Looking for Tall P Waves

Sorry for the delayed post. Internship was a bit busy last week, but I will post more regularly.  Contributions also welcome!


Motivation: "Right is height" - this is a phrase often used to describe the EKG changes caused by right atrial enlargement.  But, late last week, while listening to a lecture on EKG and description of P pulmonale, I wondered just how reliable is this sign.  Or, are there other better signs of RA enlargement? For background, right atrial enlargement is classically characterized on the EKG by "P pulmonale" or P wave height greater than 2.5 mm in lead II.

Paper: Evaluation of Electrocardiographic Criteria for Right Atrial Enlargement by Quantitative Two-Dimensional Echocardiography. Kaplan, J. D. et. al. J Am. Coll. Cardiol. (1994); 23:747-52. http://www.ncbi.nlm.nih.gov/pubmed?term=8113560%20

Methods: EKGs of hospitalized patient with mild to severe right atrial enlargement on echo were randomly selected and compared against EKGs of age and gender matched healthy controls.  There were 100 patients with right atrial enlargement and 25 control patients.  EKG were interpreted independently by two cardiologists using calipers and magnifying glasses.

Results:

Cohorts: Of the patients with right atrial enlargement, 52 were in sinus rhythm, 41 were in atrial fibrillation, 5 were in atrial flutter, and 2 in ectopic atrial rhythm.  All controls were in sinus rhythm.  Right atrial enlargement was most commonly associated with tricuspid regurgitation (30%), pulmonary hypertension (28%), and cardiomyopathy (14%).

EKG criteria:

1. P wave height>2.5 mm in lead II (P pulmonale), sensitivity: 6%, specificity: 100%
2. QRS axis > 90 degrees, sensitivity: 34%, specificity: 100%
3. P wave height>1.5 mm in lead V2, sensitivity: 33%, specificity: 100%
4. P wave height>1.5 mm in lead V1, sensitivity: 17%, specificity: 100%
5. R/S ratio > 1 in lead V1 (without RBBB), sensitivity: 24%, specificity: 100%
6. QRS amplitude <6 mm in lead V1, sensitivity: 33%, specificity: 92%

Discussion: The most interesting result of the paper, I thought, was the finding that P pulmonale is very insensitive finding though quite specific.  In fact, with the same specificity, we can look with better sensitivity at the P wave height in V2 to determine right atrial enlargement.  Another interesting finding was that often the most sensitive findings relate more to right ventricular enlargement (like QRS axis > 90 or R/S ratio >1 in V1).  Surprisingly, the specificity of this finding, was quite high meaning that almost all people with right ventricular hypertrophy have right atrial enlargement.  Thus, RV enlargement is a surrogate marker for RA enlargement.

This paper, however, has some important limitations.  First, there were 25 controls to match against 100 patients.  A few more controls would have been more desirable to capture the full range of normal.  A higher number of controls would likely have driven down the "100%" specificity of some of the findings.  Secondly, of the 100 patients with right atrial enlargement, only 52 were in sinus rhythm.  So, for findings relating to P wave morphology, the number of subjects was limited.  Overall, I think that the main points to take away are that P pulmonale has low sensitivity and RV hypertrophy can be used as a surrogate marker for RA enlargement.

The Nitro Cure

Motivation: In my first week of internship, I have seen more urinary tract infections than just about anything else.  The first treatment response for uncomplicated UTI is a three day course of Bactrim, but what happens when the patient either is allergic to Bactrim or has resistant bacteria? The second choice has been ciprofloxacin.  But, flipping open UpToDate the other day, I read to my surprise that nitrofurantoin is the recommended first line agent after Bactrim.  The issue got even more complicated when the venerable Goodman & Gilman's claimed that nitrofurantoin should be a second line agent.  So, what are the data and recommendations?

Paper: Trestioreanu, Z. et. al. Antimicrobial agents for treating uncomplicated urinary tract infection in women. Cochrane Database Syst. Rev. (2010) 10:CD007182. http://www.ncbi.nlm.nih.gov/pubmed/20927755

Methods: A meta-analysis of randomized trials with the objective of comparing the efficacy and safety of different empiric antibacterial treatments for acute, uncomplicated UTI in healthy women aged 16-65 years.  The primary outcomes were short (2 weeks) and long term (8 weeks) symptomatic cures.

Results:
Fluoroquinolone vs Bactrim: Both fluoroquinolones and Bactrim are equally effective as empiric therapy for UTI  in the short (CI: 0.97-1.03) and long term (CI: 0.94-1.05) .  Overall, there is no difference in adverse effects among the two agents, but patients treated with fluoroquinolones are less likely to develop rash (CI: 0.01-0.43).

Nitrofurantoin vs Bactrim: Both nitrofurantoin and Bactrim are equally effective for UTI treatment in the short (CI: 0.95-1.05) and long term (CI: 0.94-1.09).  Overall, no difference in adverse effects for the two agents, but patients treated with nitrofurantoin were less likely to develop rash (0.04-0.76).

The meta-analysis did not find sufficient studies to compare nitrofurantoin to fluroquinolone head-to-head.  The analysis went on to further describe efficacies of beta-lactams (which are not summarized here).

Discussion: The surprising fact of the matter is that there are generally no differences in efficacy when treating UTI empirically with fluoroquinolone, Bactrim, or nitrofurantoin.  The decision to promote Bactrim stems more from a public health perspective since resistance to fluoroquinolones is increasing.  Decreased prescription of fluoroquinolones would presumably decrease generation of resistant bacteria.  In empiric treatment of uncomplicated UTI, nitrofurantoin is an excellent alternative to Bactrim and less likely to generate the adverse effect of rashes.  Overall, both Bactrim and nitrofurantoin have equivalent severity of side-effects and equivalent efficacy.  Of note, however, this meta-analysis examined uncomplicated UTI in unhospitalized patients, and the data might be different in hospitalized patients with Foley catheters in place!

Treating a Hiccup

Motivation: Last Sunday evening, I was crouching before a window in a dim room watching the dying sun and tracing uneasy thoughts about end of summer break and impending residency.  My solemn reverie was suddenly interrupted by high pitched sounds.  My brother had been struck by a bout of hiccups - incessant hiccups.  I called out for him to stop or restrain himself.  He shot back that since I was now a doctor, I could at least tell him how to stop hiccups.  I did not know.  I lost the argument.  Well, how do you treat hiccups?

As background, a hiccup - or the medical term singultus - is a sudden inspiration ending with sudden glottic closure.  There is no known physiological reason for hiccup!  The most common cause of hiccups is gastric distension.  The hiccup reflex arc has afferent nerve pathways of phrenic and vagus nerves, central mediator in the brainstem, and efferent pathways involving the phrenic nerve.  Thus, irritation at any point in the pathway from thoraco-abdominal pathology to CNS neoplasms can give rise to hiccups.  Hiccups lasting more than 48 hours are termed persistent while hiccups lasting longer than a month are called intractable.  Traditional pharmacotherapy of hiccups involves the antipsychotic chlorpromazine though the literature is thin on evidence.  But, given the side effects of chlorpromazine, alternative therapy is needed.  Recently, gabapentin was tried.

Paper: Porzio, G. et. al. Gabapentin in the Treatment of Hiccups in Patients With Advanced Cancer: A 5-Year Experience.  Clinical Neuropharmacology (2010) 33: 179-180.  http://www.ncbi.nlm.nih.gov/pubmed/20414106

Methods: This study was conducted in Italy on patients with advanced cancer in two settings: (1) a palliative in-patient care unit and (2) home comfort care service.  Patients were assessed for presence of persistent (>48 hours) hiccups that were rated at least 7 out of 10 in the patient's subjective assessment of severity of hiccups (10 means worst hiccups of life).  These patients were treated with gabapentin (300 mg thrice daily) with titration according to response.

Results:
Patients:  In the palliative in-hospital service, 37 of 944 patients (3.9%) had severe hiccups.  In the home care setting, 6 of 134 patients (4.5%) had severe hiccups.  Most patients (31/43) had advanced abdominal cancers.  The rest had advanced cancers from other locations.

Gabapentin result: After gabapentin administration, of the 37 inpatients, 31 experienced complete resolution while 4 had improvements.  Two end-stage patients on midazolam therapy experienced worsening of hiccups.  Among the 6 home care patients, 4 had complete resolution while the other 2 had improvement.  The maximum used dose of gabapentin was 1200 mg/day.  The most common side-effect was transient drowsiness.

Discussion: The paper in its design and power has many limitations, but it does demonstrate some remarkable results.  Among the 37 in-patients with severe hiccups, 35 (>90%) had improvements in hiccups, and the majority experienced complete resolution!  Another benefit of gabapentin over current therapy with chlorpromazine is that the only major side effect of gabapentin (in this trial) is transient drowsiness.  The paper, of course, has many limitations.  There was no placebo arm to the intervention, and the trial was unblinded.  So, the observed effect could be entirely a placebo response or an observer bias.

Despite the fact that about 4% of patients with advanced cancers are afflicted with severe hiccups decreasing quality of life, there have not been large randomized trials conducted for treatment of hiccups.  The future for hiccups trials does not appear too bright either.  This prospective design may be the level of the quality of data that is going to be available for some time.  So, if I had to treat hiccups, I would probably try gabapentin first given the tolerable side-effect profile of the drug.

To Salt and Death

Motivation: Last week, I eavesdropped on the subway.  Two old men were joking about their health when one man remembered recently hearing on the news that eating less salt kills you.  The other guy did not believe him. I did not believe him.  But, after flipping out my phone, a quick search of Google News revealed many recent articles stating exactly that.  Also, the articles cited a May issue of the authoritative journal JAMA.  We know that eating more salt results in increased blood pressure in the short term but does eating less salt really kill you in the long run?

Paper: Fatal and Nonfatal Outcomes, Incidence of Hypertension, and Blood Pressure Changes in Relation to Urinary Sodium Excretion.  Stolarz-Skrzypek et. al. JAMA (2011) 305: 1777-1785.  http://jama.ama-assn.org/content/305/17/1777.abstract

Methods: In this prospective study, European patients older than age 20 were recruited who did not have diagnosed cardiovascular disease at baseline.  On initial exam, a 24-hour urinary sample was collected for sodium excretion measurement (to reflect total sodium consumption).  Baseline measurements were obtained in 3681 patients, who were followed for mortality and morbidity.  A smaller subset of 2856 patients agreed to undergo follow-up exams.  Of these patients, a cohort of 2096 patients were followed for development of hypertension (760 already had hypertension at baseline).  A third cohort of 1499 patients were followed with serial urinary sodium measurements to establish a relationship between blood pressure and urinary sodium excretion - these patients were untreated with any antihypertensive medications.

Results:
Patient selection: Overall, the cohort had 52.7% women, and all participants were white.  The mean age was 40.9 years with smoking rate of 28.4% and alcohol intake (>5g/day) rate of 24.1% .  At baseline, average BP was 124.7/76.3 and total cholesterol was 209 mg/dL.

Salt and Death: The authors divided the patients into three groups based on urinary sodium excretion: low sodium excretion (1220 patients), medium (1250), and high (1211).  In the cohort, 219 patients died.  After adjusting for other variables like age, sex, smoking, drinking, diabetes, cholesterol, and educational attainment, the hazard ratios for all mortality in the three groups (from low to high) were: 1.14, 0.94, and 1.06 (p = 0.10).  For cardiovascular deaths, there were 84 events.  The adjusted hazard ratio in the three groups from low sodium to high sodium were: 1.56 (CI: 1.02-2.36), 1.05 (0.72-1.53), and 0.95 (0.66-1.38), p = 0.02.  The trend for non-cardiovascular deaths were not significant.

Cardiovascular Events:  The rates of coronary events, strokes, and all cardiovascular events across the three sodium groups were not significantly different.

Incidence of Hypertension: The baseline levels of urinary sodium levels were not predictive of incidence of hypertension.  Incidence of hypertension was about 25% in all groups.

Sodium to Blood Pressure Levels: Systolic blood pressure was positively correlated with 24 hour urinary sodium excretion.  A 100 mEq/L increase in sodium excretion resulted in only 1.14 mm Hg increase in systolic blood pressure (!).  Diastolic pressure was unaffected by sodium excretion levels.

Discussion: In summary, I think that this paper is an excellent example of data that ought to be consumed with caution and probably not fit for broad sweeping popular media coverage.  This study has several important limitations.  First, the study design is prospective meaning that while we may see associations between urinary sodium levels (and presumably sodium consumption) and medical endpoints, it would be hard to infer causality.  For example, why do some people consume less sodium?  Other than trying to be healthy, the people eating less sodium may have been poorer or may have other habits like strenuous jobs that prevented them from eating.  The study did not really adjust for socioeconomic status.  Also, the study population was relatively young (average age 40.9 years).  In a younger cohort, other factors like genetics may be more important than dietary habits.  Finally, many patients (259) were lost to follow-up.

The most provocative point in the paper is the weak association shown between decreased sodium consumption and increased cardiovascular mortality.  The association is weak with a few events (10 events in the high sodium excretion group).  While the point is interesting, I think that it would be a stretch to say that decreased sodium consumption increases mortality.  There are many confounding variables that could affect the result, and this study probably calls for a randomized intervention trial.  Finally, the other point that I found interesting is that the incidence of hypertension did not differ among groups with varying sodium consumption.  This may suggest that while sodium consumption affects blood pressure, primary hypertension may not be caused by sodium indiscretion.

PS:  In case you are still reading, there have been some items in the news that were reviewed in this blog.  1) Fidaxomicin was recently approved by FDA as a new drug for treating C. difficile.  The drug was reviewed by Haoming here.  2) There has been a lot of buzz recently about cell phones and cancer.  The topic was reviewed in this blog here by me last year.

NSF - how real?

Motivation: A few months ago, I was lounging in the neurology office when someone called to ask about the possibility of brain MRI with contrast in a patient with kidney disease.  The resident answered, "Of course, not!"  After all, gadolinium based agents have been implicated in nephrogenic systemic fibrosis (NSF).  When the resident placed the phone on the cradle, a fellow in the room smirked and remarked that our fear of NSF is overblown.  He explained that cases of NSF occur rarely and that fear of litigation probably drove our prohibition.  Well, just how common is nephrogenic systemic fibrosis after gadolinium contrast exposure?

For background, nephrogenic systemic fibrosis was first described in 2000 and is clinically characterized by thickening and hardening of skin with "brawny" hyperpigmentation especially of the extremities.  Patients experience neuropathic pain in affected areas and suffer from flexion contractures and pressure ulcers.

Paper: Association of Gadolinium Based Magnetic Resonance Imaging Contrast Agents and Nephrogenic Systemic Fibrosis.  Bhave, G. et. al. The Journal of Urology (2008) 180: 830-835. http://www.jurology.com/article/S0022-5347(08)01225-1/abstract

Results: 
Number of known cases: The largest case registry for NSF (Yale NSF case registry) contains about 200 cases. Adding all other reported cases, at the end of 2006, there were at most 400-500 cases of NSF worldwide.

Relation to Kidney Disease: No case of NSF has been described for GFR greater than 30 mL/min.  Cases which seemingly involved patients with GFR > 30 mL/min are likely due to calculation error since in acute kidney failure, serum creatinine levels do not instantaneously reach steady state and give rise to errors in GFR estimation.

Link with Gadolinium Exposure: More than 95% of known cases of NSF have had gadolinium exposure in the weeks to months prior to presentation.  A very small minority of putative NSF cases have not had gadolinium exposure.  The upper limits of lag times from gadolinium exposure to NSF are about 1-2 years.

Incidence after Gadolinium Exposure: In single center cohorts, the proportion of patients with kidney failure developing NSF after gadolinium exposure is about 2-5%.  The authors challenge this assumption based on a simple calculation.  Before 2000, there were no restrictions on gadolinium administration.  In the ten prior years, there were 50 million MRI contrast studies, and given the 0.1% prevalence of end stage renal disesaes (ESRD) in the U.S., there have been conservatively about 50,000 MRI contrast studies performed in patients on ESRD.  Given that there are at most 500 known cases of NSF, the incidence is likely no more than 1% after gadolinium exposure in patients in ESRD.


Discussion: In summary, NSF is a rare but real danger.  What I thought was an especially valuable point to keep in mind is that in cases of acute kidney failure, the serum creatinine level is an imperfect measure, and the working assumption ought to be perhaps that gadolinium is contraindicated despite seemingly permissive creatinine levels.  On the flip side, radiologists are hesitant to administer gadolinium in patients with GFR<60 mL/min but greater than 30 mL/min.  There have been no reported cases of NSF in this group, and this cautious approach may be excessive and needs to be reevaluated. 


Finally, is there any way to prevent NSF? From a pathophysiology viewpoint, gadolinium ion is toxic to human beings, and contrast agents consist of gadolinium chelated to other molecules to facilitate excretion and limit exposure of tissue to gadolinium ions.  In renal failure, with prolonged exposure, gadolinium is lost from chelation and produces toxic effects.  In the future, the thought is that with stronger chelating molecules, we may be able to further limit gadolinium loss from chelation and prevent toxic effects before excretion through dialysis or urination.

New Weapon in the Fight Against C.Diff

Introduction

Throughout my medical school career, there are many things, both good and bad which I will never forget, among them is the pungent odor of a C.difficile infection. Its an unmistakable stink which brings back memories of awkwardly trying to write a note or perform a physical exam while wearing one of those big yellow isolation gowns. C. diff's unshakable presence in our hospitals is only expected to get worse as our careless over-use of antibiotics creates more opportunities for c.diff to out compete the normal colonic flora.

Traditionally, C.diff infections have been treated with Flagyl firstline and Vancomycin for those who failed Flagyl or have fulminant disease. However, due to increasing resistance to Flagyl, Vancomycin is rapidly becoming the only effective drug for many patients. However, even Vancomycin sometimes cannot eliminate this nasty critter causing patients to relapse almost as soon as they leave the hospital. The result is chronic, uncontrollable malodorous diarrhea that can destroy a patient's quality of life and create nightmares for hospital infection control and cost billions of dollars for our health care system.

However, hope is on the horizon with a new antibiotic -Fidaxomicin, which has recently completed phase 3 trials. Fidaxomicin is a macrolide antibiotic produced by Optimer Pharma which has shown higher in vitro activity against c.diff than vancomycin. In an article published in February of 2011 in the NEJM, Louie et al. present results from a phase III trial involving 629 patients.

Louie TJ, Miller MA, Mullane KM, Weiss K, Lentnek A, Golan Y, Gorbach S, Sears P, Shue YK; OPT-80-003 Clinical Study Group. Fidaxomicin versus vancomycin for Clostridium difficile infection. N Engl J Med. 2011 Feb 3;364(5):422-31.Methods: Patients were selected whom had diarrhea for 24 hours and either c.diff toxin A or B.

Patients could have received up to 4 doses of Flagyl or Vanc but no other therapy in the 24 hrs before randomization. Patients with fulminant c.diff or >1 recurrence of c.diff in the past 3 month were excluded. Patients were randomized to a 10 day course of either 200mg of Fidaxomicin q12h or 125mg of Vanc q6h. Patients were stratified according to whether this is their first or second episode of c.diff. All patients were followed for at least 28 days following the end of treatment.

Outcomes measured were:

1. clinical cure defined as cessation of diarrhea and no need for additional treatment.

2. Treatment failure was defined as continuation of diarrhea or need for additional therapy.

3. Global cure was defined as no recurrence within the 28 day follow up after end of therapy.

Patients were analyzed as modified intention to treat if they took at least one dose of medication.

Results:

A total of 596 patients were analyzed in the modified intention to treat analysis of which 92% actually adhered to the protocol. They did not differ in baseline characteristics.

Figure 1. mITT=modified intention to treat, PP=per protocol.

As shown in figure 1. Fidaxomicin was as equally efficacious as vancomycin in achieving clinical cure (resolution of diarrhea, no need for additional Rx). However, it achieved a 40% relative reduction in recurrence compared to vancomycin (absolute rate reduction of recurrence from 25% to 15%) (p=0.005).

s shown in Figure 2, in subgroup analysis, Fidaxomicin was superior to or as equally efficacious as Vancomycin at controlling C.diff recurrence in almost every subgroup save for 1. This subgroup comprised of patients with a certain genotype of c.diff known as NAP1/B1/027. For this group, fewer patients had relapse of c.diff after vancomycin treatment (21%) compared to fidaxomicin (27%) however this difference was not statistically significant.

Figure 2.

Safety wise, the rate of adverse and serious adverse reaction was similar in patients treated with Fidaxomicin or Vancomycin.

Discussion:

C.diff is a particularly difficult disease to treat (hence the name) because it is resistant to so many different antibiotics and because even successfully treated patients recur as much as 30% of the time within 60 days. More worryingly, strains of c.diff are becoming resistant to our two current standards of care: Vanc and Flagyl.

Louie et al. show in this study that Fidaxomicin is as efficacious and safe as vancomycin in treating acute infections of c.diff and perhaps more effective at eliminating recurrences. In addition, Fidaxomicin was previously shown to have poor system absorption through the GI tract, a positive attribute because it increases the concentration of the drug which actually reaches the colon after oral administration and also decreases systemic side effects.

The only wrinkle in all of this good news is that the 36% of patients with an aggressive strain of c.diff BI/NAP1/027 do not seem to obtain any additional reduction in recurrence with Fidaxomicin as compared to vanc. This is worrying because patients with this strain of c.diff is the most in need of a better treatment option.

Overall however, fidaxomicin represents an important addition to our current inventory of Flagyl and Vancomycin for the treatment of c.diff.

Diagnosing PE by Blood Gas

Motivation: On an early morning at 5:30 am last winter, I was huddling with my surgery team waiting for our chief to arrive and start rounding.  In the haze of early morning stupor, I heard one of the residents mention that overnight someone was short of breath.  He had called one of his fellow surgical residents, who had advised him to draw a blood gas to rule out pulmonary embolism (PE).  One of the other members of the team was about to say something when the chief arrived, and we cut short our talk to begin marching to the patient rooms.  But, I wondered how sensitive is an ABG with its characteristic signs of hypoxemia and hypocapnia in ruling out PE?

Paper: Diagnostic Value of Arterial Blood Gas Measurement in Suspected Pulmonary Embolism.  Rodger, M., et. al. Am. J. Respir. Crit. Care Med. (2000) 162: 2105-2108.  http://ajrccm.atsjournals.org/cgi/content/full/162/6/2105

Methods: For a 30 month period, consecutive patients (inpatient and outpatient) suspected of PE were subjected to a blood gas and D-dimer test.  Referring clinicians were then asked to provide a clinical gestalt of the likelihood of PE.  Subsequently, all patients underwent V/Q scanning.  For low probability V/Q scan with high clinical probability of PE or high probability V/Q scan with low clinical probability, patients were referred for pulmonary angiogram at the clinician's discretion.  Inclusion criteria for patients were essentially greater than 18 years of age with capacity for informed consent and physiological capability to undergo pulmonary angiogram if necessary.

Results: 
Patient Selection: In total, 246 patients were considered for PE.  In 49 patients, PE was accepted as the final diagnosis and ruled out in 163 other patients.  In 34 patients, the final diagnosis was unclassified.  These patients were not referred for angiogram by their clinicians but had either low probability V/Q scan with high pre-test likelihood or high probability V/Q scan with low pre-test likelihood.  The patients deemed unclassified were removed from further analysis.

Blood Gas Measurements: The clinical variable PaO2<80 mm Hg was present in 57.9% of patients with PE and 46.6% of patients without PE.  The clinical variable PaCO2<36 mm Hg was present in 44.4% of patients with PE and 39.7% of patients without PE.  Other clinical rule results are as follows:

• Abnormal (A-a) gradient- Sensitivity: 84.2%, Specificity: 27.4%, PPV: 27.4%, NPV: 84.2%
• D-dimer positive - Sensitivity: 83.0%, Specificity: 57.6%, PPV: 39%, NPV: 91.2%
• PaO2<80 mm Hg or D-dimer positive - Sensitivity: 91.9%, Specificity: 32.4%, PPV: 32.4%, NPV: 91.9%
• PaO2<80 mm Hg or D-dimer positive or Respiratory Rate>20 breaths/min - Sensitivity: 96.9%, Specificity: 21.3%, PPV: 30.7%, NPV: 95.0%
• PaCO2<36 mm Hg or Abnormal (A-a) O2 gradient - Sensitivity: 91.9%, Specificity: 14.7%, PPV: 25.6%, NPV: 85.0%

Discussion: To rule out PE, we ideally want clinical criteria with high sensitivity.  Individually, hypoxemia and hypocapnia have really poor sensitivities at 57.9% and 44.4%!  Calculating an abnormal alveolar-arterial gradient is a more sensitive indicator for PE, but the negative predictive value (84.2%) is still unacceptably low.  While evaluating the results, I was surprised to see that a D-dimer test was no more sensitive than an abnormal A-a gradient.  The D-dimer has the advantage of specificity rather than sensitivity over the A-a gradient.  If I were stuck on an island without radiology, I would choose the rule of PaO2<80 or D-dimer positive criteria to start ruling out PE though we would miss about 8% of pulmonary embolism cases.  With an estimated untreated mortality of 30%, such an approach would still be dangerous.  This study shows that at present, radiology remains an integral part of ruling out PE.

This study has some limitations.  Of the 246 patients originally included in the study, the diagnosis remained indeterminate in 34 patients.  Often these patients are the most troubling because they have a mismatch in the clinical probability and probability provided by V/Q scan.  A very interesting question is whether blood gas results could be used to further stratify risk status in these patients with intermediate risk status.  Pulmonary angiogram studies or at least further clinical follow-up history would be useful.