This 2008 study was the second large validation study of the Pulmonary Embolism Rule Out Criteria (PERC) rule. The researchers aimed to test the hypothesis that follows: patients who presented with signs and symptoms of PE but were estimated to be low risk by clinician gestalt (<15% pretest probability), in addition to being PERC negative, would have a <2% risk of VTE. Less than 2% was determined to be the point of equipose, where the dangers of workup for PE in this group (radiation exposure, incidentaloma findings) would be greater than the risk of actually finding a clinically significant PE.
Methodology and Design:
12 Large Emergency Departments in the US and 1 in New Zealand
8,138 Patients enrolled
Prospective cohort study
Patients were enrolled when a diagnostic test for PE was ordered in the ED (D-Dimer, CTA, V/Q Scan, CT Arteriography)
At time of diagnostic test ordering, clinician was asked to fill out a survey that included categorizing the patient as low (<15%), medium (15-40%), and high (>40%) risk of PE by gestalt. They also had to fill out the PERC rule.
Primary outcome –> did the patient have VTE at 45 day follow up? Did the patient die?
Patients were excluded if they had features deeming them at risk to be lost to follow up (serious mental illness, homeless, etc)
8,138 patients –> 513 found to have venous thromboembolism.
One death from PE (56 deaths not from VTE)
Low, medium, and high risk patients by gestalt had VTE at the following rates respectively: 3.5%, 10.4%, and 31.1%
PERC negative had a “sensitivity” (as in, no VTE was a positive result for a negative PERC) of 95.7%.
Patients with who were PERC negative AND deemed low risk by clinician result (aka very low risk) had VTE ~1%
PERC is a reliable tool in ruling out PE for patients deemed low risk by gestalt
Further diagnostic testing in these patients is more likely to be harmful than beneficial
There is still too much diagnostic testing being performed for low risk PE (1,843 normal CTPEs performed in this study alone! 362 of these were VERY LOW RISK patients)
Limitations and Discussion:
Generalization for our patient population at Temple (excluded any patients with potential loss for followup)
Did not discuss how many of these VTEs were clinically significant and complications of treating these VTEs. (heparin is not benign!)
What do we do with those who are low risk but PERC+? What about medium risk? Is there a role for ultrasound in ruling these out? Further work needs to be done.
This was a retrospective analysis using the population-based data of 6 US states from the State Emergency Department Databases and State Inpatient Databases from 2006 through 2012 to identify adults discharged from the ED with a diagnosis of atraumatic headache (which included tension headache, migraine, headache, acute headache, head pain) or back pain (which included lumbar disc herniation, nonspecific disc herniation, back pain including thoracic pain, lumbar back pain or sciatic, thoracic strain, lumbar strain, back strain). There was also a control group of patients with non-neurologic chief complaints. The primary outcome was a composite of return ED visit and hospitalization for primary diagnosis of a serious neurologic condition or inhospital death within 30 days of ED discharge.
They identified 2,101,081 ED patients discharged with a nonspecific diagnosis of headache and 1,381,614 discharged with a nonspecific diagnosis of back pain who were at risk for serious neurologic conditions. Overall, for patients with a nonspecific diagnosis of headache, 0.5% had a serious neurologic condition or inhospital death within 30 days of ED discharge. For patients with a nonspecific diagnosis of back pain, 0.2% had a serious neurologic condition or inhospital death within 30 days of discharge. The frequency of serious event was highest between days 1 and 3 after discharge, with a gradual decline thereafter.
For the control group (discharged with non-neurologic complaint), the incidence of outcomes at 30 days for the headache related missed neurologic emergencies was between 0.12% and 0.18% for the non-neurologic complaints.
For back pain related missed neurologic emergencies, the incidence at 30 days was between 0.08% and 0.11% for the non-neurologic complaints.
The most common missed diagnosis for headache was ischemic stroke followed by SAH, and the most common missed diagnosis for back pain was intraspinal abscess (which includes epidural abscess) followed by cauda equina. Advanced age, male sex, non-Hispanic white, several comorbidities (eg, vasculitis, hypertension, neurologic disorders, HIV/AIDS, malignancy), and diagnosis of nonmigraine headache at the index ED discharge were
associated with a significantly higher incidence of outcome across the follow-up periods.
This study showed that we do pretty well in ruling out the life threatening causes of headache and back pain, with miss rates lower than 1% and not much higher than if the patient presented with a non-neurologic complaint. The authors do suggest that atypical presentations can be prone to cognitive biases and that reflecting on our own thought processes (metacognition) can help us to avoid those. They also make the valid point of remembering the limitations of our imaging tests since SAH and ischemic stroke were the most common misses for headache. To me the largest limitations of this study were the assumptions that had to be made for data gathering, that (1) the same process that led to the adverse outcome was present at the initial visit and (2) that the same process was misdiagnosed at the initial visit, which may or may not have been true. Also patients could have returned for a totally separate reason within 30 days and had a bad outcome. Overall though, this was a confirmatory study that we have a pretty low miss rate for dangerous presentations of headache and back pain with a reminder to be aware of our own biases and the limits of our diagnostic tests. And of course, always give good return precautions!
Dubosh N, Edlow J, Goto T, Camargo C, Hasegawa K. Missed serious neurologic conditions in emergency department patients discharged with nonspecific diagnoses of headache or back pain. Annals of Emergency Medicine 2019:1-13.
The Article: The Ottawa Subarachnoid Hemorrhage Clinical Decision Rule for Classifying Emergency Department Headache Patients, AJEM 2019
The Idea: This was another study created to evaluate the Ottawa SAH (OSAH) clinical decision rule which was derived from an initial 2010 Perry paper. Since the derivation, this rule has been evaluated with very high sensitivity in several papers (2013 JAMA, 2015 AJEM).
Single-center Taiwanese retrospective cohort study from Jan 2016-March 2017
Looked at patients with the primary complaint of an atraumatic headache
2 objectives: assess the OSAH rule’s performance and to evaluate the impact of the rule for ICH and other intracranial pathology
– 913 patient included in the study (>17 years, no trauma, HA less than 14 days, no neurologic deficit)
– Incidence of SAH was 1.6%
– OSAH rule had a 100% sensitivity for SAH (95% CI, 78.2-100), sensitivity also100% (84.6-100) for all ICH
– Specificity similar to prior at ~37%
– 33.1% of all patients in the group got a head CT. If the OSAH was validated, it would raise the imaging rate of these patients to 63.6%
It is good to have another study in support of a clinical decision rule that MAY help us safely decide to avoid CT scanning. A limitation of the study is a lower incidence of SAH (1.6% in this cohort versus the 5.8% reported in the Perry validation study), lacking the power for a narrow CI (Perry study reported 100% sensitivity with a 95% CI between 94.5-100%). This cohort was also specific to a single center in Taiwan which may have contributed to the low incidence. Lastly, the paper discussed that the rate of ordering CT scans would increase should the OSAH rule be applied to all patients in this cohort which is concerning for a rule that should be helping us to safely avoid CT scans. I still believe this rule has a place in clinical practice, however, and this study is yet another data point to add towards its validation.
The Article: Cerebral ischemia and deterioration with lower blood pressure target in intracerebral hemorrhage by Andrew B. Buletko, MD, et al. Neurology, 2018.
The Takeaway: Lower systemic blood pressures in patients with acute intracerebral hemorrhage (ICH) were associated with worse ischemia on MRI and neurological deterioration, characterized by a worsening of NIHSS score during their hospital stay.
Methods: 286 patients with acute intracerebral hemorrhage were included in the study. The study enrollment period straddled the date the institution changed their ICH blood pressure target from <160 mmHg to <140 mmHg. Primary outcomes were acute ischemia on MRI within 2 weeks of presentation, and acute neurological deterioration (NIHSS score change of >4 during hospital stay). There were no significant differences in characteristics of the study groups: they were largely older (>65), with risk factors we associated with stroke- hypertension, history of smoking, hyperlipidemia, and diabetes.
Results: In terms of primary outcomes, the group with a lower SBP target (<140 mmHg) did significantly worse. There was more ischemia seen on DWI-weighted MRI (32% vs 16%, p=0.047), and more neurological deterioration (19% vs 5%, p=0.022). The group with lower SBP also spent 1 day longer in the ICU and in the hospital. Univariate analysis comparing groups that showed MRI-confirmed ischemia vs. no ischemia elucidated similar results. The ischemic group had significantly lower SBP over their first 24 hours of admission (132 +/- 7.5 vs. 141 +/- 11.6, p<0.001) and were more likely to have neurological deterioration (43% vs. 2%, p<0.001).
Thoughts: This institution changed their blood pressure targets in response to the INTERACT-2 trial, which showed blood pressures <140 mmHg did not have a reduced risk of death or disability, but did have significantly better modified Rankin scores, suggesting better functional outcomes. However, critics of the trial have pointed out some significant outcomes that were just statistically so.
This study was small, retrospective, and a single center. However, it was able to compare similar study populations before and after the intervention of a protocol change. This intervention showed worse ischemia and clinical deterioration in patients with lower SBP.
Buletko, Andrew B., et al. “Cerebral Ischemia and Deterioration with Lower Blood Pressure Target in Intracerebral Hemorrhage.” Neurology, vol. 91, no. 11, 2018, doi:10.1212/wnl.0000000000006156.
Hill, Michael D., and Keith W. Muir. “Interact-2.” Stroke, vol. 44, no. 10, 2013, pp. 2951–2952., doi:10.1161/strokeaha.113.002790.
A Comparison of Rate Control and Rhythm Control in Patients with Atrial Fibrillation
– The Atrial Fibrillation Follow-Up Investigation of Rhythm Management (AFFIRM) Investigators
Published in the New England Journal of Medicine on December 5 2002
Atrial fibrillation is the most common sustained cardiac arrythmia. Most symptoms during atrial fibrillation are caused by poorly controlled or irregular ventricular rate. Two different treatment approaches:
1. Rhythm control: use of antiarrythmic drugs (amiodarone, sotalol, flecainide, etc) in order to convert to sinus rhythm. Theoretical benefits include less symptoms, better exercise tolerance, decreased stroke risk, potential to come off long-term anticoagulation and a better quality of life.
2. Rate control: use of AV nodal blockage (beta blockers, calcium channel blockers diltiazem or verapamil, digoxin) to control the ventricular response rate.
This was a randomized multi-center comparison study conducted out of the University of Washington but included 213 clinical sites. This was an intention to treat study. They recruited atrial fibrillation patients who were either over age 65 or had other risk factors for stroke or death. They also needed clinical judgement to decide if they were likely to be in recurrent a fib, that their a fib was likely to cause death, that they needed long-term anticoagulation and that they didnt have any contraindications to anti-coagulation. Once randomized, the physician could choose which drug to use and they could use multiple agents as needed. Only after failing two trials were they allowed to consider non-pharmalogical treatment (ablation). Everyone was put on warfarin with an INR goal 2-3.
– Rhythm control: 2/3 chose amiodarone, sotalol was the next most popular. Electrical cardioversion is an option.
– Rate control: 1/2 chose beta-blockers, diltiazem was the next most popular. Goal HR
This study looked to see how both the ROCHESTER and Philadelphia Criteria do at catching INVASIVE bacterial infections. Most of the risk stratifications scores in patient between 28-60 days look for serious infections which include UTI amongst others, but none actually specify for bacteremia or meningitis. This study uses a case-control method to retroactively look at its ability to predict low-risk infants using these scores to detect invasive infections.
The charts were selected based on certain criteria: + blood cultures, + LP, or pleocytosis after Abx and meningitis. These patients could only be ED arrivals, no transfers, and no documentation that they were “Ill appearing”. These patients were then matched to two similar febrile infants with negative blood and CSF cultures.
The results showed that both had concerning misses for potential invasive bacterial infections. Rochester marked 25 patients with IBI as low risk, while the philadelphia criteria marked 11. That said, Rochester missed 2 cases of confirmed meningitis (by their definition) and Philadelphia did not miss any. Therefore, the paper suggests the Philadelphia Criteria can be used to safely rule out invasive bacterial infections without the use of an LP. They acknowledge that a potential to miss a blood stream infection exists; however, the LP would not add to the work up in these patients that are missed.
Pediatrics is a hard population to study, and even harder to study sick kids due to such a low prevalence for disease. Regardless, this study did a strong job attempting to justify the no LP argument in the 28-59 day old infant with a fever. As long as you have a long discussion and ensure proper contact information and follow up with the patient and their caregivers, one could argue that this paper provides some good evidence to discharge a patient who fits this criteria.
Aronson PL, et al. Febrile Young Infant Research Collaborative. Risk Stratification of Febrile Infants ≤60 Days Old Without Routine Lumbar Puncture. Pediatrics. 2018 Dec;142(6)
The Article: Probiotics to Prevent Clostridium difficile Infection in Patients Receiving Antibiotics
Joshua Z Goldenberg, ND; Dominik Mertz, MD; Bradley C. Johnston, PhD
A JAMA Clinical Evidence Synopsis summary of Cochrane review article of 31 articles
The Clinical Question:
In adults and children prescribed antibiotics, is co-administration of a probiotic associated with a lower risk of symptomatic C. diff infection without an increase in adverse events?
Clinical trials sited: 39 (33 adults, 6 peds)
Study years: 1988-2012
Number of patients: 9955
Setting: inpatient, outpatient, and combined
Primary outcomes: C. diff infection and adverse events
Secondary outcomes: C. diff detection and antibiotic associated diarrhea
Probiotics lowered risk of C. diff for all comers from 4% to 1.5% with a NNT of 40 to prevent one C. diff infection. NNT did not change significantly when looking at treatment setting, patient age, or probiotic species/dose.
NNT for cramping and nausea: 37
NNT for diarrhea: 17
When co-administered with antibiotics, probiotics are associated with a lower risk of C. diff infection and other common GI side effects. They are considered a cheap and low risk option that should be considered whenever we prescribe antibiotics.
Goldenberg JZ, Mertz D, Johnston BC. Probiotics to Prevent Clostridium difficile Infection in Patients Receiving Antibiotics. JAMA. 2018;320(5):499–500. doi:10.1001/jama.2018.9064
Sheen et al. Performance of Low-Dose Perfusion Scintigraphy and CT Pulmonary Angiography for Pulmonary Embolism in Pregnancy. CHEST; 2018
Pulmonary embolism is a low incidence, high morbidity diagnosis with a relatively non-specific presentation in pregnant women. In the general population, the d-dimer remains the screening laboratory value for pulmonary embolism, but this test loses its value in pregnancy. A retrospective cohort study at the Montefiore system asked if low-dose perfusion scanning (LDQ) could replace CT pulmonary angiography (CTPA) in excluding pulmonary embolism. The primary outcome was negative predictive outcome of imaging for PE diagnosis at 90 days.
The study was conducted over 5 years at multiple sites in the Motefiore system. Women with a negative CXR underwent imaging with LDQ. 2 sites in the system with L&D used CTPA as their imaging modality of choice, and the women in this cohort tended to be further along in their pregnancy, which was a notable difference between groups. 2/81 women in the CTPA negative group had a 90 day event, while 0 women in the LDQ group had a negative event. It’s important to note that these two events were after delivery, which is its own risk factor for VTE. The primary endpoint showed equivalent NPV of 100% for the LDQ group and 97.5% for the CTPA group. At face value, this suggests that low dose perfusion scans could be utilized in place of CTPA for ruling out PE.
However, all imaging studies (VQ, LDQ and CTPA) deliver safe in utero fetal radiation doses, and nuclear medicine is rarely available 24 hours per day. The study authors recommend temporary anticoagulation until nuclear medicine studies can be obtained, which imparts a significant risk to both the patient and fetus that outweighs the risk of radiation that a CTPA creates. The lack of availability of nuclear medicine makes this study very difficult to generalize to many emergency departments.
This study was very underpowered, yet findings were consistent with Cochrane review literature. The authors do not recommend utilizing a positive LDQ as confirmatory for CTPA, which would require additional imaging prior to treatment. However, if your department has access to nuclear medicine, this study found non-inferiority in LDQ scans for ruling out PE, in a study that provides less radiation to maternal tissue. This imaging study can be utilized as an alternative to CTPA in the right clinical setting, but is unlikely to be practice changing in most departments.
Are Patients Receiving the Combination of Vancomycin and Piperacillin-Tazobactam at Higher Risk for Acute Renal Injury?
Luther MK, Timbrook TT, Caffrey AR, et al. Vancomycin plus piperacillin-tazobactam and acute kidney injury in adults: a systematic review and metaanalysis. Crit Care Med. 2018;46:12-20
Vanc and zosyn in combination is a common regimen for broad-spectrum antimicrobial therapy, especially in critically ill patients. Studies have shown a 6.5-fold increase in mortality among hospitalized and critically ill patients with AKI. Many medications, including vancomycin have been associated with AKI. Rates of AKI increase wit illness severity, comorbities and concomitant medication use.
2 authors independently searched PubMed, EMBASE, Web of Science, and Cochrane databases and manually searched reference lists of included studies and abstracts for other relevant studies
authors included RCTs and observational studies evaluating nephrotoxicity or AKI in patients aged 18 years or older treated with the combination of vancomycin and pip-tazo and either vanc or pip-tazo monotherapy, or vanc plus another beta-lactam
included studies defined AKI according to serum Cr level, urine output, or need for dialysis or renal replacement therapy
the authors calculated the percentage of patients with AKI after antibiotic treatment and time to AKI
calculated pooled odds ratios and mean differences in tiem to AKI, using random-effects models
they performed a secondary analysis for critically ill patients in the ICU
included 15 published studies and 17 abstracts
24,799 patients total
6 studies compared vanc and pip-tazo with vanc monotherapy
8 compared vanc and pip-tazo with vanc plus another beta-lactam
4 compared vanc and pip-tazo with pi-tazo monotherapy
3 made multiple comparisons
AKI occurred in 2% of patients treated with vanc and zosyn, compared with 12.9% in patients with other antibiotic regimens
Vanc and zosyn increased the odds of AKI compared with each regimen
Vanc and zosyn had a number needed to harm of 11 for AKI
In subgroup analysis of critically ill patients, vanc and zosyn vs. vanc monotherapy had an odds ratio of 9.62 for AKI
difference in time to AKI was not statistically significant between antibiotic regimens
antibiotic administration, comorbidities, infection type, and conocomitant infections varied across included studies, only some studies attempted to match risk factors to control for potential confounders
also included conference abstracts, which may have poor quality, lack of peer review, and insufficient information
time to AKI was not available for ICU patients
what is the clinical significance?
There is an increased risk of AKI with the combination of vanc and zosyn, compared with vanc or zosyn monotherapy, or vanc combined with another beta-lactam, but it is unclear how clinically significant this is.
Bottom Line: When cervicitis or pelvic inflammatory disease (PID) is suspected in young, healthy women, a pelvic exam does not change sensitivity or specificity when compared to history alone.
The Article: The Additive Value of Pelvic Examinations to History in Predicting Sexually Transmitted Infections for Young Female Patients With Suspected Cervicitis or Pelvic Inflammatory Disease.Ann Emerg Med. December 2018.
Gonorrhea and Chlamydia are very common in cervicitis and pelvic inflammatory disease in women ages 15-24 years old. However, there is no specific history, physical exam, or laboratory test to confirm this infection; instead, a combination of all of these help with the diagnosis. Typically, the diagnosis of cervicitis is visualizing an inflamed cervix or discharge in the vaginal canal or at the cervix. The diagnosis of PID typically presents with lower abdominal pain without another cause and an abnormal bimanual exam.
Recently, however, the American College of Gynecology, the American Academy of Pediatrics, and the CDC, reported that gonorrhea and chlamydia laboratory tests are good enough for diagnosis in an asymptomatic woman and that a pelvic exam is unnecessary. This recommendation is for asymptomatic women and has not been tested in women with symptoms who normally present to the Emergency Department.
This study was a prospective study that aims to determine if pelvic examinations are useful in the diagnosis of cervicitis or PID. This study took place from 10/2015-10/2017 in an urban academic pediatric emergency department using a convenient sample (only when researchers were there during the day to enroll patients).
Participants were females, aged 14-20 years of age with abdominal pain or discharge at all pregnancy statuses. Exclusion criteria were anyone unstable (HR >110, SBP <90), requiring critical care, refused, no prior pelvic examination, or non-English speaking.
An attending, resident, fellow, or advanced practice provider first took a history and marked a 1st VAS score (visual analog scale). Then the provider did a pelvic exam and indicated a 2nd VAS score with information collected from a visual, speculum, and bimanual exam. This was then compared to a urine gonorrhea and chlamydia test.
Of the 848 laboratory tests ordered, 288 of those were enrolled in the study.
79 out of 288 had a positive urine, however, 123 patients were treated for cervicitis, 35 for PID, and 93 for vaginitis
166 out of 288 had a low 2nd VAS score, however, 36 patients were still treated
after pelvic examination, management changed in 71 cases, of these 35 correlated with urine tests and 36 did not correlate
There were no significant differences between the 2 groups of patients (those with + infection and those with – infection) in terms of vaginal discharge, cervical motion tenderness, uterine tenderness, or adnexal tenderness.
After the pelvic exam, the 2nd VAS score did not show a decreasing trend in STD negative cases or an increasing trend in STD positive cases. There was also no statistical difference among provider groups.
The Takeaway: Pelvic exams may not be helpful in the diagnosis of cervicitis or PID in young, healthy women.