiser K, Demarco DC. Capsule endoscopy – comparison of two strategies of bowel preparation. Aliment Pharmacol Ther 2005;22:957–62. [10] Kalantzis C, Triantafyllou K, Papadopoulos A, Alexandrakis G, Rokkas T, Kalantzis N, et al. Wireless-capsule endoscopy gastric and small bowel transit time. Effect of three different bowel preparations. Scand J Gastroenterol; in press. [11] Neu B, Ell C, May A, Schmid E, Riemann JF, Hagenmuller F, et al. Capsule endoscopy versus standard tests in influencing management of obscure digestive bleeding: results from a German multicenter trial. Am J Gastroenterol 2005;100:1736–42. [12] Selby W. Complete small-bowel transit in patients undergoing capsule endoscopy: determining factors and improvement with metoclopramide. Gastrointest Endosc 2005;61:80–5. [13] Tatar E, Shen E, Palance A, Sun J, Pitchumoni CS. Clinical utility of wireless capsule endoscopy experience with 200 cases. J Clin Gastroenterol 2006;40:140–4.
[14] Costamagna G, Shah SK, Riccioni ME, Foschia F, Mutignani M, Perri V, et al. A prospective trial comparing small bowel radiographs and video capsule endoscopy for suspected small bowel disease. Gastroenterology 2002;123:999–1005. [15] Park M-I, Camillieri M. Gastroparesis: clinical update. Am J Gastroenterol 2006;101:1129–39. [16] de Franchis R, Avgerinos A, Barkin J, Cave D, Filoche B. ICCE consensus for bowel preparation and prokinetics. Endoscopy 2005;37: 1040–5. [17] Talley NJ. Diabetic gastropathy and prokinetics. Am J Gastroenterol 2003;98:264–71. [18] Lai L, Wong G, Chow D, Lau J, Sung J, Leung W. Inter-observer variations on interpretation of capsule endoscopies. Eur J Gastroenterol Hepatol 2006;18:283–6. [19] Kockar MC, Kayahan IK, Bavbek N. Diabetic gastroparesis in association with autonomic neuropathy and microvasculopathy. Acta Med Okayama 2002;56:237–43. [20] Murray CDR, Martin NM, Patterson M, Taylor SA, Ghatei MA, Kamm MA, et al. Ghrelin enhances gastric emptying in diabetic gastroparesis: a double blind, placebo controlled, crossover study. Gut 2005;54:1693–8. [21] Asakawa H, Onishi M, Hayashi I, Fucuda A, Tokunaga K. Comparison between coefficient of R-R interval variation and gastric emptying in type 2 diabetes mellitus patients. J Gastroenterol Hepatol 2005;20:1358–64. [22] Horowitz M, Maddox AF, Wishart JM, Harding PE, Chatterton BE, Shearman DJC. Relationship between oesophageal transit and solid and liquid gastric emptying in diabetes mellitus. Eur J Nucl Med 1991;18:229–34. [23] Merio R, Festa A, Bergmann H, Eder T, Eibl N, Stacher-Janotta G, et al. Slow gastric emptying in type 1 diabetes: relation to autonomic and peripheral neuropathy, blood glucose and glycemic control. Diab Care 1997;39:823–30. [24] Holzapfel A, Festa A, Stacher-Janotta G, Bergmann H, Shnawa N, Brannath W, et al. Gastric emptying in type II (non-insulin dependent) diabetes mellitus before and after therapy readjustment: no influence of actual glucose concentration. Diabetologia 1999;42: 1410–2. [25] Schvarcz E, Palmer M, Aman J, Horowitz M, Stridsberg M, Berne C. Physiological hyperglycemia slows gastric emptying in normal subjects and patients with insulin-dependent diabetes mellitus. Gastroenterology 1997;113:60–6.
Digestive Endoscopy
Video-capsule endoscopy gastric and small bowel transit time and completeness of the examination in patients with diabetes mellitus K. Triantafyllou a,∗ , C. Kalantzis b , A.A. Papadopoulos a , P. Apostolopoulos b , T. Rokkas c , N. Kalantzis b , S.D. Ladas a a
Hepato-Gastroenterology Unit, 2nd Department of Internal Medicine, Medical School, Athens University, “Attikon” University General Hospital, 1 Rimini Street, 124 62 Haidari, Athens, Greece b Gastroenterology Department, NIMTS Hospital, 10, Monis Petraki Street, 11521 Athens, Greece c Gastroenterology Clinic, “Henry Dunant” Hospital, 107 Messogeion Avenue, 11526 Athens, Greece Received 20 October 2006; accepted 30 January 2007 Available online 12 April 2007
Abstract Background. Small intestine video-capsule endoscopy examination in patients with diabetes mellitus may be incomplete due to delayed gastric emptying. Aim. To measure video-capsule endoscopy gastric and small bowel transit time and to assess the completeness of the examination in diabetes mellitus patients. Methods. In this retrospective, case-control study, we examined capsule endoscopy videos from 29 consecutive diabetes mellitus patients. Fifty-eight matched for sex, type of preparation, age and reason for referral non-diabetic controls were selected from our video-capsule endoscopy database. Two independent experienced investigators measured transit times and assessed examinations’ completeness. Results. Video-capsule endoscopy gastric transit time was significantly longer in diabetes mellitus (87, 1–478 min) compared to nondiabetic patients (24, 4–108 min, p < 0.001). The caecum was visualized in 20/29 (69%) diabetes mellitus and 52/58 (89.6%) non-diabetic controls (p = 0.02). In 16 diabetes mellitus patients that video-capsule endoscopy reached the caecum, small bowel transit time was significantly shorter (261.2 ± 55.5 min) compared to their 32 non-diabetic matched controls (302 ± 62.7 min, p = 0.03). Conclusions. Patients with type 2 diabetes mellitus have prolonged video-capsule endoscopy gastric transit time compared to non-diabetic patients. Prospective studies are required to complete our understanding of video-capsule endoscopy transit times in the setting of diabetes mellitus. © 2007 Editrice Gastroenterologica Italiana S.r.l. Published by Elsevier Ltd. All rights reserved. Keywords: Diabetes mellitus; Gastric transit time; Small bowel transit time; Video-capsule endoscopy
1. Introduction Video-capsule endoscopy (VCE) is a widespread noninvasive modality to diagnose small intestinal diseases [1,2]. However, incomplete examination of the small intestine by capsule endoscopy occurs in about 15–20% of the patients [1–4]. A crucial factor for incomplete examinations is the long retention of the capsule in the stomach [5] and, there∗
Corresponding author. Tel.: +30 210 5832552; fax: +30 210 5326422. E-mail address: [email protected] (K. Triantafyllou).
fore, inadequate time for the capsule to visualize the small intestine, because the life span of capsule battery is 8 h. It is known that patients with diabetes mellitus (DM) have delayed gastric emptying and this may result in a higher percentage of incomplete VCE examinations. Studies using radionuclide techniques have shown that gastric emptying of both solid and liquid meals is abnormally slow in 30–50% patients with longstanding type 1 [6,7] or type 2 DM [6,8]. Therefore, long retention of the VCE in the stomach of DM patients undergoing investigation of the small bowel may limit the diagnostic yield of the examination.
1590-8658/$30 © 2007 Editrice Gastroenterologica Italiana S.r.l. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.dld.2007.01.024
576
K. Triantafyllou et al. / Digestive and Liver Disease 39 (2007) 575–580
The aim of our study was to measure both gastric transit time (GTT) and small bowel transit time (SBTT) of the videocapsule and to assess the completeness of examination in DM patients.
nal mucosa; 3 = poor: large volume of residual fluids or food residues, which disturbed visualization and interpretation [9]. 2.2. Statistical analysis
2. Methods This retrospective, case-controlled study included the videos of 29 consecutive DM outpatients who underwent capsule endoscopy examination of the small bowel in NIMTS, Attikon and Henry Dunant Hospitals in Athens, Greece, between June 2002 and April 2006. All cases had type 2 DM for 8 (5–12) years. Two of the patients used insulin, while the rest of them were treated with antidiabetic oral medications. No type 1 DM patient was detected in our database. Each diabetic patient was matched with two non-diabetic (ND) outpatient controls. Therefore, 58 matched controls in order for sex, type of preparation, age and reason for referral were selected from the VCE examinations of our institutions, performed at the same period. Matching has been performed by a blinded to the aim of the study person. Before the VCE study, each patient received bowel preparation according to the preference of the participating department: either only clear liquids diet (CL) the day before the examination (Henry Dunant Hospital) or CL plus a purge with 2 L of polyethelene glycol (PEG) solution (NIMTS Hospital) or CL plus a purge with 45 mL of sodium phosphate (PS) (Attikon University General Hospital). Neither case nor control had dyspeptic symptoms or known gastroparesis and none of them had suspected obstruction or any other reason for prolonged capsule transit times. All patients had at least once a negative upper and lower gastrointestinal endoscopy before referred for VCE examination.
In our series, mean VCE GTT in ND patients is 30 min without normal distribution [10]. We estimated that group sample sizes of 25 diabetics and 50 controls achieve 80% power to detect a difference of 60 min between the null hypothesis (both group VCE GTT means are 30 min) and the alternative hypothesis (the mean of VCE GTT of DM patients is 90 min) with unknown and unequal group standard deviations and with a significance level (alpha) of 0.05, using a two-sided two-sample t-test. Results in the text and in the tables are presented as absolute and value percent for qualitative data. Qualitative data were assessed by non-parametric tests, as appropriate. Agreement between categorical measurements was assessed by Kappa statistics. Distribution of quantitative data was assessed by Kolmogorov-Smirnov test. Normally distributed quantitative data are presented as mean value (± standard deviation) and were assessed by Student’s two-sided t-test. Non-normally distributed quantitative data are presented as median value with ranges and were assessed by nonparametric test. Correlations were assessed by regression analysis. A p-value of <0.05 indicated statistical significance. Comparison data in figures are presented as box-andwhisker plots. The box includes 50% of the results falling between 25th and 75th percentile, the median value is represented as a horizontal line inside the box and outliers are shown as open circles and extreme values as asterisks.
3. Results 2.1. Study design Eligible for evaluation studies were coded and submitted to two independent experienced viewers, for video review. Both had an experience of at least 50 capsule endoscopy studies. All videos reviewed with the PillCam SB® capsule endoscopy system (Given Imaging Ltd., Israel). Images viewed by the Rapid Reader (version 3.1) using dual view mode at maximal speed (40 frames/s). Reviewers were blinded to the patient’s status (case vs. control), preparation, indication and the diagnosis of each VCE study. Each viewer assessed VCE GTT, VCE SBTT and the overall quality of preparation independently. VCE GTT was defined as the time taken from the first gastric image to the first duodenal image. VCE SBTT was defined as the time taken from the first duodenal image to the first caecal image for cases in which the capsule reached the caecum. The overall quality of preparation for each study case was determined using a three-step scale: 1 = good: excellent visibility of the mucosa; 2 = moderate: small to moderate volume of residual fluid which did not impair visualization of the small intesti-
Demographic characteristics, indication for VCE study and type of bowel preparation in DM patients and ND controls are shown in Table 1. There was perfect matching regarding sex and bowel preparation, while there were no significant differences between cases and controls concerning mean age (p = 0.79) and indications for VCE imaging (X2 = 5, d.f. = 4, p = 0.28). The commonest indication for both groups was iron deficiency anaemia (41.4% for DM and 46.5% for ND patients, respectively). There was no difference in the two groups regarding the diagnostic yield of the VCE (X2 = 6.8, d.f. = 5, p = 0.24) (Table 2). 3.1. Quality of bowel preparation and interobserver agreement Interobserver agreement of the quality of bowel preparation was excellent (κ = 0.84, p < 0.001) (Table 2) and agreement of caecum visualization was 100%. There was also excellent correlation between the two investigators for both VCE GTT (r = 1, p < 0.001) and SBTT (r = 0.98,
K. Triantafyllou et al. / Digestive and Liver Disease 39 (2007) 575–580
577
Table 1 Patients’ demographic characteristics, indications for video-capsule endoscopy study and type of preparation DM (n = 29)
ND (n = 58)
Sex, n (%) Male Female
12 (41.4) 17 (58.6)
24 (41.4) 34 (58.6)
Age, years (mean ± S.D.)
62.6 ± 12.0
61.9 ± 11.6
Type of bowel preparation, n (%) CL PS PEG
5 (17.2) 6 (20.7) 18 (62.1)
10 (17.2) 12 (20.7) 36 (62.1)
Indication for capsule endoscopy study, n (%) Iron deficiency anaemia Recurrent melena of unknown aetiology Chronic diarrhoea of obscure aetiology Recurrent abdominal pain Miscellaneous disease
12 (41.4) 5 (17.25) 6 (20.7) 3 (10.3) 3 (10.3)
27 (46.5) 7 (12.1) 12 (20.7) 7 (12.1) 5 (8.6)
p-value 1
0.79 1
0.96
DM: diabetes mellitus patients; ND: non-diabetic controls; CL: clear liquids for 12 h, followed by an 8 h overnight fast, prior to testing; PS: CL + 45 mL of sodium phosphate (Fleet Phosphosoda® , Botania, Greece) with water in the evening before the procedure, followed by an 8 h overnight fast, prior to testing; PEG: CL + 2 L polyethylene glycol solution (Clean Prep PEG+E® , Kite Hellas, Greece) in the evening before the examination, followed by an 8 h overnight fast, prior to testing.
p < 0.001) estimations. Therefore, we analysed the mean of the two values given by the investigators for each transit time. The overall quality of small intestinal mucosal visibility was not significantly different among cases and controls (X2 = 4.1, d.f. = 2, p = 0.13) (Table 2). Two of the DM patients but none of the controls had a poor quality of small bowel preparation. Both cases had very long GTT (420 min and 478 min, respectively); the video-capsule did not reach the caecum and no firm diagnosis was made in either of them.
3.2. VCE gastric and small bowel transit times and rate of complete examination GTT of the VCE was significantly longer in DM patients (87, 1–478 min) as compared to controls (24, 4–108 min, Z = −4.97, p < 0.001) (Table 2, Fig. 1). There was no correlation between duration of DM and VCE GTT (F = 0.87, p = 0.36). The rate of caecum visualization (complete examination) was significantly lower in DM patients (20/29, 69%) than
Table 2 Overall quality of preparation, transit times, caecum visualization and diagnosis in cases and controls DM (n = 29)
ND (n = 58)
Quality of preparation, n (%) First viewer Good Moderate Poor
16 (55.2) 11 (37.9) 2 (6.9)
34 (58.6) 24 (41.4) 0
Quality of preparation, n (%) Second viewer Good Moderate Poor
15 (51.8) 12 (41.4) 2 (6.9)
32 (55.2) 26 (44.8) 0
Finding, n (%) Normal Angiodysplasia Erosions Celiac disease Miscellaneous No diagnosisa
14 (48.3) 5 (17.2) 4 (13.9) 0 3 (10.3) 3 (10.3)
15 (25.9) 12 (20.7) 17 (29.3) 3 (5.2) 7 (12.1) 4 (6.8)
VCE GTT, min (median, range) Caecum reached, n (%) VCE SBTT, min (mean ± S.D.)b
87, 1–478 20 (69) 261.2 ± 55.5
24, 4–108 52 (89.6) 302.0 ± 62.7
p-value 0.13
0.13
0.24
<0.001 0.016 0.032
DM: diabetes mellitus patients; ND: non-diabetic controls; VCE: video-capsule endoscopy; GTT: gastric transit time; SBTT: small bowel transit time. a Due to active bleeding or inadequate bowel preparation. b Only for cases (n = 16) and matched controls (n = 32) that the capsule reached the caecum.
578
K. Triantafyllou et al. / Digestive and Liver Disease 39 (2007) 575–580
Fig. 1. Video-capsule endoscopy gastric transit time was significantly longer in DM patients as compared to controls (p < 0.001). DM: diabetes mellitus patients, ND: non-diabetic controls. Data are presented as box-and-whisker plot. The box includes 50% of the results falling between 25th and 75th percentile, the median value is represented as a horizontal line inside the box, outliers are shown as open circles and extreme values as asterisks.
Fig. 3. Video-capsule endoscopy small bowel transit time was significantly shorter in DM patients as compared to controls (p = 0.032). DM: diabetes mellitus patients; ND: non-diabetic controls. Data are presented as box-andwhisker plot. The box includes 50% of the results falling between 25th and 75th percentile, the median value is represented as a horizontal line inside the box, outliers are shown as open circles and extreme values as asterisks.
in ND controls (52/58, 89.6%; X2 = 5.8, d.f. = 1, p = 0.016) (Table 2). No relation between duration of DM and visualization of the ceacum (Z = −0.95, p = 0.34) has been detected. In DM patients, the completion of VCE study was significantly associated with shorter VCE GTT (caecum reached: 68.25, 1–221 min; caecum not reached: 225, 70–478 min; Z = −3.06, p = 0.002) and the rate of complete VCE studies decreased progressively with increasing VCE GTT. However, in controls no relation was detected between the length of VCE GTT and the completion of the examination (caecum reached: 24, 4–108 min; caecum not reached: 31.75, 23–55 min; Z = −1.53, p = 0.12) (Fig. 2).
VCE SBTT was assessed in 16 cases that the capsule reached the caecum and their 32 matched controls, in order to keep the matching. VCE SBTT among DM patients was significantly shorter (261.2 ± 55.5 min) compared to ND controls (302.0 ± 62.7 min, p = 0.032) (Table 2, Fig. 3).
Fig. 2. Rate of complete examinations in relation to video-capsule gastric transit time in diabetic patients and controls. The number of cases that capsule reached the caecum/all cases at the respective video-capsule gastric transit time range is presented within the bars of the histogram. VCE: video-capsule endoscopy; GTT: gastric transit time; DM: diabetes mellitus patients; ND: non-diabetic controls.
4. Discussion Wireless VCE is a new method, superior to any known to date radiological and endoscopy investigations of the small bowel [11], to visualize the small intestinal mucosa and diagnose mucosal, vascular or tumour-like diseases. However, there are two main causes of incomplete examination, notably delayed gastric emptying [12], or even retention [5,13], of the capsule in the stomach and inadequate visualization of the small intestinal mucosa due to food residue of green viscous mucus in the terminal ileum [14]. DM patients frequently have delayed gastric emptying [6–8] and even gastroparesis [15]. The main pathogenetic factors in DM delayed gastric emptying are vagal autonomic neuropathy, interstitial cells of Cajal pathology and hyperglycaemia [15]. Until today, there is no study designed to evaluate the impact of DM on the completion rate of small bowel VCE. Selby [12], using logistic regression analysis, showed that several factors including DM had no effect on the completion of VCE small bowel visualization. However, the main aim of Selby’s study was the effect of metoclopramide and not DM on the completeness of VCE studies. On the contrary, our data clearly show that VCE GTT in patients with longstanding type 2 DM is at about 60 min longer compared to that
K. Triantafyllou et al. / Digestive and Liver Disease 39 (2007) 575–580
of ND controls. This results to significantly fewer complete small bowel examinations, since the caecum was visualized in less than 70% of the diabetics compared to 90% of the controls. Furthermore, in DM patients, we have also shown that the longer the VCE GTT is, the fewer will be the complete examinations. When VCE GTT was longer than 90 min, 50% of the capsules reached the caecum, and this percentage decreased progressively with increasing VCE GTT. No capsule reached the caecum when VCE GTT was longer than 240 min. Our findings have clinical and health-economical implications on the use of VCE in diabetics. Capsule endoscopy is an expensive (each capsule costs about 600D in Europe) and time-consuming examination. Since completion rate of the examination is reduced by 20% in DM, every effort should be made to increase the percentage of complete VCE studies. DM patients may benefit from early identification of prolonged retention of the capsule in the stomach by real-time VCE images and endoscopic placement of the capsule in the duodenum. Moreover, DM patients may actually benefit from systematic use of medications aiming to accelerate gastric emptying before and/or during the examination, although the use of prokinetics is controversial in VCE [3,12,16] and the efficacy of currently available prokinetics is not satisfactory in diabetic gastroparesis [17]. Both these approaches should be tested prospectively. A limitation of our study is that it is not prospective. However, matching was excellent, our sample was homogenous (type 2 DM patients) and we carefully excluded known causes of delayed transit times other than DM, such as gastric surgery and Crohn’s disease with small intestinal strictures. In addition, two independent investigators with excellent interobserver agreement evaluated all capsule videos blindly [18]. Another possible limitation is the mixture of bowel preparations used in the study. We have shown that 2 L of PEG solution or 45 mL of PS the evening before the study do not significantly affect VCE gastric or SBTTs in NDs, as compared with liquid diet alone [10]. Despite that DM patients were matched with controls, one cannot exclude the possibility that the effects of the three preparations could be different in DM as compared to controls. A prospective randomized study could resolve this uncertainty. Due to the retrospective design of our study, we have no data on the presence and extend of autonomic neuropathy and on glycaemic control. The lack of these data is not expected to influence our result interpretation. There are reports that gastric emptying correlates with the presence of autonomic neuropathy [19–21], but others have shown that this association is very weak [22,23] and although gastric emptying varies inversely with increasing glycaemia [8] in type 2 diabetics, a modest improvement in glycaemic control has no significant effect on gastric emptying in these patients [24]. Moreover, changes in blood glucose level within the physiological post-prandial range also affect gastric emptying [25], thus making the control for glycaemia very difficult during VCE examinations.
579
Our study has not been powered to detect differences regarding VCE SBTT between DM patients and ND controls. However, our finding that in the cases the capsule reached the caecum, DM patient’s VCE SBTT was faster than that of controls cannot be adequately explained and warrants further investigation. This could be explained by the selection of patients in whom VCE SBTT was complete. Since VCE GTT was longer in these patients, then VCE SBTT must have been faster to allow the capsule to visualize the ceacum. In conclusion, this is the first published report that in DM patients video-capsule GTT is significantly prolonged resulting in fewer complete small bowel investigations. Modalities, such as new powerful prokinetic medications or real-time VCE with endoscopic placement of the capsule in the duodenum, in case the capsule remains in the stomach, should be studied prospectively to examine whether they could enhance video-capsule small bowel endoscopy completeness rate in DM patients.
Practice points • During small bowel VCE studies in type 2 DM patients, the capsule remains in their stomach significantly longer than in controls. • The prolongation of VCE GTT in patients with type 2 DM results in fewer complete small bowel investigations.
Research agenda • In patients with type 2 DM, the effect of bowel preparation on the completeness rate of small bowel VCE study is unknown. • New powerful prokinetic medications should be studied prospectively to examine whether they could enhance video-capsule small bowel endoscopy completeness rate in type 2 DM patients. • Real-time VCE with endoscopic placement of the capsule in the duodenum should also be studied prospectively, as an alternative, in case the capsule remains in the stomach for long in type 2 DM patients.
Conflict of interest statement None declared.
580
K. Triantafyllou et al. / Digestive and Liver Disease 39 (2007) 575–580
References [1] Swain P, Fritscher-Ravens A. Role of video endoscopy in managing small bowel disease. Gut 2004;53:1866–75. [2] Rey J-F, Ladas S, Alhassani A, Kuznetsov K, ESGE Guidelines Committee. European Society of Gastrointestinal Endoscopy (ESGE). Video capsule endoscopy: update to guidelines (May 2006). Endoscopy 2006;38:1047–53. [3] Caddy G, Moran L, Chong A, Miller A, Taylor A, Desmond P. The effect of erythromycin on video capsule endoscopy intestinal transit time. Gastrointest Endosc 2006;63:262–6. [4] Sturniolo GC, Di Leo V, Vettorato MG, De Boni M, Lamboglia F, De Bona M, et al. Small bowel exploration by wireless capsule endoscopy: results from 314 procedures. Am J Med 2006;119:341–7. [5] Ben-Soussan E, Savoye G, Antonietti M, Ramirez S, Lerebours E, Ducrotte P. Factors that affect gastric passage of video capsule. Gastrointest Endosc 2005;62:785–90. [6] Rabine JC, Barnett JL. Management of patients with gastroparesis. J Clin Gastroenterol 2001;32:11–21. [7] Jones KL, Russo A, Stevens JE, Wishart JM, Berry MK, Horowitz M. Predictors of delayed gastric emptying in diabetes. Diab Care 2001;24:1264–9. [8] Horowitz M, Harding PE, Maddox AF, Wishart JM, Akkermans LM, Chatterton BE, et al. Gastric and oesophageal emptying in patients with type 2 (non-insulin dependent) diabetes mellitus. Diabetologia 1989;32:151–9. [9] Niv Y, Niv G, Wiser K, Demarco DC. Capsule endoscopy – comparison of two strategies of bowel preparation. Aliment Pharmacol Ther 2005;22:957–62. [10] Kalantzis C, Triantafyllou K, Papadopoulos A, Alexandrakis G, Rokkas T, Kalantzis N, et al. Wireless-capsule endoscopy gastric and small bowel transit time. Effect of three different bowel preparations. Scand J Gastroenterol; in press. [11] Neu B, Ell C, May A, Schmid E, Riemann JF, Hagenmuller F, et al. Capsule endoscopy versus standard tests in influencing management of obscure digestive bleeding: results from a German multicenter trial. Am J Gastroenterol 2005;100:1736–42. [12] Selby W. Complete small-bowel transit in patients undergoing capsule endoscopy: determining factors and improvement with metoclopramide. Gastrointest Endosc 2005;61:80–5. [13] Tatar E, Shen E, Palance A, Sun J, Pitchumoni CS. Clinical utility of wireless capsule endoscopy experience with 200 cases. J Clin Gastroenterol 2006;40:140–4.
[14] Costamagna G, Shah SK, Riccioni ME, Foschia F, Mutignani M, Perri V, et al. A prospective trial comparing small bowel radiographs and video capsule endoscopy for suspected small bowel disease. Gastroenterology 2002;123:999–1005. [15] Park M-I, Camillieri M. Gastroparesis: clinical update. Am J Gastroenterol 2006;101:1129–39. [16] de Franchis R, Avgerinos A, Barkin J, Cave D, Filoche B. ICCE consensus for bowel preparation and prokinetics. Endoscopy 2005;37: 1040–5. [17] Talley NJ. Diabetic gastropathy and prokinetics. Am J Gastroenterol 2003;98:264–71. [18] Lai L, Wong G, Chow D, Lau J, Sung J, Leung W. Inter-observer variations on interpretation of capsule endoscopies. Eur J Gastroenterol Hepatol 2006;18:283–6. [19] Kockar MC, Kayahan IK, Bavbek N. Diabetic gastroparesis in association with autonomic neuropathy and microvasculopathy. Acta Med Okayama 2002;56:237–43. [20] Murray CDR, Martin NM, Patterson M, Taylor SA, Ghatei MA, Kamm MA, et al. Ghrelin enhances gastric emptying in diabetic gastroparesis: a double blind, placebo controlled, crossover study. Gut 2005;54:1693–8. [21] Asakawa H, Onishi M, Hayashi I, Fucuda A, Tokunaga K. Comparison between coefficient of R-R interval variation and gastric emptying in type 2 diabetes mellitus patients. J Gastroenterol Hepatol 2005;20:1358–64. [22] Horowitz M, Maddox AF, Wishart JM, Harding PE, Chatterton BE, Shearman DJC. Relationship between oesophageal transit and solid and liquid gastric emptying in diabetes mellitus. Eur J Nucl Med 1991;18:229–34. [23] Merio R, Festa A, Bergmann H, Eder T, Eibl N, Stacher-Janotta G, et al. Slow gastric emptying in type 1 diabetes: relation to autonomic and peripheral neuropathy, blood glucose and glycemic control. Diab Care 1997;39:823–30. [24] Holzapfel A, Festa A, Stacher-Janotta G, Bergmann H, Shnawa N, Brannath W, et al. Gastric emptying in type II (non-insulin dependent) diabetes mellitus before and after therapy readjustment: no influence of actual glucose concentration. Diabetologia 1999;42: 1410–2. [25] Schvarcz E, Palmer M, Aman J, Horowitz M, Stridsberg M, Berne C. Physiological hyperglycemia slows gastric emptying in normal subjects and patients with insulin-dependent diabetes mellitus. Gastroenterology 1997;113:60–6.