Bacterial Overgrowth and IBS: Causation, Association, or Neither? Brennan Spiegel, MD, MSHS.

Bacterial Overgrowth and IBS: Causation, Association, or Neither?

Brennan Spiegel, MD, MSHS

The Brain-Gut Axis

Central nervoussystem (CNS)

Enteric nervoussystem (ENS)

Dysbiosis“Leaky” gutLow-grade inflammationVisceral hypersensitivityDysmotility

Brain-gut axis

Higher brain activation in response to stress

Thinner grey matter density

Altered amygdala reactivity

Mayer E. et al; Gastroenterol 2010;139:48

Mayer E. et al; Gastroenterol 2011;140:1943

Susceptible Host

Disease Expression

“Hit”Stress InfectionDietAllergy

Evolving IBS Disease Model

Underlying dysfunction in:• Intestinal dysbiosis • Mast cell number and function• Serotonin trafficking• HPA Axis• Cortical pain processing

Dietary factorsHigh sorbitol diet

High-fiber dietFODMAP Diet

Caffeine Alcohol

InflammationUlcerative colitisCrohn’s disease

Microscopic colitis

EndocrineHyperthyroidism

DiabetesCarcinoid

Gastrinoma

Psychological

AnxietySomatizationDepression

PTSD

InfectionSIBOC. diff

Giardiasis

MalabsorptionCeliac sprue

Carb intolerancePancreatic disease

Bile acid malabsorption

Existential Question: What Is IBS?

IBS

Is IBS an absence of other things?

Or is it some thing… unto itself?

IBS

InflammationAltered brain–

gut interactions

Visceral hypersensitivity

Geneticfactors

Psychosocialfactors

Bacterial-Host Interactions

IBSIBS

Proposed Pathophysiological Mechanisms Involved in IBS

IBS “Look-Alikes”

Bacterial overgrowth

Giardiasis

Hyperthyroidism

IBD

Lactose intolerance

Infectious colitis

Microscopic colitis

Celiac sprue

Diagnostic Battery is Extensive

Diagnostic Battery

Breath-testing

Stool Ova & Parasites

Thyroid function testing

ESR / CRP / Colonoscopy

Breath-testing

Stool leukocytes / C&S / C. diff

Colonoscopy / Flexsig

Sprue Serologies

Performing the wrong tests can lead to excessive

resource utilization and worsen patient outcomes

Performing the wrong tests can lead to excessive

resource utilization and worsen patient outcomes

IBS “Look-Alikes”

Bacterial overgrowth

Giardiasis

Hyperthyroidism

IBD

Lactose intolerance

Infectious colitis

Microscopic colitis

Celiac sprue

Diagnostic Battery is Extensive

Diagnostic Battery

Breath-testing

Stool Ova & Parasites

Thyroid function testing

ESR / CRP / Colonoscopy

Breath-testing

Stool leukocytes / C&S / C. diff

Colonoscopy / Flexsig

Sprue Serologies

“A revolutionary tool in the way irritable bowel syndrome is currently

being treated”

“A revolutionary tool in the way irritable bowel syndrome is currently

being treated”

“I am convinced that the primary cause of IBS is bacterial overgrowth –

‘The New Culprit’”

“I am convinced that the primary cause of IBS is bacterial overgrowth –

‘The New Culprit’”

duodenum jejunum ileum colon

100 103 106 1011

Normal Distribution Intestinal Flora

aerobes aerobesanaerobes

duodenum jejunum ileum colon

SIBO = >105 CFU/ml of “colonic type bacteria” in SI

e.g. gram negatives, anaerobes, enterococci

Aspirate

H2

breath tests

glucoselactulose

Normal Distribution Intestinal Flora

Lactulose Hydrogen Breath Test

H 2

H2

H2

PPM

Lactulose

min 900

cecum

colonic fermentation

LHBT criteria for diagnosing SIBO

H2

PPM

90 min 180 min

0

colonic fermentationlactulose

10

20

30

40double peak

rise H2 before 90 min

> 20 ppm H2 rise by 180 min

jejunum cecum

Pimentel M et al. New Engl J Med 2011;364:22-32

Rifaximin in IBS

Pimentel M et al. New Engl J Med 2011;364:22-32

• It’s sensible that X could cause Y (biological plausibility)

• Competing mechanisms are less plausible (coherence)

• Most studies show X and Y are related (consistency)

• X comes before Y (temporal relationship)

• Removal of X removes or improves Y

• Therapies targeting X are most effective for Y (effect size)

• Nothing gets between X and Y (lack of confounding)

Does X Cause Y? Criteria for Causation

Hill, B. Proc R Soc Med 1965;58:295

The LHBT was used to promulgate the SIBO hypothesis of IBS. But how do we know it even tests for SIBO?

Question

Yu et al. Gut 2011 60:334

“These finding demonstrate that an abnormal rise in hydrogen

measured in the LHBT can be explained by variations in oro-

cecal transit time in patients with IBS and therefore do not

support the diagnosis of SIBO.”

“These finding demonstrate that an abnormal rise in hydrogen

measured in the LHBT can be explained by variations in oro-

cecal transit time in patients with IBS and therefore do not

support the diagnosis of SIBO.”

“It seems probable that antibiotics reduce bloating in IBS by

suppressing species of bacteria in the colon, not the small

intestine. Altering colonic flora using alternative strategies to

antibiotics, such as pre or probiotics, may prove advantageou

s.”

“It seems probable that antibiotics reduce bloating in IBS by

suppressing species of bacteria in the colon, not the small

intestine. Altering colonic flora using alternative strategies to

antibiotics, such as pre or probiotics, may prove advantageou

s.”

• The test used to promulgate the SIBO hypothesis may not have measured SIBO in the first place.

• LHBT is probably a measure of colonic bacteria that are not displaced from the colon itself.

• LHBT data may simply indicate that dysmotility underlies IBS – something long established.

SIBO and IBS: Biological Plausibility?

Camilleri et al. CGH 2008;6:772Pimentel et al. DDs 2002;47:2639

Yu et al. Gut 2011 60:334

• In order for a risk factor to be the predominant cause of a disease, we should expect that competing hypotheses are less plausible.

•A causal theory is strongest when it does not conflict with other facts and there are no plausible competing theories.

SIBO and IBS: Biological Plausibility?

Hill, B. Proc R Soc Med 1965;58:295

SIBO and IBS:

Consistency among studies?

Prevalence of Abnormal* Lactulose Breath Tests in Rome I IBS

IBS Controls

100

0

50

Pimentel et al. Am J Gastro 2003;98:412

84%

20%

N=111 N=15

OR=26.2 (95% CI=4.7, 104)

*Single peak >20 ppm rise of H2 by 90 min

0

10

20

30

40

50

60

70

80

90

IBS vs. Controls: H2 rise > 20 ppm by 180%

Po

sit

ive

Pimen

tel e

t al

A

JG 2

003

<0.001IBS

Control

N=126

NS

Posser

ud et a

l

G

ut 200

6

N=204

NS

Wal

ters

et a

l

AJG 2

005

N=126

NS

Bratte

n et a

l

AJG

200

8

N=42

NS

Grove

r et a

l

NGM

200

8

N=192

Positive Lactulose Breath Test: Odds in IBS vs. Controls

0.1 0.2 0.5 1 2 5 10 100 1000

Rana 2008 12.38 (1.96, 513.13)

Grover 2008 2.29 (0.86, 7.16)

Bratten 2008 0.45 (0.18, 1.23)

Posserud 2007 1.13 (0.14, 52.89)

Parodi 2007 14.00 (3.26, 124.54)

Lupascu 2005 10.89 (3.33, 45.67)

Pooled OR (95% CI) 3.45 (0.94, 12.72)

Ford, Talley, Spiegel, Moayeddi . Clin Gastro Hep 2009

SIBO and IBS:

Dose Response Relationship?

Clinical Relevance of Abnormal SI Bacterial Counts

Comparison Group Symptom Differences

>105 vs. <105 CFU/ml No difference in IBS symptoms

>103 vs. <103 CFU/ml No difference in IBS symptoms

LHBT double peak vs. no double peak No difference in IBS symptoms

Posserud et al. DDW 2009 AB#1052

“We found no clear association between symptoms and

culture verified SIBO, mildly increased small bowel

bacteria or abnormal LHBT. Thus, the clinical

importance of abnormal small bowel bacterial counts

in IBS patients is questionable.”

“We found no clear association between symptoms and

culture verified SIBO, mildly increased small bowel

bacteria or abnormal LHBT. Thus, the clinical

importance of abnormal small bowel bacterial counts

in IBS patients is questionable.”

SIBO and IBS:

Are treatments for SIBO more effective than treatments targeting other purported

mechanisms of IBS?

Rifaximin in IBS

Pimentel M et al. New Engl J Med 2011;364:22-32

NNT = 11NNT = 11

IBS Treatment NNT vs. PlaceboAntidepressants 4Antispasmodics 5Linaclotide* 5Alosetron* 8Rifaximin* 11Probiotics 11Fiber 11Lubiprostone* 13

Comparing NNTs of Available Pharmacotherapies for IBS

Spiegel B. Clinic Gastroenterol Hep. 2011;9:461

* Higher quality studies available to support NNT calculation

Even Peppermint Oil Looks Good in IBS

Ford A, Talley N, Spiegel B et al. Am J Gastroentrol 2008 Nov 13;337:a2313.

NNT = 2.5

NNT of “Placebo without Deception” in IBS?

4

Kaptchuk et al. PLoS One 2010;22:5

SIBO and IBS:

Lack of confounders?

I plan to use antibiotics regardless of the HBT or jejunal aspirate. In fact, I don’t

check the HBT anyway. Will clearing SIBO with rifaximin lead to a long-term cure?

Or will the SIBO just come back?

Question

50

40

30

20

10

0

% G

lc H

BT P

ositi

ve

3 Months

SIBO Recurrence Following Rifaximin

13%

6 Months 9 Months

28%

44%

Lauritano et al. Am J Gastroenterology 2008;103:2031

N=80

Predictors of SIBO Recurrence:

Older age (OR=1.09)

History of appendectomy (OR=5.9)

Use of proton pump inhibitors (OR=3.5)

Predictors of SIBO Recurrence:

Older age (OR=1.09)

History of appendectomy (OR=5.9)

Use of proton pump inhibitors (OR=3.5)

Relationship between PPI Use and Foregut Bacterial Counts

Thorens et al. Gut 1996; 35: 23-6.

Species Found in Foregut of Patients on PPIs

Fried et al. Gut 1994; 35;23-6.

PPIs as the Confounder: Could IBS Be Linked to SIBO Through PPIs?

Spiegel B, Chey W, Chang L. Am J Gastroenterol 2008;103

Causation?

1. Biological plausibility?

2. Consistency?

3. Lack of confounding?

4. Removing X provides cure or robust improvement in Y?

5. Removing X is most effective treatment for Y?

SIBO IBSCriteria for causation:

Yes No

Take-Away Messages

• Test used to promulgate SIBO hypothesis may not have measured SIBO in the first place

• A dose-response effect between intestinal microbiota and IBS symptoms is lacking

• The relationship between SIBO and IBS is inconsistent

• Therapies that do not address SIBO often outperform antibiotics – even “placebo without deception”

• PPIs confound the relationship between SIBO and IBS

• Bottom line: Bacteria may play a role in some symptoms in some patients, but are unlikely to be the predominant cause of IBS.