Journal Club How to Use and Article About

Journal Club How to Use and Article About a Diagnostic Test Satyen Gada, MD LT MC (FS) USN NCC Pediatric Residency Program

Patients Patient 1: 11 yo Caucasian male seen in continuity clinic as a follow up from ER for CC of H. pylori infection. Review of the ER sheet significant for hx of burping with increased frequency and halitosis over the last two days. Family hx of uncle living in California dx with gastritis 6 months ago. He was positive for H. pylori on a blood test. Pt was started on therapy for presumptive H. pylori gastritis. CBC, Stool Guaiac, and H Pylori blood test sent.

Patients Patient 1: In the office, hx from ER is consistent. However, his snxs have completely resolved. All further hx, ROS, PE unremarkable. Lab results neg/wnl. Mother concerned that the test was performed incorrectly or not a good one. According to the internet, her son must have H. pylori, and he did get better on abx.

Patients Patient 1: Your preceptor, a GI specialist, tells you to write the note as she discusses the case with the mother. After some time, the mother storms out angry and upset. Your preceptor shakes his head and throws away two bottles of antibiotics.

Patients Patient 2: The same day in clinic, you evaluate another 11 yo male with abdominal pain. Pt has a two week hx of worsening epigastric pain with occasional emesis. Pt has had a three pound wt loss. States his stool are now black. Increased consumption of milk because it “feels better”. Mother states that he was evaluated in the ER for similar sxns 2 months ago and resolved with four weeks of tx with Zantac. H. pylori blood test at that time was negative. Mother states she also had similar sxns and was diagnosed with an ulcer 1 yr ago and given an antacid which seemed to help.

Patients Patient 2: He is currently not taking any medication. Vital signs are stable, physical exam significant for epigastric tenderness on deep palpation and stool positive for occult blood. Your preceptor, a GI specialist, tells you he/she will take care of it and to go ahead and see the triplet NICU f/u’s that have been waiting for 1 hr while you obtained the above detailed hx.

Patients Patient 2: After your NICU f/u, your preceptor tells you despite the negative H pylori blood test, this patient warrants further testing for H pylori with a “breath test” in the GI clinic. You did not appreciate the patient having bad breath, but agree with the plan. When you put the order in AHLTA, you notice there are four available tests for H pylori: blood, urine, stool, and breath test.

Patients Patient 2: The patient returns following the GI appointment with a diagnosis of H pylori peptic ulcer disease, and is being treated with the same regimen that was given to Patient #1 by the ER. The patient returns for a school physical 10 months later and has been sxn free.

Clinical Questions

Clinical Questions 1. Both patients had neg H. pylori blood tests. Why was one patient referred to more testing while the other told to discontinue therapy? 2. Why was the breath test ordered for patient #2 over the urine or stool? 3. How do I approach H. pylori testing with future patients?

Hmm… You return home a bit confused. Looking in the mirror, your reminded of how MOTO you are. You search MD consult to learn more about H. pylori and H. pylori testing and find an article titled….

Hmm… Comparison of non-invasive tests to detect Helicobacter pylori infection in children and adolescents: Results of a multicenter European study. Megraud et al. Journal of Pediatrics Vol 146: 2. Feb 2005.

Hmm… While you are deleting your junk e-mail that evening, you briefly read your Chief Resident’s messages. One is about some EBM articles on the nccpeds. com website. You print out the section on how to use an article about a diagnostic test.

H. pylori § H. Pylori is a gram-negative spiral bacteria and is estimated to infect more than half of the world’s population, predominantly in developing countries. § Organism is well established as the cause of gastritis-associated GI diseases: gastric ulcer, duodenal ulcer, gastric cancer, and MALToma.

H. pylori § H. pylori is one of the most common chronic bacterial infections among humans. § Most people are infected before age 10. § Mode of transmission is person-to-person by fecaloral or gastro-oral routes. § Primary caregivers are the reservoir for spread of infection to children. § In areas of poor sanitation, contaminated water or food may be the primary reservoir.

H. pylori § More prevalent in developing countries. § Inverse relationship of prevalence with socioeconomic status. § Estimated prevalence in the US: African Americans: 50%, Mexican-Americans 60%, Whites 26% § Prevalence in Western Europe similar to US.

H. pylori § Variable latent period of sub-clinical infection which causes gastric mucosal inflammation and progressive mucosal damage, leading symptomatic infection. § Age of onset depends on virulence of strain and age of initial infection. § Lifetime risk for H. pylori-infected individual to develop peptic ulcer disease is 1 in 6 § Risk for developing gastric cancer in countries with high prevalence of H. pylori ranges from almost 20% in China to 1. 5% in the UK.

H. pylori § Studies have shown that H. pylori does not cause GERD, and eradication therapy for H. pylori does not treat GERD long term. § Those who are positive with H. pylori will have recurrence of sxns unless H. pylori is eradicated.

Diagnostic Tests for H pylori infection Tests are divided into those which require endscopy (invasive) and those which do not (non-invasive).

H. pylori Invasive tests: 1. Gastric biopsies for H. pylori culture- gold standard. (can also be obtained with gastric brushings). 2. Biopsies for histology: evaluate severity of gastritis and density of organism. 3. Urease testing: Add solution with urea: enzyme converts urea to ammonia and change noticed on p. H indicator.

H. Pylori 1. 2. 3. 4. Noninvasive tests: Serologic testing: Ig. G ant-H. pylori antibodies in serum-genrally present 4 wks following infection. Stool antigen detection Antibody detection in urine Urea breath test: Drink solution containing urea with labeled carbon. Urease activity by H. pylori yields NH 3 and CO 2, which can be measured with spectrometry.

H. pylori Treatment: H 2 blockers/PPI and: Amoxicillin/Metronidazole + Clarithromycin x 14 d Eradication rates are 60 -80% Tx failures should be re-treated with Amoxicillin and PPI/H 2 blockers + another antibiotic.

Study: Methods 1. Open, prospective, multi-center study. 2. Inclusion criteria: Patients between ages 2 and 17 included if required an upper endoscopy and H. pylori testing. 3. Exclusion criteria: Previous H. pylori eradication tx, consumption of abx, antisecretory drugs, bismuth salts in last 2 weeks, contraindication to endoscopy/biopsy.

Methods 1. Estimated 600 children (2 -11) and 200 adolescents (12 -17) needed to have at least 30 H. pylori positive patients per age group. 2. Study stopped when more than 30 H. pylori positive patients would be included in each age group. 3. Within one week, endoscopic exam with biopsies, 13 C urea breath test, stool antigen test, serology, and antibody detection in urine were performed.

Methods Definition of H. pylori status: § Positive H. pylori defined as positive culture from endoscopy specimens. § In case of negative culture, positive results of both histology and rapid urease test. § Negative H. pylori if all invasive tests gave negative results. § Discrepant results were excluded.

Methods § § § Noninvasive testing 13 C urea breath test Stool antigen test Serology Antibody detection in urine

Methods: Are the results valid? 1. Can you accept the reference standard? Yes: Culture is gold standard for identification of H. pylori. Studies have shown histology/urease testing also to have high sensitivity (90 -95%) and specificity (95 -100%) 2. Where the reference standard and tests assess independently of each other? Yes: All patients received all tests and they were performed blinded.

Methods: Are the results valid? § Did the results of the test influence the decision to perform the reference standard? No: all tests were performed and blinded. § Were the methods of performing the test described in sufficient detail to permit replication? Yes: Name of test, company, and method of storage/performance were outlined.

Methods: Are the results valid? 1. Did the patient sample include an appropriate spectrum of patients to whom the diagnostic test will be applied in a clinical practice? (see Table 1) Pros + Age inclusion was broad 2 -17. Good exclusion criteria. + Good male/female ratio + Varied reasons for workup/findings on endoscopy

Methods: Are the results valid? Cons - Almost exclusively Caucasian. - Small number in 2 -5 age range - Author comments on number of immigrant children in test as well as the fact that all patients warranted endoscopy. - Patients from Western Europe: however, prevalence similar to US.

Interpreting Results of Diagnostic Testing • Pretest probability: Start with the patient presenting with the constellation of sxns and signs. Even though two patients may have the same results, the probability of those results meaning anything is different-you may treat one and order further studies in another. • Pretest probability, after results are obtained, alters post-test probability. Altered by the likelihood ratio of the test.

Interpreting Results of Diagnostic Testing 1. Likelihood ratio (LR): How likely is it that the given diagnostic test will raise or lower the post test probability of the target disorder. LR of 1 means that the post test probability is exactly the same as the pretest probability. 2. Ratio of the probability of a positive test confirming the disorder to a negative test reducing the chance of a disorder. 3. 1/LR = unlikelihood ratio

Interpreting Results of Diagnostic Testing 1. LR greater than 10 or less than 0. 1 generate large and often conclusive changes from pretest to post-test probability. 2. 5 -10 or 0. 1 to 0. 2 generate moderate shifts. 3. 2 -5 and 0. 5 to 0. 2 small changes. 4. 1 -2 and 0. 5 to 1 alter to small degree (rarely important).

Interpreting Results of Diagnostic Testing § Normogram allows us to go from pretest to post test probability. § Pretest, LR, and Post Test Probability. § Can use sensitivity and specificity to create a receiver operator curve. § Post test probability leads to interpretation/treatement (or lack of).

What are the results? § § Sensitivities, specificities, predictive values and diagnostic accuracy were determined in separately for each diagnostic test. ROC were also created

Results Test is positive Test is negative Patients with Patients disease without disease a b c d Sensitivity= a/a+c Specificity= d/b+d

Results Positive/Negative predictive values depend on prevelance of disease: Low prevalence will yield more false positives, and therefore lower positive predictive value. Low prevalence will yield more true negatives, and therefore higher negative predictive value.

Results § 503 recruited for study § 316 had results of all non-invasive tests and fulfilled definition of +/- H. pylori status § Of the 316, 42% positive for H. pylori. § (High prevelance of disease)

Table II. Performances of the diagnostic tests for the 316 patients with gold standard and four tests performed (UBT, Hp. SA, Urinelisa and Pyloriset EIA-G) Age group (y) Helicobacter test INFAI Hp. SA Urinelisa Pyloriset EIA-G Rapirun Urea Breath Stool Urine Blood Urine 2 ∗∗ Sensiti vity Global 96. 2 [91. 9 -98. 6] 72. 9 [64. 980. 0] 63. 2 [54. 771. 0] 88. 7 [82. 593. 3] 30. 2 [22. 538. 9] ∗∗∗ 80. 3 [73. 0 ∗∗∗ 72. 2 [64. 1 ∗∗∗ 90. 2 [84. 2 -86. 5] -79. 3] -94. 4] Specifi city Global 97. 3 [94. 0 -99. 0] 97. 3 [94. 099. 0] 93. 4 [89. 196. 4] 98. 7 [95. 799. 8] ∗∗∗ 93. 4 ∗∗∗ 93. 9 [89. 1 -96. 4] [89. 1 Accura cy Global 96. 8 [94. 4 -98. 4] 87. 0 [83. 090. 4] 82. 9 [78. 486. 8] 91. 5 [88. 094. 2] 68. 7 [63. 074. 0] PPV Global 96. 2 [91. 9 -98. 6] 95. 1 [89. 598. 2] 94. 4 [88. 097. 9] 90. 8 [84. 894. 9] 94. 7 [83. 699. 1] NPV Global 97. 3 [94. 0 -99. 0] 83. 2 [77. 787. 7] 78. 4 [72. 783. 4] 91. 9 [87. 395. 2] 64. 5 [58. 270. 4]

Authors were aware of the high prevalence of H. pylori. Conducted subgroup analysis of 76 patients in low prevalence centers for PPV/NPV Test PPV (%) NPV (%) Urea Breath 76. 4 98. 3 Serum 76. 4 98. 3 Stool 83. 3 93. 7 Urine 2 84. 7 75

Results ROC were created. Authors noted that manufacturer threshold for postive/negative results were not ideal. Sensitivity for stool could be increased from 72. 9 to 80. 3 and urine from 63. 2 to 72. 2 with minimal effect on specificity.

Are the results applicable to my patient? Is the reproducibility high? YES: Standard criteria for interpretation of biopsies and manufacturer directions for testing were used. Did comment that degradation of samples in shipment may have occurred when transporting to sites (Germany, Belgium, France).

Are the results applicable to my patient? § If you meet inclusion criteria and no exclusion criteria, results are usually applicable. Both patient 1&2 satisfy. § Due to varied military population, prevalence is difficult to obtain, therefore affecting PPV and NPV.

Patients The choice of tests depends upon issues such as cost, availability, clinical situation, prevalence of infection, pretest probability of infection, and presence of factors (PPI and abx) that may influence test results.

Patient 1&2 • There is a threshold in which a physician will accept a test, ordering no further tests. • When the probability of a target disorder lies between the test and treatment thresholds, further testing is mandated. • Once we decide what our test and treatment thresholds are, post test probabilities have direct implications.

Patient 1 § No clear diagnosis-sxns more consistent with transient dyspepsia. § Uncle lives in California § Lower prevalence in those with his sxns § Pretest probability low § Likelihood ratio/unlikelihood ratio low § H. pylori serum AB with sensitivity of 88. 7 and specificity of 93. 4% and NPV of 98. 3% in group with low prevalence. § Overall post test probability low.

Patient 2 § § § Pt with sxns and hx of recurrent ulcer. Melena and emesis on hx. Blood in stool confirmed on exam. Suggestive family history of ulcer disease. Likely high prevalence in this patient’s family. Neg serum AB, sensitivity of 88. 7%, PPV of 90. 8 %, NPV of 91%. § Urea breath test with sensitivity of 96. 2%, PPV of 96. 2% and NPV of 97. 3% in high prevalence group § Better test in this patient and warranted.