Dysmotility and Proton Pump Inhibitor Use Are Independent Risk Factors for Small Intestinal Bacterial and/or Fungal Overgrowth

C. Jacobs; E. Coss Adame; A. Attaluri; J. Valestin; S. S. C. Rao


Aliment Pharmacol Ther. 2013;37(11):1103-1111. 

In This Article

Materials and Methods

We evaluated 150 consecutive patients who presented to a single gastroenterologist between the years of 1995–2010. These subjects had unexplained gastrointestinal symptoms. All of these patients had a negative evaluation for routine gastrointestinal pathology including a normal gastroscopy, colonoscopy, CT scan, routine haematology and biochemical profiles, anti-tTG, TSH, right upper quadrant ultrasound and small bowel follow-through series. Patients with known gastrointestinal problems including previous GI surgeries (except cholecystectomy, hysterectomy and appendectomy), and those who were using medications that potentially affect intestinal motility (opioids, anticholinergics, antidiarrhoeals) and those with significant co-morbid medical problems or those who were hospitalised were excluded.

The study was approved by University of Iowa Hospitals and Clinics Investigation Review Board.

Symptom Questionnaire

A validated bowel symptom questionnaire was administered to all subjects prior to the study.[8] It enquired about the presence or absence of the following ten symptoms in the preceding 2 weeks: abdominal pain, chest pain, belching, bloating, fullness, indigestion, nausea, diarrhoea, vomiting and gas. If present, patients were asked to rate each symptom's frequency, intensity and duration on a 0–3 Likert-like scale. Intensity: 0 = no symptoms, 1 = mild, 2 = moderate, 3 = severe symptoms. Frequency: 0 = None; 1 = Less than 1 episode/week, 2 = 1 episode/week, 3 = More than 1 episode/week. Duration: 0 = None, 1 = Less than 10 min, 2 = 10–30 min, 3 = Greater than 30 min. On this scale, the total score for each symptom could range from 0 to 9. A mean total score for all 10 symptoms was calculated for each patient.

Patients' medications were documented and additionally the hospital electronic medical record database was used to confirm the use of PPIs during their initial presentation and during evaluation of their GI symptoms.

Antro-Duodeno-Jejunal Manometry (ADJM)

Manometric System. We used a 250 cm long elastic catheter that was custom-built with six solid-state pressure transducers (Koningsberg Instruments, Pasadena, CA, USA). The probe was connected to a six-channel portable solid-state digital data-logger (MicroDigitrapper 4 Mb, Medtronics; Minneapolis, MN, USA) with a sampling frequency of 4 Hz, A–D conversion, temporary storage up to 4 Mb. Upon completion of the study, data were downloaded to an IBM-compatible personal computer for analysis (Gastrosoft version 6.3, Multigram, Medtronics, Minneapolis, MN, US).

Study Protocol

Following an overnight fast, all subjects had an upper endoscopy. Next, using sterile precautions, a 2 mm Liguory catheter (COOK Medical, Bloomington, IN, US) was passed through the biopsy channel of the upper endoscope into the 3rd and 4th portions of the duodenum. Using gentle suction, approximately 3–5 mL of duodenal fluid was aspirated, and the specimen was sent to microbiology for aerobic/anaerobic culture and fungal culture. Next, the nares were numbed with 2% lidocaine gel and the 6-sensor solid-state manometry probe was placed under endoscopic and fluoroscopic guidance such that two sensors (5 cm apart) were located in the antrum, two sensors (15 cm apart) were located in the duodenum, and two sensors (15 cm apart) were located in the jejunum.

The ambulatory recorder was placed in a shoulder bag and the patients were free to ambulate throughout the study and slept at home. Six hours after probe placement, all patients ate a 600 kilocalorie standard meal consisting of a chicken sandwich, 6 oz. of milk, a cookie and a banana. The nutrient composition was 52% carbohydrates, 25% protein and 23% fat. The following morning they were instructed to wake up at 6 am. The motility recording was continued until 11 am and thereafter the probe was removed. An event marker was attached to the recorder, and the patients were encouraged to use this and mark the time of events such as eating, walking and sleeping or to indicate the occurrence of symptoms such as abdominal pain, passing flatus, etc. They were also provided with a diary, in which they described any event(s) or symptom(s), and recorded its time and duration.

Data Analysis

After completion of the recording, the data stored in the portable recorder was transferred to a personal computer for visual display and analysis. Tracings were analysed by visual inspection for motility patterns such as phase III MMCs, and for quantitative assessment of pressure activity such as area under the curve (AUC) of the pressure waves. Pressure waves that occurred simultaneously in several channels with similar amplitude and duration of ≤3 s were identified as artefacts and excluded from the analysis. Phase III MMCs were defined as propagating clusters of repetitive contractions with a frequency of 3/min in the antrum and 11–13/min in the duodenum and with a duration of at least 3 min that was followed by a period of motor quiescence.[3,9–11] The pressure activity data from each of the two sensors located in the antrum, duodenum and jejunum were averaged and were used as an overall index of motility for each segment.

Based on the manometric findings, we classified our patients as follows: (i) Normal motility (with both normal frequency and intensity of pressure activity and coordination); (ii) Neuropathy (with normal frequency and intensity of pressure activity and lack of coordination); (iii) Myopathy (With normal frequency, low intensity of pressure activity and normal coordination); and (iv) Mixed (features of neuropathy and myopathy).[3,11]

Furthermore, each patient was classified as having either dysmotility if they had two or more of the following characteristics: absence of phase III MMC activity (neuropathy), absence/diminished postprandial response (2 s.d. of normal), diminished amplitude of antral/intestinal phasic activity (≤20 mmHg) or impaired antro-duodenal coordination[10,11] (propagation of peristaltic waves between antrum and duodenal sensors).

Patients were considered to have small intestinal overgrowth if microbiology reported a positive culture for aerobic, anaerobic or fungal organisms. Bacterial concentration ≥103 CFU/mL was considered as positive for SIBO. In addition, we also assessed the prevalence of bacterial concentrations ≥105 CFU/mL (used for jejunal samples[12]). Because normally there is no fungus or very low concentrations of fungal organisms in the small bowel,[13] a diagnosis of SIFO was made if the duodenal culture yielded growth of fungal organisms.

Based on culture results, patients were categorised into four groups: SIBO, SIFO, mixed SIBO/SIFO and negative for proximal small bowel overgrowth. A second set of analysis was done on two groups: overgrowth positive (patients with SIBO and/or SIFO) and overgrowth negative (culture negative for bacteria and fungi).

Proton pump inhibitor users consisted of patients that were taking PPIs at the time of their GI evaluation at our centre. PPI use was continued throughout their motility testing. The duration of PPI use and patient compliance with PPI treatment regimen, prior to the onset of symptoms could not be accurately assessed from our questionnaire. PPI users and PPI non-users were compared for the presence or absence of overgrowth. In addition to questionnaire documentation of PPI use an independent assessment of the patient's chart was performed to confirm/refute PPI use.

The principal investigator (SSR) was not involved in the data analysis and interpretation of study results and was blinded to study ID during discussion of study results. The data analyses were performed independently by CJ and EC-A and verified by AA with assistance from JV.

Statistical Analyses

Fisher's Exact Test was used for analysis. Individual contingency tables were constructed for overgrowth vs. dysmotility and overgrowth vs. PPI use. Each individual subset of culture data (SIBO, SIFO and Mixed SIBO/SIFO) vs. dysmotility and vs. PPI use was similarly analysed. A P value < 0.05 was considered to be significant. Comparative analyses were performed for both the ≥103 and >105 CFU/mL bacterial concentration and are described under the results section. However, for the key discussion and analysis, the data analysed at a bacterial concentrations of >103 CFU/mL was used. Odds ratios were calculated for each significant contingency table. Symptom analysis was done by constructing 2 × 2 contingency tables. Fisher exact test and odds ratios were obtained with 95% confidence intervals.