Positive Predictive Value Of Finding Research Articles

Critical Care Randomized Trials Demonstrate Power Failure: A Low Positive Predictive Value of Findings in the Critical Care Research Field.

We aimed to determine the post-hoc power of randomized controlled trials (RCTs) in critical care, and describe the implications for long-term positive (PPV) and negative predictive value (NPV) of statistically significant and non-significant findings respectively in the research field. We reviewed three cohorts of RCTs. "Adult-RCTs" were 216 multicenter RCTs with a mortality outcome from a published systematic review. "Pediatric-RCTs" were 120 RCTs with a mortality outcome, obtained by search of picutrials.net. "Consecutive-RCTs" were 90 recent RCTs obtained by screening publications in 6 journals. Post-hoc power for each study was calculated at α 0.05 and 0.005, for measures of small, medium, and large effect-size, using G*Power software. Long-run expected PPV and NPV of critical care research field findings were then calculated. With α 0.05, post-hoc power for small effect-size was very low in all RCT-cohorts (eg, median 24% in Adult-RCTs). For medium effect-size, post-hoc power was low, except for Adult-RCTs (eg, median 9% in Pediatric-RCTs). For large effect-size, post-hoc power for non-human-animal Consecutive-RCTs was low (median 32%). With α 0.005, post-hoc power was even lower. The corollary was that both PPV and NPV were poor for small effect-size, unless α 0.005 was used. Even with α 0.005, with realistic (vs. optimistic) prior probability of the alternative hypothesis, the PPV was low (eg, in Adult-RCTs 57.1% vs. 92.3%). Adding mild bias (0.1) reduced the PPV even further. For medium effect-size both PPV and NPV were better; nevertheless, with α 0.05 and realistic prior probability of the alternative hypothesis the PPV was poor, and with α 0.005 and mild bias (0.1) the PPV was very low (eg, Adult-RCTs median 44.1%). To improve the predictive value of findings in the critical care research field, RCTs should be designed to have 80% power for realistic effect-size at α 0.005.

Clinicopathological characteristics and aetiological factors of granulomatous gastritis.

To examine the clinicopathological characteristics of granulomatous gastritis (GG) among different aetiologies, particularly Crohn disease (CD), and determine the contribution of Helicobacter pylori and the clinical significance of isolated GG. We identified 269 GG cases overall (0.19% prevalence): 220 had an underlying granulomatous disease (CD, sarcoidosis, tuberculosis) and only eight of these (3.6%) had H. pylori, fewer than the 10.3% rate among non-GG biopsies (P < 0.001). Conversely, among 49 GG cases without known cause (foreign body, undetermined, idiopathic), 13 (26.5%) had H. pylori more than background (P = 0.001). Most patients (n = 185/68.8%) had CD and these were more probably male (P < 0.001), younger (P < 0.001), white (P < 0.001) and had single (P = 0.010), smaller (P = 0.005) and antral (P = 0.027) granulomas amid inflammation (P = 0.005) compared to non-CD GG cases; younger age was independently associated with CD [P = 0.003; odds ratio (OR) = 1.13, 95% confidence interval (CI) = 1.04-1.22]. Among CD patients, younger age (P = 0.003; OR = 1.04, 95% CI = 1.01-1.07) and upper gastrointestinal (GI) symptoms (P = 0.017; OR = 2.53, 95% CI = 1.18-5.43) were associated with new (versus established) diagnosis, whereas multiple gastric granulomas (P = 0.003; OR = 4.67, 95% CI = 1.67-13.04) and lack of upper GI symptoms (P < 0.001; OR = 6.75, 95% CI = 2.94-15.49) were associated with lower GI granulomas. Of 86 isolated GG cases (i.e. no prior diagnosis or lower GI granulomas), 51 (59.3%) were eventually diagnosed with CD, and this was independently associated with younger age (P = 0.014; OR = 1.11, 95% CI = 1.02-1.21) and upper GI symptoms (P = 0.033; OR = 19.27, 95% CI = 1.27-293.31). The positive predictive value of finding isolated GG towards a CD diagnosis in patients aged <30 years was 91%, increasing in males (93%), with single (94%), antral (97%) granulomas or upper GI symptoms (94%). GG does not correlate with H. pylori in patients with granulomatous disease, but may be associated with the organism when such diagnosis is lacking. In CD patients with GG, younger age and upper GI symptoms are associated with a new CD diagnosis, whereas multiple gastric granulomas and lack of upper GI symptoms correlate with lower GI granulomas. GG, including in isolated cases with no prior clinical history or granuloma, probably signifies CD, particularly in younger, male patients or those with single, antral granulomas or upper GI symptoms.

Capsule endoscopy versus magnetic resonance enterography for the detection of small bowel polyps in Peutz–Jeghers syndrome

Our study aimed to assess the diagnostic utility of magnetic resonance enterography (MRE) compared to capsule endoscopy (CE) for the detection of small bowel polyps in patients with Peutz-Jeghers syndrome (PJS); with findings verified by balloon enteroscopy (BE). Adult patients were prospectively recruited across two tertiary centres and underwent MRE followed by CE, with a subsequent BE performed in patients with significant (≥10 mm) polyps. The primary endpoint was the total number of significant (≥10 mm) small bowel polyps detected. The number of patients with at least one significant polyp, correlation with BE findings, and patients' preferences were secondary endpoints. A total of 20 patients (7 male; mean age 34.9 years) underwent both investigations. The number of polyps ≥10 mm detected by CE was greater than by MRE (47 vs 14 polyps, P = 0.02). The number of patients with at least one significant polyp identified by CE was 11 (55 %) compared with 7 (35 %) identified by MRE (P = 0.25). Subsequent BE in 12 patients identified a total of 26 significant polyps in 8 patients. The positive predictive value of finding a polyp at BE was higher for MRE (100 %) compared to CE (60 %). Overall patient preferences identified CE as the preferred modality. This prospective study demonstrated that CE identifies significantly more small bowel polyps compared with MRE in patients with PJS. Correlation between the two techniques and subsequent BE however was relatively poor.

Dermatologic Manifestations as Indicators of Immune Status in HIV/AIDS

A 41-year-old man with human immunodeficiency virus (HIV) infection was admitted with dyspnea on exertion and a nonproductive cough. He was successfully treated for pneumocystis jirovecii pneumonia. Admission physical examination revealed stigmata of four HIV-associated infections demonstrated in the two clinical images: 1) scarring from herpes zoster in the right ophthalmic (V1) distribution of the trigeminal nerve (Panel A, hollow arrow); 2) seborrheic dermatitis in the bilateral nasal folds (Panel A, asterisk); 3) oropharyngeal candidiasis (Panel B, arrow); and 4) oral hairy leukoplakia (OHL; Panel B, arrowhead). Dermatologic manifestations are common among patients with HIV and may provide a clinical indication of underlying immune status. Herpes zoster and seborrheic dermatitis tend to occur early in HIV infection and are associated with relatively high mean CD4 cell counts (380/mm3 and 450/mm3, respectively)1. Oral candidiasis, the most common oral opportunistic infection in patients with HIV/AIDS, often presents at CD4 cell counts less than 300/mm3,2. OHL is relatively specific for HIV infection and is highly predictive of the development of acquired immune deficiency syndrome (AIDS)3. It is associated with a mean CD4 cell count of 143/mm3,4. The presence of both oral candidiasis and OHL together suggest more severe immune dysfunction; these concomitant infections are associated with a mean CD4 cell count of 89/mm3,5. In one study, the joint occurrence of any two HIV-related oral lesions had a mean CD4 cell count of 123/mm3 and a 75% positive predictive value of finding a CD4 cell count less than 200/mm3,4. Our patient had a CD4 cell count of 32/mm3.

A Clinical Composite Score Accurately Detects Meniscal Pathology

Five common tests were used to diagnose meniscal tears when used as a composite score. We evaluated how effectively the composite examination, when performed in the presence of an anterior cruciate ligament (ACL) injury or degenerative joint disease (DJD), determined the presence of meniscal tears. Data were collected prospectively on all patients at our clinic with a primary knee complaint. Independent variables included the presence or absence of the following: (1) a history of "catching" or "locking" as reported by the patient, (2) pain with forced hyperextension, (3) pain with maximum flexion, (4) pain or an audible click with McMurray's maneuver, and (5) joint line tenderness to palpation. Comprehensive patient demographic data were collected including ligamentous examinations and other intra-articular pathologies found at arthroscopy. Composite examination findings were correlated with the presence or absence of meniscal pathology. We evaluated 635 knees in 576 patients for historical and physical findings. Of the knees, 209 underwent arthroscopic surgery and 426 did not. Chi(2) Analysis showed a significant relation between the number of positive diagnostic tests and the presence of meniscal tears (P = .001). Five positive findings on composite examination yielded a positive predictive value of 92.3%. Positive predictive values remained greater than 75% with composite scores of at least 3 in the absence of ACL and DJD pathologies. The presence of an ACL injury decreased the positive predictive value of 5 composite findings to 67%, whereas the presence of DJD increased predictability to 100%. When all 5 symptoms and signs were positive, there was a 92.3% positive predictive value of finding a meniscal tear. Although positive predictive values decreased with a concomitant ACL injury and increased with DJD, there was a higher rate of false-positive findings (ACL) and false-negative findings (DJD). Level II, development of diagnostic criteria with consecutive patients and gold standard.

Transcranial Doppler Sonography in Children: Review of a Seven-year Experience

AIM: We report the use of transcranial Doppler sonography (TCD) examinations in children (mean age 12.7 years; range 0.5–18 years) without an open fontanelle, using TCD equipment with a 2 MHz pulse wave transducer. Our results over a 7-year period are described. MATERIAL AND METHODS: A total of 858 children were referred from the paediatric outpatient department. All patients were divided into three groups according to the reason for referral. RESULTS: We found positive pathological results in the following: (1) headache and orthostatic dysregulation (0.4%, three out of 728); (2) acute neurological symptoms (5.4%, six out of 112); (3) other indications (22.2%, four out of 18). The positive predictive value of finding an abnormality was very low (0.50) . CONCLUSION: TCD examination is ineffective in children with non-specific headache or orthostatic dysregulation due to the specific limitations of this method. We found a higher percentage of abnormal results in children referred for other indications. However, since these children usually undergo MR imaging (and MR angiography, if necessary), in our opinion the TCD examination does not add any additional information. The value of TCD in children is not in the primary diagnosis of disease but in the follow up of known vascular processes (e.g. stenoses) or in chronic diseases including angiitis and sickle cell disease. Hirsch, W. et al. (2002). Clinical Radiology57, 492–497.

Intrahepatic metastatic nodules of hepatocellular carcinoma detected at lipiodol-CT: imaging-pathologic correlation.

The purpose of this study was to determine whether tiny (< 2 cm) areas of iodized-oil retention diagnosed as intrahepatic metastatic nodules of hepatocellular carcinoma (HCC) at computed tomography (CT) following intraarterial injection of Lipiodol (Lipiodol-CT) correlate with truly cancerous foci on the corresponding sectioned pathologic specimens. Thirty-two consecutive patients with biopsy-proven HCC (26 men and six women, aged 41-72 years) underwent prospective evaluation with Lipiodol-CT before undergoing surgery (hepatic resection, 30 patients; liver transplantation, two patients). Imaging findings were correlated with intraoperative and pathologic findings. Twenty-one areas of retained Lipiodol (0.5-1.6 cm in diameter, mean +/- SD: 0.9 +/- 0.3 cm) had features consistent with intrahepatic metastatic nodules of HCC on Lipiodol-CT scans. Nineteen of the 21 areas correlated with tumor deposits at pathologic examination of the surgical specimens, whereas two of 21 were due to abnormal retention of iodized oil within noncancerous liver parenchyma. The positive predictive value of findings at Lipiodol-CT for the diagnosis of intrahepatic metastatic nodule of HCC was 90.5%. Findings at Lipiodol-CT enable a reliable diagnosis of intrahepatic metastatic nodules of HCC.