Testing for H. pylori is usually done in the same patients for whom proton-pump inhibitors are prescribed (for example, those with dyspepsia, ulcers, or gastroesophageal reflux disease) [12]. Thus, determining the length of time that patients must not receive proton-pump inhibitors before they can undergo diagnostic testing for H. pylori without concern for false-negative results is important in clinical practice. However, few data are available to guide clinicians in deciding how long to stop proton-pump inhibitor therapy before testing for H. pylori. Current recommendations range from 5 days to 1 month [6, 13].

The aim of our study was to determine the frequency and duration of conversion of urea breath test results from H. pylori-positive to H. pylori-negative in patients given a standard course of proton-pump inhibitor therapy.

Methods

Top Methods Results Discussion References

Adult patients about to begin proton-pump inhibitor therapy for gastroesophageal reflux or dyspepsia were screened for H. pylori with endoscopic biopsy or a qualitative serologic test (FlexSure, SmithKline Diagnostics, San Jose, California). Those with a positive screening test result had a ¹³C-urea breath test (Meretek Diagnostics, Houston, Texas), which was our reference standard for H. pylori infection. Patients with a positive breath test result were eligible for enrollment. Patients taking proton-pump inhibitors, antibiotics, or bismuth compounds in the 6 weeks before study entry and patients who were allergic to lansoprazole or omeprazole were excluded. Antibiotics, bismuth compounds, and antisecretory medications were proscribed during the study.

Study patients received 30 capsules of lansoprazole (30 mg) and were instructed to take one capsule each morning. Patients returned at day 28, while still taking lansoprazole, for a repeated urea breath test (patients were evaluable if their repeated breath test was done on days 26 to 30 of lansoprazole therapy). At this visit, pills were counted and lansoprazole therapy was discontinued. If the breath test result was positive, the patient was no longer followed. If the result was negative, the urea breath test was repeated 3, 7, 14, and 28 days after the end of therapy until the breath test result reverted to positive. Breath tests were performed according to manufacturer's instructions, and breath samples were sent to a central laboratory (Meretek Diagnostics) for measurement. Breath test results are presented as the increase in enriched ¹³CO₂ concentration over baseline (delta ¹³CO₂), and values less than 2.4 are defined as H. pylori-negative. A negative result on a urea breath test 4 weeks after completion of lansoprazole therapy was considered to indicate eradication of H. pylori.

Quantitative data were compared by using the paired t-test (for breath test values in the same patients before and during therapy) or a t-test. A P value less than 0.05 was considered statistically significant. The study was approved by the institutional review board at each participating institution, and all patients gave written informed consent. Our funding sources had no role in the collection, analysis, or interpretation of the data or in the decision to submit the paper for publication.

Results

Top Methods Results Discussion References

One hundred eight patients were enrolled. Nine patients did not return 26 to 30 days after lansoprazole therapy began, 3 were given antibiotics during the study (potentially suppressing H. pylori infection), and 1 had an uninterpretable urea breath test result at 28 days (the baseline breath sample had a ¹³CO₂ level < values in the acceptable range, suggesting that the sample was inadequate). Two patients were cured of H. pylori infection, as documented by negative breath test results at 28 days. Thus, 93 patients were evaluable for the study. Baseline characteristics are shown in Table 1.

After 28 days of lansoprazole therapy, 31 (33%) of the 93 patients whose H. pylori infection was not eradicated had a negative breath test result. The proportions of patients whose breath test results were positive after completion of therapy were 91% (95% CI, 83% to 96%) at 3 days, 97% (CI, 90% to 99%) at 7 days, and 100% (CI, 96% to 100%) at 14 days (Table 2).

The mean breath test value in the 93 patients studied decreased from 20.2 at baseline to 11.5 after 28 days of lansoprazole therapy (P < 0.001). The mean baseline breath test result was lower in evaluable patients whose breath test result became negative with proton-pump inhibitor therapy than in those whose test results remained positive with lansoprazole therapy (14.8 compared with 23.0; P = 0.03).

Discussion

Top Methods Results Discussion References

We found that 33% of patients with H. pylori infection develop false-negative results on urea breath tests with a standard 4-week course of proton-pump inhibitor therapy. All patients without eradication of H. pylori infection had reverted to a positive test result by 2 weeks after completion of therapy, and 97% of patients had a positive breath test result only 1 week after therapy. Although 97% of patients with H. pylori had positive breath test results 1 week after completion of lansoprazole therapy, the 95% CI (90% to 99%) means that these observed data are statistically compatible with a proportion as low as 90%. Because of the potential that 10% of test results at 1 week would be false-negative, we believe it is appropriate for patients not to receive proton-pump inhibitors for 2 weeks before a urea breath test for H. pylori is done to ensure that the therapy does not cause false-negative results. When deciding on the optimal length of time for which therapy should be discontinued before patients are tested for H. pylori infection, individual physicians must balance the importance of avoiding occasional false-negative results against concerns about patients not receiving proton-pump inhibitor therapy for an additional week.

Variable rates of false-negative H. pylori biopsy results or breath test results have been reported in previous studies of patients taking proton-pump inhibitors [2-11]; these rates range from 13% [5] to 56% [8]. Small sample sizes, as well as differences in doses and schedules, methods of diagnosis, and timing of testing, make comparison of studies problematic. In addition, most studies were not specifically designed to determine the rate and duration of false-negative H. pylori status in patients receiving proton-pump inhibitor therapy. In the largest of these studies, Hui and coworkers [8] performed histologic assessment of H. pylori status in patients with duodenal ulcers before and after 4 weeks of treatment with omeprazole (10 mg/d or 20 mg/d) or ranitidine (150 mg twice daily). Comparison of the pretreatment and post-treatment results showed that H. pylori-negative status increased by 38% in the 10-mg/d omeprazole group (n = 77), by 56% in the 20-mg/d omeprazole group (n = 76), and by 1% in the ranitidine group (n = 79). Little information has been published about the duration of false-negative results on tests for H. pylori after proton-pump inhibitor therapy. Chey and colleagues [6] evaluated 10 patients with positive (¹⁴) C-urea breath test results who received omeprazole, 20 mg/d for 2 weeks, and 13 who received lansoprazole, 30 mg/d for 2 weeks [7]; 3 of 10 and 3 of 13 patients, respectively, developed a negative breath test result 1 day after completion of therapy. All breath test results reverted to positive 5 days after therapy.

Nonendoscopic tests for the diagnosis of H. pylori are now widely available. Serologic or whole-blood antibody tests are commonly used to screen patients for H. pylori because they are easy to use and inexpensive. However, if a patient has been treated for H. pylori infection, antibody test results can remain positive for years and are therefore not generally useful in the assessment of post-treatment H. pylori status [14]. Urea breath testing is the most accurate noninvasive method of H. pylori diagnosis, with results rivaling those of endoscopic biopsy tests, and it is the test of choice to document H. pylori status after a course of eradication therapy [15]. When the breath test is performed correctly in clinical practice, the only other major cause of inaccurate results (in addition to proton-pump inhibitor therapy) is the use of antibiotics or bismuth compounds, which suppress the number of H. pylori organisms and thus decrease urease activity [15]. To avoid this confounding factor, we excluded patients taking antibiotics, bismuth compounds, or other antisecretory drugs.

The development of false-negative breath test results in patients taking proton-pump inhibitors may be due to the direct anti-H. pylori activity of these agents. Other possible mechanisms include decreased viability of the organisms at high intragastric pH with potent antisecretory therapy and direct inhibition of urease activity by proton-pump inhibitors [7].

In summary, our prospective study demonstrates that 33% of H. pylori-positive patients develop false-negative urea breath test results while taking a standard course of proton-pump inhibitor therapy. Most patients with H. pylori infection revert to positive urea breath test results by 1 week after the discontinuation of proton-pump inhibitor therapy; however, to ensure that a false-negative result does not occur, patients should not receive proton-pump inhibitors for 2 weeks before a urea breath test for H. pylori is done.

From the University of Southern California School of Medicine, Los Angeles, California; and Oregon Health Sciences University, Portland, Oregon.

Drs. Knigge and Fennerty: Division of Gastroenterology, PV-310, Oregon Health Sciences University, 3181 Southwest Sam Jackson Park Road, Portland, OR 97201.