Lactulose Breath Testing Does Not Discriminate Patients With Irritable Bowel Syndrome From Healthy Controls

Jason R. Bratten, B.S.; Jennifer Spanier, D.O.; Michael P. Jones, M.D.


Am J Gastroenterol. 2008;103(4):958-963. 

In This Article


This study demonstrates that healthy subjects and patients meeting Rome II criteria for IBS have similar patterns of breath H2 and CH4 production and comparable orocecal transit times using a standardized LBT protocol. Thus, LBT does not reliably distinguish patients with IBS from healthy controls, and the utility of this diagnostic method needs to be questioned.

The frequency of positive LBT in our patients with IBS was 74%, which is consistent with the previous reports of Pimentel and co-investigators.[10,11] In contrast to these reports, LBT was positive in 85% of our control subjects, which contrasts sharply with the 20% incidence of abnormal LBT reported by Pimentel et al. in their 15-subject control group.[7,11] Our findings are consistent with Walters and Vanner; however, they reported abnormalities in 75% of a group of 20 controls.[16] Dual breath H2 peaks were less commonly seen, occurring in 14% of patients with IBS and 26% of controls. Our data are also consistent with the recent work of Posserud and colleagues.[26] These investigators performed LBT in 46 patients with IBS and <10[5] colonic bacteria/mL and 21 controls. In all, seven (15%) patients and four (20%) controls had a positive test using the double-peak definition (P = NS). A 20-ppm rise in H2 within 90 min was observed in 35% of patients and in 45% of controls (P = NS), while a 20-ppm rise in H2 within 180 min was observed in 78% of patients and in 70% of controls (P = NS). Thus, the majority of patients and controls will have an abnormal LBT using the criteria of Pimentel and colleagues.

No differences were seen between patients with IBS and controls with respect to the rise in breath H2 over time, and our data are exactly consistent with those of both Posserud et al.[26] and Walters and Vanner[16] in this regard. Additionally, no differences existed with respect to the calculated orocecal transit times for patients with IBS and controls, further highlighting the lack of discriminatory utility of LBT in this setting.

Patients with IBS and controls also did not differ with respect to the prevalence of CH4-producing subjects. Of note, patients with IBS who were CH4 producers were significantly more likely to report constipation, and significantly less likely to report diarrhea. The association of methanogenic flora and constipation is not novel.[24,27,28,29] While this observation is potentially important, the directionality of the relationship and potential therapeutic implications are unknown. While acute animal studies have shown that intestinal transit time is slowed in CH4-perfused bowel,[30] CH4 production is also diminished with nonantimicrobial treatments of constipation, such as enemas and laxatives,[28,31] suggesting that higher levels of breath CH4 may be as much a biomarker as a pathophysiologic mechanism. Finally, the presence of methanogenic flora even in patients previously treated with antibiotics appears quite persistent over time, arguing against antibiotic therapy for CH4-producing patients with constipation as a viable treatment for a chronic disorder.[24]

The strengths of this study are its sample size, collection of patients from a variety of gastroenterology practices referring to our physiology laboratory, inclusion of the largest control group reported to date, and concomitant assessment of both breath H2 and CH4. The greatest limitation of the study is that we employed a brief symptom assessment asking patients to simply endorse dominant complaints. We deliberately chose not to employ symptom recall or attempt symptom diaries to avoid the recall bias of the former and compliance issues of the latter. While a symptom diary would have been a more robust technique, the recruitment protocol for IBS subjects made the likelihood of successfully collecting an adequate number of symptom diaries post-breath testing unlikely, and we felt that diagnostic classification could be adequately ascertained by study interview at the time of enrollment. It should also be pointed out that, while our patients with IBS were recruited from a population referred for clinical testing, enrollment criteria for this study were established a priori, and enrollment was prospective. Our controls were recruited largely from medical staff and student populations. This provided a readily accessible group that, while carefully screened, did turn out to be younger than our study population. While gut function may alter over time, neither patients nor controls were senescent, and physiologic differences between the two groups seem unlikely. In fact, we did not find differences with respect to age for breath-testing parameters in either the IBS or control group. Finally, we collected breath samples at 20-min intervals with the exception of the final two collections that occurred at 30-min intervals. It is possible that these protracted intervals between collections may have altered calculations regarding orocecal transit time. This seems unlikely, however, as the mean transit time occurred prior to the prolonged collection intervals.

Finally, our protocol for LBT differs from that used by Pimentel and colleagues, and it is possible, but unlikely, that this may explain the differences in our observations. Pimentel et al. have typically used 10 g of lactulose in 1-2 ounces of water,[11,29] while we used 10 g of lactulose in 8 ounces (240 mL) of water.[21] Lactulose itself accelerates small intestinal and colonic transit,[22,32] but the acceleration of orocecal transit is influenced more by the dose of lactulose rather than the osmolarity of the lactulose solution.[33] Also, our results are consistent with other investigators who have used slightly different LBT methods. Walters and Vanner administered 10 g of lactulose in 120 mL of water,[16] while Posserud et al. administered 15 mL of a lactulose syrup to provide a 10-g dose.[26] Two points seem clear from this. First, the technique requires standardization. Second, despite differences in technique, the three groups, using three different methodologies, have consistently demonstrated that patients with IBS and controls do not appear to differ when applying the proposed criteria for an abnormal LBT.

In summary, evaluation of a large group of patients with IBS and controls using LBT and previously reported test criteria determined that the majority of patients and controls have abnormal breath tests. Additionally, breath H2 production, CH4 production, and orocecal transit times do not differ between the two groups. We conclude that LBT, in its current form, using the proposed criteria for an abnormal test, does not discriminate patients with IBS from healthy subjects.

CLICK HERE for subscription information about this journal.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.