Thursday, October 22, 2009
Friday, September 18, 2009
Competing prediction rules for DVT
Safely ruling out deep venous thrombosis in primary care. Ann Intern Med 150 (4):229-35. PMID: 19221374
- The original Wells prediction rule was published in 1995 with 12 predictor variables (PMID: 7752753).
- Wells modified his rule in 1997 to use only 9 variables (PMID: 9428249).
- In 2003, Wells revised a 10 point rule that adds a point for history of previous DVT (PMID: 14507948.
- In 2005, a study in the Annals of Internal Medicine was titled "Wells Rule Does Not Adequately Rule Out Deep Venous Thrombosis in Primary Care Patients." However, this study was started in 2002 and so studied the 1997 version of the Wells rule (PMID: 16027451).
- Now the authors of the 2005 study have published the performance of their own rule (PMID: 19221374), named the AMUSE rule for "Amsterdam Maastricht Utrecht Study on thromboEmbolism."
We are left with the 2003 Wells rule and the AMUSE rule without a direct comparison. An insightful editorial accompanying the current publication by AMUSE notes that the two rules have similar negative likelihood ratios - actually the Wells rule has a slightly better LR- and has been externally validated.
Regarding the physical examination, both rules include the following findings:
- Distended, non-varicose collateral veins
- Calf circumference discrepancy greater than 3 cm measured 10 cm below the tibial tuberosity
The Wells rule also includes:
- Pitting edema (confined to symptomatic leg)
- Swelling of entire leg
- Localized pain along distribution of deep venous system
In addition to the physical findings, the Wells rule includes the sometimes subjective finding, "Alternative diagnosis at least as likely."
In summary, I appreciate the AMUSE group's efforts at creating a less subjective prediction rule for DVT and look forward to a direct comparison of the current versions of both rules. Until then, I will stay with the Wells rule.
This has been added to http://en.citizendium.org/wiki/Deep_venous_thrombosis.
Citation:
Büller HR, Ten Cate-Hoek AJ, Hoes AW, Joore MA, Moons KG, Oudega R, Prins MH, Stoffers HE, Toll DB, van der Velde EF, van Weert HC, & AMUSE (Amsterdam Maastricht Utrecht Study on thromboEmbolism) Investigators (2009). Safely ruling out deep venous thrombosis in primary care. Annals of internal medicine, 150 (4), 229-35 PMID: 19221374
Saturday, September 5, 2009
Postural hypotension in patients with syncope
Arch Intern Med 2009 PMID: 19636031
The authors report that in 2106 consecutive patients 65 years or older admitted for syncope, "Postural blood pressure (BP) recording, performed in only 38% of episodes, had the highest yield with respect to affecting diagnosis (18%-26%) or management (25%-30%) and determining etiology of the syncopal episode (15%-21%)."
- The lower percentages are based on 'strict criteria' for abnormal changes:
- drop in systolic BP of at least 20 mm Hg
- or
- drop in diastolic BP of at least 10 mm Hg
- The higher percentages are based on 'loose criteria' for abnormal changes:
- drop in systolic or diastolic BP of at least 10 mm Hg
- or
- systolic BP drop to 90 mm Hg or lower
- Pulse increase:11 (95CI: 9-13)
- Systolic blood pressure drop: 4 (95CI: 2 - 6)
- Diastolic blood pressure drop: 5 (95CI:3 - 8)
This has been added to http://en.citizendium.org/wiki/Syncope and http://wiki.medpedia.com/Clinical:Syncope.
Citation:
Mendu ML, McAvay G, Lampert R, Stoehr J, & Tinetti ME (2009). Yield of diagnostic tests in evaluating syncopal episodes in older patients. Archives of Internal Medicine, 169 (14), 1299-305 PMID: 19636031
Sunday, June 14, 2009
Diagnosing irritable bowel
JAMA. 2009 PMID: 18854541
This comprehensive systematic review by the Rational Clinical Examination is very helpful after a few adjustments. First, the review allows source studies to place patients with symptoms of irritable bowel who are found to have diverticulosis or polyps into the category of underlying organic illness. Patients with diverticulosis who have symptoms of irritable bowel probably have irritable bowel syndrome.(PMID: 3717113) Likewise, polyps seem very unlikely to cause symptoms of irritable bowel and these patients also probably irritable bowel syndrome and coincidental polyps. Now that the USPSTF recommends screening for polyps starting at age 50, the presence of polyps among patients with irritable bowel syndrome is less important.(PMID: 18838716)
The review cites the study of Bellentani(PMID: 2289644) to conclude that 60% of patients in primary care with symptoms of irritable bowel have irritable bowel syndrome. However, if you group the patients with polyps or diverticulosis with the patients with irritable bowel, the prevalence becomes 87%.
Likelihood ratio + | Likelihood ratio - | |
---|---|---|
History alone (Manning criteria) | 2.9 | 0.29 |
History and physical examination (Rome criteria) | 4.8 | 0.34 |
History, physical examination, and laboratory tests (Kruis score) | 8.6 | 0.26 |
Thus, the Kruis score seems good enough to diagnose irritable bowel among patients in primary care (remember that patients over age 50 probably need endoscopy to screen for polyps). The composition of the Kruis score is:
Finding | Score |
---|---|
Abdominal pain or flatulence or bowel irregularity | 34 |
Duration of symptoms >2 y | 16 |
Abdominal pain is "burning, cutting, very strong, terrible, feeling of pressure, dull, boring, not so bad" | 23 |
Alternating constipation and diarrhea | 14 |
History of blood in stool | -98 |
Physical examination or history pathognomonic for an alternative diagnosis | -47 |
ESR > 10 mm/hr | -13 |
WBC > 10k | -50 |
Hemoglobin < 12 g/dL for females or < 14 g/dL for males | -98 |
This has been added to http://en.citizendium.org/wiki/Irritable_bowel_syndrome
Ford, A., Talley, N., Veldhuyzen van Zanten, S., Vakil, N., Simel, D., & Moayyedi, P. (2008). Will the History and Physical Examination Help Establish That Irritable Bowel Syndrome Is Causing This Patient's Lower Gastrointestinal Tract Symptoms? JAMA: The Journal of the American Medical Association, 300 (15), 1793-1805 DOI: 10.1001/jama.300.15.1793
Saturday, June 13, 2009
Frequency of Acute Coronary Syndrome in Patients with Normal Electrocardiogram Performed during Presence or Absence of Chest Pain
The authors detail the findings of 387 consecutive patients with normal electrocardiograms admitted for a chief complaint of chest pain. The authors report that 17% (67/387) of patients had acute coronary syndrome ACS). However, the authors define ACS as:
- Unstable angina. Either:
- 70% stenosis (38 patients)
- positive stress test (1 patient)
- NSTEMI.
- Positive troponin (28 patients)
- Among 261 patients with electrocardiogram taken during pain, 18 (7%) had NSTEMI.
- Among 126 patients with electrocardiogram not taken during pain, 10 (8%) had NSTEMI.
This study independently confirms the findings of an earlier study (PMID 16973638) that the presence of chest pain during a normal electrocardiogram does not adequately exclude NSTEMI among a group of patient that physicians chose to admit the hospital and had a 7% prevalence of NSTEMI.
This does not mean the electrocardiogram cannot help exclude acute coronary syndrome, but means that whether the electrocardiogram is taken during pain is not important.
This does not mean the electrocardiogram cannot exclude acute coronary syndrome in patients at lower risk such as those with unusual pain and no history of ischemic heart disease (PMID 3970650).
This has been added to http://en.citizendium.org/wiki/Acute_coronary_syndrome#Electrocardiogram
Citation:
Turnipseed, S., Trythall, W., Diercks, D., Laurin, E., Kirk, J., Smith, D., Main, D., & Amsterdam, E. (2009). Frequency of Acute Coronary Syndrome in Patients with Normal Electrocardiogram Performed during Presence or Absence of Chest Pain Academic Emergency Medicine, 16 (6), 495-499 DOI: 10.1111/j.1553-2712.2009.00420.x
Thursday, May 28, 2009
History and physical examination for COPD
Fam Pract. 2009 PMID: 19423699
I do not find this review helpful. This review cautiously concludes "There is insufficient evidence to assess the value of history taking and physical examination for diagnosing COPD". This contrasts with the careful Rational Clinical Examination Project that provides a bottom line with 5 bulleted points from their interpretation of the literature. Juxtaposing these two studies demonstrates how arbitrary evidence-based medicine can be in its acceptance of evidence (Goodman, PMID: 12204023).
Part of the problem is that the authors deleted two of the studies that I thought were best - Holleman (PMID 8441077) and my own (PMID 8430714).
I believe this review particularly underplays the role of auscultating diminished breath sounds. Diminished breath sounds were the most sensitive finding in our study and in Holleman's study. While breath sounds were not statistically significant in Holleman's study, they were only auscultated in the last third of the study and so they had less observations to generate statistical power. Diminished breath sounds have been significant findings in other studies (van Schayck et al, PMID 1792447; Hepper et al, PMID 5351681; Schneider, PMID 14263096; Bohadana, PMID 684671; Melbye, PMID 9656782; Pardee, PMID 7357938).
I encourage you to read the Rational Clinical Examination for this topic instead.
Citation:
Broekhuizen BD, Sachs AP, Oostvogels R, Hoes AW, Verheij TJ, Moons KG. (2009). The diagnostic value of history and physical examination for COPD in suspected or known cases: a systematic review. Family Practice
Thursday, April 2, 2009
New bedside finding for carpal tunnel syndrome
New bedside finding for carpal tunnel syndrome. J Hand Surg [Am]. 2008 Nov;33(9):1518-24. PMID: 18984333
At most, this is a new finding to watch for in future research. This study has problems; details are below.
The authors report that in 119 patients with clinically diagnosed carpal tunnel syndrome seen by a single surgeon and 109 controls recruited from a church, the Scratch Collapse test was able to distinguish carpal tunnel syndrome with accuracy of:
Problems with this study are:
At most, this is a new finding to watch for in future research. This study has problems; details are below.
The authors report that in 119 patients with clinically diagnosed carpal tunnel syndrome seen by a single surgeon and 109 controls recruited from a church, the Scratch Collapse test was able to distinguish carpal tunnel syndrome with accuracy of:
- Sensitivity 64%
- Specificity 99%
Problems with this study are:
- Use of case control design rather than a cross sectional study of consecutive patients
- No true reference standard
- Subjective interpretation of an abnormal response
Sunday, February 15, 2009
Measuring the ankle-brachial index using a stethoscope
Can we measure the ankle-brachial index using only a stethoscope? A pilot study. Fam Pract. 2009 Feb;26(1):22-6. Epub 2008 Nov 20. PMID: 19022870
Patients:
This has been added to http://en.citizendium.org/wiki/Ankle_brachial_index
The ankle-brachial index (ABI) by ausculation is less sensitive than an earlier report of the ABI by palpation. Both methods are less sensitive than the ABI by Doppler ultrasonography.
Patients:
- Cross sectional study of 81 consecutive patients referred for arterograms of any region of the body
- Ankle-brachial index by auscultation
- ABI Doppler ultrasonography less than 0.9
- Abnormal in 17% of subjects.
- Sensitivity 71%
- Specificity 91%
- Positive predictive value 62%
- Negative predictive value 94%
- Click here to adjust the pretest probability and recalculate predictive values
This has been added to http://en.citizendium.org/wiki/Ankle_brachial_index
Thursday, February 12, 2009
Acute, traumatic knee effusions
Diagnostic value of history taking and physical examination to assess effusion of the knee in traumatic knee patients in general practice. Arch Phys Med Rehabil. 2009 Jan;90(1):82-6. PMID: 19154833
In this study of patients with traumatic knee injuries of less than 5 weeks duration, the prevalence of effusion was 31% by magnetic resonance imaging. Almost three quarters of patients with an effusion had meniscal or ligament tear.
The accuracy of ballottement in detecting effusion was disappointing: sensitivity and specificity 83% and 49%, respectively. The accuracy of self-reported swelling was similar, with sensitivity and specificity 80% and 45%, respectively.
The negative predictive value of ballottement was 86% for patients in this study with a pretest probability of 31% (click here to adjust the pretest probability).
This study has been added to http://en.citizendium.org/wiki/Knee_injury.
In this study of patients with traumatic knee injuries of less than 5 weeks duration, the prevalence of effusion was 31% by magnetic resonance imaging. Almost three quarters of patients with an effusion had meniscal or ligament tear.
The accuracy of ballottement in detecting effusion was disappointing: sensitivity and specificity 83% and 49%, respectively. The accuracy of self-reported swelling was similar, with sensitivity and specificity 80% and 45%, respectively.
The negative predictive value of ballottement was 86% for patients in this study with a pretest probability of 31% (click here to adjust the pretest probability).
This study has been added to http://en.citizendium.org/wiki/Knee_injury.
Thursday, February 5, 2009
Detecting LVEDP elevation
Diagnostic accuracy of Doppler echocardiography for determining left ventricular diastolic pressure elevation: prospective comparison to chest radiography, serum B-type natriuretic peptide, and chest auscultation. Echocardiography. 2008 Oct;25(9):946-54. PMID: 18771556
In a small study of 56 patients, including 19 with a LVEDP > 15 mm Hg (22 ± 4) by left heart catheterization, no patients had rales, 42% had radiographic pulmonary vascular redistribution, and only 11% had BNP > 100 pg/dL.
These results report even less sensitivity for the chest radiograph than we found in our earlier meta-analysis (http://pubmed.gov/894569). These results are also worse than the original large study of the BNP that found a cutoff of > 50 pg/dL gives high sensitivity (http://pubmed.gov/12124404).
It is unclear what led these patients to their catheterization, but the low sensitivities suggest these patients may have had asymptomatic, chronically elevated LVEDP.
In a small study of 56 patients, including 19 with a LVEDP > 15 mm Hg (22 ± 4) by left heart catheterization, no patients had rales, 42% had radiographic pulmonary vascular redistribution, and only 11% had BNP > 100 pg/dL.
These results report even less sensitivity for the chest radiograph than we found in our earlier meta-analysis (http://pubmed.gov/894569). These results are also worse than the original large study of the BNP that found a cutoff of > 50 pg/dL gives high sensitivity (http://pubmed.gov/12124404).
It is unclear what led these patients to their catheterization, but the low sensitivities suggest these patients may have had asymptomatic, chronically elevated LVEDP.
Subscribe to:
Posts (Atom)