Validation of a rapid whole blood test for diagnosing Helicobacter pylori infection

Introduction

Screening young dyspeptic patients for Helicobacter pylori and avoiding endoscopy in seronegative patients can reduce endoscopy workload by 30% without missing significant disease.1 Serological testing is commonly used to screen for H pylori but requires analysis in a central laboratory, which delays results. Cortecs Diagnostics has introduced the Helisal rapid whole blood test that claims to diagnose H pylori status within 10 minutes. This is the first near patient test to be available in the United Kingdom but its accuracy has not been independently evaluated. Our aims were to assess the accuracy of the rapid blood test against a “gold standard” and determine its effectiveness in screening young dyspeptic patients.

Subjects, methods, and results

One hundred and seventy seven patients undergoing routine endoscopy were invited to participate. Twenty patients taking antibiotics, bismuth salts, or non-steroidal anti-inflammatory drugs within a month or proton pump inhibitors within two weeks of endoscopy were excluded, as were three patients who had received H pylori eradication therapy. The remaining 154 patients had a median age of 49 (range 20-85 years). At endoscopy 85 had normal findings, 18 peptic ulcer, 21 duodenitis, 21 oesophagitis, and nine other diagnoses.

H pylori status was assessed using histology (modified Geimsa stain on two antral and two corpus biopsies), microbiology (one antral biopsy cultured on selective and non-selective media), rapid urease test (on one antral biopsy), and¹³ C urea breath test. Patients were defined as positive for H pylori (the gold standard) if two or more tests gave positive results (83 patients), negative if all tests gave negative results (69), and of indeterminate status if only one gave a positive result (two patients).

The rapid blood test was performed according to the manufacturers' protocol and had an 88% sensitivity (95% confidence interval 79 to 94%) and 91% specificity (82 to 97%) with a positive predictive value of 92% and a negative predictive value of 86% (table 1). Patients with an incorrect result on the rapid blood test were significantly older than those with a correct result (56.0 (SD13.8) v 47.7 (15.8); P<0.05, Student's t test).

Serum from 109 patients was analysed using a commercial H pylori serology kit (Helico G, Porton, Cambridge) with a 10 IU/l cut off (not all patients had the test performed as initially we had difficulty in obtaining funding for kits). Serology had 93% sensitivity (82 to 98%) and 87% specificity (75 to 95%), with one patient of indeterminate H pylori status (table 1).

Screening patients with the rapid blood test and investigating only those with a positive result would have avoided endoscopy in 41/62 (66%, 95% confidence interval 53-78%) patients aged under 45 and would have detected all seven with peptic ulcers. In older patients the test would have saved only 34/92 endoscopies (37%, 27 to 48%) and missed 2/11 peptic ulcers.

Comment

The rapid blood test is as accurate as a laboratory based commercial serology kit when compared with the gold standard battery of tests. It is more expensive than other serology kits, but results are available in 10 minutes, making it useful in primary care and outpatient clinics. The test must be independently evaluated in the area it is to be used because the accuracy of serology kits varies among different populations,2 3 and the predictive value of the test relates only to patients referred for endoscopy and may not be true in the general population.

In patients under 45 years not taking non-steroidal anti-inflammatory drugs screening with the rapid blood test would have reduced endoscopy workload by 66% while detecting all peptic ulcers. This has been reported previously for serological testing1 but has not been described for a near patient test. The rapid blood test missed two peptic ulcers in patients aged over 44, and endoscopy remains the investigation of choice in this age group.